ML20247L080
| ML20247L080 | |
| Person / Time | |
|---|---|
| Issue date: | 05/31/1989 |
| From: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| References | |
| NUREG-0975, NUREG-0975-V07, NUREG-975, NUREG-975-V7, NUDOCS 8906020086 | |
| Download: ML20247L080 (364) | |
Text
{{#Wiki_filter:. t l NUREG-0975 1 Vol. 7 1 Compilation of ! Contract Research for the ; Materials Engineering Branch, Division of Engineering Annual Report for FY 1988 : i U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research fr "%,, a "d [ sgg60 g g 89053.t.' 0975 R PDR t . . . . -
I l AVAILABILITY NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in NRC publications will be available from one of the following sources: l 1. The NRC Public Document Room, 2120 L Street, NW, Lower Level, Washington, DC 20555
- 2. The Superintendent of Documents, U.S. Government Printing Office, P.O. Box 37082, Washington, DC 20013-7082
- 3. The National Technical Information Service, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited in NRC publica-tions, it is not intended to be exhaustive.
Referenced documents available for inspection and copying for a fee from the NRC Public Document Room include NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement bul!stins, circulars, information notices, inspection and investi- , gation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspondence, i The following documents in the NUREG series are available for purchase from the GPO Sales l Program; formal NRC staff and contractor reports, NRC-sponsored conference proceed- ! ings, and NRC booklets and brochures. Also available are Regulatory Guides, NRC regula-tions in the Code of Federal Regulations, and Nuclear Regulatory Commission issuances. Documents available from the National Technical Information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission. Documents available from public and special technical libraries include all open literature items, such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries. Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the publication cited. Single copies of NRC draft reports are available free, to the extent of supply, upon written request to the Office of Information Resources Management, Distribution Section, U.S. Nuclear Regulatory Commission, Washington, DC 20555. Copies of industry codes and standards used in a substantive manner in the NnC regulatory process are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually copy-righted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards Institute,1430 Broadway, New York, NY 10018. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ i
NUREG-0975 l Vol. 7 R5 l Compilation of , Contract Research for the Materials Engineering Branch, Division of Engineering
' Annual Report for FY 1988 Aanuscript Completed: March 1989 Date Published: May 1989 Materials Engineering Branch Division of Engineering Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Cornmission Washington, D.C. 20555 f "e og
=.i 4 I
3 h.. 1
. COMPILATION OF CONTRACT RESEARCH l
FOR-THE l
' MATERIALS ENGINEERING BRANCH l DIVISION'0F ENGINEERING :l U.S. NUCLEAR REGULATORY. COMMISSION
- Introduction to Annual Report This compilation of annual reports by contractors to the Materials Engineering-Branch of the NRC.0ffice of Research concentrates on achievements in' safety research for the primary system of commercial light water' power reactors,
-particularly with regard to reactor. vessels, primary system piping, steam generators, nondestructive examination of primary components, and in safety research for decommissioning and decontamination, en-site storage and
- engineered safety features. Annual reports for FY-1988, from each of the branch contractors, will be quite extensive, containing many details of test-
- results, conclusions and recommendations. These reports will be published early in .FY-1989. Because they willinot.all come out at the same time, it will be difficult to assess the. total impact and value of the overall branch program by simply trying. to follow these' reports as they are published. Thus the Materials Engineering Branch assembles abbreviated reports from all the branch contractors and publishes them in a single _ annual report as soon after the end of.the year'as possible so.that the information developed throughout the year can be promptly used in the safety-regulatory process. This report, covering research conducted during-Fiscal Year 1988 is the seventh volume of the series of NUREG-0975, " Compilation of Contractor Research for the Materials Engineering Branch, Division of Engineering."
1 I iii 1
p 3 C. ! La 1 1 i E l
. TABLE OF CONTENTS 1
i FY 1988 ANNUAL REPORT OF CONTRACT'RESEARCH .' fiATERIALS ENGINEERING BRANCH R l Page Vessel and Piping Fracture Mechanics ORNL: Peavy-Section_ Steel Technology Program 1-60 MEA: Structural Integrity of_ Light' Water Reactor 61-90. Pressure Boundary Components BCL: Degraded Piping Program - Phase-II 91-113 DTRC: Elastic-Plastic Fracture MechanicsLEvaluation- 114-133 of LWR Alloys Pressure Vessel Surveillance Dosimetry ORNL: Surveillance Data Bases, Analysis, and 134-151 Standardization Program NBS: - Dosimetry. Measurement Reference Data Base for 152-169 LWR Pressure Vessel Irradiation Surveillance Steam Generators, Aging and Environmental Cracking Battelle-PNL: Steam Generator Integrity Program 170-178 ANL: Environmentally Assisted Cracking in Light 179-200 Water Reactors AllL: Long-Term Aging Embrittlement of Cast Duplex 201-216 Stainless Steels in LWR Systems ANL Aging Studies on Materials from the 217-232 Shippingport Reactor
) ,, v ;
1 2
1 I l l i i
.P.aae~-
Non-Destructive Examination ! Battelle-PNL: Evaluation and Improvement of NDE Reliability 233-256 for Inservice Inspection of Light Water Reactors PNL: Final Developments. Validation and Technology 257-266 Transfer for AE and SAFT-UT ANL: Final Evaluation of Advanced and Current 267-278 ! 1.ea k Detection Systers ORNL: Eddy-Current Inspection for Steam Generator Tubing 279-285 PNL: Review of the Status of Nondestructive Measurement 286-291 Techniques to Quantify Material Property Degradation Due to Aging and Planning for Further Evaluation Decommissioning and Fuel Cycle: W - Hanford: Evaluation of Nuclear Facility Decommissioning 292-301 Projects (EHFDP) Program BNL: The Impact of LWR Decontamination on Solidification, 302-318 Waste Disposal and Associated Occupational Exposure PNL: ESF System Fission Product Retention Effectiveness 319-332 PNL: Radionuclides Source Term Measurements for 333-346 Decommissioning Assessments EG&G: Effectiveness and Safety Aspects of Selected 347-349 Decontamination Methods ORNL: Analysis of Spent Fuel Heat Production 350-357 vi
I 1 l- , l llEAVY-SECTION STEEL TECHNOLOGY PROCRAM l h Dak Ridge National' Laboratory j Oak Ridge, Tennessee 37831 j l PROGRAM PARTICIPANTS The lleavy-Section Steel Technology (HSST) Program is administra-tively carried out through the Pressure Vessel Technology Section of the Engineering Technology Division of Oak Ridge National Laboratory (ORNL) with key technical participants from the Engineering Technology, Metals and Ceramics, Computer and Telecommunications, Chemical Technology, and Research Reactors Divisions. In FY 1988 subcontracts were in place with 4 three universities, two industrial organizations, and one laboratory within a government agency to complement the ORNL activities. The prin-cipal investigators are shown below. l HEAVY-SECTION STEEL TECHNOLOGY PROGRAM W.R. CORWIN, MANAGER ENGINEERING TECHNOLOGY COMPUTING AND TELE- CONSULTANTS DIVISION COMMUNICATIONS DIVISION H.A. Ernst S.E. Bolt C.8, Oland D.G. Ball LB. Freund R.H. Bryan J.S. Parrott G.T. Hchn B.R. Bass J.W. Bryson W.E. Pennell J.W. Hutchinson T.L Dickson ) R.D. Cheverton C.E. Pugh C.R. Loper. Jr. ' D.S. Steinert J.G. Merkle G.C. Robinson E.T. Wessel J.A. Keeney-Walker D.J. Naus K.R. Thoms G.D. Whitman METALS AND CERAMICS R & D SUBCONTRACTORS Battelle Columbus Div. D.J. Alexander R.W.McClung NBS, Gaithersburg K.V. Cook R.K. Nanstad Ohio State Univ. F.M. Haggag R.W. Swindeman Southwest Research Inst. S.K.1skander Univ. of Maryland Univ.of Tennessee l I s
. _ _ _ _ _ _ _ _ _ _ . . _ _ _ _ . _ l
i ABSTRACT f The Heavy-Section Steel Technology (HSST) Program is conduc-ted for the Nuclear Regulatory Commission (NRC). The studies relate to all areas of the technology of materials fabricated into thick-section primary-coolant containment systems of light-water- 7 cooled nuclear power reactors. The focus is on the behavior and structural integrity of steel pressure vessels containing crack-like flaws. The program is organized into 12 tasks: (1) program , management, (2) fracture methodology and analysis, (3) material F characterization and properties, (4) special technical assistance, (5) crack-arrest technology, (6) irradiation effects studies, (7) y cladding evaluations, (8) intermediate vessel tests and analysis, (9) thermal-shock technology, (10) pressurized-thermal-shock tech-nology (11) Pressure Vessel Research Users' Facility (PVRUF), and "l (12) shipping cask material evaluations. During this period, advances were made in the coordinated effort to develop the dynamic materials fracture data and the analytical tools required to construct improved fracture models for reactor pressure vessel (RPV) steels. Analytical efforts included examination of alterna- eN tive parameters governing dynamic fracture, their corresponding constitutive models and computational implementation as well as constraint and tunneling effects on crack arrest. Additional con-stitutive models and fracture parameters were incorporated into the ORNL a,ystem of fracture analysis programs, ORMCEN/ADINA/ [ ORVIRT, and alternative methods were investigated to overcome the g challenge of mesh convergence. Subcontracts at University of M Maryland, Ohio State University, and Southwest Research Institute j yielded experimental data on dynamic fracture and crack arrest, 4 high-rate stress-strain behavior, and the micromechanical aspects of brittle and ductile fracture, as well as furthering the o development of elastodynamic and viscoelastic analysis tech-niques. Final preparations for initial testing of a moderate- j] sized specimen capable of measuring high values of crack-arrest ' toughness were performed. Fracture properties were developed for a low-upper-shelf steel and commercial weld overlay cladding was examined further. Work on material and water chemistry effects on j fatigue crack propagation was published. Three new areas of NRC special technical assistance were begun: (1) the evaluation of j possible enhanced low-temperature, low-flux irradiation embrittle- ]' ment of reactor pressure vessel supports; (2) an assessment of boiling-water-reactor vessel integrity, and; (3) an overall j assessment of low-upper-shelf welds in reactor pressure vessels a with special emphasis on the reevaluation of the J-integral in assessing large amounts of crack extension. The last four wide-plate single-edge notch specimens were tested at the National y Bureau of Standards to investigate additional high-upper-shelf j materials and the effects of plate thickness in low-upper-shelf ] material under conditions leading to crack arrest near or above j the knee of the Charpy upper shelf. All testing for the Fifth j HSST Irradiation Series on the shift in the K Ic curve of high j 1 2
1 copper weldments was completed, and a draft report on their l results prepared. All irradiated fracture toughness testing was i completed for the second phase of the Seventh HSST Irradiation j Series on cladding. Plans were formulated and initiated to obtain l
- , low-upper-shelf weld metal from Unit 1 of the cancelled Midland i
[ Reactor for use in the Tenth HSST 1rradiation Series. A report i ! was drafted describing the nondestructive examinations on segments- ) of clad light-water reactor vessels. A report describing the I second series of the clad plate fracture tests of reactor vessel I steels was drafted. The experimental and analytical results from j the second pressurized-thermal-shock test, PTSE-2, on a low upper-shelf material, were published and demonstrated that low-upper-shelf material can exhibit both high crack-arrest toughness and a limited ~ benefit from warm prestressing. Initial studies on the benefits and feasibility of additional pressurized-thermal-shock , experiments were completed, and recommendations were made to con- ! duct PTSE-3 to confirm analytical predictions of the effects of cladding and PTSE-4 to clarify mechanisms of tearing in low-upper-shelf welds. The modifications of the temporary facility and detailed planning for nondestructive flaw assessment of the pres-
-surized-water reactor pressure vessel in the ORNL Pressure Vessel Research Users' Facility were completed, and planning for its overall research thrusts was performed.
OBJECTIVE The Heavy-Section Steel Technology Program is carried out to advance the understanding and validation of materials and structures behavior as they relate to light-water reactor (LWR) pressure-vessel integrity. The program had its beginning in the mid-1960's and has contributed to verifying the applicability of fracture mechanics to ves-sei integrity assessments, to advancing associated analysis tools, to data generation and correlations development, and to code criteria and rule development. The studies address the determination of the effects of flaws, variations in properties, irradiation, stress raisers, and ! residual stresses on the integrity of vessels under combined thermal and mechanical loadings. FY 1988 SCOPE In FY 1988 the HSST program was composed of a management function , and eleven technical tasks, which addressed fracture methodology and I analysis, ma: rial characterization and properties, special technical assistance, crack-arrest technology, irradiation-effects studies, clad-ding evaluations, intermediate vessel tests and analyses, thermal-shock investigations, pressurized-thermal-shock technology, a pressure vessel research users' facility, and shipping cask material evaluations. 3
Results from the program are integrated to aid in resolving major multi-disciplinary regulatory issues facing the NRC in the area of pres-sure-vessel integrity. For example, pressurized-thermal shock (PTS) is being addressed by examining crack initiation (including effects of geometry, stress state, history (e.g., warm prestressing), and irradia-tion); dynamic crack propagation (including arrest, mode conversion, stability, and rate effects); cladding effects on small-surface flaws as well as sub-surface or heavily-tunneled flaws; development, verifica-tion, and computational implementation of methods of analysis for vessel behavior; irradiation trend curve development, improvement, and correla- I tions; and post-service material evaluations. The specialized problems associated with low-upper-shelf (LUS) welds and their low resistance to ductile-crack extension, including implications to the PTS issue, are I being assessed in a state-of-the-art " white paper" and further examined by: establishing quasi-static and dynamic crack propagation and arrest behavior; validating and extending analytical procedures established for high-upper-shelf materials, and assessing the effects of irradiation on the ductile and brittle initiation, arrest, and annealing response behavior of LUS welds. Irradiation effects related to the degree of conservatism in both operating pressure-temperature limits and the low-temperature over pressurization set points for LWRs are being evaluated through irradiation experiments examining the current method of accounting for irradiation-induced shifts of initiation toughness for both high- and low-upper-shelf materials, and the degree of recovery and rate of re-embrittlement of fracture toughness associated with annealing. Other topical issues being addressed include the effects which higher-than expected embrittlement could have on pressure vessel support integrity, the effects of inservice inspection, or the lack thereof, on the failure probability of boiling-water reactor (BWR) pressure vessels, and the quantification of environmentally-assisted fatigue crack growth. In addition to addressing specific topical issues, the majority of the results from the HSST program provide information which is vital to the assessment of the continued safe operation of pressure vessels in aging LWRs. The results from the irradiation, annealing, cladding, ves-sel support, flaw density, pressurized-thermal shock, and low-upper-shelf studies as well as the development and verification of analytical methods all play key roles. Moreover, these results provide guidance and bases to evaluate applications to extend plant operating licenses. The data from the program are used in establishing and revising regulatory guides, codes and standards, and federal regulations. Indeed many of the current regulatory positions on integrity of pressure vessels and other components requiring the assessment of their fracture resistance are based on or buttressed by results derived from the HSST program. I 4
SUMMARY
OF RESEARCH PROCRESS
- 1. PROCRAM MANAGEMENT (W. R. Corwin)
In addition to administering the program, coordination was main-
)
tained with other ongoing programs and with code- and standard-writing l groups. An updated five year program plan was drafted and used as the reference document for monthly management reporting throughout FY 1988. During the year two progress reports,1,2 eight topical reports,3 10 three foreign trip reports,il 13 and twenty-seven technical papersl4 40 were issued. In addition forty-seven technical presentations 41 87 were made. i
- 2. FRACTURE MECHANICS AND ANALYSIS [B. R. Bass, S.-J. Chang, K. Hornberger, J. A. Keeney-Walker, J. C. Merkle, J. S. Parrott, and J. C. Thesken (ORNL), W. L. Fourney, C. R. Irwin, and C. W. Schwartz (Univ. ofMd)*,R.J.Dexteg,M.F.Kanninen,andP.E.O'Donoghue (SwRI)t, and A. Gilat (OSU) ] l 2.1 Introduction The USST program is continuing to improve the understanding of con-ditions that govern the initiation, rapid propagation, arrest, and duc-tile tearing of cracks in RPV steels. In PTS scenarios, inner surface cracks in an RPV have the greatest propensity to propagate because they are located in the region of highest thermal stress, lowest temperature and greatest irradiation damage. If such a crack begins to propagate radially through the vessel wall, it will extend into a region of higher fracture toughness due to the higher temperatures and less irradiation damage. Because crack initiation is a credible event in a PTS trans-ient, assessment of vessel integrity requires the ability to predict all phases of a fracture event. These phases include crack initiation, non-isothermal propagation, arrest, stable or unstable ductile tearing, and structural instability. Through the integrated efforts of several laboratory and university research groups, progress was made in develop-ing various components of the technology required to treat these phases of a fracture event. The technology includes fracture models, analysis methods, criteria and data curves and is being developed and validated )
through small and large specimen experiments.
- Work sponsored by HSST program under subcontract between Martin Marietta Energy Systems, Inc. and the University of Maryland.
Work sponsored by USST program under subcontract between Martin Marietta Energy Systems, Inc. and the Southwest Research Institute.
#Work sponsored by USST program under subcontract between Martin Marietta Energy Systems, Inc. and Ohio State University.
l l 5
4 l l The effect of viscoelasticity in dynamic fracture analysis is a key I component of the technology under development. Material properties characterization testing has been performed on A533 grade B class 1 steel, and these data were used to derive material constants for selec-ted viscoelastic models. The constitutive models, along with crack pro-pagation techniques and several proposed nonlinear f racture parameters, j have been installed in HSST-developed finite-element computer programs. The capabilities of these nonlinear techniques are being com-pared and evaluated, in part, through applications to the small- and large-specimen crack run-arrest experiments. j The following sections describe recent advances made by ORNL and subcontracting groups to develop the crack-arrest data base and the i analytical tools required to construct improved fracture models for RPV steels. 2.2 Fracture Toughness Determination and Strength Methods , 2.2.1 Stub-Panel Crack-Arrest Specimen ) Studies 88 have been conducted at ORNL to evaluate the usefulness of a relatively small panel specimen (whose size is between conventional crack-arrest specimens and the wide plate specimens) for crack-arrest and dynamic-fracture testing. A side grooved stub panel specimen with dimensions 45.1 x 99.1 x 3.39 cm was selected for these studies. A gra-dient in fracture toughness is achieved by cooling the stub region and heating the panel edge to produce a nonuniform steady-state temperature distribution across the plate. A tensile load is applied to the panel and the stub is independently loaded to provide Ky levels that are high enough for initiation of the chilled crack in cleavage. Arrest of the fast-running crack then occurs in the ductile high-temperature region of the panel. Static and dynamic analyses as carried out for the stub panel configuration confirmed its utility for producing high Ky data (>200 MPa/m)inthetemperatureregimeofupper-shelfmaterial$ehavior. ORNL's 550-kip tensile machine and associated data acquisition equipment were readied for testing the stub panel specimens. The first specimen of A533 grade B class 1 steel will be instrumented with thermocouple, strain gages, and displacement gages, similar to those employed for the wide plate specimens;89 it will be tested early in FY 1989. The test procedures are built upon the experience gained this year at ORNL by obtaining dynamic fracture data from tests of about one dozen small (6 x 6x 1 in.) specimens of 4340 steel. 6
2.2.2 Small Dynamic-Fracture Specimens l Crack-arrest toughness data were developed through small-scale l specimen testing by several HSST subcontractors. Southwest Research Institute (SwRI)90 obtained dynamic crack propagation data for A533 , grade B class 1 steel using small-scale specimens supplied by ORNL. For j this purpose, duplex A533 grade B class 1/4340 steel specimens of effec- ) tive width, w = 127 mm, were instrumented and tested at 23*C. Results from dynamic analyses performed with the SwRI code VISCRK91 are reported for two of these duplex experiments (SD2 and SD6) in Ref. 90. At University of Maryland (UM), efforts have focused on development of two rapid-loading fracture experiments.92 The goal of these studies is to determine the limits of approximating the crack-arrest toughness, Ky ,, by the dynamic initiation-toughness, kid. In ne case, explosively-loaded notched short bars are tested and the kid values so obtained are compared with K y , values for the same material. In the other case, a method is being developed to measure " lower-bound" cleavage initiation toughness in the transition temperature range using a small specimen (a notched round bar subjected to impact loading). The loading of the notched short bar (400 mm in length) with its , integral dog-bone ends is accomplished with four explosive charges that ' are detonated simultaneously. Tensile stress waves propagate to the central region of the bar, where they combine to produce a rapidly increasing K7 field that initiates a stationary fatigue-sharpened crack. Initial experiments of this type were conducted with a brittle polyester and then later with a very brittle steel (4340 hardened to Rc = 51). The current series of experiments is examining A533 grade B class 1 steel; the one experiment completed thus far gave a fracture toughness value of kid = 77 MPa/m. The impact-loaded notched round-bar experiment was developed to simulate the effects of constraint of a very thick specimen. The cylindrical shape should increase the effective thickness of the specimen by a factor of ~3. The specimen configuration used in testing has an outside diameter of 38.1 mm and a machined notch diameter of 19.1 mm. These specimens are impact loaded by 1051 J of energy when the weight of 58.8 kg drops through a distance of 1.83 m. Five specimens of A508 chemistry steel were tested, and the determinations of kid were found to be cons _istent and repeatable, yielding an average value of Kid = 54 HPa/m. A similar evaluation of the lower-bound initiation toughness of A533 grade B class 1 steel is under way. Two specimens were tested near O'C, and five specimens were tested at room temperature. The test results were very consi_ stent and yielded average frac _ture toughness values of kid = 79.2 MPa/m at 25*C and kid = 65.8 MPa/m near 0*C. l 7
2.2.3 Cleavage-Fibrous Transition Behavior Topographic analyses of selected wide plate fracture surfaces are , being performed at UM using stereo-scanning electron microscopy and I relative-height measurement techniques. In Ref. 92, these techniques were applied to regions of fracture surfaces (from WP-1.7 and WP-CE-1) where cleavage arrest was followed by fibrous-tearing reinitiation. Estimates of crack-tip opening displacement obtained from relative- ! height measurements were used to calculate K7 values at reinitiation. These studies indicated that high Ky values (>350 MPa/m) were necessary to initiate fibrous tearing from an arrested cleavage crack in the i upper-transition temperature range. Optical stereo photographs were made of the initiation region on , the fracture surfaces of the WP-1.8 wide plate specimen. Topology mea- l surements showed that the initiation occurred at separated regions of ! the crack front at times which were almost simultaneous. The dominant ! event was adjacent to a segment of the serrated pre-crack front where l the appearance of the pre-crack leading edge was quite clear. No other features, which might pertain to the higher than expected initiation I load in the WP-1.8 test, were observed. i Optical stereo photographs were also made across a large number of the run-arrest events observed in wide-plate test WP-2.2. The purpose of the topology measurements was to see whether measureable abrupt j increases of crack opening could be observed following the arrest of l general cleavage. The results suggested that, across a substantial temperature range adjacent to loss-of-cleavage, the separation resis-tance of cleavage and fibrous regions was nearly the same. Details of the topology measurements for WP-1.8 and WP-2.2 are given in Ref. 93. 2.2.4 Dynamic Fracture Toughness Relation Predictions of crack propagation and arrest in dynamic fracture problems require specification of the relation among instantaneous crack-tip velocity, 4, dynamic fracture toughness, KThis ture, T, for the fracturing material. ts arelation primary.ID, input and tempera-for predictive application-mode dynamic fracture analyses. [In an app- i lication-mode analysis, the crack tip is propagated incrementally when the relation K a lied " EID (A, T) is satisfied.] At UM, Schwartz 94 has j used data f rom [ke WP-1 series of wide plate tests to estimate the A vs K vs T relation for A533 grade B class 1 steel. This toughness rela-ID tion has been used by ORNL for pretest analysis or intermediate-size stub panel specimens of this same steel. ; l 2.2.5 Constraint Effects on Crack-Tip Yielding The HSST wide plate test specimens are usually treated in finite-element dynamic fracture analyses as two-dimensional plane stress prob-lems. Although this approximation is reasonable over most of the l 8 i
l structure, it deteriorates close to the crack tip where constraint induced by the near-tip triaxial stress gradients causes deviations from plane stress yielding conditions. Plane stress viscoelastic models need to accommodate this observation in the very region where rate effects are expected to be most significant. Triaxial constraint effects and the transition from plane stress to plane strain conditions in the crack-tip region are being investigated at UM through a series of static nonlinear three-dimensional analyses of selected fracture geometries. Results to date indicate that plane strain yield conditions dominate only for values of a yield factor, Fy, satisfying Fy < 0.5, while plane stress conditions do.uinate for Fy > 3. (The yield factor, Fy, was defined previously by Hahn and Rosenfield95 to quantify the transition from plane strain to plane stress yielding as a function of specimen thickness and load level.) This conclusion has implications for viscoelastic dynamic fracture analyses of the 10-cm- and 15-cm-thick specimens in the WP-1 series of wide plate tests of A533 grade B class 1 steel. During initiation and most of the propagation phase, the yield factor approaches the small-scale plane strain yield domain (0.5 < Fy < 0.9). However, near arrest the yield factor increases well into the plane stress yield region (3.3 < Fy < 5.0). Thus, the assumption of j plane stress conditions in two-dimensional viscoelastic-dynamic fracture ' analyses seems appropriate, particularly near arrest when nonlinear , effects are most significant. l l 2.2.6 Crack Tunneling Effects Current studies at UM are designed to produce more refined models of the influences of crack tunneling on fracture toughness determina-tions in ductile materials. Crack tunneling is commonly observed during ' the fracture of tough and ductile materials. The loss of constraint near the free surface of the specimen permits yielded ligaments to extend for considerable distances behind the leading edge of the crack. A significant portion of the apparent fracture toughness mea- t sured for a deeply tunneled crack may be due to the restraining effects of these yielded ligaments. Such tunneling has been observed in several of the WP-1 series of wide plate tests, raising questions regarding the appropriateness of two-dimensional analytical models that implicitly ignore tunneling. Calculations by Popelar86 at Ohio State University (OSU) suggest that correcting for the restraining effects of tunneling may substantially reduce the fracture toughness values inferred from the WP-1 test data, bringing all of the test data down to, and in a few instances below, the ASME Section XI reference toughness curve. The restraining effects of tunneling have been analyzed at UM for l the 10-cm-thick WP-1 test specimens using techniques previously employed ' by Popelar97 and by Smith.98 Popelar97 considered both strip ligament and parabolic ligament geometries and assumed that the restraining stresses can be averaged across the thickness of the ligaments. Smith 98 had observed earlier that thickness averaging procedures (such as those 9
l l l l l employed by Popelarss) overestimate'the restraining effect of the liga-ments for small to moderate amounts of tunneling. Consequently, Smith's solution is based upon a correction evaluated from the integral over the ligament of the solution from Tada et al.99 for a point force applied to the crack face. Results from the UM tunneling analyses are summarized in Table 2.1. These analyses have been performed independently of earlier inter-pretations of the WP-1 tests by Popelar.96 Popelar96 predicts a large reduction in the computed arrest toughness, particularly for events WP-1.28 and WP-1.4A, where the tunneling corrections are approximately 30 and 60 percent, respectively, of the uncorrected arrest toughness val-ues. .The Smith approach predicts significantly smaller corrections for events WP-1.2B and WP-1.4A that are approximately 10 and 20 percent, respectively, of the uncorrected arrest roughnesses. For all cases except WP-1.6A, the corrected data based on Smith's' approach lie between the uncorrected values and the corrected values from Popelar's approach. None of the Smith-corrected data points lie below the ASME : reference curve. Table 2.1. Tunneling effects on computed crack-arrest toughness values for the WP-1 series of wide plate tests Popelar approach d Smith approache Test T-RT Ka gg o KC AK K ("b (MP!/m) (MP1/N) (MPE/m) (MPa/E) (MPE/m) WP.1.2A 85 440 88 352 15 425 WP-1.2B 115 523 139 384 33 470 WP-1.3 77 243 29 214 22 221 52 158 99 59 20 138 WP-1.4A 83 397 56 341 9 388 WP-1.4B i WP-1.5A 79 229 22 207 3 226 WP-1.5B 95 300 56 244 N/A N/A 77 285 52 233 58 227 WP-1.6A aK g = crack-arrest toughness from generation-mode. j elastodynamic analyses. b6K = ligament correction. ( l Y
= net crack-arrest toughness. f CK = K g
-K dMethod of analysis based on Ref. 97. l 4
" Method of analysis based on Ref. 98.
10 t I
2.2.7 Viscoelastic Material Model Characterization Research efforts at ORNL and several subcontracting groups are assessing the effect of viscoelastic material behavior on cleavage crack propagation and arrest in RPV steels. This includes the examination of strain-rate and temperature-dependent constitutive models for A533 grade B class 1 steel. These models will be used in dynamic fracture anal-yses. As part of this effort, dynamic stress-strain data have been gen-erated by SwRI,100 and SRI International,101 and OSU)o2 for A533 grade B class 1 steel for use in deriving constants for proposed constitutive models. Recently, Cilat io2 has conducted tests at strain rates of ~800 and 5000 s l and at temperatures of -150 to 20*C. (All specimens for these tests were taken from HSST Plate 13-A, which is the same source plate for the wide plate crack-arrest specimens.) Results for these tests show that both temperature and strain rate have a significant effect on the stress-strain response of A533 grade B class 1 steel. Kanninen et al.100 used their dynamic stress-strain data to derive constants for the Bodner-Partom constitutive model appropriate for A533 grade B class 1 steel at test temperatures ranging from -60 to 150*C. At ORNL, Chang 103 has employed stress-strain data from Refs. 100 and 101
)
aad an extended version of the Robinson 104 viscoelastic model. The ) extended Robinson model was originally developed to describe strain- ' aFi ng effects (including yield drop) for Inconel 617 at 950 C and the reverse strain-rate effects for type 304 stainless steel at 550*C. Recent studies at ORNL indicate that the Robinson model effectively models the yield drop and strain-rate sensitivity of the A533 grade B class 1 steel at 100*C. Numerical implementation of the extended Robinson model into HSST computer programs following a technique due to Hornberger105 is currently under way at ORNL. 2.2.8 Elastic-Plastic Fracture Studies in Inhomogeneous Materials Plans were formulated for the HSST program to participate in a multi year investigation of elastic plastic fracture models for inhomogeneous structures. A major product of this investigation will be the development of an clastic plastic fracture estimation scheme applicable to inhomogeneous structural components. The research will be performed as part of a cooperative effort between the HSST program and the Atomic Energy Research Committee of the Japan Welding Engineering 1 Society. The Research Committee is chaired by Professor C. Yagawa of i the Department of Nuclear Engineering, University of Tokyo, and includes researchers from over 25 industrial and university research organizations in Japan. Through a subcontract agreement with the HSST program, the Century Research Center (CRC) Corporation (with offices in Tokyo and Santa Clara, CA) will serve as liaison between the Research Committee in Japan and the HSST program. (The elastic plastic f racture study of inhomogeneous materials is identified as the EPI Program by the Research Committee.) The research effort in the EPI Program will consist of both experimental and analytical / computational studies 11
1 1 divided among several subtasks to be performed by the participating organizations. Interim reports on developments in the EPI Program will ; be compiled'and issued annually through completion of the program in FY 1992. 2.3 Numerical Methods and Computer Programs j I
.i 2.3.1 Viscoelastic-Dynamic Analysis Methods Developments l The predictive capabilities of the nonlinear-dynamic fracture anal-' )
ysis techniques are being evaluated through applications of the ADINA/VPF38'106 and VISCRK91 computer programs to analyses of HSST I crack-arrest experiments. Recently, viscoelastic-dynamic f racture anal-yses107 of the WP-1 series 89 of wide plate tests were conducted with the ADINA/VPF Program at ORNL using finite-element models having improved mesh refinement near the plane of crack propagation. Results from these i analyses were compared with those obtained from other models having different mesh refinements along the crack plane. These combined results indicate that the viscoelastic-dynamic solutions'of the wide-plate tests expressed in terms of the inelastic fracture parameters (for example, T* (Ref. 108).have not yet converged for the mesh refinements , employed thus far in these studies. i Insight into the difficulties associated with modeling rapid crack ! propagation events in RPV steels exhibiting viscoelastic behavior are provided by two recent studies at OSU. Sheul09 studied the mode I plane strain problem of dynamic steady-state crack growth in A533 grade B class 1 steel using the Bodner-Partom model characterized in Ref. 100 and the assumption of small-scale yielding. Sheu l09 resolved the near ; crack-tip singular fields using a finite-element model with element dimensions approximately 1/1000th of the elastic plastic zone size. Furthermore, for A533 grade B class 1 steel over a temperature range i from -60 to 100*C and a crack speed of one-half the Rayleigh wave speed, Popelar210 estimates that the zone-of-dominance of the near-tip fields extends from ~5 to 55 pm compared to an inelastic region with dimensions 0.1 to 15 mm. Even if the computational capacity were available to resolve such a small region using finite elements, it is clear that elements of this size invade the micro-heterogeneity of the ; material and broach the limits of isotropic continuum analysis. 3 ( Several techniques are being explored to circumvent these stringent requirements on crack-tip mesh refinement and related difficulties asso-ciated with possible violations of continuum assumptions. Nishiokall! has proposed an exclusion zone technique that obviates the need for , l highly refined crack-tip elements. In this technique, a small rectan- l' gular domain of height 2c is defined around the crack tip to approximate the finite fracture process zone. During the dynamic analysis, this rectangle is extended in length (but not in height) to include a portion of the plastic wake behind the advancing crack. Nishioka ill advocates 12 i
i l 1 I excluding the integration of the volume term of the T*-integral 108 over this' extending exclusion zone. According to a study by Nishioka,lli the T*-integral should be essentially invariant with respect to the size of ; this extending domain provided e is sufficiently small. To investigate ] the potential of the foregoing technique for characterizing fracture behavior, O'Donoghue81 at SwRI recently performed studies of the geometry independence of the two-component parameter (T*, c) using a center-crack panel problem. Results of these analyses indicate that the I time history of T* was relatively insensitive to mesh refinement for a given height of the exclusion zone. Based on these preliminary calculations, further studies will be conducted by SwRI on the geometry independence of the (T*, c) parameter when applied to small- and large-specimen. crack run-arrest data. Moving singular element formulations represent an alternative tech-nique for achieving convergent solutions for fracture parameters in the context of viscoelastic-dynamic fracture analysis. In preliminary work, Thesken and Gudmundssontt2 have implemented a variable-order singular element proposed by Akin t l3 into an elastodynamic finite-element form-ulation and have illustrated its advantages in modeling stationary cracks subjected to dynamic loading. More recently, the formulation of Thesken and Gudmundssonll2 was extended to incorporate a moving element formulation which allows an adjustable region of convecting elements to be embedded at the crack tip within a finite body. The latter technique permits the order of the crack-tip singularity to be specified by an adjustable parameter for dynamic crack growth problems. The necessary formulations have been installed in the elastodynamic finite-element program DYNCRACK. During this report period, the DYNCRACK program was implemented at ORNL and fully tested on a number of classical dynamic fracture problems. The program has been shown to perform well, with the variable-order singular element having an extremely favorable effect on the rate of convergence. Work is currently under way in the HSST program to update the moving element formulation in DYNCRACK to accommo-date viscoelastic material behavior. The resultant formulation will be investigated for its potential in resolving the near crack-tip singular fields of the Bodner-Partom constitutive model while remaining in the size regime of a continuum element.
- 3. MATERIAL CHARACTERIZATION AND PROPERTIES (R. K. Nanstad) 3.1 Introduction l
The primary objective of this task is to characterize the mechan-ical and physical properties, with emphasis on fracture behavior, of the materials used in the structural experiments that are carried out under other tasks of the HSST program. Other supporting objectives include test method development, the use and development of metallurgical tools to support all activities, and active participation in codes and stand-ards activities. 13 1
- x. ,
i i 13.2 ~ Clad-Plate Base Metal Characterization (S. K. Iskander)
.The pretest and posttest material characterization performed pre-viously for this test series (see section 7.2).has been reported in i
Refs.'114 and 115, respectively. All eight clad plate test specimens were machined from a single, specially heat-treated A533 grade B.
' chemistry' plate. In this section, the results of crack-arrest toughness
-tests performed on base metal of the clad plates are presented.
Crack-arrest specimens were machined from the base metal.of clad plate CP-18, with an L-T orientation, corresponding to'the electron-beam (ES) induced-flaw propagating along the surface.- The base metal is A533 grade B steel chemistry with a special heat treatment to raise its
~
l RT NDT* The' material drop-weight. nil-ductility transition (NDT) temperature is 36*C. The'RT NOT determined'according to NB-2331 of the American Society of Mechanical Engineers.(ASME) Boiler and Pressure Vessel Code, Sect. III, which specifies the use of T-L orientation CVN specimens, is 7?*C. 'However, the orientation for surface crack propagation in the clad plates.is'L-T, thus;the results of.the crack-arrest tests have been
. normalized using the NDT temperature rather than RTNDT* .j The crack-arrest toughness Kj was measured according to American SocietyforTestingandMaterialsfASTM)StandardTestMethodfor Determining the Plane-Strain Crack Arrest Toughness (E1221-88) and shown
'to be conservative relative to the ASME Sect. III K y curve except at very low temperatures, .whereitappearsthat'thewefd-embrittled(WE) starter notch induces too large a driving force Kg , invalidating the~
results.- To reduce the driving force, a machined, aharp starter notch with no irittle weld, usually fatigue precracked and/or warm prestressed
~
(WPS), will be used to obtain data at low temperatures. The mean of the K y ,/KIR ratio and standard deviation derived from the present study (1.81 and 0.40, respectively) seem to be higher than those obtained in two previous investigationsils,it7 on the same material. .However, the scatter of values from each specimen size group in the'present investigation, as measured by the standard deviation, is smaller. The higher Ky ,Nyg ratio from the present study may result from the unusual heat treatment that the clad plate base material has been subjected.to in order to raise its RTNDT. On the basis of the l limited number of results, it seems that K j for this material in the i temperature range -25 to 75'C is only mildl, ydependent on the temperature. These results will be used in the analysis of the tests on clad plates. i 14 i
-l i
l I 3.3' Development of 50-mm-Thick Duplex Crack-Arrest Specimens (S. K. Iskander.and J. F. King) Successful' crack-arrest testing at temperatures 20 to 40 K above the NDT is difficult to achieve with 4fE specimens. Dup 1.ex specimens 25 to 35 mm thick, with A533 grade B and 2.25Cr-1Mo test sections, have beenusedforthis' purpose.jjgell? Also successfully tested were 50-mm- :l thick 2.25Cr-1Mo specimens. However, thus far, the number of suc- ; cessful. tests on 50-mm-thick A533 grade B or weldments at temperatures ' 20 to 40 K above NDT has been.small. In general, unsuccessful-tests on duplex ~ specimens are characterized by the running flaw either arresting in or making a right-angle turn and propagating along the heat-affected zone (HAZ) of the 4340 steel. To increase the number of successful tests on 50-mm-thick speci-mens, the EB-welding parameters have been optimized to reduce the width of the fusion-region as well as its porosity to increase the probability of the crack penetrating the tough HAZ. The variation of microhardness across the fusion zone has been exa-mined'across the specimen thickness in a 2.25Cr-1Mo specimen. Hardness
" peaks" in A533 grade B HAZ are known to be associated with higher toughness than its base metal and have indeed arrested propagating flaws in the clad plate tests.- It is not known whether the hardness peaks observed here are also associated with higher toughness than of the 4340 base metal. The hardness peaks in the 4340 HAZ, together with the porosity of the midthickness region, may be the cause of the difficulty of the cracks penetrating the fusion zone.
In the case of 2.25Cr-1Mo steel 50-mm duplex specimens, the tests were successful. However, for two other materials, A533 grade B and its weld metal, the crack-arrest tests on 50-mm duplex specimens were not successful, again because'of flaws arresting / diverting in the HAZ of the 4340. Examination of the hardness across the midthickness of duplex specimens from two such unsuccessful tests has shown the variation of hardnes, in each of these specimens is similar to that in the successful test. Inus, the reason for the arrest in the HAZ of the 4340 of duplex specimens made of one material and not in those made from another is l unknown at this time. The thermal properties of all three materials are similar. The potential gains to be made by optimizing the EB-weld are dimin-ishing. There is probably more to be gained by optimizing the specimen geometry. One approach under examination is the side groove geometry,- for example, reducing its sharpness so that crack initiation is moved from the edger toward the center. 15
1 I 3.4 Material Characterization and Fractography for Wide-Plate Series 2 , (S. K. Iskander and I. B. Johansson) The material used in the Wide-Plate Crack-Arrest Series 2 (WP-2) is a 2.25Cr-1Mo steel specially heat treated to produce LUS toughness and a high-transition temperature. The posttest characterization of this material is being conducted with specimens machined from the broken halves of wide plate specimens WP-2.1 and WP-2.5. All characterization has been performed in the T-L orientation, i.e., the fracture plane of the test specimens is parallel to the fracture plane of the WP-2 plates. l Drop-weight nil-ductility transition temperature (NDT), Charpy V-notch (CVN), tensile, and fracture toughness testing from both WP-2.1 and WP-2.5 material has been completed. The NDTs were determined to be 60 and 55"C for material from plates WP-2.1 and WP-2.5, respectively. Charpy V-notch testing has been performed on 12 specimens from the 1/2t depth of each of WP-2,1 and WP-2.5. In addition, 12 specimens were tested from the 1/4t depth of WP-2.1. All specimens were in the T-L i orientation and were tested from 25 to 250*C. The upper-shelf energy (USE) levels are very similar, ranging from 61 to 65 J. The greatest differences in Charpy impact energy (=10 J) were exhibited in the lower transition region around NDT. The knee of the upper shelf is reached at approximately 150*C when judged by the attainment of 100% shear fracture appearance. This compares favorably with the pretest results. The transition temperature (defined as approximately one-half the upper-shelf energy) from the present study ranges from 76 to 82*C and compares well to that for pretest results, 74 to 81*C. Examinations of the fracture surface of wide plate crack-arrest specimen WP-2.5 have been performed with a scanning electron micro- ! scope. Primarily, cleavage events associated with the first crack ini-tiation site and each reinitiation site following a crack arrest have been examined. It was obvious that significant ductile tearing preceded the first cleavage crack initiation event. The ductile tearing had a typical dimple structure and extended about 2.2 mm beyond the electron-beam weld. The second cleavage event was also preceded by ductile tear- ; ing, and the cleavage initiation site was associated with a particle I located a small distance ahead of the ductile tearing region. Ductile tearing also occurred upon all the subsequent six crack-run arrest events. Of the remaining crack initiatior. aites, only the sixth initia-tion site could be accurately determinet Similar examinations will be performed on WP-2.4. 3.5 dc-Potential Drop JR Curve Round Robin (D. J. Alexander and j R. K. Nanstad) The JR curve for four specimens, two steel and two aluminum alloys, were determined using a computer-controlled test procedure and a de-potential drop (de pd) technique for monitoring crack extension. The tests were performed as part of a round-robin project on the use of 16 a
s u (
)
.i de pd coordinated by the' David Taylor Research Center (DTRC). Specimen's -
were-fabricated and precracked by DTRC and were tested at ORNL without modifications. The ORNL computer-controlled procedure is one which allows the operator to choose unloading compliance or'de pd as the con-trolling test method. For.this round-robin project, unloading compli-
.ance was selected for test control and de pd was also'used to monitor crack extension.- This method was chosen to allow for direct comparison of the'two techniques.
Results from all four tests showed fairly good agreement between 1 the compliance.and'de pd crack extension predictions and good' agreement of both techniques with the measured crack extensions.. .It-is clear that' the selection of normalization voltage.is crucial to the final analy-ses. For the two steel specimens, agreement between J y values and the JR curves was good. Forthetwo'aluminumspecimens,a[argediscrepancy was observed and subsequent discussions.with=DTRC point toward specimen orientation differences. 3.6 Low-Upper-Shelf Weld Material Investigations (D. J. Alexander, R. K. Nanstad, and C. M. Goodwin) In preparation'for the conduct of the Fourth Pressurized-Thermal-Shock Experiment (PTSE-4), 2.25Cr-1Mo weld metal and flux were procured for the fabrication of a submerged-arc weld intended to have low CVN
' impact energy on the ductile shelf and a high-transition temperature.
~
The target ranges of properties are a yield strength of 520 to 760 MPa, a CVN ductile. shelf' energy'of 54 to 68 J, and'a CVN transition temperature (average of upper- and lower-shelf energies) of 90 to 140*C. The weldment was prepared in 150-mm-thick plate cnd sectioned into slabs of sufficient thickness for machining of CVN and tensile speci-
. mens. Some slabs were given postweld heat treatments (PWilT) of 510*C (P510) or 565*C (P565) for 6 h followed by furnace cooling to 150*C.
Specimens were also fabricated from as-welded (AW) slabs. Tensile results for the AW, P510, and P565 conditions gave average yield strengths of 726, 707, and 689 MPa, respectively. The ultimate strengths were'similar, ranging from 804 to 821 MPa, while total elong-ations range from 11 to 13%. Charpy impact results revealed transition temperatures of 40, 34, and 60*C with upper-shelf energies of 48, 42, and 42 J, respectively. Furthermore, the lateral expansion values on the upper shelf are around 0.81'mm. These preliminary results show the desired yield strength, but CVN transition temperatures and ductile shelf values lower than desired. Chemical analysis of the weld is under way as are additional PWHTs to define a wider treatment versus behavior window. ;
-17 l
I 1
- 4. SPECIAL TECHNICAL ASSISTANCE 1
4.1 Environmentally Assisted Crack-Crowth Technology * (W. H. Bamford) This task has been carried out through subcontract with Westing-house Electric Corporation, Plant Engineering Division. Experimental work was concluded at the end of FY 1986. Follow-on work will be i performed by other ongoing Nuclear Regulatory Commission (NRC) programs to the degree necessary. The first priority of the task was to complete the characterization of two major known influences on the enhancement of crack growth rates in water environments: the influences of the material chemistry through sulfur content and the influence of the water environment. Both of these tasks were concluded. Another important task was to develop a relationship between cyc-lic-fatigue and static-load crack growth. This work is interrelated with the mechanisms of environmental enhancement of crack growth and should considerably improve understanding in this area. This work will continue elsewhere. The ultimate goal of the task has been to propose, as appropriate, revisions to the ASME reference crack growth gurves based on environmental considerations. A final report summarizing the results of this program from its inception (early 1970s) was published this year. 4.2 LWR Vessel Supports (R D. Cheverton, W. E. Pennell, C. C. Robinson, R. K. Nanstad and F. B. Kam) 1 Structural supports for most pressurized-water-reactor (PWR) pres-sure vessels are located in the cavity between the vessel and the biological shield (Fig. 4.1). Within the cgvity ghe fast neutron flux ($) for energies (E) > 1.0 MeV is 52x 10 n/cm *s, and temperatures are <66*C. The corresponding calculated increase in the NDT temperature by 32 effective full power years (EFPY), based on the radiation embrittlement data available from materials testing reactors (MTRs) prior to 1987 (Refs. 119,120), is quite small, if the difference in the MTR and the PWR cavity fast neutron energy spectra is neglected. ! However, late in 1986, data from the High Flux Isotope Reactor [ (HFIR) vessel surveillance program 121,122 indicated that the embrittle-l ment rates of the several vessel materials (A212-B, A350-LF3, A105-II) l weregygstantiallygreaterthananticipatedonthebasisofMTR l data. Further evaluation of the HFIR data suggested that a fluence rate effect was responsible for the apparent discrepancy. As a result
- Work sponsored by USST program under subcontract between Martin ;
Marietta Energy Systems, Inc., and Westinghouse Electric Cort tation, Plant Engineering Division. l 18 i i
of this new information, the NRC requested that ORNL evaluate the impact of the apparent embrittlement rate effect on the integrity of LWR vessel supports. The scope of the ORNL evaluation included correlation of the HFIR data for application to the evaluation of LWR vessel supports; a survey and cursory evaluation of all U.S. vessel support designs; selection of two plants in accordance with established criteria for specific plant evaluation; and a specific plant evaluation of both plants to determine critical flaw sizes for their vessel supports. Two sets of radiation damage trend curves (ANDT vs dpa), based on the HFIR vessel surveillance data, were developed, and 32-EFPY ANDT values were calculated for " typical" Ceneral Electric (CE), Babcock and Wilcox-(B&W), Westinghouse, (W) and Combustion Engineering (CE) plants, assuming that a critical portion of a support existed in the cavity at mid-height of the core. The results, presented in Table 4.1,. indicate much larger shifts in NDT based on the HFIR data than on the MTR data. Many of the vessel supports are not located at mid-height of the core and thus experience smaller shif ts in NDT than those indicated in Table 4.1. However, the two plants selected for specific plant evaluation, Trojan (Portland General Electric) and Turkey Point Unit 3 (Florida Power and Light), both of which are PWR plants, do have vessel supports similar to that shown in Fig. 4.1 which are located near the mid-height of the core. Loading conditions for the evaluations were provided by the utilities and included large- and small-break loss-of-coolant accidents (LBLOCA and SBLOCA), seismic loading, thermal loading and dead-weight loading. i Table 4.1. Vessel support ANDT valves corresponding l to 32 EPPY and mid-height of core ; (" typical" LWR plants) { r 1 1 MSSS ANDT *C I
& dpa rate designer (type (E > 1 MeV) (E > 0.1 MeV) dpa n/cm2 s s1 MTR HFIR data reactor) data Ad Bd CE (BWR)b 2.9 x 107 5.8 x 10 24 5.8 x 10 5 0 d d B&W (PWR)C 2.0 x 108 6.1 x 10 13 6.1 x 10 4 11 100 72 W (PWR) 5.9 x 108 3.9 x 10 12 3.9 x 10 3 28 133 122 CE (PWR) 1.8 x 109 4.5 x 10 12 4.5 x 10 3 39 139 122 ANDT vs dpa (E > 0.1 MeV) correlations A & B.
b Boiling water reactor. C Pressurized water reactor. d ANDT not estimated for dpa < 10 -4 19
p ,L. (:
}l. ,
F d i I J
.j ORNL DWG 884837 ETD .l l
.I, g ;
.v .- r. g.t. ,
\ll4 Vny '
. %,4 9
- 4. ,g:ei, '
1.... .
..y a ?
f.d BIOLOGICAL p(4 SHIELD s h e'. , uf
;. e
.f..
9:' sn 3.,. 9
$*, I
#is: l I
- c. v:#
,J Ls _ _ _ _ _ _ _ _ _ _
I -
. _____.3 .
u _ _ __ s i i a __ _ _ _ _c -
.I ,___. . I. .__,
," SUPPORTM -
f,"]. i]I II
- i. .
' d LJ u LL v;, l .'
s' l REACTOR N A' (.. p'.! VESSEL
.j.$
y)
\
, e, > ;
u;
$. ' r.N ,4l.
.N$ f,:;i
.Ne n i ,. .
"i! \ CAVITY sj,M.a .
j Fig. 4.1. A PWR vessel suport . located in the cavity between the vessel and the biological shield , 1 i i 20 i
1 For both plants the flaw location resulting in the smallest criti-cal flaw depth was on the upper flange of the horizontal cantilever beam at a point within the concrete biological shield. The best-estimate values of T-NDT at these locations are -7 and -14*C for 32 EFPY and -7 and 16*C for late 1988 for Trojan and Turkey Point, respectively. A number of uncertainties exist in the analysis including the radi- I
'ation damage trend curve deduced from the HFIR surveillance data, the !
initial NDT of the material, and the normal operating temperature of the structure, resulting in a wide variation in calculated critical flaw size. Critical flaw sizes (depth / surface length) were calculated for a reasonable rar.ge of conditions and 32 EPPY "best estimate" values for the most sensitive location and type of flaw. For the most severe cred-ible loading conditions, the values are 10/64 mm for Trojan and 8/15 mm for Turkey Point (Table 4.2). Corresponding values for late 1988 are 23/406 mm and 8/15 mm, respectively. The propagation of flaws by low-cycle fatigue was calculated to be negligible. Table 4.2. Summary of best-estimate minimum critical flaw sizes and values of T-NDT for Trojan and Turkey Point vessel supports (cantilever beam) Critical flaw size T-NDT (depth x surface length) Plant EFPY
# s LBLOCA SBLOCA [ggEC Trojan 7.5d -7 -7 23 x 406 28 x 406 32 -7 -65 10 x 64 30 x 64 Turkey Point ll.8d 16 6 8 x 15 25 x 51 32 -14 -14 8x 15 23 x 46 d
At location of minimum-depth critical flaw. bat inner surface of biological shield (cavity interface). C Safe shutdown earthquake. d Late 1988. Based on the results of our study a recommendation was made for continued work in this area. An effort should be made to reduce the uncertainties in the material and operating data used as input to the evaluation, including (1) interpretation and extrapolation of the HFIR surveillance data; (2) fracture toughness of specific materials, includ-ing NDT n ; (3) credible static and dynamic loading conditions (Turkey 21
Point); and (4) operating temperature. The likelihood of the existence 4 of flaws of estimated critical size and larger should also be eval- l uated. An examination of inspection procedures and results, if and when they become available, may be useful in this regard. Because shield-tank and long-column supports are exposed to the maximum fluxes in the cavity, a review of utility / industry evaluations of these structures or an independent evaluation should be conducted. The variation in horizontal loads resulting from unequal friction in the components that accommodate thermal radial expansion should be quantified. 4.3 BWR Vessel Integrity Assessment (R. D. Cheverton and D. C. Ball) In November 1987 the NRC requested that ORNL begin to evaluate the integrity of BWR vessels when subjected to PTS loading. Two specific reasons for concern are that (1) surveillance data from BWRs indicate that the embrittlement rate may be greater than expected and (2) the 10-year inspection for flaws specified by the ASME code is not required by the NRC for BWR vessels. Previously, it was believed that radiation-induced embrittlement of BWR vessels and thus the potential for propaga-tion of flaws in the belt-line region of these vessels were essentially negligible. The approach selected involved the calculation of the conditional probability of vessel failure in accordance with the methodologies devel g d for the Integrated Pressurized-Thermal-Shock (IPTS) Pro-gram. This approach permits consideration of the influence of in-service inspection on the calculated probability of failure, because it is necessary to specify flaw density and flaw size distribution. For the analyses conducted, values of these parameters were consistent with no in-service inspection. Details of the method of analysis for estima-ting the conditional probability of vessel " failure" (P(FlE)] and of the parameters simulated are discussed in Refs. 123 and 124. On the basis of discussions with NRC staff and the description of the Design Basis Accident (DBA), Ref. 125, it was tentatively concluded that BWR vessels may be subjected to thermal-shock loading as a result of sudden loss of pressure, but presumably there is no means for rapid repressurization. Thus, the conditional probability of failure (P(FlE)] was calculated for thermal-shock loading conditions only. An exponen g tial decrease in coolant temperature of the form Tc = T (288 4p)e was considered, where cT is the coolant temperature in g+TC, p is the final coolant temperature in *C, and t is time in minutes. Calculations were made for 8 = = (step change in temperature from 288*C to (T p ), 8 = 0.15 min 1 (AT = 167"C for t = 13 min), and Tp = 93 and 121*C. For all cases, the pressure was assumed to be essentially atmospheric. Two va}uesoftheflujd-fil,m, heat-transfercoefficient were used (1.70 x W/mC), the lower value representing natural con-10 and 3.41 x 10 vection and the higher value moderate forced convection. l 22 1
Thus far, two sets of calculation have been made. For the first set, radiation embrittlement was deduced from the PTS embrittlement { trend curve used for the IPTS studiesi l?3 fluence data 126 were taken { from a very early compilation; copper (Cu) and nickel (Ni) concentrac-tions of 0.35 and 0.80% were used; and the initial value of RT was UDT 18'C. For both sets of calculations the flaws were axially oriented and two dimensional (2D), and the vessel dimensions were 6.50-m inside dia-meter, 0.156-m wall, and 3.81 mm of cladding. Also, for both sets of calculations, the flaw density was assumed to be 0.1 flaws / vessel, con-sistent with what was considered to be reasonable for the IPTS stud-ies.123 Results for the first set of calculations indicated for a fluence (e) of 1 x 1019 neutrons /cm2, the conditional probability of crack ; initiation [P(IlE)]rangesfrom4x 10 6 to 8 x 10' , and the condi-tional probability of a flaw initiating and not arresting within the j inner 95%ofthewall[P(FlE)]rangesfrom2x 10 7 to 7 x 10 9 [0f course, the slower the thermal transient and the smaller the value of the surface heat transfer coefficient (h), the lower the probability of
" failure."] For fluences of 3 x 1018 and 1 x 1018 neutrons /cm2, the probabilities are about 2 and 5 orders of magnitude less, respec-tively. It was apparent that even for the very severe conditions considered, all of the plants except possibly Big Rock Point and La Crosse have low estimated frequencies of failure.
For the second set of calculations an updated complication of ves-sel chemistry, initial RTNDT, and fluence data were used; on the basis of information in Ref. 127 and discussions with Neil Randall,i Regulatory Culde 1.99 Rev. 2 was used in lieu of the PTS trend curve. (However, even Regulatory Culde 1.99 Rev. 2 may underestimate the damage rate in BWRs.) For the first set of calculations referred to above, copper and e were simulated and used as independent variables. In the second study, ART DT was used as an independent variable instead, and the dis- ' tributionforART was obtained using the Regulatory Guide 1.99 Rev. 2 trendcurveandMNeCarlotechniques. In this study Cu, Ni, and 4 were used as independent variables, and ART was, of course, the dependent variable. CalculatedvaluesofPkE)andP(IlE),basedon Regulatory Culde 1.99 Rev. 2 and using ART NDT as the independent vari-able, have been obtained for step changes in coolant temperature from 288 to 93*C and from 288 to 121*C, a fluid-film heat-transfer coeffic- ! ient of 3.4 x 103 W/m2**C, and three values of RT ' '
- NDT o
*H. E. Mayfield, Personal Communication to R. D. Cheverton, Oak Ridge National Laboratory, Oak Ridge, Tn., March 22, 1988.
i P. N. Randall, Personal Communication to R. D. Cheverton, Oak Ridge National Laboratory, Oak Ridge, Tn., January, 1988. 23
I Based on the recenft compilation. of chemistry, RTNDT "" ~ i fluences for all BWR vgssels* and the information from the study, it l appears that the most suscept. e BWR vessel examined'is one with RTNDT
= 157'C and RT NDT
= d4 C. F his vessel and a step change in coolant temperature of 198'C, P(F E) < 1 x 10 9, and P(1 E) = 1 x 10-7 1 for a step change of 167'C, P(F E) < 1 x 10-9, and P(I E) < 1 x 10 9 If the actual time dependence of the coolant temperature were included, the results would rot be significantly different because the initial decrease in temperature is so rapid. However, if a lower value of the final bulk coolant temperature were used, P(FlE) could be sub-stantially greater.
AdecreaseinP(FlE)canbeachievedbyusingam'orerealisticflaw shape [three dimensional (3D) as opposed to 2D]. However, results of the IPTS studies 123 indicated that the effect was not large. Perhaps l the greatest reduction in the calculated frequency of failure can be achieved by thoroughly examining the frequency of events. 4.4 Evaluation of Impact of Low-Upper-Shelf Weldments on Pressure Vessel Integrity (J. C. Merkle) This year the HSST program undertook the task of examining the impact.of the existence of LUS weldments in RPV's. The effort has consisted of_two major thrusts: (1) participation in an ad hoc task group convened to examine both current and improved controlling parameters for ductile crack extension, and (2) compilation of a state-of-the-art assessment of the entire LUS issue. The specific problem concerns the continued operation of PWRs with low-upper-shel f weld toughness. Already, 5 units have fallen below the legal screening limit of 68 J, and it is anticipated that an additional 10 units will do so before reaching the end of their original design lives. It is expected that continued operation below the 68-J limit can l be justified using an instability (J-T) analysis demonstrating that ) unstable crack propagation will not occur. A draft ASME document using this approach suggests that crack extensions up to 18 mm may be required. Furthermore, it is proposed that Jmod should be used as the parameter controlling ductile crack extension rather than the con-ventional deformation J (JD ). H wever, the use of Jmod may be prema-ture, because some J-A a data are known to exhibit an upward curvature that is physically unacceptable. l l l
,M. E. Mayfield, Personal Communication to R. D. Cheverton, Oak Ridge National Laboratory, Oak Ridge, Tn., March 22, 1988.
24 1
HSST has been active in the series of ad hoc working group meetings through participation of the ORNL staff as well as that of John Landes of Univ. of Tenn. and flugo Ernst of Alpha Research who were placed under j subcontract for this purpose. The meetings have been held with two objectives: (1)-to examine which J parameter is the more correct and define ranges of valicity, in particular, the allowable amount of crack j extension; and (2) to determine how small specimen data may be extra- { polated and applied to structures. Conclusions from the discussions can i be summarized as follows.
- 1. Considerable uncertainty exists regarding which J-like parameter is fundamentally more correct and its range of applicability. l
- 2. At present, for conservatism, it would appear prudent to continue with deformation theory J, although with further advances in defin-ing its limits of applicability, J mod may be preferred.
- 3. For many materials. existing crack growth (A a) limits in testing standards now appear unduly conservative: 30% of the remaining ligament appears more appropriate.
- 4. In testing high-toughness materials, it is more difficult to obtain
" valid" data covering a fixed amount of crack extension. For these materials, J limitations are likely to be more restrictive than crack extension limits; thus, it is unlikely that any increase in allowable crack extension would be of particular benefit. This is likely to be the case with-modern high purity reactor pressure ves-sel steels and austenitic stainless steels.
- 5. There is a continuing need for development of elastic plastic frac- !
ture-mechanics (EPFM) testing techniques to assist in the evaluation of different EPFM parameters, their ranges of applicability, and ; means of extrapolation to larger crack extensions.
- 6. Current approaches to defining allowable J and A a limits are incom-plete because they do not include a material sensitivity factor l relating crack growth to changes in crack-tip constraint. i An outline of the white paper on the LUS issue was compiled. The white paper covers a number of items, including: the causes, history, scope, and status of the problem; the historical and fracture mechanics bases for the current 68J legal lower limit for charpy upper-shelf energy; the code and regulatory guidance which exists and an appraisal of its adequacy; an incorporation of the initial evaluation of appropri-ate fracture parameters from the ad hoc task group; and a der iption of relevant materials data for irradiated LUS welds. It is antici pated that the white paper will provide a basis for the recommended courses of action necessary to reach an improved regulatory position on the con- l tinued operation of vessels obtaining LUS welds. i l
I i 25 i i l l L.____._._.___ _ _ _ _ . . - -_
- 5. CRACK ARREST TECilNOLOGY [D. J. Naus, J. A. Keeney-Walker, C. B.
Oland, C. C. Robinson, and B. R. Bass (ORNL), R. J. Fields, R. I deWit, and S. R. Low III (NBS)*] f 5.1 Introduction The primary objective of the crack-arrest studies is to generate data and associated analysis methods for understanding the crack-arrest behavior of prototypical RPV steels at temperatures near and above the ; onset of the Charpy upper-shelf region. Program goals include (1) ex-tending the existing K ,g data bases to values above those associated with the upper limit in the ASME Code; (2) clearly establishing that crack arrest occurs prior to fracture-mode conversion; and (3) valida-ting the predictability of crack arrest, stable tearing, and/or unstable tearing sequences for ductile materials. Wide plate tests and analyses are being utilized to provide bases for obtaining and interpreting dynamic-fracture data (with relatively long crack runs) and bases for validation of viscoelastic fracture mod-els and analysis methods. Supplemental tests are also planned using ; intermediate-size, stub panel crack-arrest specimens. During this reporting period four wide plate crack-arrest tests were completed, con-cluding the wide plate testing series. Checkout of the loading and heating-cooling systens for the stub panel crack-arrest specimens was also completed as well as development of a performance specification for a high-speed data acquisition system. 5.2 Wide-Plate Crack-Arrest Experiments The wide plate crack-arrest experiments have been performed at the National Bureau of Standards (NBS), Caithersburg, Maryland, under an interagency agreement. The tests are designed to provide fracture- l toughness measurements at temperatures approaching or above the onset of the Charpy upper-shelf regime, in a rising toughness region, and with an i increasing driving fcirce. In addition to providing crack arrest data, the wide plate tests provide information on dynamic fracture (run and arrest) processes that are being used by researchers at ORNL, SwRI, and the UM to develop and evaluate improved fracture-analysis methods.
- Work sponsored by llSST program through an interagency agreement between the Department of Energy and Department of Commerce.
26 l l l l -- - _ _ _ _ _ _ _ _
Ofi?llI'llCIO 6710 65 g
-su.
O r 1. A
?. L 1
\p;.
3.-
,j n: '
g .
\y1, c
, ,e i.
s a NB
.. f 5r ,
pf i i . l l _g- . v 1,.
.9 {
j Fig. 5.1. Wide plate crack-arrest specimen under test. 27 i
)
l l 5.2.1 Specimen Preparation Each test specimen is fabricated by welding a precracked, single- ) edge notch test article (1 x 1x 0.1 m or 1 x 1x 0.15 m) to pull plates having the same cross-section geometry as the test article to form a test specimen approximately 10 m long. Up to 40 thermorouples and 25 strain gages are positioned on each specimen, with the concentration of 1 instrumentation being adjacent to the plane of crack propagation. The ! specimen is then placed into the NBS testing machine and the heating-cooling-insulation systems installed as well as the balance of the instrumentation (crack-opening-displacement gages, acoustic emission transducer and displacement transducer). More details on specimen prep-aration can be obtained from Ref. 89. l 5.2.2 Testing Procedure i After checkout of the instrumentation systems to demonstrate oper-ability, a thermal gradient is imposed across the plate by liquid nitro-gen cooling of the notched edge while heating the other edge of the plate. When the desired thermal gradient is established, final calibra-tions on the instrumentation systems are completed and tensile load is applied to the specimen until the crack run-arrest event (s) occurs. Figure 5.1 presents a specimen under test. 5.2.3 Test Summary The wide plate crack-arrest tests have utilized three materials: (1) prototypical A533 grade B class 1 material from il3ST plate 13A (WP-1 series), (2) a second heat of A533 grade B class 1 material supplied by Combustion Engineering, Inc. (WP-CE series), and (3) a degraded (simula-ted) low upper-shelf base material (WP-2 series). Properties of these materials are presented in Refs. 89, 128 and 129. A total of sixteen tests have been conducted, completing this experimental series, with four having been conducted during this reporting period: WP-1.8, WP-CE-2, WP-2.2, and WP-2.6. t 5.2.4 Wide-Plate Analyses and Results Crack-arrest toughness values are computed for each wide plate specimen using both static and dynamic analyses as well as handbook l techniques. Figure 5.2 presents generation-mode (fixed-load boundary I condition) crack-arrest toughness values versus arrest toughness temperature minus RTNDT f r each of the wide plate tests conducted to date (generation-mode results for specimen WP-1.1 are not available). The data show that cr est can and does occur at temperatures up to and above that which corresponds to the onset of Charpy upper-shelf energy values extending T-RT NDT = 78 C for WP-1 material), with measured Kia l above the limit included in Section XI cf the ASME Code. 28
~
- 73 .
r E n 8"A 4 u .
" o B
E " r 4 0 C a
- R . 2 D p L r
F 1 i
' i
= s
% e.
S R ' D> E E T ' V E 2 m c mO M aO - T N
$ 93GR 4
E 2 3 E . C
==O _
N S O O t OB ;S t F " " I S S - N S . - O ' E - S N K 6 1 N E C 0 I H
$ 1 I1 M T - D S ^
0 D - 6 T _ E - _ - E X I _ 1 A M 0 2 8 9 l SN i 4 1 4- A O I 7 8 T I T 0 i 2 MC l G k E 1 W A LI S O L 4 3 0 0 a' I l 1 N o - 1 R c 8 ^a A ) O
" 0 C l
m % hIo*
< m 1
8 O ( ( 1 3452 E 2
- *0 m o
g 4 l 0 D 6 N T 2- 2222 C- e T I P PpPp P O*> ^O 0 - R 4T I W WWWw W S e >oA4 a o 0 1
- I 2
E 2 3 4 5 6 7 8. C 1111111 I 0 PPPPPPPP WWWWWWWW 0 2
>030InAa' _
I
- - - - - - 0 4-0 0 0 0 0 0 0 0 0 0 0 0 6 - 5 4 3 2 1 s
Ef c$ mnszIo=>ow [ eu gm=m}o4eU oo rs [l-l t! l I'
Table 6.1. Summary and status of the ten irradiation series of the HSST Program I' Neutron fluence Irradiation Status as of Series Objective and Specimen um ure l number materials types n/cm2 x loa November 1988 (>l MeV) 1 Upper transition' O.4 4TCS 2.2-7.0 270-300 Complete plate.and weld CVN, TEN metal 2,3 Ductile shelf 0.5-4TCS 0.4-1.2 233-343 Complete low CV shelf, CVN, TEN high-copper welds 4 Ductile shelf 1TCS 0.5-2.7 288 5 Complete HSST Plate 02 and CVN, TEN state-of-the-art low-copper welds, FRG material 1-4TCS 1.3-1.8 288 i 10 Irradiations 5 KcI curve shif t and testing high-CV shelf, CVN, TEN high-copper weld DWT completed i metal Ka 1.0-1.3T Target: 288 10 Irradiations l 6 l curve shift and unirra-high-CV shelf, CCA 1.7 high-copper weld diated tests j metal completed ! 7 Stainless steel 0.5TCS 2.0-2.5 288 10 1-wire cladding single- CVN, TEN and complete, and multinire 5.3-5.8 3-wire irradiations and testing completed curve shift 1 4TCS Target: Target: Planning 8 K Kla 1Io,wupper-shelf CVN, TEN 2.0 288 welds DWT 9 Thick-section 1-4TCS Target: Target: Planning annealing CVN, TEN 2.0 288; low upper-shelf DWT anneal welds 10 Low upper-shelf weld 1-2TCS Target: Target: Planning from Midland CVN, TEN, 0.5 and 288 reactor vessel DWT 5.0 30
5.3 Stub-Panel Crack-Arrest Tests 1 i Studies to evaluate the usefulness of a relatively small panel specimen (Fig. 5.3) for crack-arrest experiments have been initiated at ORNL. All test hardware components required for applying load to the specimen have been designed, procured, fabricated, and received. The ,l heating-cooling-insulation systems, which will be used to develop the desired thermal gradient in the specimen, also have been designed, fabricated, and checked out. A specification for a 40-channel high-speed data acquisition and signal-conditioning system has been completed and procurement initiated. i
)
i 6. INVESTIGATION OF 1RRADIATED MATERIALS (R. K. Nanstad) 6.1 Introduction The HSST Irradiation Effects Program now includes ten series, and a summary and status statement (as of November 1988) is provided for each ; series in Table 6.1. The program is designed to provide information ~ regarding fracture toughness of reactor vessel materials, including base materials, welds, and stainless steel cladding. The Fifth Series is intended to validate the amount and shape of the ASME K Ic curve shift as a consequence of neutron irradiation. Currently, estimates of the K curve shift are based on results from Charpy impact testing. TheSix!h Series will determine the effect of irradiation of the material's ability to arrest a rapidly propagating flaw. This series is being conducted immediately following the Fifth Series. The Seventh Series is determining the effects of irradiation on pressure vessel stainless steel cladding. Analyses of certain thermal-shock scenarios have been inhibited by a lack of information regarding the fracture resistance of the cladding. The Eighth and Ninth Series were conceived to determine the K ic curve shif t and shape for low upper-shelf welds and evaluate the effects of thermal annealing in thick sections. The Tenth Series will characterize low upper-shelf weld metal from the Midland reactor vessel j (never placed in service) and include irradiation rate and saturation effects. 6.2 Fifth HSST 1rradiation Series (R. K. Nanstad, F. M. Haggag, and S. K. Iskander) The primary objective of the Fifth Series is to obtain valid frac-ture toughness, K curves for two nuclear pressure vessel materials irradiated at 288b,toashighatoughnesslevelaspractical. The irradiation parameters are an irradiggion temperature of 288*C and a j moderate neutron fluence of 1.6 x 10 neutrons /cm2 (>l McV) chosen to i complete irradiations in a reasonable time yet provide an adequate shift. The materials are submerged-arc weldments of 0.23% and 0.31% Cu 'i 31
$', ,a I
f
.I ORNL-OWG 884948 ETD 250 ,
FFTH HSST
- a UNIRRADIATED IRRADIATION SERIES WELD 73W ' ~
200 -'
- IRRADIATED . (0 m.CW 2
-1.75 X 1023 n'm {,jy,y)
~2880C 8
a
$ 150 -
m; " - g "
.; 7 m ,
c .- e - 100 ~ u ---v=3
- V ,Sg e
l . *A
- M%
a s(.- / 82
- 0C 50 - ;e- ,
0
.Lm , ,
300 0 100 200
-200 100 TEMPERATURE ( C)
Fig. 6.1. Charpy V-notch impact tests showed transition shift and shelf drop and a decrease in slope of-curve in transition region. i ORNL DWG 884949 ETD !' 1 l l I 6 4 al a HSST WELD 73W
- OJl (0.31% Cu)
IRRADIATED. 28BOC, 300 - -1.5 X 109 O n/cm 2 (gg,y) 7 " Ao ~ 250 - A iTC$ ' RILLED SYMBOLS- EN D VALID E399 NO E D 2TCS
/ AAA 0 A l O ATCS U
[ ~ *I 41RRAD. b2 UNIRRAD.
.I l ASME 4 .!
DATA "' yl D / j _ f"f 150
+ POPIN .
' O ,
W
//j / e
- M g h ..
A p 4 As i RAD
~
~hY
,j u
- E185 ff j w ~~ % $ e Z .
,/) ' )
j ( 50
~
I __
#,# g ,
RT*g NOT t I I du! i 0 50 100 150 N 50 0 I TEMPERATURE (OC) l Fig. 6.2. Preliminary analyses of fracture toughness results show five data-points below ASME KIc curve shifted by Charpy 41 J shift. A lower-bound analysis , (dashed line) clwws a lower slope for irradiated case similar to Charpy results. { i 32
content (0.60% Ni in both weldments). Irradiation of all 12 capsules was completed in December 1985. Dosimetry results indicate final fluences very near'the target. Both ORNL and Materials Engineering
. Associates (MEA) participated in the testing program.
l All planned unirradiated and irradiated testing for the Fifth Irra-diation Series has been completed. Tests performed included tensile, CVN impact, DWT, and fracture toughness. Unitradiated compact specimens of 25.4, 50.8, 101.6, 152.4, and 203.2 mm thick (1TCS, 2TCS, 4TCS, 6TCS, and 8TCS) were tested while irradiated testing was' limited to ITCS, I 2TCS, and 4TCS. l Results show that the CVN 41-J temperature and DWT NDT temperature shifts are very similar for both welds following adjustment for fluence differences. The CVN 41-J temperature shifts were 72 and 82*C, while the adjusted DWT NDT temperature shifts were 70 and 83*C, for welds 72W and 73W, respectively. Room temperature yield strength increases were ! about 25 and 30% for 72W and 73W, respectively. Regarding fracture toughness, the scatter in both unirradiated and i irradiated fracture toughness results using elastic plastic fracture ' mechanics is large. Figures 6.1 and 6.2 show the CVN and the irradiated fracture toughness results for weld 73W, respectively. Preliminary j analyses of the data, without statistical analys_es, indicate that the lower-bound K y curve shift (101*C at 125 MPa*/m) for weld 73W is greaterthanLSeCVN41-Jshift (82*C) and the postirradiation KIc curve is of shallower slope than the curve for unirradiated material (dashed line in Fig. 6.2). The change in shape, however, is similar to the shape change shown by the CVN results (Fig. 6.1). Final analyses of all results will be dependent on the interpretation of cleavage pop-ins. Because of the relatively large data scatter, wide variations in specimen sizes, and the need to use elastic plastic fracture mechanics, the use of statistical analyses is an important part of this program. The use of Weibull failure statistics is incorporated-in the program as are standard statistics of variance procedures. Additionally, the use of empirical adjustment schemes and other statistical distributions are being investigated, all aimed toward the development of a rational scheme for constructing the irradiated K curves. Ic 6.3 Sixth IISST Irradiation Series (S. K. Iskander) The primary objective of this irradiation series is to establish the amount of radiation-induced temperature shift that occurs in the ASME K ia crack arrest curve of the two high-copper weldments typical of those in older light-water reactor pressure vessels. In conjunction with results from the Fifth Series, the Sixth Series results will allow an assessment of the currently used correlation between the ASME K yg . curve shift and the shift in RTNDT. Toward this aim, the high-copper I weldments and irradiations of the Sixth Series are identical to those of the Fifth Series. 1 53 i _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ .- 1
The two capsules that contained 30 specimens each began irradia-tions in late December 1985 and were completed on May 2, 1986. All unirradiated crack-arrest tests of the small- and intermediate-size specimens of welds 72W and 73W have been completed, but most of the large (203 x 203 x 51 mm) duplex specimens and weld-embrittled control specimens remain untested. Final fabrication and testing of the large duplex control specimens will take place once the specimens have been machined and the appropriate welding parameters for final specimen fabrication have been established. The remotely operated fixture developed for use in testing crack-arrest specimens in the hot cell has undergone testing in the laboratory using specimens of the same size as irradiated ones of the Sixth Ser-ies. The uniformity of temperature distribution in the test specimen has been evaluated for temperature ranges in excess of those envisioned to be used inside the hot cell. The accuracy of the fixture's tempera-ture-indicating device, as compared to temperatures measured by thermo-couples fixed to a specimen, has been verified. Moreover, it has been used to perform unirradiated Ky , testing, and the remote device has functioned without major problems, although some minor ones still need to be addressed prior to its installation in the hot cell. 6.4 Seventh IISST Irradiation Series (F. M. Haggag and R. K. Nanstad) The primary objective of the Seventh USST 1rradiation Series is to examine the radiation-induced degradation of fracture properties of stainless steel weld overlay cladding. In the first series (now complete) good quality cladding typical of what would be expected in a , reactor pressure vessel showed only slight embrittlement, whereas clad- l
' ding highly diluted by the ferritic base plate, exhibited generally poorer properties and enhanced radiation sensitivity. I In the second phase, currently in progress, a commercially produced three-wire series-arc weldment was irradiated by MEA at the University of Buffalo. The three-wire series-arc procedure, used by Combustion Engineering, Inc., Chattanooga, Tennessee, was similar to their commer-cial process used to fabricate early pressure vessels. It produced highly controlled weld chemistry, microstructure, and fracture proper-ties in all three layers of the weld. The weldment was given a postweld heat treatment typical for nuclear pressure vessels. The three layers of cladding were required to allow the fabrication of tensile, Charpy impact, and h5TCS from the cladding. Irradiations of the specimens to 2 and 5 x 10 neutrons /cm2 (>l MeV) have been completed.
Irradiation of the three-wire series-arc staini gs steel cladding specimens at 288"C to fluence levels of 2 and 5 x 10 neutrons /cm2 (>l ' MeV) decreased the CVN upper-shelf energy by 15 and 20% and increased the 41-J transition temperature by 13 and 28'C, respectively, as shown in Fig. 6.3. Irradiation also degraded the CVN lateral expansion signi-ficantly; expansion on the upper shelf was reduced by approximately 40% l \ 34
ORNL.DWG 87C 17216R 90 - - f
.' 20% 15 % 1 70 -
/ '. ----------- ...- --
, 9 60- - ,/ , - h >- / / , m 50 -
./ /
13'c /- -y,15'c j 2 40 - _ W
/ i 30 -
j l _
/ ----- lR R ADI ATED: 5x10" n/cm2(>gy,y) l f'
.20 -
IRR ADI ATED: 2x10" n/cm2g>g y ,y) _ f ~. # --- UNIRR ADI ATED 10 - LS ORIENTATION - I I I 0
-150 -50 50 150 250 TEMPERATURE ('C)
Fig. 6.3. Charpy toughness of three-wire stainless steel cladding is somewhat degraded by irradiation. ORNL-DWG 88-4428 ETD 1.50 , , , UNIRRADIATED 2
---. lRRADIATED: 5 X 10" neutrons /cm 4,3 y,y) -
1*25
-.-IRR ADIATED: 2 X 10" neutronsicm2 g,, 'y,yy E
- LS ORIENTATION 1.00 43% 41% -
Q z I 0.75 -
........ m....m...
g , a ,,-
< 0.50 -
o 1 y = 41 c./,L._. s*c b 0.25 - ll'
/ o s ' o' O.00
' 150
-150 -50 50 TEMPER ATURE ('C)
Fig. 6.4. Effect of irradiation on CVN laterial expansion of three-wire stainless steel cladding. I 35 ., l 4
for both fluence levels, as shown in Fig. 6.4. Furthermore, the 0.38-mm transition temperature shifts were 41 and 46*C for the low and high flu-ences, respectively. These results generally agree with those for the single-wire cladding produced with good weld practice. Preliminary analysis of the 0.5TCS fracture toughness specimens ! fabricated from three-wire series-arc stainless steel cladding indicated thefolloy{ng. Irradiation exposure to an average fluence of 2.41 x 10 neutrons /cm2 (>l MeV) resulted in a consistent decrease in the initiation fracture toughness (J or J at test temperatures of
-75'C, room temperature, 120*C,andhk8"C.g)hisisinagreement T with the reduction in both the CVN upper-shelf energy and lateral expansion discussed above. However, the percent reduction in initiation toughness ,
I of the 0.5TCS is greater than that of the CVN impact energy but closer to the percent reduction of the CVN lateral expansion. Irradiation l exposure has also reduced the tearing modulus at test temperatures of l
-75 to 288"C. Both unirradiated (control) and irradiated specimens !
exhibited the same trend of slightly increased initiation toughness from
-75'C to room temperature, followed by a significant toughness decrease ;
from room temperature to 288"C. l
- 7. CLADDINC EVALUATIONS (S. K. Iskander and K. V. Cook) 7.1 Introduction The objective of this task is to demonstrate the effect of stain-less steel weld cladding on the resistance to and extent of crack propa-gation for small surface cracks subjected to stress gradients similar to that produced by a thermal shock. The occurrence, size, and distribu-tion of such small cracks in representative segments from actual boil-ing- and pressurized-water RPVs is also being studied to broaden the data base used in probabilistic studies.
7.2 Fracture Mechanics Testing of Clad Plates (S. K. Iskander, C. C. Robinson, C. B. Oland, D. J. Alexander, and K. V. Cook) Six plates were commercially clad with stainless steel using the three-wire series-arc technique (used on some of the older reactor pres-sure vessels). The base metal on which the cladding was applied, an A533 grade B chemistry steel, was heat treated to raise its transition temperature so that it would be brittle at temperatures at which the cladding is tough. Two similar plates were left unclad to provide comparison data. l-The tough surface layer of cladding and llAZ seemed to contribute i significantly to the load-bearing capacity of the plates by arresting flaws at loads and temperatures that have ruptured unciad plates, as seen in Fig. 7.1 by comparing the results of the tests on clad plates i 36 ( i i
I l l 1
)
ORNL-DWG 884947 ETD 1000 - CP 19 O O INITIAL LOAD
- A, ST LOAD 900 -
CP 17 0 1 If o
- 800 -
_$ d6% O 4 0 i
' 700 -
v CP 21 C P-15 o O (FAILED) O
& ' - >,/
6 ,I, /; , y, 600 - UNCLAD CLAD PLATES PLATE 500 - i Fig. 7.1. Pop-in, arrest loads, and Corresponding Crack lengths for the four plates tested at room temperature.
)
1 37
i l l l CP-15, CP-17, CP-19 and unclad plate CP-21. In fact, the clad plate CP- l 19 arrested a flaw subjected to a driving force (as measured by the { initial load) almost 50% higher than that which broke an unclad plate. . Moreover, the residual load-bearing capacity of plates (even with f airly large flaws) as measured by the arrest loads was generally greater than required to break the unclad plate, Fig. 7.1. The HAZ played a prominent role in enhancing the load-carrying capacity of the clad plates. As measured by the Charpy V-notch impact energy, the HAZ is the toughest of the three metallurgical zones of the clad plate specimens at 25"C, while the cladding is toughest at -25"C. It is not clear at this time whether cladding alone, without benefit of the tough strong HAZ which played a pronounced role in arresting propa-gating flaws, would have also elevated the load-bearing capacity beyond that of the unclad plate. In the case of radiation-embrittled reactor pressure vessels, the HAZ will most likely undergo toughness degradation similar to that of the base metal, and therefore may not play such a prominent role in arresting propagating flaws. A topical report, Effect of Stainless Steel Weld Overlay Cladding on the Structural Integrity of Flawed Steel Plates in Bending-Series 2, by S. K. Iskander et al., has been written and is under review. 7.3 Flaw Characterization Studies of Clad LWR Vessel Material (K. V. Cook, R. W. McClung and R. A. Cunningham) Nondestructive ygesting tasks on segments of LWR vessels were completed. A paper was published in Nuclear Engineering and Design. , 1 l 1 A topical report, Detection and Characterization of Indications in Segments of Reactor Pressure Vessels, by K. V. Cook, R. A. Cunningham, Jr., and R. W. McClung, has been' prepared, reviewed, and will be pub-lished soon. This report describes the studies that have been conducced to estimate flaw density in segments cut from two LWR pressure vessels. Objectives were to evaluate these LWR segments for flaws with ultrasonic and liquid penetrant techniques. Both objectives were successfully completed on segments from both a BWR and a PWR vessel. The topical report describes in detail that one significant indication was detected in the BWR seam weld segment by ultrasonic techniques and characterized by further analyses terminating with destructive correlation. Interpretations are given relative to assumptions now made in LWR applications. Additionally, the report describes that the PWR vessel segments did not contain as much weldment as originally understood to be the case; thus, the ultrasonic examination was limited to the cladding and subcladding regions. Only one indication of note was detected ultrason- ! ically at the clad-to-base metal interface. 38 l
N ', l ,
, y f
g
. Fluorescent liquid' penetrant inspection of the cladding surfaces
'> for both LWR vessel segments detected no significant_ indications over a 2
L total ~of approximately 6.8.m of cladding surface.
- 8. INTERMEDIATE VESSEL TESTS AND ANALYSIS (R. H. Bryan)
The most recent intermediate vessel test (V-8A) was petformed in 1982todemonstratethefracturebehaviorofst99{withlowtearing pl .resi stance at ductile upper-shelf temperatures. Finite-element anal-Wa yses of that experiment were the basis for estimates of tearing insta-7 bility based on J resistance curve data. Although it was known that the W results of the deformation plasticity finite element analyses were quite sensitive to details of geometrical modeling and the stress-strain rela-Li . tionship~, the calculated instability conditions (pressure = 134.5 MPa; pre-instability tearing = 12.5 mm) agreed reasonably well with the experimen58 1 . observations (pressure = 140 MPa; pre-instability tearing = 5.0 mm). The discrepancy in the calculated extent of tearing was not considered surprising in view of the known sensitivity of computed crack extension to' slight variations in J resistance curves. i
' Fracture' behavior observed more recently in two pressurized-thermal-shock experiments 3*l31 indicated that posttest analyses of the ; later experiments produced unsatisfactory estimates of ' stable ductile tearing. Consequently, further finite-element analyses of several of the intermediate'v essel tests as well as the pressurized-thermal-shock experiments are being made. In addition to the effects of the finite-H . element mesh'and stress-strain models, the influences of large deforma-tions and non proportional loading, which are generally not well repre-sented in. deformation plasticity calculations, need to be evaluated.
- - Finally the assumption that tearing is simply controlled by the J-resis-e tance curve deserves reconsideration.
i
- 9. THERMAL-SHOCK' INVESTIGATIONS (R. D. Cheverton) i'
[ A program is under way at ORNL to evaluate the behavior of subclad L and through-clad flaws in PWR pressure vessels during thermal-shock loading conditions. The effort, which is sponsored by both the NRC and DOE, includes, among other things, the development of fracture-mechanics _ (FM) methods of analysis for subclad cracks; an experimental investigation of the validity of the method of analysis; and, if and when validated, the application of the method or analysis to determine L the. benefit of cladding in restricting the propagation of flaws. The development and validation of the methods of analysis are progressing in the evaluation of a confirmatory thermal-shock experiment E previously conducted. The experiment was performed in the ORNL thermal-shock' test facility in which the gaench medium is liquid nitrogen. The test specimen was a thick-walled steel cylinder similar to those used previously132,133 but with cladding on the inner surface. Two cladding 39 H b . - ~ u_ _ - _ _ _ _ _
^
? l
'4
, e i s
i [O t materials and six semielliptical subclad flaws were included in the
~
p first phase of-the recent experiment. A second phase of the experiment p~ utilized. surface flaws. .Results of the experiment agree reasonably well with the~ predictions. A series of smaller, clad plate experiments are planned as follow-on work. ;
.10. PRESSURIZED-THERMAL-SHOCK 1ECHNOLOGY '(R. H. Bryan, C. C. Robinson, D. J.' Alexander, and J. A. Keeney-Walker)
-10.1 Introduction The potential utility and feasibility of additional pressurized-
' thermal-shock experiments were . investigated during FY 1988. The investigation involved the consideration of two distinct classes of fracture problems: the effects of stainless steel cladding on crack behavior'and the behavior of cracks in low-upper-shelf welds. Results.
and conclusions were presented to the NRC, and it was recommended that two additional' experiments be performed, PTSE-3 to confirm analytical predictions of the effects of cladding and PTSE-4 to clarify mechanisms of tearing of low-upper-shelf. welds. Preparations for PTSE-4 will begin in FY 1989.
'10.2 An' Experiment on Effects of Cladding on Crack Propagation Effects of cladding on crack propagation were recognized during the conception of the pressurized-thermal-shock experimental program to be ! an uncertain factor in vessel integrity evaluations with the possibility that conclusions drawn from overcooling accident analyses could be made better or worse. Accordingly, in 1981, cladding investigations involving clad plate tests, thermal-shock experiments, irradiation-effects experiments, and a pressurized-thermal-shock experiment were considered as potentially usef ul elements of a research plan. Clad--
plate tests!34,335 and. irradiation studiesl36 were initiated in 1981 and 1982, respectively, as the preliminary steps of the plan. The clad plate. tests were devised as a simple means of determining whether cladding could serve as a potential' deterrent to longitudinal extension of'short surface cracks. The objective of the irradiation studies was to measure the effects of irradiation on cladding toughness as a
< contribution to resolving questions about whether the properties of irradiated cladding could promote or inhibit crack propagation.
The clad plate tests have been completed and showed that the clad-ding hadLsome capacity for limiting crack propagation.137 139 The respective' contributions of fracture toughness and ductility to cladding behavior in these tests are not yet well known. L L Early studies of irradiation effects of austenitic stainless steels showed that toughness can be reduced by fast neutron fluence levels that ( \L 40 _ _ . _ _ _ - ____1________.__._________________________ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . . . _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _
7 l, I l are pertinent to pressurized-water reactor pressure vessels. 140,141 More recent studies were made specifically of the effects of irradiation ; on cladding. In the HSST program 7th Irradiation Series,136 cladding produced by an automated single-wire submerged-arc welding process indi-cated that very little degradation in toughness would be expected in good quality cladding, but toughness was appreciably reduced in highly diluted cladding.142 In the second phase of the 7th Irradiation Series, three-wire series-arc cladding has been investigated. Irradiation of this cladding, which is typical of some commercial light-water reactor materials, resulted in a decrease in toughness, an increase in yield strength, and no substantial change in ultimate tensile strength and ductility.18 With the conclus.on of these preliminary cladding studies, the 2 principal cladding issue of 1981 remains unresolved. It has not yet been determined whether the presumption of an infinitely long flaw is always more conservative than consideration of a short flaw. In one comparison of these alternative assumptions Yang and Bamford analyzed a surface crack in a clad vessel under pressurized-thermal-shock loading in which the variation in flaw shape was taken into account.143 They concluded that, for a particular loading sequence, accounting for shape changes leads to more favorable consequences than are predicted by NRC-accepted proceduresl44 and that including the effects of cladding gives slightly more favorable results. It is now understood that some loading sequences could produce more severe consequences if one starts with a short crack, even though the initial effect may be to inhibit crack propagation. Therefore, the next logical step should include establishment of a credible experimental basis for verifying analytical methods that now exist or may be deve-loped for evaluating propagation of short cracks in cladding. To support the investigation of the utility and feasibility of a clad-vessel experiment, modifications to the O weremadeandtheSHAPEprogramwasdeveloped.gg{USAcomputerprogram The SHAPE program
)
uses stress-intensity factors calculated by OCA/ USA for a pressurized-thermal-shock transient to generate the modes of crack shape changes that are attainable. Three-dimensional finite-element analyses were performed to generate the sets of influence coefficients needed for this task.146,147 Potential experimental configurations for a clad-vessel test (PTSE-3) have been explored with OCA/ USA and SHAPE analyses. Cladding parameters and initial crack geometries that arecompatibleyjththe capabilities of the pressurized-thermal-shock test facility 3 have been identified. The recommended experiment has an initially short axial surface crack that penetrates the cladding layer. In the transient experiment the crack will be made to propagate depthwise and then axially in a predictable succession of shape changes. Tunneling beneath the cladding will occur during the latter phase of propagation. The crack will finally arrest within the test section of the vessel. 41
The vessel for this experiment has already been fabricated, except for cladding. This vessel (originally ITV-7) and the vessel for PTSE-1 (Ref. 131) were prepared by the Babcock & Wilcox Company with identical welded-in-inserts of specially tempered material.iwe The base material for the clad vessel test is fully characterized by small-specimen tests and by the PTSE-1 experiment itself. Therefore, more pertinent tough-ness data are available for planning PTSE-3 than for planning any l previous thick-vessel fracture test. Current plans for PTSE-3 call for cladding with properties representative of irradiated cladding to be applied to the area where the flaw will be located. It is expected that the influence of the cladding will depend more on its strength and duc-tility than on intrinsic toughness. The experiment will be designed to reveal the dependence of crack behavior on these cladding characteris-tics. I 10.3 An Experiment with a Low-Tearing-Resistance Weld j l The second pressurized-thermal-shock experiment, PTSE-2 (Ref. 3), was concerned with behavior of a crack in material with low-tearing resistance, which is one aspect of the NRC pressurized-thermal-shock ^ issue.144 The principal objective of the PTSE-2 experiment was to observe the transitional behavior of steel with low-tearing resistance to examine the interplay of cleavage and ductile modes of f racture. The , experiment produced four phases of stable tearing and two of unstable j tearing.6 It is obvious from the results that the PTSE-2 material had both cleavage and tearing resistances much higher than the limiting value of 220 MPA* /m implied by Section XI of the ASME Code 849 and used in developing the NRC pressurized-thermal-shock rule. 150,151 This con-clusion confirmed that the limiting value may be quite conservative. However, the incremental depths of stable tearing observed in both PTSE-1 (Ref. 131) and PTSE-2 were inconsistent with calculations based on small-specimen J-resistance tests and deformation plasticity analysis. Several factors may contribute to this inconsistency. Two impor-tant possibilities are that a deformation plasticity model may not be adequate, particularly for some phases of PTSE-2, or that structural and material characteristics more complex than a yJ resistance curve rela-tionship may govern tearing behavior. Confirmation of a reliable means of predicting the amount of stable tearing is important, because this type of tearing could in some situa-tions inhibit cleavage crack propagation and in others promote propaga-tion. Furthermore, the influence of warm prestressing on initiation of cleavage fracture is very sensitive to the extent of stable tearing. The experiment (PTSE-4) proposed for clarifying this problem is similar to PTSE-2. However, the initial crack in PTSE-4 will be placed in an insert of low-tearing-resistance weld metal. Certain types of weldr. 9ete ncted by the NRC as having relatively high sensitivity to radiation damage in terms of shifting the brittle-ductile transition temperature and reducing the Charpy V-notch upper shelf energy.is2 Thus the PTSE-4 test will examine flaw behavior in the type of material of greatest concern among reactor pressure vessel s t eel s . 42
To perform the PTSE-4 experiment the PTSE-2 vessel must be repaired with a long insert of weld metal with appropriate properties. Trial weldments have been fabricated with a 2.25Cr-lHo weld wire with suitable composition. Tensile and impact properties of the welds are being determined for a range of tempering treatments so that optimal proce-dures for vessel preparation can be specified.
- 11. PRESSURE VESSEL RESEARCH USERS FACILITY 11.1 Facility Development (W. E. Pennell, C. E. Pugh, and C. C. Robinson)
In FY 1987 ORNL undertook an initiative in concert with the NRC and DOE to establish a Pressure Vessel Research Users' Facility (PVRUF). The facility is to be centered around a complete PWR pressure vessel and is to provide unique research and development (R&D) opportunities for a number of organizations. It is anticipated that domestic and foreign organizations will supplement their pressure vessel integrity research programs by conducting PVRUF studies in several technical areas through cooperative or independent programming. In addition to the NRC and DOE, the domestic interests are anticipated to include the Electric Power Research Institute, the Nuclear Utility Management and Resources Council, LWR component vendors, and utilities. Internal ORNL funds were used to procure the PVRUF vessel from Combustion Engineering, who had fabricated it for use in a four-loop 1100 MWe PWR plant. The basic approach adopted in development of the PVRUF concept was to focus on those PWR pressure vessel R&D needs which require the use of a full scale prototypic test article in order to generate high quality, directly useable program output. Initial development of the program plan for PVRUF was begun this year. Essential elements of this logic are (a) the use of a mission survey to determine the high priority PVRUF R&D tasks; (b) development of functional requirements, conceptual designs, and cost and schedule estimates for PVRUF based on the mission survey results; and (c) preparation of proposal and sponsor interaction leading to a definition of the time phased funding commitment for the project. Actions on the initial phase of the PVRUF mission survey are well advanced. This phase of the survey used appropriate segments of the ORNL technical community to evaluate and rate potential PVRUF R&D tasks. A total of 14 R&D Lasks were included in the review package. The most important categories for task rating were (a) the need for the output from the candidate R&D Lask and (b) the need for PVRUF to generate the required R&D task output. 43
Response to the PVRUP mission survey has been good. A statistical ~ analysis of the numerical R&D Lask-rating data was begun; however, non-destructive evaluation of the vessel to obtain improved flaw density estimates has already been identified as a high priority, early task. (See,Section 11.2). Work on the PVRUF will be continued at ORNL in the future but not as part of the HSST program. At the request of the NRC, PVRUF has been established as an independent program. 11.2 Flaw Density Studies (R. W. McClung, R. A. Cunningham, Jr., K. V. Cook) A plan for the nondestructive evaluation of the PVRUF vessel for flaw density was prepared this year. The PVRUF vessel is located near the south end of Building K-702 at the Oak Ridge K-25 Plant. It has an outside diameter (beltline) of 4.5 m, weighs over 3.5 MN, and is lying on its side (horizontal orientation on its shipping skid). The prelimi-nary plan addressed complications for the nondestructive examination because of the satellite (field) inspection requirement and the vessel orientation. To meet the satellite inspection requirement, support ser-vices and facilities have been added to the site. Since the vessel will ' not be housed in an appropriate facility until funding is arranged, it was concluded that manual ultrasonic scanning procedures will be used while the PVRUF is in the temporary location. Oncu the vessel is inside and in a normal vertical position, it will be possible to apply commer-cial or advanced automated in-service inspection (ISI) equipment. Phase I studies completed during this period follow: (1) familiar-ization of personnel with drawings and inspection data as available, (2) preparation of vessel for internal inspection, and (3) initiation of the nondestructive evaluation of the cylindrical portion of the PVRUF vessel. In order to implement the manual inspection activities, several modifications were made to render the interior of the vessel accessible and habitable. These modifications include access stairs, platform and double entry doors, movable interior scaf f olding, exterior and interior lighting, air conditioning, electrical heating, water spray exterior cooling feature, and an array of 110- and 220-V electrical general ser-vice receptacles. Also provided is a nearby trailer to provide conven-ient office space. Ultrasonic contact angle-beam search units (transducers) were pro-cured for initial manual inspection of zones contained in the cylindri-cal section of the vessel. The protective coating on the inner (clad) surface was removed from portions of the area in preparation for non-destructive inspection of the cladding for flaws. l 44
?
A small area.near the core support-structures positioning keys was used to demonstrate that a fluorescent penetrant inspection could be performed successfully. This demonstration also established procedures that can be used for penetrant examination. Preparation was begun on detailed written quality assurance (QA) plans. Additional material was also obtained - f rom the vessel f abricator to enhance the nondestructive examinations. This includes the seven ori-ginal NDE calibration blocks for the vessel and approximately 6 linear 'l meters of fabrication weld seam removed from a comparison vessel.
- 12. SHIpp1NC CASK MATERIAL EVALUATIONS (J. C. Merkle)
In FY 1987, the NRC added a task to the HSST program, the primary objective being to evaluate the suitability of candidate materials for use as primary structural materials in nuclear spent-fuel shipping casks. Further, a near-term objective was to concentrate evaluation on the use of nodular cast iron (NCI) for such an application. A meet.ing of the HSST Expert Panel on Shipping Casks was held at the Oak Ridge National Laboratory late in FY 1987. The purpose of the meeting was to consider the proposed use of NCI as a primary structural material for nuclear spent-fuel shipping casks. Within the information available, it was determined that there were no overriding reasons for NCI to be disqualified as a candidate structural material for spent fuel shipping casks at the current time, but that several factors pertaining to its use needed substantially more study before a regulatory decision could be made. Based on the recommendations of the expert panel and the interests of the NRC, detailed plans were formulated this past year for the continuation of the materials evaluation for spent fuel shipping casks. The bulk of this work would be to perform an analytical study of the potential for failure of casks fabricated from materials subject to frangible failure modes, specifically NCI and ferritic steel forgings. This study would be similar to that performed by Lawrence Livermore National Laboratory which examined lead-stainless steel casks,153 but would also incorporate relevant flaw-size and fracture-toughness fre-quency distributions, as well as explicitly examine the dependency of failure on the effectiveness of cask impact limiters. However, this work was halted, af ter the detailed planning phase, in accordance with revised programmatic guidance from the NRC. REFERENCES I
- 1. W. R. Corwin, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. l Lab., Heavy-Section Steel Technology Program Semlann. Prog. Rep. l April-September 1987, USNRC Reporg, NUREC/CR-4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2), April 1988 45 !
l
7
- 2. W.'R. Corwin, Martin Marietta Energy Systems, Inc., Oak Ridge Nati. Lab. , Heavy-Section Steel Technology Program Semiann. Prog.
Rep. October 1987-March 1988, USNRC Repogt, NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988. 1 l
- 3. .R. H. Bryan et al., Martin Marietta Energy Systems, Inc., Oak ~I Ridge Nat1. Lab. , Pressurized-Thermal-Shock Test' of 6-in. -Thick -
Pressure Vessels. PTSE-2: Investigation of Low Tearing Resis-tance and warm Prestrgssing, USNRC Report, NUREC/CR-4888 (ORNL- j 6377), December 1987 i
- 4. D. B. Barker, R. Chona, W. L. Fourney, and C. R. Irwin, University of Maryland for Martin Marietta Energy Systems, Inc., Oak Ridge Nat1. Lab., A Report on the Round Robin Program Conducted to Evaluate the Proposed ASTM Standard Test Method for Determining the Plane Strain Crack Arrest Fracture Toughness, K of Ferritic NUREC-CR-4996(ORNL/Sub/79b,778/4),
Materials, USN,RC Report, January 1988.
- 5. K. V. Cook and R. W. McClung, Martin Marietta Energy Systems, ;
Inc. , Oak Ridge Nat1. Lab., Flaw Density Examinations of a Clad \ USNRC Rep Boiling NUREG/CR-4860, Water Reactor Rev. Pressure Vessel Segment,1 (ORNL/TM-10364/R1), February
- 6. R. deWit et al., National Bureau of Standards for Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Wide-Plate Crack- ,
Arrest Testing: A Description and Discussion of the First Two l Wide-Plate Tests and the Results of Six, Full Thickness, Bend Bar Tests, NBSIR 87-3627, February 1988.*
- 7. L. F. Miller, C. A. Baldwin, F. W. Stallman, and F. B. Kam, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Neutron Exposure Parameters for the Metallurgical Test Specimens in the Fif th Heavy-Section Steel Technology Irradiation Series Gapsules, USNRC Report, NUREC/CR-5019 (ORNL/TM-10582), March 1988.
- 8. W. H. Bamford, Westinghouse Electric Co. for Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., A Summary of Environmentally Assisted Crack-Growth Studies Performed at Westinghouse Electric Corporation: Unde.r Funding from the Heavy-Section Steel Technology Programy USNRC Report, NUREC/CR-5020 (ORNL/Sub/82-21598/1), May 1988
- 9. J. J. McCowan, R. K. Hanstad, and K. R. Thoms, Martin Marietta Energy Systems, inc., Oak Ridge Natl. Lab., Characterization of Irradiated Current-Practice Welds and A533 Grade B Class 1 Plate for Nuclear Pressure vessel Servige, USNRC Report, NUREG/CR-4880, Vol. 1 (ORNL-6484/V1), July 1988.
- 10. J. J. McCowan, R. K. Nanstad, and K. R. Thoms, Martin Marietta i Energy Systems, Inc., Oak Ridge Natl. Lab., Characterization of Irradiated Current-Practice Welds and A533 Grade B Class 1 Plate for Nuclear Pressure vessel Servige, USNRC Report, NUREG/CR-4880, Vol. 2 (ORNL-6484/V2), July 1988.
46
- 11. - S. K. Iskander, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., ORNL Foreign Trip Report, ORNL/FTR-2759, November 1987.
i', 12. B. R. Bass, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., ORNL Foreign Trip Report, ORNL/FTR-2780, December 1987.
- 13. C. E. Pugh and C. S. Kramer, Martin Marietta Energy Systems, Inc.,
Oak Ridge Natl. Lab., ORNL Foreigh Trip Report, ORNL/FTR-2907, June 1988.
- 14. B. R. Bass, C. E. Pugh, J. Keeney-Walker, R. J. Dexter, P. E.
O'Donoghue, and C. W. Schwartz, " Evaluation of Viscoelastic Frac-ture Criteria and Analysis Methods," pp. 59-92 in Proceedings of the Fifteenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, October 26-29, 1987, ProceedingNUREC/CP-0091,Vol.2(Feb.1988).ySNRCConference
- 15. R. H. Bryan et al., " Performance of Low-Upper-Shelf Material Under Pressurized-Thermal-Shock Loading (PTSE-2)," pp. 41-58 in Proceedings of the Fif teenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, October 26-29, 1987, USNRC Conference Proceeding NUREC/CP-0091, Vol. 2 (Feb. 1988).*
- 16. K. V. Cook and R. M. McClung, " Detection and Characterization of Flaws in Segments of Light-Water Reactor Pressure Vessels," pp.
, 139-54 in Proceedings of the F1f teenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, October 26-29, 19g7, USNRC Conference Proceeding NUREC/CP-0091, Vol. 2 (Feb. 1986>.
- 17. J. G. Merkle, " Constraint and Strain Rate Effects in Fracture Toughness Testing," pp. 5-16 in Proceedings of the Fifteenth Water 4 Reactor Safety Information Meeting, Gaithersburg, Maryland, October 26-29, 1987 USNRC Conference Proceeding NUREC/CP-0091, Vol. 2 (Feb. 1988) 4
- 18. F. M. Haggag, W. R. Corwir., D. J. Alexander, and R. K. Nanstad,
" Effects of Irradiation on Strength and Toughness of Commercial ;
LWR Vessel Cladding," 1,p. 177-94 in Proceedings of the Fifteenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, October 26-29, 1987 USNRC 2 Conference Proceeding NUREC/CP-0091, Vol. 2 (Feb. 1988)."
- 19. S. K. Iskander et al., " Effects of Commercial Cladding on the Fracture Behavior of Pressure vessel Steel Plates," pp. 123-38 in Proceedings of the Fifteenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, October 26-29, 1987, USNRC Conference Proceeding NUREC/CP-0091, Vol. 2 (Feb. 1988).*
- 20. M. F. Kanninen et al . , Viscoelastic-Dynamic Analyses of Small-Scale Fracture Tests to Obtain Crack Arrest Toughness Values for PTS Conditions," pp. 93-112 in Proceedings of the Fifteenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, Octo-ber 26-29, 1987, USNRC Conference Proceeding NUREC/CP-0091, Vol. 2 (Feb. 1988)."
47 1 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - - _ _ _ _ 1
- 21. D J. Naus, B. R. Bass, J. Keeney-Walker, R. J. Fields, R. deWit, !
and S. R. Low III, " Summary of HSST Wide-Plate Crack-Arrest Tests and Analyses," pp.17-40 in Proceedings of the Fif teenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, Octo-ber 26-29, 1987, USNRC Conference Proceeding NUREC/CP-0091, Vol. 2 L (Feb. 1988).~
- 22. C. W. Schwartz and B. R. Bass, " Evaluation of the Presence of Con-straint in Crack Run/ Arrest Events." op. 113-22 in Proceedings of the Fif teenth Water Reactor Safets *sformation Meeting, . ,
Gaithersburg,. Maryland, October 26-29, 1987, 1 ProceedingNUREC/CP-0091,Vol.2(Feb.1988).ySNRCConference !
- 23. J W. Dally and D. B. Barker, "A Method to Heasure Crack Initia- ;
tion Toughness in Steel at Very High Loading Rates," pp. 59-65 in Proceedings of the Fall 1987 Society for Experimental Mechanics Meeting, Savannah, Georgia, October 25-28,'1987.
- 24. C. W. Schwartz, H. C. Lee, and B. R. Bass, " Dynamic Fracture Prop-agation Relations Inferred from Wide Plate Crack Arrest Tests of !
A333B Steel," pp. 66-71 in Proceedings of the Fall 1987 Society for Experimental Mechanics Meeting, Savannah, Georgia, October 25-28, 1987,
- 25. J. J. McCowan and R. K. Nanstad, "A Statistical Analysis of Frac-ture Toughness of Irradiated Low-Alloy Steel Plate and Welds," pp.
569-89 in Influence of Radiation on Material Properties: 13th International Symposium (Part II), ASTM STP 956, F. A. Carner, C. H. Henager, Jr., and N. Igata, Eds., American Society for Testing and Materials, Philadelphia, Pa., 1987. !
- 26. R. J. Sanford and J. W. Dally, " Strain Cage Methods for Measuring the Opening-Mode Stress Intensit K y , Experimental Mechanics 27(4), 381-87 (1987).$y Factor,
- 27. X. J. Zhang, R. W. Armstrong, and C. R. Irwin, " Cleavage Fractur-ing in the Upper Transition Temperature Range for Pressure Vessel Steel s," p. 242 in Materials Week '87, Extended Abstract, ASH International, Metals Park, Ohio, 1987.T
.1
- 28. B. R. Bass, C. E. Pugh, and C. W. Schwartz, "An Evaluation of Vis-coplastic-Dynamic Fracture Criteria and Analysis Methods in Crack Run/ Arrest Events," pp. 9.viii.1-2 in Proceedings of International Conference on Computational Engineering Science, Atlanta, Georgia, April 10-14, 1988.
- 29. R. J. Dexter, "The Significance of Crack-Tip Characterization in Dynamic Fracture," pp. 9.vi.1-4 in Proceedings of International ,
Conference on Computational Engineering Science, Atlanta, Georgia, April 10-14, 1988.
- 30. C. W. Schwartz and B. R. Bass, " Computational Finite Element-Boun-dary Element Analysis for Numeric Fracture Problems," pp. 9.ii .1-4 48
>>c in Proceedings of International Conference on Computational Engineering Science, Atlanta, Georgia, April 10-14, 1988.
- 31. .B. R. Bass, C. E. Pugh, and J. Keeney-Walker, " Computational Meth-ods for Viscoelastic Dynamic Fracture Mechanics Analysis," Vol. 3, l pp. 127-36 in Proceedings of the ASME 1988 International Computers l in Engineering Conference, July 31-August 4, 1988, i
- 32. A. R. Rosenfield, P. N. Mincer, and C. W. Marschall, "High Temperature Crack-Arrest Toughness Measurements Using Compact ]
Specimens," pp. 73-85 in Fracture Mechanics: Eighteenth Symposium, ASTM STP 945, D. T. Read and R. P. Reed, Eds., American Society for Testing and Materials, Philadelphia, Pa., 1988. ?
- 33. D. B. Barker et'al., "A Method for Determining the Crack Arrest ,
Fracture Toughness of Ferritic Materials," pp. 569-96'in Fracture 'l Mechanics: Nineteenth Symposium, ASTM STP 969, American Society j for Testing and Materials, Philadelphia, Pa., 1988.
- 34. B. R. Bass et al., " Fracture Analyses of Heavy-Section Steel Tech-nology Wide-Plate Crack-Arrest Experiments," pp. 691-723 in Fracture Mechanics: Nineteenth Symposium, ASTM STP 969, American Society for Testing and Materials, Philadelphia, Pa., 1988.
- 35. R. H. Bryan et al., " Pressurized Thermal Shock Experiments with Thick Vessels," pp. 767-86 in Fractures Nechanics: Nineteenth Symposium, ASTM STP 969, American Society for Testing and Materials, Philadelphia, Pa., 1988.
- 36. R. D. Cheverton, S. K. Iskander, and D. C. Ball, " Review of Pressurized-Water Reactor-Related Thermal-Shock Studies," pp. 752-66 in Fracture Mechanics: Nineteenth Symposium, ASTM STP 969, American Society for Testing and Materials, Philadelphia, Pa.,
1988.
- 37. R. deWit, S. R. Low III, and R. J. Fields, " Wide-Plate Crack Arrest 'J e s t i ng : Evolution of Experimental Procedures," pp. 679-690 in Fracture Mechanics: Nineteenth Symposium, ASTM STP 969, American Society for Testing and Materials, Philadelphia, Pa.,
1988.
- 38. B. R. Bass et al., " Late-Event Viscoelasticity in Wide-Plate Crack-Arrest Tests," international Journal of Pressure Vessels and Piping, 31, 325-48 (1988).i
- 39. C. E. Pugh et al., " Wide-Plate Crack-Arrest Tests Utilizing a Pro-totypical Pressure Vessel Steel," International Journal of Pressure Vessels and Piping, 31, 165-85 (1988).$
- 40. K. V. Cook and R. W. McClung, " Flaw Density Examinations of a Clad Boiling Water Reactor Pressure Vessel Segment," Nuclear Engineer-ing and Design, 108, 211-19 (1988).%
- 41. J. W. Bryson, " Finite Element Fracture Analysis on a Microcompu-ter," presented at the ASME Pressure Vessels and Piping Division Fall Conference, Knoxville, Tenn., October 23, 1987.
I 49 = _ _ _ _ _ - .
i f42. C.E. Pugh, "An Overview of the Pressure' Vessel Research Users' Facility,"' presented.at the ASME Pressure Vessels and Piping Diva- J son Fall Conference, Knoxville, Tenn., October 23, 1987. 1
- 43. D. A. Steinert, " Dynamic Fracture and Craphic Analysis," presented I at the ASME Pressure Vessels and Piping Division Fall Conference, Knoxville, Tenn., October 23', 1987.
- 44. B. R. Bass, " A' Comparison of Dynamic Viscoelastic Fracture Cri-teria," presented lat the Fall 1987 Society for Experimental Mechanics Meeting, Savannah, Ca., October 25-28, 1987.
- 45. J. W.' Dally and D. B. Barker, "A Method to 'ansnre Crack Initia-tion Toughness in Steel at Very High Loading kates," presented at, the Fall 1987 Society for Experimental Mechanics Me'4ing, Savan-nah, Ca., October 25-28, 1987.
- 46. C. W. Schwartz, !!. C. Lee, and B. R. Bass, " Dynamic Fracture Propagation Relations Inferred from Wide-Plate Crack Arrest Tests of A533B Steel," presented at the Fall 1987 Society for
. Experimental Mechanics Meeting, Savannah, Ca., October 25-28, 1987.
- 47. R. K. Nanstad, " Reactor Pressure Vessel Materials," presented as an invited talk at the Conference on Materials for Nuclear Appli-cations, University of Missouri-Rolla on October 26, 1987.
- 48. B. R. Bass, C. E. Pugh, J. Keeney-Walker, R. J. Dexter, P. E.
O'Donoghue, and C. W. Schwartz, " Evaluation of Viscoelastic Frac-ture Criteria and Analysis Methods," presented at the Fifteenth Water Reactor Safety Information Meeting, Caithersburg, Md., Octo-ber 26-29, 1987.
- 49. R. II. Bryan et al . , "Perf ormance of Low-Upper-Shelf Material under Pressurized-Thermal-Shock Loading (PTSE-2)," presented at the Fif-teenth Water Reactor Safety Information Meeting, Caithersburg, Md., October 26-29, 1987.
- 50. K. V. Cook and R. M. McClung, " Detection and Characterization of Flaws in Segments of Light-Water Reactor Pressure Vessels," pre-sented at the Fifteenth Water Reactor Safety Information Meeting, Caithersburg, Md., October 26-29, 1987.
- 51. J. W. Dally and C. R. Irwin, " Dynamic Fracture Initiation Stud-ies," presented at the Fifteenth Water Reactor Safety Information Meeting, Gaithersburg, Md., October 26-29, 1987.
- 52. F. M. Haggag, W. R. Corwin, D. J. Alexander, and R. K. Nanstau,
" Effects of Irradiation on Strength and Toughness of Commercial LWR Vessel Cladding," presented at the Fifteenth Water Reactor Safety Information Meeting, Caithersburg, Md., October 26-29, 1987.
- 53. S. K. Iskander et al., " Effects of Commercial Cladding on the Fracture Behavior of Pressure Vessel Steel Plates," presented at the Fifteenth Water Reactor Safety Information Meeting, Caithers-burg, Md., October 26-29, 1987.
50 I
- 54. M. F. Kanninen et al., " Viscoelastic-Dynamic Analyses of Small- ;
Scale Fracture Tests to Obtain Crack Arrest Toughness Values for l PTS Conditions," presented at the Fifteenth Water Reactor Safety Information Meeting, Caithersburg, Md., October 26-29, 1987. j i
- 55. R. K. Nanstad, " Effects of Irradiation on K Curves for liigh- j ic Copper Welds," presented at the Fifteenth Water Reactor Safety j Information Meeting, Caithersburg, Md., October 26-29, 1987.
- 56. Le J. Naus, B. R. Bass, and R. J. Fields, " Summary of HSST Wide-Plate Crack-Arrest Tests and Analyses," presented at the Fifteenth Water Reactor Safety Information Heating, Caithersburg, Md., Octo- I ber 26-29, 1987.
- 57. C. W. Schwartz and B. R. Bass, "An Evaluation of the Presence of
-Con s t ra i nt in Crack /Run Arrest Events," presented at the Fifteenth Water Reactor Safety Information Meeting, Caithersburg, Md., Octo-ber 26-29, 1987.
- 58. B. R. Bass, "An Overview of Ongoing Large-Scale Crack-Arrest Tests l
of Pressure Vessel Steels in The U.S.A.," presented at the Nordic Materials Research Seminar, Copenhagen, Denmark, November 25-26, 1987; on December 1, 1987, at the Royal Institute of Technology, Stockholm, Sweden.
- 59. R. K. Nanstad, " Comparison of Charpy Impact and Fracture Toughness Transition Temperature Shifts Due to Irradiation (HSST Irradiation Series 4 and 5)," presented at a minisymposium, "Is the Charpy Shift Representative of the Fracture-Toughness Shift?" sponsored by the E10.02 Task Croup on Fracture Toughness, Albuquerque, N.M.,
January 1988.
- 60. R. J. Fields, " Plastic Zone Formation Around an Arresting Crack,"
presented at the First Symposium of the International Union of Theoretical and Applied Mechanics on Recent Advances in Nonlinear Fracture Mechanics, California Institute of Technology, March 14-16, 1988.
- 61. B. R. Bass, " Development of Dynamic Inelastic Fracture Analysis Methods for Light-Water Reactor Applications," presented at the Engineering Technology Division Information Meeting, Oak Ridge Nat l . Lab. , Ma rch 17, 1988.
- 62. W. R. Corwin, " Behavior of Low-Upper-Shelf Toughness Steels in I Light-Water Reactor Pressure Vessels," presented at the Engineer-ing Technology Division Information Meeting, Oak Ridge Natl. Lab.,
March 17, 1988.
- 63. C. C. Robinson, " Nuclear Reactor Pressure Vessel Jntegrity," pre-sented at the Engineering College, Arkansas State University, March 31, 1988.
51
l'
- 64. B. R. Bass, C. E. Pugh, and C. W. Schwartz, "An Evaluatn.of Vis-coplastic-Dynamic Fracture Criteria and Analysis Methods in Crack Run/ Arrest Events," presented at the International Conference on Computational Engineering Science, Atlanta, Ca., April 10-14, 1988.
- 65. R.'J. Dexter, "The Significance of Crack-Tip Characterization in Dynamic Fracture," presented at the International Conference on Computational Engineering Science, Atlanta, Ca., April 10-14, 1988.
- 66. C. W. Schwartz and B. R. Bass, "Comptoational Finite Element-Boun-dary Element Analysis for Numeric Fracture Problems," presented at the International Conference on Computational Engineering Science, Atlanta, Ca., April 10-14, 1988.
- 67. C. E. Pugh, J. C. Merkle, R. H. Bryan, W. R. Corwin, and B. R.
Bass, "Some Lessons About Fracture-Hechanics Methods from Heavy-Section Steel Technology Program's Large-Scale Tests," presented at the Joint Meeting of ASTM Task Croups E24.08 and E24.06.02, Reno, Nev., April 27, 1988.
- 68. C. E. Pugh, "A Few Observations on the Consistency of Crack-Arrest Toughness Measurements for Reactor Pressure Vessel Steels," pre-sented at the Joint Meeting of ASTM Task Groups E24.08 and E24.01.06, Reno, Nev., April 27, 1988.
- 69. C. E. Pugh, "Some Observations from HSST Thermal-Shock and Pres-surized Thermal-Shock Experiments," presented at CSNI Fracture Assessment Group Meeting, Stuttgart, FRG, May 24, 1988.
- 70. C. E. Pugh, D. J. Naus, and B. R. Bass, " Crack Arrest Behavior of Reactor Pressure Vessel Steels at High Temperatures," presented at IAEA Specialists' Meeting, Stuttgart, FRG, May 25-27, 1988.
- 71. C. E. Pugh et al., " Evaluation of Fracture Models Through Pressur-ized-Thermal-Shock Testing," presented at IAEA Specialists' Meet-ing, Stuttgart, FRC, May 25-27, 1988.
- 72. A. Cilat, " Status of High Strain-Rate Testing for A533B Steel at Ohio State University," presented at the Fourth Annual HSST Pro-gram Workshop on Dynamic Fracture and Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
73.. J. H. Giovanola, " Viscoelastic Stress-Strain Characterization of A533B Steel," presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture and Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 74. R. J. Dexter, " Development of Bodner-Partom Constitutive Model,"
presented at the Fourth Annual USST Program Workshop on Dynamic l 52
.\
Fracture and Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 75. B. R. Bass, " Development of Other Constitutive Models," presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture and Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 76. S. J. Hudak, "Small-Specimen Testing at Southwest Research Insti-tute," presented at the Fourth Annual HSST Program Workshop on
-Dynamic ~ Fracture Crack-Arrest Technology, Caithersburg, Md.,
June 1-2, 1988.
- 77. B.'R. Bass and A. Pini, "Small-Specimen Testing at ORNL," pre-sented at the Fourth Annual HSST Program Workshop on Dynamic Frac-ture Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 78. J. W. Dally, "Small-Specimen Testing at the University of Mary-land," presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture Crack-Arrest Technology, Caithersburg, Md.,
June 1-2, 1988.
- 79. R. J. Fields, " Wide-Plate Testing at National Bureau of Stand-ards," presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 80. D. J. Naus, " Summary of Wide-Plate Test Results," presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 81. M. F. Kanninen, R. J. Dexter, and P. E. O'Donoghue, " Development of VISCRK and Applications to Small- and Large-Specimen Tests,"
presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 82. B. R. Bass, " Development of ADINA/VPF and Applications to Small-and Large-Specimen Tests (Elastic and Viscoelastic)," presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture Crack-Arrest Technology, Caithersburg, Md., June 1-2, 1988.
- 83. C. W. Schwartz, " Investigation of Constraint and Yielding in the Crack-Tip Region," presented at the Fourth Annual HSST Program Workshop on Dynamic Fracture Crack-Arrest Technology, Caithers-burg, Nd., June 1-2, 1988.
- 84. F. M. Haggag, " Tensile and Charpy Impact Behavior of an Irradiated 2 Three-Wire Series-Arc Stainless Steel Cladding," presented at the 14th ASTM International Symposium on Effects of Radiation on Materials, Andover, Mass., June 27, 1988.
53
- 85. R. K. Nanstad', " Effects of Irradiation on KIc Curves for liigh-Copper Welds," presented at the 14th ASTM International Symposium -l on Effects of Radiation on Materials, Andover, Mass., June 27, ;
1988.. 1
'86. R. D. Cheverton et al., "An Embrittlement Rate Effect Deduced from liFIR That May. Impact LWR Vessel Support Life Expectancy," pre- -
sented at the American Nuclear Society Nuclear Power Plant Meet- , ing, Snowbird, Utah, July 31-August 3, 1988. ! l
'87. R. K. Nanstad, " Irradiation Effects and Associated Advances in Ferritic Pressure Vessel Steels!' presented at the TMS-AIME Pall Meeting and World Materials Congress, Chicago, Ill., September 27, 1988.
- 88. J. Keency-Walker and B. R. Bass, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., " Analysis of a Stubbed-Panel Crack-Arrest Specimen," pp. 8-17 in Heavy-Section Steel Technology j Program Semiann. Prog. Rep. April-September 1987, USNRC Report NUREC/CR-4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2), April 1988.,
- 89. D. J. Naus et al., Martin Marietta Energy Systems, Inc., Oak Ridge Nat1. Lab. , Crack Arrest Behavior in SBN Wide Plates of Quenched and Tempered A533B Steel Tested Under Nonisothermal Condf tions, USNRC Report, NUREC/CR-4930 (ORNL-6388), September 1987.
- 90. M. F. Kanninen, Martin.Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., "Elastodynamic and Viscoelastic-Dynamic Fracture Mechanics," pp. 21-31 in Heavy-section Steel Technology Program Semiann, Prog. Rep. October 1987-March 1988, USNRC Report ,
NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988.,
- 91. R. J. Dexter et al., " Dynamic-Viscoelastic Analysis and Small-Specimen Experimental Methods for the Study of Fracture in A533B Steel," in Proceedings of the Fourth International Conference on Numerical Methods in Fracture Mechanics, San Antonio, Texas, March 23-27, 1987.
- 92. W. L. Fourney et al., Martin Marietta Energy Systems, Inc., Oak i Ridge Natl. Lab., " Investigation of Damping and Cleavage-Fibrous Transition in Reactor-Crade Steel at the University of Maryland,"
pp. 32-61 in Heavy-Section Steel Technology Program Semiann. Prog. \ Rep. October 1987-March 1988, USNRC Repopt, NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988.
- 93. W. L. Fourney et al., Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., " Investigation of Damping and Cleavage-Fibrous Transition in Reactor-Crade Steel at the University of Maryland,"
pp. 32-43 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. October 1987-March 1988, NUREg/CR-4219, Vol . 5, No.1 (ORNL/TM-9593/V5&N1), August 1988
- 94. C. W. Schwartz, Martin Marietta Energy Systems, Inc., Oak Ridge
) Natl. Lab., " Dynamic Fracture Propagation Relations Inferred from I l-54
.l
;-)
I
]
l l WP-1 Test Series," in Heavy-section Steel Technology' Program i' Semiann. Prog. Rep. April-September 1987, USNRC Repor L 4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2), April 1988 }, NUREC/CR-il
- 95. C. T. Hahn and A. R. Rosenfield, " Plastic Flow in the Locale of Notches and Cracks in Fe-3Si Steel Under Conditions Approaching Plane Strain," Report No. SSC-191, prepared for the Ship Struc-tures Committee, U.S. Coast Guard (1968). { '
- 96. C..H. Popelar, informal presentation of analysis for crack tunnel-ing effects, Fourth HSST Program Workshop on Dynamic Fracture and Crack Arrest, Caithersburg, Md., June 1-3, 1988.
- 97. C. H. Popelar, "A Quasi-Three-Dimensional Model for Crack Propaga-tion in Materials that Exhibit Extensive Crack Tunneling," pp.
753-66 in Proceedings of the Fourth International Conference on Numerical Methods in Fracture Mechanics, San Antonio, Texas, March 23-27, 1987.
- 98. E. Smith, "The Restraining Effect of Ductile Ligaments on Plane Strain Crack Propagation and Arrest in Ferritic Steels," Engineer-ing Fracture Mechanics, 19 (4), 601-04 (1984).T
- 99. H. Tada, P. C. Paris, and C. R. Irwin, The Stress Analgsis of Cracks Handbook, Del Research Corporation, Hellertown, Pa., 1973.
100. M. F. Kanninen et al., Martin Marietta Energy Systems, Inc., Oak Ridge Natl . Lab., " Viscoelastic Characterization of A533B Steel," in Heavy-Section Steel Technology Program Semiann. Prog. Rep. April-September 1986, USNRC Report, NUREG/CR-4219, Vol. 3, No. 2 (ORNL/TM-9593/V3&N2), December 1986. 101. J. H. Giovanola and R. W. Klopp, SRI International, " Viscoelastic Stress-Strain Characterization of A533B Class 1 Steel," pp. 62-70 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. April-September 1987, NUREC/CR-4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2), April 1988.* 102. A. Gilat, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., "High Strain Rate Testing of A533 Crade B Class 1 Steel at Various Temperatures," pp. 61-71 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. October 1987-March 1988, USNRC Reporg', NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988. 103. S.-J. Chang, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., "0RNL Unified Inelastic Deformation Model," pp. 72-74 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. October 1987-March 1988, USNRC Report, NURgC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988." 104. C. E. Pugh and D. N. Robinson, "Some Trends in Constitutive Equa-tion Model Development for High-Temperature Behavior of Fast- l Reactor Structural Alloy %,s " Nuclear Engineering and 1>esign, 48 (1), 269-76 (June 1978). 55
l. I, K. Hornberger et al., " Numerical Integration and Implementation of 105. . Visco plastic Models into Finite Element Codes," pp. 477-90 in l
. Proceedings of Third International Conference on Computational l Plasticity,, Barcelona, Spain, April 6-10, 1987, Pineridge Press, j J
Swansea, U.K. l 106. K. J. Rathe, ADINA - A Finite Element-Program for' Automatic Dyna-mic Incremental Nonlinear Analysis, Report AE 84-1, Massachusetts Institute of Technology, Cambridge, Mass., December 1984. 107. .B. R. Bass et al., " Summary of Viscoelastic-Dynamic Fracture l Analyses of the WP-1 Series of Wide-Plate Tests," pp. 9-20 in .! Heavy-Section Steel Technology Program Semiann. Prog. Rep. l October 1987-March 19BB, NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM- 4 9593/V5&N1), August 1988.' 108. S. N. Atluri, T. Nishioka, and M. Nakagaki, " Incremental Path Independent Integrals in Inelastic and Dynamic Fracture Mechan-ics," Engineering Fracture Mechanics, 20(2), 209-44 (1984).*' 109. J. C. Sheu, " Dynamic Elastic-Viscoelastic Crack Crowth," Ph.D. ! Dissertation, School of Engineering, Ohio State University, 1988. i 110. B. R. Bass et al., Martin Marietta Energy Systems, Inc., Oak Ridge ; Natl. Lab., "Inclastic Fracture Model Development," pp. 10-14 in : Heavy-Section Steel Technology Program Semiann. Prog. Rep. April-September 19FB, USNRC Reporg, NUREC/CR4219, Vol. 5, No. 2 (ORNL/TM-9593/V5&N2), March 1989 111. T. Nishioka, " Finite Element Analysis of the T*-Integral in Non-linear Dynamic Fracture Problems," pp. 9.v.1-4 in Proceedings of International Conference on Computational Engineering Science, Atlanta, Georgia, April 10-14, 1988. 112. J. C. Thesken and P. Gudmundsson, " Application of Variable Order Singular Element to Dynamic Fracture Mechanics," Computational Nechanics, 2, 307-16 (1987).i 113. J. E. Akin, "The Generation of Elements with Singularities," International Journal of Numerical Nethod Engineering 10, 1249-59 (1976).T I 114. W. R. Corwin, Martin Marietta Energy Systems, Inc., Oak Ridge l Natl. Lab., " Stainless Steel Cladding Investigations," pp. 70-79 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. October 19BS-March 1986, USNRC Rep
- ort , NUREC/CR-4219, Vol. 3, No.
1 (ORNL/TM-9593/V3&N1), June 1986. j d 1 115. S. K. Iskander et al., Martin Marietta Energy Systems, Inc., Oak { Ridge Natl . Lab. , " Post test Material Characterization of Clad ) Plate Materials," pp. 115-26 in Heavy-Section Steel Technology 1 l Program Semiann. Prog. Rep. April-September 1987, USHRC Report NUREC/CR-4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2), April 1988.g l 1 56 :
l i 136. W. R. Corwin et al., Martin Marietta Energy Systems, Inc., Oak
' Ridge' Natl . Lab. , " Crack-Arrest Test Resul ts .in Support of Wide Plate Testing," pp. 54-56 in Heavy-Section Steel Technology Prod ]
gram Semiann. Prog. Rep. April-September 1985, USNRC R NUREC/CR-4219, Vol. 2 (ORNL/TM-9593/V2), January 1986.gport, 117. A. R. Rosenfield et al., Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., " Crack Arrest Studies at Battelle Columbus," pp. 102-109 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. April-September 1984, USNRC NUREC/CR-3744, Vol. 2 (ORNL/TM-9154/V2), December 1984.geport, 118. R. K. Nanstad et al., Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., " Low-Upper-Shelf Material Characterization," pp. i 51-54 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. October 1986-March 1987, USNRC Repogt, NUREC/CR-4219, Vol. 4, 3 No. 1 (ORNL/TM-9593/V4&N1), August 1987. l 119. J. R. Hawthorne, Studies of Radiation Etfects and Recovery of Notch Ductility of Pressure vessel Steels, British Nuclear Energy Conference, Iron and Steel Institute, London, November 30, 1960. 120. L. E. Steele, J. R. Hawthorne, C. Z. Serpan, Jr., E. P. Klier, and H. E. Watson, Irradiated Materials Evaluation and Reactor Pressure Vessel Surveillance for the Army Nuclear Power Program, NRL Memo-randum Report 1644, September 1, 1965. 121. J. R. McWherter, R. E. Schappel, and J. R. McGuffey, Union Carbide Corp. Nuclear Div., HFIR Pressure Vessel and Structural Corpponents Material Surveillance Program, ORNL/TM-1372, January 1966.' 122. R. D. Cheverton, J. C. Merkle, and R. K. Nanstad, Eds., Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Evaluation of HFIR Pressure-Vessel Integrity Considering Radiation Embrittle-ment, ORNL/TM-10444, April 1988.* 123. D. S. Selby et al., Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Pressurized Thermal Shock Evaluation of the H. B, Robinson Unit 2 Nuclear Power Plant NUREG/CR-4183 (ORNL/TM-9567), September 1985.* , USNRC Report, 124. R. D. Cheverton and D. C. Ball, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., OCA-P, A Deterministic and Probab-11istic Fracture-Mechanics Code for Application to Pressure Ves-sels, USNRC Report, NUREC/CR-3618 (ORNL-5991), May 1984.* 125. L. C. Usu, General Electric Corp., An Analytical Study of BriLLle Fracture of GE-BWR Vessels Subjected to the Design Basis Accident, NEDO-10029, June 1969. 126. C. D. Whitman, C. C. Robinson, and A. W. Savolainen, Eds., Union ! Carbide Corp. Nuclear Div., Oak Ridge Natl. Lab., Technology of I 57 I
\
l I
Steel Pressure Vessels for Water-Cooled Nuclear Reactors, ORNL/NSTC-21, December 1967.# 127. P. P. Stancavage, US BUR Vessel Aging Mitigation, NEA/UNIPEDE Specialist Meeting on Life-Limiting and Regulatory Aspects of Core
. Internals and Pressure Vessels, Stockholm, Sweden, paper No. 12, October. 14-16, 1987. i 128. D. J. Naus, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. ;
Lab., " Crack-Arrest Technology", pp. 155-156 in Heavy-Section l Steel Technology Program Semiann. Prog. Rep. April-September 1987,
.USNRC Report y NUREC/CR-4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2),
August 1988." 129. D. J. Naus, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. I Lab., " Crack-Arrest Technology," p. 162 in Heavy-Section Steel l Technology Program Semiann. Prog. Rep. October 1987-March 1988, USNRC Report NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988.3 130. R. H. Bryan, B. R. Bass, S. E. Bolt, J. W. Bryson, J. C. Merkle, R. K. Nanstad, and C. C. Robinson, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Test of 6-in. -Thick Pressure Ves-sels. Series 3: Intermediate Test Vessel V-BA - Tearing Behavior of Low Upper-Shelf Material, USNRC Report, NUREC/CR-4760 (ORNL-6187), May 1987.# i 131. R. H. Bryan, B. R. Bass, S. E. Bolt, J . W. Bryson, D. P. Edmonds, R. W. McCulloch, J. C. Merkle, R. K. Nanstad, C. C. Robinson, K. R. Thoms, and C. D. Whitman, Martin Marietta Energy Systems, Inc. , Oak Ridge Natl. Lab. , Pressurized-Thermal-Shock Test of 6-in. -Thick Pressure Vessels. PTSB-1; Investigation of Warm Pre-stressingandUpper-She1{ Arrest, USNRC Report, NUREC/CR-4106 (ORNL-6135), April 1985. 132. R. D. Cheverton, D. C. Ball, S. E. Bolt, S. K. Iskander, and R. K. Nanstad, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Pressure Vessel Fracture Studies Pertaining to the PWR , Thermal-Shock issue: Experiments TSE-5, TSE-SA USNRC Report, NUREC/CR-4249 (ORNL-6163), June 1985.* , and TSE-6, 1 133. R. D. Cheverton, D. C. Ball, S. E. Bolt, S. K. Iskander, and R. K. Nanstad, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab. , Pressure Vessel Fracture Studies Pertaining to the PWR Thermal-Shock Issue: Experiment TSE-7, USNRC Report, NUREC/CR-4304 (ORNL-6177), August 1985." 134. C. C. Robinson and J. C. Merkle, Union Carbide Corp. Nuclear Div., Oak Ridge Natl. Lab., " Stainless Steel Cladding Investigations - Scoping Studies," pp. 118-123 in Heavy-Section Steel Technologg Program Quarc. Prog. Rep. October-December 1981, MSNRC Report, NUREC/CR-2141, Vol. 4 (ORNL/TM-8252), April 1982." 135. R. K. Nanstad et al., Union Carbide Corp. Nuclear Div., Oak Ridge Natl . Lab. , " Stainless Steel Cladding Investigations," pp. 109-129 in Heavy-Section Steel Technology Program Quart. Prog. Rep. April-June 1982, USNRC Report, NUREC/CR-2751, Vol. 2 (ORNL/TM-8369/V2), December 1982." 58
(t 136. R.' C. Berggren and R. K. Nanstad, Union Carbide' Corp. Nuclear JDiv., . Oak Ridge Natl . Lab. , " Investigation of Irradiated Materials," pp. 78-80 in Heavy-Section Steel Technology Program Quart.' Prog. Rep. Apr11-June 1982, USNRC Report,;NUREC/CR-2751, Vol. 2-(ORNL/TM-8369/V2),. December 1982.*
' ~
137. W. R. Corwin et al.. Martin Marietta Energy Systems, Inc., Oak Ridge .Nat t . Lab. , Effects of Stainless Steel Weld Ovtriay Cladding on the Structural Integrity 'of Flawed Steel Plates in Bending, , Series 1, USNRC Report, NUREC/CR-4015 (ORNL/TM-9390), April 1985. l
.138. .S. K. Iskander et'al., Martin Marietta Energy Systems.,Inc., Oak ; . Ridge Natl. Lab.,'" Crack Arrest Behavior in Clad Plates " pp. 222-42 in Heavy-Section Steel Technology Program Semiann. Prog. Rep.
April-September 1987,_USNRC Repor1, NUREC/CR-4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2), April 1988." 139. S. K.' Iskander et al.,' Martin Marietta Energy Systems, Inc., Oak Ridge Natl . Lab. , " Crack Arrest Behavior- in Clad Plates," pp. 212-226 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. October 1987-March 1988, USNRC Reporp, NUREC/CR-4219; Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988." 140. J. R. Hawthorne and H. E. Watson, " Exploration of the Influence of Welding Variables.on Notch Ductility.of Irradiated Austenitic Stainless Steel Welds," pp. 327-36 in Proceedings of the International Conterence on Radiation Etfects in Breeder Reactor Structural Materials, Scottsdale, Arizona (June 1977) . 141. W. R. Corwin, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Assessment of Radiation Effects Relating to'Neactor Pressure V USNRC Report, NUREC/CR-3671, ORNL-6047, July 1984.gssel Cladding, 142. W. R. Corwin, R. C. Berggren and R. K. Nanstad, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., Charpy Toughness and Tensile Properties of a Neutron-Irradiated Stainless Steel Sub-merged-Arc Weld Cladding Overlag, USNRC Report, NUREC/CR-3927 (ORNL/TM-9309), September 1984 143. C. Y. Yang and W. H. Bamford, " Variable Flaw Shape Analysis for a Reactor Vessel Under Pressurized Thermal Shock Loading," pp. 41-58 in Fracture Mechanics: Seventeenth Volume, ASTM STP 905, J. H. Up erwood, R. Chait, C. W. Smith, D. P. Wi t hern. W. A. Andrews, and J. C. Newman, Eds. American Society for Testing and Materials, Philadelphia, Pa., 1986. 144. NRC Policy Issue, Enclosure A, "NRC Staff Evaluation of Pressur-ized Thermal Shock," SECY-82-465, U.S. Nuclear Regulatory Commiss-ion, Washington, D.C., Nov. 23, 1982. 145. R. H. Bryan, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., " Preliminary Investigations of Future Pressurized-Thermal-Shock Experiments," pp. 233-34 in Heavy-section Steel Technology Program Semiann. Prog. Rep. October 1987-March 1988, USNRC Repo.(t, NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988." 59
J. Keeney-Walker, D. A. Steinert, and J. S. Parrott, Martin 146. Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., " Stress-Intensity Factor Influence Coefficients for Surface Flaws in Pres- i sure Vessels," pp. 235-236 in Heavy-Section Steel Technology Program Semlann. Prog. Rep. October 1987-March 1988, USNRC Repo5L, NUREC/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988. 147. J. S. Parrott, J. Keeney-Walker, and D. A. Steinert, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., " Mesh Conver- l gence Study for PTSE-3 and PTSE-4' Pretest Analysis," pp. 236-244 in Heavy-Section Steel Technology Program Semiann. Prog. Rep. October 1987-March 1988, USNRC Repori, NUREG/CR-4219, Vol. 5, No. 1 (ORNL/TM-9593/V5&N1), August 1988." 148. 11. A. Domian, The Babcock & Wilcox Company, Research and Development Division for Martin Marietta Energy Systems, Inc., Oak. Ridge Natl . Lab. , Vessel V-7 and V-8 Repair and Characterization of-Insert Material, ORNL/Sub/82-52845/1, The Babcock & Wilcox Com-pany, Research and Development Division, Alliance, Ohio, May 1984. 149. " Rules for in-Service Inspection of Nuclear Power Plant Compo-~ nents," ASME Boiler and Pressure Vessel Code, Section XI, American Society of Mechanical Engineers, New York, 1983. 150. Title 10, Code of Federal Regulations, Section 50.61, " Fracture Toughness Requirements for Protection Against Pressurized Thermal Shock Events, July 23, 1985. 151. U.S. Nuclear Regulatory Commission, Regulatory Guide 1.154, Format i and Content of Plant-Specific Pressurized Thermal Shock Safety Analysis Reports for Pressurized Water Reactors, January 1987. 152. R. Johnson, Resolution of the Task A-11 Vessel Materials Toughness Safety Issue, NUREC-0744, Vols. 1 and 2, Rev. 1, October 1982 ,USNRC Report, 153. L. E. Fischer et al., Lawrence Livermore Natl. Lab., Shipping Container Response to Severe Highway and Railway Accident Condigions, USNRC Report, NUREC/CR-4829, Vols. 162, February 1987.
*Available for purchase from National Technical Information Service, Springfield, Virginia 22161.
4A vailable in public technical libraries.
- Copies are available from U.S. Covernment Printing Office, Washington, D.C. 20402. ATTN: Regulatory Guide Account.
60 l l E_.m___ __ __ _ _ _ _ ._ . . . . _ _ _ . __ _ _ _ _ _
I I-l CONTRACT TITLE: STRUCTURAL INTEGRITY OF LIGHT WATER REACTOR PRESSURE BOUNDARY COMPONENTS CONTRACTOR: Materials Engineering Associates, Inc. Lanham, MD 20706-1837 PRINCIPAL INVESTIGATORS: F . J . Lo s s , Program Manager W. H. Cullen J. R. Hawthorne ! A. L. Hiser i D. E. McCabe B. H. Menke J. B. Terrell
, ABSTRACT Experiments that' simulate a flaw embedded in the clad layer of irradiated RPV steel have identified the important material properties that control crack instability during transient conditi.ons. Warm prestress effects' were shown to strongly influence the effective fracture toughness performance of flawed structures. Its impact is !
important to the analysis of accident scenarios and to service life considerations. New data confirm that the elevation in transition temperature (AT) from C y tends to underestimate AT from K , for base metals, although the correspondence at high AT levels (>100 4 C') is much improved over that found previously, J R curve characterization of nuclear piping stccis was completed, with all data contair.ed in the NRC's Piping Fracture Mechanics Data Base. The primary finding is l that thermal-aging of cast stainless steels can result in large reductions in fracture toughness. Characterization of the fracture toughness of a low Cy upper shelf A 302-B plate revealed a severe size dependence on J-R curves such that small specimens, of the ::ype used for survellience, would not yield conservative results. Restalts from tests of Gundremmingen archive material do not confirm the dependence of radiation effects sensitivity on test specimen orientation reported for specimens cut from the vessel itself, thus presenting an anomaly. The first investigation into the effect of PWR environment in fatigue life of piping steels showed that the margin of safety offered by Section III of the ASME Code is nearly completely used up. The resulting data base can be used to formulate revisions to Section III which account for the fatigue bahavior of piping steels in PWR environments. Completed studies of environmental-assisted cracking under simulated crack-tip conditions show that a strong environmental component of the overall crack growth rate exists whenever manganese sulfide inclusions dissolve at the crack tip. This accounts for the evidence that many of the older reactor steels, with higher sulfide inclusion contents, exhibit elevated crack growth rates, near the ASME water environment reference line. L The International Group on Radiation Damage Mechanisms (IG-RDM) was officially formulated to mount a coordinated effort for the identification and modeling of embrittlement mechanisms in RPV steels. i b
Small Angle Neutron Scattering (SANS) studies of mechanisms of the copper and phosphorus content contributions to radiation sensitivity provide further evidence of precipitation phenomena. Formation of copper-rich precipitates. (or clusters) appears to be a radiation-- enhancement process while that for phosphorus appears to be radiation-induc ed'. Field Ion Microscopy (FIM) studies saw evidence of-phosphorus-rich regions in service-irradiated and test reactor-irradiated'Gundremmingen RPV material. Irradiation assessments of the ASTM A 302-B reference plate and a MEA A 533-B reference plate (0,2% Cu, each) revealed only a small sensitivity to a fluence rate of j 8 x 1020 n/cm2 .3-1 (low) vs. 9 x 1012 n/cm2 -s-1 (high) for a total exposure of 5 x 1018 n/cm 2, E > 1 MeV. The difference. in embrittlement by the two fluence rates was somewhat greater for the A 533-B plate; the lower fluence rate was the more detrimental rate. Six IAEA-supplied RPV materials have been 288 C irradiation tested for embrittlement sensitivity vs. Cu, P and Ni content. The 41-J temperature increase for the two materials with a high phosphorus content greatly exceeded projections by RG 1.99 Rev. 2. OIRECTIVE This program of applied research will provide analytical and experimental methods and data that are necessary to ensure the structural safety and reliability of pressure boundary components in commercial light water reactor power systems. Emphasis is placed on characterization of material properties performance in a nuclear environment for application to plant aging, plant-life extension, and mitigation of the consequences of postulated accident scenarios. Current work is organized into three maj or tasks: (1) fracture mechanics investigations, (2) environmentally-assisted crack growth in high temperature, reactor primary water, and (3) radiation sensitivity and postirradiation properties recovery. The approach centers on experimental characterization of nuclear grade j steels and an asssessment of fracture and fatigue behavior under conditions of a nuclear environment where investigation of irradiated materials is a key element of each task. Experimental studies generally involve small, laboratory specimens whose behavior forms the basis to develop predictive methods for structural assessment. These studies are supported by analytical models of the mechanisms responsible for the observed behavior. The data will provide the basis for codes and standards, e.g., fatigue and fracture limits for i design in the ASME Boiler Pressure Vessel Code (Sec. III and XI), ASTM test methods, and revisions to NRC Guides. j
~
l 62
4 FY-1988 SCOPE TASK 1 - FRACTURE TOUGHNESS CRITERIA Complete experiments that characterize the behavior of surface flaws
- l. in irradiated, clad RPV steel. Complete a study to model WPS effects
!- in relation to a PTS scenario. Expand the data base for (a) C y vs. KIcMJc correlations and (b) piping fracture mechanics. Develop J R cume data from a low C yupper shelf RPV plate in order to evolve I a means for J-R curve extrapolation. Report on characterization of decommissioned Gundremmingen RPV steel. TASK 2 - ENVIRONMENTALLY-ASSISTED CRACK GROWTH IN LUR MATERIALS i Complete stress-life tests of smooth and notched piping steel specimens in PWR environments. Conduct stress-life tests of girth butt welded steel pipes containing a PWR environment. Complete a data base which can be used in revisions to the ASME Secion III Code. Complete study of environmentally-assisted crack growth rates in l prefatigued piping steel in simulated crack-tip environments. l Evaluate effects of corrosion potential and effects of managanese sulfide contaminants on crack-tip corrosion potential.
'l TASK 3 - IRRADIATION SENSITIVITY AND POSTIRRADIATION RECOVERY Investigate potential synergisms between impurity and alloying elements in radiation sensitivity development and identify associated radiation effects mechanisms. Formulate an international cooperative ,
group of scientists to advance the mechanisms of radiation damage in 1 RPV steels. Experimentally investigate effects of neutron fluence-rate on the magnitude of irradiation-induced changes in notch ductility, strength and fracts.re toughness properties of steels and weld deposits. Explore possible causes of anomalous dependence of , radiation sensitivity on specimen test orientation seen in trepanned 1 Gundremmingen RPV material. I 1 63
TASK 1 -' FRACTURE TOUGHNESS CRITERIA FRACTURE RESISTANCE OF 1RRADIATED STAINLESS STEEL CLAD VESSELS
Background
This research task has been designed to evaluate quantitatively the severity of a surface flaw in a clad layer of RPV steel in both the non-irradiated and irradiated conditions. The objective is to develop data that can be used in assessing the role played by cladding in either enhancing or mitigating crack initiation in a postulated PTS scenario. The complete' background information on the fabrication of the clad layer and the resulting material properties were reported in Reference 1. Baseline behavior for similar flaws in irradiated base metal is used for comparison. Clad material was prepared for this study using a three-wire process typical of older vessel construction. Summary of Results The unique feature of this test program was that relatively large specimens with surface cracks were tested in simulation of service conditions. The material fracture toughness properties of the consti-tuent materials and of the composite region as a whole was determined for both irradiated and unirradiated conditions. Elastic-plastic analysis of surface cracks was applied to calculate fracture toughness. Tests of clad layered bend bar specimens showed improved transition temperature fracture toughness over that of the virgin base metal. Tests on unirradiated specimens indicated that the fracture behavior of the composite clad metal-HAZ-base metal region was controlled principally by the properties of the HAZ material. In the present case, the HAZ material had about 40 C lower transition temperature than base metal. If the initial surface crack is sufficiently small so that it is embedded entirely in the clad metal, the crack must grow in size through slow-stable growth as illustrated in Fig. 1. These M ACHINED NOTCH FATIGUE CR ACK FUSION LINE CL ADDING 'w ' SLOW STABL E GROWTH Fig. 1 Slow-stable growth in surface SPECIMEN C10, -120* C Kgg =163 MPaV m
'^' **
M ACHINED NOTCH e dded in clad metal. ,:pe w FUSloN Li C L ADDING SLOW STABLE GROWTH HA2 - SPECIMEN C12, -60 C K gg =223 pa Vm 64
two examples were for surface cracks that were nominally half through-the-thickness of the clad layer. Hence, the crack growth resistance toughness (R-curve behavior) of the clad metal provided a first line of defense against crack instability. However, the clad metal had relatively low R-curve toughness and the resistance against growth may not be sufficient to ensure protection against unstable crack growth when the HAZ is reached in a PTS accident scenario. One of the objectives of the present proj ect was to develop an analytical basis for calculating crack instability conditions for j embedded surface cracks in cladding. A model developed for the i prediction of stable growth and for crack instability in HAZ material
- 1. is' illustrated in Fig. 2. Both R-curves shown (clad metal and HAZ) were determined from 1/2T compact specimens of thickness appropriate l to the depth of the metallurgical layer, about 5 mm. The R-curves are positioned on the abscissa according to initial flaw size iei the clad metal and where the growing crack enters the HAZ material. Crack l drive was calculated using an elastic-plastic model suggested by Merkle, Reference 2.
The example presented in Fig. 2 appears to have provided a reasonable prediction while at the same time explaining the mechanism of fracture control. There are two other noteable observations suggested by the above model: (a) that residual stresses in the clad layer are irrele-vant to the analysis, ....,,,,,,,,,, and (b) that total crack ;; *-""""""**' size is used to assess
~ .."qp;n, crack drive in a metal- ,, _. W '**I'* W
lurgically complex region. Specifically in u. - a PTS scenario, residual stresses in clad metal { will self equilibrate at f
, ,,, _ gg,,g.ay a stress level well I below that needed to ,, _
initiate slow-stable ~M crack growth. Secondly, a - for initial flaws that , 'j , , , penetrate into the HAZ . e. .. i. ,. i,. or into the virgin base CRACK SIZE mm metal, it is not accu-rate to mathematically strip away the clad layer part of the Fig. 2 Instability prediction using flaw. K-crack drive (K- lad metal and HAZ J applied) is derived from fr m C mPact spe cimens .R -curves the total flaw, indepen-dent of material prop-erty c'ifferences . 65
I. ! I' l l j J Tests on clad bend bar AS RECEIVED JRMDIATED 1t5 s M* n/cmay specimens- that were " -
. J
' irradiated reinforced , ,
the argument. that HAZ j # fracture- toughness con- t [ trols crack - instability y= - I 1 j
' of- the . composite g I I
region. Figure 3 shows f ,/ , _ , _ _ , , that the transition tem- w f. f ,,,,,,,,,,,,,,,,, perature. shift. indicated . .
",,'Q so is = =muen by 1/2T compact' specimen f
/ /
of HAZ material (dashed- / /
- lines) - and that of clad layered bend bar speci-
" ~" -" ~" ' " " "
mens with surface cracks (open '** triangles) is about the same. The transition temperature and temperature shift Fig. 3 Summary comparison of tran-shown for base metal sition temperature shift (76 C) was the same for between base metal and HAZ. both compact specimens and surface cracked bend bars. WARM PRESTRESS UNDER SIMULATED TRANSIENT LOADING
-Background The phenomenon of warm prestress (WPS) and its effect on the apparent fracture toughness performance of steel- has been .known and demonstrated experimentally for more than 20 years. If this phenomenon could be utilized in the analysis of loss-of-coolant scenarios in PWR reactors, it would represent a clear advance in efficient and safe use of current operating units.
The initial obj ective of the present proj~ect was to evaluate the features of the available mathematical models used to predict WPS effects. The obj ective of a second phase of study was to apply the i knowledge developed in phase 1 to the special conditions that are j generic to the pressurized thermal shock scenario. Here, stress l: intensity levels of WPS can be high and the beneficial effects potentially quite strong. A program was set up to predict trends in K f (post-WPS fracture) as a function of loading and thermal history using three of the most promisu s analytical models. This was followed by experimental w rification of Kf , and then by development of the necessary model refinement to improve the accuracy of prediction. Summary of Results AK Ic trend curve for the a 533-B virgin material was positioned on the temperature coordinate of a transition curve using IT compact 66
I' I L WPS t.OADING specimens. The tests for this UXFOA000
" ^^"
purpose were run at -23 C and b
-95.5 C. Two types of warm prostress cycles were chosen as , x 'c illustrated in Fig. 4. These s were load unload-cool-fracture, e.,, - - - - - - - - - - - - - -o A LUCF, and load-partial unload-cool-fracture, LPUCF. Warm pre- *1 _
0'*" n stress Kups was kept constant,at l 192 MPa/m and at T - 177 C. x ,c _ _ _g / l LUCF fracture, K g waT3 performed o
- e. 4 l (LPUCF) ---o--------o '
-95.5 C.
at 'T - -23 C and " MIN LPUCF ad 1/3 p'artial unload and " MIN " e o staar fracture temperature, T, f was r, rg
-95.5"C. Each test condition TEMPERATURE was replicated about 10 times to establish statistical signi- Fig. 4 Schematic representation ficance. One LUCF test was made of LUCF and LPUCF warm at -110 C to further explore prestress cycles test temperature sensitivity.
The results of the experimental work suggest that a modified KIc-like transition curve could be developed to model allowable post warm prestress loading. Two such transition curves were suggested, for LUCF and LPUCF loading. See Fig. 5. For the LPUCF cycle employed in , the present experiment, there was at least 100 percent retention of ! I fracture strength (corresponding to Kwps) down to (NDT -95 C) test temperature. Such curves can represent an appropriate replacement for K Ic transitional temperature curves for the assessment of transient scenarios. 220 Pot , l e t Un i ted pp 200 - _[ [ *PS I C 4',:3,, Q 'ieo - p iu:r o i4o - l g _ l p -_ __ - - _ _ - - -
*']
2 100 - l l
,o -
l a...... _ s. j lJL O tu:r
~
50 - 1 4o - j '" a , n w .oy in w.e m s o l
" l l
. L t _,_ _ L _ m 120 -100 -80 60 -do -20 o 20 i27 j
TEMPERATURE *C
)
Fig. 5 Implied transition temperature shift after K wps 1 ading The optimum means for the development of WPS - modified transition curves would be by computation, using a proven WPS model. Three models to predict K g fracture were evaluated. The most accurate one 67
was the Curry small scale yield-critical stress model (Ref. 3). A final recommendation of the present program was, however, that more work could be expended on improving the methodology for establishing . l the appropriate critical stress and critical distance used in that l particular model for various materials. i An unequivocal alteration of the effective fracture toughness of l A 533-B steel through WPS loading was demonstrated by the present experiments. These results suggest that service performance of reactor vessels could be entirely unrelated to virgin material fracture toughness trends established via present experimental practices. From an engineering standpoint, it would be of interest to know if a vessel retired from service would have higher effective toughness (WPS enhanced) than that indicated by surveillance specimens. CORRELATION OF DYNAMIC C y AND STATIC KIc/EJc TRANSITION TEMPERATURE INCREASES DUE TO IRRADIATION Backgound Reactor pressure vessel (RPV) surveillance capsules contain Cy specimens. but many do not contain fracture toughness specimens; accordingly, the radiation-induced shift (increase) in the brittle-to-ductile transition region (AT) is based upon ' the AT determined from notch ductility (C y) tests. Since the ASME K Ic and K IR reference fracture toughness curves are shifted by the AT from C, y assurance that this AT does not underestimate AT associated with the actual irradiated fracture toughness is required to provide confidence that safety margins do not fall below assumed levels. , Summary of Results To assess this behavior, comparisons of AT values defined by elastic-plastic fracture toughness and Cy tests were made using data from RPV weld and base metals in which irradiations were made under test reactor conditions (Ref 4). These comparisons represented the first systematic consideration of the AT from C y and AT from fracture toughness relationship. Using "as-measured" fracture toughness values (KJc), AT(KJc @ 100 MPa[m) were found te be less than AT(Cy@ 41 J) for weld metals but greater than AT(C y @41J) for base metals. This trend was also followed when the fracture toughness data were adjusted using the Ic-correction. To assess chemical composition effects on transition response, four plates were characterized in the pre- and post-irradiation conditions using Charpy-V, tensile, and compact tension (for fracture toughness) specimens. The fracture toughness tests were performed at quasi-static and rapid-loading conditions using unirradiated material, while irradiated specimens were tested using only quasi-static loading conditions. All of the fracture toughness test data were evaluated using the Kyc approach outlined in (Ref. 4). 68
Overall, the loading-rate-induced AT was found to be virtually independent of the chemical composition level of Cu, Ni and P. In contrast, the. irradiation-induced AT was highly dependent upon composition level, with the plate composed of high Cu (0.3% wt.), high Ni (0.7% wt.) and high P (0.028% wt.) yielding the highest AT. One encouraging sign from these results is the generally good agreement. iound between AT(Cy ) and AT(KJc) en ' ' ' ' ' for these four plates, in _ '/ contrast to previous re- Em - RPV Steel Plates j sults (Fig. 6). In par- * / ticular, the good agreement 5 "' am [* found at large AT levels p/ provides the first indi- t , . cation that AT(C y) can give [ f *4
,a ,
reasonable agreement with at high AT
?3
[ ,, M'
'/ ',
AT(KJc) _ 9 New Data levels. However, the fact ; ,
^
that AT(C y) tends to be { as - "''o"' Da - less than AT(Kyc) on , ,
,a a n an balance implies that an se in in isa in additional margin on AT may L* T**Pt" *t kJc
- 200 MP'6 ( C) be required to ensure that
.a non-conservative estimate Fig. 6 AT(Cy ) and AT(KJc) are in much of irradiation fracture better agreement for these four toughness does not occur. plates than for previous results.
PIPING FRACTURE MECHANICS DATA BASE
Background
Knowledge of material properties such as strength and fracture toughness is an important consideration for evaluations of structural integrity, such as for nuclear piping systems. The objective of this task was to establish such a data base for nuclear piping materials, with additional testing used to supplement available data. Summary of Results A computerized data base, called the Piping Fracture Mechanics Data Base or PIFRAC, was assembled for nuclear piping materials (Ref. 6). PIFRAC contains data from ferritic and austenitic stainless steels, welds, HAZ's, and cast stainless steels in various thermally-aged condition. PIFRAC currently contains data for 74 material conditions (including thermally-aged and mechanically-processed conditions), with j a total of 350 J-R curves. As an example of results contained in FIFRAC, Fig. 7 illustrates J-R l curves for a heat of cast stainless steel (CF8) pipe in the unaged condition and in an aged condition at 400 C for 10000 h (Ref. 7). The decrease in J-R curve level illustrated is somewhat larger than that found for other heats and different aging conditions. I 69
A report on ferritic and austenitic stainless steels is being prepared. All of the data in this report are given in PIFRAC. EL TR & 1;r-
- o,,,,,o __4L *:.lL1,a' 8 **
, *]* #L1lL reo enut.fnL5tedna* Pn 945MTB, f%e -
4aw Fig. 7 Comparison of J-R curves * , for HEAT PL at 25 C for i s' a "i ~ the unaged condition and // o ^ --
- an aged condition (at 7"' /// #
' ' " * ""* " 'i ',, f 400 C for 10000 h)
/'/
f lo o o ,.
= -
of p.*"* ^^',
. ieno
~ g,p,=!14 4 l
's [e de de cel nm. w CHARACTERIZATION OF GUNDREMMINGEN RPV MATERIAL
Background
The 250 MW boiling water Gundremmingen Reactor, KRB-A located in the Federal Republic of Germany (FRG) has been decommissioned. The vessel presents a unique opportunity for a critical correlation test of power vs. test reactor environment effects. A joint USA-FRG study, conceived by the NRC, is underway to evaluate material cut from the vessel beltline and above-core regions. Initial objectives of the joint study were examination of: in-depth embrittlement, notch ductility vs. fracture toughness relationships, service-induced vs. test reactor-induced embrittlement, and postirradiation annealing behavior. MEA's tasks in support of these obj ectives included the development of baseline mechanical properties for the preservice vessel cendition using archival material and the determination of fracture resistance changes wrought by a light-water test reactor environment for comparison against changes depicted by material trepanned from the vessel. MEA's irradiations are being conducted in the 2 MW test reactor (UBR) which is an NRC benchmark reactor, j Qualification of exposure rate effects is being made jointly by MEA and MPA, its counterpart in the FRG. Initial MEA irradiation tests of the archival material revealed smaller notch ductility changes than fevnd for the vessel material, : even though the test reactor fluence was about three times that received by the vessel inner wall surfaca at the beltline. Equally significant, an anomally was observed by MPA for the vessel material relative to its apparent radiation embrittlement in strong vs. weak test orientations (ASTM C-L vs. L-C). Here, the weat orientation indicated a much greater C y 41-J transition temperature eievetion than the strong orientation and a greater reduction in upper sheli energy level. The findings raised concerns on the application of NRC l Regulatory Guide 1.99, especially in its method for upper shelf l analysis.. 70
.]
L' l I In 1987, MEA undertook additional UBR irradiations 'of - the archive material to help clarify the anomalous results for the KRB- A trepan
- material. In 1988, the follow-on irradiations were completed and evaluations made of the C y,. tension, and fracture toughness specimens representing both test orientations.
Summary of Results The ' initial' postirradiation comparisons of C y specimens representing the strong and weak test orientations of the archival material indicated equal embrittlement tendencies. The 41-J transition temperature elevations with the (low) fluence of 2.7 x 1028 .n/cm 2 (E > 1 MeV) were about 18 C; the upper shelf reductions were less than 5 J. A more stringent 288 C irradiation test however, using a fluence of .2.3 x 1028 n/cm2 , showed a large difference in upper shelf reductions. The greater reduction was associated with the strong test orientation of the forging. (36 J or 23% vs. 19 J or 18%, weak orientation). This finding is diametrically opposed to the test orientation dependence of the upper shelf reduction found for the vessel trepan material by MPA. Thus, the anomaly persists. Other MEA irradiation tests of the archive material demonstrated that the anomaly is not rooted in the KRB-A vessel service temperature being somewhat less than 288 C (currer}t estimate for the beltline. region: 279 C). Data were developed for lower temperature irradiation conditions as part of the follow-on ef fort . .The uncertainty ir. the irradiation service temperature is judged not of practical significance since the transition temperature elevations with 8.6 x 1028 n/cm2 at 279 C and 260 C were greater than that at 288 C by. only 9 C and 12 C, respectively. In turn, the estimated 279'C service temperature should not have a maj or bearing on the anomalous C-L vs. L-C orientation results. To resolve the anomaly directly, the NRC has agreed to an MEA proposal that a special test reactor experiment be undertaken wherein specimens from the outer ligaments of the vessel trepan (low fluence exposure, essentially nil embrittlement region) and specimens from the archive material are irradiated together. Both test orientations of each material will be included for a 1:1 comparison of radiation sensitivity. If the same orientation dependence of embrittlement ) sensitivity is indicated ira this case, it can be concluded that the ) anomaly is rooted in the material tested and not the difference in fluer'ce rates between UBR and KRB- A service. On the other hand, if the same orientation dependence is not indicated, it will confirm the !. fluence-rate effect. Either finding will prove significant to Regulatory Guide 1.99. A report summarizing all investigations made by this task, including , fracture toughness vs. notch ductility comparisons, has been issued (Ref. 7) . 71 !
SIZE EFFECTS ON J-R CURVES FOR A 302-B PLATE
Background
of the various concerns facing utilities with nuclear power plants, a requirement of Part 50 of Title 10 of the Code of Federal Regulations ] (10 CFR 50) specifies a minimum Charpy-V upper shelf level of 68 J < (50 ft-lb) for the belt line materials of RPVs. Recent calculations l by members of American Society of Mechanical Engineers (ASME) Section XI indicate that J-R curve data for crack growth increments ranging up to 15 mm or 20 mm may be required to demonstrate structural integrity for such low upper shelf materials, using procedures under consideration. by ASME Section XI. From standards adopted by Committee E 24 of the American Society for Testing and Materials (ASTM), extremely large specimens, ranging in thickness up to 150 mm or 200 mm, are needed to give valid data for the large crack growth increments described above. Obviously, any use of such large specimens in irradiation programs such as surveillance capsules is generally impossible due to size constraint in the RPV. Therefore, procedures and methodology for taking the available valid data from small specimens (for crack growth intervals from ~ l mm to 2.5 mm), and extrapolating that data in a conservative manner up to the large crack growth intervals required, would fill the void needed to accurately and conservatively evaluate safety margins for cases in which low upper shelf material is a key concern. As initially conceived, the goal of this program was to develop J-R curvo data for a material exhibiting a low Charpy upper shelf energy level, using small and large specimens. These data would form a reference data set for the development and validation of procedures to extrapolate small specimen data at small crack growth (Aa) levels to the large crack growth levels required to assess the structural stability of RPVs. The large test specimens would provide valid data ' to large crack growth intervals to validate or permit refinement of the extrapolation procedures. Summary of Results Although most of the limiting low upper shelf toughness materials are in fact weld metals, a base plate was selected for this program. A weld metal was not chosen since RPV welds are typically fairly narrow, less than 51 mm (2 in.), and the large specimens, in particular, therefore would be composite or duplex specimens, composed primarily of base plate with the relatively narrow band of weld metal. The small specimens would instead tend to be composed of all weld metal. In contrast, the use of base plate as the material means that all specimens, both large and small, would be m.entially homogeneous. Of particular concern in the use of weld metal is the difference in deformation characteristics among weld metal, base plate, and even the intermediate zones (such as heat affected zone and fusion line), and the effect that this would have on the apparent measured toughness. The material used in this study is an ASTM A 302-B plate. The heat treatment was determined from a review of the metallurgical histories 72 I
1 l l, I of many production A 302-B plates used in early vessel construction. The sulfur content of the ingot used was 0.025% (wt.), leading to a high content of manganese-sulfide inclusions. In addition, a minimum of cross-rolling was used to maximize differences in properties for the high (L-T) and low (T-L) toughness orientations. Charpy-V upper shelf energy for the T-L orientation was found to be 74 J (54 ft-lb). At the test temperature of 82 C or 180 F (used for the fracture toughness tests), the 0.2% offset yield strength was 459 MPa
- (66.6 ksi) and the ultimate strength was 584 MPa (84.8 ksi) .
The surprising result from the J-R curves of this plate was the extreme size dependence found in contrast to the expected uniformity of R curves from the different size specimens (Ref. 8). As illustrated in Fig. 8 for J D* and J M*, increased specimen size yielded large reductions in J-R curve level. In particular, the decrease in J-R curve slope implies substantial reductions in toughness, as one might determine from a tearing instability of J-T analysis. The various specimen sizes (ranging from 0.5T- to 6T-CT) did give similar J levels, but the tearing trends after crack initiation were quite dbferent, with the largest specimen size providing almost no resistance to tearing. The behavior of this material in a large structure such as an RPV cannot be estimated, although one would not expect it to exceed that of the 6T-CT specimen. Should this steel be used in a RPV, the data from the small surveillance specimens (CT) would not be expected to yield an accurate nor conservative prediction of the R curve associated with full thickness sections. cam .,, w n . . n, . , 0 0.50 I aid I f.3 2.80 2 0 1 0 0.50 iN 1.50 2.80 2 0 L OO SM;E Me_Ld0 n.4d!-E th a'e Q e ie: s N' 62 * % O.5 T g ~$ 1T wg
~
, 0.5 T .
g q p 1T f q 98 [ 2T I O ;eg # 7 ~ Nf [ 2T - I h p %-., ,j.[ sy "{* A m' , ., , m av7)" sT f
, _ , , , , , , M, w * " * * ,. +.,,+ +.s+ +&. * * " ' * ~ ""' x ios .
v '
- - 0. W t t - 4T CT a e -4Hf WO e U<1 + - L Y- ci . 0.51-;t e IT <T e -(14T s 2 7-<1 4 - 2 T-CT 0 id d 40.0 EP e ed 0 2P_ B 42.0 L 0 wa . n. m . o.. .
Fig. 8 Using both J*D and J*, g large reductions in toughness are found with increasing specimen size for this A 533-B plate. The fracture surface resulting from these tests are, at best, unusual. As illustrated in Fig. 9, the fracture surface of virtually j every specimen exhibited a severe splitting behavior. Initially j thought to develop at the Mn-S inclusions, the splits may instead occur due to a banded, alloy-rich structure. The distribution of the splitting correlates with both the banding and the inclusions, such that neither can be positively credited with causing the splitting. 73
.).
As well, the fracture surfaces of the large and small specimens exhibit similar split densities, such that the total number of splits may be responsible for the large reductions in fracture toughness in the large specimens. No intuitive explanation for this is obvious. Reference 8 gives several topics for follow-on work, covering needs for explaining the extreme size dependence in the A 302-B plate and for a data extrapolation study using a low C upper y shelf energy weld metal. - ba
.N
"' t . .
l 1 i-ms
;gf/plpi:e -,fNJ f y+ ril;fPt 79:'[m e?'
- ;,e: . 6 t Yig' f. nt$f.
a*xyn'.agC a?~ gf'4?p
; k '1;[ J . ;g ^ l y
n ci:
"' kn,
. ^
; Y._ y .- r y;j
,1
, h h'; . 4 :
(
- n 1l) f ,,yy b w( b , $ $ $ '
q k Fig. 9 Fracture surfaces for the high toughness , small specimens and the low toughness, large specimens are similar in overall j appearance, principally distribution of the splits. l l l 74
l TASK 2 - ENVIRONMENTALLY-ASSISTED FATICUE CRACK CROWTil IN l LWR MATFRIAIS l STRESS-LIFE TRENDS FOR CARBON PIPING GTEELS IN PWR ENVIRONMENTS
Background
In Section III of the ASME Boiler and Pressure Vessel Code (Ref. 9), environmental degradation is incorporated into stress-life (S-N) calculations without any significant experimental basis from which a designer can infer the degree of environmental effects. Tests have been conducted in PWR environments on smooth, notched, and 102-mm , welded pipe specimens in order to evaluate any reduction in cycles to ! failure which may be attributed to environmental effects. Summary of Results The testing of ASME SA 106-B . steel specimens at both 24 C and 288 C air and in PWR environments at 288 C are summarized in Table 1. The results for S-N testing in air and PWR environments show the following trends: (1) Smooth and notched base metal specimens tested at PWR environ-ment temperatures showed a reduction in low cycle fatigue strength, which results in a decrease in the intended safety margin of the ASME Section III design curve for carbon steels (Ref. 10, Fig. 10). 10 4 ' h ASME SA 106 5 Steel Z
;;;y we= :
^
e Air Enyt . , R 100, smooth ^ M b3 " l!$.- t 00, si$tf PWR Eavt . . R Kg-6
"'"C 10 4 a$
E j
; : , . ::=:::::n:: 1 i a
+'
$0 "'"" b 10 3 +'
/ b [
Cs 5 e
! io ' -
- v= :::- i !
Rean Data Ciarve, 24*C AJr Tests - 10 0 ' ' ' ' ' ' '"' ' "" ' 10' 10 10 ' 10
- 10
- 10 ' 10 Cycles To Failure Fig. 10 ASME Section III pseudostress amplitude vs.
i cycles to failure for small smooth and notched { specimens of SA 106-B steel in 288 C air and l PWR environments. Though the application of I Neuber's rule to the data for notched l specimens yields slightly conservative results, the data for notched specimens nearly completely uses up the margin of safety offered by the Section III design curve. 75 i i 1 w______,_______.____--____m--__-_- _ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ' - - - - ~ ~ ~ ' - - - " - ' ' ~ - - - - ~ ~ ~ ~ ~ - - ~ ~ ~ ~ ~ ~ ' ~
l-t c l (2) Notched base metal specimens tested at 0.017 Hz in 1.0 part per billion , (ppb) dissolved oxygen environments nearly L completely used up the safety margin of the Section III design curve, whereas smooth specimen tests from base metal and weld metal under' similar conditions showed considerably less ' degradation. These results show that fatigue crack initiation is not significantly affected by 1.0 ppb dissolved oxygen, and that most of the observed degradation of the notched specimens may be . attributed to crack' growth acceleration (Ref. 11, Fig. 10). (3) Girth-butt welded 102-mm diameter pipe ' specimens tested in 288 C air exhibited an extrapolated fatigue strength reduction factor (K g) of 3.47, and data analysis using the procedures outlined in the Section III article NB-3653.6 resulted in pseudostress amplitudes which were conservative when compared to the Section III mean data line for carbon steels. The only specimen tested which contained a pressurized' 288 C PWR environment showed insignificant degradation of fatigue life (Ref. 12, Fig. 11). 10 4 : _ na GL tb lielded Pipe Specimens -
,.,. res*c ,,
~
10 3 --- 6 * ' 3 - a
= .
0 0O : 4 3 2 _
+' - 10 3 +'
! 10 2 7 E Z 5
[ j : = [ C [ ~ fe7cNEsN a -
/ .
=
j io ' y :::e:=:-
- er,r l '*' -j 10 e t t t orn! ! I t unl t ! t o"l ! tinul ? ! t nnl e i i A 10 8
10 10 ' 10* 10 0 10' 10 Cycles To failure
'l Fig. 11 ASME SEction III pseudostress amplitude vs. cycles to failure for 102 mm girth butt welded , pipe specimens of SA 106-B steel in 288 C air. One data point was generated for a specimen which contained a 13. 8 MPa , 288 C PWR environment. The l-procedure outlined in article NB-3653.6 l in Section III was used to determine the pseudostress amplitude values.
76
c Table'1 Summary.of Testing in PWR Environments Specimen R-e Strain Ratio Test ' Environment Number of
, Type Frequency Specimens (liz) ' Tested Base Metal 1 -1.00 2 10 24 C Air' 19' Base Metal 1 -1,00 2-10 288 C Air 16 Base Metal 1 0.05 2-10 288 C Air; 3 Base Metal 1 0.50 2-10 288 C Air 3 Base. Metal 2 -1.00 2-10 288 C Air 7 Base Metal 3 -1.00 2-10 288 C Air' 8 Base Metal 6 -1.00 2-10 288 C Air 7 Weld Metal 1 -1.00 2-10 288*C Air 7 Base Metal 1 -1.00 1.0 288 C PWR 9 Base Metal 1 -1.00 0.1 288 C PWR 'S Base Metal 1 -1.00. 0.017 288 C PWR' 4 Base Metal 1 0.05 1.0 288 C PWR 8 Base Metal 1 0.50 1.0 288 C PWR 7 Weld Metal 1 -1.00 1.0 288 C PWR 6 Weld Metal 1 0.05 1.0 288 C PWR 6 Weld Metal 1 0.50 1.0 288 C PWR 5 Base Metal 6 -1.00 1.0 288 C PWR 6 Base Metal 6 -1.00 0.1 288 C PWR 7 Base Metal 6 -1.00 0.017 288 C PWR 5 Base Metal- 6 0.05 0.017 288 C PWR 6 Base Metal 6 0.50 0.017 288 C PWR 1 Base Metal 3 -1.00 0.017 288 C PWR 2 Welded Pipe -
-1.00* 0.2 288 C Air 8 Welded Pipe * -
-1.00* 0.017 288 C PWR 1 Welded Pipe * -
0.05* 0.017 288 C PWR 0 Welded Pipe * - 0.50* 0.017 288 C PWR 0
- Did not complete
- Load ratio The results of this investigation include the creation of the first i known data base of fatigue life data for SA 106-B piping steel small i specimens in simulated PWR environments. Based on the results of this investigation, it should be clear that Section III should be revised before plant life extension activities are continued. Re-certi-fication of pressure-retaining components for operation beyond their original 40 year design lives will require that utilities must
{ recalculate, with greater precision, the stress analysis of pressure J vessel and piping components and new fatigue usage factors, based on a revised Section III methodology which properly takes into account all of the known variables which degrade fatigue life, f l 77
MICR0 MECHANISMS AND CALCULATIONAL MODELS FOR ENVIRONMENTALLY-ASSISTED CRACKING IN PWR ENVIRONMENTS
Background
The potential environmentally assisted cracking problems- of LWR materials in PWR environments are complex functions of the loading variables, including their dependence on time, and material and environmental chemistries. Attention has been focused on a number of crack-tip microprocesses which might explain the contribution of the environment to the enhancement of crack growth. Among these l possibilities are hydrogen-assisted cracking models and oxidation film rupture / repair models which are in various states of conceptual development and experimental verification. In conjunction with the micromechanism development, various calculational models are being derived and evaluated, many of them based on strain-rate at the crack tip and . the relation of this strain rate and various time-dependent corrosion processes on the environmentally-assisted cracking phenomenon. Summary of Results Slow strain rate tests (SSRT) using cyclically prestrained piping steel and environments which were rich in manganese sulfate were completed and analyzed during the past year. These tests were intended to simulate both the cyclically plasticized material found at crack tips and the contaminant-rich aqueous environment which evolves with the crack enclave. For comparison, tests were also conducted using non-prestrained steel and bulk water PWR chemistry. All tests were conducted in fully deoxygenated environments; many were conducted with corrosion potential control. The as-received and prestrained materials were characterized through transmission electron microscope examinations of the dislocation structure. Figure 12 shows a comparison of the as-received and cyclically prestrained microstructure of A 516 Gr. 70 carbon steel. Restraining resulted in a small elevation of the yield and ultimate stress values, but little thange in ductility. In helium and bulk PWR-water environments, all SSRT specimens at the free corrosion potential failed in a ductile manner, i.e., there was no measurable component of stress-corrosion cracking (SCC). In environments containing Mns, all specimens cracked by SCC. The number of initiation sites decreased with an increase in the level of prestrain. Figure 13 shows the fracture surface from a specimen tested in MnS-containing PWR water. Corrosion assisted fracture properties were not affected by the cyclic restraining alone. However, the presence of MnS in the aqueous environment promoted a large degree of corrosion cracking and a reduction in the time-to-failure of the SSRT specimens. Another l important factor in the MnS-saturated, PWR water experiments was that the corrosion potential was -590 mV (vs. standard hydrogen electrode (SHE)) as compared with values in the pure PWR water experiments of
-750 to -850 mV (vs. SHE). This places the reaction in an area on the Pourbaix diagram in which H2S, FeS2/FeS are stable. The latter two 78
sulfides have been observed in previous X-Ray photoelectron spectro-scopy (XPS) studies of corrosion fatigue specimens of pressure vessel and piping steels. g
* "* ?hf x .,4 <, ' f4 llld.,, , y;'? , .. a N -
j s, ( Q
.: u ,n 4,1 - ,
. , y, ,t, a '
t y g 3 ;;
,A,Wjp;,*;',);;,/)1'Wg'y,:.
n ,.1,! ' / ..y
~
~
* * '; in '
'y,' ,
rj! *: .s * *
' V, .
'f, /@ rg?f'$ p
- s .- : , ,. :' , Y. }. al} , .o/
s, i, . ' , ' 7j + l' ,2
*^
g<y,,:2 yf.'* [';. ..l p.; , , c : g ~/ ..
; g h ..,,, ,, , j m
1 f) , I:Y*l W'.N*f YThjk '! - l 1. Ipm Fig. 12 Transmission electron micrograph of dislocation structure in A 516 Cr. 70 steel as-received (a) and deformed -1 1%. 4 a ,r- '
,yh-t
"*$;/ ;
) .
?
F f PW - 4f
]hls .
(a) (b) Fig. 13 Scanning electron micrograph of notched and prefatigued A 516 Gr. 70 steel tested under slow strain rate conditions in manganese sulfide contaminated PWR environment. Specimen overview is shown in (a), and an enlargement (b) shows the notch depth and corrosion assisted fracture regime. 79
p I l
.I I,
These results demonstrate that cyclic restraining, similar to that , which'might occur during system shakedown, does not affect subsequent i corrosion-cracking . susceptibility. Additionally, a significant corrosion cracking conponent is evident in PWR water only when the applied corrosion potential is high (near 0.0 mV ( SHE) ) ., or when a con' aminent species (such ' as MnS) is available to elevate the local potential, i l 80 I (
w l TASK 3 - IRRADIATION SENSITIVITY AND POSTIRRADIATION PROPERTIES RECOVERY MECHANISMS OF IRRADIATION DAMAGE FOR REACTOR VESSEL STEELS
Background
,. This task is directed to the isolation and modeling of radiation effects mechanisms in RPV steels. The experimental phases are being performed largely by subcontractors under MEA guidelines. Additional work is being done through cooperative research programs with certain overseas laboratories including the Technical Research Center of Finland (VTT), the Centre for Study of Nuclear Energy (CEN) in Belgium and the Paul Scherrer Institut (PSI) in Switzerland and with certain USA organizations such as Carolina Power and Light Company (CP&L), the Tennessee Valley Authority (TVA) and the Oak Ridge National Laboratory (ORNL).
Initial objectives were the application of advanced microscopy for purposes of (a) confirmation of the basic mechanism for the copper content contribution to radiation sensitivity development, (b) definition of the mechanism by which nickel alloying reinforces the copper content effect, and (c) isolation of the mechanism of the phosphorus contribution to radiation sensitivity development. Signi-ficant progress was made in fulfillment of these obj ectives in 1987 and was reported in Reference 13. Recently, the task was extended to a now set of objectives described in the Section: Future Plans. Summary of Results During 1988, the Small Angle Neutron Scattering (SANS) investigations subcontracted to GKSS were completed. Also, a set of investigations by Field Ion Microscopy (FIM) and Imaging Atom Probe (IAP) represent-ing a cooperative effort with ORNL (DOE-sponsored participation) were completed. Concurrently two new thrusts . proposed by MEA were under-taken with NRC encouragement. One depicting a cooperative program with CP&L and TVA, pursued the anomalous differences in irradiation behavior observed in test reactor vs. service-induced embrittlement studies of the Gundremmingen vessel material. The second thrust was the initiation of a new technical activity which would serve as a forum for work on radiation damage mechanisms identification and modeling. This objective has been realized with the formation of the International Group on Radiation Damage Mechanisms in Reactor Vessel Steels (IG-RDM) described below. The IG-RDM was officially installed and a Charter adopted during the 1988 International Workshop on Mechanisms of Embrittlement. of Pressure Vessel Steels at Harwell, (UK). The Workshop was conceived by MEA and co-organized by J. R. Hawthorne (MEA) and Dr. C. A. English (UKAEA Harwell Laboratory). The MEA-sponsored SANS investigations proceeded with two groups of materials: laboratory-melted A 533-B steel plates having statistical variations in copper and phosphorus content and laboratory-melted iron 81
(model) alloys depicting variations in copper, nickel and phosphorous content. Both groups were 288 C irradiated and tested for strength and notch ductility changes and then analyzed with SANS . The iron alloys were analyzed in the present reporting period; results confirmed certain trends observed with the steel alloys in the earlier tests. Specifically, the results indicate that low volume fractions of fine phosphorus clusters (or phosphides) form in low copper materials which independently contribute to radiation hardening and embrittlement. In high copper materials, phosphorus appears incorporated in the copper clusters formed by irradiation wherein the independent hardening contribution of phosphorus is essentially lost. Phosphorus, however, refines the size distribution and slightly increases the volume fraction of the copper clusters without having major effects to the mechanical properties of the material. The SANS investigations of the iron alloys also gave indications that nickel has a minor effect on the radiation-enhanced precipitation of copper-rich particles but phosphorus appears to influence the kinectics of the process possibly by interactions with vacancies. In copper-free iron alloys, the clusters formed are richer in phosphorus than is necessary for the precipitation of Fe3P; therefore, the clustering process could be taken as one of a radiation-induced effect rather than a radiation-enhanced effect. Copper-rich cluster forma-tion is judged a radiation-enhanced process, based on the evidence available. The cooperative FIM program on the Gundremmingen vessel material detected phosphorus-enriched regions in the actual KRB-A pressure vessel material and the test reactor-irradiated archive material (Ref. 14). The fluences of these materials were 2.7 x 1018 and 8.5 x 10 18 n/cm2 , respectively. It appeared that the fluences were sufficiently low to observe the phosphorus pre-clusters or atmospheres. The difference in microstructure produced by the rapid vs. slow irradiation rates is still under investigation. The materials will be investigated further using SANS and other techniques. The formation of the IG-RDM and its specific obj ective s was the original proposal of MEA. The intent of forming this activity was to promote or encourage (a) timely information exchanges between F laboratories engaged in mechanisms isolation, identification and modeling, (b) new cooperative programs for mechanisms identification among laboratory sites, and (c) broad application of the highly specialized or unique microscopy equipment to research materials having wide interest. Building on these aims, the new IG-RDM is viewed as a means to (a) schedule informal but structured periodic meetings where recent results could be discussed in an open, informal atmosphere, (b) establish cooperative ventures, through research material exchanges and/or cross-applications of microscopy equipment or radiation facilities, (c) coordinate research on special problems, and (d) arrange cost sharing activities, where possible. One expected benefit to the NRC from the IG-RDM activities is illustrated in Fig. 14 Advancement of embrittlement prediction capabilities necessarily involves a close working relationship between and mutual 82
MEA APPR_QACH To oUALIFYlNQ,JAD.lA.IjQN EM3RITTLEMENT FOR PLEX _ RQ 1,99 PREDICTIVE CAPABILITY w- ~ f g (hN VTT SER T N
/ N UNDER N ING~
Ed ksc COMPOSITION - ~ HARDENING e DOSE RATE - - REVERSIBILITY
]
1AR / ANNEAL -- - INCUBATION TIME 'l STATIC / DYNAMIC - --- PROPERTY SELECTIVITY FLUENCE DEPENDENCE - - TIME / TEMPERATURE SPECTRUM DEPENDENCE - - STRAIN RATE sensitivity
- IRON BASE VS. STEEL ALLOY Fig. 14 Anticipated interaction of the new International j Group on Radiation Damage Mechanisms in RPV Steel (IG RDM) conceived by MEA. The IG-RDM is designed !
to promote cooperative programs and information l exchange to speed research developments of high value to Regulatory Guide 1.99 reinforcement of mechanical properties determination efforts on one hand and mechanisms identification efforts on the other. The IG-RDM is designed to speed the latter while reducing costs of the total effort. The Workshop itself was the second of its kind and the successor to the 1985 Workshop sponsored by the NRC(MEA) and EPRI(UCSB) and held in Monterey, California Representatives from seven countries participated in the present Workshop; comments received by the co-organizers were very positive . The three-day meeting had technical sessions focusing on four topics: experimental techniques, mechanisms isolation and modeling, fluence-rate effects and PLEX, and postirradi-ation annealing and reirradiation. The overriding impression was of buoyant activities within Europe and the USA with a marked increase in the number of active research groups interested in mechanisms since 1985. Also, it was clear that major advances in microscopy equipment and its application had been made since 1985. These advances played significant roles in the reported research progress. Follow-on workshops are planned by MEA at roughly one year intervals. INVESTIGATION OF FLUENCE-RATE INFLUENCE ON IRRADIATION EMBRITTLEMENT SENSITIVITY OF RPV STEELS
Background
Data comparisons evolving from long term (power reactor surveillance) vs. short term (test reactor experiments) suggests a time-at-tempera-33
ture dependency of irradiation-induced embrittlement. A second inference from surveillance data is that, contrary to some recent proj ec tions , the embrittlement process does not saturate at fluences that are less than end-of-life values for vessels made of highly ; radiation-sensitive steels. To verify both indications for the NRC, a j set of special experiments was mounted in a light-water-cooled test ; I reactor (the UBR) in 1983, i The experiments were designed to provide the closely-controlled 4 conditions that are unobtainable in a power reactor. Three fluence j rate levels, were selected for study and span the range of power reactor-to-test reactor exposure rates. Target fluences for the two i highest fluence rate exposures were 0.5 x 1028, 1.0 x 1028 and ; 2.0 x 1018 n/cm2 ; the target fluence for the lowest fluence rate set was 0.5 x 1018 n/cm 2 (E > 1 MeV). Materials are two reference steel plates (A 302-B and A 533-B) and two reference submerged-arc welds. ; I Summary of Results The reactor exposures to all but two of the 18 capsules forming the irradiation matrix were completed by 1988. Materials in the two I remaining capsules are the submerged arc welds. Postirradiation ! evaluations have also been completed for the available Charpy-V (C y), compact tension (0.5T-CT) and tensile specimens. References 15 and 16 i presented the results for the intermediate fluence rate exposures l (5-6 x 1011 n/cm2 .3-1) and the high fluence rate exposures (8-9 x 1012 n/cm2 -s-1). Findings for specimens of the reference plates irradiated at 288 C at the lowest fluence rate (8 x 1010 n/cm2 -s'1) are now available. The specimens irradiated at the lowest fluence rate received a fluence of 0.54 x 1019 n/cm 2 (E > 1 MeV) at 288 C nominal). The fluence ; variation was 6% for the Cy and tensile specimens; the variation was somewhat larger for the CT specimens because of the larger capsule cross-sectional area. For the ASTM A 533-B reference plate, the data j can be compared to data for a high fluence rate experiment receiving 0.45 x 1018 n/cm2 . For the A 302-B reference plate (ASTM correlation l monitor material), the data can be compared to data for a high fluence j rate experiment receiving 0.56 x 1019 n/cm2, j Referring to the A 302-B results (Fig. 15), the data sets for the lowest and highest fluence rates essentially form a common data band. Treated individually, the lowest fluence rate data describe a , C y 41-J transition temperature elevation of 53 C which is only 9 C higher than that found for the highest fluence rate test. The yield strength elevations by the two exposure conditions differed by less than 12 MPa (1,7 ksi). Accordingly for this fluence level, the 100X 3 difference in fluence rate is shown to have an insignificant effect on f embrittlement sensitivity of this steel. This plate contained i 0.21% Cu (high) but only 0.18% Ni (low). i 84 j I
- 10 0 0 10 0 200 2884'Cl ItHbl PL ATE CODE 23 F UI I
( A 302 - B .LTI N-
^
b 93
~
<N N ____'_ . 84 0" UNMAADIATED
\
\ f' O UNIRAADIATED 4Q 0; 8
, 0 54 0 n
=> c a">
fe p f ... . , g ,* g D;N y E
-16 0 -80 0 80 16 0 240 320 400 5504* F)
TEMPERATURE Fig. 15 Comparison of Charpy-V notch ductility change with a high fluence rate vs. a low fluenge rate exposure at 288 C to 0.5 x 10 19 n/cm . The fluence rates differ by a factor of 100 and are illustrative of test reactor vs. power reactor vessel exposure conditions. , i The results for the A 533-B reference plate (0.20% Cu, 0.63% Ni) do not follow the trend pattern for the A 302-B plate exactly. In this case, the difference in embrittlement produced by the lowest fluence rate vs. the highest fluence rate appears larger; that is, the transition temperature elevations differ by 20 C and the yield strength elevations differ by 56 MPa (8.1 ksi) . A portion of these differences can be attributed to the fluence difference between the two irradiation test. Considering this factor, the fluence-rate effect does not appear large (unless considered on a percentage basis). Of additional interest, the A 533-B plate exhibits about the same sensitivity to radiation-induced change of notch ductility and yield strength as the A 302-B plate at a fluence level of 2 0.5 x 1018 n/cm . It is expected that the two weld metals (0.35% Cu, 0.65% Ni) now being irradiated to this fluence level will show a higher sensitivity to radiation induced embrittlement. The particular welds involved are designed to test the significance of weld vs. plate and that of welding flux type (Linde 80 vs. Linde 0091). Preliminary results (Ref. 15) suggest that the welds are much more sensitive to fluence rate than plates in this fluence regime. Results from fracture toughness tests of the two welds and the A 302-B plate generally are consistent with the Cy notch ductility test findings. The data for the A 533-B plate appear to be an exception. i kThereas the C y data for the low vs. high fluence rate exposures show a difference in 41-J temperature elevation of about 15 C, the companion fracture toughness data describe a difference of about 40 C from preliminary assessments. The fracture toughness comparison is hampered by data scatter. 1 85
INVESTIGATION OF IMPURITY ELEMENT - IMPURITY ELEMENT INTERACTIONS IN RADIATION EMBRITTLEMENT SENSITIVITY OF RPV STEELS
Background
MEA experiments have demonstrated that maj or . element-element interactions exist in radiation sensitivity development in reactor vessel steels and welds. The studies under the current program have revealed that ' copper and manganese, copper and molybdenum and copper and . phosphorus can interact to influence the C y 41-J transition temperature elevation produced by a given fluence (Ref, 17, 18). Such interactions have an important bearing on the prediction of radiation-induced embrittlement by NRC Regulatory Guide 1.99 or other. methods. I In the case of the copper-phosphorus interaction, MEA experiments E established that the level of the phosphorus contribution to 288 C radiation sensitivity depends greatly on the amount of copper present; l that is, the phosphorus contribution is most pronounced when the copper content is low, for example 0.02% Cu. At a high copper content, say at 0.25% Cu, the phosphorus is " tied up" in the form of a Cu phosphide and does not form a phospacrus precipitate, and its individual contribution to radiation sensitivity will be small. This ! raises _ the question of the influence on embrittlement sensitivity of a high (or intermediate) phosphorus content when coupled with an intermediate copper content. The ASTM recommends a phosphorous content of no more than 0.015% P and a copper content of no more than 0.12% Cu for reactor vessel beltline materials (product analysis specification). The availability of certain ' steel compositions in the materials inventory of the 1 International Atomic Energy Agency (IAEA) was viewed by MEA as an excellent opportunity for clarifying the practical implication of the ASTM-allowable limits on the two elements. Access to the materials was possible through MEA's representing the USA and the NRC in the IAEA International Working Croup on Reliability of Reactor' Pressure Components (IWG-RRPC). The NRC authorized the addition of certain IAEA materials to the MEA irradiation effects investigation in'1987. The materials were secured and irradiation-tested in this reporting period. Summary of Results The six IAEA-supplied materials were 288 C irradiated in the UBR reactor using two experiment assemblies. One containing Charpy-V (Cy ) specimens for notch ductility determinations , received a fluence of 2.7 x 102' n/cm2 ; the other containing tensile specimens for strength determinations, received 2.2 x 1018 n/cm 2 (E > 1 MeV). Three of the materials were thick-section RPV plate, forging and weld materials produced commercially; the remaining materials were 32-mm thick A 533-B- steel plates from laboratory melts depicting increasing contents of phosphorus and copper impurities. The commercial plate and forging materials had about the same copper, l nickel and phosphorus content (~ 0.15% Cu, 0.81% Ni and 0.016% P) and 86 ; i i _______________________U
1 showed about the same 41-J transition temperature elevation , (~ 79 C). . Embrittlement projections for these materials by Regulatory ' Guide 1.99, Rev. 2, were within 3C of the measurements. The weld metal containing a higher content of each of the three elements (0.26% Cu, 1.10% Ni, 0.026% P). Its 41-J temperature elevation was approximately twice that of the plate and forging. While a higher sensitivity to radiation embrittlement was expected based on the weld's composition, the measurement of 41-J transition temperature
' increase was much greater than the Regulatory Guide 1.99 projection.
The difference equates to two standard deviations in this case. The yield strength elevation of the weld metal by irradiation also was about twice that for the plate and forging. The results clearly demonstrate the need for minimizing copper and phosphorus in RPV materials and further study of the nickel content contribution for amounts greater than 1% Ni. The contribution of nickel content to radiation sensitivity has been tied to a reinforcement of the copper content contribution (and possibly to the phosphorus contribution in low copper steels). The results for one of the three laboratory-melted plates (0.01% Cu, 0.8% Ni, 0.007% P) support this general proj ection; the transition temperature elevation by 2.7 x 1028 n/cm 2 was on the order of 22 C. That' is, the 0.8% nickel content did not contribute in a major way to the radiation sensitivity. The addition of 0.017% P doubled the irradiation effect, attesting to the importance of this impurity to radiation embrittlement projections. Lastly, the findings for the third plate of the laboratory melt series (0.16% Cu, 0.8% Ni, 0.017% P) demonstrated the overriding effect ,of copper content; the transition temperature elevation was 2.5 times that of the low copper, high phosphorus plate. While Regulatory Guide 1.99 proj ec tions for the low copper, low phosphorous and high copper, high phosphorous materials agree relatively well with experimental measurements, the Guide underpredicted the performance of the low copper, high phosphorus plate by more than 2 standard deviations. It is noted that the current formulation of Regulatory Guide 1.99 does not contain a term for a phosphorus contribution; MEA recommends that a suitable term be developed and added to avoid underpredictions in the case of low copper content steels typical of improved steelmaking products. The experimental phases of the task are now completed; a final report is in preparation. Irradiated specimen materials are being , distributed world-wide to interested laboratories engaged in the ! search for radiation damage mechanisms. I 87
FUTURE RESEARCH PLANS TASK 1 - FRACTURE TOUGHNESS CRITERIA This task has been completed, l TASK 2 - ENVIRONMENTALLY-ASSISTED CRACK GROWTH IN LWR MATERIALS I I This task has been completed. l l TASK 3 - IRRADIATION SENSITIVITY AND POSTIRRADIATION RECOVERY l Investigations on effects of neutron exposure variables, primarily fluence rate effects and service temperature effects, have been extended by new task assignments to further support Regulatory Guide 1.99 and PLEX considerations. In addition, investigations will continue on mechanisms responsible for ' steel composition effects and interactions apparent in irradiation sensitivity, postirradiation annealing behavior, and reirradiation embrittlement trends to preclude future technical surprises. The investigations on fluence rate will compare notch ductility and tensile properties behavior of several prototype welds and plate materials at fluence rates approximating BWR service, PWR service and test reactor irradiation experiments. Fluence targets range from 1 x 1018 to 2.5 x 1028 n/cm 2 illustrative of BWR and PWR vessel exposures before EOL. Materials will represent a variety of copper, nickel and phosphorus contents as well as different primary compositions and/or welding materials in the interests of testing material variability. A special MEA experiment will irradiate specimens of the Gundremmingen vessel and Gundremmingen archival material together for a direct determination of the cause of the currently anomalous results for the service-irradiated material. A second group of new experiments will evaluate the significance to data bank analyses and usage, of 260 C vs. 288 C vs. 316 C irradition conditions and the effect of fluence rate on 260 C vs. 316 C irradiation effects differences. Mechanisms identification and modeling thrusts will continue to be pursued through subcontract efforts and through cooperative (j oint) programs with other ; laboratories worldwide. The new International Group on Radiation ' Damage Mechanisms in RPV Steels, put in place by MEA in 1988, has I objectives of promoting and guiding cooperative programs on mechanisms and the acceleration of research progress while reducing potential costs to countries via collaborations. 88 l l
l l REFERENCES
- 1. D. E. McCabe, " Fracture Evaluation of Surface Cracks Embedded in Reactor Vessel ' Cladding: Material Property Evaluations,"
USNRC Report NUREG/CR-5207, Sept. 1988. ..
- 2. J. C. Merkle, " Approximate Analysis of Ductile Crack Growth in a Nozzle Corner Region," a presentation to NRC Vessel and Piping Integrity Review Workshop, Oak Ridge, TN, June 1-5, 1981.
- 3. D. A.' Curry, "A Micromechanistic Approach to the Warm l Prestressing of Ferritic Steels," CERL Laboratory RD/L/N103/79, Central Electricity Generating Board, Berkeley, U.K.
Sept. 1979.
- 4. A. L. Hiser, " Correlation' of Cy and KIc/EJc Transition Temperature Increases Due to Irradiation," USNRC Report NUREG/CR-4395, Nov. 1985.
- 5. A. L. Hiser and G. M. Callahan, "A User's Guide to the NRC's Piping Fracture Mechanics Data Base (PIFRAC)," USNRC Report NUREG/CR-4894, May 1987.
- 6. A. L. Hiser, " Tensile and J-R Curve Characterization of Thermally Aged Cast Stainless Steels," USNRC Report NUREC/CR-5024, Sept. 1988.
- 7. J. R. Hawthorne and A. L. Hiser, " Experimental Assessments of Gundremmingen RPV Archive Material for Fluence Rate Effects Studies," USNRC Report NUREG/CR-5201, Oct. 1988.
- 8. A. L. Hiser and J. B. Terre 41, " Size Effects on J-R Curves for A 302-B Plate," USNRC Report NUREG/CR-5265, Jan. 1989.
- 9. ASME Boiler and Pressure Vessel Code, Nuclear Power Plant Components, Section III, Division 1, Subsection NB, Class 1 Components, American Society of Mechanical Engineers, New York, issued annually.
- 10. J. B. Terrell, " Fatigue Life Characterization of Smooth and Notched Piping Steel Specimens in 288 C Air Environments,"
USNRC Report NUREG/CR-5013, May 1988.
- 11. J. B. Terrell, " Fatigue Strength of Smooth and Notched !
Specimens of ASME SA 106-B Steel in PWR Environments," USNRC Report NUREG/CR-5136, Sept. 1988.
- 12. J. B. Terrell, " Fatigue Strength of ASME SA 106-B Welded Steel Pipes in 288 C PWR Environments ," USNRC Report NUREG/CR-5195, Dec. 1988.
l'
- 13. " Compilation of Contract Research for the Materials Engineering Branch, Division of Engineering - Annual Report for 1987,"
89 j l i
V e
'USNRC Report NUREG/CR-0975, Vol. 6, June 1988, pp. 98-100.
- 14. ~ M. G. Burke and M. K. Miller, " Solute Clustering and Precipitation in' Pressure Vessel Steels Under Low . Fluence Irradiation- Conditions," Journal de Physique, Vol. 49-C6, 1988, 1 p.283. l l
- 15. " Compilation of Contract Research for-the Materials Engineering i Branch, . Division of Engineering - Annual Report for 1987,"
USNRC Report NUREG/CR-0975,' Vol. 6, June 1988, pp.101-102.
- 16. J. R. Hawthorne, " Status Report on NRC/ MEA Dose-Rate Experiments in the Buffalo Reactor," ASTM STP 1001, . American Society for Testing and Materials, Philadelphia, PA, May 10, 1989.
- 17. J. R. Hawthorne, " Exploratory Studies of Element Interactions.
and Composition Dependencies in Radiation Sensitivity Develop-ment," USNRC Report NUREG/CR-4437, Nov. 1985.
- 18. Structural . Integrity of Water . Reactor Pressure Boundary Components ' - Annual Report for 1985," F. J. Loss, Ed., USNRC' I Report NUREG/CR-3228, Vol. 4, June 1986, pp. 174-181.
!I
.i 90
i
-DEGRADED PIPING PROGRAM - PHASE-II BATTELLE COLUMBUS DIVISION Principal Investigator: Gery Wilkowski Key Staff: J.:Ahmad, C. Barnes, F. Brust, D. Guerrieri, G. Kramer, M. Landow, C. Marschall,' R. Olson, V. Papaspyropoulos,-
P. Scott, and P. Vieth r l -. ABSTRACT This report summarizes the results from the Degraded Piping Program - Phase II. The program started in March 1984 and ended January.1989. .A
. summary of the research progress, its significance, and future needs are given. Recent efforts included two large diameter cold-leg pipe experiments which reached limit-load, and three artificially aged centrifugally cast stainless steel pipe experiments that also reached
-limit load.
I OBJECTIVE The overall objective of the Degraded Piping Program is to verify and l improve simple estimation schemes to predict'the fracture behavior of circumferentially cracked pipe. The program is limited to quasi-static fracture and cracks in straignt pipe. There are a variety of materials,
,- flaw geometries, pipe sizes, and loading conditions evaluated.
FISCAL YEAR 1988 SCOPE In 1988, efforts concentrated on completion of several prototypical circumferentially cracked pipe experiments. The results of these efforts completed the Degraded Piping Program - Phase II. These results and some of the other program highlights are described in the following sections.
SUMMARY
OF RESEARCH PROGRESS AND ITS SIGNIFICANCE The NRC's Degraded Piping Program - Phase II started in April of 1984. ( Its main objective was to validate the fracture behavior of { circumferentially cracked nuclear piping at quasi-static loading rates, j The validation involved conducting pipe fracture experiments at LWR conditions which were used to assess various fracture analyses. These data'were then used to validate and develop analysis procedures for leak-before-break or inservice flaw inspection criteria such as ASME Section ! XI IWB-3640 and 3650. 1 I 91
The full-scale experimental efforts involved conducting pipe fracture experiments on various nuclear grade pipes at LWR temperatures. A listing of piping materials investigated is given in Table 1. The pipe sizes ranged from 102-mm (4-inch) to 1067-mm (42-inch) diameter. Generally, the pipes were procured from cancelled nuclear power plants. In one case, a 711-mm (28-inch) diameter main recirculation pipe, which was removed from a BWR, was tested. Many of the experiments involved pipe pressurized with water at 288 C (550 F) under bending loads. Over , 61 experiments were conducted and documented in detailed data record books. The data record books consist of four volumes containing data for assessment of current analyses and any future analyses that may be developed. The material characterization efforts involved two aspects. The first aspect was to evaluate the material properties of each pipe necessary for flaw evaluation analyses. This typically included: Chemical anal Charpy V-notch tests, tensile tests, and compact (Tension) [C(T)]tests. yses, The C(T) specimen tests were conducted to obtain J-integral fracture resistance curve data for the pipe material at the pipe experiment test temperature. These data have subsequently been input into the NRC's data base on pipe material properties called PIFRAC. The second aspect of the material characterization efforts involved specific investigations, such as
- an evaluation of crack instabilities in ferritic steels at LWR temperatures (which are believed to be due to dynamic strain aging),
. development of a test method to evaluate the toughness in the through-thickness direction of a specimen which simulates the constraint condition in a surface cracked pipe, a evaluation of methods to extrapolate J-R curves from small C(T) specimen data,
+ assessment of material anisotropy on ductile fracture toughness, and a various round-robin activities to validate calculation and testing techniques.
l At the beginning of the Degraded Piping Program, it was believed that generally all nuclear piping failure stresses could be predicted by a limit-load analysis. The pipe fracture results from this program showed that this is not always true. A plastic-zone screening criterion was developed to show when limit-load failure was expected, and when elastic-plastic fracture mechanics would be needed to predict the potentially l lower failure stresses than those predicted by the limit-load analyses. Since that time, various engineering fracture mechanics analyses have been developed. These analyses use the J-integral fracture parameter and are frequently referred to as J-estimation schemes. The verification of these J-estimation schemes using pipe fracture data and more detailed finite element analyses have been a central focus of this program. The 92
finite element analysis efforts have also been evaluated in round-robins for both laboratory specimen and cracked pipe experiments. The material characterization, pipe. experiment, and analytical efforts from the program are summarized in the following sections. Afterwards, the significance of some of these results and future needs are discussed. Material Characterization Efforts
-The main focus of this activity was to provide material characterization data from laboratory-specimen tests for pipes subjected to full-scale fracture tests. Included were chemical composition, tensile stress-strain curves, Charpy V-notch transition curves for ferritic steels, and J-resistance curves. The mechanical property tests were designed to t simulate conditions. existing in the full-scale pipe fracture test, including notch acuity, crack plane orientation, test temperature, and rate of loading. Specimens were machined'from the same lot of pipe used in the pipe experiment; no mechanical flattening of the pipe was permitted. In addition to the data being used for analyzing pipe tests, they were transmitted to Materials Engineering Associates for inclusion in the NRC's Piping Fracture Mechanics Data Base (PIFRAC). Pipe materials that have been characterized in the Degraded Piping Program are indicated in Table 1.
In addition to conducting tests to characterize materials, several specific tasks were undertaken within the program to provide needed data. These tasks included: (1) Study of methods for predicting large-crack-growth J-R curves from small-specimen data. The Battelle study compared the . usefulness of deformation J (JD) and modified J (JM) for extrapolating J-R curves. It also developed an empirical-method for extrapolating J-R curves and revealed possible size effects on J-R curves, especially in weld-metal tests (Ref. 1). (2) Development of a special single-edge-notch test [SE(T)] to simulate a surface-cracked pipe bending experiment. Figure 1 is a schematic of the SE(T) specimen subjected to loading using rigid wedge grips. Note that the entire ligament is under tensile stress, just as it is in a surface-cracked pipe subjected to bending. In addition to developing experimental techniques for the SE(T) test, Battelle developed an estimation [- formula for' calculating a J-R curve from the test data, verified the formula by finite element analysis, and applied the test method to evaluation of cracks in welds (Ref. 2). An interesting result of SE(T) tests at 288 C (550 F) is shown in Figure 2. Notice that the crack in Specimens A8-5 and A8-7 began in the HAZ and displayed a crack-opening angle similar to that for the base metal specimen. However, when the crack reached the fusion line in Specimen A8-5, it continued to extend along that boundary while exhibiting a much smaller crack-opening angle. This behavior may be indicative of a 93
' minimum tou'ghness region at the fusion line in austenitic c stainless steel weldments.
(3). Study'of anisotropy effects on crack-growth direction in ferritic-steel pipes. This study was prompted by observations.
'in both pipe tests and C(T) specimen tests of cracks growing at a-large angle to the intended direction in ferritic' steels..
. Examination of a seamless pipe revealed nonmetallic inclusi_ons oriented ~at 20 to-30 degrees from the pipe axis,.apparently the result of twisting of the pipe during' hot-forming.. Testing of C(T) specimens machined from the pipe in several different orientations revealed that the fracture resistance had a-minimum value in the direction of.the inclusions.
(4) Participation in round robins with.other NRC contractors on-tensile testing, J-R curve. calculations, and use of the direct-current electrical potential method to monitor crack growth in C(T) specimens. During the course of the material characterization . studies, a number of
. interesting and, in some cases,. unexpected results.were obtained. One of the. unexpected findings was dynamic crack jumps in some of the carbon steels at LWR operating temperatures of 550 F (288 C), possibly.
associated with dynamic strain aging. These crack jumps occurred
' in intermittently both laboratory between periods specimen of slow fracture stable tearing toughness tests Figure (and 2, were observed upper curve) and full-scale pipe tests (Figure 3). Although the fracture mode was ductile during the. jumps, such instabilities are indicative of reduced toughness and could potentially led to sudden large leaks in flawed reactor pipes subjected to accident. conditions. If the jumps are associated with dynamic strain aging, it is likely that their ocurrence will be.a function of both strain rate and temperature.
Other. interesting results included: (1) flux welds are much less. tough' than welds made using inert gas in both austenitic and ferritic steels, and (2) most of the carbon steel pipes tested in the Degraded Piping Program exhibited tensile strength values at 300 and 550 F (149 and 288 C) that were greater than those at room temperature; this result indicates that many carbon steel pipes used in nuclear plants are susceptible to dynamic strain aging. Susceptibility to dynamic strain aging may be accompanied b several undesirable characteristics at LWR temperatures, including: 1) dynamic crack jumps, (2) reduced crack-initiation toughneu , and 3) reduced tearing modulus (Ref. 3). On the other hand, a beneficial effect from dynamic strain aging is an increase in the material's strength at elevated temperatures, i Pipe Fracture Experiments The scope of the full-scale pipe experiments included: (1) obtaining data on circumferential1y cracked pipe, (2) using representative pipe materials,' crack geometries, and loading conditions, (3) performing all L l 94
experiments at elevated temperatures, (4) determining crack initiation, maximum load and. crack growth in each experiment, (5) comparing experimental loads relative to net-section-collapse predicted failure loads, and (6) establishing a comprehensive pipe fracture database for evaluating existing analytical fracture models. In addition to verification of existing analysis methods, the extensive pipe fracture database will be useful for evaluating fracture mechanics l parameters developed or modified in the future. This database has i already been useful in evaluating flaw assessment criteria for cracks i found in service, and has been used to benchmark the criteria presented in Section XI of the ASME Code; Article IWB-3640 for austenitic pipe and proposed Article IWB-3650 for ferritic pipe. The test matrix consisted of 61 full-scale pipe fracture experiments conducted at Battelle's laboratories in Columbus and West Jefferson, ; Chio. Flaw geometries included through-wall cracks, surface cracks and complex cracks. (A complex crack, as defined in this program, is a long internal surface crack that has propagated through the wall thickness for a short distance.) Figure 4 illustrates the number of experiments performed as a function of pipe diameter and crack geometry. Evident from this figure is the wide range of pipe diameters studied in this program. Figure 5. presents the data in a similar format, except that the number of experiments is shown as a function of loading method rather than crack geometry. Loading methods included four-point bending, pressure, pressure and bending, and bending with a high system compliance. Quasi-static loading rates were used in all cases. A number of representative materials were evaluated in the pipe fracture experiments as indicated in Table 1. Figure 6 shows the breakdown of experiments by material type. Pipe material wall thicknesses ranged from 0.25 to 3.41 inches (6.4 to 86.6 mm). The test matrix rationale was to perform the simplest experiments early in the program and increase the complexity of the loading conditions and flaw geometries as the program progressed. Thus, many of the first year experiments evaluated through-wall-cracked pipe under simple bending, while the third and fourth year experiments evaluated surface cracks in welds under combined pressure and bending. An example of a high-energy experiment is shown in Figure 7. This photograph shows the decompression behavior of a 16-inch (406-mm) { diameter pipe after an internal surface crack has broken through the wall of the pipe under combined pressure and bending loads. In this particular experiment, the test pressure was 2,250 psig (15.5 MPa) which is representative of PWR primary piping. The crack was embedded in the center of a low-toughness submerged-arc weld (SAW). Two recent experiments have evaluated the load-carrying capacity of large diameter cold-leg pipes under simple bending loads. The first experiment 95
evaluated a through-wall crack in the base metal, while the second experiment evaluated a similar crack in a shop manufactured weld. These experiments were challenging due to the extremely high bending moments required to propagate a crack in such a heavy-wall pipe. Figure 8 is a post-test photograph of the second cold-leg pipe experiment. A through-wall crack was tested in the centerline of an SAW in this experiment. A close-up of one crack tip is shown in Figure 9. Both crack tips initiated at the mid-wall of the pipe weld, but quickly grew out of the SAW weld and into the pipe base metal. As is evident from Figure 9, the ! crack growth direction was erratic, changing in a zig-zag fashion across the weld. Results from all pipe experiments have been incorporated into an extensive pipe fracture database. Data collected and reported in this database consist of applied loads, load-line displacements of the test machine, internal pipe pressure, crack-opening displacements, direct-current electric potential measurements (for crack growth monitoring), rotations of the cracked pipe section, and pipe temperature. In addition, the pipe fracture database also summarizes pertinent material property data, such as chemical analyses, tensile results, Charpy V-notch impact data, and J-resistance curve data. Five experiments from other research programs have been incorporated into the database in addition to the Degraded Piping Program experiments. Crack initiation and maximum moment data from all pipe experiments have been compared against limit-load analyses such as the net-section-collapse method. Figure 10 illustrates such a comparison for the two cold-leg pipe experiments discussed earlier. Net-section-collapse calculations were based on two values of flow stress for these materials. Both the actual tensile property data and the ASME code design stress were used to define the flow stress of the material. In the case of the weld metal experiment, both the base metal properties and the weld metal properties were used in the limit-load calculations. Thus, two sets of bars are shown in Figure 10 for the weld metal cold-leg pipe experiment. Comparisons to net-section-collapse predictions showed that, in both , cases, experimental maximum loads exceeded the predicted loads when j either the base metal data or design stress were used. j These comparisons to limit-load analyses have led to the development of a plastic-zone screening criterion to determine when net-section-collapse analyses are valid and when more complex elastic-plastic analyses are j warranted. Such a criterion is shown in Figure 11. In this figure, experimental stress data are normalized against net-section-collapse predicted stresses for both through-wall-cracked and surface-cracked pipe. The data arc plotteo as a function of a non-dimensional plastic l zone parameter. A statistical analyses was performed on these data and a l lower bound failure curve wc, defined with a 95 percent confidence level. l This is shown by the solid line in Figure 11. Such a criterion shows l that even high-toughness stainless steel materials can fail below net-section-collapse predictions if the pipe diameter is sufficiently large. l l i 96 l l
It also shows that surface-cracked pipes are less sensitive to toughness than are through-wall- cracked pipes. A variety of other experiments were performed within this phase of the program. The crack instability behavior of pipe under bending was evaluated in several experiments. Conditions of unstable crack propagation were observed once maximum moment was achieved in pipes with long internal surface cracks or with complex cracks (such as found in the Duane Arnold nuclear plant). This instability behavior resulted in complete double-ended guillotine breaks in two experiments, and was found to be quite sensitive to changes in the compliance of the four-point-bending test machine for these two crack geometries. Encouraging results were obtained from experiments in which circumferentially cracked pipe had been repaired by the weld-overlay-repair technique. In these exp-eriments, a fatigue crack was grown completely through the wall of a stainless steel pipe and half-way around the circumference. A multiple pass TIG weld overlay repair was fabricated over the crack. Post-test results from these pressure and bending experiments showed tremendous plasticity in the base metal before crack initiation occurred. The results were also used in a round-robin design analysis for weld-overlay repairs. The round-robin predictions showed that various vendors and engineering firms made consistent predictions that agreed well with the experimental results. Another series of experiments evaluated cracks in low-toughness submerged-arc welded pipe. These pipe fracture experiments had lower failure stresses than predicted by J-estimation scheme analyses using base metal material property data. These results suggested that both the material properties of the base metal and the weld metal must be considered when analyzing cracks in a non-homogeneous material. Finally, an interesting phenomenon was observed in a number of ferritic steel pipe specimens containing through-wall cracks under bending. Nearly all of these experiments showed significant out-of-plane crack growth behavior. In the case of one cold-leg pipe specimen (similar to the experiment shown in Figures 8 and 9), the through-wall crack turned from the circumferential direction and began propagating along the axis of the pipe. Thus, the crack followed the direction of lowest toughness rather than the direction of greatest principal stress. This result reveals that the material anisotropy due to the rolling and forming process is a significant factor in predicting the fracture behavior of through-wall-cracked pipe under bending. Analytical Efforts The fracture mechanics analysis effort in the Degraded Piping Program - Phase II was aimed at providing the NRC with simple engineering models for predicting the integrity of nuclear power plant pipes containing cracks. The majority of the work was focused on predicting the behavior of circumferentially oriented cracks in pipes subjected to predominantly 97
bending loads. A limited amount of effort was also devoted to combined internal pressure and bending loads. Through-wall cracks, part-through-wall internal surface cracks, as well as complex cracks in different pipe sizes and materials were considered. Both base-metal and weld cracks were analyzed. The majority of the modeling effort was based on the J-integral tearing instability (J/T) approach of elastic-plastic fracture mechanics (EPFM). The application of the J/T approach requires two types of analytical l models. These are for (a) establishing curve using experimental data, and (b)g calculating an appropriate applied J and J-resistance the (J-R) tearing modulus (T) for the crack-structure configuration of interest. In the Degraded Piping Program, the work within Type (a) analyses included several activities. For example, finite element analyses were performed to aid the development of a simple method for extrapolating J-R curves for large amounts of crack growth from laboratory specimen tests (Ref. 1). Elastic-plastic finite element analyses of 1.0 inch (25.4 mm) thick 1T, 3T, and 10T planar dimension specimens of both austenitic and carbon steels were performed (Ref. 1). An example of the (far-field) J-resistance curves resulting from these analyses on Type 304 stainless steel at 550 F (288 C) are shown in Figure 12. Such results were used in deciding how a J-resistance curve from a 1T specimen can be extrapolated to larger crack growth amounts normally encountered in through-wall crack , analyses of pipes. The 10T specimen was also analyzed by a number of participants in pn international analysis round-robin organized by . Battelle (Ref. 41. Figure 13 shows a key result of the round robin in l the form of predicted load versus applied displacement plots. The round-robin results also indicated that in general the far-field J values for large crack extension were in better agreement with the modified (JM) values rather than deformation (JD) values obtained by estimation metnods. The Type (a) analyses also included the evaluation of a single-edge notch specimen loaded under fixed-grip boundary conditions as a candidate for characterizing radial growth behavior of surface cracks in pipes j (Ref. 2). The study suggested that in this specimen, as well as in circumferential1y surface-cracked pipe in bending, there may be little or no J-controlled growth. If this is the case, further research aimed at more accurate prediction of surface-crack behavior in pipes may be l needed. ! Another activity within Type (a) analyses was aimed at developing and comparing J-R curves for cracks in welds by finite element analyses and by estimation methods. The results of this activity are reported in = Reference 5. The work in Type (b) analyses resulted in improved methods for analyzing pipes with through-wall cracks in bending. These methods, called LBB.GE and LBB.ENG (Ref. 6), and several other estimation analysis methods were evaluated by comparing their predictions with experimental data. Figure 14 represents an example of such evaluations in terms of predicted load versus displacement curves for a specific experiment. In terms of 98 l 1
1 I i initiation and maximum load predictions, it was found that, with a few exceptions, the EPRI/GE estimation methods for through-wall cracked pipes gave underpredictions when used in conjunction with JD resistance curves. Altogether, six different through-wall crack analysis methods were evaluated. A computer code called NRCPIPE, which allows convenient application of these methods for LBB evaluations, was developed for a personal computer. In addition to estimation analyses, three-dimensional finite element analyses of through-wall cracked base metal as well as welded pipes were performed. A stainless steel 16-inch (406-mm) diameter pipe fracture experiment was analyzed by a number of participants as part of one of the three international round robins organized by Battelle. Figure 15 shows a key result of the round robin in the form of predicted load versus applied displacement plots. Note that a general trend from this effort, as well as other efforts, is that the FEM analyses underpredict the experimental results for circumferential through-wall-cracked pipe in bending. For circumferentially surface cracked pipes in bending, new J-estimation methods were developed (Ref. 7). These developments effectively utilized an existing two-dimensional estimation method to solve the three-dimensional surface crack problems. The predictions of these methods were compared with experimental data. Figure 16 shows an example of such a comparison. Significance of Results At the beginning of the Degraded Piping Program - Phase II, the status of the pipe fracture mechanics methodology was quite limited. For example;
- the net-section collapse (limit-load) analysis for circumferentially cracked pipe was developed and verified on small diameter stainless steel pipe,
- the EPRI/GE J-estimation elastic-plastic fracture mechanics analysis was developed, but had little verification, a the NRC.LBB analysis method was under development, I
l
- there was a limited amount of material property data, and !
a there was a limited amount of pipe fracture data. Some of the developments during the course of the Degraded Piping Program
- Phase II are:
a the material property data base was greatly expanded, e a detailed data base for circumferentially cracked pipe under slow monotonic loading was expanded, 99
i
- new J-integral based engineering analyses were developed and j L verified for circumferential1y through-wall-cracked pipe stability evaluations used in LBB assessments, a a finite length surface-cracked pipe elastic-plastic fracture )
mechanics engineering analysis was developed, j
. the accuracy of, finite element analyses of cracked pipe was assessed, e an energy balance method was developed to' predict the start of an
. instability ~and estimate the_ magnitude of crack growth in an instability event for both through-wall and surface-cracked pipe under' combined load-controlled and displacement-controlled stresses,
. a-statistically based fracture analysis method to predict maximum loads-of circumferential1y cracked pipe was developed,
. a methodology to predict the fracture of circumferential1y cracked piping systems subjected.to dynamic loading such as earthquakes was initially conceived,
+ a constraint correction for circumferential complex-cracked pipe was ~
developed, and
- an assessment of Charpy versus JIc correlations was made using data developed in this program.
These developments have subsequently been used to assess Leak-Before-Break (LBB) fracture' analyses, and in-service flaw inspection methods. LBB developments include evaluation of through'-wall circumferential cracked pipe J-estimation schemes, and development of experimental data to verify crack opening areas for leak-rate predictions. In-service flaw inspection criteria included assessments of the ASME Section XI IWB.-3640 , austenitic flaw evaluation procedure, and the IWB-3650 ferritic flaw > evaluation procedure. This involved assessment of the inherent safety factors in the ASME analyses and development of fracture toughness data for development of reasonable lower bound material properties in the ASME 1
. analyses. Although numerous changes and improvements have been made to the ASME criteria, further improvements to unify and make secondary ;
corrections are still needed. Another assessment was on the weld-overlay repaired pipe fracture analyses. Fracture experiments were conducted on prototypical weld-overlay repaired pipe that initially had large circumferential through-wall cracks. This effort validated the NRC's NUREG-0313 Rev. 2 criteria for weld-overlay fracture analyses. 100
r. FUTURE RESEARCH PLANS The Degraded Piping Program - Phase 11 is now completed. However, there are several areas of piping integrity that are in need of further efforts. Some of these resulted from the Degraded Piping Program results, while others were not within its scope. These are briefly described below. The first area in need of further evaluation is fracture and crack-opening-area analyses of pipe fittings. Effort is needed to develop verified analyses for the LBB and in-service flaw inspection criteria of
-fittings. LBB and in-service flaw inspection criteria are currently based on the predictions of cracks in straight pipe. Pipe systems, however, may have cracks in fittings, i.e., elbows, tees, etc. Analyses verified by experimental data are needed for complete LBB analyses, and for future ASME Section XI Code flaw acceptance criteria. l The second area in need of further study is integration of the pipe fracture program results with the Piping Reliability Program results.
The objective would be to assess the significance of possible ) interactions between pipe fracture studies and suggested code changes i from the EPRI/NRC Piping Reliability Program. The EPRI/NRC Piping Reliability Program has recently completed efforts to suggest changes in the design rules for allowable stresses in nuclear piping. The basis for this suggestion comes from uncracked pipe fitting and pipe system tests. The Degraded Piping Program has developed considerable methodology for assessment of cracked pipe under quasi-static loading, while the NRC's IPIRG program is examining the behavior of cracked piping systems under dynamic loading. There has been very little consideration of acceptable flaws (which are much smaller than the Degraded Piping Program or IPIRG flaw geometries) on the suggested changes from the Piping Reliability Program. Thus, additional study is needed to examine the safety margins using the suggested pipe design stress rules if there are small flaws in critical areas. A third area in need of additional efforts is development of a replacement criterion for the DEGB design rule. The elimination of the DEGB design criterion for pipe whip restraint and jet impingement removal has been a major step forward in improving piping safety and at the same time making plants more economical to run. The next logical step is to apply the elimination of the DEGB toward defining a maximum credible leakage area for more realistic designs for equipment qualification, pipe I support design, redesign of internal reactor core supports for lower ! depressurization loads, etc. This effort will develop methodology that ! could be used to assess future licensing requests for replacement of the DEGB design rule. A fourth area of further activity is to resolve issues from the Degraded Piping Program - Phase 11 that have not been sufficiently addressed. The Degraded Piping Program - Phase Il program contained over 70 technical subtasks. Within those efforts many concerns were resolved, while others could not be resolved within the scope of the program. Furthermore, some 101
of these issues were discovered during the course of the program. This effort is needed to assess the more significant issues for LBB or in-service flaw inspection criteria assessments. Some of the more significant issues are:
. improvements in circumferential through-wall cracked pipe analyses for short crack lengths applicable to LBB analyses, a improvements in the understanding of dynamic strain aging on the fracture behavior of carbon steels, a assessment of the effects of anisotropy on fracture behavior of ferritic piping steels, a generation of additional quasi-static straight pipe fracture data (i.e., large diameter pipe with surface cracks),
. generation of additional ferritic steel weld, bimetallic welds and fusion line toughness data, etc.,
a refinements to surface-cracked pipe estimation schemes for combined loading,
- generation of a data base on cyclic loading and dynamic rate effects on the strength and toughness of nuclear piping material, a resolution of discrepancies between finite element analyses of cracked pipe experiments and experimental data, and a development of improved criteria to uniformly handle crack-opening area and fracture predictions for welds.
The last area of effort that needs to be continued is validation of ASME Section XI Code procedures as well as coordination with NRC-NRR, ASTM and other NRC contractors. The objective of this effort is to assess proposed changes to the ASME Section XI Flaw Evaluation Working Group, and to coordinate pertinent research developments with the Working Group if they could impact the Code. For instance, at the September 1988 Pipe ! Flaw Evaluation Task Group meeting, a list of over 20 " maintenance" items for changes to the stainless steel and ferritic piping criteria was made. In order to implement these changes, input from the past Degraded Piping Program and additional new items is necessary. l REFERENCES
- 1. Papaspyropoulos, V., Marschall, C., and Landow, M., " Predictions of J-R Curves with Large Crack Growth from Small Specimen Data",
Technical Report for NRC Degraded Piping Program, prepared by Battelle Columbus Division, NUREG/CR-4575, September 1986. l 102
i
- 2. Wilkowski, G. et al., " Degraded Piping Program - Phase II", Sixth Program Report, October 1986-September 1987, prepared by Battelle Columbus Division, NUREG/CR-4082, Vol. 6, April 1988.
- 3. .Marschall, C., Landow, M., and Wilkowski, G., "Effect of Dynamic Strain Aging on Fracture Resistance of Carbon Steels Operating at Light-Water-Reactor Temperatures", paper presented at 21st National Symposium on Fracture Mechanics, Annapolis, MD, June 28-30, 1988.
- 4. Ahmad, J. et al., " Elastic-Plastic Finite' Element Analysis of Crack Growth in large Compact Tension and Circumferentially Through-Wall-Cracked Pipe Specimen", Technical Report for NRC Degraded Piping Program, prepared by Battelle Columbus Division, NUREG/CR-4573, October 1986.
- 5. Nakagaki, M., Marschall, C., and Brust, F., " Analysis of Cracks in Stainless Steel TIG Welds", Technical Report for NRC Degraded Piping Program, prepared by Battelle Columbus Division, NUREG/CR-4806, l
December 1986.
- 6. Brust, F., " Approximate Methods for Fracture Analyses of Through-Wall Cracked Pipes", Technical Report for NRC Degraded Piping Program, prepared by Battelle Columbus Division, NUREG/CR-4853, February 1987.
- 7. Scott, P. and Ahmad, J., " Experimental and Analytical Assessment of Circumferentially Surface-Cracked Pipes Under Bending", Technical Report for NRC Degraded Piping Program, prepared by Battelle Columbus Division, NUREG/CR-4872, April 1987.
i l 1 l 103
TABLE 1. PIPES SUBJECTED TO FULL-SCALE TESTS AND TO MATERIAL CHARACTERIZATION STUDIES IN DEGRADED PIPING PROGRAM Diameter in Inches Wrought Ferritic Pipe ASTM A1068 6 and 16 ASTM A333, Gr. 6 4, 10, and 24 ASTM A516, Gr. 70 28 and 37 SA Welds in Wrought Ferritic Pipe ASTM A516, Gr. 70 37 ASTM A106B 8 and 16 Wrought Austenitic Pipe Type 304 stainless 4, 6, 16, and 42 Type 316L stainless 16 Inconel 600 6 Cast Austenitic Pipe CF8M stainless .12 and 16 SA Welds in Wrcught Austenitic Pipe Type 304 stainless 6 and 16 GTA Weld in Wrought Austenitic Pipe 4 GTA Weld Overlay on Wrought Austenitic Pipe Type 304 stainless 6 i l
\
104 ______ L
/
-e- c:I -
s *
/,
w-c: a i
/ {
FIGURE 1. SCHEMATIC OF SE(T) TEST TO SIMULATE SURFACE CRACKED PIPE BENDING EXPERIMENT
.t Type 304 <:> g 5 t a i r. l e b s 5' I U P I i; a s e Mttal
' 'g ~5 j-
- AB-12 '"',-
5
@g
~
~e 1
.N o Heat Affected Zone *'
~
\
As-5 m q ._ wejo metal e' p-E _.. v FIGURE 2. H0 PH 0 C ~ (D AT 550 F 28 C) 105 i
Displterment, mm 10000 , Pipe DP2-F26 (ASTM A155) : 9000 - made fr'om ASTM A516 Gr'. 70 steel - 40000 ;
~
Tested at 288 C (550 F) - ; 7000 - Ba '= 16.9 mm (0.664 in) ' .i 30000 6000 - j - 25000 2 g_5000 g -
~$ -
20000 3 - 4000 - I _I n a 48 (300 0
~
N 3000
, Bu = 15.4 mm (0.605 in) 10000
-2000 1000 -
5000 0 O. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Displacement. In FIGURE 3a. UNSTABLE CRACK GROWTH IN COMPACT SPECIMENS LORD-LINE DISPLACEMENT, mm e se les Ise zee ese see ase 1s8888 1 4 4 4 I i I Crack initiation - See 125een -
.3 z sea a plesees -
f 3 - 4ee 3, < a m l g 75ees - W g - 3ee d
& e 500se -
3 d - roe Et F e 25ees - les a e e 3 6 S 12 15 LORD-LINE DISPLACEMENT, Inches FIGURE 3b. UNSTABLE CRACK GROWTH IN PIPE FRACTURE TEST AT 550 F (288 C) (THE PIPE IS THE SAME PIPE AS IN FIGURE 3a) 106 1 l j
.______________--______-__--__-__N
i i 4
~32 30 - -
LOADING CONDITION ~ 28 - - G Bordng 26 - M Pressure
' 24 - M Pressure and W
-3 22 - gg4 20 -
'f 18 -
U 16 - 14 - 12 -
. 10 -
8-6-
-0 i .i 4 6 10 . 12 16 28 37 42 Pipe Olometer. Inches FIGURE 4. TEST MATRIX FROM THE FULL-SCALE PIPE FRACTURE EXPERIMENTS SHOWING THE NUMBER OF EXPERIMENTS BY. DIAMETER AND LOADING TYPE 22
~
CRACK GEOMETRY Q Through-Wall Cracks M Surface Crocks
@ Complex Crocks 20 -
f S 18 - ' 16 - 14 - 12 - l $ 10 - ( 8-b b M W W W/
~
0 7 77, / i i i i i i i > 4 6 10 12 16 28 37 42 Pipe Diometer, lnches FIGURE 5. TEST MATRIX FROM THE FULL-SCALE PIPE FRACTURE EXPERIMENTS SHOWING THE NUMBER OF EXPERIMENTS BY DIAMETER AND CRACK GEOMETRY 107
32 30 - 28 - / 3 * "'S'"'- ' 26 - St intess Steel - Welds 24 -
\ E Carbon Steel- Base Metal 22 -
Carbon Steel - Weeds 20 - 18 - c - W 600 ~! [ i 16 -
, 14 -
12 -
$ 10 -
b$r 4 8 6 h m ? m m m= 10 12 16 Ploe Diometer, Inches 28 37 42 FIGURE 6. TEST MATRIX FROM THE FULL-SCALE PIPE FRACTURE EXPERIMENTS SHOWING THE NUMBER OF EXPERIMENTS BY DIAMETER AND MATERIAL TYPE
, p b
t>lg , 1
)
p 4
'$ ^
~' /i
-[.
p . f ., :
- d , + = _ ~
t4 -
~:or.m _.
i g i e IR ., . q){;y p>7 v gw ;n ,
' < i
( j l i . I l _- _ . ~ , , , , .. FIGURE 7. HIGH-ENERGY PIPE FRACTURE EXPERIMENT CONDUCTED UNDER ; PRESSURE AND BENDING AT ELEVATED TEMPERATURES 103 l ___-__________ _ a
_ _ . , _ - , _ - _ _ _ _ - _ , - . , _ _ . - _ _ _ - _ . _ , .,,,,.--..-y_,.--,__ --7-_,___-m
-.y a
4 i-18 "
%se% :6;/ ;.s 9 E, ,
kW +h ,
, . - , . , "-'*edyg
~~a que
~ m ;s m- .. .
=-
. e. , .. ..
g r= E . y - .7 #Y: --+;; .[Shzi NN
..:rtM;&
-m
' M u a[,l %
. ,., j FIGURE 8. POST-TEST PHOTOGRAPH OF A LARGE DIAMETER COLD-LEG l EXPERIMENT CONDUCTED UNDER BENDING AT ELEVATED !
TEMPERATURES. l ,,O f,f ~~
\, .q.3%
~ .
s tr pi.g-
~
. .j- k y# -_ %'[
.s' y ~' b j
Q ;
- g. ge .g
- ; - n he y-..#[fyMd l ;
y. If ,.,, A ~M V-
.,~
, ,pum ;ighi%,]p;j f y #.e .
s ,. '
~'
x\'
- _ , p; g sgg 4 .- _
~
.)
FIGURE 9. CLOSE-UP OF ONE CRACK TIP FROM THE COLD-LEG EXPERIMENT SHOWN l IN FIGURE 8. THE THROUGH-WALL CRACK IN THIS EXPERIMENT WAS MACHINED INTO THE CENTERLINE OF A SUBMERGED-ARC WELD 109 1 l l
- l. . . . . .
1.2 1.1 - 1-0.9 -' , 0.8 - 0.7 - 0.6 - 0.5 - 0.4 - 0.3 - 0.2 - 0.1 - 0 , . Bose Me i r rtes Met i ere h tes
@ (Yeid + Ultanatel/2 M 24mA E Desgn FIGURE 10. COMPARISON OF EXPERIMENTAL LOAD DATA FROM TWO COLD-LEG EXPERIMENTS TO NET-SECTION-COLLAPSE PREDICTED LOADS I4 o a $
i _. o i G O o O O to g 5 8 o I os 6 3- @ e l 95% confdence leel for in-sec expresson for att eta: C=3 4 o6 s o4 02 - Sur e croc do o I I I I I I ' o 25 30 35 40 O 05 10 45 20 Simplifed Dmensionless Plaste-Zore Pararneter, FIGURE 11. EMPIRICAL SCREENING CRITERIA DEVELOPED FROM THE FULL-SCALE PIPE FRACTURE EXPERIMENTS USING A DIMENSIONLESS PLASTIC-ZONE PARAMETER AND A FLOW STRESS VALUE = (oy + uo )/2. 110
.CrEck Extension,-mm 25 se 7s see 125 ise 17s 3,, a I i i i i r.
2e A o IT ~ o 3T les - A IST 4 A A A 15 .j A se - l a 1 m ' A N E 88 - A - is c m'
' 4
$ D j{
4e _ g Ao
- o. -
5 g AD 2e -e l a 1 2 3 4 5 6 7, Crack Extension, inches FIGURE 12. COMPARISON OF FAR-FIELD J-RESISTANCE CURVES FROM FINITE ELEMENT ANALYSIS FOR THE 1T, 3T, AND 10T NONSIDE-GROOVED TYPE 304 STAINLESS STEEL SPECIMENS AT 550 F (288 C) DISPLACEMENT, mm 50 75 100 125 158 175 200 225 a 25 188 i i e i i i i i Emperiment - 40a se - Finite Eiernent Results by: 88 - X Participant I - 350
% $KMgy o
- l* Xx M Participant 2 78 -
D g
- Participant 3 -
300
, x
.o o o Participant 4 l
Se - x o Porticipont 5 - 250 3 M D 0
- d 3- 59 D W X -
200 @ a bD o . ,,,J f cr 4e p X 3 J P X 150 se x
,, ( * " -
100 50 le ' !- l )
. . . . e 5 6 7 8 9 l
W l 2 3 4 l DISPLACEMENT, inches q l FIGURE 13. COMPARISON OF THE FINITE ELEMENT ANALYSIS RESULTS FOR THE l C(T) SPECIMEN WITH EXPERIMENTAL DATA 111 , _______.__._-_m _ _ _ _ _ _ _ _ _ _ _ _ _ - *
' 60 0 G.E. a LBB.GE 50 - + PARIS X LBB.ENG2
- LBB.NRC V EXP. DATA E
s e-s 30 -
- a. !
20 - a sN ' 10 - d 0 , , , , , , 0 10 20 30 LOAD-UNE DISPLACEMENT. inch FIGURE 14. COMPARISONS OF PREDICTED LOAD-DISPLACEMENT CURVES USING POWER-LAW EXTRAPOLATION OF J -R CURVE TO EXPERIMENTAL RESULTS(16-INCH [406-MM]DIkMETERTP304 STAINLESS STEEL PIPE WITH TWC IN CENTER OF SAW) LORD-LINE DISPLACEMENT, mm s_ tes 2ee 3ee 4ee see see 70s see i i i i i i i oND
- seese -
a a
, o D
o o g 5a , 2ee 49ees - "N a , o 9 + ^ d d M 3Mee -
,+#
A M a 15e f c Corrected esperrnental resJts l
,, _ re tw %
see $ a o Pr,tc i d h D Portcoont 2 $
$ o Portccont 3
+ Portcoant 4 -
Se 8-leses '
, p,,,,,,, 5
,a .
5 1e 15 2e 25 3e 35 , LORD-LINE DISPLACEMENT, Inches l FIGURE 15. COMPARISON OF THE FINITE ELEMENT ANALYSIS RESULTS TO 16-INCH (406-MM) DIAMETER CIRCUMFERENTIALLY THROUGH-WALL CRACKED PIPE EXPERIMENTAL DATA 112
.l
( 2 ( .1.8 - G - t 1.s
)
j1.4 -
)
~= -
t' 1.2
; 1 y
3~ o n 5'.s 3 e h, o u g o& n." .4 -
.2 -
g , , 9 2 4 6 8 10 12 14 16 18 20
~ Rg /t FIGURE 16. PREDICTION OF MAXIMUM LOAD BY THE SC.TKP METHOD FOR CIRCUMFERENT.lAL SURFACE-CRACKED PIPE IN PURE BENDING l
l l 113
FY-88 ANNUAL REPORT l Elastic-Plastic Fracture Mechanics Evaluation of l Light-Water-Reactor (LWR) Alloys f I David Taylor Research Center Code 2814 Annapolis, MD 21402 E.M. Hackett and J. A. Joyce ABSTRACT: A series of J-R curve and blunt notched specimen tests were performed on several alloys with varying levels of tou6 hness, to examine the extent of J-controlled crack growth. Based on the concept of an experimentally defined loss of singularity, crack growth on the order of 30% of the remaining ligament'has been shown to be within the J-controlled
-region. Beyond this region, small specimen J-R curve extrapolations, based on fitting a simple power law to the data over the extended 30% validity region, have been shown to provide accurate to conservative approximations of large specimen J-R curves. Evaluation of the fracture behavior of aged cast stainless steel at 5720F has shown JIC values in the range of 1500 in-lbs/sq.in.
to 1700 in-lbs/sq.in. and little effect of notch acuity on the J-R curve behavior. Ductile crack growth beginning from electrical discharge machined (EDM) notches in compact specimens of the stainless steel was shown to exhibit J-R curve behavior which.was consistent with crack growth emanating from a fatigue precrack. Results from the NRC contractor DCPD J-R curve round robin have been analyzed and show excellent inter-laboratory agreement for two different alloys despite significant differences in experimental and analytical test procedures. The round robin will serve as the basis for j modifying the ASTM J-R curve standard (E1152) to include a suggested DCPD test ; procedure. Investigation of the transition fracture behavior of A533B under dynamic loading rates has shown that specimen size has only a small effect on the measured fracture toughness when the tests were performed at impact loading rates. The position of the ductile to brittle transition region defined from these tests was shown to correlate well with that defined previously by crack arrest tests at the Oak Ridge National Laboratory (ORNL) I and the National Bureau of Standards (NBS). OBJECTIVES: Extend current J-R curve crack growth validity limits and develop R curve extrapolation techniques; resolve selection of either J or Jg as a fracture characterizing parameter for pressure vessel and piping alloys; analyze results of NRC DCPD J-R curve round robin to assess accuracy and consistency of the various DCPD techniques; evaluate effects of aging and notch acuity on the fracture behavior of cast stainless steel; report results of characterization of the fracture behavior of A533B steel in the transition region. l 114 L
--___-_____-____a
FY-88 SCOPE:
- 1. Extend current J-R curve crack growth validity limits and develop 1 techniques for extrapolating R curves from small specimens to large crack extensions. Coordinate research among NRC contractors and other interested parties on 'the general topic of use of J/Jg-R curves for characterization of
'the fracture resistance of nuclear pressure vessel and piping materials,
- 1. Quantify the effects of aging on the fracture toughness of cast stainless steel. Evaluate the influence of notch acuity on the ductile fracture l behavior of the alloy. l R
L l 3. Complete NRC contractor round robin for DCFD J-R curve testing to assess repeatability and accuracy of the test procedures. Analyze and report the results. l
- 4. Report results of static and dynamic fracture toughness tests of A533B steel in the upper transition region. Correlate results with Dak Ridge National Laboratory (ORNL) wide plate crack arrest tests.
l
SUMMARY
OF RESEARCH PROGRESS: EXTENSION OF THE LIMITS TO J CONTROLLED CRACK CROWTH The objectives of this investigation were to; (1) present experimental results and development of an experimental analysis which attempt to define the limits of the J singularity controlled region of crack extension in bend type laboratory fracture mechanics specimens; and (2) develop and examine a J-R curve extrapolation technique. Current limitations on J-controlled crack growth as described in ASTM E1152-87 are highly restrictive (crack extension is required to be less than 10% of the initial specimen remaining ligament). This is particularly limiting in fracture analyses of nuclear plant structures which require development of J-R curves with-large amounts of crack extension for use in instability analyses. Moreover, the E1152 limitations have not corresponded to observed experimental phenomena that could be taken to identify the loss of the singularity in a bend type fracture mechanics specimen. Also, there are at present, no standardized approaches for extrapolating J-R curves obtained from small specimens to predict the crack growth resistance of larger specimens or structures. The experimental test matrix included J-R curve tests on 1/2T, IT and 2T compact [C(T)] and three-point-bend [SE(B)] specimens of several medium to high strength steels (ASTM A710, A516, A106, and A533B steels and a high yield strength, 3%Ni steel). The J-R curve tests were conducted in accordance with ASTM E1152g87 with the exception that crack extensions on the order of 60% of the initial remaining ligament were achieved. The data were analyzed both in terms of the deformation theory J integral and the modified J integral (Jg) . Tests of blunt-notched C(T) specimens were also performed to examine the accuracy of deformation plasticity assumptions at large deformations and crack extensions. l 115
'J-R Curve Tests The J-R curve testing performed for this investigation was extensive and only partial results on one representative alloy will be presented here. The test matrix for the entire test series is presented in Table 1. All J-R curve testing was performed using the crack growth corrected J equations found in ASTM E1152-87. The -equations for Jg were taken from Ernst (1). Figure 1 provides J-R curves for different scale compact specimens of A533B steel. The corresponding Jg-R curves are provided in. Figure 2. A key feature of these data sets is the consistency of the J-R curves and.the inconsistency.of the JM-R curves. The distinct upward sweep to the Jg-R curves shown in Figure 2 has been observed for the J M-R curves of specimens of the majority 'of the other alloys that have been investigated. Development of the singularity zone analysis described in a later section indicates that the upward sweep in the Jg-R curves is indicative of the loss of the controlling singularity for the J-integral and-the continued consistency of the J-R curves beyond this point is fortuitous.
Blunt Notch Tests Blunt notch tests of IT compact specimens were conducted and the final loads and J values were compared with results from corresponding elastic compliance J-R curve tests. For deformation plasticity conditions to prevail, the final loads and J values between.the blunt notch and J-R curve tests should be in good agreement. A comparison of the load-displacement records for several A533B blunt notch and J-R curve tests are presented in Figure 3. A summary of the comparisons performed with /533B, A710 and the 3%-Ni steel is presented in Table 2. This comparison shows good agreement for the A533B and 3%-Ni steels and poor agreement for the A710 steel. The percentage differences for the final load and J values for the A710 alloy were found to increase progressively with crack extension. These specimens were characterized by larger deformations than either the A533B or 3%-Ni steels at similar amounts of crack growth. The blunt notch tests for materials other than the highly ductile A710 alloy indicate that deformation theory is applicable for clastic-plastic materials of the type described herein, even when large crack extensions and corresponding material unloading has occurred. This provides support for extension of J-R curve crack growth validity limits beyond that currently provided in ASTM E1152-87. Complete details of the J-R curve and blunt notch testing were provided in NUREG/CR-5143. Singularity Zone Analysis To follow-up the blunt notch tests, an analysis, aimed at quantifying the extent of the J-singularity controlled region of crack growth in a laboratory specimen, was developed. It was observed that real (engineering) limits to i the applicability of the J-integral were apparent in a plot of the plastic component of crack opening displacement versus crack extension. Such a plot is shown for a 3%-Ni steel 1/2T compact specimen in Figure 4. This plot shows a region of initial crack blunting, a region of singularity controlled crack growth, and finally, a gradual return to crack blunting. While the presence of a singularity is difficult to verify experimentally, it is known that it j l should produce the most intense conditions for crack growth. The plot in L Figure 4 shows the most intense crack growth occurring in the central linear l 116 i _ _ _ _ _ _ _ _ _ _ _ _ _ _ . .L
re gio n .' The presence of blunting in-the beginning of the test and the weakening of the singularity as the test progresses are reflected by the fact that the increment of crack growth per increment of plastic displacement is i greatly reduced in these regions. The. delineations between these zones are I much more apparent on this plot than on a standard J-R curve. The limit of l the singularity controlled crack growth region is taken to be at a 5% deviation from the linear region shown on the plot. Application of this type of analysis to the materials investigated, has shown the loss of singularity controlled growth to be at approximately 30% of the' remaining specimen ligament. This loss of singularity region has also been found to be consistent with the beginning of the upward sweep observed in the Jg-R curves. The specimen size dependence demonstrated by the modified J, therefore, appears to be a clear indicator of a loss of singularity in these~ specimens. Details of the development of the singularity zone analysis were provided in NUREG/CR-5238. J-R Curve Extrapolations The experimental singularity zone defined above extends the applicable region of the J resistance curve well beyond the present limits of ASTM E1152-
- 87. As shown above, the results indicate that crack extensions of 30% of the initial remaining ligament can be justified by the singularity limit defined directly from experimental data.
When even this extended J-R curve data is not adequate to address crack growth resistance, it is proposed that an extrapolation can be made in the following manner. First, only small specimen data in the singularity zone should be utilized. Second, a simple power law of the form J = C(Aa)m where C and m are fitting parameters and Aa is the amount of crack extension, should be fit to the small specimen singularity zone data for purposes of extrapolation. Such an analysis can be applied to either deformation theory J-R curves or to Jg-R curves. Typical results are shown in Figure 5 for the A533B steel. The smooth power law curves have been fit to the 1/2T CT specimens in the region 0.015 < Aa < 0.150. This region was determined to be the singularity zone using the analysis described previously. The power law was then extended and compared with IT CT data also shown in Figure 5. The comparison is excellent to a crack extension of 0.500 inches on the IT specimen. Conclusions (1) Blunt notch test results on several alloys indicate that J-R curves are generally useful well beyond the present crack extension validity limits in E1152-87. (2) An analysis was developed to define an engineering limit to the extent of J-singularity controlled crack growth in the alloys examined. This analysis indicates loss of the controlling singularity at approximately 30% of the initial specimen remaining ligament for a majority of the materials investigated, li7
(3) J-R curve extrapolations based on fitting a simple power law to the data in the extended validity region provide accurate to conservative estimates of large specimen behavior in the case of deformation theory J, and accurate to non-conservative estimates of large specimen behavior in the case of Jg. l l J/Jg-R Working Group Meetings ) A working group, established in November 1987 to work on the J/Jg-R curve issue, met several times during the year to assess progress on extension of J-R curve validity limits and extrapolation techniques. These working group meetings were aimed at coordinating research in this area with particular regard to resolution of the ASME A-11 low upper shelf vessel toughness issue. Reports on the group's progress were presented to the ASME Section XI Working Group on Flaw Evaluation and the ASTM J-R curve Task Group (E24.08.03). Interactions between the working group members and the ASTM and ASME groups have resulted in significant progress on this issue. An example of this progress was distribution of J-R curves from the ORNL V8-A low upper shelf weld material to enable group members to examine the effects of J-R curve extrapolation techniques on the A-11 1/4T flaw analysis. Results of this investigation will be available early in 1989. EVALUATION OF THE FRACTURE BEHAVIOR OF AGED CAST STAINLESS STEEL The deformation and fracture behavior of austenitic stainless steels is known to be influenced by aging occurring due to exposure to elevated temperatures in service [2]. The objective of this investigation was to quantify the fracture behavior of an aged, cast austenitic stainless steel at
$720F. A secondary objective was to evaluate the effects of changing the notch acuit'y on the fracture behavior.
A pipe section with 2 inch wall thickness that was artificially aged (700 hours at 752oF) by FRAMATOME was sent to DTRC via the Battelle Columbus Laboratories. The cast macrostructure of a cross section through the pipe wall is shown in Figure 6. Fracture test specimens were removed from the pipe wall with the notches machined in the L-C orientation. Tensile test specimens were removed along the longitudinal axis of the pipe. Preparation of the fracture test specimens included both fatigue precracking and notching with electrical discharge machining (EDM) in order to evaluate notch acuity effects on the J-R curve. Tensile test results for the stainless steel at 5720F are presented in i Table 3, where they are compared with results from the FRAMATOME ]j investigation on the same material. One of the tensile specimens (G00-1) was accidentally prestrained, thereby invalidating comparisons on a yield strength { basis. The other properties were consistent with those obtained by FRAMATOME i on the same material at 6080F, J R curves for two IT compact specimens tested j at 572 F are presented in Figure 7. Specimen G00-6 was EDM notched to a total l crack depth of approximately 1.3 inches. The EDM notch root radius was l approximately .010 inches. Specimen G00-7 was fatigue precracked in accordance with ASTM procedures to the same crack depth as G00-6. The resulting J -R curves , obtained using a DCPD technique, were in excellent agreement with JIC values of 1530 in-lbs/sq.in. for G00-6 and 1631 in-lbs/sq.in, for G00-7. This agreement leads to the conclusion that the reduced 118 l ______-______a
l i i i notch acuity produced by the EDM technique had little effect on the overall J- l R curve behavior in comparison with a fatigue precracked specimen. This l conclusion is applicable only for the conditions investigated (aged cast i stainless steel at 572 F) and is likely to be different for other material / temperature combinations. A further observation pertinent to the fracture behavior of the stainless steel is the occurrence of static and dynamic strain aging at the temperature (5720F) and displacement rate (.02 L inches / minute) used for testing. The manifestation of strain aging is illustrated by the serrations on the load displacement record shown in Figure
- 8. Strain aging has been shown to significantly decrease the ductile fracture toughness of reactor pressure vessel steel [3].
DCPD TESTING TECHNIQUE STANDARDIZATION FOR PRESSURE VESSEL AND PIPING ALLOYS l The objective of this task is to pursue standardization of a DCPD technique ; for determining J-R curves, within the ASTM (E24.08.03). To this end, DTRC I initiated a round robin among several NRC contractors in FY-87. The materials chosen for the round robin were A106 steel and a high strength Aluminum- i Magnesium alloy used by the Navy. The A106 steel exhibits a significant amount of blunting prior to fracture initiation, while the Al-Mg alloy is closer to being elastic-perfectly plastic in its fracture behavior. The specimen chosen for the round robin was the IT compact with the current input and potential output points located as shown in Figure 9. The specimens were . precracked at DTRC to an initial crack depth of 1.3 inches. Each participant l' was directed to un their own DCPD J-R curve technique. Contractors included in this round rabin were DTRC, the Oak Ridge National Laboratory (ORNL), Materials Engineering Associates (MEA), and the Battelle Columbus Laboratories (BCL). All of the round robin results from the participants were received at DTRC by September, 1989. The results were analyzed and presented to the ASTM Task Group on J-R Curves (E24.08.03) during the committee week meetings in Atlanta (November, 1987). A comparison of the experimental / analytical procedures employed by the participants is presented in Table 4. With the exception of DTRC, all of the participants used the DCPD signal to develop an "on-line" J-R curve using a procedure referred to as a "one-point pin". The DTRC technique requires the final measured crack length after test completion for generation of the J-R curve. This procedure is referred to as a "two point pin." Details of these procedures are presented in NUREC/CR-4540. Estimated crack extrnsions using the DCPD techniques were in excellent agreement with the firal measured values for both the A106 steel and the Al alloy. The percent error in predicting the actual measured crack extensions was generally less for the DCPD techniques than for elastic compliance estimates on the same specimens. Comparison of the J-R curves on the basis of JIC and tearing modulus (T) values for the two alloys is shown in Tables 5 and 6. The agreement for the Al alloy in both cases was excellent. The agreement for A106 was good on the basis of the tearing modulus and poor on the basis of the JIC values. The poor agreement observed on the basis of JIC can be explained by differing techniques used to account for specimen blunting behavior and the overall material variability of the alloy. 119 l l I
UPPER TRANSITION FRACTURE TOUCHNESS OF A5338 STEEL During the past three years, DTRC and the USNA have developed a test method for dynamic testing of fracture mechanics specimens which is specifically designed for application to the upper transition temperature range [4]. The method uses impact loading rates of approximately 100 inches /second and obtains a JIC and/or a J-R curve using an analytical key curve approach. During FY-87/88, this approach was applied to IT and 2T three-point bend specimens of A533B steel to assess the transition fracture behavior of the steel and to compare with wide plate crack arrest test data. This effort was completed in early FY-88 and reported in NUREG/CR-5142. Major conclusions were the following. First, it was demonstrated that the J integral can be evaluated from fracture specimens loaded at drop tower rates (100 inches /second), yielding both JIC and J-R curve data. It was also shown that the higher loading rate produces a significant shift on the ductile.to brittle transition region for A533B steel, in this case the shift was approximately 4000. More surprisingly, it was shown that in terms of the J at initiation of cleavage fracture, specimen size had only a small effect on the measured fracture toughness when the tests were done at drop tower rates, i.e the effects of size and rate did not seem to accumulate in terms of the position of the ductile to brittle transition region. This conclusion was further supported by crack arrest results obtained by ORNL and the National Bureau of Standards on this material which were shown to be consistent with the dynamic initiation toughness's measured in this study. The observation of the occurrence of cleavage fracture in the A533B specimens after significant amounts of ductile tearing is cause for concern as there are no standardized methods currently available to har.dle the scatter in fracture toughness data associated with this phenomena. FUTURE RESEARCH PLANS: Research plans for FY-89 include the following: (a) Complete investigation of J/Jg-R curve validity extension / extrapolation techniques and report on procedures and recommendations. i (b) Pursue resolur.lon of J/Jm-R curve issues within ASTM and ASME. l (c) Pursue standardization of PCPD J R curve testing within ASTM based on the , results of the NRC round robin. l (d) Issue summary report on the fracture behavior of cast aged stainless ' steel. (e) Initiate investigation of cracking in flawed pipe elbows. 120 l I
4 i l
.J REFERENCES
, (1) Ernst, ii.A. in ELASTIC-PLASTIC FRACTURE, ASTM STP 803, Vol. 1, 1983, pp. ! 191-213.
. (2) Hiser, A.L. , " Tensile and J-R Curve Characterization of Thermally Aged Cast Stainless Steels," NUREG/CR-5024, September, 1988.
- (3) Ostensson, B., "The Fracture Toughness of Pressure Vessel Steel at
- ,
- Elevated Temperatures," Proceedings of The Reliability Problems of Reactor Pressure Vessel Components, 1978, Vol. 1, International Atomic Energy Association, Vienna.
(4) Joyce, J..A. and Hackett, E.M., " Development of an Analytic Key Curve Approach to Drop Tower J-R Curve Measurement," NUREG/CR-4782', December, 1986, 121
TABLE 1 - J-3 CUSVE TEST MAthl* NUMBER OF SPECIMEk5 PER COND!ilDN L THREE-POINT COMPACT BEND C(f) SE(B) M*TERIAL 1/21 li 21 1/27 17 27
.......................................................r..............
A710 3 8 3 3 3 3 A5338-02 3 3 3 3 A5338-M13 3 8 3 3 1 A516 3 3 3 3 3 3 A106 8 3% Ni 3 8 3 6 TABLE 2 BLUNT WotCH TEST
SUMMARY
MA1ERIAL/ Test FINAL FINAL % DIFF %DIFF CRACK SPECIMEN TYPE LOAD J LOAD J EXTENSION (pomds) (tbs /in) (%) (%) (inches) 3 Ni STEEL FYB A3 BLUNT-MOTCHED 8100 1700 -4 2 FYB-A12 J-R CURVE 8400 1667 0.107 FYB A7 BLUNT-NOTCHED 6000 2205 5 0 FYB All J-R CURVE 6300 2216 0.202 FYB A6 BLUNT NOTCHED 4000 2435 -13 -1 FYB A10 J-R CURVE 4500 2405 0.3 A5338 E74 BLUNT NOTCHED 6880 2907 0 -6 E71 J R CURVE 6870 2749 0.096 E 72 BLUNT-h01CHED 5130 4463 -1 1 E75 J R CURVE $190 4401 0.209 E 73 BLUNT NOTCHE0 3900 4278 -8 -2 Eio J R CURVE 4200 4208 0.274
................................re....... 4.........................
A710 GFF-J13 BLUNT-NOTCHED 6150 7150 -7 1 CFF-J11 J R CURVE 8700 7210 0.1 CFF J14 BLUNT-NOTCHED 5800 9512 -17 6 CFF J10 J R CURVE 6800 10170 0.207 GFF-J12 BLUNT N01CHED 3600 9450 -39 16 CFF-J9 J R CURVE 5000 112^4 0.298 LFF-J6 BLUNT-NOTCHED 2400 10750 -50 19 GFF J5 J-R CURVE 3600 13250 0.387 122
k j j., ,
~ TABLE 3 - TENSILE 7EST DATA'(OR CAST ACED STAINLESS STEEL <
i ORIENTATION:.LONGIVUDICAL;, TEMPERATURE: 5F2*F SPECIMEN . ULTIMATE TENSILE YlELD STRENGTH ELONGAil0N . REDUCTION IN 1 10
.L . STRENGTH. -(0.2% OFFSET) (IN 2 INCHES) AREA (KSI) (KSI) (%) (%)
o..............................................r...........S.................F DTRC G00I t 92 ' 54* 24 23 l 1 DTRC G00 2 - 88.5 33.5 24 30 FRAMATOME B3+ - 90 L 30- .'22.4 33 U.
'FRAMATOME a 84* 85 29.6 24.8 36 s
- Specimen was' accidentally pre-strained approximately 1.6%
therefore the yield strength is not representative
' ' + - FRAMATOME data at' temperature of 608*F TABLE 4
.NRC DCPD J-R CURVE ROUND ROSIN EXPERIMENTAL / ANALYTICAL PROCEDURE VARIACLE/ l_ l .l l l CONDITION l MEA -l BCL l ORNL l DTRC - l
............l............l............l............l............l l 1 I I I INPUT l l20A-A106l l l CURRENT' l 15A l 40A-Al.l 10A j .60A l l .. I I I I INPUT- l .l- l l l LOCATION l W/2 l W/2 l W/4 l W/4 l
- l. I -l i I POTENTIAL.l l l l l OUTPUT lFRONTFACElFRONTFACElFRONTFACElFRONTFACEl LOCATION l l l l l l l I I I
; CAL!BRATIONl JOHNSON l MODIFIED l EXPERIMENTAL l EXPERIMENTAL l l l JOHNSON l j l
/ I i i i l I' BLtdTING l.ASSPEf> l EXPERIMENTAL) ASSUMED 'l ASSUMED j l GEFAVIOR 'l l BE HAVIOR l BE r,AVIOR l 1 l l I
- .y , , ISOLATION l-KACH:WE. lMACHlhEANDlMACHINEANDl MA KINE l j ,..
l CELY lELIPGAGE lCLIFGAGE l OMY l I'#'
- l. l l l l I 1 j ANA!YTICALl l l 4 TECnNICUE l ONE PMNT j DNE PolNT l ONE POINT lTWOP0aWT l l .PI N l Plk l PIN l PIN l (
r ! I i i t
...mna,. p. mmm.pm ,m.: .. p.,m um pm m my p 123 I
=
A 7 3 7 0 3 7 8 =
^ /
q "" 5 2 4 2 1 2 2 2 2 2 2
=
G S E * = N U
- I L x * "
R U
. A D e " "
E O d
- Y T M / "
J " d " O = L L A l l l l l l I l I l l I l l l l I l l l l 1 1 l l
* =
0 0 3 l A
- s. 4 8 2
- s. ' O. - "
e )
" s 5 3 5 6 .4 s6 6
"=
R O d i " F / s J k " " N O d ( I G ~ l l I l I i l I l l l l I l I l l l 1 I l E l l l l . R - 0 0 0 Y
)
n" 0 6 0 8 0 7 0 0 3 0 8 3 s 5
' . T 1 I
6 I _ - 7 " 7 4 4 3 0 3 2 6 1 3 4 6 ' 2 2 2 D I c - s 8 q
- 2 2 "
L i 3 / " " s
- WV T
A J 1 8 b E t " 2 s 1 n" 1 i( " l I l l I l 1 Il I l l l l I l I I l 1 l ! l E l l R . " E / *
- Y D. "
V O R 1 " . T O T N " AE L C R t c d A E R M "* M R T o " n O I 2 s O 4 D 2 3 T
- 4 4 4 6- 7 8
.a B C A E " - - - - - 3- '- -' "
a J J N N J J J ic L P S " c C C G G c cC " J " G G G F M F G G G "
= =
4 5 6 = a;; == 0 3 2 0 9 0 7 3 3 2 3 1 1 8 4 4 4 9 9 =
/ol 2 2 2 2 2 2 1 2 2 1 2 =
G S E = =
= =
N U = I L x = = R U = A D e = = L E O d = E T M / = E J = = T d == = S =
= -
l l l l l l l l l l l l l l I i l l l 6 l l l l l l l l l l
=
0 =
= 0 0 0 0 0 0 0 =
1 A = s. 0 0 s. =
= 3 9 0 4 2 0 6 5 5 8 2 =
R ) = 3 2 3 3 3 3 2 3 3 2 4 O de i = = F / s = = J k = = N d ( =
=
O I = . 3 G = l l l l 1 1 i l l l 1 l l l l l 1 l l l l I l 1 1 E l l l l
=
=
R . 0 0 0 0 0 =
) = 0 0 0 0 0 0
_ Y n == 8 3 0 0 5 0 1 8 4 6 4 = i= T 7 0 4 0 = s t 7 5 2 7 4 9 2 3 - 0 8 q 1 8 1 0 0 1 3 5 2 9 7 =
=
E 1 c - s = 1 1 1 1 1 1 1 1
=
i 3 / = L L B A J 1 s = = A V 8 bt =
=
_ T 2 E
- = =
=
. 15 in ==
( .
=
1 g l l l l 1 i i l : ; l 1 ] l l l 1 l 1 l _ E l l l j l ll l
=
=
- R . = . E / = _ V Y D. _ C R t = _ - O = t c t _ T T N = A c A E E u R - M = r A a . M m72 d R 9 6 8 0 n O I 4 s 9 9 o 0 1 _ a = 3 4 3 2 3 1 3 s . _ B C 1 I A E = - - - - - - - - - - - = r F F F F F F F r _ c L P = F A F A A A A A A A A A = _ t S = A J = G c G G G G G G G G c = ga
6 _ _ _ _ - - - _ 0 I 5 l
) 0 L
L A T T T ( 2 2 1
/ /
6 1 1 0 4 ' 1 5 9 A ^
= 2 2 T A ^
T W G G G 4
, / 3 A B B B 0 3 3 3 )
G 3 5 3 5 3 5 6 A A A ^ i ( N
@ A V O O I
S 3 N
'0 TE X
E E G K N C A R A R Y 2 C T I D
'0 U
A V M r T S A 1 y '0 T 1 ug/ /
- - y
- - )
- - ,.- - O 0 0
_ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 E, 0 0, 0, 0, 0, ' 0, 0 0, 0 6 5 3 2 1 _ 9 8 7
&CciC= 7 fir
I ,' l l!!j li!> 6 _ - - 0 g ) L L 5 A T T T T i ( 2 2 1 1 0 6 / / 1 1
@ 0
=
5 2 7 2 4 1 A 5 1 A T T A A W G G G G
/
A B B B B 3 3 3 3 3 3 3 3 l fS 5 A A 5 A V 5 A O 5 A i 4 0 ) i n t S B e e _ g ( N 3 3 5 O A _ O I r g g i 3 E 0 T S N f o s e E X v r G E u
^
e N A R K C C R g o Y A J Oo T R - I D 2 C 2 I i g L 0 e A r u Q V g i M F T S A i 1 g
. 0 T
2 _ T f 1!
- - - - - - y3 0 0 0 0 0 0 0 9 0 0 0 0 0 0 0 0 0 0, 0, 0 0 0, 0, 0, 0, 6 4 2 0' 1
8 6 4 2 _ 1 1 1 ET , m~ i ll
5 2 5 0 7
/
2 7 6 E 7 8i' I
/
3 v 3 3 7 5 E A 2 w8 ; 3 0 3 N 5 A 4 7
/
1 7 E 5 B g 1 3 . 3 0 5 ) A . n i ( D O C 1 1 O 5 1 0 0
/
,, / /j // -
_ 1
/,
/ -
/.'
- V b
_ 0 0 0 C 0 C O _ 0 0 0 0 0 0 0 0 0 0 0 6 5 ~- 2 _ 1
^ 8c 2:Q C - ,-
dm
i 3 0 G N I T N U 5 L i 2 B . 0 L o A N 0 _ I ' Q . F - T I O. M I 2 O L i 0 Y
-o T I
o R A oo L U G
. N 5 I i 1
.I*D oo ,/o S O G
N 9 I T O E N U N T L O i 1 C B Z . 0 T L Y 2 A T
/ I I 1 T R I A N L 2 I U S G 5 N i 0 B I .
Y i S 0 F -
/ /
Q
/
e.
/
/
k7 4 G 4 2 1 8 6 4 2 00 I 1 1 . 0 0 0 0 0 . . . . O 0 0 0 0 0 0 R. - g8 5
) ) ) ) 0 2222 0000 9888 5 3 333 i 4 3333 .
5555 0 A AAA s ( ( ( ( n e m TT TT i 4 CCCC c e i p 0 TT TT S 22I 1 T
/ / 2 1 1 / 5 1
i 3 9554 r . 221 1 o 0 f AAAA e AAAA n . GGGG o n z i 3 i I y . AVOO i t 0 N r s a O I n l e u 5 S m g i n i 2 N c i S
. E e 0 T p '
S f X o E T t I i 2 K r m i
. C o i A
f L 0 R e C n o Z y t ( b I i 5 1 i r 0 a l 0 u g 0 n 1 i ' i S 0 f o t i m 5 i i0 L . O r
- 1
- - - - - R_
0 C 0 0 0 0 C 0 O 0 C 0 0 0 0 0 0 0 G 0 0 0 0 0 0 0 ' 6 5 4 3 2 1 m c.. E
a Outer Diameter m W p pm h 4
- i 1:
ll,, N
. fj,. Each Division - 1/8"
. 4 N
p k- , 4 n as
;Ei' pg i
Inner Diameter Figure 6 - Cast Macrostructure of Aged Stainless Steel Pipe 130
O 1 2 3 4 O 1
- q y -
F i g u r e 7 J R C , u r v e s f o r C a C C s R O t A O A C 6 g K e O , d E 2 ( X E S D t T M a E i N N G n S o O l I t O e O c - s h s N e 7 i d S n ) ( t c F e h a e l e t i s , g ( 5 u 7 e 2 P r F e
) c r
a c k e d
)
_ 0 4 _ , ,! i 1 iil , , l i ' !t m-j l
~ i : ,c 6 0 , ij ! m , ' - - <]'
F ikb i i!i g u l!l l llhj5
!t li (
/
r 'i;1 e jN j.I *jo
/
d 8 ll ll
- )
i, GL ( Oo ; Or [-P
- sd 7 -
D i i S s I hp
- i l w c ic ne g m e
En ft f eR c t c so q r S e r r cd a f t f 0
. i i
So C 2 l o t r O n r s aC D ia I ns i n n t d A c i g A h c ig ni e a g d s t i v S e t a o i f n l i S e t s r s a i S n e c A o g l i n S g p e c i m e 0 n . i 4
- i t ! {i "5
m., n ; i lii,( - (; 1' e i frp ' , , L. >~ J . .L-
)
l 1 l I j l 4
*i COMPACT TENSION SPECIMEN C(T).
If 'd t a F
< POTENTIAL OUTPUT IDCATIONS 2Y(U).
5 y t 1 I r U W/4 - ORNL/DTRC V/2 - MEA /BCL CURRENT INPUT LOCATIONS Figure 9 - Schematic cf DCPD Current Input / Potential Ouput Locations for JT i Compact Specirnen Used in the DCPD Round Robin 133
IJ l 1 . j SURVEILLANCE DATA BASES, ANALYSIS, AND STANDARDIZATION PROGRAM . l 1988 ANNUAL REPORT (October 1987 - September 1988) j F. B. K. Kam Project Manager NUCLEAR REGULATORY COMMISSION PROGRAMS OAK RIDGE NATIONAL LABORATORY NRC TECHNICAL MONITOR: A. Taboada November 1988 By acceptance of this article, the pubhsher or escipient ackriowledges the U.S. Governinent's right to retain a nonexclusive. toyalty f ree hcense en and to any copyright COvefing the at tscle. Prepared for the U.S. Nuclear Regulatory Commission Office of' Nuclear Regulatory Research Washington, D.C. 20555 l under Interagency Agreement DOE 188680415B 1 NRC FIN No. B0415 Prepared by the j OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37831 operated by MARTIN MARIETTA ENERGY SYSTEMS, INC. for the U.S. DEPARTMENT OF ENERGY under Contract No. DE-AC05-640R21400 134 b '
{
}
CONTRACT TITLE: SURVEILLANCE DATA BASES, ANALYSIS, AND STANDARDIZATION PROGRAM CONTRACTOR AND LOCATION: Oak Ridge National Laboratory, Oak Ridge, TN 37831 PRINCIPAL INVESTIGATORS: F. B. K. Kam, R. E. Maerker, F. W. Sta11mann, and M. L. Williams OBJECTIVE (S): . Validate and improve methodologies and data bases which are used to predict neutron exposure and radiation embrittlement para-meters
- Prepare and update' voluntary consensus standards (or NUREGs) which can be used to support recommendations for data bases and methodologies in codes and regulatory guides relating to the surveillance of light water reactor (LWR) pressure vessels Maintain, upgrade, and develop (if necessary) data bases, pre-dictive procedures, and standards relating to reactor pressure vessel (RPV) fluence spectra determination and embrittlement assessment FY 1988 SCOPE: Program activities are conducted in the areas of neutron dosi-metry and transport calculations, the Embrittlement Data Base (EDB) and damage correlation studies, and coordination and standardization activities.
Analyses were performed on the VENUS-3 Partial Length Shield Assembly (PLSA) benchmark experiments to validate a procedure to predict the fluence rates from a three-dimensional low-leakage core similar to that in the H.B. Robinson-2 power reactor. The
. procedure involves a combination of a synthesis and a super-position techniques. A preliminary comparison of calculated-to-experimental values (C/E ratios) is reported pending completion of the measurements in December 1988. New revised VENUS-2 results are reported based on new estimates for the source nor-malization from CEN/SCK, Mol, Belgium. New thermal cross sec-tions for the SAILOR cross-section library have been calculated.
This new data should give improved thermal fluxes in LWR config-urations and improved heat-generation rates in RPVs and cavities. E Calculations have been initiated to calculate the thermal neutron fluxes (E < 0.414 eV) and the fast neutron fluxes (E < 1.0 MeV) in a boiling water reactor (BWR) and a pressurized water reactor (PWR) from the inside surface of the pressure vessel to the inner surface of the concrete biological shield. Preliminary results for a BWR Mark VI reactor are reported herein. A significant number of data points has been added to the EDB l in FY 1988. The design of the EDB format has been modified slightly to provide detailed documentation for each data point and thus facilitate the process to Q/A this data base. ) 135 t
Software. programs have been added to facilitate' data retrieval and manipulation of the raw data base files. A demonstration of the ease with which the software and data base design format can be used to generate information quickly for regulatory use was provided to P. N. Randall of NRC. A summary of the program's coordination and standardization activities is presented.
SUMMARY
OF RESEARCH PROGRESS Research is being performed to assist in the following NRC regulatory objectives:
- 1. establishing reliable pressure vessel (PV) surveillance program criteria as required by 10CFR50, Appendix H;
- 2. validating, in benchmark experiments, RPV fluence determination procedures and data bases as required by NRC's Standard Review Plan;
- 3. verifying compliance with Regulatory Guide 1.99, Section 10.61 of 10CFR50, and technical specifications; and
- 4. providing support documentation for ASTM voluntary consensus Standards and Regulatory Guides.
Task A. Neutron Dosimetry and Transport Calculations A review of the various methodologies used by industries and research institutes for RPV fluence determination shows that most organizations employ an analysis sequence consisting of three steps.1 These include transport calculations, dosi-metry measurements, and a statistical procedure to combine the calculations and measurements to arrive at a fluence value which has a smaller uncertainty than the original calculations. An accurate determination of damage fluence accumu-lated by the RPV as a function of space and time is essential in order to ensure . the vessel integrity for both pressurized thermal shock transients and end-of-life considerations. The desired accuracy for neutron exposure parameters such as. displacements per atom or neutron fluence (E > 1.0 MeV) is of the order of
-110% to 15% (Ic).2 These types of accuracies can only be obtained realistically by_ validation of the entire analysis sequence in benchmark experiments. This task deals with the validation of the analysis sequence proposed in Ref. 1.
A.I. Recalculation of NESDIP2 Experiment (R. E. Maerker) In view of the alarming discrepancy between calculated and measured rhodium and indium foil activities on the centerline in the void box behind the NESDIP2 radial shield at the ASPIS facility in Winfrith,3 a recalculation using a superior flux extrapolation procedure was performed. This option, called theta-weighting, provides greater accuracy when the spatial mesh contains cells that ' depart greatly from square. Comparisons of measurements with the revised and original cairulations, both of which used the superior updated iron cross-section set of Fu,4 are shown in Table 1. 136
) i l
Table.l. Comparison of new and original values of C/E along the centerline through up to 20 cm of additional mild steel followed by the void box (NESDIP2) C umula t ive thickness 103Rh 115In 323 of steel penetrated' Position old- new old new old new (cm) 4 9 14 0.87 0.90 0.89 0.93 0.85 '0.89 14 15 0.93 0.97 0.90 0.94 0.92 0.96 19 16 0.96 '1.00 0.89 0.93 0.94 0.98 24 17 0.92 0.99 0.85 0.91 0.91 0.96 29 18 (void box) 0.69 0.79 0. 74 0.82 0.90 0.99 A subsequent measurement of the rhodium activity behind the radial shield was performed by the British after closing up the cavity, and a new calculation of I this configuration also made. The resulting C/E value of 0.92 that was obtained
'is considered acceptable.
The conclusion may be drawn that the original discrepancies have been reduced to the vicinity'of 20%, but that they are still significant. It seems that the most likely cause of the remaining disagreement lies in the open void box )
. measurement, where neutrons below about 2 MeV may have leaked around the radial 1 r shield before entering the void box. Although the British disagree with this 1 conjecture, it. is supported by the relatively good agreement attained when the void box is closed up.
A.2. VENUS-3 Analysis - Preliminary Results (R. E. Maerker) I l In the VENUS-3 mockup at Mol, fuel pellets were replaced by stainless steel throughout the lower half of the pin length over the outermost five rows of an opposite pair of peripheral assemblies. The configuration was chosen to mock up the essential features of PLSAs which have been inserted into several of the
~
j older PWRs to reduce core leakage to lower levels than simply a revised core I loading scheme can do. The resulting geometry is truly three-dimensional (3-D), and would require use of some extremely expensive analytical techniques (i.e., Monte Carlo or 3-D discrete ordinates) unless a simpler procedure involving manipulation of the results of 1-D and 2-D codes can be devised. One such pro-cedure involves a combination of a synthesie and a superposition technique, whereby the two source components, one lying above the PLSA source discontinuity l ! and the other below, each give rise to flutes that are synthesized from calcula-tions involving penetration through that part of the PLSA geometry that controls the contribution from that particular component. The midplane VENUS-3 geometry l may be well described by takir.g full advantage of the features of the variable-mesh option offered in DOT-IV (see Fig. 1). 137
aj F 1 HORIZONTAL PLANE (X-Y) GEOMETRY FOR VENUS-3 B 3, _ w 5 =0 e
.@ PLSA-4 y -i 3 oooo oooa
= o 8 aaoa i L' o o D o o f~l .
64 9 63 6 DSTANCE IN CENTNETERS F. Horizontal plane (X-Y) geometry for VENUS-3. Fig. 1. Source measurements in the form of 3-D tables of a relative source distribution normalized to a pin average power of 1 fission /s have been recently provided by Hol,* and were filled out by ORNL to cover all locations in the core. To con-form to the synthesis procedure adopted, source components lying above and below . ; the midplane were obtained by partial integration over the appropriate vertical (Z) intervals for the two XY flux calculations. The sources for the two XZ calculations involved integration of each component over the cross (Y) direc-L tion, and those for the two X calculations were obtained by further integration of the XY sources over Y, and as a check, integration of the two XZ sources over the appropriate Z intervals. The pairs of XZ sources and pairs of X sources, corresponding in each case to components above and below the midplane, were calculated for two different cuts through Fig. 1, one corresponding to the X axis and the other to Y. This permits flux synthesis calculations to extend to dosimeter locations lying near eith~r axis. Discrete ordinate calculations have been made for all ten sources and geometries described above. At the moment of this writing, the only dosimeter measurements available from Mol are 58Ni(n,p) activities performed in April 1988** and, at this time, only relative vertical profiles can be compared until additional Mol measurements ate made to relate the absolute source operating during these measurements with the arbitrarily normalized source used in the transport calcu-lations. A comparison of the vertical shapes is shown in Table 2 for location 1 ; in Fig. 1 (facing the simulated PLSA) and in Table 3 for location 2 in Fig. 1 ] (facing the unmodified assemblies). 1 Comparison of the profiles in the strongly asymmetric case in Table 2 and the almost symmetric case in Table 3 indicates excellent agreement, but until abso-lute levels are compared, no conclusions should be drawn concerning the accuracy l of the calculational method, i
*1.. Leenders, CEN/SCK, personal communication to F. B. K. Kam, Oak Ridge National Laboratory, September 8, 1988. ]
**A. Fabry, CEN/SCK, personal communication to F. B. K. Kam, Oak Ridge National l Laboratory, October 19 and 21, 1988.
138
l Table 2.. Comparison of relative vertical nickel profiles at location 1-(PLSA) Calculated fraction Distance above due to source below midplane Measured Calculated midplane (cm)
-23.5 0.254 0.263 0.816
-19.5 0.333 0.351 0.761
-15.5 0.423 0.451 0.690
-11.5 0.539 0.550 0.604
-7.5 0.657 0.653 0.505
-4.5 0.739 0. 74 9 0.391
-1.5 0.830 0.850 0.319 1.5 0. 91 2 0.901 0.272 4.5 0.967 0.969 0.222 7.5 1.000 1.000 0.157 11.5 0.964 0.966 0.114 15.5 0.880 0.888 0.089 19.5 0.739 0.760 0.067 23.5 0.563 0.579 0.057 Table 3. Comparison of relative vertical nickel profiles at location 2 (unmodified)
Calculated fraction Distance above due to source below midplane Measured Calculated midplane (cm)
-23.5 0.449 0.446 0.912
-19.5 0.604 0.611 0.888
-15.5 0.736 0.739 0.845
-11.5 0.853 0.848 0.790
-7.5 0.929 0.934 0. 719
-4.5 0.978 0.975 0.617
-1.5 0.991 0.997 0.526 f 1.5 1.000 1.000 0.467 4.5 0.980 0.980 0.375 7.5 0.942 0. 942 0.275 l: 11.5 0.853 0.860 0.204 15.5 0. 746 0.753 0.149 L 19.5 0.604 0.625 0.106 l 23.5 0.449 0.463 0.082 139
A.3. Transport Calculations for a Boiling Water Reactor (M. L. Williams and M. Asgari) Two-dimensional DOT transport calculations have been performed to determine the fast and thermal flux distribution throughout the RPV and cavity of a BWR, The analysis used the SAILOR cross-section library which has 47 neutron groups and 20 gamma groups. The standard flux synthesis method - based on combining RO, RZ, and R calculations - was utilized. A high order biased quadrature set consisting of 130 directions was used in the RZ cavity calculation in order to accurately treat the cavity streaming. Table 4 summarizes the values at a 45* azimuth (approximate peak) and at the reactor midplane for *(>1.0 MeV), $(<0.4 MeV), dpa rate, and percent of dpa rate from neutrons greater than 1.0 MeV, greater than 0.1 MeV, and less than 0.4 eV. The fast flux incident on the vessel is about 1.5E+9 which is more than an order of magnitude lower than observed for most PWRs. One also observes that neutrons above 0.11 MeV contribute at least 90% of the dpa rate at the different locations. Finally, preliminary calculations indicated that the fast flux and dpa rate, at locations above the top of the active core in the cavity, peaks at the back rather than the front of the RPV. On the other hand, the thermal flux in the cavity behaves much dif ferently than the fast flux - the thermal source is obviously located within the concrete shield, where the fast neutrons leaking from the vessel are thermalized. The thermal neutrons emerging from the concrete travel across the cavity and impinge on the outer side of the RPV, causing a peak in the thermal flux within the vessel at the outer radius. Table 4. Integral values at midplane, 45' Percent of dpa rate from neutrons o(>1.0 MeV) e(>0.1 MeV) $(>0.4 eV) dpa E> E> E< [n/(cm2.s)] [n/(cm2*s)] [n/(cm2.s)} (dpa/s) 1 MeV 0.11 MeV 0.4 eV RPV l 0-T 1.46E+9 2.88E+9 2.66E+9 2.25E-12 80 97 ~0 1/4-T 9.43E+8 2.51E+9 1.17E+8 1.54E-12 69 97 ~0 3/4-T 2.88E+8 1.28E+9 1.06E+7 5.83E-13 49 96 -0 Mid cavity l l 7.41E+7 3.67E+8 1.19E+8 1.66E-13 45 92 0.6 Concrete i i 0-T 4.41E+7 2.17E+8 1.64E+8 1.01E-13 44 90 1.5 l 140 i
A.4, New Cross Sections for SAILOR (M. L. Williams and F. B. K. Kam) The SAILOR S cross-section library assumed a Maxwellian thermal spectrum at 300'K in averaging the thermal values in the energy range below 0.4 ev. In order to obtain a more realistic thermal spectrum, a transport calculation utilizing seven thermal groups with upscatter in the range below 0.4 eV has been performed for a 1-D PWR model using the 27-group CSRL Criticality Library (based on ENDF/B-V). ! Results of this analysis were used to correct the thermal group cross sections of the water and ex-core structural materials in SAILOR to improve the calculation of the thermal neutron flux as well as gamma production in RPV transport calculations. The correction factors were obtained by: l
- 1. collapsing the 27-group absorption cross sections into the appropriate two thermal groups corresponding to the energy intervals [0, 0.1] and [0.1, 0.4]
using the calculated multigroup thermal spectrum;
- 2. dividing the new absorption cross sections by the original SAILOR values to get a correction factor (f,) for the neutron thermal absorption and thermal neutron gamma production cross sections:
(Dag)new ; and ) f, = (Dag) original
- 3. assuming a constant scatter cross section in the thermal range, the correc-tion factor for the total cross section is calculated, f t = 1 + (fa-1) [ ag} .
\Otg / original l
l In addition to the modified thermal cross-section values the updated iron cross-section set of Fu4 was added to the original SAILOR ENDF/B-IV iron data. This modified set labeled "new SAILOR" was used to assess the impact on the transport calculations of a PWR configuration. Two 1-D transport calculations were per-formed, one using the " original SAILOR" cross-section set and the other using the new SAILOR cross-section set. Tables 5 and 6 summarize the results of this new SAILOR cross-section set. 1 Table 5. SAILOR thermal correction factors based on 27-group results for absorption (f a) and total (f ) cross e sections Group 46 (0.1 - 0.4 ev) Group 47 (0.0 - 0.1 ev) l N uc lide f a It f a f t H- 1.00 1.00 0.91 1.00 Cr* 1.06* 1.01 0.79* 0. 91 Ni 1.07 1.01 0.78 0.96 j Mn 1.06 1.04 0.80 0.82 Fu-Fe 0.97 1.00 0.95 0.99 i
*The Cr results are estimated from the other values because an error was found ;
in the Cr data on the 27-group CSRL Library. ; i 141
- Table 6. Impcet of new SAILOR cross-section set on neutron flux calculations Ratio of- Thermal-to-fast Thermal fluxes new-to-original flux ratios using original SAILOR cross Thermal Fast
. Interval Radius section set fluxes fluxes Original New (cm) (E<0.4 eV) (E>l MeV) 48 (RPV 0-T) 217.42 5.69E+12 1.18 1.12 1.09 1.15 49 219.20 2.20E+12 1.26 1.14 0.50 0.55 50 222.26 7.66E+10 0.99 1.17 2.44E-02 2.07E-02 51 225.32 2.80E+10 1.08 1.20 1.36E-02 1.23E-02 52 228.38 1.30E+10 1.14 1.23 9.90E-03 9.20E-03 53 231.43 1.28E+10 1.20 1.26 1.52E-02 1.50E-02 54 234.49 1.80E+10 1.34 1.29 3.43E-02 -3.27E-02 55 237.55 3.49E+10 1.31 1.33 1.13 1.11 56 (cavity) 265.67 8.33E+11 1.24 1.37 4.24 3.85 57 318.85 8.89E+11 1.23 1.37 5.22 4.72 A.S. Revised VENUS-2 Results (M. L. Williams)
New estimates from Mol for the source normalization in VENUS-2 were received.* A value for the average midplane linear fission rate per pin was recommended to be 1.8-1.9E+08 fission /s-cm. In Ref. 8, Leenders cited several measurements which were made to establish the appropriate VENUS-2 normalization. Based on these results, the following value was chosen to be used in normalizing calculations: q iss on Linear fission rate (average pin, core midplane) = 1.88E+08 _ According to the recommended range of values provided by Mol, there is about a j 5% uncertainty in the normalization (and hence, in the calculated results). In the previous VENUS-2 results reported in the 1987 NRC Annual Report, a value of 2.07E+08 was assumed for the linear fission rate. The new value will, there- ( fore, reduce the earlier calculated results by a factor of 0.908. In Table 7, the new calculated values are compared with the VENUS-2 experimental results. Overall, the new normalization has improved the results somewhat and made them more consistent with the VENUS-1 results. There appears to be something wrong at < the point (-17,-14) -- the C/E values for all dosimeters at this location are about 10% higher than neighboring locations. Perhaps, there is some positioning inconsistency between the calculations and measurements. The same trend as in the earlier results for the C/E values is seen (especially for the high threshold dosimeter) to decrease with increasing water penetration slong the 45* traverse through water gap 1. Table 6 compares C/E values for VENUS-1 and VENUS-2 dosimeters at corresponding locations. In most cases, the C/E values seem fairly consistent. Hewever, again the results for the (-17,-14) location in VENUS-2 appear to be about 10% high. In addition, the 58Ni C/E values in the barrel are much dif ferent in VENUS-1 and VENUS-2 -- in VENUS-2, the calculations agree well with the measured values, while they are up to 30% low in VENUS-1. Such a large variation does not seem to be physically reasonable.
*L. Loanders, CEN/SCK, priva te c ommunica t ion t o F. B. K. Kam, Oak Ridge National Laboratory, October 26, 1988.
l IL: i l I 3
l
)
1 l l l Table 7. Comparison of calculated (C) and measured (E) equivalent ' fission flures in the steel regions of VENUS-2 237Np(n,f)F.P. 27Al(n,n)248 , 58Hi(n.p)58Co ll51n(n,n')115 min VEN L's Loc a t ion C* E* C/E* C* E* C/E* C* E* C/E* C* E* C/E* Inner Baffle (-l .* 2 ) 9= 8.1
- 2.716E91 2.491t9 1.090 1.584E9 1.426E9 1.111 1.651E9 1.51]E9 1.093 1.974E9 1.443E9 1.016 ;
( -1.-1 ) 9=45.0* 3.202E9 2.952E9 1.085 1.813E9 1.699E9 1.067 1.935E9 1.825E9 1.060 2.331E9 2.357E9 0.989 I outer raffle (-29, *2) 9* 0.9* 9.188E8 -- 5.185E8 5.919E8 0.876 5.465E8 5. 799 E 8 0.942 6.616E8 7.144 E 8 0.926 ( -29, ~ 2 ) 6
- 8. l
- 8.663E8 -- 4.906E8 --
5.174E8 5.515E8 0.938 6.275E8 -- ( -29. -7 ) 9= 16. 8
- 7.064E8 6.884E8 1.026 3 993E8 4.516E8 0.884 4.196E8 4.467E8 0.939 5.103E8 5.481E8 0.931
(-29,-12 ) 9=24. 7* 4.360E6 4.336E8 1.006 2.407E8 2.690E8 0.895 2.479E8 2.624E8 0.945 3.077E8 3.330E6 0.924 (-27.-l=) G=29.2* 4.601E8 -- 2.548E8 2.756E8 0.925 2.617E8 2,753E8 0.951 3. 244 E 8 3.530E8 0.919 (-22, 141 9= 34.0
- 9.246E8 -- 5.137E8 -- 5.405E8 5.439E8 0.994 6.616E8 --
(-17 -14) 8 40.2* 1.602E9 I.455E9 1.101 8.606E8 8.300E8 1.037 9.246E8 8.692E8 1.064 1.142E9 1.12109 1.018 Core Barrel (-37, *2) 9* 0.7* 1.076E8 -- 7.990E7 9.621E7 0.831 6.805E7 7.355E7 0.925 7.895E7 8.967E7 0.881 (-37. -5) 8=10.8* 9.818E8 1.031E8 0.952 6.734E7 8.503E7 0.79 2 5.821E7 -- 7.00417 7.901 E 7 0.887 ( -35.-12 ) Ga 21. l
- 1.02118 1.037E8 0.984 6.350E7 7.333E7 0.866 5.651E7 6.024E7 0.938 7.044 E 7 7.91017 0.891
(-34.-15) 0 25.6* 9.Il8E7 9.035E7 1.009 5.262E7 6.0llE7 0.875 4.725E7 5.058E7 0.934 6.099 E 7 6.896E7 0.884 (-33.-1$) 9 28.8* 8. 29 7E 7 -- 4.863E7 5.48257 0.887 4.31217 4.498E7 0.959 5.550E7 6.04857 0.918 (-31.-20) 4*33.9* 7.013E7 7.020E7 0.999 4. 79 5E7 5. 714 f.7 0.839 3.986t? -- 4.870E7 5.370E7 0.907 (-28.-24) 0-41.0* 5.460E7 5.485E7 0.996 4. 582 f.7 5 188E7 0.883 3.494 E 7 3.673E7 0.9513 3.992E7 4. 36 3E 7 0.915 (-26,-26) G=45.0* 5.329E7 5. 3!.8 r 7 0.997 4.629E7 5.121E7 0.904 3.489E7 3.670E7 0.9507 3.938f.7 4. 38 2 E 7 0.899 Neutron Pad (R=62.8) G=21.l* 9.918E6 1.122E7 0.884 7.650E6 1.049E7 0.729 5.310E6 -- 6.605E6 8.187E6 0.807 (8*62.8) 6 45.0* 7.309E6 7.704E6 0.949 6.585E6 8. 393E6 0.786 4.148E6 -- 4.932E6 5.896E6 0.836 Center Hola (+2.5.*2.5) Water Cap 1 ( -16 . - 16 ) 1.050E7 9.767E8 1.075 6.883E8 6.658E8 1.034 6. 78 7E8 6.452E8 1.052 7.739E8 7.911 E8 0.978 (-18.-18) 4.972E8 4.617E8 1.077 4.001E8 4.056E8 0.986 3.$82E8 -- 3.799 E8 4.002E8 0.949 (-20,-20) 2.4 79 E8 2.309E8 1.074 2.356E8 2.533E8 0.9301 1.930E8 1.930E8 1.004 1.944E8 2.067E8 0.941 (-22,-22) 1.28]E8 1.238E8 1.034 1.408E8 1. 519 E 8 0.927 1.068E8 -- 1.026E8 1.113E8 0.922 ( *24,-2= ) 7.284f7 --
- 8. 215 E 7 9.713E7 0.84 6 5.985E7 6.265E7 0.955 5. 818 r.7 6. 664 E 7 0.873 Wster cap II (9=10.75*) 3.075E7 -- 2.880E7 --
2.090E7 2. 24 7 E 7 -- (0 16.63*) 3.147E7 -- 2.802E7 3.148E7 0.890 2.064E7 2.262E7 2.585E7 0.875 (0 21.14*) 3.ll3E7 -- 2.694E7 2.518E75 1.070 1.992E7 2.210E7 2.532E7 0.873 (0 25.62*) 2.942E7 -- 2.41 ) E 7 2. 6 74 f. 7 0.902 1. 805 E 7 2.04 6 E 7 2.302E7 0.889 (G=28.78*) 2.737E7 -- 2.318E7 2.535E7 0.914 1. 708 E 7 1.918E7 2.133 E 7 0.899 (G=33.89*) 2.370E7 -- 2.217E7 2.496E7 0.828 1.562E7 1. 699 E 7 1.870E7 0.908 ( G.3 7. 44 * ) 2.140E7 -- 2.207E7 -- 1.494E7 1.574E7 -- l (G=40.99*) 1.978E7 ~ 2.19 8 E 7 2. 29 8E 7 0.956 1.444E7 1.485E7 -- l (0 45.00*) 1.9 2 5 E7 1.957E7 0.981 2.194E7 2.298E7 0.955 1.4 24 E t 1.450E7 6.f67E7 0.926
]
evalues given in (n/(cr2 s)]. IIIe ad 2. 716 E9 a s 2. 716 a ld SThe calculat ion may have an inconsistent: radius with measurement. l l l l 143 E __ - - - _ _ _ _ _ _ _ _ _ - - _ _ _ _ _ __
Table 8. Comparison of cateulated (C) and measured (E) values in VENUS-1 and VENUS-2 237Np(n.f)F.P. 38Hi(n.p)$8Co ll5!n(n n')ll5er, VENUS VENUS-] VENUS-2 VENUS-1 VENU$-2 YENUS 1 VEN U$ -2 Lec a t ion C/E C/I C/E C/E C/E C/E Inne r Waf f le j l ' (-l,+2) 0- 8.l* 1.090 1.093 1.016 (-l,-l) 6 45.0' l.074 1.085 0.987 1.060 0.980 0.989 Outer Baffle (-29, +2) 0* 0.9' O.983 0.974 0.942 0.987 0.926 (-29, -2) 9a 8.l* 0.968 0.938 0.991 (-29. -7) 9 16.8* 1 .00 3 1.026 0.945 0.939 0.981 0.931 (-29,-12) 6 24.7' l.00 8 1 . 00 6 0.924 0.945 0.980 0.924 (-27.-14) 0 29.2' l.01 3 0.919 0.951 0.984 0.919 i (-22 -14) 0 34.0* 1.012 0.961 0.994 0.964 (-17,-14) 0 40.2* 0.993 1.101 0.945 1.064 0.94 9 1.018 Core Barrel (-37, *2) G= 0.7* 0.999 0.808 0.925 0.900 0.881 (-3 7. -5) 0 10.8* 0.976 0.952 0.887 (-35,-12 ) 9= 21. l
- 0.909 0.984 0.781 0.938 0.900' O.891
(-34,-15) 6 25.6' O.915 0.934 0.884 (-33,-17) 0 28.8' O.930 0.959 0.918 (-31.-20) Ga33.9* 0.973 0.999 0.719 0.896 0.907 (+28 -24) G=41.0' O.896 0.996 0.951 0.915 (-26,-26) 0 45.0* 0.935 0.997 0.698 0.951 0.917 0.899 Neutron Pad (R*62.8) 0=21.l' O.884 0.807 (R*62.8) 9=45.0* 0.949 0.836 Center Hole (+2.5,+2.5) 0.970 0.965 0.984 Water Cap I (-16.-16) 1.075 1.052 0.930 0.978 (-18,-18) 1 . 04 5 1.077 0.927 0.949 ( 20,-20) 1.056 1.074 1.004 0.923 0,941 (-22 -22) 1.038 1.034 0.903 0.922 (-24,-24 ) 0.955 0,889 0.873 Water Cap 11 i (G=10.75') 0.925 j (Gal 6.63') 0.896 0.875 ) (9=21.14') 0.873 (G 25.62*) 0.94 3 0.889 (G*28.78') 0.907 0.899 (0 33.89') 0.453 0.908 i ( 6-3 7. 44 * ) 0.902 ( G e &O .9 9 ' ) (G=45.00') 0.874 0.926
~3/0 reel
(-3.5,-3.5) 1.050 (-6.5,-6.5) 1.025 (-9.5,-9.5) 1.044 (-12.5,-12.5) 0.948 144
Task B. Embrittlement Data Base (EDB) and Analysis (F. B. K. Kam and [ F. W. Sta11mann) f The EDB is a computerized data-base system which contains information pertaining , to radiation-induced damage of reactor vessel steel, base material, and weldmer.ts collected from surveillance reports of operating power reactors and from materials test reactor experiments. The purpose of the EDB is to assist the NRC in establishing guidelines, such as Reg. Guide 1.99, for assessing the safety of RPVs under routine heat-up and cool-down conditions and the more severe pressurized thermal shock conditions caused by loss-of-cooling accidents. The data can also be used in the analysis and validation of theoretical embrittle-ment models which may be proposed as alternates to the NRC guidelines. B.I. Current Status of the Embrittlement Data Base The original MPC data were the starting point for the EDB. In addition to the 177 power reactor data points which were used in Reg. Guide 1.99, Rev. 2, 250 additional power reactor data points have been added. A considerable amount of time and effort was devoted to the design of the EDB format, to the documentation of each data point, and to the software which will provide the end user with a convenient means to retrieve, view, manipulate, plot, and fit the data. In addi-tion to the power reactor data, at least 400 test reactor data points have been added to provide a means of correlating test and power reactor data. New data are continually being added to the raw data files. An overview of the NRC/ORNL EDB was presented to the Materials / Surveillance Work Group of the EPRI Reactor Vessel Embrittlement Management Project. The group recommended that the EDB be used as the industry-wide embrittlement data base. EPRI and NSSS vendors agreed to Q/A the EDB data before it is released. A draft copy of the EDB writeup and data base was released to them in order that they ran
- 1. scope effort required to Q/A this data base;
- 2. scope effort required to define required improvement, if necessary;
- 3. scope a periodic update program including Q/A requirements and schedules;
- 4. start effort to Q/A the EDB; and
- 5. establish, in cooperation with NUMARC and the NRC, a program for long-term maintenance of the EDB to assure that the data base does not become outdated as have previous data bases.
B.2. Comparison of Itegulatory Guido 1.99, Rev. 2, Data with New Data in the EDB Two hundred and fifty data points, 79 welds and 171 plates, were retrieved from the EDB to compare with the 177 data points used in Reg. Guide 1.99, Rev. 2. Reg. Guide values were calculated from the chemistry and fluence values for each of the 250 points. The residuals, measured shift value minus the Reg. Guide value, were determined and plotted against fluence , copper content , nickel con-tent, and the 30 f t-lb Reg. Guide CVT shif t values for both plates and welds i separately. The resulting plots are shown la Figs. 2 through 5 with the 2 boundaries (i34*F for plates and 56'F for welds ). From the figuees, one sees that the weld data behave as expected, but there are considerably more outliers for plates than the 5% expected from a 2o boundary. I 145 l L._______________.________. _ _ . _ . . . . _ _ _ _ . . _ _ _ _ _ _ _
f 5 , L a w ggv+y 7:h f.' e I' Y ,
'", _ S.', o 3
i s m, e J Residualsvs. Fluence m PlateMaterials J '!n , ', i 100 j,
+ +
l e
+
+ '+
+ +
J" 50 __________ + _3__ _g_ ;_ + ++
+ _ ___ _ _g__________. __g y, ;
a + .
. + 1 e e+dd . + m 8- j
______ __________ s "#_#k__________. _t n .g ?f; , S o Original Database + + 5
+ New Data fron EDB +g E' -100 Mean value +
.E --- 134 'F Uncertainty
, a
-158 '
b.e3 ' 16.8 17.8 18.8 19.0 28.8 21.0 ; n/en*
.i Fluence > 1 MeV (logarithm) 1- .
Residualsvs. Fluence m WeldMaterials 188
-l: ?
se n J.
+ + ++# 4 +A +
3 .$ o o + gC+g+e j + g+g T9 o o t +o 6 -se _ _ _+_ _ _ _ _ _ _ ___s__ _ _ _ _ _ _ +_ _ _(+ g+ ++ o
'8 +
j o Original Database
+ New Data from EDB E' -100 Mean Value J --- ?. 56 *F Uncerta inty +
158 - ~ 16.0 17.0 18.6 19.0 20.0 21.0 e/csa Fluence > 1 MeV (lostrlthm) Fig. 2. Plots of residuals versus fluence for Reg. Guide 1.99, Rev. 2, and rew data from EDB. 146
I.'- s! a i.} I jg 1 3 Residuals vs. Copper _ m. Plate Materials Dj ' 100 6- - e t +
'3 + + .
+ o
)
m 58 + J - t-- 1 _[ 2----- o-t 4-- v---{j g---- g s--------- 8
+2 ' '- '
- 8 M^ * #+
-j b --.4----
-- - S --
hh- t p + 3.--- + -------
-Se p
- 4 a.
w .5 ,8, o Original Database + + '+
+ New Data from EDB+ +
3: - k-'-108 Nean value + h 5- ---- 134 *T Uncertainty .) '
-158
@m .0.00 8.05 0.10 0.15 0.20 0.25 0.38 1,( j Weight Percent Copper m ? y c4 N! l 'h 'i hi ' j Residualsvs. Copper m VeldHaterials i 1ee (: <
, o
, .e
~
g ..------------------y------m------------. L,! $ +++ + + $ oft + 0 .m $ +0 o
$ + S o^ ++o+ 49 C A #, +
0 ' '#o m, .
+ '"
" 0
+ +o o@o e : ++ +
g+4o o .*g . 0 8 -58 ..-----+.----,.------..-o--y-.---- - - . - - - - - - ?! + 7 5' o Original Database
; E ~100 + New Data f rom EDB U d + Hean Value
---* 156 *F Uncertainty a ' '
- p. .-158 0.00 0.10 0.20 8,30 0.48 0.58 A
q Uelght Percent Copper
- d. l 1-l
[ f: Fig. 3. Plot.s of residuals versus copper for Reg. Guide 1.99, Rev. 2, and new data from EDB. 4 5 147 I l
Residualsvs. Nickel m PlateMaterials 100 e j . so t+. 4._._ _ _ _ _ _ _ _ _ f __ _o_ o+- :: _ _i dy,, __________. l- i +c #t 34b +
* +
m -50 +
-g 4 o%
$8- o Origina1 Database 4 +
i l
+ Hew Data from EDB + l E .. -100 Hean value . +
$ --- 134 *F Uncertainty
-158 '
8.8 B.2- 0.4 8.6 B.8 1.0 Weight Fercent Nickel j Residualsvs. Nickel m VeldMaterials 100 e O
.m d
&+g
____________________S__0+ .+o h e.,
+
f
.f g *+m+G + % +1 4% o. -
I
+
'+ +'++ + o +g +
3 g + + + +g w -se __ 5t _ _ _ _ _ _ _+_ _.
+
____________+______4 ._.___________ 3 + Y o Original Database E. ~100 + New Data f rom EDB
~$ + -- Mean Value
--- 156 *F Uncertainty
-150 3,0 0.2 0.4 0.6 8.8 1.0 1.2 1.4 Velnht Percent Nickel Fig. 4. Plots of residuals versus nickel for Reg. Guide 1.99, Rev. 2, r.nd new data from EDB.
148 i e m__ m._______ __. __ _ _ _ . . _ _ . __ _ _
t Residualvs. Reg. Guide m'flatematerial 109
+ + o Original Database
.g '.+ + New Data from EDB
=
+ + Mean Value<
*'3 50 + +- o ++
. - - p+ $ $+g b + - g 6 .-e.----------. . - 1 34 *T Uncertainty 4; .
t -- 4 e o g 64+4 +Ro+ _9hh:3 o A et.go.' 4 k$. + +---------s_.
- o u. -50 o ++ , +
5
.t. ++ + +
+
l- 2. e
-200 .
M .
-150 '
l~ 0 50 100 150 200 250
~
CUT Shif t at 30 f t-lb According to Reg. Guide 1.99 Rev.2 Residualvs. Reg. Guide m Weldnaterial 100 e o 2 -
----v---r----------------
50 3 -- y g+-- + -- + &j+ + +o + 0
, 3 o #o n oA+, +jo0% o +
'+* ++ *$ ,D h + {+ o ~ +, oo o 4
2 + l +++g*o+ ~ o o o + +o
.c -
w -50 _ _ _++,_9 0_+++____+_
+
1 6 o Ocisinal Database E. -100
+ New Data f rom EDB
,c - + Mean Value
--- 156 *F Uncertainty
-250 - "-
8 50 109 150 200 258 500 350 400
*F CUT Shift at 30 ft-lb According to Reg. Guido 1.99 Rev.2 Fig. 5. Plots cf rc.sidaala versus CVI shft at 30 ft-lb according to Reg. Guide 1.99, Rev. 2. ,
149 1 I 1
i Task C. Coordination and Standardization Activities Standardization activities in ASTM E10 Committee on Nuclear Technology and Applications continue to be an important effort to provide and update voluntary consensus standards for the LWR PV program. Three standards, E944-83 (Adjustment Methods), E1006-84 (Test Reactors), and E482-82 (Transport Methods) , I have been updated and submitted to E10 for balloting. E1006 has been approved, but E944 and E482 required changes. Coordination efforts to establish cooperative research safety programs with the international community, other U.S. laboratories, and industry continued in FY 1988. NRC's cooperative research program with CEN/SCK, Belgium, is continuing on the VENUS-3 benchmark experiments. Extension of this program to use the BR-3 shield tank to confirm the dose-rate ef fect issue that arose in the High Flux Isotope Reactor (HFIR) and independently confirmed in the Shippingport neutron shield tank material is in progress. Joint efforts with D. Pachur of KFA, West Germany, A. Fabry of CEN/SCK, Belgium, and F. W. Sta11mann of ORNL to provide a significantly improved capability for RPV embrittlement predictions based on both test and power reactor data are continuing. Activities with AEE Winfrith on the NESDIP Benchmark Experiments have been delayed due to other pressing priorities by the U.K. group. Coordination efforts with the J. Stefan Institute on a proposed joint research program in the areas of neutron dosimetry and radiation embrittlement surveillance are in progress. Establishment of the research agreement is expected to take effect in FY 1989. Considerable success has been attained with EPRI and U.S. industry to set up an industry-wide embrittlement data base. The NRC/ORNL EDB will be the focal point of this data base which will be made available to the industry. Coordination ef forts with EPRI and U.S. industry have improved significantly, and should make this effort a success. The following reports were published in FY 1988 in support of the program.
- 1. F. B. K. Kam, R. E. Maerker, M. L. Williams, and F. W. Sta11mann, Pressure Vessel Fluence Analysis and Neutron Dosimetry, NUREG/CR25049, ORNL/TM-10651, U.S. Nuclear Regulatory Commission, Washington, DC, December 1987.
- 2. P. Chowdhury, M. L. Williams, and F. B. K. Kam, Development of a Three-Dimensional Flux Synthesis Program and Comparison with 3-D Transport Theory Results, NUREG/CR-4984, ORNL/TM-10503, U.S. Nuclear Regulatory Commission, Washington, DC, January 1988.
- 3. F. W. Sta11mann, Analysis of the A302B and A533B Standard Reference Materials in Surveillance Caysules of Commercial Power Reactors ,
NUREG/CH-494', ORNL/TM-19459, U,S. Nuclear Regulatory Commission, Washington, DC, January 1988.
- 4. L. F. Miller, C. A. Baldwin, F. W. Stallmarn, and F. B. K. Kam, Neutren Exposure Parayeters for th Q qtallurgical Test Specimens in the Fifth
_ _ l , Heavy -Sectica Steel Technology Ir radiation Series Capsules , NUREG/CR-5019, ORNL/TM- 10582, l' S. Nuc lear Regulatory Commiss ion , Washington , DC, March f 1 I 1988. 150 l 1
-. h w + ->
F g FUTURE'RESEARCH PLANS j
- 11. Continue data coll'ection for the EDB to obtain missing baseline data for.com-mercial power reactors, additional test reactor data, and materials test results. such as fracture toughness and hardness data.
?
- 2. LVerify,'with the' assistance of.EPRI and NSSS. vendors, the data in the EDB, m ,
identify, inconsistencies,.and establish procedures for the determination of uncertainties in the data. L
;_ -3. Continue software development for the EDB to facilitate data extraction, curve fitting, and graphic representations for regulatory use.
- 4. . Develop and test new models for embrittlement. predictions.
- 5. Upda'te the LSL-M2 program package and existing dosimetry cross-section data-in the associated data files.
- 6. Document calculational results to predict the fluence rates in the VENUS-3 benchmark experiments.
.7. Establish long-term continuation program to maintain and upgrade the RPV analysis. methods andl data bases in the areas of dosimetry, transport calcu-lations, and' damage correlation related to containment' performance. As part of this tash, continue active coordination with industry and'with foreign and
~
' domestic research. institutions in the area of reactor safety.
- 8. Apply the procedure validated in the VENUS-3 benchmark experiments to an operating. power reactor with the PLSA-low-leakage core design.
REFERFNCES
- 1. F B. K. ' Kam et al. , Pressure Vessel Fluence Analysis and Neutron Dosimetry, D*;EG/CR-5049,'ORNL/TM-10651, USNRC, Washington DC, December 1987.
- 2. ASTM. Standard E706-84', " Master Matrix for Light Water Reactor Pressure Vessel Surveillance Standards," 1987 Annual Book of Standards, Section 12, Volume 12.02,'American Society for Testing and Materials, Philadelphia, PA, 1987.
- 3. R. E. Maerker, Analysis of the NESDIP2 and NESDIP3 Radial Shield and Cavity Experiments, NUREG/CR-4886, ORNL/TM-10389, USNRC, Washington, DC, May 1987.
-4. C. Y. Fu and D. M. Hetrick, Update of ENDF/B-V Mod-3 Iron: Neutron-Producing Reaction Cross Sections and Energy-Angle Correlations, ORNL/TM-9964, I ENDF-341, Oak Ridge National Laboratory, Oak Ridge, TN, July 1986.
- 5. G. L. Simmons and R. Roussin, SAILOR - A Coupled Cross Section Library for Light Water Reactors, DLC-76, March 1983.
151 l
v: LWR PRESSURE VESSEL SURVEILLANCE i DOSIMETRY IMPROVEMENT PROGRAM: l I 1988 NIST Annual Report i Prepared by: E.D. McGarry
. lonizing Radiation Division Center for Radiation Research National Institute of Standards and Technology Gaithersburg, MD 20899 152
I OUTLINE OF NRC 1988 ANNUAL REPORT CONTRACT TITLE: DgSIMETRY MEASUREMENT REFERENCE DATA BASE FOR LWR PRESSURE VESSEL IRRADIATION SURVEILLANCE j CONTRACTOR AND LOCATION: NATIONAL INSTITUTE OF STANDARDS AND TECHNOLOGY (formerly NBS), GAITHERSBURG, HARYLAND ; NRC TECHNICAL MONITOR: A. Taboada ! PRINCIPAL INVESTIGATORS: E. D. McGarry-and J. A. Grundl OBJECTIVES: TASK STATEMENTS
- 1. CALIBRATION AND BENCHMARKKARK REFERENCING OF PV SURVEILLANCE DOSIMETRY 1.1 Certified Neutron-Fluence and Certified-Fissions Standards 1.2 NESDIP Program Participation 1.3 VENUS Program Participation i 1.4 Presentations, Reports and Meetings
- 2. LONG TERM OUALITY ASSURANCE AND MAINTENANCE OF PRESSURE SURVEILLANCE-DOSIMETRY q 2.1 Radiometric Dosimetry Measurement Assurance 2.2 SSTR Measurement Assurance 2.3 B&W Owners's Group Davis-Besse Benchmark
SUMMARY
OF 1987 RESEARCH PROGRESS ]
- 3. CALIBRATION AND BENCHMARK REFERENCING OF PV SURVEILLANCE DOSIMETRY ]
l . 3.1 Certified Neutron-Fluence and Certified Fission Standards 3.2 NESDIP Program Participation 3.2.1 Certified 2ssU Fluence Standards 1 3.2.2 The " ANSWERS" 3.3 VENUS Program Participation 3.3.1 NIST 252 Cf-to-Belgian 235 U Standard Field Calibration Transfer 3.3.2 VENUS Absolute Power Determinations 153 l
1-
)
3.3.3 Certified Fluence Standards for the VENUS II and VENUS III Campaign 3.3.4 Repair of BR-1 Cavity Fission Source Monitor 3.3.5 Re-interpretation of Emulsion Data from VENUS-I 3.4 Presentations, Reports, and Meetings 3.4.1 Compendium of Benchmark Fields for PV Irradiation Surveillance 3.4.2 Cancelled Meetings / Trips Instead 3.4.3 Report of Dosimetry for MEA Metallurgical Test Irradiations 3.4.4 Benchmark Referencing Standard Guide
- 4. LONG-TERM OUALITY ASSURANCE AND MAINTENANCE OF PRESSURE VESSEL SURVEILLANCE DOSIMETRY 4.1 Radiometric Dosimetry Measurement Assurance 4.1.1 PUD Neutron Sensor for Out-of-Core Reactor Dosimetry 4.2 SSTR Measurement Assurance 4.3 B&W Owner's Group Davis Besse Benchmark 4.4 Analyses of Transport Calculations Done In Support of MEA / Buffalo Irradiation Experiments
- 5. CONSULTATION 5.1 Shippingport Post-Mortem Dosimetry
- 6. FUTURE RESEARCH 6.1 NRC Dosimetry Reg. Guide 6.2 NUREG/CR: Calculational Benchmark 6.3 VENUS Program Participation 6.4 IAEA AG Meeting on Nuclear Data for Radiation Damage
- 7. REFERENCES 154
_ _ - . - _ _ _ . _ _ . _ _ _ _ _ _ _ - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ = _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ __ l l l l 1988 ANNUAL REPORT CONTRACT TITLE: DOSIMETRY MEASUREMENT REFERENCE DATA BASE FOR LWR PRESSURE VESSEL IRRADIATION SURVEILLANCE CONTRACTOR AND LOCATION: NATIONAL INSTITUTE OF STANDARDS AND TECHNOLOGY, (formerly NBS), GAITHERSBURG, MARYLAND NRC TECHNICAL MONITOR: A. Taboada PRINCIPAL INVESTIGATORS: E. D. McGarry and J. A. Grundl OBJECTIVES: Provide calibration of LVR-PV related surveillance dosimetry traceable to NIST standard neutron fields including distribution of neutron fluence standards. Maintmin a compendium of reference data for LWR-PV related benchmark neutron fields. Participate in the preparation of ASTM standards for routine LWR-PV dosimetry and in establishing consensus methods of data
' interpretation. Provide experimental assistance for LWR-PV dosimetry benchmark measurement programs.
TASK STATEMENTS
- 1. CALIBRATION AND BENCHMARK REFERENCING OF PV SURVEILLANCE DOSIMETRY 1.1 Certified Neutron-Fluence and Certified-Fissions Standards Prepare a NUREG report on the results of activation analyses of the 137 Cs radioactivity in 237 Np and 23eU fluence standards prepared and distributed for the LUR-PV-SDIP.
1.2 NESDIP Provram Participation Continue to follow and to support (with additional NBS fluence standards) NESSUS benchmarking of NESDIP/RPV cavity dosimetry measurements. i, 1.3 VENUS Pronram Participation 1.3.1 In support of the VENUS II campaign, make a fluence-rate transfer
-from the U.C. 252Cf standard neutron field to the Belgian Mark III 235g standard field before 1 February 1988.
1.3.2 In support of the VENUS II campaign, provide neutron fluence standards to Mol for: (1) Ba-La radiometric counting validation of fissionable foils; (2) 27Al(n ,a)2 ' Na radiometric counting validation in both thick and thin foils; (3) 64 2n(n , p)6 ' Cu radiometric counting calibration of coincidence counting technique. 1.3.3 Repair the BR-1 Cavity Fission-Source run-to-run monitor and return it to Mol by 1 March 1988. 155
f 1.3.4 Coordinate between HEDL and ORNL an interpretation of the HEDL emulsions irradiation in VENUS-I based upon approximately 50-group transport calculations of spectra at seven locations in the VENUS-I benchmark experiment. 1.4 Presentations. Reports and Meetinzs 1.4.1 Continue development of the Compendium of Benchmark and Test-Region Neutron Fields for PV irradiation surveillance. 1.4.2 Contribute sections to one Progress Report and submit a Branch Annual Report. 1.4.3 Participate in NESDIP Workshop Meeting to be held at Winfrith (ENGLAND) approximately 1 February 1988. 1.4.4 Attend the LUR-PV-SDIP Program Meeting May 1988, Toledo, Ohio, 1
- 2. LONG TERM OUALITY ASSURANCE AND MAINTENANCE OF PRESSURE SURVEILLANCE DOSIMETRY 1
l 2.1 Radiometric Dosimetry Measurement Assurance For quality control checks on neutron activation determinations by vendors and service laboratories, distribute foils irradiated to certified fluences in the 252Cf and 235 U Standard Fission Neutron Spectra. 2.2 SSTR Measurement Assurance j 1 Implement a minimum-requirement NUREG, as defined in FY-87, for quality assurance of SSTR dosimetry masses. 2.3 B&W Owners's Group Davis-Besse Benchmark Continue support of this benchmark activity in the areas of radiometric and SSTR neutron-dosimetry measurements and LiF-chip gamma-ray dosimetry measurements.
SUMMARY
OF 1988 RESEARCH PROGRESS
- 3. CALIBRATION AND BENCHMARK REFERENCING OF PV SURVEILLANCE DOSIMETRY 3.1 Certified Neutron-Fluence and Certified Fission Standards Table 3.1.1 gives an NIST analysis of data received from the Idaho Nuclear Engineering Laboratory (INEL) for neutron fluence standards supplied to INEL by NIST to benchmark the radioactivity counting of dosimetry from the MEA / Buffalo metallurgical irradiations. The table is part of a NUREG-in-preparation for fluence standards issued for the LWR-PV-SDIP and related irradiations experiments. Considerable detail is purposely included on the table so that nuclear parameters for the gamma-ray counting system get documented with the analyses of the fluence standards. For this annual I report, the columns of interest are those specifying the types of fluence 156 l' . - - - _ _
is 6 l 1
-J TABLE 3.1.1 RISULTS OT NEUTRON TLUENCE STANDARD COUNTINC BASED ON REDUCTION TO MEASURED CROSS SECTION Reporting Laboratory: Idaho Nuclear Engineering Laboratory (INEL) -
A. .Neasured Activity at E0I and Derivation of Average Reaction Rate I . D .' observed Activity QE01 Number of Decay -Average Dosimetry Nuclei Decay Correction Reaction T1uence Reaction Reported Stand ard a Yield IC) Constant Factor (d) RateI ') Stand ard Irrad. Format (s) Forcat A N) NY . A(s~l) C- (D Te-NI-A Ti/Te-2 547e(n,p)54Mn 4.407E+00 2 222E+03 2.000E+20 2.567E-0R' O.9957 1.2A2E Te-Ni-A Ti/Te-2 ssNi(n p)ssco . 1 630E+02 8.22 E+04 1.260E+21- 1.133E-07 0.9810 . 1.731E-15
! Ni-C U/Te-1 sagi(n,p)SBCo 5.209E+02 1.195E+05 1.98 E+21 1.133E-07 0.9810 1.994E-15 T1-3 U/Te-1 h6Ti(n,p)*6Sc 6.362E+00 1.944E+03 3.07 E+20 9.570E-08 0.9839 1.974 E-16 UN-51 U/Te-3 23 s U(n , f)l C 3Ru 9.141E+01 7.49 E+04 1 34 E+20 2.035E-07 0.9474 5.40 E-15 UN-51 U/Te-3. 238U (n.f)s5Z r 4.664E+01 3.82 E+04- 1.10 E+20 1.252E-07 0.9672 5.35 E-15 UN-51 U/Te-3 235U (n f)l"0Ba 2.344E+02 1.92 E+05 1.27 E+20 6.273E-07 0.8492 5.29 E-15 UN-31 U/Te-3 23sU(n,f)3a-La- 2.341E+02 1.92 I+05 1.27 E+20 6.273E-07 0 8492 5.23 E-15 UN-51 U/Te-3 288U (n,f)l37Cs 3.311 E+01 2.71 E+02 1.27 E+ 20 7 160E-10 0.9998 5.56 E-15 B.' Derivation of Observed Cross Section and Comparisons with Published Experimental values, and with Calculated Values for Neutron Dosimetry Standardization.
I.D. NSS Average Cross Section Experimental Ration Calt.ulated Ratio: Fluence Deduced from value Deduced Deduced Fluence Rate. Reported Data (835 Experiment Cross calculated Standard . Reaction (c)=t/T < D/(4) Cor.pe nd ium) Section U) ; I Te-Ni-A Skfe(n,p) 1.553E+10 82.5 eb 81.7 cb 1.010 81.0 mb 1 019 : Te-Ni-A $8N1(n,p) 1.553E+10 !!!.4 111.0 1.004 105.0 1.060 'I ' Ni-C 58Ni(n,p) 1.798E+10 110.9 111:0 0.999 105 0 1.056 Ti-3 65Ti(n.p). 1.713E+10- 11 5 11.8 0.975 11.2 1.027 UN-51 23sU(n f)Ru 1.712 E+ 10 314.5 312.2 1 008 305 2 1.030 UP51 238U (n f)2r 1.712 9 10 312.2 312.2 1.001 305.2 1.023 0951 21se (n,f)sa 1.712E+10 308.5 312.2 0.989 305.2 1 011 UW5] - 2 3 US (n, f)Ba-La 1.712E+10 308.4 312.2 0.988 305.2 1.010 UN-51 2 eU(n,f)Cs 1.712I+10 324.8 312.2 1.041 305 2 1.064 I')0uantity reported (with saca attenuation correction included): observed dps of reaction pro:!uct at EDI per c3 of foil. (b) Tree-field dps of reaction product at E01 = (Reported Tor: rat) = (foil eass)/(1+p,c). The scattering l correction, (1+p,g), is given in the test report. A 235U fission correction (2 2%) is included for the ! 238U fluence standard (UN-511 (c)Nusber of reaction isotcpe atoms in foil x fission yield when appropriate. (d)specified in the test report. For an uninterrupted irradiation of length T at a censtant fluence rate, j- C is equal to {(1 - exp(- AT))/ AT). t I, (*) Average reaction rate t <D = o<6) = A/( ACTNY), where (4) is the NBS certified fluence divided by the I length of the irradiation T as specified in the test report. As a measured quantity, <D may be identified with the " saturation activity" per nucleus as employed in most ASTM standards, notably E261. .l II)Value calculated with 235U fission spectrum shape and dosimetry cross sections from ENDF/5-V. 157 i
standards (see the " Dosimetry Reaction" column) and the last five columns
-of Part-B of the table, which provide an evaluation of " Cross Section Deduced from Reported Data <R>/<4>." <R> is an average saturated reaction rate derived f rom . the INEL data reported in Part A and <d> is the NIST (formerly NBS) supplied fluence rate.
3.2 NESDIP Pronram Participation Since the ' start of the United Kingdom's NESDIP Program in 1983, NIST has played only a minor role in the NESDIP dosimetry efforts. That role, to date at least, has been work to better understand the qualification and use of the NESSUS Irradiation Facility (1) in the NESTOR Reactor (which
" drives" the fission plate source for the NESDIP irradiations) as a i dosimetry-benchmark irradiation facility.
3.2.1 Certified 235U Fluence Standards Five certified 235U fluence standards, including 2 3 eU(n, f)t 3 7Cs . one of the standards with a 30-year half life to be recycled .. were sent to Winfrith in 1987. NIST has still not received results of the gamma-ray counting and analyses. l 3.2.2 The " ANSWERS" NIST has recently learned that technology exchange with Winfrith may be at an end. It is understood that J. Butler's Radiation and Shielding Group has changed into a commercial venture, with Peter Miller as business manager. It seems the group's name is now ANSWERS: " Application codes for Nuclear Systems using services based on Hinfrith Expertise in Beactor physics and Shielding." Information and irradiation services are now for sale. Even so, NIST feels it is to Winfrith's advantage to complete work on the fluence standards. At'least they should return the 30-year 137 Css standard. 3.3 VENUS Program Participation NIST.has been involved with the CEN/SCK VENUS series of LWR-PV benchmark experiments since the start of VENUS Phase I in 1983, through VENUS Phase II, which started in 1986, and now into VENUS Phase III, started in 1987. ) The primary NIST function has been to provide supplementary benchmark . calibrations and backup (but " hands-on") verification of the CEN/SCK ' benchmark referencing which hac provided the common denominator for the dosimetry through all phases of the VENUS experiments. i 3.3.1 NIST 252 Cf to Belgian 235 U Standard Field Calibration '"ransfer NIST personnel traveled to Mol, Belgium to re-establish the tie between the , NIST standard neutron fields and the CEN/SCK 235 U Cavity Fission Neutron ! Field. The basic calibration measurements were made with an NIST dual l fission chamber containing two of the same known-mass fissionable deposits as were used in the calibration accomplished at Mol by NIST (then NBS) in 1983. J i The measurements are straight forward. The fission rates of known mass 158
_ _ , _ _ _ _ _ _ = _ _ _ _ , deposits are measured in the CEN/SCK 235 U fission spectrum relative to the absolute reactor power, which is continuously (i.e., "over the years") tracked with a separate fixed fission chamber that serves as a run-to-run monitor. The run-to-run monitor operation is independently calibrated by NIST against a known 239 Pu fission rate per Certified fluence rate in the field of a 252 Cf source of known source strength. The source strength is determined at NIST by the MnSO,, Bath Technique. In short, NIST establishes ) l the calibration of the run-to-run monitor and all other calibration B measurements are normalized to each other in terms of their response per monitor count rate. This is why repair of the CEN/SCK run-to-run monitor was such a major undertaking. (see Sect. 3.3.4) Confirmation (to former results. .1983) were made, as mentioned previously, with two of the same masses as were measured in 1983. Agreement is within the 1% (la) level. 3.3.2 VENUS Absolute Power Determinations To properly compare absolute transport calculations with fluence rate measurements in the VENUS Facility, the absolute core power, or equivalently the volume-integrated fission rate in the core, must be known. The VENUS Program requires at least two independent determinations of the power. NIST became involved in the VENUS power determination when, in 1980, the mass of 23 5U in a miniature sealed fission chamber (Mol I.D. 1879) was determined by experiments in the NIST reactor thermal column (2). With a samll scaled fission chamber it is possible to probe down inbetween the fuel plates in a low-power core and, without disturbing the thermal fluence rate, make vertical and horitontal scans of the absolute fission rate over the core volume. This powerful technique for determining absolute core-power levels was first employed in PCA (3), using the known-mass fission chamber. The reported mass of 4.72 2% micrograms of 235U was determined by comparing fission rates with a known-mass deposit from the NIST inventory. That deposit was #25A-03-1 with a 235U mass of 31.22 1.4% micrograms. Because of corrections for such things as gradients, relative spatial locations, and thermal fluence perturbations, this mass has been the subject of many late-night discussions. It was, after all, a fundamental constant of the PCA Experiments in that it was used as the one method to determine the absolute core power of PCA and a bias in its magnitude, or in its assigned uncertainty, would directly change all experimental results. This dependency was a primary reason for VENUS requiring at least two independent methods of power calibration. Now, after eight years of additional work in which the Belgians have considerably advanced the state-of-the-art in the determination of absolute power levels in experimental reactor facilities, the "4.72 g mass" stands as the constant that links three VENUS campaigns to PCA, because the cross-calibrated companion to the subject miniature chamber (the compamion chambers identification is Mol I.D. 1880) has been extensively measured in all of the experimental configurations throughout ten years of effort. This conformation of power scales has special significance at this time. With reference to the discussion in Sect. 6.2, there are plans to turn the , 1 159 I
. - - - _ _ - _ _ _ _ . - - _ _ . . - _ - _ _ - _ _ _ ________________a
ORNL/ PSF (4/12 + SSC) Benchmark Experiment Series, particularly the SDMF-4 Dosimetry LIntercomparison Experiment, into an improved calculational benchmark which will have " benchmarked" experimental data for the same isotopic reaction rates as are used in commercial power reactor surveillance. There is a direct connection, of the power-level calibration at PCA to power-level determinations for the various PSF Experiments. 3.3.3 Certified Fluence Standards for the VENUS II and VENUS III Campaign The " fluence standards business" with CEN/SCK is a lot more interesting than with most other laboratories because we can exchange radio-artifacts l 1
-(radiometric fluence standards) from each other's standard fission spectra.
This was done in 1983 and again in 1988 for the seNi(n.p) and 225 In(n,n')
)
reactions, j The 1983 inter-calibration resulted in an adjustment of about 0.8% in the calibration which had been in effect since before the start of LVR-PV-SDIP - . in 1977. The December 1988 intercalibration is not yet complete but I preliminary analyses suggest no changes. The 1988 inter-calibration also included 235 U spectrum irradiations of 27 Al (n,a) and will include some work with 8'Zn(n,p), where the latter is a newer dosimetry technique with a unique coincident counting method. 3.3.4 Repair of BR-1 Cavity Fission Source Monitor The repair of the Mol 235 U Cavity Fission Source monitor was completed at NIST before the February 1989 deadline. In fact, it was repaired, and then delivered and checked out by an NIST representative in early December in conjunction with Fluence-Level Transfer from the NIST 252 Cf Standard Field. 3.3.5 Re-interpretation of Emulsion Data from VENUS-1 The issue here was that HEDL (Gold) feels that Nuclear Research Emulsion 1 (NRE) technique has the capability to serve as a differential neutron- ; energy spectrometer, especially in the energy range 0.3 MeV to 1.0 MeV. 1 This requires selective processing of integral data in some pre-defined energy-bin structure. Gold felt that some of the NRE data from VENUS-I inoicated that there were more neutrons (by about 20%) in the mentioned ] energy range than was predicted by the transport theory. However, the analysis at that time was questionable because the energy group structure of the available transport calculations was too coarse for proper data interpretation. McGarry was commissioned to get a finer-group transport spectrum from ORNL so another analysis could be made. According to private communications with W. McElroy, the spectrum was received, an analysis was accomplished, but the results are not conclusive. There are no finer-group transport calculations available so it seems that l the issue will be dropped. 3.4 Presentations. Reports. and Meetings 3.4.1 Compendium of Benchmark Fields for PV Irradiation Surveillance 160
.l l
j l No new entries were made to the Compendium this year. It is planned to have the Compendium provide a " summary description" of the newly-proposed PSP (4/12+SSC) Calculational Benchmark when it is ready (see Sect. 6.2). 1 3.4.2 Cancelled Meetings / Trips The February'1988 NESDIP Workshop at Winfrith, England and the May 1988 LWR-PV-SD1_ Program Meeting at Toledo, Ohio were cancelled. Instead, McGarry traveled to Richland, WA in October 1988 to assist with the publication of NUREG/CR-3321(4). 3.4.3 Report of Dosimetry for MEA Metallurgical Test Irradiations Two papers were presented by McGarry at the 14th Symposium on The Ef fects of Radiation on Materials at Andover, Massachusetts in June 1988. The paper by principal author Prillinger (University of Stuttgart) dealt with transport calculations of the ex-core MEA experiments at Buffalo. See Sect. 4.4 The paper by principal author McGarry presents the effects on spectrum-averaged dosimetry and dpa cross sections because of different radiation environment factors considered in the Prillinger calculations. 3.4.4 Benchmark Referencing Standard Guide A revised draf t of this standard, E706 IIB (7) was accepted for ballot by the ASTM E10.05 Subcommittee meeting in Orlando, Florida, January 1988.
- 4. LONG-TERM OUALITY ASSURANCE AND MAINTENANCE OF PRESSURE VESSEL SURVEILLANCE DOSIMETRY 4.1 Radiometric Dosimetry Measurement Assurance 4.1.1 PUD Neutron Sensors for Out-of-Core Reactor Dosimetry Second only to the fission of 237Np for good spectrum coverage of neutron with energies E > 1.0 MeV, 238 U is the desirable reaction for PV surveillance dosimetry. However, in partially thermalized spectra, the fission in the natural 235U contaminant of 238U can require a significant correction. A method to avoid the correction is to use very depleted uranium; but this approach is very expensive. The Paired Uranium Detector (PUD) dosimetry technique was developed at NIST to circumvent the need for the highly depleted 238 U material.
4.1.1.1 Brief Outline of the Method The scheme of using paired uranium detectors involves simultaneous irradiations of a not-so-highly depleted uranium (approximately 200 pm 235 0) dosimetry with a natural uranium foil, which is readily available and has 0.7 w/o 235U. The latter serves as the low-energy-neutron fission-rate monitor. The small concentration of 23 5U, has substantially less self absorption than an enriched uranium dosimeter. Therefore, the 23 sU-to_2 3 s U spectral index can be measured and corrections for the 235U fission in depleted 2 3 8U can be easily and accurately made. 161
~
l
) a m
- r e % % % u
( ut 3 9 7 i . e t a . . n d t t aR 1 1 3 a e n a N r d e R u d m a e n l c o a e n i d r . b a t e u d h c t % % % t e t n a e e 8 6 4 a s s E e l t . . n a u R pa 0 0 0 e m _ y eR d r t e D n c h n v a n c i A i i a d . h t e e w e r l) t b s e ab e ) - n c rt 3 4 8 l t 2 2 o n t x . . p s AA p U c e 1 1 1 e u / s ed d m 3 r e pn A o R S I e 2 h A n 1 r t / i A o 1 t n A ys c i t e 1 e ) o ni t
- l 1 4 9 5 i i t 1 e ( aU 8 7 5 3 t ai D e r5 . 2 a t v 4 s u3 0 0 0 U r ri D n t2 et E U o a d c cc L P p N n i na B s a f u A f e i c T o R 8 c % i d 3 e 5f s e eU _
2 p 1i e p t8 9 3 6 U s c r o e3 8 9 9 l e u t l2 . . . m e ap t o p 0 0 0 o r h ns a e F e I s D f t n imoi'n
- g i n l s ) n _
y
'a e d) i y at r f ci el 1 8 3 m t n e oit t a 2 2 4 o n oi n fi e r c i t a e oiv l u 0 0 0 a gr t
G i ci pt e t t et e a s r ne apc d n n e i c RS A ( ( o c d n 3 2 p n nu o AA s u e pe r e sh l 8 0 0 0 h et ax 2 0 0 0 l t r re o 6 4 2 a ro td / t h ot cn 5 o c c eI 2 t i ( e p a h r S f w yu o . t t e s y na R c d nr b i r W y n e oo ad P t a t i t r t a i l an t c o ru l v ll e ro ce t eq aa ail ei at c c cC c eu l t re a nn i iprvs ei Fd F Ui p p cs ) ) ) y yua co
- b
- T TSC AP ( ( (
_ 0
The zasU detector of the PUD pair may be directly calibrated in the 235U fission neutron spectrum of the Cavity Fission Source at NIST and the ( quality. assurance of the depletion checked in a thermal neutron beam. General features of PUD detector measurements for typical out-of-core positions are displayed in Table 4.1.1. These positions are characterized by the zasU/ 2asU spectral indexes shown in column two. The ratio of specific activities for the detector pair, shown in column three, give observed spectral indexes directly in a simple expression involving isotopic atom fractions for the two uranium materials. The composite zaag and 235 U fission product activity in each detector is indicated in columns four and five of the Table 4.1.1. The 23sU response in the depleted uranium dominates at all positions which, of course, is the desired response. In the natural uranium, however, the 235U and 238U become equivalent near the accelerated position. Because of composite 23eU and 235U response in both detectors, the uncertainty in a derived spectral index is enhanced. The factor by which the uncertainty in the specific activity ratio, and hence the spectral index, must be enhanced is given in column six of Table 4.1.1. A similar increase for the derived average reaction rates depends upon the value of ; the specific activity ratio. In this case an uncertainty is added in quadrature to the uncertainty in specific activities. For a nominal activity ratio uncertainty of 5%, the uncertainty to be added in quadrature is given in columns seven and eight of Table 4.1.1. Uncertainty enhancement for the depleted uranium is small compared with other typical experimental uncertainties. For natural uranium, however, it may become significant as the spectrum hardens. 4.1.1.2 First Application in Surveillance Capsule Perturbation Experiments PUD detectors were first employed in the SDMF-4 Surveillance Capsule Perturbation Experiment. Detector pairs furnished by NIST were placed, along with conventional single isotope radiometric fission detectors, at five positions in the assembly. Specific activities were measured by HEDL I and are reported in NUREG/CR-3321. Derived spectral indexes and average reaction rates are shown in Table 4.1.2. Results are reasonably consistent among the three fission products counted, including the sometimes troublesome 1 3 Ru activity. The observed spectral indexes indicate an expected spectrum hardening in going from the water / steel interface (OT) to the quarter thickness (QT) as epithermal i neutrons from the water are absorbed. Continued degradation of the spectrum occurs deeper in the steel as inelastic scattering in iron transfers neutrons to below the assU response threshold. I l l The calculated spectral indexes show a quite different trend not in ; l agreement with experiment (column five), the mismatch of spectral indexes is to be associated with the 235U detector. Inadequate treatment of low- ! energy neutron transport in the calculation may be responsible. Such l shortcomings are often dismissed as unimportant because low-energy neutrons do not seem to contribute to steel embrittlement. The lesser disagreement of calculation with the 2aaU response, however, cannot be neutron-induced i embrittlement. Column five of the table shows a diverging trend between i 163 . l i l (
)1 11i Il 1 ;Il! ;]i 3Ii ilI il !jl j]li1!
_ ~ h 1 _e l u s p
, , a rc . e t ss t e ct e rh ucv . at n : . d eis u o oD c 6 2. 1 8 2 ofl t qd i iU l 9 0 1 1
- 9. '
rf c .n tU tP a ; . . pef u =i c8 a C 0 1 1 ' 1 0 l d h a3 R nl ao Te e2 oehr p Qb R iv T e . s er- ,xo
'e t 3 3 4 4 5 sl an ga 1 1 1 1 1 i - eo Tb aR - - - - - f eyi H r 'D E E E E E t - s ,d e U 8 3 0 1 9 ua0 s i s v P 5 3 9 3 R r3i To t A . 7~ 3 - f Qv n 2 1 5 2 2 0 en e 1 cau ;
=
m - nhR e i d e t3 rB r nu 0 oV e al r1 c p f e ; x , td re E rt rn ol Z noa t u 4 5 eh cs ap
- 9 rsr F e Z ea M ,f e5 rc D : .
038 482 578 900 779 s f h9 ge 2 S o c 666 034 380 034 867 Cit
~ iD l -
7d e nh 1 n tU a 122 1 1 1 1 56 1 22 022 3 fh it
. i ) aP C 1 wot c 4 x8 R ro ao s e3 oh er f t E t d2 f s so n L l nU uf ei B u I ( sd a c A s a tl c) as T e l/ l uey f s R a) uobl re r5 sh e un r t3 esnvoi sk o c2 r c t eU sit ri c p( D 1 1 3 235 502 261 303 eet c eh e S a U 446 21 1 344 889 722 ruae t t t P 222 222 222 222 677 al i p u e . ad s or D rvae e rr =t
- e D d er r U ) rei or - y T a P a O u q
( hf a n nt od e-o i re i l d9 ae t sl n3 r i - eAe sh s T i d nt o T T T Q B r en o - P O Q H T V t .h a id ngt t n en y i a id a s et rd o , ena Pf ruw4 ) l h oo6 a a Tct( ( h s2
- 3 t lli fu
I l measurement and calculation that amounts to 23% over the thickness of the surveillance capsule and then changes drastically in the water behind the capsule. 4.2 SSTR Measurement Assurance Measurement assurance activities with respect to SSTRs continues. NRC monies have been used to keep NIST actively involved in the general trend by industry to implement SSTR dosimetry in ex-vessel cavity measurements. The bulk of this work is taking place in Westinghouse PWRs and in the B&W Owner's Group Davis Besse Dosimetry Benchmark Experiment. NIST Process and Quality Control Funds are supporting a very active program with Westinghouse Research to establish the mass scales for ultra-light masses (sub-nanaogram to picogram) SSTRs. The idea of a minimum-requirement NUREG for quality assurance of SSRT j dosimeter masses is still active. The emphasis will increase as more ' experimental data becomes available to support the current progress with measuring ultra-light masses. In the interim, W/R&D and NBS are co-authoring a paper on the " state-of-the-art" results (see Sect. 6.5). 4.3 B&W Owner's Group Davis Besse Benchmark In addition to direct NIST involvement in the dosimetry for Davis Besse, which has been reported (see FY-87 NBS Annual Report for NRC LWR-PV-SDIP) and involves PUD radiometric detectors (see Sect. 4.1.2, this FY-88 report) and LIF chips, or advanced technology lithium fluoride gamna dosimetry, NIST supports some QA of B&W SSTR masses throagh the tie with Westinghouse Research. Furthermore, as the time for radiometric assay of an extensive amount of dosimetry approaches, benchmark referencing of the counting and interpretation of specific activities is planned. The plans are dependent upon some unanswered questions about which laboratories will do such analyses. 4.4 Analyses of Transport Calculations Done In Suonort of MEA / Buffalo Irradiations Experiments As indicated in Sect. 3.4.3, neutron transport calculations have been done ( by Prillinger, University of Stuttgart) in support of the PV steel irradiation experiments carried out by Materials Engineering Associates (MEA) at the Buffalo Materials Research Reactor. The calculations were a study to determine spectrum average cross sections to be used for interpretation of in-situ dosimetry data from particular MEA experiments. The study involved a series of transport calculations because during the course of irradiations, which lasted over three years, there were a number of neutron environment perturbations of previously unknown significance. 1 The calculational study addressed the issues of determining "best" spectrum l average cross sections and uncertainties which represented the effects of these uncontrollable conditions in the irradiation environment. NIST has examined the calculations and reported on the significance of the perturbations to spectrum averaged cross sections used to reduce the dosimetry data. I 165
When a series of transport calculations are done to examine the effects of variations of the radiation environment on the spectrum in a praticular experiment, the absolute fluence scale of the calculations is usually not a meaningful quantity comparison with dosimetry measurements. Table 4.4.1 shows data frc 1 dosimetry experiment at the center of a water-shielded dry-standpir facility. Four isotopic reactions are involved: 63 Cu(n,o), '6 Ti(n,p), S ' Fe (n , p) , and $sNi(n,p). The dosimeters were irradiated in the form of wires and the identification numbers in the table are for selected cuttings, chosen at different heights, from each of the different dosimeter wires. The next to the last column gives the saturated f reaction rates per nucleus for the various cuttings. The magnitudes of these individual reaction rate differ significantly. This makes selection of an average, or some representative value for comparison with the i calculations, diffucult. However, when ratios are taken with respect to a particular reaction for similarly positioned cuttings (in Table 4.4.1 ratios are formed with respect to the 5 ' Fe (n , p ) reaction), the normalization, as shown in the last column, allows the ' derivation of a meaningful average and a standard deviation of at most 2%. These ratios are spectral indices; ratios of spectrum averaged cross sections. It is the spectral indices which serve as the basis for confrontation with the calculated quantities.
- 5. CONSULTATION 5.1 Shinoinroort Post-Mortem Dosimetry Early in FY-88, NIST was quite active in obtaining documentation and discussions with people at Argonne and Westinghouse concerning the possibility of establishing a fast-neutron data base for the current post-mortem metallurgical testing of cuttings from the Shippingport ex-vessel shield tank and the pressure vessel itself. This activity has ceased, because of the following:
(1) There is literally a ton of paper documenting Shippingport. Most of the information is still in the archives. Because it consists of a wide variety of reports, memos, calculations, etc., about 95% of the information is not applicable to the fast fluence problem. It would require an estimated man year of effort to arrive at an average pressure vessel fluence and a power-time history with an estimated accuracy of 25% to 30%. Both pieces of information are required for the PV and the ex-vessel structure. Whether the accuracy would be as good as 30% for the latter is not known; probably not. (2) There is already some total fluence information available (Greenwood, Argonne) from key memos. This is estimated to be uncertain to about a factor of two. A lot of the uncertainty is associated with large azimuthal variations which resulted from a number of different unique-geometry cores. Coupling the need for an azimuthal fluence-distribution history with the need for a power-time history, may double the effort required. 166
1 i- \. 1 1
> 1
.1 1
4 TABLE 4.4.1 VALUE OF USING SPECTRAL INDICES TO EXAMINE THE CONSISTENCY OF DOSIMETRY DATA DURING THE DATA REDUCTION PHASE OF THE EXPERIMENT 1 FISSION FISSION CROSS RATIOS
' DOSIMETER FLUX SECTION REACTION W.r.t. MAX REACTION .I.D. (>l MeV) (>l MeV) RATE FE-54 MIN (x.E+10) (mb) .(x E-15)
CU-2 7.990 0.867 6.9273 0.00924
'CU-63(N,A)CO-60 CU-6 8.100- 0.867 7.0227 0.00975 CU-8 7.660 0.867 6.6412 0.00958 CU-12 5.990 0.867 5.1933 0.00986 AVERAGE 0.00961.0.009845 STD. DEV. 0.00024.0.009370 I
TI-2 7.460 17.3 129.0580' O.17209 TI-46(N'P)SC-46 TI-6 , .7.230 17.3 125.0790 0.17372 TI-8 6.840 17.3 118.3320 0.17063 TI-12 5.320 17.3 92.0360 0.17482 AVERAGE- 0.17281 0.174406
'STD. DEV. 0.00159 0.171221 FE-2. 6.510 115.2 749.9520 1.00000 TE-54(N,P)MN-54 FE-6 6.250 115.2 720.0000 1.00000 FE-8 6.020 115.2 693.5040 1.00000 FE-12 4.570 115.2 526.4640 1.00000 AVERAGE 1.00000 STD. DEV. 0.00000 NI-2 6.380 156.8 1000.3840 1.33393.
NI-5 B (N , P) CO-58 FI-6 6.200 156.8 972.1600 1.35022 NI-8 5.910 156.8 926.6880 1.33624 f'-
- NI-12 4.580 156.8 718.1440 1.36409 AVERAGE 1.34612 1.35822.*
L STD. DEV. 0.01210 1.334017 H H \" } '. l 167 l i g . - -__ ___ ____ __-__ - - _____-___ -
- 6. FUTURE RESEARCH 6.1 NRC PV Surveillance Dosimetry Reg. Guide
~
Review of the first (Carew) draft, considerable thought, attendance at one meeting, and grappling with an overall outline were NIST activities in CY-88 regarding this much needed Reg. Guide. NIST feels that there are still unresolved issues about how the industry will realistically handle ,
" reconciliation of differences between measurements and calculations" and l " extrapolations of 2-D calculations from measurement locations to ... the l 3-D world". NIST will submit an overall outline for consideration, which at least hopes to address these issues. However, top priority is being given to the text, on Dosimetry Measurements and Benchmark Referencing. !
The existing FY-89 programm brief is being modified to reflect emphasis on this task without an increase in FY-89 funding. 6.2 NUREG/CR: Calculational Benchmark A maj or accomplishment of the LUR-PV-SDIP was to develop specialized benchmark neutron fields and carry out experiments in them to validate dosimetry methods and procedures. The PCA Benchmark Experiment was used for a calculational blind test of transport theory (3). Subsequently, a modified configuration of the PCA (the PCA 4/12 + SSC) was built and neutron-spectrum characterized to serve as a low-powered " critical facility" for the ORNL/ PSF. Seven benchmark experiments were irradiated in a considerably higher fluence of neutrons from the Oak Ridge Research Reactor operating at 30 MW. Five of the seven experiments were benchmarks in the mentioned (4/12 + SSC) configuration and all of these contained a variety of neutron dosimetry and had associated transport calculations. Consequently, it is now apparent that an improved calculational benchmark j is a universal version of the PSF (4/12 + SSC) benchmark field, j one immediate problem that is solved by adoption of the above idea is that of providing " proper" dosimetry results for comparison with calculations. Since the PCA had a fluence environment about 4000 times less intense (i.e., 7 kW/30 MW) than the comparable PSF version, the dosimeters used are not the same reactions as were used in the PSF, and more important, that are used in power reactors. Therefore, a second formerly unidentified task (not in the original FY-89 Program Brief) is to collect all core- and structural-design information, collect and critically review all dosimetry measurements and their related transport calculations, and document (in a NUREC) the information to be supplied to the Reactor Shielding Information Center (RSIC) for distribution as an NRC Dosimetry Calculational Benchmark Problem. The existing FY-89 programm brief is being modified to also reflect emphasis on this task. j l 6.3 VENUS Program Participation ] l With NRC support, NIST continues to interact with the VENUS Program. The I current areas of emphasis are benchmark activities which inter-relate f j dosimetry and fluence-rate calibrations and measurements in the various l l phases of the VENUS experiment. One hold over from FY-87 is work to be I l completed on the benchmarking of the " Zn(n,p)" Cu coincident-counting ] 168 I _ _ _ _ _ _ _ _ _ _ _ _ __ l
1 l s dosimetry technique in use at Mol, Belgium. 6.4 IAEA- AG Meeting on Nuclear Data for Radiation Damage The IAEA Advisory Group Meeting on Nuclear Data for Radiation Damage Assessment and Related Safety Aspects will meet in September 1989 in Vienna, Austria. This is the first meeting since October 1981 and NIST participation has been requested. Consistent with NRC Regulatory Research plans, one person from NIST could participate. In any event, NIST (E. D. McCarry) will co-author a paper with Westinghouse Research (F. H Ruddy) entitled, " Benchmark Referencing of Ultra Low-Mass, Solid State Track Recorder Neutron Dosimeters in NBS Standards Neutron Fields." ,
- 7. REFERENCES (1) M. D. Carter, I. J. Curl, M. F. Murphy and A. Packwood, "The NESSUS Reference Field in the NESTOR Reactor at Winfrith", Proc. of the 5th ASTM-Euratom Symposium on Reactor Dosimetry. Geesthacht, Federal Republic of Germany, September 24-28, 1984, EUR-9869, Commission of the European Communities, 1985.
(2) E.D. McGarry, " Determination of the Effective Mass of 235U in the Fission Chambers Used for Absolute Core-Power Measurements in PCA," 1RR-PV-SDIP Ouarterly Procress Report. April 1980 - April 1980 - June 1980, NUREG/CR-1241, Vol. 2, HEDL-TME 80-87, NRC, Washington, DC, pp. NBS-13, July 1981. (3) W. N. McElroy, Ed. , IRR-PV-SDIP- PCA Experiments and Blind Test, NUREG/CR-1861, HEDL-TME 80-87, NRC, Washington, DC, July 1981. (4) ASTM E706-87, " Master Matrix for LWR Pressure Vessel Surveillance," 1988 Annual Book for ASTM Standards, American Society for Testing and Materials, Philadelphia, PA, Section 12, Vol. 12.02, 1988. I l 1 169
l 1988 NRC BRANCH ANNUAL REPORT REGULATORY ANALYSIS FOR REVISION OF REGULATORY GUIDES 1.83 and 1.121 Principal Investigators: R. J. Kurtz R. C. Hanlen T. V. Vo-December 1988 Prepared for the Division of Engineering Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission under Contract DE-AC06-76RLO 1830 NRC FIN 82097 i 1 Pacific Northwest Laboratory Richland, Washington 99352 170
[.
. STEAM GENERATOR INTEGRITY PROGRAM L REGULATORY ANALYSIS FOR REVISION OF REGULATORY GUIDES 1.83 and 1.121' Pacific Northwest Laboratory P.O. Box.999 Richland, Washington 99352' Principal Investigators:
R. J. Kurtz, R. C. Hanlen, and T. V. Yo ABSTRACT H This report summar)izes Laboratory (PNL)(a on work performed regulatory in FYfor analysis 1988 theby Pacific of revision Northwest Regulatory Guides 1.83 and 1.121. ~ Additional evaluation of sampling / inspection
. strategies for in-service' inspection of steam generator tubes by Monte Carlo simulation techniques is described. Initial work on a value-impact assessment of proposed revised Regulatory Guides 1.83 and 1.121.is discussed.
- Work with the American Society of Mechanical Engineers (ASME) Code to incorporate. improved eddy-current inspection techniques and performance demonstration requirements is summarized. Interaction with Program for
- Inspection of: Steel Components (PISC)-III to communicate final results from the Steam Generator Tube Integrity Program / Steam Generator Group Project (SGTIP/SGGP).'is mentioned.
OBJECTIVE The objectives of the program performed at PNL are the following:. 1) to develop recommendations for revision of Regulatory Guides 1.83 and 1.121 utilizing research results from the SGTIP/SGGP and the relevant technical-literature, 2)- to conduct a value-impact assessment of the revised regulatory guides to determine safety and cost benefits, and 3) to pursue adoption of improved eddy-current' techniques and performance demonstration requirements by the ASME Code. FY 1988 SCOPE The FY 1988 scope included: 1) Monte Carlo simulation analysis of other possible sampling plans, such as those proposed by industry for comparison to the proposed revision of Regulatory Guide 1.83; 2) a value-impact analysis of-revised Regulatory Guides 1.83, Rev. 2 and 1.121, Rev. 1 to determine the safety and cost impacts compared to existing NRC regulations, current or c proposed field practices, and the impact of backfitting; 3) work with the l ASME Code to incorporate improved eddy-current inspection techniques and (a) Operated for the U.S. Department of Energy by Battelle Memorial Institute under Contract DE-AC06-76RL0 1830. 171 I
.----_.__..~._--_..-----n.- . - _ . _ . - - - - - . . - - . - - - _ . . - - - - . . .- - - . -
performance demonstration requirements into Section XI; and 4) interaction with PISC-III to communicate final results of the SGTIP/SGGP.
SUMMARY
OF RESEARCH PROGRESS MONTE CARLO SIMULATION ANALYSES In FY 1987, a major NRC research program was concluded that determined the margin-to-failure of degraded steam generator tubes and the reliability of field-practice eddy-current (EC) in-service inspection equipment and procedures. One of the Steam Generator Tube Integrity Program / Steam 1 Generator Group Project (SGTIP/SGGP) objectives was to evaluate and compare l various sampling / inspection schemes for in-service inspection of steam l generator tubes. Information on EC inspection reliability was available from ~ four round robin examinations of the retired Surry 2A steam generator. Analysis of this information was used to select the input parameters for the sampling plan evaluations. An analytical evaluation of 20% and 40% l systematic sequential sampling plans was conducted using the ranges of probability of detection (P0D) and EC sizing reliability. Monte Carlo simulation analyses designed to supplement the analytical work and test key assumptions were performed. Results of these evaluations were described by Bowen et al. (1988) and Kurtz et al. (1988). During the past year, in a cooperative program with the Electric Power Re. search Institute (EPRI), additional Monte Carlo simulations were performed to expand on the results obtained during the SGTIP/SGGP and address a number of issues not considered in that program. First, additional sampling plans were considered. In the present study, 20% and 40% random initial sampling schemes were compared with the equivalent systematic scheme. Second, it was of interest to determine if a 33.3% random or systematic initial sampling scheme would give performance equivalent to the 40% systematic sampling scheme, assuming some clustering of degraded and defective tubes. Results from the SGTIP/SGGP indicated that 40% systematic sampling was almost as effective as 100% inspection. Third, in the earlier work, any of the initially sampled tubes giving nonzero EC indications triggered a second stage of inspection in the region surrounding the suspect tube. Second-stage inspection continued until a 2-tube wide " buffer" zone with no positive EC indications was observed that completely surrounded the suspect tube. In the present study, the effect of a 20% through-wall EC threshold on triggering of second-stage inspection was investigated. The last topic considered was a i comparison of current plant technical specification requirements with the sampling / inspection schemes described above. This comparison was made by performing additional simulation runs using a 3% random initial sample followed by second-stage inspection according to result categories C-1, C-2, and C-3. The simulations were accomplished by application of a previously developed computer program that simulates 100% inspection and the various ; sampling / inspection schemes mentioned above. A generator with the same number of tubes as the retired steam generator (3,388) was assumed, and six l different tube maps (each with a specified number and distribution of degraded and defective tubes) were considered. Assuming one flaw per tube, ! 172 i I i
- __ _ _ _ _ _ _ _ - _ _ - _ _ _ _ - _ _ _ _ = _ _ _ - _ _ _ _ _ - -
Table 1 gives brief descriptions of the flaw size distributions represented by the tube maps used in this study. Each map has a specified number of flaws in each through-wall size category. Map numbers 1, 6, 8, and 13 correspond to the tube maas presented in the report by Bowen et al. (1988), . except that all tubes witi flaws <20% are blanked. The modified maps are l distinguished here by an 'A' after the map number. Maps 20 and 21 were furnished by EPRI and represent actual 100% in-service inspection results from two different plants. Two different POD curves (Figure 1) and EC sizing models (Figure 2) taken from the previous work (Bowen et al. 1988) and one plugging limit (40%) were considered. For a given combination of tube map, P0D curve, EC sizing model and plugging limit, 25 independent applications of i all 14 sampling / inspection schemes were simulated. For each combination, i results from the 25 simulated inspections were summarized in tables and plots. , 1 TABLE 1. Tube Maps for Simulation Analysis j
)
Flaw Size Cateaory, % Description i Map 20-50 50-75 75-100 1A 5 5 5 3 Isolated Clusters 6A 40 40 10 10 Isolated Clusters j 8A 38 40 10 1 Large Cluster ! 13A 247 160 116 Predicted Surry (large cluster) 20 0 4 12 12 Isolated Defectives ' 21 0 1 3 3 Isolated Defectives Tables 2 and 3 give summary statistics from the simulation runs for all six maps using the best combination of POD curve (curve 5) and EC sizing model (Team V). Similar results were obtained for the other P0D curve /EC sizing I model combinations. The inspection plans in the tables are coded with three l digits. The first digit denotes sampling type--S for systematic and R for 1 random. The second digit denotes the initial sample size--2 for 20%, 3 for j 33.3%, and 4 for 40%. The final digit denotes the second stage inspection i threshold value--0 for 0% and 2 for 20%. The current plant technical ! specification requirement is denoted as TSR and 100% inspection is denoted as 100. Table 2 gives the sampling plan effectiveness, which is the average number of defective tubes (defined as a tube with degradation 275% through-wall) plugged divided by the total number of defective tubes in the tube map. ; Table 3 provides the sampling plan efficiency, which is the average ratio of the number of defective tubes plugged by a sample plan to the number of defective tubes plugged by 100% inspection. The efficiency results provide an idea of how well a sampling plan performs relative to the "best" possible (i.e.,100% inspection). In short, if 100% inspection cannot find a defective tube, subsampling cannot be expected to find it. There are several important points to note from the results shown in Tables 2 and 3. Both the effectiveness and efficiency are related to cluster size and the number of flawed tubes present. The results show that systematic 33.3% sampling was equivalent to 40% random and systematic plans, and gave 173 1 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - _ _ _ _ - _ _ _ _
1.0 c 08-o j Curve 5 2 e0 .6 - h i OO h Curve 4 0.4-1
.o 1 e
0- 0.2 - 1 1 0.0 - , , , O 20 40 60 80 100 ' True Flaw Depth, % FIGURE 1. P0D Curves 4 and 5 l 100 Y
. 80 -
0- Team V O i 3: 60 - l m E I
'O
{ 40 - E
.m U.S. Team Average I
o 20 - 1 w 0 i i , , 0 20 40 60 80 100 True Flaw Depth, % FIGURE 2. EC Sizing Models--U.S.. Team Average and Team V 174
l
~
TABLE 2. Inspection Plan Effectiveness for Best POD and EC Sizing Model i Inspection Map Plan lA 6A 8A 13A 20 21 i S20 0.42 0.77 0.96 0.94 0.19 0.19 S22 0.42 0.77 0.96 0.94 0.19 0.19 S30 0.58 0.87 0.96 0.94 0.31 0.31 S32 0.57 0.87 0.96 0.94 0.31 0.31 S40 0.62 0.92 0.96 0.94 0.38 0.39 S42 0.62 0.92 0.96 0.94 0.38 0.39 R20 0.35 0.68 0.96 0.94 0.19 0.23 R22 0.35 0.67 0.96 0.94 0.19 0.23 R30 0.61 0.84 0.96 0.94 0.29 0.27 R32 0.60 0.83 0.96 0.94 0.29 0.27 R40 0.58 0.84 0.96 0.94 0.36 0.56 R42 0.58 0.84 0.96 0.94 0.36 0.56 TSR 0.23 0.84 0.69 0.95 0.33 0.00 100 0.96 0.95 0.96 0.95 0.95 0.96 TABLE 3. Inspection Plan Efficiency for Best P0D and EC Sizing Model Inspection Map Plan 1A 6A 8A 13A 20 21 S20 0.44 0.81 1.00 0.99 0.20 0.19 S22 0.44 0.81 1.00 0.99 0.20 0.19 S30 0.60 0.92 1.00 0.99 0.33 0.31 S32 0.59 0.91 1.00 0.99 0.33 0.31 S40 0.64 0.97 1.00 0.99 0.40 0.40 S42 0.64 0.96 1.00 0.99 0.40 0.40 R20 0.36 0.71 1.00 0.99 0.21 0.23 R22 0.36 0.70 1.00 0.99 0.21 0.23 R30 0.63 0.88 1.00 0.99 0.30 0.28 R32 0.63 0.87 1.00 0.99 0.30 0.28 R40 0.60 0.89 1.00 0.99 0.39 0.57 R42 0.60 0.88 1.00 0.99 0.39 0.57 TSR 0.24 0.88 0.72 1.00 0.36 0.00 performance nearly equal to 100% inspection when tube degradation was clustered (note results for maps 6A, 8A, and 13A). Examination of the tables indicates that 33.3% and 40% inspection results are very close and that 100% inspection is distinct from the other sampling schemes for instances where defective tubes are isolated (note for maps 1A, 20, and 21, the , effectivenessand efficiency values are less than 0.65 for all inspection I plans considered except 100% inspection). l l 175 ,
p If a 90% efficiency' rating for an inspection plan is sufficient, then the - systematic 33.3% plan performs as well as 100% inspection for maps 6A, 8A, and 13A. Using .the same 90% efficiency criterion, only 100% inspection will suffice for maps 1A, 20,- and 21. It is important to recall from the evaluation by Bowen et al. (1988) that even with 100% inspection, most round robin teams that inspected the retired Surry 2A steam generatcr could not detect and plug more than 65% of the defective tubes present. Thus, if there are a large number of isolated defective tubes, even results from 100% inspection may not be adequate. The evaluation of an EC threshold for triggering second-stage inspection was inconclusive since many of the tubes that would have given rise to EC values less than 20% were removed from the tube maps studied. l 1 Simulated inspections following technical specification requirements yielded 1 very inconsistent performance. .In most cases, where degradation was sparse, this sampling schene tended to miss everything. Only when degradation was f copious did this scheme perform effectively. VALUE-IMPACT STUDIES Value-impact analyses of proposed Regulatory Guides 1.83, Rev. 2, and 1.121, Rev. 1, were initiated during the past year. These analyses shall examine the safety and. cost impacts of the revised regulatory guides compared to existing NRC regulations and current industry practice. The safety and cost will be evaluated in terms of factors such as public risk, occupational exposure, and cost to industiy and the NRC. "A Handbook for Value-Impact Assessment" (Heaberlin et al.1983) will be used to perform the assessment. Another basic reference to be used is " Generic Cost Estimates" (Science and Engineering Associates 1986). The work during FY 1988 involved comparing the proposed guides to existing NRC regulations and current industry practices to determine differences in requirements and practices related to steam generator tube inspection and plugging criteria. In addition, the base case risk assessment was completed. Existing risk analyses described in NUREG-0844 (U.S. NRC 1988) and NUREG-1150 (U.S. NRC 1987) were used as the basis for the risk analyses. The product of these efforts will be used to evaluate the safety and cost impacts of the proposed revisions. Other activities included collection and evaluation of information for characterization of risk and cost factors. Analyses will be conducted to quantify the public risk and occupational exposure expected to result from implementing the revised guides. The " base case" public risk and occupational exposure is evaluated before any changes , to the steam generator system or operational requirements are considered. The " adjusted case" of public risk and occupational exposure is evaluated for the proposed revised regulations. From these results, the change in public risk and occupational exposure caused by the revised regulations is determined. Another major element of the value-impact assessment is the determination of the cost impacts of the proposed revised regulations. This effort shall include an assessment of off-site property, on-site property, industry 176
-)
i implementation, industry operation, NRC development, NRC implementation,: and NRC operational cost impacts due to the proposed changes in the regulatory requirements. As a last step in the analysis, the above information will be combined and summarized and'a report prepared. Further, the regulatory efficiency of the proposed requirements will also be qualitatively assessed at this stage to 1 evaluate the expected improvements resulting from the current industry practices. ASME CODE ACTIVITIES The objective of this task is to participate in the ASME Section XI Task Gr_oup on Eddy-Current Examination for revision of Appendix IV and for development of performance demonstration qualification criteria. The goal is to seek adoption of improved EC inspection techniques and performance 'i demonstration qualification criteria by the ASME Code. Two PNL staff (R. J. Kurtz and R. H. Ferris) are members of the task group and have contributed towards the Appendix IV revision. To guide development of EC inspection qualification requirements, work performed at PNL for qualification of ultrasonic inspection of primary system pressure retaining components will be used. l_ PISC-III INTERACTION The program was also involved with the Program for Inspection of Steel Components (PISC) III to transfer information and technology developed by the SGTIP/SGGP. During FY 1988, an information meeting was held at PNL to communicate results of the SGTIP/SGGP to members of PISC-III. The purpose of ' this meeting was to give PISC-III members a comprehensive review of ;
'SGTIP/SGGP research findings to enable them to more effectively plan their steam generator research program.
FUTURE RESEARCH PLANS The focus of FY 1989 activities will be on conducting the detailed safety and cost assessments of the proposed regulatory actions. In this work, information from existing probabilistic rink analyses (PRAs) will be used to estimate the public risk. The need for new, or revised, steam generator tube rupture (SGTR) accident sequences and their frequencies from existing PRAs j will be evaluated based on current industry practices and steam generator design and operations. Information developed from the SGTIP/SGGP concerning l in-service inspection reliability will be incorporated into the study to account for uncertainties in degradation detection and sizing. The purpose of this evaluation is to ensure that the modeling and data used are applied j in a consistent and appropriate manner. Involvement with the ASME Code Task Group on Eddy-Current Examination and with PISC-III will be continued. 177 j l
-i i
REFERENCES Bowen, W. M., P. G. Heasier and R. B. White. 1989. Evaluation of Sampling Plans for in-Service Inspection of Steam Generator Tubes. NUREG/CR-5161, Vol .1, PNL 6462, Pacific Northwest Laboratory, Richland, Washington. Heaberlin, S. W. et al. 1983. A Handbook for Value-Impact Assessment. NUREG/CR-3568, Pacific Northwest Laboratory, Richland, Washington. Kurtz , R. J. , et al . 1988. " Steam Generator Tube Integrity Program / Steam Generator Group Project." In Compilation of Contract Research for the Materials Enaineerina Branch, Division of Enaineerina - Annual Report for FY 1987. NUREG-0975, Vol. 6, pp. 210-240. Science and Engineering Associates. 1986. Generic Cost Estimates. l NUREG/CR-4627, Science and Engineering Associates, Inc., Albuquerque, ! New Mexico. U.S. Nuclear Regulatory Commission (U.S. NRC). 1987. Reactor Risk Reference Document (Draft). NUREG-1150, Washington, D.C. U.S. Nuclear Regulatory Commission (U.S. NRC). 1988. NRC Intearated Program for the Resolution of Unresolved Safety Issues A-3, A-4, and A-5 Reaardina Steam Generator Tube Intearity. NUREG-0844, Washington, D.C. 173
( 1 Environmentally Assisted Cracking in Light Water Reactors-W. J. Shack, T. F. Kassner, P. S. Maiya, J. Y. Park,~ and W. E. Ruther . Materials and Components Technology Division Argonne National Laboratory Argonne, Illinois 60439 L, l '. December 1988 Contribution to Comoitation of Contract Research for the Materials Enaineerina Branch. Division of l Enaineerir,o Technoloav: Annual Reoort for FY 1988 Ito be oublished by the Office of Nuclear ! - Reaulatorv Research. U.S. Nuclear Reaulatorv Commission). 1
- Work supported by the Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission; FIN No. A2212; Project Manager: J. Muscara.
179
-{
Environmentally Assisted Cracking in Light Water Reactors W. J Shack, T. F. Kassner, P. S. Maiya, J. Y. Park, and W. E. Ruther i Materials and Components Technology Division l Argonne National Laboratory { Argonne, Illinois 60439 l l 1 Objective . Piping in light-water-reactor (LWR) power systems has been affected by several types-of environmental degradation. Intergranular stress corrosion cracking (IGSCC) of austenitic stainless steel (SS) piping in boiling-water reactors (BWRs) has required research, inspec-tion, and mitigation programs that will ultimately cost several billion dollars. As extended lifetimes are envisaged, other potential environmental degradation problems such as corro-sion fatigue must be considered. The objective of this program is to develop an independent capability for the assessment of environmentally assisted degradation in light water reactor (LWR) systems. 2 Scope Research during the past year focused on (1) stress corrosion cracking (SCC) of . austenitic stainless steels (SS), (2) fatigue of Type 316NG SS, and (3) SCC of ferritic steels used in reactor piping, pressure vessels, and steam generators. The suitability of proposed alternative materials and the potential effectiveness of proposed actions to solve or mitigate SCC by modifications in BWR water chemistry have been assessed. The studies have emphasized Type 304 SS, which was used for original construction for most U.S. BWRs, and Type 316NG SS, which has been widely used to replace Type 304 SS piping in both the U.S. and Japan. In addition, the modified Type 347 SS developed in Germany has also been examined. The effects of operating temperature and environment on the fatigue behaviar of Type 316NG SS are being studied. The data will be used to assess the degree of conservatism inherent in the ASME Code Section ill Fatigue Design Curves for this material. The environmental and material conditians that can produce SCC susceptibility in the ferritic steels commonly used for vessels and piping are being studied. Although it is known that these materials become susceptible to TGSCC in high-temperature water containing dissolved oxygen, the ranges of dissolved oxygen and impurity concentrations i that can lead to SCC in these materials remain ill defined. 3 Summary of Research Progress
' 3.1 Stress Corrosion Cracking of Austenitic Stainless Steels i Critical Strain for initiation of Stress Corrosion Cracking Most investigations of the SCC susceptibility of BWR piping materials involve fracture-mechanics and constant-extension-rate (CERT) tests. These tests are primarily measures of crack propagation rather than crack initiation. Although crack initiation, the precursor to crack growth, is of critical importance, it is difficult to define and study experimentally, j 180
l The present studies have focused on the determination of the strains required to produce j measurable cracks in CERT tests. Interrupted CERT tests have been used in previous studies to estimate the strains j required to initiate cracks (1,2]. However, the procedures required that initiation be defined by either large, easily observed cracks or extensive examination of specimens by scanning electron microscopy (SEM). A modified CERT specimen geometry was developed in which the strain in the specimen was localized, and hence, the region in which crack initiation was likely to occur was greatly reduced in size. A small-diameter (1.0 mm) through-hole was drilled in the center of the gage length (Fig.1). The tests were inter-rupted at relatively low plastic strains and then the specimens were cross sectioned by electrical discharge machining (EDM), as shown in Fig. 2, so that the entire region of crack initiation could be examined relatively easily by SEM. This approach provides a convenient technique for identifying crack initiation, i.e., detecting cracks less than 10-20 m in length. The plastic strains required to initiate cracks in a simulated BWR environment were determined for a Type 316NG SS (Ht P91576) and two heats of Type 304 SS (Hts 53319 and 30956). The composition of these materials is given in Table 1. The heat treatments for the di!Terent stainless steels, the degree of sensitization determined by the electro-chemical-potentiokinetic-reactivation (EPR) technique, and the CERT test parameters and results are summarized in Table 2. The heat treatments did not sensitize Type 316NG SS. The nominal plastic strain was determined a posterfort by measuring the change in length between two marks placed at the ends of the gage section. The local strain (cloc ) was then determined by multiplying the nominal strain (enom ) by a strain concentration factor, Both elastic-plastic finite-element calculations
- and Neuber's rule [4] suggest that the strain concentration factor is -8, and this value was used in the analysis of the results.
Scanning electron microscopy (SEM) photographs of the regions of strain concentration are shown in Figs. 3. Crack initiation occurs at local strains of approximately 2.0%, 3.0%, and 4.0% for Types 304 (sensitized), 316NG, and 304 (solution annealed) SS, respectively. Initiation of SCC occurs at relatively low plastic strains for all the materials, although Type 316NG and solution-annealed Type 304 SS tend to be slightly more resistant. The Type 316NG and solution-annealed Type 304 SS exhibited transgranular stress corrosion cracking (TGSCC); the sensitized material exhibited intergranular SCC (IGSCC). The effect of dissolved-oxygen level and the corresponding changes in electrochemical potential (ECP) on crack initiation was also studied. Tests were performed on solution-annealed Type 304 SS in water containing 0.005-O.25 ppm dissolved oxygen and 0.1 ppm sulfate, with corresponding ECPs ranging from -560 to 100 mV (SHE). Figure 4 shows the cracking observed in solution-annealed Type 304 SS at several different combinations of ECP and local plastic strain. Even at potentials below -400 mV (SHE), SCC occurred at strains of ~4%, i.e., about the same level as in the tests at -100 mV (SHE). l
- 11tpin Pai, Purdue University at Calumet Private communicauon to P. S. Malya. ANL, September 1987.
)
181
O 3 0 25132 3/16 R ,
.1 4 1 U2 %
se-5/8 9 r.... 1 U s .....i a f.' . . . ,
'fi ., ....?, ~- ~- y ONE THROUGH FOLE 0.040 DIA .250 DIA Figure 1. Geometnj of CERT Specimen Usedfor Crack initiallon Studies. Dimensions in
~ inches.
< \
/ /
N b - N h (MMDMW "WMMMM i I t \
! 1 1 N}))])}))))lll
! rl
} () l l l l l 1 l d)])]l])))))
l l l. Figure 2. Cross-Sectioning Scheme for CERT Specimen Shown in Fig.1. i 182 l
I V 1 l Table 1. Chemical Composition (wt.%) of Test Alloys 316NG 304 304 316NG 316NG Element (P91576)s (53319) (30956) (467958) (NDE-28) ] 0.015 0.06 0.06 0.02 0.014 , C
'1.63 1.69 1.54 1.53 1.77 Mn 0.02 P 0.02 0.024 0.019 0.023 0.01 0.013 0.007 0.008 0.002 S
0.42 0.59 0.48 0.65 0.52 S1 10.95 8.88 8.0 12.60 13.58
.Nt 16.42 18.33 18.99 17.29 17.79 Cr 2.59 Mo' 2. I 4 0.I4 0.44 2.51
-' O.06 0.19 0.16 0.11 OJ 0.068 0.029 0.1 0.0682 0.11 N 0.0006 D 0.002 0.0005 - 0.002 Iblarre llalance llalance Balance Fe 11alance alical or identification numlxrs are indicated in parentheses.
Table 2. Crack Initiation Observations on Typen 316NG and 304 SS In interrupted CElfrTests after Varlous Strains . at 289"C. i = 2 x 10-7 s-4. Environnrnt: 0.25 ppm O2 + 0.I ppm so, EPR. ECPb onom. Eno m, s toc.' Specimen % SCCid mV(SilE) MPa % Material Numbers C/cm2 Type 31GNG SS.1050*C/0.5h + 650T/24h 229 6.04 48.3 Yes 310NG PM9-54 0 78 147 1.43 11.4 Yes 316NG PM9-58 0 lil 144 1.36 10.9 Yes 316NG PM9-59 0 27 125 0.59 4.7 Yes 316NG PM9-52 0 24 0 12 124 0.38 3.0 Yes 316NG PM9 53 No PM9-50 0 45 107 0.22 1.8 316NG Type 304 SS,1050*C/0.Sh 0 94 244 5.79 46.3 Yes 304 5-57 12.8 Yes 0 31 171 1.60 304 5-85 8.0 Yes 0 109 144 1.00 304 5-82 0 49 Ii8 0.57 4.6 Yes 304 3-15 3.0 No 5-83 0 62 102 0.38 304 Type 304 SS. 8050T/0.5h + C00"C/24h* 94 225 5.80 46.3 Yes 304 5-34 24 Yes 70 144 1.40 11.2 304 5 24 Yes 120 128 0.29 2.3 304 3 11 2 1.5 No 5-70 24 83 102 0.19 304 a PM9XX: . IIT P91576 SXX: ilT 53319
' DOC llT 30956 b Electrochemical potentia 1.
cilased on an estimated strain concentration factor of 8. dStress corrosion crack initiation detected at 1000X magntlicatton.
'Except Specimen 311,1050T/0.5h + 700"C/0.25h 4 500aC/24h.
183
- v. ,
t 3 , 670 pm, i< . f, , ~ t r _, ,1epm i F , , , , 1 (o ) Ep= 5.8 % (b) Ep = l.27 % I l (c) Ep = 0.29 % (d) Ep = 0.19 % Figure 3. SEM Photomicrographs of the lloie Regions of CEllr Specimens of Sensitized 7}]pe 304 SS Sulgected to Crack Infilallon Tests. Nominal plastic strains in the specimen are Indicated: localized strain in the hole region is higher by a factor of ~8. Effect of Crevices on Stress Corrosion Cracking Susceptibility l A specimen geou.etry similar to that used for the initiation studies was used to study the effects of crevices on SCC. In this case, two holes, which penetrated only halfway through the specimen, were drilled in the gage length, and Type 304 SS pins (0.84 mm in dia) were inserted in the holes. The tests were conducted to failure, and fracture always l occurred at one of the drilled holes. The fracture surface was examined by SEM to determine the length of the largest crack and whether the failure occurred by SCC. Average l crack growth rates were estimated from the measurements of maximum crack length. 1 Two heats of Types 316NG SS (His P91576 and 467958) and one heat of Type 304 SS (Ht 53319) were studied. Type 304 SS was tested in both the sensitized and nonsensitized conditions. The heat treatments were similar to those used for the crack initiation studies (see Table 2) and produced no measurable sensitization in the Type 316NG SS. The results j of the tests on the crevice specimens are summarized in Table 3 along with corresponding I results obtained from standard, smooth CEffr specimens without the drilled holes 184
l l'l l l l ' l ' l ; TYPE 304 SS,SA (289'C) ! 200 0.005-0,25 ppm 02+ 0,1 ppm SO2 - - 10 0 - D g g y 0 0 g 0 - O - [ j 00
< -100 -. O _
G c-5 g -200 -- o o NO SCC INITIATION Q-O SCC INITIATION
$ -300 - 0 0
~
U go
-400 -
g O O
-500 - _
-600 i i i - t i i I l l l 1 0 2 4 6 8 10 12 14 16 18 LOCAL PLASTIC STRAIN, Ejo, (*/o)
' Figure 4. Effect of Open-Circuit ECP on Crack Initiation in Solution Annealed
'n)pe 304 SS.
Types 316NG and 304 SS are very sensitive to SCC under crevice conditions. Under non-crevice conditions HL 467958 of Type 316NG SS was resistant to TGSCC not only in high-purity water containing 0.2 ppm dissolved oxygen, but also in impurity environments, even at very slow strain rates (Table 4). liowever, under crevice conditions, even Ht , 467958 exhibited TGSCC, as is illustrated in Fig. 5. Similarly, smooth specimens of solution-annealed Type 304 SS exhibited ductile fail-ure, but the same material cracked transgranularly under crevice conditions (Fig. 6a) with an average crack grmvth rate of 1.39 x 10-9 m s-1 (Table 3). Unlike the solution-annealed Type 304 SS, sensitized Type 304 SS failed by IGSCC even under non-crevice conditions in high-purity water, but the presence of an artificial crevice (Fig. 6b) slightly aggravated IGSCC, as indicated by the enhanced crack growth rates (see last two lines of Table 3). The test results clearly demonstrate that SCC can occur under crevice conditions in environments where no SCC was observed under non-crevice conditions. Since the elec-trochemical conditions at the tip of a stress corrosion crack may be similar to those in a crevice of the same composition, these results may explain why some fracture-mechanics i tests show crack propagation for materials and environmental conditions under which no cracking was observed in CERT tests on a smooth specimen [41 185
l CREVICED SCC IN TYPE 316NG SS (HT467958) IOSO*C / t/2 h + 650'C/24 h d = 2 x 10-7 s-l O.25 ppm O2 Figure 5. Ferrite Morphology of Cast Materials frorn loop A (a) and (b) loop B (c), and loop C (d) Hot-Leg Manual isolation Valves. x;
/[ :c -
l E
)
l go 335 pm, m, m 3 . {o) SOLUTION ANNEALIO Figure 6.
.. (a) Morphologies of M23Co
'% Carbides on the Austenite- 1 Ferrite Boundaries: (b) Sigrna a;s Phase Formed on Slip Dands in Austentile of Cold-Irg Check Valves.
I
,1335 pm, ,670 pm, (b) SENSITIZED CREVICED SCC IN TYPE 304 SS (289'C)
(r . 2,30-7 s-1 q 0.25 ppm Og l l 186 l .
1 l l
)
i Crevice bent-beam SCC tests were performed on two heats of Type 316NG SS (Hi ' P91576 plate and HL NDE-28 pipe) in high-purity water with 0.2-0.3 ppm dissolved oxy- ) gen at 289 C. The plate specimens were shot peened to three different levels, represent- 1 Ing differents degrees of surface cold work. Some of the specimens were furnace heat j treated for 24 h at 500 or 600 C after the shot peening. The specimens from the pipe were l fabricated from the inside diameter near the weld heat-affected zone. The specimens and specimen holders were designed to provide a 0 to 0.25-mm crevice and 15% total strain. ; The specimens were tested for ~5000 h. Transgranular cracking was observed in one specimen from Ht P91576. Intergranular and transgranular cracking was observed in one l specimen from Ht NDE-28. In this case the crack actually propagated through the thick-ness of the specimen, it is possible that this particular crack may have initiated from a weld defect. Additional tests are planned to better assess the actual potential for cracking , induced by cold work. Heat-To-Heat Variations in Susceptibility of Types 316NG and 347 SS to Stress Corrosion Cracking CERT results for Type 316NG SS revealed significant heat-to-heat variations in susceptibility to SCC. Typical results are shown in Table 5. Susceptibility to TGSCC in chloride solutions depends on the concentration of Ni, Si, Be, and Cu [5,6). However, no simple correlation was found between resistance to crack growth and minor variations in chemical composition among these heats. The heats were also analyzed to determine whether the origin of heat-to-heat variations can be attributed to the presence of trace elements such as P. S. Cu, V, Ti, and Pb. Again, no conclation between cracking suscepti-bility and the concentration of residual elements was found. CERT tests have also been performed on five different heats of Type 347 SS. The results are summarized in Table 6. Except for HL 316642, all the heats of material were susceptible to TGSCC. However, susceptibility was observed only at at strain rates
<5 x 10-7 s-1, which is lower than that generally required to induce TGSCC in Type 316NG SS in similar environments. The strain rate at or below which SCC occurs is the (upper bound) critical strain rate. Above this critical strain rate, purely mechanical failure dominates. In terms of this measure of resistance to SCC Type 347 SS is some-what superior to Type 316NG SS.
l Effect of Heat Treatment on Susceptibility of Type 347 SS to Stress Corrosion Cracking Table 7 summarizes results on both base metal and weldment specimens from Ht 1 316642 of Type 347 SS. The base metal specimens, which are designated as K2 ... are extremely resistant to SCC after heat treatment at 1050 C/O.5 h followed by 650 C/24 h. The weldment specimens, which are designated as K2W ..., were tested both in the as-welded condition and after a subsequent 500 C/24-h heat treatment. No cracking was observed in the weld specimens at strain rates of either 1 x 10-6 or 5 x 10-7 s-1, although l the strains-to-failure are lower by approximately 50% for the weldments, as expected. However, at a strain rate of 2 x 10-7 s-1, cracking occurred in a weld specimen but not in the base metal specimen tested at this strain rate. The crack was located in the base metal relatively far from the weld. Additional tests were performed on material from Ht 316642 in the solution-annealed condition. These results also are given in Table 7. The material, which was extremely resistant to TGSCC in the aged condition, is susceptible to TGSCC in l I i 187
l I
. Table 5. ficat-to-Ifeat Variations in SCC Susceptibility of Type 316NG SSab t(. tr. Omas. Avg. Crack Gnsth Hate
- lleat No. h % MPa m. ri i i
k P91756 474.0 34.1 461 7.35 x 10-10 i 467958 577.0 41.6 501 0 08056 641.0 46.0 390 4.38 x 10-10 59076 398.1 28.7 487 3.83 x 10-30 ) N DF,-28 653 0 47.0 470 1.28 x 10-8 D440104 712.0 51.0 464 3.49 x 10-10 D472701 647.0 46.6 455 2.97 x 10-10 D442604 669.0 48,2 463 0 D450905 664.0 47.8 468 7.41 x 10-15 s All spectmens were heat treated at 1050*C/0.5 h + 650*C/24 h. bTests were conducted in water containing 0.25 ppm dissolved oxygen with 0.1 ppm sulfate at 289"C aruf's strain rate of 2 x 10-7 s-1 Table 6. IIcat-to-ficat Variations in SCC Susceptibility of Type 347 SS E. tr. o mas, a .,. SS Potential. Test No. Ileat No. s-I h MPa Failure m. r i mV(SIIE) 263 174100 1 x 10-6 65.5 432 Ductile 0 21 I 274 174100 5 x 10-7 ! !4.5 417 TGSCC 1.63 x 10-8 94 272 174100 2 x 10-7 301.5 448 1CSCC 1.10 x in 8 -8 275 174100 t x 10-7 !.74.5 451 1ESCC 7.58 x 10-10 55 301 170162 1 x 10-8 55.7 427 Ductile 0 64 305 170162 5 x 10-7 114.1 430 Ductile 0 91 310 170162 2 x 10-7 250.5 471 'IUSCC 5.50 x 10-10 22 ! 349 869962 1 x 10-e 94.5 460 Dact11e 0 22 350 869962 5 x 10-7 182.8 466 'IUSCC 3.80 x 10-10 34 348 869962 2 x 10-7 442.0 472 'IUSCC 2.97 x 10-80 80 364 316642' 1 x 10-s 100.0 438 Ductile 0 -1 365 316642 5 x 10-7 198.0 444 Ductile 0 3 367 316642 2 x 10-7 487.0 443 Ductile 0 -87 382 LPN I x 10-e 93.8 452 Ductile 0 86 j 90 380 LPN 5 x 10-7 174.8 459 TGSCC 1.22 x 10~9 377 LPN 2 x 10-7 530.5 460 1ESCC 9.83 x 10-30 -54 l l I eThe tests were performed in water with 0.25 ppm dissolved oxygen and 0.1 sulfate at 289*C. 188 l
the solution-annealed condition as well as the as-welded state. The results suggest that
~ heat treatment and thermomechanical history have a significant effect on the susceptibility of Type 347 SS to TGSCC.
To better understand the possible effects of heat treatment on susceptibility to SCC, the precipitate size and morphology of specimens from three heats of Type 347 SS (Ht . 174100,170162 and 869962) were examined by SEM. All the heat-treated specimens contain precipitates (Fig. 7) that were Nb-rich, as determined by energy-dispersive x-ray analysis. The precipitates are less than a micron in diameter. The crack growth rates , appear to decrease with an increase in precipitate size. The increased resistance to crack growth with an increase in precipitate size is consistent with the idea that precipitates offer resistance to slip and hence can retard the film rupture process that is generally involved in the crack growth process. . 1 Fracture-Mechanics Crack-Growth-Rate Tests l Fracture-mechanics crack-growth-rate tests have been perfonned on two 0.7T com-pact-tension specimens, one of CF3M cast SS (Specimen CTC24-2) and the other of Type 316NG SS (Ht D440104, Specimen 104-2). The ferrite level of the cast material was 5%, I as determined by magnetic permeability measurements. The specimens were fatigue pre-cracked in high-temperature water with 0.2-0.3 ppm oxygen and 0.1 ppm sulfate at 289 C under a cyclic load with R = 0.25. Crack growth tests were then performed with a cyclic load with R = 0.95 (sawtooth wave shape,12-s rise and 1-s fall time). Crack lengths were measured by de electric potential and by compliance measurement techniques. Measure-ments by the two techniques were in good agreement. The crack lengths as a function of time for the two specimens are shown in Figs. 8 and 9 for Type 316NG and CF3M SS, respectively. Because the fatigue crack growth rate was higher in the cast specimen dur-ing precracking, the initial stress intensity factor K under the R = 0.95 loading was
~21 MPa ml/2 (19 Ksiinl/2) in the Type 316NG SS specimen and ~22 MPa ml/2 (20 Ksiinl/2) in the CF3M SS specimen. As is shown in Fig. 8. for K<~22 MPa ml/2 (20 Kst inl/2), sustained crack growth did not occur in the Type 316NG SS specimen. The bursts of transient crack growth shown in the figure are associated with the compliance l measurements. To make the compliance measurement, the usual load history must be l interrupted by ~10 cycles of relatively low R loading (R = 0.4). Apparently this change in I loading history is sufficient to initiate crack growth, which however, is not sustained. For K
> ~20 Kst inl/2 (22 MPa ml/2), steady-state crack growth was achieved under R = 0.95 loading. Although compliance measurements were still made, no transient behavior was observed, presumably because the steady-state rate under these conditions was relatively high. This suggests that ~22 MPa ml/2 (20 Ksiinl/2) represents a threshold stress inten-sity for growth of SCC in this material under this loading condition. The stress intensity of the cast specimen was always 220 Ksiinl/2 (22 MPa ml/2) because of the longer precrack l produced by the initial fatigue loading. In this case steady-state crack growth occurred i
under the R = 0.95 loading, but there was a significant decrease in the crack growth rate under. constant load, R = 1. The effect is related to environmental crack growth and is not simply due to a change in mechanical fatigue crack growth; under hydrogen water chem-istry conditions, no measurable crack growth has been observed under R = 0.95 loading. 189
(a) HT 174100 AW + 600*C/24 h 4., . 1.10 x 10-'m.e d ,
.h
. g !
(b) HT 170162 AW+ 600*C/24 h 4 I.,* 6.60 x 16'*m.e*' , p, s, s. (c) HT 869962 , .,',. y-SA+650*C/24 h * ' 8., 2.97 x 16m.." - . L '. . - ' s ~ 4,4
;\ . o.
Figure 7. Precipitate Distributions in DJpe 347 SS and Corresponding Crack Growth Rates. Effects of Cuprous lon and Ethylene-diamine-tetraacetate Salts on Stress Corrosion Cracking In plants with admiralty brass, aluminum brass, or copper-nickel condenser tubes and/or feedwater heaters, copper is one of the metallic impurities present in reactor coolant water. Previous work indicated that the minimum concentration of Cu2+ required for severe IGSCC of sensitized Type 304 SS decreased from ~1.0 ppm at 289'C to -0.1 ppm at 150'C. These concentrations are higher than those typically encountered in the reactor coolant water in BWRs and in secondary system water of recirculating or once-through . steam generators in PWRs. However, since the thermodynamic stability regime of Cu+ ) increases relative to that of Cu2+ as the temperature of the water increases, Cu+ and Cu* are i the most likely copper species to be present at the potential-pH conditions in BWR and i PWRs, and tests were performed to determine the minimum concentration of Cu+ needed to promote SCC of sensitized Type 304 SS. I 190 l l
s i > , . , i i i i R-1 0.794 ^ R = 0.95
~
% ~
R - 0.25 - S & /* g 0.786 R = 0.95 / i a - gm N .e _ ; 0.778 - K<22 MPa m 1
/2 4--> K>22 MPa.m1/2 0.77 O 200 400 600 800 1000 1200 1400 1600 1800 2000 Test Time, h.
Figure 8. Crack Growth in a 'n.jpe 316NG SS Specimen under R = 0.95 (0.25 ap_d 1.0) Loading in Water with ~0.25 ppm Dissolved Oxygen and 0.1 ppm SO 4 I I 4 i i 0.87 - - R = 0.95 _ R=1 0.85 - j - d . j _ In .i\ E 0.83 -
. R = 0.25 -
U 0.81 - - R = 0.95 _ 0.79 O 400 800 1200 1600 2000 2400 Test Time, h FTgure 9.' Crack Growth in a CF3M SS Specimen under R = 0.95 (0.f5 and 1.0) Loading in Water with ~0.25 ppm Dissolved Oxygen and 0.1 ppm SO4 191
The feedwater chemistries ECP values, and test results are summarized in Table 8. As in the case of CuCl2, the time-to-failure of the specimens exhibits an abrupt change over a narrow Cu+ concentration range, as is shown in Fig.10, and the fracture mode changes from predominantly ductile fracture to IGSCC. The dependence of the crack growth rate . on Cu+ concentration is shown in Fig.11, For the most part, the data lie along a line of f slope ~1, which is consistent with the hypothesis that the cathodic-reduction partial l I process is Cu+ + e = Cu' and the heuristic result that the crack growth rate is proportional to 1/[e-] when the process is controlled by the rate of cathodic reduction [7,8]. Previous tests on Cu2+ indicated that the crack growth rate increased as the 1/2 power of the Cu2+ concentration, which is consistent with the reduction of Cu2+ to Cu', which consumes two electrons 19]. The temperature dependence of the crack growth rate of the steel in low-oxygen water I (<5 ppb) containing ~1-2 ppm Cu+ is shown in Fig.12, The broad maximum in the crack growth rate between ~170 and 250*C and the decrease in IGSCC susceptibility at both lower and higher temperatures is similar to the behavior observed in water containing 0.2 ppm dissolved oxygen with and without sulfate additions 110,11]. The crack growth rate of the steel at ~200 C in water with ~1-2 ppm Cu+ is higher by factors of 10 and 3 than . I It is in high-purity water with 0.2 ppm dissolved oxygen and in similarly oxypnated water with 1 ppm sulfate, respectively. The effect of pH on SCC behavior in Cu+ solutions was also investigated. Hydrochloric acid and NH40H additions were used to obtain pH2 5'c values ranging from 3.6 to 8.7. Tests were also conducted in solutions containing only hcl for comparison with the results in solutions containing Cu+. At a pH 25'c of ~4.7 (e.g., ~1.0 ppm hcl), the crack growth rates in water containing 0.1 and 1.0 ppm Cu+ are larger by factors of ~7 and 45, respectively, compared to those in solutions without Cu+. In a basic solution containing 0.1 ppm Cu+ and NH40H (pH25'c = 8.7), the effluent copper concentration was lower by a factor of 10 than in the feedwater and no SCC occurred. These results and similar results, obtained in water containing Cu 2+ and reported pre- l viously [11], demonstrate that cuprous and cupric ions at concentrations >0.1 ppm in high-temperature, low-oxygen water strongly promote IGSCC of sensitized stainless steel. These species undergo cathodic reduction on the surface of the steel, which couples with anodic f dissolution at the crack tip and leads to rapid advance of the stress corrosion crack. Since the crack growth rates are, in general, higher by a factor of 10 at a given strain rate and temperature than for other species at similar concentrations (e.g., dissolved oxygen or vari- l ous oxyanions), it appears that the kinetics of the cathodic reduction process for the I cuprous ions are faster than those associated with the other species, i As part of the investigation of the effects of organic contaminants on the SCC of sensi- . tized Type 304 SS in oxygenated water, CERr tests were performed in water containing 1 0.2 ppm dissolved oxygen and four EDTA salts at an anion concentration of 1 ppm. Although some transgranular cracking was observed, the crack growth rates, which are i given in Table 9, are very low compared to those observed in high-purity oxygenated water [13]. Since the dissolved-oxygen concentrations in the effluent were <5 ppb and the ECP values of the steel and the platinum and copper electrodes were also quite negative [-510 to -680 mV(SHE)], these substances apparently react with dissolved oxygen and thereby 192
...___.._3 7
l 1 1 160-A o _ y g 0 14 0 - TEMP {gQRL 6 a 289*C W 120- 0 289*C W o 150*c 3 100-g JjNSITifATION 2 LL- 80- , EPR = 2 C/cm O ' 60 STRAIN R ATE i = 1x10"'s"I p 40-q_ . FR A QJ1LR,LM_Qp,L 20- OPEN = D + TG CLOSED = D + IG 0- : . 0 0.01 0.1 1 10 Cu+ (ppm) Figure 10. Dependence of the Tune to Failure of Lightly Sensitized (EPR = 2 C/cm2) DJpe 304 SS Specimens on Cuprous Ion Concentration in CERT Experiments at 150 and 289*C. 16 -
- ./.
's ..
5 E h16' , stope =1 ) I . F ,,JQfSIIgATION 2 EPR = 2 C/cm e 16 8- JTRAIN R ATE M i = 1x10-6,-1 l _fR._AflVRE M0pL l h I O OPEN = D + TG CLOSED = D + IG 16' O.01 0.1 1 10 Cu+ (ppm) Figure 11. Dependence of the Crack Growth Rate of Lightly Sensitized (EPR = 2 C/cm2) DJpc 304 SS Specimens on Cuprous Ion Concentration in CERT Experiments at 200*C. 193
i34 E i i l I i 6 [ g[ . a2o too 250 20o
. no iso too T,*C
? : , a Cu' CONCENTR ATioN O 1.0 ppm
~
\s O 2.0 ppm g - \
g .[ \ SENSITIZATION 2 g -9 =
\. EPR = 2 C/cm ._ ;
o E \ E' - STRAIN RATE i = 1x10:d U10" 4 FRACTURE MODE. , K - O - OPEN = D + TG
- CLOSED = D + IG 10"' , , , , ,
1.6 1.8 2.0 2.2 2.4 2.6 1000/T ( K") i Figure 12. Efect of Temperature on the Crack Growth Rate of Lightly Sensitized (EPR = 2 C/cm 2) 'lype 304 SS in LounOxygen (<5 ppb) Feedwater Containing ~1-2 ppm Cuprous Ion as CuCL decrease susceptibility to IGSCC. This contrasts with the behartor Observed with for carboxylic (acetic, formic, lactic, and oxalic) and short-chain aliphate (propionic and butyric) acids, which also decreased the crack growth rates significance, but did not' decrease dissolved oxygen or ECP values [13]. Fracture-mechanics crack-growth-rate tests are in progress to confirm the relatively innocuous (or even potentially beneficial) effects of several of these organic acids. Initial results indicate that the addition of 1 ppm of
. propionic acid to the feedwater decreased the crack growth rate of a sensitized Type 304 SS specimen by an order of magnitude compared to that observed in high-purity water; however, the crack growth rate of a Type 316NG SS specimen was unaffected by the change.
3,2 Fatigue of Type 316NG SS In LWR primary piping, low-cycle fatigue is potentially a significant degradation mech- ' l anism which must be considered to justify extended operation of the plants [14]. Current fatigue design is based on the ASME Section III fatigue design curves, which do not explic-illy consider environmental effects. Instead, the design curves are obtained by introducing a factor of 2 on the strain range, or 20 on the cycles from the mean life curve, whichever is more conservative. It has been shown that the effect of the standard BWR environment on the fatigue life of A106-GrB, and A333-Gr6 steels can completely erode the "2 or 20" mar-gin in the Code design curve [15). Tests on Types 304 and 304L SS in the Dresden I reac- , tor [161 showed that the BWR environment had a significant effect both on sensitized mate- ] rials and on as-received materials, at least under some loading conditions, even in relatively !
-high plastic strain ranges. It is possible that environmental degradation would be even I h
more significant at lower strain ranges. Although the current data for Type 316NG SS 117] I 194 I 1 i
._ _ _ _ ____ a
are above the ASME Section 111 design curve, it appears that the environment eliminates a substantial portion of the safety margins built into the Code. The objective of the current work is to provide additional information on the effects of operating temperature and environment on the fatigue behavior of Type 316NG SS. Specimens were fabricated from a Type 31GNG SS 22-in.-diam pipe, manufactured by Sumitomo, with no additional heat treatment. Baseline in-air tests were performed under strain control with a triangular wave form and a strain rate of 5 x 10-3 s-1 with the same specimen design and loading systems that were to be used for the tests in the environment. The results of the in-air tests are summarized in Table 10. Figure 13 compares these results with the design curve for austenttic stainless steels in ASME Section til and the ASME mean data curve. The lives at 320 and 288 C are somewhat shorter (22-35%) than those observed at room temperature, but the differences are not large. The relatively small difference in temperature between the 320 C and the 288 C tests has little effect on fatigue life in air. However, in an aqueous environment, somewhat larger differences may be observed. In stress corrosion tests on sensitized Type 304 SS, susceptibility to cracking can decrease quite markedly as the temperature increases over this range. The results are in good agreement with the mean data curve for fairly short lives corre-sponding to plastic strain ranges of greater than 0.5%, but they fall below the mean data curve at longer lives. However, this does not necessarily indicate that the fatigue strength of the Type 316NG SS is less than that of Type 304 SS. The mean data curve is based almost entirely on tests with lives of less than 105 cycles (18); the portion of the curve for lower linear stress amplitudes was obtained by extrapolation. The results for Type 316NG SS are close to the actual data for Type 304 SS. The departure from the mean data curve occurs only for the portion of the curve that is extrapolated beyond the range of the sup-porting data. 3.3 Stress Corrosion Cracking of Ferritic Steels CER1' tests were continued on specimens fabricated from several ferritic steels (A333, A106, A155, A516, and A533B). The tests were perfonned in oxygenated water at 289 C without and with 0.01 and 0.1 ppm sulfate (as H 2SO 4) at strain rates of 1.0 x 10-6 and 2.5 x 10-7 s-1 The dissolved-oxygen concentrations were ~0.2-0.3 ppm. The results of the tests are summarized in Table 11. Except for one specimen of A533B (Ht A5401 Specimen W7-3), all materials showed transgranular cracking on the fracture surfaces. The average crack growth rates varied widely (1 x 10-10 to 3 x 10-8 m s-1), even for same heat of material. The addition of 0.1 ppm sulfate did not seem to have j an appreciable effect on crack grcwth rates in these materials. Presumably, crack tip impurity levels from the inclusions overwhelm the contributions from the bulk water ! chemistry. 1 ! As expected from related work on the fatigue crack growth of ferritic steels in reactor environments [19. 20), there appears to be a strong correlation between the sulfur content of the steel (particularly the presence of sulfide inclusions) and susceptibility to SCC. Figure 14 shows TGSCC on the fracture surface of Specimen 30C-1 (A106B Ht DP2-30), Figure 15 l I shows ductile fracture of Specimen W7-3 (A533B, Ht No. A5401). Micrographs of the cross 195 i
t .. Table 10. Cycles to Failure as a - Function of Strain Range for Strain-Controlled Fatigue Tests on Type - 316NO SS Test- Strain Hange Cycles . Number - % to Failure 25 %
'390 1 0.75 ' 25.736
'1391 1.0 -13,56I.
, 1392 0.5 60,741 1393 0.4 127,388 1394 1.5 4.649 1395 0.35 - 183,979-1396 0.75 ' 30,000 1397 0.30 347,991 1398 0.27 -666,000
-1399 0.25 ,
e 1400- , 0.25 ~ -1.775.000 320*C 1404 0.5 47.011 1405 'O.75. 20.425 l 1406 . 0.4 82.691 288'C 1407- 0.4 - 82,691 1408 0.75 21,548 1409 0.5 54,144
*The test failed at -1.900,000due'to an
, ' equipment analfuncuon.
-1000 , , , . , ,
o ' 22 C : o 320*C 1 .
. . 2ee C - $
- lw: 1
. o -
g ASME Mean Data
~
5 o i E 100 f
- c. 5 m 80 :
o
.l3 0 o o 1
ASME Design I 10 ' ' ' ' 103 104 105 106 107 108 Cycles to Failure Figure 13. Comparison of Current Results on Fatigue of Djpe 316NG SS in Air with the ASME Section III Design and Mean Data Curves. 196 i { L__----_-.____---
- 1
. k9 \ A I
i
'sp,
,oh_
y ,. , ,4. - @.: , A, n.. " ;- .e pu. , o g.
...'.g
- . c:a .
.C V ;y -
f r, '?^:;:T'Yhlk0
. : ... . ,Wa,* - e.
g.
- .g,'~
'#4
* ' ' ' 'M ' *
,. . . . . "*'g# . . ' . f,~ 3 g j *L ',' : &,.. .
,. - .;8 _ y ,.. ,s
.4
'. + 4 A, ,
h,~ J T _{t / . . ,j. . . Q.. r -* -
.p .y \
gj -
.s
- ty, , * , < ... '
~
1 O,. , < :. . @ - ??; m. 0.5 m m Rgure 14. 'lYansgranular Cracking on the FYacture Surface of A106B Ferritic Steel. n
.t
.e V- 4, y.
+ 7;;
i fje f . k(> e' -' m :v Q '
. :t j
3LR;t .qe I g3.-g ,
.Wp, ' 'j} j'- l?
'O f d..,,.. A;,-; A y ~^ Q .:j (Il '^
y I ;Y' 1mm _ i Figure 15. Fracture Surface of A533B Ferritic Steel Showing Only Ductile Failure. 197
i 1 h$?fbhkhhbkh' My3% V ('; E ' WY;bf%:y$c@k&G;$, f[5
% 4( ?
k,$!fB%Wf h. ifiR Y
?Nuh?p&g 4 nl;,gx& %@g wnr:w y nm g
-1 tryyJayP;g;p;.,pq}
. WD y gyp 4v.w,;;a m g, ?"8.g i nD b-y w p'%%
y h@3_ ,;. 9 gf. r ($kl$l#%gl9X.Nh$khN$hNh$$h? h;;% b MSD{k,pp M $. f,pM %ygy 2<g s"lM
- m
-1 d iQN N S $vg*;:dG.cIWM
- mun & $
- M.;y;c N '<$-
7.a%wWn o +c w WW
$ M$$$ hM-MM.%
a 9n . ~ g.: 9;k
@ah % %g a,4.- m edW w@$F wb p,.:: we - W ' m , ?,.;. % ,~y)4z ;? 4;wlw:qy?
.~n. e.. soR;vy:
rn : t fpa pgg;,dnu, su
- g. h wp < p%pi:x.;
v 3 .. .. w 4 ,. a u ; ~ eggg- m <3%y w . W: > ,l7l*. W.lR$a+s Y jM e ;we.m 3; qw pax;. sy i.f Qp
- $ W .: ,m
,9p4yw n q*
'.a%;p?N&WQ3$.$
w $n,4,.#.;wm@%er: m-
., me -
e,4yclg 16h m.f cmy y-1 e,.,,co w> .8 Nrm". @g; .g g..A nr yc *;'.,m;.
< ws Ji% < J i_ ,
nn ~
- N?:h k h ny.~q &w Axa;$qhpheQM
;;:n . a "W ti
%v wwym.m g%,
.bN 4g @W agwpag
;gggpgggpqhggdhga Q,AQ O q .m mg m 4 . w h n u .~.g ']
iMgure 16. Longitudinal Cross Section of A106B Ferritic Steel (Heat No. DP2-F30, Spect-rnen 30C-1) Showing Distribution ofinclusions. 4 a 4 ' # 4% 4 W s
.. .. a
'4
' N.r i;i *
- c' .
E.
- g ;..e 3
T 5 i 200gmk _ Figure 17. Longitudinal Cross Section of A533B Ferritic Steel (Heat No. A5401. Specimen W7-3) Showing Distribution ofinclusions. l 1 i 198
section of Specimens 30C-1 and W7-3, are shown in Figs.16 and 17, respectively. There are far fewer inclusions in the less susceptible A533B steel than in the more susceptible A106B material. The nonuniform distribution ofinclusions may also be responsible for the specimen-to-specimen variation in crack growth rates in same heat of material. 4 Future Research Plans l
- 1. Continue studies of crack growth behavior of alternate materials for BWR piping j systems such as Types 316NG and modified 347 SS and duplex austenitic-ferritic stainless steels to determine their inherent crack growth resistance and their ;
susceptibility to cracking during water chemistry upsets. I
- 2. Complete studies on the effect of surface cold work on the initiation of cracks in Types 316NG and modified 347 SS in BWR-like environments.
- 3. Continue study of environmental and loading history effects on the susceptibility of ferritic structural steels to stress corrosion cracking.
- 4. Infllate studies on the effect of corrosion potential, impurities, and strain rate on 1ASCC of stainless steel and the characterization of microstructural features responsible for susceptibility.
- 5. Continue environmental fatigue tests under simulated nominal BWR water chemistry conditions.
5 References
- 1. P. L. Andresen, " Crack Initiation in CERT Tests on Type 304 Stainless Steel in Pure Water," Corroston Houston 28(1), pp.153-58 (1982).
- 2. P, S. Malya, " Prediction of Environmental and Strain-Rate Effects on the Stress Corro-sion Cracking of Austenitic Stainless Steels " J. Pressure Vessel Technol.,109. pp.
116-123 (1987).
- 3. Not used.
- 4. W. E. Ruther, W. K. Soppet, and T. F. Kassner, in Enotronmentally Assisted Cracking in Light Water Reactors: Semiannual Report, April-September 1985, NUREG/CR-4667 Vol.1, ANL-86-31, pp. 27-54 me 1986).
- 5. A. J. Sedriks, Corroston of Stainless Steels, John Wiley & Sons, New York 1979. pp.
145-150.
- 6. J. Hochmann and J. Bourrat. " Stress Corrosion in Austenitic Stainless Steel," Mem.
Sci. Rev. Met., 60, 551-563 (1963).
- 7. W. E. Ruther, W. K. Soppet, and T. F. Kassner, in Light-Water-Reactor Safety Research Programs: Quarterly Progress Report, January-March 1985, NUREG/CR-4490 Vol. l.
ANL-85-75 Vol. I, pp. 25-42 (March 1986). 199
- 8. W. E. Ruther, W. K. Soppet, and T. F. Kassner, in Envimnmentally Assisted Cracking in Light Water Reactors: Semiannual Report, April-September 1985 NUREG/CR-4667 Vol. I, ANIc86-31, pp. 27-41 (June 1986). i l
- 9. W. E. Ruther W. K. Soppet, and T. F. Kassner, in Envimnmentally Assisted Cracking in l Light Water Reactors: Semiannual Report, October 1986-March 1987, NUREG/CR-4667 Vol. IV, ANIe87-41, pp. 35-47 (December 1987).
- 10. W. E. Ruther, W. K. Soppet, and T. F. Kassner, in Light-Water-Reactor Safety Materials Engineering Research Programs: Quarterly Progress Report, January-March 1984, NUREG/CR-3998 Vol. I, ANIc84-60, pp. 38-49 (September 1984).
- 11. W. E. Ruther, W. K. Soppet, and T. F. Kassner, "Effect of Temperature and Ionic Impurities at Very Low Concentrations on Stress Corrosion Cracking of Type 304 ,
Stainless Steel" Corrosion Houston 44, 791-799 (1988). '
- 12. W. E. Ruther, W. K. Soppet, and T. F. Kassner, in Environmentally Assisted Cracking in Light Water Reactors: Semiannual Report, October 1986-March 1987, NUREG/CR-4667 Vol. IV, ANIe87-41, pp. 47-53 (December 1987).
- 13. V. N. Shah and P. E. MacDonald, Residual Li f e Assessment of Major Light Water Reactor Components-Overview Vol.1. NUREG/CR-4731 EGG-2469 Vol.1 (June 1987).
- 14. H. S. Mehta, S. Ranganath, and D. Weinstein, Application of Environmental Fatigue Stress Rules to Carbon Steel Reactor Piping, EPRI NP-4644M, Vols. I and 2 (July 1986).
- 15. D, A. Hale, S. A. Wilson, E. Kiss, A. J. Gianuzzi, low Cycle Fatigue Evaluation ofPrimanj Piping Materials in a BWR Environment, GEAP-20244 (September 1977).
- 16. J. Alexander et al., Alternative Alloysfor BWR Pipe Applications, EPR1 NP-2671-LD, pp. 5-43 to 5-55 (October 1982).
- 17. Criteria of the ASME Boiler and Pressure Vessel Codefor Design by Analysis in Sections III and VIII, Division 2. American Society of Mechanical Engineers, New York 1969,
- 18. P. M. Scott, A. E. Truswell, and S. G. Bruce, " Corrosion Fatigue of Pressure Vessel Steels in PWR Environment-Influence of Steel Sulfur Content," Corrosion Houston 40, 350-357 (1984).
- 19. J. H. Bulloch, "The Effect of Sulfide Distribution and Morphology on Environmentally Assisted Cracking Behavior of Ferritic Reactor Pressure Vessel Materials " in 1 Proceedings of the Third International Symposium on Environmental Degradation of l Materials in Nuclear Power Systems-Water Reactors, G. J. Theus and J. R. Weeks. eds.,
American Nuclear Society, LaGrange Park, IL,1988. 200
LONG-TERM AGING EMBRITTLEMENT OF CAST DUPLEX l
. STAINLESS STEELS IN LWR SYSTEMS * ;. Principal Investigators
- 0. K. Chopra and H. M. Chung
]
I Materials and Components Technology Division ARGONNE NATIONAL LABORATORY 9700 South Cass Avenue Argonne, Illinois 60439 1 1 OBJECTIVES The primary objectives of' this program are (1) to investigate the sig - nificance of in-service embrittlement of cast duplex stainless steels under light water reactor-(LWR) operating conditions, and (2) to evaluate possible remedies to the embrittlement problem f or existing and future plants. l 4 SCOPE The scope includes the following: (1) characterize and correlate the .I microstructure of reactor-aged and long-term laboratory-aged material with y loss of: fracture toughness to identify.the mechanism of embrittlement, (2) conduct Charpy-impact, tensile, and J-R curve tests on these materials ! to characterize the loss of f racture toughness, and (3) evaluate the effects of key compositional and metallurgical variables on the kinetics and degree of embrittlement.
SUMMARY
OF RESEARCH PROGRESS Effort during the past year has focused on characterization of the micro- l structure and f racture morphology of reactor- and laboratory-aged material, .; recovery anneal of embrittled material, mechanical properties of cast ! stainless steel aged up to 30,000 h, and a preliminary assessment of low- ! temperature embrittlement under LWR conditions. I Microstructural Characterization l I Results of microstructural characterization by transmission electron microscopy (TEM), small-angle neutron scattering (SANS), and atom probe field ion microscopy ( APFIM) have been analyzed, and five metallurgical processes L were identified in association with aging embrittlement of the ferrite phase ofduplexstainlesssteel,{.g.,spinodaldecomposition,precipitationofa', Approximate time-temperature-transformation
- G, Y2 , and spherical M23 6' C '
l (TTT) diagrams for the onset of the five metallurgical processes have been
- l. also constructed on the basis of the microstructural characterization of the CF-8 heats.1 It was shown that the primary mechanism for the thermal aging embrittlement for 280-400'C is the spinodal decomposition of ferrite and M23 C6
- Work supported by the Of fice of Nuclear Regulatory Research, U. S. Nuclear Regulatory Commission; FIN No. A2243; Project Manager: J. Muscara.
201 l i l c :
carbide precipitation on ferrite-austenite boundary is the secondary mechanism for high-carbon CF-8 and -8M heats. An investigation was conducted to provide a mechanistic understanding of the activation energy for extrapolation of the accelerated aging data. Evidence was obtained which indicates that the kinetics of the spinodal decomposition of ferrite are strongly influenced by the synergistic effect of G-phase precipitation.. In a manner similar .to the acceleration of a' precipi-tation by Ni, Ni appears to accelerate the spinodal decompos1 tion. Since the precipitation of Ni-rich G-phase will deplete the ferrite matrix of Ni or because of Ni-Si clustering prior to C-phase precipitation, spinodal decompo-sition seems to be slower for materials and aging conditions in which G phase precipitation is significant. This effect of retarding the spinodal decompo-sition is expected to be pronounced for aging at ~400*C in which the G phase precipitation is fast, but is negligible for ~300'C aging. Consequently, the-activation energy of embrittlement will be significantly smaller than the activation energy of normal spinodal decomposition.in an Fe-Cr binary alloy (i.e., the activation energy of Cr diffusion), but instead it will be comparable to the activation energy of Cr diffusion minus the activation energy of the G phase precipitation. This model of activation energy being strongly influenced-by the synergistic effect of G-phase precipitation (in Ni-Si clustering) has been verified for 15 different heats of CF-3, -8, and -8M grades. G phase pre-cipitation characteristics were analyzed f or seven high-activation energy
,(40-55 kcal/ mole) and eight low-activation-energy (15-28 kcal/ mole) heats, which include specimens aged at G. Fischer Co., the KRB BWR, and at ANL. For all the high .and low-activation-energy heats, G phase precipitation for 300*C aging was minimal or absent. Forfall the high-activation-energy heats, the
-C phase precipitation for 400'C aging (up to 30,000 h) was either absent or limited to a minimal extent of heterogeneous precipitation on dislocations (which should not influence the kinetics of the ferrite matrix decomposi-tion). In contrast, all.the eight low-activation-energy heats were charac-terized by copious amounts of homogeneous G phase precipitation for 400*C j aging. When the M 2 '
significant for 400}C6 precipitation C aging in additionontothe the austenite-ferrite homogeneous G phase boundary is precipita-tion in ferrite, medium-level (30-38 kcal/ mole) activation energies appear to } resul t. Results of the microstructural characterization showed that the G-phase precipitation behavior was difficult to predict from bulk chemical compositions only, but it appears that the actual ferrite chemical composi-tions (which may be influenced by the f abrication and cooling processes) are important factors. Examination of the long-term-aged CF-8M heats (containing high Ni) revealed an unexpected decomposition process of austenite. The decomposition is referred to as "spinodal-like" decomposition (of austenite) in contrast to spinodal decomposition of ferrite. It is caused by segregation of Fe and N1 on a scale of 100-300 nm compared to Fe and Cr segregation in the spinodal decomposition of ferrite on a scale of 2-5 nm. Preliminary microhardness measurements indicate that local regions of austenite undergo the "spinodal-like" decomposition in high-nickel CF-8M heats and such local regions are significantly hardened by the decomposition. Overall effect on the material toughness is not well understood at present. 202
1
-Charpy-Impact-Tests j Impact tests were conducted. on standard Charpy V-notch specimens' machined
- from the aged and unaged materials according to ASTM specification E 23.
The. data for room temperature impact kineticsandextentofembrittlement.gngrgywereanalyzedtodeterminethe *
,.The Charpy-impact energy, KCV, -is expressed as KCV = K,+ SU.- tanh [(P -.0)/a]} , (1)'
is the minimum impact energy reached after where long-termP.isaging, the. aging parameter,-K,imum S. is half the max decrease in impact energy (i.e., half' the difference'between initial and. minimum impact energy),L0 is the log of the time to achieve 6 reduction in impact energy, and a is a shape factor repre-senting the time. between 'the start and end of the decrease'in. impact energy.
'e The time, t, at different aging temperatures is expressed by the Arrhenius relationship r 5 P
t= IO exp ()
- 673 j_'
where Q is the activation energy, R the gas constant, and T the absolute
~
l temperature. The aging. parameter, P, represents the degree of aging reached after 10F h at'400*C. The values of the constants in Eqs. (1) and (2) for various heats of cast stainless steel are given in Table I and the best fit curves for some of the heats are shown in Fig. 1. The Charpy-impact data are plotted as a function
.of the aging parameter in Fig. 2. The actual time and temperature of aging are shown on five separate axes below the figures. The service time, in years, at the hot-leg temperature of LkRs is shown at the top of-the figure.
The effect of aging temperature and time on the shifts in upper-shelf energy (USE) and transition temperature of the three grades of cast material have been presented earlier. The impact energy data were analyzed with the hyperbolic tangent function given by KCV = Ko + B{l + tanh [(T-C)/D]} , (3) where Ko is the lower-shelf energy, T-is the test temperature, B is half the distance between upper- and lower-shelf energy, C is the mid-shelf transition temperature in *C, and D is the half width of the transition region. The values of B, C, and D change with aging time whereas K o is assumed to be unaffected by aging. The results indicate that thermal aging decreases the impact energy and shifts the ductile-to-brittle transition curves to higher temperatures. However, different heats exhibit different degrees of embrittlement. In general, the low-carbon CF-3 grades of cast materials are the most resistant, and the molybdenum-containing CF-8M grades are least resistant to embrittle-ment. For all grades of cast materials, the extent of embrittlement increases with an increase in ferrite content. The significant results are summarized below. 203
300
^ CF-8 HEAT 60 s "E O M 0 3
p :o B 8 o - 0 200- 3 u R g-
^
-[G 0:21%
O 0 0 A 0 0 K O' 0 W D 450 C' O o 0 3 6 g 100~ O 400 C O a 350 C e h
'(
j2 0 320 C e UNAGED D D 0 , ,
,, s 8 8
'10 10' 10 1hs $f4 10 10 !
AGING TIME (Hours). 1 1 300-CF-8M HEAT 64 E E 3 -
$ 0 g O g g g 200 - e O g
^ S 5 6=28%
E W D 450 C D Z 100- O g 3 W o 400 C 0 A 350 #C h g 9 2 ( o 320 C e UNAGED O
$ h 6 0
0 , ., m -, 8 10' 10' 10 8 10 10' 1b8 10' AGING TIME (Hours) Fig. 1. Ef f ect of Aging Time and Temperature on the Room Temperature Impact Energy of CF-8 and CF-8M Cast Stainless Steel. l 204
0 0 0 0 0 0 2 - 1
- 8 - e _
4-2 . e-
'u, 0
8 T A a 3 5 '01
's-, ,0 E 1 H e 5L
)0 8
F C
.$3 og 0 4 '01 77,,'0 )
s r u s 0 '" r1 " '
. 1 o
a H c' 3 uC 3 e
'/
( Y 1 3 . E ( 0
$3 G 3 '0 ,'0 M ee 1'-
1 rt C l i g' g l
. 1 I E
E . T uS L' t 0 ] o 2 3 2 b 3 h O'
/M'!2 0
1
'7m,'0.
G IN as r e T 1 ' f1 L' :- 1 G pl m A A A' e
' s.
3 E 3 3 DlC' T M % 'C *C * C C C ElK' G b, 1 1!t 0 m T I 2 0 5 00 50 2 09 0 Al Nl 1 '
, 0 o
= 4 4 3 3 2 U 5. ' 1 o
6 17' R s 7
=
- O O a o_7 e14' e
- - - - - . 0 h t
0 0 0 0 0 , 0 00, 0 0 0 a 8 4 2 52, 4 2 1
=
4 83, n P o E >oX yegWzW y g2- IWp m" -
- g n
i g A l a m E >oM yemWzW h g 2- r e 0 0 h 0 0 0 0 0 0 2 1 e e 4 g T . e ' f o l 4 6 7. e k AT c E l l E 5 7' y, n et 0 l I H M gg o 8 0 ' 1 7,7 , 0
, 1 uc l a f p 0'I 3 nm 1 - I 8-
)
s r a I I F C p ' B gg 4
'0 '
7,)r
,'0 o s
u II 7
,1 .
e0' L 1 H 2 Y ( 1
, '3
(
^ p ' 7,5- M E g C a_-{ 3
- I '0 i 0
2 1" b g 0 E L' 1 7- s 0I 1 T F 3 O' G N T 3 M'
/' 2
'0
' 7,,*G I A L' A 1
- ,0 1
A E ' B h 3 1K' 1C' M I T %C 2 05 00C50C 8' *
- cC
- OG 16 '!
A1 1
' 7,'0 2 09 0 N1 0. ' '0 -m
= 4 4 3 3 2 U 1
' 6 14' 1
-- --O--- -O - a o_ V e13'
- 0 0 0 e 0 0 0 0 0 0
'm 000 p
0 0 e 6 4 2 = 0 529 4 2 1 4 332 p aE >oM yegWzW g {2- - d*3p y* L 1 l ,
,_ . - _ - _ _ _ - _ _ - _ _ - _ = _ _ _ _ _ - _ -_-__-_
'1
' - i; s
TABLE I.: Values (of the Constants Representing the Kinetics of Embrittlement for Cast Stainless Steel o Constants Activation K S 0 a . Energy. Heat- (J/$m)2 (yje,2) .(kcal/ mole)
- 47. 163.6 38.5 2.89 1.11 21.90-51- 153.0 31.0 3.06: 'O.58 44.06 56' 100.4 51.3 4.22 1.05 56.05 59L .91.2 63.5 3.26 1.55 46.94 60- ,64.7 63.9 2.82 0.63 47.51 L 63 140.2 58.4' 2.43 0.92 24.30 L 64 53.3 73.9 2.47 0.66 .34.06 65 54.3' 78.9 2.84 1.07- 36.39 66 94.9 74.9 2.72 1.73 30.23'
;(a) <High-carbon CF-8 stainless steels exhibit low lower-shelf energy and high mid-shelf transition temperature relative to the low-carbon CF-3 steels. - ' The~ lower impact energy for CF-8 steels is attributed .to M23 C6 carbides which form' at the: ferrite /austenite phase boundaries during produc-
' tion heat treatment of the casting.
(b). The mid-shelf transition temperature of unaged CF-8M steels is lower than that of unaged CF-8 steels. The difference is due to the absence of phase boundary carbides in the as-cast material. (c) Additional precipitation of phase boundary' carbides and/or growth of existing carbides occurs in the high-carbon' steels during aging at 450 or 400*C. Although phase boundary carbides are not present in the as-cast CF-8M material, they form during aging. (d) Phase boundary carbides have a strong influence on'the transition temperature, but have little or no effect on USE. The presence of carbides at the_ phase boundaries leads to phase boundary separation and/or initiation of cleavage of the ferrite at low temperatures, while ductile fracture at high temperatures, by void formation and growth, is not influenced by the phase boundary carbides. (e) The results suggest a " saturation eff ect" for USE after aging. The values of USE decrease significantly after aging for 2600 h at 400*C and do not. change for longer aging times. This behavior is observed for all grades of material. (f) Thermal aging leads to a decrease in the ferrite content of all grades of cast stainless steel, particularly after aging at 450*C. The decrease in ferrite content is significantly greater for CF-8M steels, Fig. 3, than for the other grades. The larger decrease in the ferrite content for , CF-8M steels and during 450'C aging may be attributed to the formation and/or l 206
.I i
. _ _ _ _ _ _ -_-__-_-__n
t ; ' l m ~! 1 q i o L( r g , , , , o ,o - , , , , o ui ,,,,,ui
, ,,,o,,
t
- CF-8M CAST SS
,30 -- g , 3 , ,
J;
't d
.U . .
I HEAT 64 -. , p j F 20 - _
;h U
y - D O~ 6 0 1 HEAT 63
.lo . _
AGING TEMP 'C O 45o - 0400 A 350
' ' ' ' ' ' ' "! ' ' ' ' ' i l "! ' ' l ' ' "! - ' 'tu O'
lo 102 go s 10" 10 5 g ; AGING TIME (h) Fig. 3. DecreaseLin Ferrite Content of Thermally Aged:CF-BM. Cast Stainless Steel. growth,of phase boundary carbides. The migration of phase boundaries is often
' observed in aged CF-8M steels.
(g) Charpy-impact data for CF-3 and some CF-8M steels indicate an
. . . inversion fn impact energy at temperatures in the transition region, i.e...the impact er mpy. of samples aged at 450*C is higher than that of. samples aged:for
'equivalenc times'at 400 or 350'C. The room temperature impact energies of l heats.52, 47, 51, and P2, aged for 10,000 h at 450*C, were 10'to 15% higher thancafter aging for 10,000 h at 400*C. This behavior was observed for cast materials which contain no' phase boundary carbides and have a very low mid-shelf ' transition temperature in the unaged condition, i.e., for most of the low-carbon.,CF-3 grades and some CF-8M grades of cast stainless steel.
.(h). The kinetics of embrittlement vary significantly f or the various heats of' cast stainless steel; the activation energies range between 20 and 56 kcal/ mole. The activation energy is lower for.the molybdenum-containing
'CF-8M steels or for steels with higher nickel content. The values obtained for activation energy in the pr those observed in the GF study,gsent study areenergies e.g., activation significantly higher between than 17 and 25 kcal/ mole.
(i) The shape of the impact energy vs aging time curves also varies considerably for the various heats, e.g., shape factor a varies between ~0.6 and 1.7. These results indicate that the kinetles ad extent of embrittlement are f- controlled by several mechanisms that depend on material parameters and aging L temperature. Data obtained at 450*C aging are not representative of reactor operating conditions; materials aged at 450*C show significant precipitation and growth of phase boundary carbides and a large decrease in ferrite content of the material. Consequently, extrapolation of the 450*C data to predict the extent of embrittlement at reactor temperatures may not be valid. 207
The results also indicate that the published correlations for estimating ( the kinetics of embrittlement are not accurate. The activation energy for the ]' [ processofembrittlementiscurregtlydescribedasafunctionofthechemical composition of the cast material, and is given by the equation Q(kcal/ mole) = -43.64 + 4.76(% Si) + 2.65(% Cr) + 3.44(% Mo). (3) i The activation energy from Eq. (3) ranges between 15 and 25 kcal/ mole for the various grades of cast stainless steels. These values are significantly lower than those observed in the present study. The predictions based on Eq. (3) l will be conservative for most heats of cast stainless steel. However, Eq. (3) may be nonconservative for some heats since it does not include the effects of other elements, such as nickel, carbon, and nitrogen, on the kinetics of embrittlement. Kinetics of Embrittlement The kinetics data from FRA,6 cy,5 and the present study were analyzed to develop a correlation between the activation energy for embrittlement and the chemical composition of the cast material. Initially, all major elements and carbon and nitrogen were included in the correlation. Elements with poor coefficients of correlation were then excluded. The analyses yielded two separate correlations: one for the Argonne and FRA data, given by Q(kcal/ mole) = 21.64 + 2.30 Cr - 1.94 Ni - 1.8 Mo
+ 4.92 Si - 29.40 Mn + 75.93 N (4) and the other for the GF data, given by Q(kcal/ mole) = -15.93 + 1.65 Cr - 1.30 Ni + 1.93 Mo
+ 4.10 Si + 10. 54 Mn + 71. 00 N. (5)
The observed and predicted activation energies for the two data sets are plotted in Fig. 4. The coefficients for chromium, nickel, silicon, and nitrogen show the same behavior in the two correlations, however, the I constants and the coefficients for molybdenum and manganese have opposite sign. The different effects of constituent elements in the two correlations are not clearly understood. Extent of Embrittlement The Charpy-impact data were analyzed to obtain a correlation between the , extent of embrittlement (i.e., minimum impact energy, K,, achieved after long-term aging) and material variables. The minimum impact energy is plotted in j Fig. 5 as a function of a material parameter consisting of the measured ferrite content (6 ,in %); chromium, molybdenum, silicon, carbon, and nitrogen content (in %) of the steel; and the mean ferrite spacing (E in um). The I results for all heats for which the material variables were known are shown in the figure. Thedatashowagoodcorrelationwiththematerialparamgter. The results indicate that the impact energy will be less than 50 J/cm (~30 ft lb) for those cast stainless steels for which the material parameter is greater than ~60. 208
I a 60 l I l 1 i j i l l l O. - O ANL - - GF ~ E AFRA q
- 40 -
00 cy- - - f 6 o 0 - a, - - - O A w g 20 - CF-8 0 6-- - CF-8 o-5 CF-3 OA CF-3 o CF-8M e A - - CF-BM e - n. 1 I l l I l l O 20 40 60 0 20 40 60 OBSERVED 0 (kcal/mol) Fig. 4. Observed and Predicted Activation Energy for Low-Temperature Embrittlement of Cast Stainless Steel. ANL: Argonne National Laboratory, FRA: Framatome (Ref. 6), and GF: Georg Fischer Com (Ref. 5).
@ l l I I _
i00 g MIN. RT IMPACT ENERGY OF AGED CAST SS - 150 , 4- ANL GF FRA EPRI BAT ;; 80 .-
$ CF-3 @ G 4 CF-8 O O O _ g
~
CF- 8M e 5 A Y > h,100 60 W e >-
>- ~
e .
- e W -
40 $W b gO o - g 50 -
, g E e -
20 g L
- o 3 . -
E { m I I I I O 150 200 250 O 50 10 0 sk (C + 0.4N) (Cr + Mo + Si) f (x 10-3) Fig. 5. Correlation between Minimum Room Temperature Impact Energy and Material Parameter for Aged Cast Stainless Steel. EPRI: Electric Power Research Institute (Ref. 9), BAT: Battelle Columbus Laboratory. 209 l _______________.______________._______.________o
l For cast stainless steels containing >10% ferrite, the mean ferrite spacing is in the range of 40 to 200 um; Cr + Mo + Si concentration is ~22% for CF-8 or CF-3 and ~24% for.CF-8M; and nitrogen content is typically 0.04%. Thus, for cast materials with 0.06% C and 100 pm ferrite spacing, the impact j energy will be below 50 J/cm2 when the ferrite content is above 20%. Cast materials with 10 or 15% ferrite can also reach very low impact strength when the ferrite spacing or the nitrogen content is high. For all cast stainless steels in service, the variables in the material parameter are readily available. The composition is known, ferrite content can be calculated f rom the composition or measured with a f errite scope, and the ferrite spacing can be determined by a surface replica technique. Thus, Fig. 5 can be used to estimate the extent of embrittlement for any cast i stainless steel component. J-R Curves The results of the J-R curve tests7 indicate that thermal aging decreases JIC and the tearing modulus of cast stainless steel at room temperature as well as at 290*C. The reduction in toughness is greater for materials aged at 400 or 450*C than for those aged at 350*C for similar lengths of time. The fracture toughness of the high-carbon CF-8 steels is lower than for the low-carbon CF-3 steels. The f racture toughness results are consistent with the Charpy-impact data, i.e., unaged and aged materials that show low impact strength also exhibit lower fracture toughness. The JIC values and Charpy V-notch impact energies obtained at room temperature and at 2 6. Results f rom the gtudies at Westinghouse (WH),g0*C are plotted in F Electric Power Research Institute (EPRI), and Framatome (FRA) are also shown. The dashed lines represent the lower-bound values. Thermal aging decreases the JIC values, and ! the relative reductions in JIC are similar to the relative decreases in impact energy. Figure 5 shows that the minimum impact energy can be as low as 20 J/cm for some heats of cast stainless steel. ThecgrrespondingJIC value f rom the lower-bound curve in Fig. 6a would be ~40 kJ/m . Thus, the correla-tions in Figs. 5 and 6 can be used to predict the minimum JIC values for any heat of cast stainless steel aged for a long time. The tearing modulus also decreases with thermal aging. The tearing modulus and JIC value for various heats and aging conditions are shown in Fig. 7. At both test temperatures, the tearing modulus decreases with a decrease in J 1C. Fracture toughness data for other aging conditions as well as other heats are being obtained to better establish the correlation between JIC, tearing modulus, and Charpy-impact energy. Preliminary Assessment of Embrittlement under LWR Conditions The embrittlement of any cast stainless steel component during reactor service can be estimated from Fig. 5 and Eq s. (1), (2), and (4). The material information needed for the assessment is the chemical composition, ferrite content and spacing, and initial impact strength of the cast material. When the material parameter is known, the minimum room temperature impact energy, K,, is determined f rom Fig. 5. The constant S is obtained from the difference between the initial and minimum values of impact energy, and the activation 210
( l i 1 I I IMPACT ENERGY,KCV (ft tb) .I-O 2000 , , , , CAST STAINLESS STEEL TESTED AT ROOM TEMP ; . 10000
- 1500 - .;
" - 8000 -[
- f
~
., "E ^ "I 6000 h -1000 - AoiNo TEMP (*C) ANL WH FRA EPRI KR8 g g _
' SSo A o 400 o o v _ 4000 -
500 - W .
- W /' 4 284 0 o
/j CF-8:OPEN SYMBOLS
- 2000 CF-3: SPLIT SYMBOLS
#j#j I i l. l CF-8M: CLOSED SYMBOLS 1 i O
O 200 250 -300 O 50 10 0 150 2 IMPACT ENERGY,KCV (J/cm )' IMPACT ENERGY,KCV (ft.lb) O 40 80 12 0 16 0 2000 , g i g i g i g i, CAST STAINLESS STEELS - 10000 TESTED AT 290*C AGING _ 1500 - TEMR ("C) ANL WH EPRI 8AT KRB 350 a 8000 400 0 V o -' 427 * "!!
"3
-6000
)1000 -
28 o
./ - C 9 CF-8:OPEN SYMBOLS CF-3: SPLIT SYMBOLS 3 j -4000 3 9
CF-BM: CLOSED V - - SYMBOLS - 500 - g "7, - -2000 O 'l I I I I I O O 50 100 15 0 200 250 300 lMPACT ENERGY,KCV (J/cm ) i' Fig. 6. Correlation between JIC and Impact Energy for Aged Cast Stainless Steel Tested at Room Temperature and at 290*C. WH: Westinghouse (Ref. 8). 211 j
. _ . . -- -___ _ ___ _ - _ _ _ _ _ A
i l I i awO , , , , TESTED AT ROOM TEMP 4 AGING TEMR (*C) A V'284 #
'o 350 u
~
o 400- o 9ED O D 450 g oo O g 1000 - O UNAGED , o O REANNEALED
-i;- #
l, pA g. - SYMBOLS o OPEN:CF-8
~
f0 0 0 0% SPLI T: CF-3 CLOSE: CF-BM
~ fIO . O.
I I l i O O 200 400 600 800 T (AVERAGE) 2000 , , , , l TESTED AT 290*C 4 AGING TEMR ('C)
- V '284 -
~
o 350 u 0 400 E O 450 O O UNAGED o
$ 1000 g 8
a g SYMBOLS 9g o OPEN:CF-8 O g SPL I T: CF- 3 - 4 g ,10 A CLOSE: CF-8M , O I l I I l I l O 200 400 600 800 T (AVERAGE) Fig. 7. Correlation between JIC and Tearing Modulus for Cast Stainless Steel Tested ar, Room Temperature and 290*C. l 212
-i l - _ _ _ _ _ _ _ _ _ _ - . ..
1
)
i energy, Q, is determined from Eq. (4). The average values of the constants 0 and a in Eq. (1) are 2.8 and 1.0, respectively. The decrease in impact energy during service at reactor temperature is determined from Eqs. (1) and (2). Examples of the predicted embrittlement bnhavior of heats susceptible to embrittlement (A and C) and typical heats (B and D) of CF-8M and CF-8 cast j stainless' steel'are shown in Fig. 8. The theoretical chemical composition and 1 the ferrite content and spacing of the heats are given in Table II. All 1 compositions are within ASTM specifications. The compositions of heats A and C are selected to give high ferrite content and fast kinetics of embrittlement i.e., low activation energy. The mean. ferrite spacing for most cast' stainless steels with >10% ferrite varies becween 40 and 200 pm. A large value of the ferrite spacing is selected for heats A and C to get a conservative estimate of the extent of embrittlement. TheresugtsshowthattheimpactenergyofheatsAandCwilldecreaseto below 40 J/cm (~20 ft*1b) after 4 or 5 yr of service at 320*C. Heats B and D, with lower ferrite content (15%), exhibit much less embrittlement, i.e., the impact energy will not decrease below 90 J/cm2 even after long-time service. The kinetics of embrittlement are also slower for these heats; the activation energy is >40 kcal/ mole, compared to 18 kcal/ mole for heats A and C. The results also show that the minimum impact energy is the important factor in estimating the embrittlement behavior. Slow kinetics of embrit-tiement, i.e., high activation energy, delay the decrease in impact strength. This behavior is seen for heat E, which has the same material parameters.as heat C, but the activation energy was arbitrarily assumed to be 45 kcal/ mole rather than the calculated value of 18 kcal/ mole to illustrate the effect of slower kinetics. With very slow kinetics, the impact energy 2 decreases to ~40 J/cm after 40 yr of service. The decrease in fracture toughness, i.e., values of JIC and tearing modulus, during reactor service can be estimated from the room temperature impact energy and Figs. 6 and 7. For example an' impact energy of 25 J/cm 2-corresponds to a lower bound vglue of 70 kJ/m for J IC. The tearing modulus can be estimated from Fig. 7; however, the correlations between JIC and tearing modulus are not well established at present. Recovery Anneal The low-temperature embrittlement of cast sta f byannealingat550*Cfor1handwaterquenching.gniesssteelcanbe. reversed The ductile-to-brittle transition curves for the KRB pump cover plate material, after reactor service and after reannealing for ~1 h at 550*C, show that the USE of the material increases from 247 to 330 J/cm2 af ter reannealing and the mid-shelf transition temperature decreases from 37 to -16*C. The annealed material was aged at 320, 350, and 400*C to investigate the reembrittling behavior. The results are shown in Fig. 9. The material reembrittles in a relatively short time. For example, aging for 100 h at 400*C or 3000 h at 320*C decreased the impact energytothevalueobservedafterreactorsergice. After 3000 h of aging at 400*C, the impact energy decreased to ~20 J/cm , a value close to the lower-shelf energy for the material. 213 I i l
)
f i l 1 i TABLE II. Theoretical Chemical Composition and Ferrite Morphology of Cast Stainless Steel used for Predicting the Extent of Embrittlement under LWR Conditions i Ferrite i Qxposition (wt %)' Content a Interceptb y . d Heat Grade C N bh Si Ni Cr Fb .(%) (um) (kcal/nole) (J/m2) i A G-8M 0.05 0.02 1.2 1.2 10.0 21.0 2. 6 28 180 18 20 B CF-&1 0.05 0.05 0.5 1.0 9.0 19.5 2.0 15 80 40 90 F C CF-8 0.04 0.02 1.3 0.5 8.4 21.0 0.4 24 200 18 - 3G D. CF 0.05 0.05 0. 5 1.0 8.5 20.5 0.4 15 80 45 90 E T-6 0.(4 0.02 ~1.3 0.5 8.4 21.0 0.4 24 200 45 30 a calculated frm chadcal cmp ition with Hull's equivalent factor. bk s e d e lues. cCalculated frm Fq. (4), value for heat E was arbitrarily assuned. d DetermLned frm Fig. 5. I I I Ilill l I I I lill l iIII4 :40 PREDICTED ROOM TEMR IMPACT ENERGY - 200 12 0
- N _
"E
\ 3
& K 10 0 :-
f h \ N - I k
~ \ \
\
\D --
80 k b E E z 100 60
\ z
\ \ 8 W
_ N _ W U g _
\ h 40 b- l 1
2 N- 2
- - CF-8 CF-8M N { - -- --- 20 l
l
- A !
O I I I I Illl I I I IIIII I I I I III O O.I I 10 10 0 SERVICE TIME AT 320*C (years) Fig. 8. Predicted Embrittlement Behavior of CF-8M and CF-8 Cast Stainless Steel. 214 i
4 i
~400 l- l -
200 N
'~O REEMBRITTLEMENT OF RECOVERY ANNEALED KR8 MATERIAL m l
00 g 0 -- 100f O o o 80 = g ? LOO n o - -- 60 7 l- -320 ! 's: 80 -- W
~
= 9, -
40 >
'N ~ 60 E
a: - 350*C - - W 40 - 400*C w1
.$ 20 3 b 20 - e REACTOR SERVICE $-
f - 10 g 2 2 "E l l - 1010' 10 2 10 3 10 4 AGING TIME (h) Fig. 9. Reembrittlement Behavior of Recovery Annealed ' KRB Pump Cover Plate Material. 1~ lJ It is not. clear at present whether this behavior is typical of all reannealed cast' stainless steels or is unique to this material.. Recovery annealed material f rom other heats and grades of cast stainless steel are being aged to better establish the reembrittlement behavior.- I Future Re' search Plans Microstructural characterization will be continued on long-term-aged materials to establish the mechanism of embrittlement and to investigate the
.' synergistic: effects of G phase precipitation and spinodal decomposition.
Select heats of cast' material are being tested in the embrittled and reannealed condition to determine the extent to which G-phase precipitation and spinoda1' decomposition influence the kinetics of embrittlement. Tests to determine ductile-to-brittle transition temperature, tensile properties, and J-R curves will be conducted on reactor-aged material and long-term-aged materials to characterize the fracture toughness. The results will be analyzed to validate the preliminary correlations for evaluating and predicting the toughness loss suf fered by cast stainless steel components during the service life of reactors. References
- 1. H. M. Chung and O. K. Chopra, in Properties of-Stainless Steels in Elevated Temperature Service, M. Prager, ed. , MPC-Vol. 26/PVP-Vol. 132, ASME, New York, p. 17 (1988).
- 2. H. M. Chung and O. K. Chopra, in Proc. Second Intl. Symp. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, September 9-12, 1985, Monterey, CA, American Nuclear Society, LaGrange Park, IL, pp. 287-292 (1986).
215
e i I l l
- 3. 0. K. Chopra and H. M. Chung, in Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, G. J. Theus and J. R. Weeks, eds., The Metallurgical Society, Warrendale, p. 737 (1988).
- 4. O. K. Chopra and H. M. Chung, in Properties of Stainless Steels in Elevated Temperature Service , M. Prager, ed. , MPC-Vol. 26/PVP-Vol. 132, ASME, New York, p. 79 (1988).-
- 5. A. Trautwein and W. Gyoel, " Influence of Long Time Aging of CF-8 and CF-8M Cast Steel at Temperatures Between 300 and 500*C on the Impact Toughness and the Structure Properties," Spectrum, Technische
, Mitteilungen aus dem+GF+Ronzern, No. 5, May 1981; Stainless Steel Castings , V. G. ' Behal and A. S. Helilli, eds. , ASTM STP 756, Philadelphia, p. 165 (1982).
- 6. G. .Slama, P. Petrequin, and T. bbgep, "Ef fect of Aging on Mechanical Properties of Austenitic Stainless Steel Castings and Welds," presented at SMIRT Post-Conf erence Seminar 6, Assuring Structural Integrity of Steel Reactor Pressure Boundary Components, August 29-30, 1983, Monterey CA.
- 7. A. L. Idser, Tensile and J-F Curve Characterization of Thermally Aged Cast Stainless Steels, NUREGICR-5024, MEA-2229 (September 1988).
- 8. E. I. Landerman and W. H. Bamford, in Ductility and Toughness Considera-tions in Elevated Temperature Service, MPC-8, ASME, New York, p. 99 (1988).
- 9. P. McConnell and J. W. Sheckherd, Fracture Toughness Characterization of Thermally Embrittled Cast Duplex Stainless Steel, EPRI Report NP-5439 (March 1987).
i l l l l l 216
L, A'ging Studies on Materials ( l from the Shippingport Reactor W. J. Shack, O. K. Chopra, and H. M. Chung Materials and Components Technology Division ' Argonne National Laboratory ) Argonne, Illinois 60439 - j 1 9 5 December 1988 Contribution to Compilation of Contract Research for the Materials Enoineerino Branch. Division of Enoineerino Technoloov: Annual Reoort for FY 1988 (to be oublished by the Office of Nuclear Reaulatorv Research. U.S. Nuclear Regulatorv Commission).
' Work supported by the Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission; FIN No. A2256-6; Project Manager: E. Woolridge.
217 l t _ _ - . _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _
i
.i Aging Studies on Materials from the Shippingport Reactor W. J. Shack, O. K. Chopra, and H. M. Chung l 1
Materials and Components Technology Division ] Argonne National Laboratory i Argonne, Illinois 60439 ! g 1 Objective l The objective of this program is to develop an understanding of the metallurgical phe-nomena that may occur in nuclear reactor structural materials as a consequence of extended service at operating temperatures within and outside of the radiation environ- .t ment and to assess the impact of these phenomena on structuralintegrity. Although many aging phenomena such as the embrittlement of cast stainless steel, low-temperature sensitization of austenitic stainless steels, and radiation embrittlement of pressure vessel steels have been studied in the laboratory, most of the studies have been based on simulation of actual reactor conditions. The Shippingport reactor offers a unique opportunity to validate and benchmark the laboratory studies, and thereby, provide a sound basis for evaluating the integrity of structural components near the end of the projected life - of a plant. Additional work will be undertaken to identify possible ncw mechanisms of component degradation. As the opportunity arises, additional materials will be procured from other reactors, and the extent of degradation in these materials will be evaluated and compared with' results from the Shippingport components. 2 Scope The work in the program is currently divided into four tasks: (1) Procurement of mteterials from Shippingport and other operating reactors: (2) Study of the aging degradation of cast-duplex stainless steel: (3) Irradiation embrittlement and SCC of irradiated components; and (4) Exploratory aging studies. The work during the current year has focussed on the procurement and characterization of samples from the neutron shield tank of the Shippingport reactor, and the characterization of aged cast stainless steel ! sampics obtained from the primary piping of the Shippingport reactor. Characterization of the neutron shield tank material includes mechanical testing (Charpy, tensile, and fracture toughness), microstructural characterization by transmission , electron microscopy (TEM) and small angle neutron scattering (SANS), and annealing and ) reirradiation studies. The sampling and study of the neutron shield tank has been carried ; out in cooperation with the US Department of Energy (DOE) through a program on plant life extension at Sandia National Laboratory. Characterization of the aged cast stainless steel le includes mechanical testing (Charpy, tensile, and fracture toughness) and microstructural l l- characterization by transmission electron microscopy (TEM) and atom probe field ion I microscopy (APFIM). I' 218 j
l j Future work may include sampling and characterization of materials from the pressure f vessel of the Shippingport reactor. f i 3 Summary of Technical Progress ) 3.1 Primary and Secondary-Coolant System Materials Samples of all materials from the primary- and secondary-coolant systems at Ship-pingport that were likely to have undergone metallurgical changes during aging have been obtained. They include primary coolant system valves, sections of a coolant pipe, a main steam pipe, a feedwater pipe, an instrument pipe, a purification pipe, and a fuel pool pipe; and two cast pump volutes. One volute had never seen service; the other was in use over the entire life of the plant. The components and materials obtained from the primary- and secondary-coolant systems are listed in Table 1. In most cases the valve-to-pipe weldments were obtained with the valves. The actual time at temperature for the cast components was ~13 years at ~281 C (538 F) for the hot-leg components and -264 C (507 F) for the cold-leg components. The components were at a hot stand-by condition >204*C (400 F) for an additional time of ~2 years. The chemical composition, hardness, and the amount and distribution of ferrite for the check valves and the spare pump volute are given in Table 2. All materials are CF-8-grade cast stainless steel. The hardness of the materials increases with an increase in ferrite content. For some of the materials an effect of orientation was observed; the hardness and ferrite content of the axial cross sections was higher than that of the circumferential cross sections. Table 1. Primary- and Secondary-Coolant System Materials Obtained from the Shippingport Reactor Cast Stainless Steel llot-l.eg Valves loops A. II. C Cold-txg Valves loops A.13. C Check Valves Loops A.13. C, D Pump Volutes loop A. spare Miscellaneous Materials Primary coolant piping and weldment (wrought) Feedwater pipe Main steam pipe i l t 4 219 I l
i All the valve materials have a radially oriented columnar grain structure. 'lypical exann ples of the grain structure for the check valves and manualisolation valves are shown to Figs. I and 2, respectively. Figure 1 also shows the weld between the valve and piping. All the valves also contained repair welds; an example is shown in Fig. 2. The pump volute has i a mixed grain structure, i.e., columnar and equiaxed, Fig. 3. The ferrite morphology of the different materials is shown in Figs. 4 and 5. The mate-rials contain lacy ferrite with a mean ferrite spacing, I, in the range of 150 to 300 pm. The check valve materials show that a significant amount of ferrite has been transfonned into austenite. Most of the ferrite /austenite phase boundaries have migrated. The original phase boundaries are decorated with carbides, which most likely have fonned during the produc-tion heat treatment of the material (Fig. Ga). The new phase boundaries are also decorated with carbides. Transfonned islands of austenite are also seen within the ferrite regions. llecent investigations suggest that embrittlement of the ferrite phase in cast duplex stainless steel may occur after 10 to 20 years at reactor operating temperatures. Such embrittlement could influence the mechanical response and integrity of pressure boundary components (1-31. Laboratory studies have identified a number of metallurgical processes that produce embrittlement in accelerated (higher temperature) aging tests. These pro-cesses include: (1) spinodal decomposition involving segregation of Fe, Cr, and N1; (2) nucleation and growth of the a' phase out of the spinodal structure: (3) precipitation of Ni- and SI-rich G phase; (4) growth of small spherical M23Co carbides; and (5) precipitation of n austenite within the ferrite [3]. The primary embrittling process appears to be the spinodal decomposition, with carbide precipitation on austenite-ferrite boundaries as the secondary embrittling process. The components from Shippmgport are I being examined to detennine if corresponding changes can be observed in reactor-aged materials. Examination of specimens from the valves by transmission electron microscopy (TEM) showed very finely scaled mottle images (~1-2 nm) in the ferrite, which are known to be characteristic of n' prime fonnation by spinodal decomposition. Atom probe field ion j microscopy is needed to provide conclusive evidence of spinodal decomposition. G phase { was also observed in the ferrite. The G-phase precipitates are ~5-15 nm in diameter, with ! a particle density of ~1021/m3, These obsen'ations are consistent with studies on low-temperature (-300 C), laboratory-aged materials. An unexpected microstructural feature, a new phase that precipitated on slip bands and stacking faults, was observed (Fig. Ub) in the , austenite phase. It has been tentatively identified as o phase, which precipitated on stack-ing faults in the austenite. Microhardness measurements, made directly on the ferrite and austenite, are summarized in Table 3. Tensile and Charpy V-notch impact specimens have been fabricated from materials containing >6% ferrite, but the tests have not yet been perfonned. The microhardness lev-cls observed in the ferrite are consistent with those observed in aged cast stainless steels that have shown significant loss of toughness. Because of the relatively low ferrite levels in the Shippingport components, the actual loss of toughness is expected to be relatively modest. Ilowever, the microstructural changes in the ferrite are consistent with a much larger toughness decrease in materials with higher ferrite levels and more continuous dis-tributions. 220
,l l h .
(: .
,4 , ,
_ ,3
, , 4 ,a j.,
.lNf +
f, g^4
. k
.. . ,J ' . 6, .. ,
, (#. _ . -
n .. R,,2$L .:;.
'4 l l .-
f 5.P.9 g* r
' * ' ,j g ,
, . i j .
9 .i 1 4 . t
?
i 4 ] Y- :
.. L ,
i
'Q.W' . . .
Figure 1. Microstructure along an Axial Section of the Isop A Check Valve. (a) Valve body and (b) weld. 221
l l l
.+ . s
.c
, .; l
- 4
.' - . . , i
+
3 !-
^~
;s e t. , y ._ . s
. f)'. . 1 .
k r.h
,. (*; f '; * '
s
.y g' -
/~ [. . g' i
_ s..f. , y - , y i
*[ 'r
* : f 'l
.4
-2 ,.
4 .
.. 9
,~
.h
./ . r . - f Figure 2. Microstructure along an Axial Section of the Loop B Manual Isolation Valve. A repair weld is also seen on the outer diameter of the value.
-5 .: h. ,
k .
,A.
_ fr -
.45 - +w ;
g* , g g -
~
~y . - .
.m {
. +'. ^w ;..'
...- , \
e % ,, . ' ' q ..y..* tu S, -_,, q, . o .c ~ Figure 3. Microstructure along an Axial Section of the Sparc Volute. l 222
1* - e y , r, ; l jf ' ' " '
*/
7
.]
, ; ,3 -
40, s / l
' fi' ' 93 // y ys .' a , .y, .
o.- 4,, r 6/ , ;, ,- gp #i' l
~
. L /..
d 6 - fa [g[g% g'. [ ~; ,
- s. x. W
.' ,P{,.
. \ s .m o= . , . ' .mo p=. ,
.t , T,,, _ # . bmim. i Figure 4. Ferrite Morphology of Cast Materials from (a) Loop A, (b) Loop D, and (c) Loop C Cold-Leg Check Values.
Table 3. Microhardness Measurements on Aged Components from Shippingport ANL Valve Loop Ferrite Austenite Designation ID Vickers liardness Vickers liardness CA7 0 5-117-1 A 243 165 CA4 0 5-117-1 A 252 167 CB7 0 5-117-2 B 230 142 CC4 0 5-117-3 C 234 159 MAI 05-116-2 A - - MA9 05-116-2 A 310 176 MU2 05-116-4 B 286 178 VR Volute - 260 199 l 3.2 Neutron Shield Tank Material The embrittlement suffered by the HFIR vessel at Oak Ridge National Laboratory has ! raised the issue of whether low-temperature, low-flux irradiation can produce an unex-l pectedly high degree of embrittlement of reactor support structures. To help resolve this question, samples were obtained from the neutron shield tank of the Shippingport reactor. l This sampling effort was sponsored jointly by the NRC and the DOE Plant Life Extension 1 Program (PLEX) at Sandia National Laboratory. The actual sampling was performed by Pacific Northwest Laboratory under subcontract to Argonne and Sandia. The samples are approximately 6 in. In diameter and were obtained from the inner wall of the shield tank along with the corresponding samples from the very slightly irradiated outer wall. The I locations of the samples from the outer and inner shell of the shield tank are shown in Figs. I 7 and 8. The material from the Shippingport neutron shield tank is an A212 Gr-B steel l 223
. _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ ____-_-__a
.y g- .,
.f.
', [ T
[ . > O $[ 2
/ V. % e sfc{r9U$h - i-
,r.
A s.- %W . b;p, t < , y
- . .g ;4 ;.
>u. 3- 1,.
%pd [: Y j, , , , , a 8y
.s- - ~ w[
.mw m . ,,
r ;: ;s; ,,
,r a.
?q . , _ 'I i
?f0 :yblQ. - . - .
;[l:
.z.o ;
g gg #- e.
*' t.,
g,,
, jw
. ,.u.
..9 , , . -
,s. j,,okgk.
. w:. m p.
,<.;z z .
~ m av 9 , p. 8
, yf.yle
+ '
,f ) .g L y ,
( W.- ., o'.< m, m'a,og e e. . . .g,--.. .-
- ,n#s. ,m . d. ., /a<u . Wp lL.
m l. ' # .a 4,
,' .M y
' N .h N ' ,,
) ,
- A # h wg] h -
As v.R ny o.a gf.,. - 4 ww%_; ~,:, p .. gg A. ..%j m% . w ,., 4.y.gmyga - g
- a. , . ..
7%,Q g d - - i e . f h, 7 4h;g g h{,,q) p ,
% g %gl'* WfMqy , &c;:,h,y).g 49g gy(;psg s MdOM o
, ' N ,,NGCM 4 k s n -
- mm.w " $ W< W, 200 m, E,w w:p*1 AwtN MA disNc ,,200 +/d4(#' Fm, '
9- cumsRa . , immsw dj gata , ,
.a m e W v W WMu,4 s gm:3, A;mNgmNngm N tM Q %W \~N' n '
.US # -
vyE WkM . b* h&q p$ p$ W:1p&
$996 6uu:p,,D.Y,r.hfM@M s J
?MQwfV eQ W . &sQ, a WE
&- M. onyi p b[pps,l@wgk< f GQH
$$M[NNMC@ w%gu h-N fh%',<h,m @ % M] N i$n4 N ( Y J [Nq
$g.;yah.%wsm.$i4%g%w%: qv
.m ' v y, e s 4 p.
g $,,,)f .,a
- k. w! (! q%m. y $.. s s]m n p) h Q~ h ( m% m -n. ; ' .. .s l
. . +w :q.(:we -M.g , y~
O ,. ys~y w%my?>s: e , c ;me .wm., 3Mmm;p':m.m.g:y;g,. y m;4w v w vSv p , ~, . ,w Qs u ~s
!NAf s 3ggdy%,
.d, .. D . A ,,,,We /
g g g% r. $ ,.' 9gv< e g.c , o
", a =,.3 40e m .%py,e'% wen app,g.pN g;W ..
7 m a ,' ,
)
.n. +
ce
$g, .
..: a
[Ih .a , I-h.
- 4. g
,m m Q QL,n. w eQ.,QWknf,
. ?Q M, n fQ}M ig
- 6. & >
c lp4Q> , f "l$y., r /a ~p= g M. - 7 ow.g . u l pp.m .,f,y@WM%w#mp,itzgxm o c . L i V~ b Q 200 pm
- n. ?Yll5%i%$Ml4
,i # 200pm'l . /1
- a?WI Figure 5. Ferrite Morphology of Cast Materials from loop A (a) and (b) loop B (c), and loop C (d) Hot-Leg Manual Isolation Valves.
224 i
- i k
_ . _ _ _ . . _ _ _ . _ _ _ _ _ . . _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ . _ _ _ _ _ _ _ _ __ d
., F E RR 6 T E k< 1 m y _-4 a
l h - e e . AUSTENITE .k < f ( . t l r
'e Figure G.
(a) Morphologies of M23Ce Carbides on the Austenite-Ferrite Boundaries; (13) Sigma Phase Formed on Slip Bands in Austenitte of Cold-Leg Check Valves. f I' 225 i
l l similar to the material used for the HFIR reactor. The shield tank was fabricated by the Pittsburgh and Des Moines Steel Company. As indicated in Figs. 7 and 8, the vessel is com. l posed of a number of plates welded together. The available records are not adequate to determine whether all the plates are from the same heat of steel. Ilowever preliminary chemical analyses of samples taken from each of the plates strongly suggest that both the inner and outer shells of the vessel were fabricated from a single heat. Additional analyses and metallographic studies are being performed to confirm this conclusion. A typical chemical analysis of specimens taken from the vessel is given in Table 4. North South Weld i Weld Deval 300" O' & 1ma 2?S* b h b h b h 699'
, i j i 14 -D-9-$~ 15 -p Bottom of I 3 -$- - 692* 6"
, Lifting skirt --. / ,
8-$- l 2 -$- - 690'6-
,1-$- 13 4- - 690'
, l 11 -$- I 5-p ,
- 689' I - 687 6" gg Horizontal Weld 6 !
- 684' O
GE1 - O m2+ O =+ O e+ ~ Figure 7. Sample Locations on the Outer Shell of the Shippingport Shield Tank. Also shown are locations of additional samples obtained by the General E;tectric Company, which is the contractor for the Shippingport Decommissioning Pro-ject. 1 Charpy-impact, tensile, and Jn curve tests will be conducted on material in two orientations (rolling and transverse direction) and at three fluence levels. Tests will be performed on material from the outer shell to obtain baseline mechanical property data. Charpy-impact and tensile tests will also be conducted on weld metal specimens from the inner and outer shells to characterize the embrittlement behavior of welds at two fluence levels. A cutting diagram for a typical base metal sample is shown in Fig. 9. Some preliminary Charpy tests have already been performed. The results from these tests are summarized in Table 5. Estimates of the fluence and flux levels for the most highly irradiated samples are given in Table 6. The available results for the shield tank are summarized in Fig.10. along with data Data from the HFIR reactor
- and data on another heat of unirradiated A212 Gr-D steel (5).
obtained on A212 Gr-B steel at ORNL in the 1960s show that this material exhibits a very wide range of transition temperatures and upper shelf toughness for nominally similar compositions and heat treatments [6). The low toughness observed at room temperature for the specimens from the outer shell indicates that, even in the unirradiated condition, j the transition temperature of this material is on the high end of the distribution observed . in the early ORNL tests. The data at 55 C (131 F) demonstrate that irradiation has lowered :
*S. Iskander. ORNL. Private communicauon to W. J. Shack. ANL. September 1988.
226
p North South Weld Weld aton 300' 0' 45* 180' 225, l h b # v \ l- 699' i I l l l 9 15 4F 3 -1F 14
- 602" 6" 8HF l
i 2-h i 13 1
- 690 6"
--------g------L-------------~~t--r-------$----------- 600' 11 HF I
5 -$- 6 I I
- 689' i
Horizontal Weld 12 6-h - 687'6' i D D R I R 3 Figure 8. Sample Locations on the Inner Shell of the Shippingport Shield Tank. Table 4. Typical Composition of A212 Gr-B Plate from the Neutron Shield Tank Element level (wt.%) C 0.23 Mn 0.78 P O.02 S 0.03 Si 0.20 Cb 0.05 Ni 0.04 Cr 0.04 ' O 0.01 l l N 0.004 Ta 0.03 }. ! Mo, Ca. Ti Al <0.01 B. Sc. Sn <0.01 I V,Zn.Zr <0.005 f l the toughness of the material. The increases in the NDT temperature obtained from the HFIR surveillance specimens and corresponding changes in A212 Gr-B steel irradiated in a l j ! high flux research reactor are shown in Fig.11 [4]. The current data from the shield tank permits only rough estimates of the changes in transition temperature under irradiation, and there is considerable uncertainty in the actual irradiation conditions. Estimates of the l change in transition temperature based on the limited testing to date, are shown in Fig.10 for comparison with the data from the IIFIR reactor. Although large uncertainties exist at present, these preliminary results suggest that the changes in transition temperature are l not as severe as might be expected on the basis of the changes observed in HFIR. However, the actual value of the transition temperature is high, and the toughness at service temperature is low, even when compared with the HFIR data. i 227 1 1
Rolling Direction l
*
- 1 d
First Cut v LT 1.20 LT U V v o LT LT(or LS) 0 56 O ' o O O "
\ Third Cut
% r \
> O ~
O .) % ^ o O o o l 8
)
. 1 ao O(r a O 1 p Second Cut j ENgurc 9. Cutting Diagram for Base Metal Samples from the Inner Shell of the Neutron Shteld Tank.
4 Future Research Plans An attempt is being made to organize a cooperative effort to obtain samples from the Shippingport pressure vessel. Expressions of interest in the endeavor have been obtained from DOE, EPRI, the D&W Owners Group, Yankee Atomic, Electricit6 de France, and Bettis Atomic Power Laboratory. Although the materials and operating conditions of the Ship-pingport reactor are not exactly prototypical of modern PWRs, valuable information can be obtained on several important issues such as the throughwall variation of toughness, com-parisons among different types of specimens (CT, Charpy, EPRI miniature specimen designs) used in surveillance studies, and activation and radioactive product studies for decommissioning. The actual sampling operation would be performed after shipment of the pressure vessel to the Hanford site in Washington, which is currently scheduled for the second quarter of FY 1989. Characterization and mechanical testing of the material from the neutron shield tank is the highest priority task during FY 1989. Activation analyses are currently in progress to try to get a batter understanding of the actual irradiation conditions. Microstructural analy-ses will be perfonned to determine the actual mechanisms of embrittlement. 228
b 1 i l 1 1 Table 5. Charpy Toughness Values of Samples frorn the Inner and Outer IIalves of the ' Shteki Tank inner and Outer Shells j Temperature Cp inner Italf Cy Outer llalf Sample Number /Shell location / Plate aC 'F joules /cm2 ft Ib joules /cm2 ft. Ib 9/l - Top North 25 77 8.1 - 4.7 11.3 6.52 9/1 55 131 39.0 22.5 48.5 28.1 8/l 25 77 10.3 6.0 12.8 7.4 8/I " 55 131 43.5 24.9 53.5 31 l 3/1 Top South 25 77 10.9 6.3 14.5 8.1 3/l " 55 131 45.5 26.1 61.0 35.3 2/1 25 77 13.5 7.8 11.6 6.7 i 2/l " 55 131 52.8 30.6 68.6 39.7 a 9/O Top 25 77 43.2 25.0 46.4 26.9 j 9/O 55 131 85.7 49.5 87.7 50.8 3/O 25 77 35.1 20.3 48.0 27.8 3/O 55 131 94.8 54.9 98.3 56.9 2/O 25 77 29.4 17.0 48.0 27.8 2/O 55 131 114.3 66.2 92.0 53.3 6/O tiottom 25 77 24.0 13.9 42.5 24.6 6/O 55 131 89.1 51.6 93.1 53,9 12/O 25 77 - - 49.4 28.6 12/O 55 131 84.0 48.6 96.5 55.0 1 i l i Table 6. Estimated Fluence Levels for the Inner Shell of the Shield j Tank , Sample Fluence Flux (neutrons /cm2 , > IMcV) (neutrons /cm2. s. > J McV) 9 8 x 1017 4 x 100 3 8 x 1017 4 x 100 2 5 x 1017 3 x 109 8 5 x 1017 3 x 109 I l 229 1 l
14.. . , ('
.e.
b- 3 $: $e E 8C 'sm D. o .g
-6. - o (p .
{
- T. T O E3T2 E- E 3
.I CD :I LI (D C 2 LU $
y'
'0 'o . < - . O o E
.i i i l iiijiiil i i i
.l i5 i
- j.e
+ E
- ..%k o
te _. O -e - W e
-Q
.e
=
B
~
~
D # O 000 'g
~
, .g
_ o..o .9
~
d ooo o oo M O-
- b. g e +
3m . 4
< + + 5 t-k@ 4 o
e a E e gmi F de. su E $R 5 S- gg - N ;2 0
~
00 oo o. s-QS gd B$ a ao .g .l
.5& *5 u a a be
@3 -
ee gQ oc a .@t eE
- n. x i , i l i i il i i i l i i i I i i i o o o o o o c; o CD O T N w e
qi 1; 'A6;eug g L 230 l ': \
h t id wna k nB a T Gr-d 2 l ie1 9 h 2 _ l 1 SA i 0 nR t 1 oR r H i i tu ee i i Nh t t rfo i o ps gn i
) ino i s pt
- ipia
- 8 2 h da l
1 m Sr r
~ t i 0 c eI 1 /
n h r R i ( t t o R R i r c oa RO O i v fe N i e t sR N I , M ft i I D i h s D E 1 SeT E T i e r x T IA k e uu E A D dn c C a D t N I D A t ea 7 n e rh g A A R ad t l e L R R ml i '0 lu pi L R I ie t shi t 1 F mH e I I L n Tdn i E L L E S i o V L E f o or r na R E H ef i t u o ia U H S gs i e it t S S R ni t f s N sa R R C ah 2-i nD I I RS a e r F F G m H H E / c i Tc nB . n B B B d a -r l 2 2 2 8 tel i G 1 1 1 o 6 1 ae 2 A A A 2 2 t x l 0 i mrv2 u 1 2 i 1 t sS 2 A i AEG C
= s i
E fo r-D of i i n Gt a o2 e
- - - - i s1 H r2 r
. 0 0 0 0 0 paAe h 0 5 2 0 2 5 1 1 5 mRt oRon CHA CLo E ggE5aE$ HQZ 1. 1 e r u g R O-
_ = _ _ - _ - _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Mechanical testing of the cast stainless steel materials will be perfonned to detennine whether the losses in toughness are consistent with those predicted by correlations devel-oped from the laboratory database. Stress corrosion and corrosion fatigue tests will be i performed to investigate relative susceptibility to environmentally assisted cracking. j Selected materials will be thermally aged further at temperatures between 290 and 400'C l in the laboratory to obtain additional information and to identify possible artifacts intro- l duced during laboratory studies. Thennal. aging will also be carried out on material which has been annealed for I h at 550'C to recover toughness. 5 References
- 1. G. Slama, P. Petrequin S. H. Masson, and T Mager, "Effect of Aging on Mechanical Properties of Austenitic Stainless Steel Castings and Welds," SMIRT Post Conference Seminar 6 Assuring Structural Integrity of Steel Reactor Pressure Boundary Compo-nents, August 29-30, 1983, Monterey, CA.
- 2. O. K. Chopra and H. M. Chung, "Effect of Low-Temperature Aging on the Mechanical Properties of Cast Stainless Steels," in Properties of Stainless Steels in Elevated Tem-perature Sercice - MPC-Vol. 26/PVP-Vol.132, M. Prager, Ed., American Society of Mechanical Engineers, New York,1988.
- 3. H. M. Chung and O. K. Chopra, " TEM, APFlM, and SANS Examination of Aged Duplex Stainless Steel Components from Some Decommissioned Reactors," American Nuclear Society Annual Meeting, June 12-16, 1988, San Diego, CA.
- 4. R. D. Cheverton, J. G. Merkle, R. K. Nanstad, Eds, Evaluation of HF7R Pressure Vessel !
Integrity Considering Radiation Embrittlement. ORNL/TM-10444, Oak Ridge National l Laboratory, August 1988.
- 5. M. S. Wechsler, R. G. Berggren N. E. Hinkle, and W. J. Stelzman, " Radiation liardening and Embrittlement in a Reactor Pressure Vessel Steci," in Irradiation EfTects in Strue tural Alloysfor Thermal and Fast Reactors, ASTM Special Technical Publication 457, i American Society for Testing and Materials, New York,1968.
i 1 l 232 i l l
6
)
p, CONTRACT TITLE i Evaluation.'and Improvement'in [
- Nondestructive Examination (NDE)' Reliability for
~
i Inservice. Inspection of Light Water Reactors CONTRACTOR AND LOCATION' o q Pact fic Northwest . Laboratory P. O. Box 999, Richland, Washington- 99352- , l PROJECT MANAGER S. R. Doctor
-PRINCIPAL INVESTIGATORS J. D. Deffenbaugh, M. S. Good,'E. R. Green, P. G. Heasler, F. A. Simonen, J. C.- Spanner, T. T. Taylor, T. Vo .
ABSTRACT l
-The Evaluation'and Improvement of NDE Reliability:for Inservice Inspection Lof Light Water Reactors (NDE Reliability) program at the Pacific Northwest Laboratory was established by the NRC to determine the reliability of current inservice inspection ~.(ISI) techniques and to develop recommendations that will ensure a' suitably high. inspection reliability. The objectives of: this program include determining the reliability of'ISI performed on the primary systems of commercial light-water reactors (LWRs); using probabilistic frac-ture mechanics analysis to determine the impact of NDE unreliability on system safety; and evaluating reliability improvements that can be achieved with.
improved and advanced technology. A. final objective is .to . formulate' recommended - .
- revisions to ASl4E Code-and Regulatory' requirements, based on. material. proper-ties, service conditions, and NDE uncertainties. ~ The program scope is limited to ISI of the primary systems including the piping, vessel, and other inspected -
components. - This is a progress report covering the programmatic work 'from . October 1987 through September 1988. f h OBJECTIVE 1' C The Evaluation and Improvement of NDE Reliability for Inservice Inspection L of Light Water Reactors (NDE Reliability) program at Pacific Northwest Labora-
. tory (PNL) was established to determine.the reliability of current ISI tech-niques and to develop recommendations that will ensure a suitably high inspec-tion reliability. The objectives of this NRC program are to:
~
a determine the reliability of ultrasonic ISI performed on commercial light-p water reactor (LWR) primary systems n [ 233 L 1' L.
- using probabilistic fracture mechanics analysis, determine the impact of NDE unreliability on system safety and determine the level of inspection reliability required to ensure a suitably low failure probability
- evaluate the degree of reliability improvement that could be achieved using improved and advanced NDE techniques
- based on material properties, service conditions, and NDE uncertainties, recommend revisions to ASME Code, Section XI, and Regulatory Requirements that will ensure suitably low failure probabilities.
FY 1988 SCOPE The scope of this program is limited to ISI of primary systems; the results i and recommendations may also be applicable to Class II piping systems. ! The scope for FY 1988 included:
- Work to move program results into ASME Code.
- Conduct a re-analysis of the PISC-II data base.
- Conduct studies on UT equipment interactions to develop operating toler-ances.
- Assess adequacy of existing ISI requirements and develop, as needed, probabilistic-based ISI requirements.
- Provide expert consulting on field problems.
- Complete mini-round robin report.
- Conduct studies into the inspectability of coarse-grained materials.
- Conduct studies into developing surface preparation requirements to ensure effective UT can be performed.
~
- Complete analysis of ISI results versus true state of piping removed from the field.
SUMMARY
OF RESEARCH PROGRESS The program consists of three basic tasks: a Piping task, a Pressure Vessel task, and a New Inspection Criteria task. Because of the problems associated with the reliable detection, correct interpretation, and accurate characterization of defects during inservice inspection of piping, the major efforts were concentrated in the Piping task and the New Inspection Criteria task. However, work did begin this year on the Pressure Vessel Task. This report is divided into six sections. The first section highlights the Management activity, the second section covers the ASME Code Activities, the third section covers work related to Inspection of Reactor Pressure Vessels, ( 234 l l
i l
- J the fourth section covers New Inspection Criteria, the fifth section covers ,
one fast response for Consulting on Field Problems, and the sixth section l covers Piping Inspection activities. I i 1.0 NDE MANAGEMENT During this past year, many task activities made significant progress towards reaching conclusions and in developing the technical basis needed to establish positions for ASME Code recommendations and regulatory recommenda-tions. 3 2.0 ASME CODE ACTIVITIES Proactive participation in ASME Section XI activities continued toward achieving Code acceptance of NRC-funded PNL research to improve the reliability of nondestructive examination / inservice inspection (NDE/ISI). The objective of this task is to develop upgraded criteria and requirements for qualifying . ultrasonic testing / inservice inspection (UT/ISI) systems. i During the past year, PNL representatives attended four different series of meetings held in conjunction with the ASME Section XI Subcommittee on Inser-vice Inspection of Nuclear Power Plant Components. PNL staff hold memberships on the Working Group on Volumetric Examination and Procedure Qualification, Working Group on Surface Examination and Personnel Qualification, chair a Task Group on Acoustic Emission Monitoring, and serve as Secretary and member of the Subgroup on Nondestructive Examination (SGNDE). In May, a joint meeting of the ASME Boiler and Pressure Vessel Code Committees and the National Board of Pressure Vessel Inspectors provided an opportunity to attend ASME Section V Subcommittee meetings and serve as technical liaison between Section V and the SC-XI SGNDE. Following each SGNDE meeting, minutes were prepared and dis-tributed to a mailing list of about 85 addressees, along with agenda materials for the next meeting. A proposed revision to Code Case N-409 (N 409-1) received final approval from Section XI, the Main Committee, and the Board on Nuclear Codes and Stand-ards. Code Case N 409-1 consists of an expansion of N-409 that describes a statistically designed performance demonstration to qualify the personnel, equipment, and procedures used for UT/ISI of all light-water reactor piping welds in accordance with Section XI requirements. The proposed Appendix VII on Personnel Training and Qualification was formally approved by the Main Committee (M.C.), although two letter ballot negatives were received during second consideration of this item. A response to these two negatives was prepared, along with a proposed editorial revision to accommodate concerns expressed by the negators, resulting in withdrawal of
- one negative. Reaffirmation of the proposed Appendix VII, including the i editorial change, was approved by the cognizant Working Group, the SGNDE, and the Section XI Subcommittee. This item was then submitted for consideration by the Board on Nuclear Codes and Standards (BNCS). Four negatives were received from the initial BNCS ballot on Appendix VII, and an extensive response was prepared to address concerns raised in these negative ballots. Two BNCS members were contacted regarding their ballots, and both tentatively agreed ;
235 1... ...._________..____.m.. .____m..m __.m___.____m______----__. ..
I to withdraw their negative votes on this item. It is expected that the proposed Appendix VII will be approved by BNCS on a second consideration ballot to be issued in early October. The proposed Appendix VIII on UT/ISI Performance Demonstrations was approved by the SGNDE and Section XI Subcommittee for submittal to the Main Committee. It is expected that this document will be considered by the M.C. during the December 1988 meeting. This document includes essentially all of the provisions of Code Case N 409-1, plus it extends the performance demon-stration concept to other Section XI applications such as clad / base metal interface of pressure vessel shell welds, nozzle inner radius areas, pressure vessel shell welds other than the clad / base metal interface, nozzle-to-shell welds, and bolting and studs. When adopted, this Appendix will represent a significant enhancement in the performance demonstration requirements for all of the key Section XI UT applications. A proposed rewrite (restructuring) of IWA-2300 was approved by the SGNDE and SC-XI and was included as an introductory element in the proposed Appendix VII package. PNL staff gave technical presentations on a) new inspection criteria, b) the SAFT technology, c) Surry steam generator program results, i and d) acoustic emission technology overview to various SC-XI and/or SC-V ' groups. A document entitled " Qualification Process for Ultrasonic Testing on Nuclear Inservice Inspection Applications" (NUREG/CR-4882) was revised to accommodate NRC and PNL review comments. This document has now received PNL clearance and has been submitted for final NRC review. 3.0 PRESSURE VESSEL INSPECTION TASK 3.1 ANALYSIS OF PISC II PNL received a complete set of the Programme for the Inspection of Steel Components (PISC-II) round-robin data on the four plates from the Joint Research Centre, Ispra, Italy in June 1986. The initial objectives of this task were to review the data present and attempt to duplicate some of the results present in the PISC-II reports so that we can be sure that we understand the data and that it is correct. The specific tasks completed to date include:
- Assembling and computerizing complete information on the true-state data.
The original computer data did not contain a complete description of the flaws or the blocks. We had to extract the relevant information from PISC reports and communications with Ispra.
- Duplicating selected defect detection probabilities from PISC-II Report No. 5. This was an attempt to identify exactly what set of data was used in the PISC-II reports and verify the procedures Ispra used to calcu-late defect detection frequency probability (DDF).
- Implementing a scoring procedure for PISC data. We want to utilize dif-ferent scoring methods than those employed by PISC and also want to verify their results.
Review of PISC-II Scored Results in " CLEAN. PROC". After reviewing the data Tile " CLEAN. PROC" that was received from Ispra, we have determined the summary tables on pages 23-27 in Report No. 5 were developed as follows: 236
I The best detection results from each team (the selection was normally I made by computer program that compiled the results of several inspection i reports) were compared with intended defects only. A list of intended defects for each specimen may be found in PISC II Report No. 2. Other defects such as unintended weld fabrication defects or implantation defects, etc. were not used in scoring. None of the inner radius cracks (which were intended defects) { were used in scoring the data. Also, no false calls were reported in the 1 tables. i The following three examples provide a comparison of our analysis vs. the results in Report No. 5. When reading the examples, remember PNL did not score any inspection results - " CLEAN. PROC" contains scored results (i.e., PISC personnel have associated intended defects with each team's inspection results). All we have done is divide the total number of intended defects that should have been detected into the defects that " CLEAN. PROC" indicates were detected. Example 1: Team EC005499 Plate 1 PISC II Results 1.00 Our Results 1.00 Number of Defects 15 Team EC005499 is a computer-compiled selection of the best results of all procedures / techniques used by Team EC. PISC Report No. 2 indicates that Plate 1 had 15 intended defects. Reviewing data in " CLEAN. PROC" indicates that EC005499 did indeed detect all 15 flaws; therefore, its DDF is 1.00. In this example, our results and the results reported in Report No. 5 for Plate 1 are in agreement. Example 2: Team DB005599 Plate 1 PISC II Results 0.92 Our Results 0.93 Number of Defects 14 Note in this example that the number of defects is 14 instead of 15 for Plate 1. Fourteen defects were used in this instance because the inspection coverage coordinates in " CLEAN. PROC' indicate that team DB005599 (again a computer selection) did not entirely scan Plate 1 and an area where one defect was located was not scanned; therefore, only 14 defects were used in determining DDF. Reviewing the detection data in " CLEAN. PROC" indicates that 13 of the 14 defects were detected. Our results and the PISC Il results are in close agree-ment (13 + 14 = 0.9286) -- perhaps PISC II rounded down. 237
I 1 Example 3: Team ES000799 ' Plate 1 PISC II=Results -0.46 Our Results 0.40 Number of Defects 15
. Team ES000799 illustrates an example.where our.results and PISC II-do not agree and we cannot ascribe a logical reason for the disagreement. The' data in " CLEAN. PROC" indicates that the. entire block was scanned and that six flaws were detected. . 6 + 15 = 0.40 DDF, yet PISC II results. indicate a DDF of 0.46.
If one assumes that only 13 flaws should be used, then a DDF of 0.46 is correct; however, using 13 flaws in inconsistent with the inspection coverage data. Perhaps the data in " CLEAN. PROC" is wrong -- who knows?- Scoring Algorithm. PNL has. developed a software algorithm that will-
. score data from the RAW. PROC. data file. The algorithm compares the dimensions of each indication for a. specific inspection with flaw dimensions given in true-state data for'the. test plate that was examined. When all indication dimensions x, y, and z intersect with true-state flaw d_imensions, the algorithm associates that specific indication with a specific flaw. This scoring algo-rithm will be applied to the PISC-II data base, an analysis performed, and reported next year.
4 3.2 EQUIPMENT INTERACTION MATRIX The objective of this subtask is to' evaluate the effects of frequency
; domain, UT/ISI equipment interactions, and determine equipment tolerance values for controlling inspection reliability. An integrated computer model for the entire inspection system including the pipe section and flaw has been developed to ' explore the effects of frequency interaction on flaw detection and sizing.
Work in FY88 concentrated on the validation of the model for the acoustic components -- the perspex shoe and the flawed pipe section:
. Excellent agreement was-found between the model predictions and measured '
single frequency. beam patterns for through-transmission, 90 corner reflec-tion, and tandem-probe scanning of strip flaws. The model is based on two-dimensional, elastodynamics physical optics (EPO) ray tracing theory. 1
. Comparisons between model predictions and frequency domain measurements of specular reflection signals from large, smooth flaws showed good agree- 1 ment. An example is the frequency domain comparison shown in Figure 1 '
for specular reflection within a block with an end cut to 42 measured ! using a 45* SV transducer.
~
l
. Model calculations showed that flaws can have transfer functions that would I produce decreased detection and sizing reliability due to frequency domain
-equipment interactions. Several worst-case flaws can be identified in j the calculated transfer functions shown in Figure 2. j l
i 238
1
-)
i
,.)
1 1.0 L I i ' j ~ ' K- --- Predicted j - N_ Measured a N ia '\ g -
;\
o { _ 0.5 - E 4 - r ~ 1 -
\ s
.j N
%/ N O I I I l O 2 '4 6 8 10 Frequency, MHz FIGURE 1.Normalized - Measured vs.respect with Predicted to 45Transfer Flaw Functions for 42 Flaw.
10
% Smm 10 mm Y'%\
~~--
0 - \ - . . - 20 mm
-- 50 mm c,o
\n .N g
-N -10 i
\g \.Ns \
m f .. \ P'j '
< -20 -
*8./% g '
s z j
\s_/~'~ .'
2 30 - 40 - T '. :
-50 t I e I \l O 2 4 6 8 10 j Frequency, MHz j l
FIGURE 2. Frequency Responses of Various Sizes of 85 Flaws i i 239 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ a
- The model results showed the primary cause of undesirable, flaw transfer functions to be phase cancellation at the receiving piezoelectric element face due to the wavefront and face not being parallel. Thus, frequency i domain equipment interactions might be greatly reduced through the use of a phase-insensitive receiving probe. The worst-case flaws were selected based on placing nulls in the center of the equipment passband.
. Calculations for equipment center frequency and bandwidth sensitivity studies were completed using several worst-case flaw transfer functions calculated by the ray tracing model. It was found that the present equip-ment bandwidth tolerance of *10% as given in ASME Code N 409-1 is accept-able and that it may be possible to relax this standard slightly. The results for the center frequency study have not been analyzed yet.
It was concluded that the present equipment tolerances for things like pulser amplitude can be relaxed; however, tolerances that affect the frequency characteristics will need to be tightly controlled. It should be noted that ASME Code currently controls all operating characteristics at the *10% level. 4.0 NEW INSPECTION CRITERIA Work continued on the assessment of the adequacy of existing ASME Code requirements for ISI and on developing technical bases for improving these requirements to assure safe nuclear power plant operation. Several inter-related activities on this task have been directed to the development of probabilistically based inspection requirements. Particular attention has been directed to requirements for inspection intervals and weld inspection sampling plans. DEVELOPMENT OF PROBABILISTIC APPROACHES During this past year PNL has continued with the development and assessment of alternative approaches for probabilistically based inspection requirements. , This activity has emphasized interaction with NRC staff, other laboratories, ! ASME groups, and industry efforts as performed by EPRI. An overview document has been written that outlines a comprehensive prob-abilistic approach for the development of improved inspection requirements. This document provides a flow chart that relates inspection requirements to quantitative goals for improvements in systems safety. Computational methods and sources of data for quantifying the approach are described and are critic- 1 ally reviewed. The document will guide future efforts on the NDE Reliability program. PNL has actively participated in the startup of an ASME Task Group on Risk-Based Inspection Guidelines, which is to be an ASME research activity. Participation in this group is expected to further the goals of the NDE Reli-ability program, and to lead eventually to specific recommendations for the introduction of probabilistic methods as a basis for ASME Section XI require-ments. While the initial focus will be on nuclear power, the group will also be concerned with insights to be gained from applications in other industries such as aircraft, navel, petrochemical, and civil engineering structures. While a planning meeting was held on February 18, 1988 at Washington, D.C., 240
l e> , further activities have been delayed as funding for the task group is being sought. l A one-day workshop was held at PNL on January 20, 1988 to permit an exchange of views between PNL and NRC staff on the topic of PRA methods and !' their potential use in establishing priorities for inservice inspection. Parallel applications to' prioritize plant aging issues were noted. Also dis-cussed was the approach to be taken by the proposed ASME Task Group on Risk-Based Inspection Guidelines. DATA BASE ON PLANT OPERATING HISTORIES This activity responds to a recommendation from a 1987 workshop with NRC staff,.which suggested that data bases and industry records be searched for information on piping failures and repairs, and also for information on findings of inservice inspections (effectiveness). During the past year, we established sources of such information and determined the potential usefulness of the types of information that is most readily available. Contacts with utilities indicated that suitable records are maintained at plant sites, but.that costs of on-site visits to extract the information was beyond the scope of the NDE Reliability program. Further discussions with NRC staff revealed two potentially useful data bases; namely, Licensing Event Reports (LERs) and an industry-maintained data base available through the Nuclear Power Plant Reliability System (NPRDS). A trial search of the potentially more useful NPRDS data base was per-formed. It was determined that most of the information relates to " failures" of minor consequence (gasket leaks, cracks in small diameter fittings, etc.) and that these " failures" were typically found visually through evidence of leakage. Nevertheless, inservice inspection has in many cases been effective in detecting weld cracks. The NPRDS data were determined to be relatively accessible, and. easy to interpret. Therefore, a complete evaluation of all the piping related failures (400 items) in the NPRDS data base was initiated. Useful trends will be extracted from this information and reported during the next year. OCONEE-3 PILOT STUDY l The objective of this study was to initially assess and demonstrate the feasibility of using data from existing Probabilistic Risk Assessments (PRA) to establish inspection priorities for pressure boundary systems and components. A pilot application of PRA methods to the Oconee-3 plant was completed during the past year. The study was based on PRA data from an EPRI study (NSAC-60) and on data for failure probabilities from NRC-funded evaluations of actual observed failure data (NUREG/CR-4407). Based on the results of the pilot study, the proposed use of PRA methods has been demonstrated to be a useful tool for identifying those systems and piping sections or welds that need to l be inspected with the highest priority. Table 1 lists a number of Oconee-3 systems and rankings that provide insight into which systems should be given the highest priority for inservice I inspection. Two alternative ranking parameters were employed. The Birnbaum parameter addresses the consequences of failure, given that a failure does 1 241 l l
a l TABLE 1. Rankings of Systems and. Components for Inspection Priority as Based on. Risk Considerations for Oconee-3 Weld Inspection. Birnbaum-Imp rtance Importance f System (a) Rank Value Rank Value I Low-Pressure / Injection (b) 1 (5.9E-06) 2 (1.5E-02)
' High-Pressure Injection 2 (5.1E-06 5 (5.4E-03)
Reactor Pressure Vessel 3 (5.0E-06 1 (1.0)~' Steam Generators: 4 (1.5E-06 9 1.5E-04).
. Emergency _Feedwater . 5 (7.2E-07 3 1.5E-02)
' Service Water 6 (3.6E-07 4- 7.7E-03 .{
7 (1.7E-07) 6 3.6E-03 Reactor Coolant (c) Power Conversion 8J 8 2.1E-04 E
' (8.0E-08).
Standby Shutdown Facility 9 (3.0E-08) 7 6.9E-04 Instrument Air 10 (7.0E-10) 10 -(1.5E-05 (a)0nly systems of interest to Code-Type ISI' are listed. (b)Under normal conditions, the most frequently used function of the LPI system is decay-heat removal (OHR) after a shutdown. (c)'The PCS system consists of the following: main-feedwater, main steam, condens' ate, condenser circulating water, and vacuum systems. occur. This parameter focuses inspection towards the most safety critical systems (importance to preventing core melt), even if they have performed very reliably in the past. In contrast, the Weld Inspection Importance para-meter makes use of estimates of system reliability to focus additional attention towards systems that are more'likely to experience service failures. In general, the two parameters 'give. similar priorities. However, there are notable 1 exceptions such as the steam generator, which moves up in priority when the. relatively poor service performance of steam generator tubes is taken into , consideration. ; A further step in the pilot study involved a much more detailed assessment of the emergency feedwater system. Failure Modes and Effects Analysis (FMEA) was applied to identify and prioritize the most risk-important piping sections within this system. The results were then compared to the current inspection-requirements'specified by ASME Section XI. 5.0 CONSULT ON FIELD PROBLEMS This is a summary review of the round-robin studies that have been con- , ducted to quantify the performance of ultrasonic procedures to detect, locate, i and size defects in pressure vessel shells and nozzles. The results are exam-ined in light of the inspection conditions and ultrasonic procedures in use in~the United States. 242
j w
-i tc In 1965, the U.S. Pressure Vessel Research Committee (PVRC) began an NOE .!
program.in which teams of trained inspectors conducted inspections on thick-section-test blocks containing various welding defects. During the early stages of this program, it was found that the inspection procedures were not
.well defined and the~ published results revealed large differences in performance between teams using nominally the same inspection procedure. A generally small. defect detection probability was found which was worse when the inspec- 1 tions were performed through cladding. A new program was initiated in 1974 to test procedures related closely to those required by the 1974 ASME Code Section XI. This-round robin study consisted of three unclad test blocks provided by.PVRC and used in the aforementioned.PVRC' program (two butt-welded '!
plates and one welded forged nozzle) and was under the guidance of the Plate j Inspection Steering Committee (PISC I). The PISC I program was designed to investigate the. capability of the 1974 ASME Section XI. inspection requirements to detect, locate,.and size defects. In addition, several European agencies applied alternative'. procedures, some of which were employed for ISI-in Europe at that time. A total of 34 inspection agencies took part in the PISC I pro-gram, with 28 teams employing the ASME-based procedure. In 1981, another round-robin testing program was started under the guidance of .the Program .for the Inspection of Steel Components (PISC II) and was com-pleted in 1986. The PISC II program was designed to evaluate the capability of procedures to detect and locate defects.as well as to accurately size the defects. The PISC II program had 50 inspection agencies involved-in the inspec-tion of the four plates (two butt-welded flat plates and two plates containing nozzles). The'U.S. had five teams that employed manual inspections and several teams that employed advanced ultrasonic techniques. The PISC-I and II results can be summarized by the plot in Figure 3. The vertical-axis is the probability that a defect will be detected and correctly sentenced, and the horizontal axis is the through-wall extent of the defect. The term " sentenced" pertains only to defects which were rejectable according to ASME Section XI IWB-3510 1974 Edition and include only those defects which were not undersized by more than 3 mm. The 50% DAC curve is the performance for t.SME 1974 Code, while the 20% DAC is for the 1986 version of ASME Section XI Code. The error bars on the two curves show the range of variability of the teams employing procedures in the spirit of ASME at the respective DAC levels. These results clearly show the ineffectiveness of procedures based on early ASME Code. This point is further seen in Figure 4, where the detection probability (MDDF which is the mean value of the defect detection frequency for ASME procedure on all the defects in the specified plate) is plotted versus the'DAC level for three of the plates. The trends are pretty similar for I each plate, but show the improvement that can be achieved by the 10% and 20% DAC levels. Clearly, the 50% DAC level provides very low defect detection rates. The results from the two PISC studies can be summarized by the plots shown in Figure 5. The two studies are consistent with one another and show a clear division of the defects into three categories. The data in Figure 5 , is only for Plate 3 and for 50% DAC. The mest difficult defects to detect are the smooth, sharp planar defects that would include cracks. f , 243 i _ __ . _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ __ _____A
J 1.o 3
- Special '/,,-
p - Procedures . , _ (European) / y
- / Marshall Committee f
Y - /
~
1
/
~
0 / i
~
' x
-a 0.5 - /
g _ / ASME 20 or 10% DAC l (New Proposal) 3e _ i
/
O - I
-; - x
.e _
ASME 50% DAC 5 (industrial) h a _ 0.0 M i
- f f Y l 0- 10 20 30 40 50 60 70 Defect Through Wall Size (mm)
FIGURE 3. Plot of the Defect Detection and Correct Sentencing of Defects for Plate 3 for Various Inspection Procedures as a Function of Through-Wall size h 1 -
- u. 2 2 o 2- -9 9-O 9 3I 3-
=
.5 -
17 - 1 l 3 I l 0 I I I I
> J 50% DAC 35% DAC 20% DAC 10% DAC ]
i Cut-Off Level ] FIGURE 4. Mean Value of the Defect Detection Frequency for ASME Procedures ) for All the Defects in Each Specified Plate J I l 244 l _ ._-______---O
,A - f- PISC 11 Results (Plate No. 3) h- ASME 50% DAC ,e di. ASME 50% DAC g1
- p. p j
. .y .
1
^U ,#
.5 C' l ' C ,5 C O
l ,' A 2 ,e' A i s B- p
' #_L-w--
1 -e - r - - m U _ J# _ L _ i_ .. _ L .A . m 10 50- 10. 50
. Defect Size (mm) Defect Size (mm)
PISC l Results (PISC 1 Procedure = ASME 50% DAC) E JL di 5 e -1 j y'1 ,.
.9 s .* '/
5 .5 Cl B J' 5 C D < e ,- f y .5 , O l _
- o. A ,a* _
_./ A
$ ', i i d 'r m o L'$ A --W-*- 7
. 50 10 . 50 Defect Size (mm) Defect Size (mm)
A: Smooth sherp pioner defects B: Volumetric defects C: Rough cracks or combination of smooth planer defects and volumetric defects FIGURE 5. ' Plot of'Results from PISC I and PISC II for ASME 50% DAC Procedures of Defect Detection Probability Versus Defect Size when Classifying the Defects into Three Categories 0 100 5 5 50 $ 6 E
.E -
35 gE 10 % 7or g ,; n -eN o H eL_. ._ el 20 Q .,, g k EE io l' 20 - 45*_ to k e y ,- -e . y
= <
e s* - Q 5 0 50 100 150 200 Position in Depth of the Defect, mm FIGURE 6. Amplitude Response (DAC Level) from the Lower Tip of Smooth, Planar Defect that is used for Detecting this type of Defect showing that for the 45' Inspections, 5% DAC must be. Employed 245
.The smooth,. sharp planar defects are detected by the lewer crack tip response', and Figure 6 shows some PISC-II data that shows the DAC levels
_ required to detect this lower tip signal for defects with a 10- to 25-mm through-wall height. The response is plotted as a function of the defect i
; location from very near the surface to 200-mm deep. These data show that if ' ! .a 45* probe. is to be used to detect this class of defect (i.e., cracks), then it will have to be performed at 5% DAC. These data clearly show the necessity for using a 60' and 70 inspection at 20% and 10% DAC, respectively.
l The five U.S. teams (not all inspected each plate) which participated in
. the PISC-II round-robin exercise used a procedure that was developed by the PVRC and was called Procedure.C. A comparison of ASME. Code minimum requirements and the PVRC Procedure C is shown in Table 2. The significant difference noted is the inclusion of a near-surface technique, the increased scan sen- .
sitivity, and the lower recording level. The results for the U.S. teams are j shown below in Table 3: ! TABLE 2. Comparison of ASME Code Minimum Requirements Contrasted to the PVRC Procedure C that the U.S. Teams Employed in the PISC-II Program Code (Minimum)- Procedure C
- Techniques (Minimum) 0 , 45 , 60 0 , 45 , 60 , 70 RL Dual Element '
(InnerNearSurface)
*- Recording Level 50% DAC 20% DAC
= Scan Overlap Overlap Minimum Beam 50%, 1/4" Near Surface Dimension, or 50%
- Scan Sensitivity i 6 dB 14 dB and 12 dB (70 )
i TABLE 3. Detection Performance Plate Number Defect Detection Probability Number of Teams
#2 94-100% 5
#3 50-84% 3
#9 75-92% 3 ;
l 246
Since the U.S. teams performed manual inspections, the results found in the PISC study may not be directly comparable to field ISI work because much of the field work is performed with mechanized scanning systems. However, the PISC-II data does show that the average inspection performance was nearly the same for both manual and mechanized inspections for the plates in this study. This was somewhat of a surprise, but probably is a result of the human being yielding high performance under the attention of an important international study. Thus, these data are useful and the U.S. teams' performances were found to be consistent with other teams using comparable procedures. The U.S. teams provided useful data in showing that reasonably effective examinations can be performed on vessels, but it also demonstrates that there will be a considerable variation in inspection results from teams using the same inspection procedure. These data indicate that considerable increase in inspection reliability could occur if the reasons for the variability in inspection performance were under-stood and corrective actions taken. In the PISC-II round robin, it was found that the procedures employed had a lot to do with the performance that was achieved in the detection of defects. As an example, one team (denoted as EC) performed a number of inspec-tions according to a variety of conditions. These results are summarized in the table below: Team Procedure and Code DDF EC005499 10% M,B,S,TD 1.0 EC000699 20% M,U 0.94 EC002599 20% M,C,TD 0.94 EC001699 20% M,V,S,TD 1.0 EC005399 20% M,B,S,TD 1.0 EC005099 35% M,B,S,TD 1.0 EC001199 50% M,U 0.56 EC003099 50% M,C 0.56 EC001299 50% M,U,S 0.61 EC005699 50% M,U,S,TD 0.78 EC005799 50% M,C,S,TD 0.83 EC004899 50% M,B,S,TD 0.83 The legend for this table is: M -- Manual examination B -- Examination from both sides of the plate 10-50% -- Percent DAC reporting level l S -- Use of near-surface technique U -- Unclad-side inspection C -- Clad-side inspection TD -- Tandem technique DDF -- Defect Detection Frequency These data clearly show the use of a 50% DAC level is unacceptable result-ing in a defect detection frequency of 0.56. The utilization of a 20% DAC with a tandem procedure resulted in a defect detection frequency of 0.94. These data are very useful since it was all obtained by a single team and, thus, allows one to focus on the effectiveness of various procedure parameters. l 247
. Conclusions. It .is' apparent from the PISC-I and PISC-II data bases that important conclusions can be drawn about defect detection performance. However,- ,it must be remembered that the data _have some qualifications that have to be ' recognized 'in order to keep the results -in perspective.
- 1. The PISC round-robin tests were conducted in the laboratories of the teams and, thus, were conducted under' optimum conditions. For.this reason, these results represent an upper bound on performance.
- 2. The PISC-round-robin tests were designed to measure procedure capability and,.thus, are an upper bound of performance and should not be confused with procedure reliability.
- 3. Many of the inspections were conducted with teams having researchers'and other-highly qualified staff that do not normally conduct field ISI:
thus, these results again form an upper bound for performance.
- 4. The' condition of the cladded surfaces were to the high standards in use in Europe and are not typical of.the conditions that exist in U.S. pressure vessels. Thus again, the results represent an upper bound for performance.
The significant conclusions that pertain to ISI as practiced in the U.S. are:
- Even under the most optimum conditions of the PISC exercises, the use of the 50% DAC and other minimum requirements from the 1974 Edition (including the Summer of 1975 addenda to ASME Section XI Code) clearly demonstrate that vessel examinations performed to these requirements will not be effective for detecting defects.
* -The PISC-II round-robin data base indicates that significant improvement in the effectiveness of pressure vessel examinations can be achieved by working at 20% DAC and including an effective near-surface technique.
Thus, pressure vessel examinations should be conducted, at a minimum, to the requirements of the 1986 ASME Code Edition (including the Winter 1987 addenda) since this code has been upgraded to contain these require-ments.
- Given the caveats above, the performance of U.S. teams (using the upgraded requirements that are in the 1986 ASME Code Edition) during the PISC-II trials indicates that high . levels of defect detection can be achieved.
The data also shows large variability in the performance of the U.S. teams using the same procedure. If the reasons for the variability could be understood and corrective action were taken, then considerable improve-ment in pressure vessel inspection reliability would result. 6.0 PIPING TASK 6.1 MINI-ROUND ROBIN-A complete description and analysis of this study can be found in NUREG/CR-4908. The Mini-Round Robin (MRR) was conducted to provide an engineering data base for UT/ISI that would help: L 248 f?
quantify the effect of training and performance demonstration testing that resulted from IEB 83-02, e quantify the differences in capability between detecting long (greater than 3-in.) cracks versus short (less than 2-in.) cracks, and quantify the capability of UT/ISI technicians to determine length and depth of intergranular stress corrosion cracks (IGSCC). I Data from the depth sizing portion of the MRR was analyzed using both absolute and relative statistics. An overview of this analysis is presented below. This section examines flaw sizing capability using absolute measurements. Table 4 presents a summary of the MRR sizing data expressed in absolute units. TABLE 4. Summary of Absolute Sizing Performance Destructive UT Crack Depth Estimate (inches) Depth, in. Team 1 Team 2 Team 3 Team 4 Team 5 Team 6 Team 7 Team 8 0.000 0.550 0.165 0.220 0.234 0.516 0.072 0.131 0.062 O.000 0.103 0.124 0.330 0.117 0.172 0.048 0.344 0.200 0.000 0.103 0.124 0.323 0.144 0.344 0.227 0.399 0.310 0.033 0.352 0.171 0.110 0.160 0.413 0.099 0.132 0.198 0.077 0.088 0.248 0.028 0.061 0.028 0.061 0.209 0.171 0.127 0.132 0.099 0.028 0.000 0.055 0.165 0.330 0.193 0.143 0.253 0.138 0.127 0.055 0.413 0.074 0.209 0.110 0.172 0.275 0.158 0.220 0.069 0.344 0.034 0.261 0.179 0.209 0.039 0.160 0.149 0.127 0.440 0.088 0.209 0.116 0.227 0.578 0.124 0.248 0.151 0.516 0.034 0.172 0.158 0.242 0.649 0.346 0.381 0.069 0.649 0.606 0.502 0.199 0.255 0.103 0.055 0.385 0.076 0.344 0.048 0.117 0.110 0.450 0.675 0.476 0.485 0.113 0.649 0.433 0.589 0.598 If all inspections were performed on material of the same thickness, both relative and absolute sizing measurements would give the same results. However, in the MRR specimens, two are constructed of thicker material (0.9 in versus 0.6 in.); and furthermore, one of these thicker specimens also contains the largest crack in the data set, a circumstance that*could cause differences. It is also important to note that the regressions performed with the two types of measurements rest on different assumptions about measure-ment error variability. The absolute model assumes that variability remains the same as thickness changes; the relative model assumes that variability increases as wall thickness increases. Consequently, the absolute model " weights" observations from the thick sections more heavily than the relative. In the current data set, absolute regression will weight the 0.45-in. and 0.242-in. cracks more heavily than the relative. 249
Another important difference between the two models occurs when results of a regression fit are usd to predict sizing performance in other thicknesses of material, particularly thicker material. The absolute sizing model, when 3 extrapolated to thicker material, produces a smaller variance for the measure- ' ments than the relative. Thus, extrapolations using the absolute model are less conservative than the relative. In Table 5 regression parameters obtained from the absolute measurements are presented. These regressions utilized inspections on cracks larger than 0 04 in. From these results, we see that the slope, and more importantly, the R2 statistic is consistently larger for the absolute regressions, indicating a better fit to the data. TABLE 5. Summary of Team Regression Fits for Absolute Sizing Data with Small Cracks (( 0.040-in.) Removed Absolute Results Relative Results Standard Intercept, Deviation' 2 2 in. Slope in. R Slope R* Team 1 -0.02 1.57 0.21 0.43 0.96 0.12 Team 2 0.04 0.78 0.11 0.38 0.25 0.03 Team 3 -0.06 1.36 0.08 0.80 1.37 0.62 Team 4 0.03 0.22 0.04 0.27 0.27 0.18 Team 5 0.04 1.61 0.15 0.59 1.77 0.44 Team 6 -0.05 1.05 0.18 0.30 0.58 0.08 Team 7 0.10 0.91 0.13 0.38 0.27 0.03 Team 8 -0.03 1.12 0.10 0.62 0.73 0.25 2
*R = square of multiple correlation coefficient The residuals of both sets of regressions were examined in an attempt to determine exactly why the absolute model fit the data better than the relative model. We found the difference to be entirely attributable to a single crack (the crack with depth 0.45 in. in Table 5). When this crack was deleted from the regressions, both models fit equally well. As explained before, this crack is weighted more heavily in the absolute regression because it is in a thicker section. The absolute model does better because the teams can size this one crack better than the rest.
Selection of the superior model depends entirely on whether or not this particular crack is " anomalous." It so happens that the two cracks in the I thick sections were manufactured with a different process than the others and were consequently much stronger reflectors. Because of this and the conserva-l i tive nature of the relative model, we favor it in spite of the better fit shown by the absolute model. 250 _______-_________L
l 6.2 INSPECTION OF CAST STAINLESS STEEL The objective of this task is to evaluate the effectiveness and reliabil-ity of ultrasonic inspection of cast materials used within the primary pressure boundary of LWRs. Due to the coarse microstructure of this material, many inspection problems exist and are common to structures such as cladded pipe, inner-surface cladding of pressure vessels, statically cast elbows, statically cast pump bowls, centrifugally cast stainless steel (CCSS) piping, dissimilar metal welds, and weld-overlay-repaired pipe joints. Far-side weld inspection is an inspection technique included in the work scope since the ultrasonic field passes through weld material. Activities included weld-overlay-repaired pipe joints and CCSS. Weld-overlay repair is being used as a temporary repair mechanism for BWR piping weakened by IGSCC and is being sought as a longer-term repair mech-anism. NUREG/CR-4484, Status of Activities for Inspecting Weld Overlaid Pipe Joints, was published in 1986. Activities thereafter were monitored and a status update provided in the form of a Research Information Letter (RIL). i The primary conclusion of the redrafted RIL (April 27,1988) was that i much work has been performed to demonstrate the effective ultrasonic inspection of weld-overlay-repaired pipe joints; however, insufficient data exists to classify this inspection as effective and reliable. The RIL also described an evaluation test that was recommended for providing sufficient data for deter-mining if the technique is effective and reliable. CCSS piping is used in the primary reactor coolant loop piping of 27 pressurized water reactors (PWRs) manufactured by the Westinghouse Electric Corporation. However, CCSS inspection procedures continue to perform unsatis-factorily due to the coarse microstructure that characterizes this material. Past work began with an evaluation of distortion incurred by an ultrasonic field after it had propagated through the pipe wall thickness. A matrix of different wave modes (both longitudinal and vertically polarized shear waves) at different resolution settings (6, 3, and 1.5 mm) were passed through either a pure equiaxed or a pure columnar CCSS microstructure and the resulting ultra-sonic field maps were collected [4,5]. To expand this work to the more complex microstructure that exist in CCSS, an upgrade to the data acquisition and analysis system was made. This was needed since the signal patterns of the complex material forms have numerous spurious signals that interfere with the data acquisition process of the previous system. Another factor was that the majority of field material was thought to be of a mixed microstructure. i Activities for the past work period included acquiring three additional CCSS pipe sections, acquiring a second scan matrix of ultrasonic field maps with the upgraded system, and submission of an article to the annual Review of Progress in Quantitative Nondestructive Evaluation [7]. Three CCSS pipe sections were on loan to PNL from Southwest Research Institute (SwRI) and reported at the last LWR meeting. The ultrasonic field map system upgrade was implemented by digitizing the RF signals and implementing a post-gating process. The data acquisition technique was also changed so that the field map of a 45 longitudinal-wave field was essentially not degraded by receiver directivity. This was accomplished by applying the microprobe to 251
Carbon Steel Equiamed Microstructure Columnar Microstructure l , 0
.____.__-_..._.-0 --
'0 'O Diffusely Mixed Microstructure Layered Microstructure' Color Key IdB)
O Scan Aperture.114 x 114 tmmf 137 x 137(mmf' FIGURE 7. Ultrasonic Field Maps of 1-MHz, 45 , Longitudinal-Wave Fields a 45 facet which increased reception uniformity about a broad angular zone centered about 45 . The previous technique (applying the microprobe normal to the sample) biased sensitivity toward 0 . Four maps were taken from material of each microstructural classification, each in unique material volumes (Figure 7); however, only one field map for each microstructural classification was displayed. The objectives of this work were to determine if a 1-MHz, 45 , longitu-dinal field maintained spatial coherence in all the microstructural forms of CCSS, to quantify the degree of distortion incurred by the ultrasonic field, and to evaluate if an effective ultrasonic inspection could be performed in all the CCSS microstructure. Spatial coherency was evaluated by examining all 20 field maps. Each field map except one (from the diffusely mixed micro-structural sample) displayed an ultrasonic field in which the 0 to -3 dB region was contiguous. Thus, the spatial coherency of the transmitted field was rated as high for the pure microstructural forms and as moderate for the mixed microstructural forms. 1 Field distortion was evaluated by measuring the refracted angle and the l positional variation of the field (Figures 8 and 9). Significance of the i refracted-angle results was the increased standard deviation of the equiaxed l sample and the extremely large standard deviations of both materials having a l mixed microstructure. j Another parameter selected to quantify field distortion was the field-position variation normalized to field width. When inspecting a selected material volume, sufficient field overlap is designed into the scan procedure to ensure detection sensitivity remains high for a defect anywhere within the 252 l \ l l
i 48 -"---- 47 , 46 e
- e. e a
6 45 g 44 e e 43 e * . f 42 i g 41 j 40 E 39 . , 38 I
)
f/ c,) ' o' o# o O # ** o Material Classification FIGURE 8. Effective Refracted Angles of 1-MHz, 45 , Longitudinal Wave ? NFPV .~ - , where MFW is the Mirumum of Measured 3 dB Fie s d Widths and RVFP is the Range of Vasiation in Feeld Position Normahned Field Position Var,ction 10 --- - - - 09 -- Circumletemial ; Measurement i Ansal Measurement c 07 I C E. O6 , s Os [- o - C g 04 . g
~j o3 _-
l 02i 0v r 00 ! I' . . .I ___E: _ . . _ _... . _ . _ _ _
+*$/ /$** ,/
P y,/$t* y,fj/ pV s'j/ Material Classibcation FIGURE 9. Positional Variation of Longitudinal-Wave Fields 253
material volume of interest. Obviously, if a small field width exists, then the allowable field displacement error must be small. Likewise, if a large field width exists, then the field-displacement error may be larger. Therefore, the field-position variation was normalized to field width and are plottec in Figure 9. The variations associated with the reference samples are assumed to be indicative of set-up variation. For CCSS material, increased variation was expected for the equiaxed material and ranged between 27% and 28% for values pertaining to measurement along both pipe axes. A low circumferential value of 5% was obtained for the columnar samples; however, the axial value was 23%. This latter value might initially seem high since the standard deviation of the refracted angle was small; however, the axial, -3 dB, field width of the columnar scans also is reduced and produces a higher normalized value. The two mixed microstructural forms had values ranging between 45% and 55%. This is alarming since scan patterns on a pipe may use circumferential increments as high as 50%. If two successive measurements are made and the field misdirection is outward from the two positions, then a material volume thought to be inspected by past procedures might be skipped. Due to the difference in field distortion, the worst-case material classi-fication (mixed equiaxed-columnar microstructure) should be assumed for an ; inspection. An alternative is to continuously determine the microstructure ' as a scan is performed and to interrupt the data acquisition process and imple-ment an appropriate technique customized to the detected microstructure when the probe passes to a different microstructure. This latter choice assumes an effective microstructural classifier and that an effective inspection tech-nique exists for each of the possible microstructure. ! 6.3 SURFACE ROUGHNESS CONDITIONS The objective determined for this work was to establish specifications such that an effective and reliable ultrasonic inspection is not prevented by the condition of the exposed surface. Past efforts included an attempt to 4 quantify the effect produced by an outer surface irregularity. This approach I was then modified through coordination of the NRC with EPRI in establishing a mathematical model to be used as an engineering tool to derive guidelines for surface specifications. Activities for the past work period included the formulation of a coordination plan between EPRI, NRC, the Center for NDE (CNDE) ! at Ames Laboratory, and PNL; a visit by CNDE personnel at PNL; an exchange of l data between CNDE and PNL; and development of better experimental procedures for obtaining quantitative data for comparing to the model predictions. Both EPRI through the CNDE at Ames Laboratory and the Research Branch of the NRC through PNL have developed capabilities that are uniquely suited for establishing a validated model. First, CNDE has extensive experience in the computational modeling of ultrasonic wave propagation fields in solid materials. l For this reason, EPRI and the NRC have established a three-year time frame in which the two organizations, through the referenced institutes, will cooperate i in attempting to determine and validate an ultrasonic computer model. To l facilitate the cooperation between CNDE and PNL, a coordination plan was for-mulated. This plan assigned individual and joint responsibilities to both ; CNDE and PNL. 254 i , i i
1 The first-exchange of data was completed in February, which involved an immersion scan with isotropic materials. The bottom line of this work was that the model and expeirmental results were in good agreement for the longi-
/ . tudinal mode except for low amplitudes (i.e., f -14 dB). . There are still some discrepancies with the shear-wave data that are being studied further.
PNL began an effort of refining ultrasonic microprobes and data acquisition techniques so that the collected experimental data would accurately measure the parameters predicted by the model. This effort resulted in longitudinal-a wave microprobe refinements, a presentation concerning the development of a L shear-wave microprobe at the " Review of Progress in Quantitative Nondestructive Evaluation," and submittal of a manuscript for publication in the conference proceedings ~[8]. 6.4 CHARACTERIZATION OF FIELD PIPE The objective of this task is to provide pipe weld specimens that can be used to help determine the effectiveness and reliability of ultrasonic ISI that is being performed on BWR piping. This goal will be accomplished by obtaining BWR recirculating piping from a variety of power plants. These weld specimens will also serve to support PNL laboratory studies and such programs as PISC III. Weld specimens were acquired from Monticello and Vermont Yankee BWR nuclear power plants. The welds were sectioned from the pipe remnants in FY 1986. Due to high amounts of alpha contaminations on the Monticello specimens, it was-decided to decontaminate only the 11 Vermont Yankee specimens and wait until FY 87 to have the 28 Monticello weld specimens decontaminated. A complete characterization of the 11 Vermont Yankee weld specimens was performed by PNL personnel; this included ultrasonic and penetrant examinations. After'the decontamination of the 28 Monticello weld specimens (14-12", 2-22",12-28") PNL representatives performed a complete characterization on each individual weld specimen. This included penetrant testing on the ID of the weld region, ID and OD weld profiles, counterbore and penetrant pictures.
. In some cases, ultrasonic verification was performed on some of_ the specimens.
After the evaluation was finished, the weld specimens were packaged and returned to PNL. I L During FY88, a complete UT evaluation was performed on a selected number ' l of weld specimens (20 welds) using manual ultrasonics and the PNL SAFT-UT field i system. When the UT evaluation was completed, a thorough analysis was performed L on the data that was gathered. A weld specimen summary report (which included l' PT and UT information, and specimen characteristics) was put together on 42 weld specimens. A selected number of these specimens will be sent to Europe to be l used in the PISC III international round robin on austenitic steel. The remain- 4 ing weld specimens will remain at PNL to be used to support other future NRC-
- l. related programs.
In FY88, five of the Monticello safe-end weld specimens were packaged in an overseas shipping container and are pending shipment to Europe. These o will also be used in the PISC III exercises. The remaining five safe-ends were packaged and sent to a DOE burial site in Richland, Washington. 255
l 6.5 PISC III Work continued in support of the Programme for the Inspection of Steel' Components to ensure that the results will- be useful and reflect the practice, conditions, and defects of interest to the U.S. light-water reactors. Specific thrusts were in the area of the round robin studies on austenitic stainless steels. FUTURE RESEARCH PLANS j Continue the research described in this review to reach completion and prepare the research results for recommending and defending regulatory positions and/or code revisions. Continue providing technical consultation to the NRC on field problems as they arise. REFERENCES
- 1. Plate Inspection Steering Committee (PISC). Full Reports, Euratom Report, EUR-6371-EN-(Vol. I/VI, 1979).
- 2. Plate Inspection Steering Committee, " Analysis of the PISC Trials Results for Alternative. Procedures," Euratom Report No. 6, EUR-6371-EN (1980).
- 3. Program for the Inspection of Steel Components (PISC II), Organization for Economic Cooperative and Development, Nuclear Energy Agency, Committee on the Safety Of. Nuclear Installations, CSNE No. 117-121, September 1986.
- 4. Ultrasonic Inspection of Heavy Section Steel Components, The PISC II Final Report, edited by R. W. Nichols and S. Crutzen (Elsevier Applied Science Publishers Ltd, England, 1988).
- 5. Good, M. S. and L. G. Van Fleet, " Ultrasonic Beam Profiles in Coarse Grained Materials," in 8th International Conference on NDE in the Nuclear Industry, edited by D. Stahl (American Society for Metals International, Metals Park, Ohio, 1987) pp. 657-666.
- 6. Good, M. S. and L. G. Van Fleet, " Mapping of Ultrasonic Fields in Solids,"
in Review of Progress in Quantitative Nondestructive Evaluation, edited by D. O. Thompson and D. E. Chimenti (Plenum Press, New York, 1988), j Vol. 7A, pp. 637-646.
- 7. Good, M. S. and E. R. Green, " Mapping of 1-MHz, 45 Longitudinal Fields in Centrifugally Cast Stainless Steels," in Review of Progress in Quanti-tative Nondestructive Evaluation, edited by D. O. Thompson and D. E.
Chimenti (Plenum Press, New York, 1989), Vol. 8.
- 8. Good, M. S. and E. R. Green, "A Shear-Wave Microprobe Utilizing Surface-l Wave Mode Conversion," in Review of Progress in Quantitative Nondestruc-l tive Evaluation, edited by D. O. Thompson and D. E. Chimenti (Plenum l-Press, New York, 1989), Vol. 8.
256 l 4
b-CONTRACT TITLE I Final Developments, Validation and Technology Transfer for AE and SAFT-UT l CONTRACTOR AND LOCATION Pacific Northwest Laboratory P. O. Box 999, Richland, Washington .99352 l l PROJECT MANAGER S. R. Doctor 1 PRINCIPAL INVESTIGATORS P. H. Hutton, J. C. Spanner, L. D. Reid ABSTRACT The program for Validation and Technology Transfer for AE and SAFT-UT is designed to accomplish the final step of moving research results into beneficial application. Accomplishments for FY88 in the areas of Acoustic Emission (AE) and Synthetic-Aperture Focusing of Ultrasonic Test data (SAFT-UT) under this , program are discussed in this paper. The information is treated under the I topics of Code Activities, Field Validation, and Seminars. Projected FY89 activities will continue to focus on these three areas. OBJECTIVE The program concerning Validation and Technology Transfer for AE and SAFT-VT has the objective of developing field procedures, performing field validation testing, and providing training for NRC headquarters and regional staff for the use of advanced NDE technology. The objective also includes l ASME Section XI Code acceptance of the AE and SAFT-UT technologies to facilitate I implementation of the methods by Regulatory staff and the utilities. This program will benefit nuclear regulation in the following areas: l
. . Provides a mechanism for technology . flow from PNL to the NRC staff.
- Supports the process of moving advanced technology into the ASME Code.
- Provides technology that can solve ISI problems and, thus, can aid the NRC staff in assuring the structural integrity of components.
l 257
11 FY 1988 SCOPE
-The FY88. work scope consisted of:
- Conduct = a workshop (s).with NRC headquarters and Regional staff in the .
SAFT, AE, and eddy current. technology. 1
- Pursue the acceptance of technology through ASME Code changes.
. Work with'NRR and Regional staff.to identify places where.these advanced ;
technologies can be applied to resolve ISI problems for the NRC. ; r
. Pursue validation field testing.of AE and SAFT-UT through international L agreements and the PISC program.
= .Begin field' validation of advanced techniques.
1 The program for each of the technologies is structured under three' general headings:
-1. Code Activities. Establish the technologies as a recognized and approved part of the ASME Code through the appropriate instrument (Code Case, Code' Appendix,. Code modification, etc.).
- 2. Field Validation. Apply the technologies to problems on operating reactors to demonstrate validity for the intended applications of the technology.
- 3. Seminars. Present technical information and demonstrations to NRC Regu-latory. staff to. acquaint them with the technologies and the application methodology..
l
SUMMARY
OF RESEARCH PROGRESS FY88 results are presented under the three program structure headings described above. AC0USTIC EMISSION Code Activities A Nonmandatory Appendix to the ASME Section XI Code titled " Acoustic Emission' Monitoring of Nuclear Reactor Pressure Boundaries During Operation" was approved by the Working Group on Volumetric -Inspection and the Subgroup on Nondestructive Examination in the November 1987-January 1988. time frame. It was subsequently tabled by the Subcommittee on Nuclear Inservice Inspection with.the-recommendation that it should be resubmitted as a Code Case. A Code Case version of the material was approved by the Subcommittee in April 1988 and is on the Main Committee agenda for December 1988. l 258 l
. Field Validation Most of the on-reactor ' demonstration of technology developed under the -
NRC. AE program has taken place at the TVA Watts Bar Unit 1 Nuclear Power Plant. Selected areas of the reactor pressure boundary have been AE monitored during cold hydrostatic [1] and hot functional [2] . testing of the reactor system with significant benefit to the AE technology. A third set of tests has been performed to evaluate the signal source location capability of the installed AE system. It is important to evaluate the signal source location capability of the i AE. system on a reactor because of the large wall thickness used in pressure boundary components. AE signal source location algorithms available today for practical application locate in two dimensions. Many reactor pressure vessels have a wall thickness of 8 to 12 inches or more. This is particularly true of the vessel nozzle region. With a third dimension of this magnitude, there is a need to evaluate the accuracy of the two-dimensional source location. The Watts Bar Unit i vessel, being completely empty, offered an opportunity to evaluate the source location function using simulated AE signals injected on the inside of the vessel and nozzles. Signal attenuation by the vessel cladding and the clad-to-base metal interface in the 400 to 500 kHz frequency range being used to monitor for AE precluded use of the normal pencil leak break signal source to inject signals at the.inside of the vessel and nozzles. In real application of AE monitoring, this problem would not be present because one would be seeking to detect signals originating in the ferritic vessel wall. An electronic pulser with a transducer was used to-inject acoustic signals at known points around the inside of the
#2 inlet nozzle and on the inside surface of the vessel.
Figure 1 shows the location of the AE sensors on the #2 inlet nozzle and. I the location of the signal inputs on the inside surface of the nozzle. Source I location for signals injected on the nozzle ID was accurate to the projected signal input location within about 3 to 5 inches in the circumferential direc-tion in all cases. In the front-to-rear direction, the location was accurate to within about 5 to 7 inches for signals originating within the region under the AE sensor array. Figures 2 and 3 illustrate representative results obtained from input signals within the AE sensor array. The front-to-rear accuracy
. deteriorated progressively as the source point moved farther outside of the AE sensor array. There are methods of improving the front-to-rear location accuracy for signals outside of the sensor array by using additional sensors; however, since the signals are being detected and the radial location is rea-sonably accurate, it may be difficult to justify the additional hardware for the increase in source localization accuracy.
No usable data was obtained on the vessel wall. In its present operating mode, the AE monitor instrument requires that all four sensors in an array l receive a signal before it will be processed for source location. Signal attenuation in the cladding and at the clad-to-vessel interface was large enough to prevent detection of the injected signals at all four AE sensors mounted on the vessel wall. Many of the signals were detected by two or three of the sensors which would be sufficient to perform a less-than-optimum source j
' location. This suggests that the operating mode of the AE instrument needs j r
259
seanos, o a sea,on A A
, g
, I . ,
_-e.__._- ( . 4 2 in f q ,o ,n 3_ _,,,, d,,nn,ut une'p i 7g ,'/, ; _ _ J9.e 1"ft'ia*.___p,,i; ,n _ -l / _L_ loa g_, _ / c _igtL'te__. D- ,
'4 -
?- 1 ,,,,,,,
3, ,n n.d.o. i,.gu, t,n. j, sensor
/> l
, , 3 4 ,n
"' {
U""_'_ ; __ __3J a_'ap"' 'et $"g"ao' . . _3ha 'nilet _ _ / o' i 80.' 270 90
<_. 4-0 Ad
= 30,n 56in FIGURE 1. Signal Input Locations on #2 Inlet Nozzle, Watts Bar 1, for AE Source Location Calibration Location Plot Array 1 g 3 ,1 I . . 6 F
=__,_,Ao_ 31,__,,,_,_,_____,;
12 e . in input Line C 10 =
) 82 Cold Leg Pipe 1
.5 Norste Rollout m Syrnbol Key )
8 - f i A 1. 1
= 9 8 1 2.
5 C 2. 2 h6 -
] D 2 E 3 3
4. 0 . 4
> i
=
4 j A Vessel Shell j a I I O J 270' 0- 90' 180' 270- ) X Coordinate Date 12-14 87 Tape Posn.an 1.10 1,23 Filter EV6417 FLT Tape i De 0 Inlet #2 Watts Bar 31 en e AE Sensors i FIGURE 2. AE Source Location Indications - Input at 0 on 31" Input Line 260 _________a
j :: 1 s 1 Locatson Plot Array 1 e 3 i input Lin g 10 .^ #2 Cold Leg Pipe
;A ;
4 Symbol Key g j
- g ;8 , B 2 4.
Norrie Rollout C'4 '6 Vessel Shell
. ~4 t . .1 i
'I I '
0 270' 'O' 90" 180" 270' X-Coordinate i ter EV7629 FLT a D Inlet #2 Watts Bar 31 in. e AE Sensors FIGURE 3. AE Source Location Indications'- Input at 270 on 31" Input Line c.- to be changed.to accept data from less than four sensors to give minimum' source location information in such cases. The fact that these test signals were not detected by all four sensors-does not imply inability to detect AE signals originating from crack growth in the vessel. wall belt-line region. That_ capability has been demonstrated separately in two previous tests [3,4]. In summary, the test results demonstrated that acceptable two-dimensional L location of acoustic signal sources can be achieved on a reactor structure even in'the relatively' complex geometry of the nozzle-to-vessel joint. Seminars Material describing the status of AE technology development and potential applications beneficial to regulatory needs was presented to NRC Region III in a technical workshop held in February 1988. A similar presentation will be. made at NRC Region I, King of Prussia during a technical workshop scheduled for: October 10-14, 1988. The workshop is being organized by Region I with attendance'and participation by Regions II and IV plus NRR and RES staff members. These workshops provide an opportunity to discuss current regulate;y. needs vs. NDE technology capabilities. 261 _ . __ . _ - - _ - - - _ _ - _ _ _ _ _ _ _ _ _ - _ - _ _ _ _ _ _ _ _ _ _ = _ - - - _ _ _ _ _ _ - - _ - - - - _ -
___- _ _ _ _ _ _ = _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ I 1 i SAFT-UT Code Activities I The emphasis of this activity was to get the SAFT technology into ASME Section XI Code. The position of Section XI has been not to include prescrip-tive procedures in the Code because this has been found not to work. Section , XI has taken the position that performance demonstrations will be used, and any l system passing the demonstration requirements can be used for ISI. Hence, we I have discussed where else would it be appropriate for the technology to be l included in the Code. Section V was then approached and was receptive to including SAFT in a new appendix to Article 4 called " Computer Processed Imaging." This happened at the end of the fiscal year and no other progress can be reported. Field Validation Two activities occurred this year that stand out with regard to field validation. The first was participating in the inspection of an 0.D. indica-tion in the belt line region of the reactor pressure vessel at Indian Point Unit No. 2. The second was the inspection of several defects in a cast elbow on the hot leg of the Trojan Nuclear Power Plant. In addition, a report was published detailing the operational principles and implementation of the SAFT-UT real-time inspection system (NUREG/CR-5075). This report provides a full description of the hardware and software used in the real-time SAFT-UT system, how it operates, how to use it, and where to go to find more information. A final report was drafted for the SAFT-UT program and was approved by the NRC program manager for publication once a new section is added based on the work to be performed on thick sections early next fiscal year. This final report pulls together all the relevant information on studies and documentation that completes the work performed at PNL to develop and engineer a fieldable SAFT-UT real-time system. This final report also contains the procedure that has been developed to conduct inspections for defect detection, characterization and sizing. The inspection that was performed at Indian Point Unit No. 2 was conducted through the use and cooperation of Westinghouse and their vessel scanner and Dynacon and their data acquisition system. PNL provided the transducer to be used in conducting the examination, and the other companies conducted the exam-ination and provided PNL with the digitized data for subsequent processing at PNL. Some problems were encountered with the equipment during the data acqui-sition and the number of scans was reduced to 2 and 1/2. The 1/2 was a scan with only 1/2 the aperture in the increment direction. The data turned out to be quite useful and provided some very good images. There were several limitations of the equipment that prevented us from doing all of the work that we wanted; for example, the scanner could not increment reliably in one wavelength intervals; and thus, we were limited to only doing 2-D or line-SAFT processing. The results obtained by SAFT processing are shown in Figure 4. If the indication was surface connected then the image in Figure 5 would be expected; and hence, it is concluded that the indication was not surface connected. 262
o i 4
- ;: . ~ , . . .,
.: u'
~
~
~
- y;' ' t . . .
[- s 7 \ s ;-
. ./ .
O
. _ .. ~.. . . _ . _ . .
u_.t .a'q ;i citi-t_ . H.,.;G i-ii l i
...........<.a......_,. -
FIGURE 4 SAFT Processed Image of the Indication Located in the Belt Line Region Near the Outside of the Indian Point Unit 2 Nuclear Power Plant i
^ '
N j , 3 ,, 1 FIGURE 5. SAFT Image of a Surface-Connected Notch that is 0.3 Inches Deep that was Located in a Calibration Block and Used to Assess the Capability of the Westinghouse Scanner and the Dynacon Data Acquisition System with the PNL SAFT Technology 263 l
This is extremely important and needed to be determined to aid in assessing the significance of the defect. The data supports the position that the indi-cation was not parallel to the 0.D. surface but was at a taper. The inspection at Trojan was similar in that the data acquisition was handled by Amdata using their I-98 system. PNL provided a probe for the data l acquisition, and the data was subsequently processed and analyzed at PNL. The results were rather dramatic and provided some very good images of the indica- l tions. The SAFT images had about a 10 dB increase in the signal-to-noise l ratio versus the images collected and displayed with the conventional inspection using the I-98. This was a statically cast component, and it had a relatively fine grain structure that made it easy to image and to obtain good signal-to-noise ratios. The data was collected with a 500-kHz shear-wave probe at a 45 inspection angle. The images of the defect areas shown in Figure 6 produced some very good l I results that made it possible to cluster the indications. The indications were disbursed; and through analysis, it was pretty apparent that there were a number of indications and we chose to combine them in a logical manner. These two field trials illustrate that other commercial equipment can be used to collect SAFT-type data and can give very good results that help to resolve inspection problems. The results were very good, but there has not been any destructive correlation to ascertain the bottom-line effectiveness of the SAFT performance. These tests were very encouraging in the sense that the SAFT technology is compatible with many pieces of commercial equipment.
' gg y
- g r 'uM
~ '
u"
's =
l ') . g
.,, ' 4 b'@}j'$,III%E [*$j*kIUI%$ '
. ===<= = = - = ..a FIGURE 6. SAFT Image of One Zone Scanned on the Cast Elbow at the Trojan Nuclear Power Station showing the Clustering of the Indications and the High Signal-to-Noise Ratio 264
j
-Seminars The participation in seminars includes all the activities already high-~
lighted in the AE section and includes the SAFT field system demonstration at the ' October 1988 seminar. In addition, Region.I personnel had some hands-on experience during " June 1988 while they were at PNL to gain training on IGSCC i in some field samples at PNL. l
SUMMARY
Significant progress has been made in obtaining ASME approval of a Code Case 'for' inservice AE monitoring to be incorporated into the Section XI Code. Calibration of source location capability of the AE monitor system installed-at Watts Bar Unit I reactor showed that good two-dimensional. location of an acoustic signal source can be accomplished even on the complex nozzle-to-vessel geometry. Progress has been.made in transferring AE technology to NRC Regional and NRR staff. Substantial progress has been made in finalizing the SAFT field system documentation, initial field trials with other commercial UT equipment with extremely good resulting images, and finding a receptive audience for acceptance
.of the SAFT technology within the ASME Code.
FUTURE RESEARCH PLANS Program plans for FY89 include:
- Complete ASME approval of AE Code Case.
. Follow up on expression of interest from Taiwan Power Company and Toledo-Edison Company in AE monitoring reactor components.
- Participate. in technical workshop scheduled to be held at NRC Region I.
- Develop input for Appendix to ASME Section V Code to the SAFT technology.
- Participate in field trial of SAFT technology evaluation on thick-section steel at the EPRI NDE Center.
=' Work to develop a means whereby the SAFT technology can be used in the PISC III program on the Full Scale Vessel.
. Publish the final report.
REFERENCES
- 1. Hutton, P. H. T. T. Taylor, J. F. Dawson, R. A. Pappas, and R. J. Kurtz.
1982. Acoustic Emission Monitoring of ASME Section III Hydrostatic Test, 1 Watts Bar Unit 1 Nuclear Reactor, NUREG/CR-2880. Pacific Northwest Laboratory, Richland, Washington. 265
I
- 2. Hutton, P. H., J. F. Dawson, M. A. Friesel, J. C. Harris, and R. A. Pappas.
1984. Acoustic Emission Monitoring of Hot Functional Testing, Watts Bar Unit 1 Nuclear Reactor, NUREG/CR-3693. Pacific Northwest Laboratory, j Richland, Washington. l
- 3. Ibid, pp. 18-19.
- 4. Hutton, P. H., R. J. Kurtz, R. A. Pappas, J. F. Dawson, L. S. Dake, and J. R. Skorpik. 1985. Acoustic Emission Results Obtained from Testing the ZB-1 Intermediate Scale Pressure Vessel, NUREG/CR-3915. Paci fic
' Northwest Laboratory, Richland, Washington.
i 1 I 266 ______-__-__-_________a
h*f , ,
. FINAL EVALUATION OF ADVANCED AND CURRENT LEAK DETECTION SYSTEMS
- D. S. Kupperman, R. Carlson, W. Brewer,.and R. Lanham Argonne National Laboratory 1 Materials and Components Technology Division Argonne, Illinois. 60439 December 1988 i-he satssi.ittaJ manuerript t.a6 twen authored by a conleartur of the U. S. Giwernment un.hrt contract - No. W-31 109 E NG-30.
AcLovdin31 y, the U $ Governewn! reimiris e nonexclusive, royalty free hcente to publish ur reproduce trie published lorrn of this contribution, os dilow othess to riu no, for U S Governmem purposes. l, { 1 l 1 Contribution to Compilation of Contract Research for the Materials EnnineerinF. Branch. Division of Ennineerinn Technolor.v: Annual Report for FY 1988, to be published by the Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission, Washington, D.C.
- Work supported by the U S . Nuclear Regulatory Commission, Office of Nuclear Regulatory Research, FIN NO. A2250; RSR
Contact:
J. Muscara. 267 1
;- i
,I 1
FINAL EVALUATION OF ADVANCED AND CURRENT LEAK DETECTION SYSTEMS
- by D. S. Kupperman, R. Carlson, W. Brewer, and R. Lanham Materials and Components Technology Division ARCONNE NATIONAL LABORATORY 9700 South Cass Avenue Argonne, Illinois 60439 ABSTRACT This report presents the results of a study to evaluate the adequacy of leak detection systems in light water reactors. The sources of numerous reported leaks and methods of detection have been documented. Research to advance.the state of the art of acoustic leak detection is presented, and procedures for implementation are discussed.
OBJECTIVE Assess' the reliability of leak detection technology for LWRs and evaluate advanced acoustic leak detection concepts for detection, location, and characterization of leaks'. FY 1988 SCOPE Complete and update the review of the reliability of leak detection systems. Complete the evaluation of acoustic leak detection for in-reactor detection, discrimination, and location of leaks through laboratory testing of valves and cracks (including ICSCC) with leak rates of 1 to 10 gal / min. Complete evaluation of techniques to discriminate crack leaks from valve and seal leaks. Prepare a comprehensive document (guidance manual) for acoustic leak monitoring of reactor primary systems. Prepare a draft for revision of NUREG Guide 1,45. I l.
- Work supported by the Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission, FIN NO.: A2250; RSR
Contact:
J. Muscara. l 268 .l 1 i
)
l i
SUMMARY
OF RESEARCil PROGRESS l Current Practice l l According to the Code of Federal Regulations 10CFR part 50, means must i he provided for detecting and, to the extent practical, identifying the locat Ion and source of leakage. The safety significance of a leak can i vary, and prompt and quantitative information is necessary to permit imme-diat e act ion should a leak he detrimental to the safety of the f acility, U.S. Nuclear Regulatory Commission Guide 1.45 (1) recommends the use of at least t hree dif ferent methods to detect Icakage in reactors. Monitoring of both sump-flow and airborne-particulate radioactivity is recommended. A third method can involve either monitoring of condensate flow rate from air I. coolers or monitoring of airborne gaseous radioactivity. Although the current methods for leak detection reflect the state of the art , other techniques may he developed and used Regulatory Guide 1.45 also recom-mends that leak rates from identifie and unidentified sources he monitored separately to an accuracy of 3785 cm / min (1 gal / min), and that indicat ors and alarms for leak detection be provided in the main control room. Since the recommendat.lons of Regulatory Guide 1.45 are not mandatory, the technical specifications for 74 operating plants that include pres-surized-water reactors (PWRs) and boiling-water reactors (BWRs) have pre-vlously been reviewed [2] to determine the types of Icak detection methods employed, the range of limiting conditions for operation, and the sur-veillance requirements for the leak detection systems. The results are repeated here for completeness. All plants use at least one of the two systems specified by Reguintory Guide 1.45. All'but eight use sump monitoring, and all but three use par- l ticulate monitoring. Monitoring of condensate flow rate from containment i air coolers and of atmospheric gaseous radioactivity is also used in many plants The limit on unidentified leakage (" identified" leakage go generally that collectedfrommonitoredvalven)i~orallPWRsis37gScm/ min (1 gal / min), wherens the limit for most BlRs is 18930 cm fmin (5 gal / min). Thelimitsongotalleakagearegenerally37850cm/ min (10 gal / min) , for ITRs and 94630 cm / min (25 gal / min) for BWRs. (Regulatory Guide 1.45 does not specify leakage limits, st that the leakage detec.
- t. ion system should be able to detectbut does stggg/ min Icak in 1 h. )
a 3785-cm In some cases, ed in the plant l ! technicallimitsorratesofincreaseinleakagearealsostag/ specifications. TwoBWRsgavealimitof379cmmin /h(0.1 gal / min /h); four have a limit of 1893 cm / min /h (0.5 gal / min /h). Leakage is checked every 12 h in most PWRs, and every 4 or 24 h in most BWRs. One BWR specifies that a continuous monitor with a control room alarm shalI he operational. For BWRs, calibration is generally performed a t- 18-mont h int ervals; functional tests are performed every month. l 269 i t l 1
The estimated sensitivity of leakage monitoring systems is occasion-ally addressed in the technical specifications. For example, one speci-z fication indicates that air particulate monitoring can, in principle, detect a 0.013-gal / min leak in 20 min, that the sensitivity of gas radio-activity monitoring is 2-10 gal / min, and that the sensitivity of condensate l flow monitoring is 0.5-10 gal / min. Continuous sump pump monitoring appears i capable of detecting a 1-gal / min leak in 10-60 min. One other safety-related aspect of improved leak detection systems concerns radiation exposure of plant personnel. Improved systems with leak location capability could reduce the exposure of personnel inside the con-tainment and could present an attractive alternative to augmented in-service inspection (ISI) . Improved leak detection is consistent with the defense-in-depth philosophy of the Nuclear Regulatory Commission (NRC) and would lead to earlier detection of system degradation. Unidentified leakage ultimately passes through the sump pump unless it is trapped in the system. In addition, condensate from the containment air cooling systems passes through a flowmeter and then through the sump, adding to the unidentified leakage. Identified leakage, primarily that which is selectively collected from leaking valves, flows to a drain tank that is also pumped out. The total leakage is the combined unidentified and identified leakage. Estimates of leak rates are obtained from the cooling system flow meter, the level indicators, and the frequency of operation of the pumps. There are no requirements for monitoring leakage outside the contain-ment. These leaks are detected by a variety of indicators, such as tem-perature and pressure rises and changes in background radiation, and by visual examination during routine maintenance. Many methods can be used to detect a leak. These methods include ; radiation, sump, and condensate flow monitoring, coolant inventory, and l measurement of variations in temperature, pressure, and dew point. Generally speaking, reactor operators reIy on sump pamp monitorjng to establish the presence of leaks. Other methods appear to be less reliable or less convenient. Ingostreactors, the surveillance periods are too long to detect a 3785-cm / min (1-gal / min) leak in 1 h, as is suggested by Regulatory Guide 1.45, but it appears that this sensitivity could be achieved if monitoring procedures were modified. However, simply tightening the current leakage limits to improve sensitivity is not adequate, since this might produce an unacceptably high nuaber of spurious shutdowns because of the inability of current leak detection systems to identify leak sources. None of the systems currently in use provides any ! information on leak location, and leaks must be located by visual ' examination after shutdown. In order to help characterize more quantitatively the cause of leaks i in reactors and to obtain information regarding the adequacy of leak detection technology, Licensee Event Report (LER) Compilations from June l l l l 270 l i i l _ _ _ _ _ _ - _ _ _ _ _ -
1985 to March 1988 (e.g., LER Compilation for March 1986, NUREG/CR-2000, ORNL/NSIC-200) were reviewed. These compilations contain summaries of in-formation submitted by the nuclear power plant licensees in accordance with federal regulations. Each summary includes the date of the incident; the i reactor, component, and system involved; and, if a leak occurred, usually the leak rate and the action taken. Of over 4000 reported events, 91 were identified as relevant to the problem of detecting leaks in the primary systems of light water reactors (LWRs). Differences between PWRs and BWRs with regard to leak detection have been analyzed. The primary source of leaks, accounting for 56% of reported leaks, is valves and pumps.[1} The sump monitor accounts for 46% of the detection methods reported. Events have been divided into those in PWRs and those in BWRs. FWRs account for about 71% of the reported leaks. This is slightly higher than the percentage of PWRs in the U.S (about 70%). Although pumps and valves are the main source of leaks regardless of the reactor type (58% of leaks occur in PWRs and 49% in BWRs), a greater percentage of small-line leaks occurs in BWRs than in PWRs (39% vs. 14%). With regard to detection methods, the greatest differences between reactor types are as follows: (a) The sump pump is reported as the detection method more frequently in BWRs than in PWRs (66% vs. 37%). (b) The radia-tion monitor is reported as the detection device (excluding false alarms) more frequently in PWRs. In fact, for the events studied, the radiation monitor never correctly detected a leak in a BWR (it did, however, initiate four BWR false alarms). Another point of interest is that inventory bel-ance was reported as the method of detecting a leak in 16% of the PWR cases. For both BWRs and PWRs, about one false alarm occurs for every three actual leaks. Anomalous signals from radiation monitors are the cause of these false alarms. Tables 1.1-1.3 summarize the analysis presented above. Numerous questions arise in connaction with an assessment of the ade- l quacy of leak detection. One concern is whether the flow path for uniden- ! tified ' leakage to the sung pump is uninpeded. All indications are that < fluid from a leak will pass directly to the sump pump if it is not absorbed f by the environment or insulation. Levels in the containment are separated ! l by gratings that permit the fluid to pass to the sump (s). Another concern j is the time it takes to locate a 2eak. In general, leaks are located by ! visual examination, which is a slow process. (For this reason, an j important benefit of improved leak location capability would be reduced I exposure of personnel to radiation.) In addition, in the case of BWRs, the start of the examination can be delayed by up to six hours while the inert a gas is removed from the drywell. ! l 271 L
l Table 1.1. Leak Sources for LVRs
-Leak PWR + BWR PWR BWR INR BWR Source (% of total) (% of total) ( S. of total) (% of PWR) (% of BWR)
Valves 46 35 10 48 37 Pumps 10 6. 3 10 12 Small 20 10 10 14 39 Lines ICSCC 3' 2 2 2 6 Misc. 21 20 2 26 6 TOTAL 100 73 27 100 100 Table 1.2. Leak Detection Methods for LWRs PWR + BWR PWR BWR PWR BWR Detection (% of (% of (% of (% of (% of Method total) total) total) PWR) BWR) Sump Pump 46 27 19 37 66 Radiation 19 19 0 26 0 Monitor Visual 14 7 6 11 20 Inspection Inventory 12 12 0 16 0 Balance Other 10 6 4 10 14 TOTAL 100 71 29 100 100 Table 1.3. False Alarms Obtained with Leak Detection Systems in LWRs (% of actual leaks) i PWR + BWR PWR BUR 31 23 8 l 1 l J 272
l 3 l The issue of whether a significant delay in leak detection could result from the absorption of leakage by the environment or insulation has also been addressed. A simple calculation based on the ideal gas law (PV - I nRT) has indicated that even in the worst case, i.e., with an ambient i temperature of 120*F (323 K) and the cooling condenser off, a delay of only I afewhourswouldresultfromtheabsorptionofmogsturebytgeenviron- ! ment. Assuming a containment volume of 500,000 ft (14,000 m-) and a vapor 1 pressure of 12 kPa, the maximum amount of water that can be absorbed by the j l air is about 300 gal. At a leak rate of 1 gal / min, saturation would be reached in about 5 h. With the condenser on, moisture from a leak would be ; collected at the sump in a much shorter time. The question of whether a significant amount of moisture could be held in the insulation is more difficult to answer and has not been addressed. Although sump monitoring can be reliable if conscientious surveillance is maintained, the reliability of radiation monitors is questionable, pri-marily for two reasons: (a) The high background radiation level in some reactors forces the alarm trip point to be set so high that the monitor is potentially insensitive to a rise in radiation level due to a leak; in one case, the radiation alarm was not activated by the presence of a 25-gal / min leak. (b) Spurious electrical signals cause false alarms to occur at a relatively high rate. Also addressed was the issue of whether action is taken before leaks exceed the flow rates recommended in the plant technical specifications. The answer to this question is "not necessarily," according to the LERs that were reviewed. Reported flow rates ranged from 0.3 gal / min to ">32 gal / min total;" sometimes, reports simply described leakage as " excessive." Problems._ Associated with Current Leak Detection Technology Although current leak detection systems are adequate to ensure a leak-before-break scenario in the great majority of situaticns, one must also consider the possibility that large cracks may initially produce only low leak rates. This situation could arise because of corrosion plugging or fouling of relatively slowly growing cracks or the relatively uniform growth of a long crack before penetration. In such cases, the time re-quired for a small leak to become a significant leak or rupture could be j short, depending on crack geometry, pipe loading, and transient loading l (due to a seismic or water hammer event). The shortcomings in existing leak detection systems are not simply a matter of conjecture. The Duane-Arnold safe-end cracking incident [2] indicates that the sensitivity and reliability of current leak detection l l systems are clearly inadequate in some cases. In the Duane-Arnold case, l the plant was shyt down on the basis of the operator's judgment when a leak rate of 11360 cm / min (3 gal / min) was detected; however, this leakage rate i is below the required shutdown limit for almost all BWRs. Examination of the leaking safe end showed that cracking had occurred essentially com-pletely around the circumference. The crack was throughwall over about 20% of the circumference and 50-75% throughwall in the nonleaking area. i 273 4
l i The concern about potential problems with current leak detection tech-nology extends beyond the U.S. borders. The experience with PWRs in France has been discussed in a paper [3] presented at an international conference on surveillance of reactor coolant boundaries. French regulations related to primary coolant systems are based on NRC Regulatory Guide 1.45. In practice, however, leak detection is largely based on the chemical and control volume tank level and (to a lesser extent) on the sump level and flow monitor. Locating leakage is generally difficult and is done by local inspection after a leak is detected. The main components involved in leaks of' primary coolant systems in France have been valves and, to a lesser ' extent, primary pump casing seals. During transient operation, the leak detection capability is reduced; as a result, the French Safety Authority has required that primary coolant leakage detection and quantification methods be improved. Otherwise, few problems have arisen in France from the primary coolant leakage detection system in the past few years. Future Needs and Current Developments It has become apparent that no single currently available leak-detection method for LWRs combines optimal leakage detection sensitivity, leak-locating capability, and the desired level of accuracy in leakage measurement. For example, although quantitative leakage determination is possible with condensate flow monitors, sump monitors, and primary coolant inventory balance, these methods do not provide adequate location infor- ! mation and are not necessarily sensitive enough to meet regulatory-guide goals. Leak detection capability can be improved at specified sites by acoustic monitoring or moisture-sensitive tape (MST).[5] However, current acoustic emission monitor (AEM) techniques provide no source discrimination (e.g., to distinguish between leaks from pipe cracks and valves) and no leak rate information (a small leak may saturate the system). The MST provides neither quantitative leak rate information nor specific location information other than the location of the tape; moreover, its usefulness with " soft" insulation needs to be demonstrated. Since tne issuance of NRC IE Bulletins 83-02 and 82-03 and the traim ng of ultrasonic inspection persor:ncl, the probeb!.lity of de%cting IGSCC under field conditions has increased. However, many cracks are missed during ultrasonic ISI and are detected only because of leakage, thereby raising doubts concerning the capability of ultrasonic ISI to detect cracks. The present ultrasonic testing procedures for ferritic weldments (ASME Code Sections V and XI) do not appear to be adequate for the detection and evaluation of ICSCC in austenitic stainless steel (SS) piping. The detection of IGSCC before the cracks have grown large enough to cause a leak, and the detection, location, and sizing of leaks once they occur, are very difficult technical goals to achieve. Leak detection techniques need further improvement in the following areas: (a) identi-fying leak sources through location information and leak characterization, so as to eliminate false alarms; (b) quantifying and monitoring leak rates; and (c) minimizing the number of installed transducers in a " complete" system through increased sensitivity. l l 274 l 1
Acoustic Leak Detection System . The first step in the implementation of an acoustic leak detection (ALD) and location system is to identify acoustic receiver sites and determine the spacing required between waveguides to meet the sensitivity needs of the system. The spacing scheme will differ with the reactor being monitored (i.e., PWR or BWR) and will depend on the required level of sensitivity. Estimates of S/N ratios for IGSCC leaks as a function of distance and acoustic background levels are presented in Ref. 5, Section III.C.3, Fig. 3.28. The level of background noise-must be established for the various regions to be monitored in order to optimize the number of sensors used. The figure in Ref. 5 can be used to estimate the optimum sensor spacing once the desired sensitivity and background noise levels are established. 'While the figure is for 10-in pipe, the data will be assumed to be valid for all piping systems unless alternate data are available. Attenuation measurements will have to be obtained for other piping systems in the field to obtain more precise sensor spacing information. The results presented in Fig. 3.28 of Ref. 5 are for BWR conditions. Because of the higher pressure in a PWR, the acoustic signals for a given leak rate are higher. Adding 6 dB of S/N to the results of the figure ) should provide a conservative estimate of acoustic signal vs. leak rate for a PWR. As an example, assume that 100 m of monitored piping in a BWR (the approximate length of the primary pressure boundary) is divided into low, moderate, and high levels of acoustic background noise covering 40, 40, and 20 m of piping, respectively. Also assume that the desired detection sensitivity is 1 gal / min. For a 3-dB S/N ratio, the required sensor spacings (for a signal in the 300- to 400 kHz range) are approximately 10, 2, and 1 m. Four sensor locations are required to cover the low-acoustic-background area (40 m of piping), 20 sensors for the moderate-noise area (40 m of piping), and 20 sensors for the high-noise region (20 m of piping). For location analysis, three sensors are required at each site to carry out the correlation averaging routine. Therefore, 132 sensors are 4 needed to adequately caver the reactor primary pressure boundary under the proposed conditions. For a PWR, assume that 150 m of piping has a 60-m/60-m/30-m set of piping lengths at low , moderate , and high-background levels. With an increase of 6 dB in signal intensity for a P4R compared to I a BWR, one has sensor spacings of 12, 4, and 2 m for a 3 .B d S/N ratic. Approximately 150 sensors will be required [3 x (5 + 15 + 30)) to com-pletely monitor the plant under this scenario. Obviously, the nunber of sensors can be significantly reduced if only isolated sections of the plant i are monitored. l l l Sensors with a resonance frequency of 375 kHz but with good sensi-l ivity down to 100 kHz were used by ANL. Other sensors may be selected. However, to take full advantage of the information in the acoustic signal, tests with these sensors on an ALD facility will be required to establish their response characteristics and develop curves of leak rate vs. acoustic signal intensity. The considerations in the selection of transducers are the center frequency, bandwidth, ruggedness, response to temperature and 275
1 humidity, capacity of cables and preamplifiers to withstand the reactor environment, and obtaining approval to place materials used to fabricate [ the equipment inside the reactor. A plan to bring out tho' electrical l signals to the control room for analysis by a computer needs to be estab-l 1shed. Feed-through cables can be minimized by use of a multiplexer inside l the reactor. The next step is the selection of waveguides and coupling schemes. Considerations for waveguides are the length, diameter, material, and surface finish. Stainless steel is preferred. The long, thin waveguides minimize heat transfer from the pipe but reduce the acoustic energy trans-ferred to the receiving transducer. Thus a compromise must be made. Waveguides ranging in diameter from 3 to 13 mm and up to 250 mm in length seem most effective. Two coupling schemes seem adequate. In the first, the waveguides are screwed into a plate to mechanically press the waveguide to the pipe outer wall. In the second, a spring loading device is used to press the waveguide to the pipe. Generally speaking, a rounded tip on the end of the waveguide in contact with the pipe is most effective. For averaging of correlogram, three waveguides with transducers will be required at each site, separated by a minimum of 10 cm around the circum-ference. This allows nine correlogram to be generated and averaged for each pair of detector locations. The selection of electronic filters and amplifiers is important. High-gain, low-noise preamplifiers and amplifiers are desirable. Ampli- { fiers built into the casing of the transducer housing are available though ; they have not been extensively evaluated in this program. Preamplifiers
^
i feeding a signal to a multiplexer followed by a quality linear amplifier in the control room seem moct efficient. Selection of filters is also impor-tant. Filters should have variable adjustments so that the frequency of the window of operation can be manipulated as needed. Signals are filtered after amplification and then are transmitted to the analyzing computer. Receivers should be calibrated with a pencil lead breaking scheae or electronically simulated ultrasonic waves.(5) Variations in sensitivity of the numerous transducer-waveguide systems can then be stored in the computer, and thus rms signal intensities can be normalized. Computer hardware and interfaces similar to those described in Ref. 5 can be used. Modifications that include state-of-the-art computer hardware and software will be desirable along with upgrading the system to handle data from more than two waveguide-transducer systems. The system should include off-site monitoring of the rms signal from all sensors and proper interfacing between computer and acoustic signals. Computer programs must incorporate the correction for frequency dependent attenuation if signifi-cant differences in attenuations between 100 and 400 kHz exist. This will ensure that the analysis required to identify the leak source will be carried out correctly. 276
In addition to calibration of sensors,- a permanent self-checking system may be installed. This could consist of a pulser generating an acoustic wave that can be detected by the various probes to check for sys-tem deterioration. If acoustic background noise levels are relatively constant, they may also be used to determine whether a probe is failing. Background noise data from each transducer should be stored for future reference. The ALD system should be validated on a laboratory facility with leaking cracks such as the one ANL used for this investigation. The signal vs. leak rate and the frequency spectrum vs. leak type should be evaluated on the laboratory test loop. Calibration procedures could also be verified on the laboratory apparatus. Tests with field equipment must be carried out to account for differences in receivers. Attenuation vs. frequency can be determined by means of pencil lead breaking techniques or by use of a broadband or variable narrowband electronic source to simulate acoustic waves from 100-400 kHz. In order to determine the attenuation, the-source should be placed 1 and 5 m from the sensor, and acoustic signals should be compared. The equations used for the analysis of frequency-dependent sound waves are presented in Ref. 5. This information is important because it is a factor in making'a decision regarding the source of the leak. Once the source of the leak is established, the leak rate'is estimated from the graph (or equivalent) in Section III.C.3, Fig. 3.29.[5] With the system installed as described in that report, and with the analysis carried out as indicated, an ALD system will be in place that can make low-level decisions and detect, locate, and characterize leaks in LWRs. This system can be set up to monitor the resctor contiaucusly by maans of a computer, alerting reactor personnel to problems, and automatically analyzing acoustic signals. The results of the ANL work on such a system is described cocpletely in Ref. 5. ACKNOWLEDGMENTS Wer are grateful. for the invaluable assistance of R. Popper and J. Killelea, who helped in the design and data acquisition phases of this proj ec t . I 277
l REFERENCES (1) Reactor Coolant Pressure Boundary Leakane Detection Systems, U.S. ! Nuclear Regulatory Commission Guide 1.45 (May 1983). ' (2) D. S. Kupperman, T. N. Claytor, D. W. Prine, and T. A, Mathieson,
" Evaluation of Methods for Leak Detection in Reactor Primary Systems and NDE of Cast Stainless Steel," in Proc. of U.S. Nuclear Reculatory Commission Twelfth Water Reactor Safety Research Information Meeting, Gaithersburg, Maryland, Oct. 22-26, 1984, NUREG/CP-0058, Vol. 4, l pp. 342-362 (1985).
[3] P. Cassette, C. Giroux, H. Roche, and J. J . Seveon, " Evaluation of Primary Coolant Leaks and Assessment of Detection Systems in Contin-uous Surveillance of Reactor Coolant Circuit Integrity," in Continuous Surveillance of Reactor Coolant Circuit Intenrity. Proc. CSNI Specialist Mtg., London, England (August 1985), Nuclear Energy Agency, Paris, France, 1986, pp. 165-178. [4] Report of the U.S. Nuclear Reculatory Commission Pipinn Review Commit-tee (Summary), NUREG-1061, Vol. 5, p. 33 (April 1985) and Vol. 1, pp. 4-52 (August 1984). [5] David S. Kupperman, David Prine, and Thomas Mathieson, Aonlication of Acoustic Leak Detection _Technolony fpr the Detection and Location of Leaks in Linht Water Reactors, Argonne National Laboratory Report NUREG/CR-5134, ANL-88-21 (October 1988). l l 278 l l l l E - - - - - - - - - - - . - - _ - - - - - - - - - _ _
EDDY-CURRENT' INSPECTION FOR STEAM CENERATOR TUBING PROGRAM ANNUAL' PROGRESS REPORT FOR PERIOD ENDING DECEMBER 31. 1988 i a C. V. Dodd, W. E. Dee'ds, and J . R. Pate Metals and Ceramics Division OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37831 l -
SUMMARY
j 1 The steam generator is a crucial part of the boundary between the primary
.and secondary containment of a pressurized water nuclear power plant. The maintenance of the steam generator tubes is essential to the system integrity and a good inspection is important to this maintenance. The reflection probe I has demonstrated an improved capability to detect and. size defects in degraded-steam . generators. We have been optimizing the reflection probes and ' the instrument operating parameters for steam generator inspections. We have developed models that accurately calculate the changes in the eddy-current signals produced by the many property variations. These property variations include tube sheets, tube supports , copper deposits, magnetite deposits , denting, and defects. Other programs compute the accuracy and sensitivity of eddy-current measurements of defects in the presence of these property variations. These programs have been run for reflection, pancake and circumferential coils. The-optimum coil for each of these types was determined. A comparison between the optimum of each type showed that the reflection coils are best, the pancake coils -
next and the circumferential coils are worst. We have designed both single reflection coil probes and an array of 16 reflection coils. The former has been constructed and tested. It gives results that agree closely with the computed values. We improved the coil ruggedness so the probe lasted for many tube inspection cycles in our simulated inspection rig. Since the computation of the signals produced by a defect is important to the above analysis we performed a study of the accuracy of the defect theory for
' reflection, pancake and circumferential coils. Experimental verification was performed for all three probe types and improvements were made in the way that the defect signals are calculated. The studies led to the development of a technique that directly calculates the defect size and depth from the signal produced as the probe scans over the defect. This technique, which we call inversion, was experimentally verified for simple defect geometries. The newest inversion technique shows good potential for a more direct determination of defect size and depth.
INTRODUCTION ! Among the many types of serious tubing degradation are long-known ones, l' such as wastage, pitting, and fretting, and recently discovered ones, such as ( intergranular attack (ICA), intergranular stress-corrosion cracking (ICSCC), and l fatigue. However harmless artifacts, such as tube supports, tube sheets, small dents, and buildup of copper or magnetite, also affect eddy current signals and must be distinguished from the dangerous ones. This can be accomplished by taking eddy current readings at several different frequencies and using the additional information to eliminate the unwanted variables , provided the various artifacts affect the readings differently at the different test frequencies. 279 l
i To optimize the sensitivity to the properties of interest, we must find which combination of test frequencies is best able to determine any particular property and also which probe design is most sensitive to defects in the region of interest. These optimization are performed by. computer modelling, using programs ' developed at Oak Ridge National Laboratory (ORNL) . When the appropriate hardware has been constructed, then the actual test equipment is calibrated, or
" trained," by using it to measure an array of standards containing all combinations of expected property variations. Then the algorithms can be determined that will measure the desired properties and ignore the others.
Finally , - the optimized hardware and computer algorithms are used to make measurements on actual steam generator (SG) tubes. DEVELOPMENT OF BETTER PROBES Figure 1 shows the three types of probes that have been investigated for D i f f erent i a l Pancake Refiection Coil Coi! CoiI
)
=" em EC3M H b b b i
Figure 1 Coil types that have been analyzed inspecting SG tubes. The differential coils are ideally coaxial with the tube. They must be made smaller than the tube inner diameter (ID) in order to pass dents or corrosion buildup. This decrease of fill factor limits their sensitivity , and also allows probe wobble, which makes variable sensitivity around the l l circumference. Differential coils can detect changes in tube properties which occur in distances comparable to the coil separation, but are insensitive to slowly varying changes along the tube. The slow changes can be detected by monitoring the output of one of the coils, as an " absolute" coil. However, l either type of circumferential coil arrangement averages the tube properties -1 i 280 i
around the circumference of the tube and is therefore relatively insensitive to the effect of a small defect at some point around the circumference. Small pancake probes can be pressed against the inner wall of the tube, minimizing lif toff problems while still being able to pass small dents and I corrosion buildup. Since they interrogate a smaller volume of tube, they are ; much more sensitivo to small localized tube variations and are also much less sensitive to artifacts outside the tube, such as tube supports. The penalties are that either the inspection speed is decreased or the probe must have an array of pancake coils to cover the circumference of the tube in one pass. In addition there is more data to store and analyze. A small reflection coil can also be pressed against the inner ' wall of the tube, like a pancake coil, but it has several additional advantages as well. For one thing, it can be made insensitive to liftoff over a finite range; this is useful if the probe has to slide over small dents or a buildup of corrosion products. For another thing, it can be " nulled," i.e. , constructed to give zero reading, when the coil is in air, so that it measures only changes from nominal conditions, which can then be amplified more than the non-nulled signal from a single coil. Both theoretically and experimentally, reflection coils are an order of magnitude more sensitive to small defects than circumferential coils and. have a better signal-to-noise ratio than pancake coils. They are also less sensitive to artifacts outside the tube. 1 The defect sensitivity factor at a given location in a conductor is a measure of the eddy current flaw signal produced by an infinitesimal flaw located there. It is generally a maximum near the coil and falls of f exponentially with distance into the conductor. Figure 2 shows a plot of the defect sensitivity factor for a pancake coil against a flat plate (this is approximated by a very small pancake coil against the inner wall of a much larger tube). Note that the defect sensitivity is spread over a wide area, making it less sensitive to a localized small flaw. Figure 3 shows the corresponding graph for a reflection coil. Note that the sensitive region is more sharply focussed under the coil, so that a defect located there will produce a larger signal change. IE$ i Coll l AXIS i 1
- l i i' l ,
i N 9/ i \ /
}
W l i Figure 2 Defect sensitivity factor for Figure 3 Defect sensitivity factor for a pancake coil above a plate a reflection probe above a plate 281
y _ - _ _ _ - _ - _ - - - F V. , j i e Figure 4 s' hows a . comparison of the accuracy of depth measurements made with .
. pancake and ' refection coils for a set of property variations including defects, l
tube' supports, cop'per coating, magnetite and wall thinning. 0.00'9' - Pancake Probe , O.008 cr _ O - p $ 0.007 - l m u - O.006 - e 3 - (1-W O.005 - Reflection Probe o _ H 0.004 - o - w 0,003 - u. w - o 0.002 - O.001 - 1b 2b 3b 4b Sb '6b 7b COIL MEAN RADIUS Figure 4 Comparison of defec t depth measurement error foc reflection prebe and pancake probe with the same property variations i DEVELOPME11T OF REFLECTION COIL ARRAYS Since reflection coils are superior to the pancake or circumferential
' coils, we have investigated several types of multi-coil arrays to make it possible to scan the entire circumference of a tube simultaneously. The size of reflection coil that has the best compromise between defect sensitivity and coverage ability is one with a mean driver coil radius of 1 mm (0.04 in.),
- l. However an array of sixteen of these coils is needed to cover the inner circumference of a SG tube in one pass with sufficient overlap to ensure detection of every flaw of detectable size. Actually, only eight such coils can be mounted in a single ring inside the tube, and the other eight must be mounted in another ring indexed 22.5 degrees around the axis. A sixteen-coil array was made using pancake coils in Zetec mounts, but it proved too fragile for practical use. More rugged arrays of reflection coils have been made, using fewer than l 282
1 1 sixteen coils, and they have survived torture testing. But we have not yet been able to procure a sixteen-coil array of the more rugged design. FLAW INVERSION, OR DETERMINING FIAW PROPERTIES FROM EDDY CURRENT SIGNALS l j There are many ways to calculate the flaw properties from the eddy current signals. The most straightforward and elegant method is to make a Fourier-Bessel transform of the eddy current readings, which involve the flaw size and depth. By scanning across a flaw, multiplying the readings by the appropriate Bessel ! function and integrating the result, one can obtain a complex number from which ' the flaw size and depth can be extracted. Unfortunately, the method loses accuracy if the flaw has finite size (since the defect sensitivity factor varies appreciably over the flaw volume) or if the experimental data are noisy. Much effort has gone into various schemes for averaging the eddy currents over the flaw volume , with limited success. Figure 5 shows a plot of defect depth calculated by this inversion method from actual experimental measurements versus the actual flaw depth. Note that the agreement is better for small flaws (i.e. , little depth) than for large flaws. Figure 6 shows the results of various methods of averaging the eddy current signal over a large flaw to try to reproduce the actual experimental signal, marked "EXP," as the probe is moved past the flaw. The curves marked S.PT, DEP, and VOL respectively represent the , results of using, the oddy current at a Single Point at the center of the flaw, ! averaging the eddy current signals over the depth of the flaw, and averaging over the Volume of the flaw. _ O. 0 - O 10-4 ii,o
-0. 1 -
gov ed*d
-0. 2 -
- xP l $ r ..
fl 51 -0. 3 -
} ll }
l 4,4 L.) E O
- D. 4 -
% 2,y 1' i ,f
, .c
-0. 4 --0, 3
i
-0. 2 -0.
i 1
- 0. 0 $
2
'W'
' O z
t ACTUAL DEPTH 0 2.2 44 6.6 8.8(Y10*h NORMAtlltD REAL IMPEDANCE i Figure 5 Defect depth determined by Figure 6 Normalized Impedance I inversion plotted against actual defect Change Due to a 60% Farside Flat depth Bottom Hole Another method, which is less elegant, but faster and more satisfactory at present, is to build a lookup file containing the results of theoretical calculations at different depths throughout the wall of the tube. Then, when the tube is scanned by a reflection or pancake probe, the lookup file can be searched for a match with the signals measured by the probe, and the depth and volume of any defects can be found very quickly. Another advantae,e of this method is that we can choose to consider only the interval where the defect 283 I
l' signal is strongest. since this J s the. region with the highest signal-to-noise ratio, the effect of noise on the results is minimized. Other flaw inversion techniques have been described and tried, but all of ( them so far are very sensitive to noise in the data, and some require inordinate amounts of computer time as well. DEVEh0PMENT OF TESTS FOR FATIGUE OF SG TUBES 1 It has recently become apparent that metal fatigue can contribute to failure of SG tubes in aging reactors, and it has been shown that fatigue can l change the electrical and magnetic properties of materials. Therefore we have designed tests to measure these properties on samples of Inconel tubing subjected l' to repeated flexing. To measure the changes of conductivity and permeability most accurately, we have used through transmission measurements, as shown in Figure 7, with the transmitting and receiving coils on opposite sides of the tube wall. 1 l lNCONEL TUBE I [f /_ [f M 'N-
,/ \
/ \
/
AECEIVEA TRANSMITTER COIL COtL f
- m. is
/
'\ [/
'N 'x -
/,
'N N ,f .
\w%_ ./
Figure 7 Measurement of conductivity and permeability using a through .. transmission eddy current test . I Inconel tubes 0.46 m (18 in) long, clamped at one end, were flexed 0.63 J
~
cm (0.25 in) back and forth at the other end by a motor-driven eccentric cam as indicated in Figure 8. Two different tubes broke completely in two after each was flexed about 5.6 million cycles. Tube samples will be tested at different intervals to determine if a measurable change in the conductivity and permeability occurs in Inconel. This kind of failure can lead to catastrophic SG failure, such as occurred at the North Anna reactor site. 284 l l l I
MOTION
~
4 __ A I N CO N E L_ TUBE l CLAMP Figure 8 Diagram of Inconel Tube Fatigue Test SUtOIARY AND CONCLUSIONS The importance of reliable steam generator inspection has increased as the j generators have aged. At the same time, the number and complexity of the known problems involved in steam g(nerator inspections have also increased. The problems have become so complicated that conventional equipment is no longer capable of distinguishing many newly discovered types of defects from other artifacts that may be completely harmless. Therefore it is important to have equipment that can determine the critical properties reliably. Small reflection coils pressed against the inner wall of the tube have significant advantages over conventional bobbin coils or even small pancake coils in their ability to distinguish various tube properties and be relatively insensitive to harmless artifacts outside the tube. Although the reflection coils are more expensive and difficult to construct, they are the most accurate and sensitive eddy current { probes available at this time. Finally, metal fatigue has now taken its place as one of the critical tube properties that should be monitored, i l t 285 i
- - - _ _ _ _ _ . _ _ _ _ __ ._ .______ _ J
y-,, , 6
+
CONTRACT" TITLE-Review of:the' Status of Nondestructive Measurement Techniques
- t. , to. Quantify Material. Property Degradation Due to Aging -
h4 , and' Planning 1for:Further Evaluation ; l CONTRACTOR AND LOCATION , Pacific. Northwest Laboratory a P. . 0.. Box 999,.Richland, Washington 99352 m-PROJECT MANAGER j S. R.-Doctor ;
, 1 PRINCIPAL INVESTIGATORS D. M. Boyd, S. M. Bruemmer, E. R. Green, G. J. Schuster, E. P. Simonen j ABSTRACT-I The materials used in nuclear reactors are inspected periodically during the. service life of the power plant' to detect. degradation that might occur.
. These' inspections. follow the. rules specified in Section XI of the ASME Boiler and Pressure Vessel Code. These inspections are designed to detect service . 1
.induc'ed failure mechanisms. This program is designed not to look at the detec-- ,I tion of defects but the' making of nondestructive measurements.to quantify.the
~
material properties that a defect may reside in or the incipient condition (s)' that may initiate a defect. l This program is intended to provide an assessment of the. technologies that are available to quantify with nondestructive measurements material pro- . perties or material property' changes related to degradation due to aging of -j structural components in light water reactors. In addition, a program plan ci will be developed that describes the work necessary to create adequate engi- l neering data bases for demonstrating and validating prototypic systems for 1 making these measurements. ; The main thrust this year has been an extensive review of literature and an assessment of the technology. The second major activity was the planning of a workshop to bring together 30 leading experts in materials and NDE to discuss the state-of-the-art and to address where future work should go. OBJECTIVE To review the literature, current expertise and related activities, and prepare a plan for development of engineering data bases and validation of prototypic systems for nondestructively measuring material properties and , degradation due to aging effects. O l 286 i i i a ,
FY 1988 SCOPE
= Conduct workshop (s) to gain an understanding of a) the microstructural changes in materials as a function of aging; b) property degradation; c) the nondestructive measurement problem to identify property changes:
fatigue life usage, tougnness loss, etc.; and d) the potential measurement techniques and the correlation between the measurements and material properties.
- Develop a detailed program plan and set priorities for work efforts including the following:
- Develop a statistically designed matrix for aging specimens under temperature, radiation, and stress conditions as a function of time for single and/or multiple combinations of these conditions (as appropriate) and for fatigue testing for important primary system component materials. The matrix of specimens will include and bound the range of variability in material chemistry and properties due to different heats and manufacturers; weldments and base material will be included. Thus, all conditions should be covered but the resulting test matrix will be practical. Develop a priority for the specimens and aging exposures.
- Develop a list of mechanical tests to be performed to quantify mate-
. rial properties before and after various aging exposures and fatigue testing and define size and quantity of specimens for conducting all needed destructive and nondestructive tests.
- Develop a matrix of NDE techniques to be evaluated for each aging condition (s) and for fatigue.
- Assess other research programs concerning NDE and aging as ASME, EPRI, industry, National Research Council in Canada, etc. and factor into this program plan.
- Based on the workshop (s) and technical experts' input, begin developing procedures to age materials in an acceptable accelerated manner that !
simulates reactor aging with respect to microstructure and propet ties. j . From the workshop (s) and a literature review, write a white paper on possible technologies for nondestructively making material property mea-surements. Recommend the most promising techniques for evaluation.
- Design testing procedures, protocol, and sample sets for quantifying and validating NDE baseline and material property measurement techniques. - Begin to identify available materials from other programs that have under-gone different radiation exposures that could be used for NDE measurements.
- Initiate a survey of any material that has already been aged in a manner that is similar to the types of aging in a light-water reactor environment for potential use in this program.
I 287
l' ( o . Define representative fabrication steps for. materials used in light-water
. reactors so that the test specimens prior to aging and fatigue testing
't will undergo the representative fabrication processing steps as actual plant materials; this will consider forming, welding, and heat treatment steps.
- Assemble a team of experts to provide a peer review of the program plan and provide technical input.
SUMMARY
OF-RESEARCH PROGRESS
, 1.0 NDE METHODS FOR MEASUREMENT OF MATERIALS PROPERTIES AND PROPERTY CHANGES Introduction The objective of this project has been to review the literature, current expertise, and related activities; and to prepare a plan for ' development of engineering data bases and validation of prototypic systems for nondestructively measuring material properties and property changes due to aging-in structural components of light water reactors. In order to achieve this task, the scope-of work for the program has been divided into the literature review, preparing.
a white paper on the nondestructive measurements-(NDM) and materials evalu-ation of aging-induced microstructural changes, holding a workshop on the topic, and preparing a program plan for the next phases of the program. This report
- will review the status of the program and discuss the upcoming workshop.to be held on October 27-29.-
' State-of-the-Art NDM
~The inservice inspection (ISI) role requires NDE measurements to detect and size' flaws in order to provide input for fracture mechanics analysis.
The present minimum flaw size is set by the fracture analysis performed on the initial reactor design. The fracture mechanics analysis requires assump-tions about.the material properties such as yield strength, fracture toughness, and fatigue parameters. The goal .of improving the fracture mechanics informa-tion requires use of either destructive or' nondestructive methods to assess the materials properties and materials properties changes. The destructive . techniques of material properties measurements such as tensile testing, fatigue testing, and fracture toughness testing are not always practical or cost. effec-tim. Destructive methods also require making test coupons from either sections of the actual component or aging the coupons in a manner similar to the com-ponent loading conditions and service environment. Nondestructive testing methods-have the potential for obtaining in-situ the needed material property data for input to the fracture mechanics analysis. The NRC has identified that additional information on fracture toughness and fatigue (remaining fatigue life) are two important areas for future development in support of power plant life extension. As part of this program, a library search and literature review has been performed. The library data base system was used to locate references to nondestructive evaluation techniques to measure the materials properties of fatigue life and fracture toughness, when subjected to radiation embrittle-p 288 \ !
t = j 1 ment, fatigue. and thermal. embrittlement. From this' data base, the abstracts were reviewed and~ pertinent papers were obtained for detailed review. One of the papers reviewed was by R. Bruce Thompson and Donald O. Thompson, j
" Ultrasonics in Nondestructive Evaluation," Proc of the IEEE, Vol. 73, No. 12, ,
December 1985, on the use of ultrasonic waves in NDE. One of the sections included 'a review of Material Property Measurements. They pointed out that the microstructure of materials such as grain, porosity, inclusions, micro-cracks, second phase content and morphology, and texture are important to material performance and failure modes in a structure. They also pointed out that each of these microstructural properties affect the ultrasonic velocity and the attenuation. Their concluding statement was "... material property mea-surements are only possible in samples in which only a few of the metallurgical parameters are unknown." However, in the section on material property measure- ; ments, they provided a review of two NDE measurements that have made progress !' in quantitative property measurements. The first was a review of ultrasonic stress measurements using velocity and birefringence methods. One comment on the stress measurements was that " considerable work remains before high reso-lution images of spatially varying stress patterns can be obtained." The second material property technique discussed was grain size determination. Grain size measurements have been demonstrated using scattering of ultrasound and frequency-dependent attenuation measurements. A paper by E. P. Papadakis, " Physical Acoustics and Microstructure of Iron Alloys," International Metals Reviews, Vol. 29, No. 1, 1984, is a review of ultrasonic grain scattering in iron alloys. The effect of microstructure on the frequency-dependent scattering is discussed. The three regions for ultra-sonic scattering are reviewed: Rayleigh Region (aF4 when X >> D Region (aDF2 when X D),andDiffusionRegion(aF/DwhenX((l), 2 D.Stochastic Scattering theory and experimental verification show that acoustic scattering is a function of frequency. He provides an excellent review of research in grain scattering with the salient points as follows: Theory agrees with experiments for:
- 1. Homogenous bulk specimens
- 2. No preferred orientation
- 3. Equiaxed grains
- 4. No twins, intergranular transformation products or precipitates
- 5. Nodular graphite in ductile iron
, Theory does not agree with experiments for:
- 1. inhomogeneous specimens
- 2. wires and thin sheets
- 3. preferred orientation
- 4. elongated grains
- 5. complex intergranular structure ;
The paper then provides a detailed review of specific frequency versus attenu-ation experimental data on the different scattering regimes and iron alloy specimens. 289
The literature review has looked at many NDE techniques which may have potential as a method for material property and property change measurements. The methods reviewed include acoustic elasticity, internal friction, acoustic emission, electrical resistivity, small angle neutron scattering, thermal, mag-netic hysteresis, Barkhausen noise, eddy current, holographic interferometry, and micro hardness methods. This is not an all-inclusive review; however, an attempt has been made to identify the initial techniques which have potential. Table I shows a summary list of potential techniques for NDM. TABLE 1. Potential NDE Methods for NDM Fracture Hardness Yield Toughness Radiation) Fatigue Strength Magnetic Hysteresis X X X I Eddy Current X X Inttrnal Friction X X X X Holog.opJ c Interferometry X X Acou:.ti cs X X X X Acoustic Emission X X X X i This table identifies only that the NDE technique has potential as a method for material property measurements. This information has been used to provide an understanding of the present
" state-of-the-art" of NDE for material property measurements. This information will be used to help guide the workshop and discussion sessions on the future direction for the program.
2.0 MATERIAL AGING MECHANISMS A literature review was conducted to assemble relevant reports and docu-ments related to the aging that occurs in light water reactor materials. The best reference uncovered was a report by the Idaho Nuclear Engineering Labora-tory for the NRC Aging Program (NUREG/CR-4731, Vol.1). This report dealt in great detail with the materials and environmental aging factors that are present in light water reactors. A summary of the ranking and three major degradation i drivers are shown in Table 2. The components that are highlighted with two , asterisks are the ones of more structural significance because of the implica-tion to operational functionality. It is very evident that the reactor pressure vessel is by far the most important reactor component. The second ranked component is the cast stainless steel piping. 290 l 4 l
ry '
/ ,,
l-f ' q TABLE 2. Primary' LWR Degradatiori Mechanisms ,
.(Condensed from NUREG/CR-4731, Vol. 1) 4
- e' Radiation Embrittlement
..t
**- Reactor Pressure Vessel RPV Supports.
fieutron. Shield Tank L Internal Reactor Components i. I' ,
- Thermal' Embrittlement.
**- Cast l Stainless Steel Piping i- Pump Bodies and Elbows-y Ferritic Stainless Steels Fatigue (Mechanical, Thermal, and Corrosion)
~
** Reactor Pressure Vessel Piping and-Nozzles.
Pumps-Steam Generator.
"3.0 WORKSHOP-
- The scheduled workshop is designed to bring together a cross-section of
' NDE and material. experts to. discuss the work and knowledge with regard to' the changes in material microstructure and material property changes that result -
from the light water reactor aging environment. The NDE experts are to be present to. provide them with a basic understanding of.the problem, to place
'into perspective.the work that has been conducted, and to see where the future work should be performed. It was also hoped that material specimens which have already been aged can be identified. This is extremely,important because one of'the-major costs with this kind of program will.be the development of a series-of well-characterized specimens with known material properties and aging history.
The workshop is by invitation only and is to be held from October 27. through noon on October 29, 1988 at the National Institute of Standards and Technology (formerly NBS). The plan is to have one day on materials, one day on NDM, and the-third day on integration and getting out things that attendees had not had a chance to-say before. FUTURE RESEARCH PLANS-The workshop will be held and then the major work to be performed next
. fiscal year will be to write' the report containing an assessment of the NDM-
~ technology and a program plan to. conduct the needed work to develop and validate f ..- techniques to nondestructively measure material properties.
291
FISCAL YEAR 1988
SUMMARY
REPORT TITLE: Evaluation of Nuclear Facility Decommissioning Projects (ENFDP) Program CONTRACTOR: Westinghouse Hanford Company Richland, Washington PRINCIPAL INVESTIGATORS: D. R. Haffner D. H. Doerge ABSTRACT: Major studies have been undertaken in recent years by the U.S. Nuclear Regulatory Commission (NRC) and others, on the technology, safety, and costs associated with the decommissioning of nuclear facilities. The Evaluation of Nuclear Facility Decommissioning Projects (ENFDP) Program described in this report is being undertaken by the NRC to compile and evaluate the -t activities of ongoing decommissioning projects both within the United States ; and abroad. Assessment and evaluation of the methods, impacts, and costs will provide bases for evaluating licensee's proposed decommissioning plans, and for future decommissioning guidance and regulation. This report discusses Fiscal Year (FY) 1988 work scope, research progress made in FY 1988, and future research plans of the ENFDP Program. OBJECTIVE: The objective of the Evaluation of Nuclear Facility Decommissioning Projects (ENFDP) Program is to provide the Nuclear Regulatory Commission (NRC) staff data ~which will allow an assessment of manhours expended, radioactive wastes-generated by type and volume, alternative methods of decommissioning and occupational doses incurred during decommissioning activities. The data provided will also include available cost information to assist the NRC in determining the proper amount of funds which must be available to ensure timely and safe decommissioning of licensed facilities. Additionally, detailed information on waste-characterization from reactor decommissioning projects will provide support for agreement state regional waste compact guidance. FISCAL YEAR 1988 WORK SCOPE: The work effort for Fiscal Year (FY) 1988 was divided into three categories for proper activity definition and cost accounting. Those categories are: o Program Management o Data Collection i o Data Processing and Reporting , 292 l l
l
.u For aldetailed description of the overall program plan and definition of the activities _under the three categories, see NUREG/CR-2522, Rev 2, " Evaluation of Nuclear Facility Decommissioning Projects, Program Plan," submitted for publication September 1987.
Specific; tasks and objectives established for FY 1988 were- i i o Procram Manaaement 'l l
. Prepare the annual summary report for FY 1987 ongoing projects by 4 December'15, 1987.
Revise the ENFDP Program Plan as required. Prepare an. updated Program Summary Status Report for the NRC Division of Engineering Safety, Materials Branch. Prepare monthly summary reports which will present, as a minimum: the progress during the period; work planned'during the next period; budgeted and actual costs by task; milestone status; and cost / milestone variances. j o Data Collection Collect data at TMI-2, Shippingport, Humboldt Bay, Lacrosse, Federal Republic of Germany lingen, Gundremmingen, and Niederaichbach sites and'the Northrop TRIGA facility. o Data Processina and Reoortina Prepare draft status reports on TMI-2 recovery activities. Issue the Project Summary Report for Humboldt Bay decommissioning (contingent upon completion of safe storage preparations). Prepare draft status report on the Shippingport Station Decommissioning Project (contingent upon receipt of adequate data). Prepare draft status report on the three Federal Republic of Germany decommissioning projects following translation of pertinent j information from data collected. i Prepare a final report on the Northrop TRIGA reactor dismantling activities (contingent upon data acquisition). 1
~
I 293 J
i I I l
SUMMARY
OF RESEARCH PROGRESS: Proaram Management The Program Management activity includes routine management, project control, project reporting, and revision of the Program Plan as necessary. A camera-ready copy of the annual summary report for the FY 1987 ENFDP Program was prepared and sent to NRC-HQ in December 1987 (Reference 1). The report included the status of the three ENFDP program tasks: program management, data collection, and data processing and reporting. A draft of the ENFDP Program Summary Status Report (Reference 2) was provided to the NRC Division of Engineering, Materials Engineering Branch, as part of their contribution to a publication on the status of all NRC research programs. The report provides a summary of the ENFDP Program since inception - its objectives and results, current status, and planned activities. Summary reports covering work progress and costs were provided monthly to the NRC Program Manager. Total program expenditures through the end of FY 1988 were $2,068,000. Actual costs of each of the three program work categories in FY 1988 were: PROGRAM COSTS FOR FISCAL YEAR 1988 Task FY 1988 Cost ($000's) Program Management 5 38 Data Collection 19 Data Processing and Reporting 94 TOTAL $151 Data Collection The Data Collection activity includes site data collection, subcontracts or agreements with licensees, and search and recording of reference published data. Three Mile Island-2 (TMI-2) Data collection at the TMI-2 site was ongoing throughout the year'. The L data obtained from TMI-2 for the period April 1979 through May 1986 has been divided into nine categories to correspond to the nine major tasks identified for the facility's recovery efforts. The categories are: 294 l
.-_-_-____-__-___A
p p. l I o Reactor Coolant System and Systems Decontamination o Reactor Building Decontamination o Reactor Defueling and Disassembly o Auxiliary and Fuel Handling Ruilding Decontamination o- Common Support Facilities and Systems Operations o Plant Stability and Safety Activities
- o. Liquid Waste Handling o Solid Waste Handling o Radioactive Waste and Laundry Shipments Data collected at TMI-2 during FY 1988 included information in all nine l categories identified above. !
In 1986, the exposure tracking system at TMI-2 was revised. The first eight categories were realigned into four major categories for exposure tracking purposes. These four new categories for the activities after May 1986 are: o Reactor Building Decontamination o Auxiliary and Fuel Handling Building Decontamination o Balance of Plant (Outside Reactor Building and Auxiliary and Fuel Handling Building) Decontamination o Radioactive Waste Handling The fifth and last category, Radioactive Waste and Laundry Shipments, will remain unchanged. Data collection with this revised work breakdown was initiated in FY 1988.
- Shiocinaoort Station Decommissioning Proiect (SSDP)
Data collection efforts at SSDP were limited in FY 1988 due to the limited release of actual decommissioning data. Although 00E-SSDP0 formally agreed to support the ENFDP Program, they stated data will not be released for ENFDP use until the completion of the various Activity j Specifications. Samples of data to be released were obtained during the one SSDP site visit in FY 1988. Data released for ENFDP use included supporting data for topical reports released on the asbestos removal task and activities associated with:the irradiated component transfer to the reactor vessel. l - Humbpidt Bay-3 (HB-3) Data collection at the HB-3 site is nearing completion as is the site decommissioning work. Final data collection at Humboldt Bay was completed in the spring of 1988 with only minor additions for project completion expected near the end of 1988. L 295
L 3 1 l Lacrosse (LACBWR) Data collection at the LACBWR for the ENFDP Program was initiated in the spring of 1988. Contact persons were identified and a six-month site visit frequency was established for the collection of decommissioning data derived from preparations for the SAFSTOR mode at LACBWR. ff eral d Republic of Germany (FRG) Reactors c The annual data collection trip to the FRG scheduled for the spring of 1988 was delayed due to FY 1988 NRC foreign travel restrictions, but was . rescheduled for October 1988. The three FRG reactor decommissioning projects to be visited include Lingen (KWL), Gundremmingen (KRBA), and' Niederaichbach (KKN). ' Northroo TRIGA ENFDP requirements for the Northrop TRIGA facility data collection and analysis were reduced in FY 1988 to include only the final radiological survey data. An improved survey methodology was implemented at the Northrop TRIGA facility which claims a 90% confidence level. Acquisition of the final survey data is expected the first quarter of FY 1989. Data Processino and Reportina The Data Processing and Reporting activity includes development of computer programs to store and allow further analysis of decommissioning data, development of a standardized reporting format, compilation and processing of site data, and issuance of facility decommissioning reports. Three Mile Island-2 (TMI-2) Analysis and entry into the Decommissioning Data System (DDS) of TMI-2 data collected on site cleanup and recovery activities continued during FY 1988. The majority of efforts involved the plant stability and safety activities, auxiliary and fuel handling building decontamination and waste handling information. Five draft reports were prepared in FY 1988; the final status report on Auxiliary and Fuel Handling Building Decontamination (Reference 3), the final status report on Plant Stability l and Safety Activities (Reference 4), the status report on Solid Waste Handling (Reference 5), the status report on Liquid Waste Handling (Reference 6) and an updated final status report on Reactor Building
]
Decontamination (Reference 7). ! l i l 1 1 296 l
- Humboldt Bay-3 (HB-3)
All HB-3 decommissioning data collected to date, except final preparations for SAFSTOR, have been entered into the DDS data base for analysis and summarization in the HB-3 final project summary report. Preparation of the HB-3 final project summary report, scheduled for publication in FY 1988, was not completed and has been rescheduled following completion of preparations for safe storage, anticipated near the end of calendar year 1988.
- Shioninoport Station Decommissioning Pro.iect (SSDP)
Since the release of decommissioning data from SSDP was limited as described in the data collection section above, a project summary status report scheduled for publication in FY 1988 was not completed due to lack of sufficient information. However, a letter status report on the SSDP project was prepared for NRC internal use in FY 1988. Lacrosse (LAC 8WR) Minimal decommissioning activity occurred at LACBWR during FY 1988 as the licensee was awaiting approval of their decommissioning plan. Radionuclides inventory information has been entered into the DDS data base; however, insufficient data was available to prepare a LACBWR status report in FY 1988.
- Federal Republic of Germany (FRG) Reactors Information on three reactor decommissioning projects in the Federal Republic of Germany (FRG) was obtained from the Department of Energy (D0E) as part of DOE's International Technology Exchange Program. Two of the reactors, Kernkraftwerk Lingen (KWL) and Kernkraftwerk Gradremmingen Block A (KRBA) are being decommissioned to safe storage.
The third, Kernforschungszentrum Karlsruhe Niederaichbach (KKN) is in the final planning stages for dismantlement as a demonstration project. General information on the three FRG reactor decommissioning projects have been entered into the DDS data base. Preparation of a status report on the FRG projects has been, rescheduled to FY 1989 to include information obtained on the October 1988 data collection trip. c Northroo TRIGA ENFDP requirements for Northrop data collection were reduced in FY 1988 to include only the final detailed radiological survey data obtained with an improved survey methodology. This has delayed the analysis and reporting on the Northrop TRIGA project into FY 1989. 297 i
m L FUTURE RESEARCH PLANS: The ENFDP Program will continue through FY 1989. The planned work, as in FY 1988, will be divided into three activity categories; program management, data collection, and data processing and reporting. Specific tasks for FY 1989, subject to final budget approval, will be: o Procram Manaaement Routine management and control of all program activities will continue. Monthly summary reports will be prepared, and as a minimum will present progress during the report period, work planned during the next period, costs by task, and milestone status. These reports will be prepared by the 15th working day for the preceding month. The ENFDP Program Plan, NUREG/CR-2522, will be updated, if necessary to reflect current program objectives and changed scope of the program. An FY 1988 Annual Summary Report for the program will be prepared. The report will summarize the activities for the fiscal year and will include a status of any ongoing data collection effort such as those on the TMI-2 recovery efforts and at the Shippingport Station Decommissioning Project. An updated ENFDP Program Summary Status Report will be prepared for the NRC Division of Engineering, Materials Engineering Branch, for publication as part of their compilation of all research programs within the NRC. o Data Collection Collection of data; i.e., facility design and configuration, operating history, decommissioning techniques, personnel exposure, labor requirements, lessons learned, waste management and project costs (where available) will continue for ongoing projects.
- Data from the TMI-2 cleanup effort will continue to be collected as available. One data collection trip.is planned in FY 1989.
- Data acquisition for the Shippingport Station Decommissioning Project will be scheduled as the information is released by the Department of Energy. No data collection trip is planned in FY 1989.
- Final data collection for the Humboldt Bay Unit 3 decommissioning t project will be scheduled upon completion of preparations for safe storage. No data collection trip is anticipated in FY 1989. !
298
Data.from the Lacrosse preparations for SAFSTOR will continue to be collected as available. Two data collection trips are planned in FY 1989. Data collection at the three Federal Republic of Germany (FRG) reactor decommissioning projects will continue through the Department of Energy's International Technology Exchange Agreement 4 with the FRG. The FRG reactor decommissioning projects are Niederaichbach (dismantle), Lingen (safe storage), and Gundremmingen (safe storage). Since the Lingen facility has completed the preparations for safe storage, only surveillance and maintenance information will be collected at that project, if available. One ; data collection trip is planned in FY 1989. 1 Data acquisition for the Northrop TRIGA facility will be scheduled upon the completion of the final survey data package by Chem-Nuclear Systems, Inc. No data collection trip is planned in FY 1989. o Data Processina and Reportina Data collected from the above-mentioned decommissioning projects will be processed and prepared in the proper format for input into the computerized data base. The information will be arranged into the eleven separate reports dealing with reactor characteristics, personnel exposure, project costs in dollars (where available), manhours by major task, waste volumes, waste characteristics,' shipment of waste including destination, cost, curie content, and container type, and dose rates at various points throughout the facility that may affect the project. In addition, the radionuclides inventory obtained from characterization of n e site prior to the start of the decommissioning project will be placed into the data base and not only serves as a reference for the amounts and types of radionuclides associated with a project, but also provides a measure of the magnitude of the project as compared to other projects. As sufficient data becomes available from each of the nine original tasks, a final status report will be prepared for publication. During FY 1989, a final status report on Common Support Facilities and Systems Operations will be prepared. Depending on the availability of resources, status reports will also be prepared on Reactor Defueling and Disassembly, Reactor Coolant System and Systems Decontamination, Liquid Waste Handling, Solid Waste Handling, and Radioactive Waste and Laundry i Shipments. 299
i l Analysis of TMI-2 data will be directed toward the five remaining major' tasks identified for the recovery efforts. The major tasks for recovery ' 1
'are*
- Reactor. Building Decontamination
- Auxiliary and Fuel Handling Building Decontamination Balance of Plant' Decontamination
- Radioactive Waste Handling
- Radioactive Waste and Laundry Shipments ;
Primary efforts will be directed toward processing the Shippingport 1
-Station Decommissioning Project (SSDP) data, contingent upon data release-by DOE. Of particular interest will be the' waste characterization and disposition from the project. Information obtained on specific tasks already completed will be processed and input into the DDS data base to allow comparison of estimated versus actual values for waste volumes, manhours, and personnel exposure. The SSDP information will be summarized in a. status report upon release of sufficient data.
A final project summary report on the SAFSTOR preparations.at the Humboldt Bay Unit 3 facility will be scheduled for publication upon
-completion of the project. -The report will provide a chronicle of the project with summary data for major tasks. Appendices to the report will contain the detailed information gathered during the project.
A final project summary report on the final radiological survey data of the Northrop TRIGA reactor facility will be scheduled when the data is. obtained from the decommissioning contractor. Subsequent to translation of the FRG reactor decommissioning project reports obtained in FY 1987 and 1988, that information will be entered into the DDS data base. A status report on these projects will then be scheduled for submittal to NRC. A summary status report on the SAFSTOR preparations at the Lacrosse facility will be prepared pending the accumulation of sufficient data.
REFERENCES:
- 1. " Fiscal Year 1987 Summary Report," December 16, 1987.
- 2. "ENFDP Program Summary Status Report," draft report dated July 19, 1988.
- 3. NUREG/CR-4315, Vol. 4, " Summary Status Report - TMI-2 Auxiliary and Fuel Handling Building Decontamination," draft report dated September 29, 1988.
- 4. NUREG/CR-4315, Vol . 6, " Summary Status Report - TMI-2 Plant Stability and Safety Activities," draft report dated March 31, 1988. l 300 i
L i
- 5. NUREG/CR-4315, Vol . 7, '" Summary Status Report - TMI-2 Solid Waste Handling," draf t report dated January 8,1988.
- 6. NUREG/CR-4315, Vol. 8, " Summary Status Report - TMI-2 Liquid Waste Handling," draft report dated January 8, 1988.
- 7. NUREG/CR-4315, Vol . 2, Rev.1, " Summary Status Report - TMI-2 Reactor Building. Decontamination," draft report dated June 23, 1988.
i l I l l-301 _____________y
THE IMPACT OF LIGHT-WATER REACTOR DECONTAMINATION ON SOLIDIFICATION, WASTE DISPOSAL. AND ASSOCIATED OCCUPATIONAL EXPOSURE C. R. Kempf, P. Soo, L. Milian Brookhaven National Laboratory
- 1. INTRODUCTION AND BACKGROUND During operation of light-water reactors, corrosion of metallic components in the primary system occurs. Corrosion products are circulated through the system by the cool ant , and some become radioactive as a result 'of neutron activation in the core. After years of operation, deposition of the corrosion products within the primary system leads to a steady increase in radiation l evel s. This, in turn, causes increasing difficulty during routine maintenance of the plant because of worker exposure to radiation.
Decontamination reagent protocols have been developed for application to corrosion products / oxides in both the oxidizing-chemistry environment of boiling water reactors (BWRs) and the reducing-chemistry environment of pressurized water reactors (PWRs). These different processes will generate characteris-tically different types and volumes of radioactive wastes. All of them involve the use of complexing agents because they fonn selective and strong water-soluble complexes with corrosion products. The decontamination solutions are flushed through anion and cation exchange resin beds after decontaminating the reactor coolant system. This process is carried out to remove excess decontami-nation reagents (chelating / complexing agents) as well as nonradioactive and radioactive ions / complexes. Species that could be expected in spent decontami-Fe+2, nation solutions include the cations Mn+2, g+, Cr+3, Cr+ 6, Fe+ 3, Ni+2, Co+ 2; and the anions NO 3 , citrate, oxalate, picolinate, formate, and ethylenediaminetetraacetic acid (EDTA). Depending on pH conditions, the metal compl exes (metal ion plus complexing / chelating agent) could be cationic, anionic or neutral. This is a consequence of the multiple electron-donating groups on various compl exing agents. These factors make decontamination waste a unique and complicated type of low-level waste.
- 2. RESULTS AND DISCUSSION 2.1 Evaluation of Chemical and Physical Degradation in Decontamination Wastes (Task 1)
The purpose of this task is to determine chemical and physical conditions which could lead to thermal excursions, gas generation, and/or general degra-dation of waste ion-exchange resins used for clean-up at nuclear power plants. This task was initiated as a consequence of concern about three anomalous incidents. These were: a thermal excursion in resins undergoing dewatering at Arkansas Nuclear One (suf ficient heat was produced to bring the temperature of the wastes to at least 365 F); and two gas generation / pressurization events in resin wastes undergoing transportation from Millstone Nuclear Station and from the James A. Fitzpatrick Nuclear Power Plant (gas pressures in the wastes were suf ficient to result in the lifting of the lid of the high integrity container 302
i; L i H l shipping cask in .both cases). In all three cases. resin wastes were involved and the dewatering process had been (or, in the case of the thermal excursion
> wastes, was in the process of being) performed. The resin wastes were quite heterogeneous and had not been. thoroughly characterized. The specific causes of these events could not be identified.
This work will provide information to allow determination of whether such events could. happen in the future, either during storage or processing at the plant, during transportation or at the final disposal site. The plan for Lnis task has involved setting up a simpli fied experimental system in which heat and/or gas generation as well as color changes, precipitates or other signs of , chemical reaction can be observed. Speci fical ly , IRN-78 and IONAC A-365 anion and IRN-77 cation resin batches were regenerated and their moisture contents in regenerated form vere determined. Then, these resins were " loaded" with typical reagents or species that would be expected to be caught on the resins from a decontamination campaign; for anion resins, picolinic- acid and EDTA were used, while for cation resins, ferrous ions were used. The equilibrium moisture contents of these loaded resin forms were also determined. Once batches of regenerated and decontamination reagent-loaded (or, in the case of the cation resins, metal ion-loaded) resins had been prepared, they were subj ected to addition of oxidizing chemicals, in particular nitric acid and potassium permanganate solution. These additions were carried out in several ways: (1) in small increments coupled with monitoring of changes in pH of the resin slurry to allow observation of the exchange with nitrate and with perman-ganate for the regenerated form of the anion resins; (2) dropwise and in bulk to allow observation of the ef fect of oxidizing agent amount and also of the heat generation and absorbance taking place in the resin slurry; and (3) with inter-mittent dewatering by vacuum aspiration between additions of nitric acid and potassium permanganate to simulate the dewatering which ns known to have occurred in the heat and gas generating incidents described earlier. The results of these procedures are given in the following sections. Regeneration, Reagent Loading, and Moisture Content Determinations [Brumfield and Kempf, 1988; Kempf, et al, 1988J The moisture content of the regenerated resins (hydroxide and. hydrogen ion forms for anion and cation resins, respectively) and the " loaded" resins was taken for each resin type as the dif ference in weight between the dewatered (vacuum-aspirated) state and the oven-dried state. Figure 1 shows the results of this determination for the regenerated and " loaded" resins. The average. moisture content for IRN-78/0H- resins was 69.9%; for IRN-77/H+, 55.9%, and for 10NAC A-365/0H- resins, 46.5%. The average moisture content for IRN-78 resins loaded with EDTA was 47.0%; for IRN-77 resins loaded with Fe+2, 47.7%, and for IONAC A-365 resins loaded with picolinic acid, 32.8%. A comparison of values in Figure 1 shows that, for IRN-77, the moisture content of the H+ form is -8% higher than that for the Fe+2-l oaded form. The
+2 charge on the iron means that only one-half as many ions (Fe+2) may occupy the fixed ionic sites of the resin as compared to the +1 charge on the hydrogen ion. This may lead to a decrease in total associated " hydration" moisture attached to the Fe+ 2 versus that attached to H+.
303
I-l l l I g l IRN-77 IRN-78 10NAC A365 (H+) (OH-) , (OH-) {
\ /
Regenerate Regenerate w/1H hcl w/1N Na0H
\ / t Wash with deionized water to pH 7-8 Holsture Content of Regenerated Determine moisture content (Vacuum Oven-drying Aspiration vs 6 days in oven at 9C'C 55.9% 69.9% 46.5%
o LOADING Fe 50g - Te+ 2 Na2EDTA Picolinic w/" Reagents
- Acid I
Wash to pH 7-8 - moisture content 24 hrs in 1 47.7% 47% 32.8% i i Figure 1 Flowchart of Resin Regeneration, fioisture Content Determinations and EDTA, Picolinic Acid and Fe+2 Loading, 1 i 304 l
1 , ;
-l The anion resin, IRN-78, exhibited a moisture content of 69.9% for the OH- form versus 47% for - the EDTA form. The EDTA molecule is considerably.
1arger than the ' hydroxide ion. It is also. capable of existing in a number of ionic states, +2 to -4, depending on the pH [ Peters, et al, 1974].
' Around. neutral pH, the principal forms of EDTA are the -2 and -3 states.
Compared to hydroxide ion (whose charge is -1), two or three times as many 1 fixed ion sites could be occupied by EDTA as by hydroxide . ion. There would f thus be expected to be less total associated " hydration" water with a lower 1 net counter: ion population. Similar results occurred for the 10NAC A-365 resins loaded with picolinic a'cid. The picolinic acid group is expected to have a -1 charge identical to hydroxide ion, however, it is a much larger molecule and may therefore allow accommodation of less associated water in the resin structure than' hydroxide ion. Picol.inic Acid Loading of Resins - The picolinic acid decontamination reagent loading of the IRN-78 and IONAC A-365 resins used for this task has been studied in detail because picolinic acid itsel f has a very low dissociation constant, 6, Under conditions such as these, achieving even a 50% resin loading would-require a tremendous amount of picolinic acid sol ution unless, as is. theoretically expected, the . uptake of picolinate by the resins drives the picolinic acid . equilibrium toward dissociation. The process, as it is thought to occur, is summarized below: (A) Picolinic Acid (aq) *-+ Picolinate (aq) + "(aq) -104 (B) Resin-[0H] + Picolinategq) 4* Resin [ Picolinate] + OH- Ka? The cycle (A)(B)(A)(B) proceeds until the resins have taken up as much l picolinate as they - can; in the process, the picolinic acid dissociates significantly. Ten-gram regenerated, vacuum-aspirated samples were taken of IRN-78 and IONAC A-365 resins. These were equilibrated with two di f ferent concentrations of picolinic acid, one corresponding to a theoretical 100% p loading of the 10-gram sample and the other corresponding to a theoretical 50% loading of the 10-gram sampl e used. The theoretical ' loadings were calculated based on reported exchange capacities of 1.76 meq/ gram of IRN-78 and 5.28 meg / gram of IONAC A-365. Extents of picolinate loading were determined through (spectroscopic) measurement of picolinate remaining in the supernatant above the 10-gram resin samples after equilibration for 1, 2, and 9 days with and without stirring. The longer the equilibration time, the more picolinate was loaded on the resins and the lower the moles of pic'olinic acid remaining in the supernatant. Table 1 provides a summary of the extent of loading of picolinate that can be achieved on IRN-78 and IONAC A-365 resins. These results show that 50% loading may be accomplished to 99.6% and 94.7% completion and 100% loading can only be achieved to 83.2% and 68.4% for IRN-78 and 10NAC A-365 resins, respectively. 305 i
m, p , I 4 j.
~
= Table 1 Uptake of Picolinate for-IRN-78 and 10NAC A-365 Resins Equilibrated with Picolinic Acid solution
- Average . % PA Loaded
, Theoretical Supernatant Total Picolinic on Resins Follow-Type Loading Picolinic Acid Acid Added ing a 9 day Equil E Resin .% (Moles) (Moles) Period 50 3.16 x 10-5 8.77 x 10-8' 99.6 IRN-78 100 2.96 x 10-8 1.76 x 10* 2 83.2 l-
- 50 1.41 x 10-8 2.64 x 10-2 94,7 s!0NAC A-365 100 1.67 x 10-2 5.28 x 10-2 68.4
. Exchange with Nitric. Acid and with Potassium Permanganate IRN-78 ' and 10NAC 'A-365 anion resins were equilibrated with small increments of nitric' acid while the pH was being monitored. The resins were
. originally in the- 0H- form. When nitric acid was added, they exchanged their OH- ions fo r . NO '- ions. The liberated ~ 0H- ions were neutralized by zthe'-H+ from the ' nitric acid. These reactions may be described by the following three equations: -!
HNO 3 (aq) **' N (aq) + NO 3 -(,q) (1)
- NO 3-(resin) + OH-(aq) (2)
OH-(resin) NO-(aq) 3 OH-(aq) + H+(aq) +HO 2 (3) 10NAC A-365 and IRN-78 resins titrated with 3M HNO3 produced the titra-tion curves given in Figure 2. These are similar in shape but different in -l relative position. The N0 3 ions were taken up by both the IRN-78 and the .
. IONAC A-365 resins, while the OH- ions were being released. At the same .
time. . the 0H- ions were being neutralized by the H+ ion .of the nitric acid, .thus decreasing . the pH of both resins. The . shi f t of the titration-curve of the IRN-78. resin to the right indicates that 'the IRN-78 resins are capable of taking on nitrate ion more readily than the 10NAC A-365 resins. The initial pH' of the IRN-78 resins was -13. Tnis would indicate that compared to the IRN-78, the 10NAC A-365 resins were somewhat hesitant about l giving up their OH ions; the IONAC A-365 initial pH was about 9. !
/
306
l
)
l IRN-78 and 10NAC A-365 resins were also titrated with 0.04M potassium permanganate (the initial pH of the permanganate solution was 6.6). The results of this experiment are given in Figure 3. The resins were original-ly in the OH- form. When permanganate ions were added, the resins exchanged their OH- Ions for Mn 09 - ions. The characteristic purple color of permanganate disappeared as the resins exchanged OH- for Mn0 9 . When the purple color of Mn0 9 persisted for several minutes, it was assumed that the maximum amount of permanganate had been taken up by the resins and thus, the permanganate remained in the supernatant layer. These reactions may be described by the following expressions: 4 KMn0q *- K+(aq) + b (aq) b) OH-(resin) + Mn0 3 -(aq)
- OH-(aq) + Mn0g-(resin) (5)
The liberated hydroxide ions were not neutralized in this titration. As a result, the pH of both the IRN-78 and 10NAC A-365 resins increased. The IRN-78 resins remained at high pH for extended periods during these studies. Sorption phenomena are enhanced on these types of resins under these conditions [ Moody and Thomas,1972]. It is believed that some sorp-tion of both permanganate and nitrate ions occurred during these studies, since exchange capacities indicated from the titration curves are considerably larger than expected. Results of Nitric Acid and Potassium Permanganate Addition to Regenerated and to Reagent-Loaded Resins [Kempf, et al, 1988] In preliminary reaction studies, oven-dried and vacuum-aspiration dewatered samples of regenerated (OH- form) IRU-78 and 10NAC A-365 resins were subjected to bulk (5 ml) and dropwise additions of potassium perman-ganate ( -0. 04M) and nitric acid (3M). Thi s wa s carried out in a test tube. The results of these experiments are summarized in Figure 4. I A comparison of the reaction resul ts for oven-dried resins versus dewatered (vacuum-aspirated) resins upon nitric acid addition shows that the drying of the resin does have an effect on the extent of the reaction of the resins with HNO3. A comparison of the reaction results for dropwise addi-tion of nitric acid versus bulk addition of nitric acid shows that the way in which the nitric acid is added also has a strong bearing on the extent of I the reaction. Re sins treated with potassium permanganate reacted to a lesser extent than those treated with nitric acid. Both chemical s are oxidizing agents but a comparison of their relative oxidizing " strength" is not appropriate since the concentrations varied by nearly two orders of i magnitude, i.e., the nitric acid was much more concentrated than the ; permanganate. The next round of experiments involved two new parameters compared to resins were " loaded" with picolinic j these prel iminary + studies , namely: acid, EDTA or Fe 2; and dewatering by vacuum aspiration was performed ] 1 l J l 307 1 i
1 l
)
1 1 13'00 - G ' N IRN-78 and IONAC-A365 Resin
% Titrations with 3M Nitric Acid i
L y 9.00 -
\,\ As i
5.00 - 4 = 10NAC-A365 Resin ! o = IRN-78 Resin I (5-gram vacuum-as trated samples of both resins, orf inally in 011- form) 1.00 .,,,,,,,,,,,,,,,,r.i,,,,, ,,,,,,,,,,,,,i,,,,,,,,,,, ....... . ...... 1 0,00 0.01 0.01 F 0,02 0.02 f 0.03 0.03r 0.04 ! Moles of Nitric Acid Added Figure 2 IRN-78 and 10NAC A-365 Resin Titrations with Nitric Acid. i. 13.00 IRN-78 and 10NAC-A305 Resin Titrations with Permanganate Ion _ _g H a 12.00 - Z a ,-- 11.00 - a = 10NAC-A365 Resin 34- o = IRN-78 Resin A (5-gram vacuum-aspirated samples of both resins, originally in 011- form) 10.00 vr , , , , , , , , , , , , , , , , , , , , , , ,,,,,, ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, .,,,, ,,,, 0.00 E + 000 4.00E-004 0,00E-004 1.20E-003 Moles of Perinatiganate Added Figure 3 IRN-78 and IONAC A-365 Resin Titrations with Permanganate Ion. 308 i
. Ih4Ctloat with EMn0[ and HNO, ,
c a . 1
/
d k) r er a k E p e
~
edh5 Swelled l li"%5"~ed) I":b 5'", re,C.:e"n't4 l1:1.
~0 l l
'io"UdsI'l[ e iUfa) $"Nls'Oened
, n.d ed) - ft) s5 g s e dw re Opw15e)
TOM";di *
@,w e) 10:03;;it".:i" en re opulse)
MNai8/0H* , HMO , traces of smote (oven-drted) (dropwile) Figure 4 Summary of Reactions of Resins with Potassium Permanganate and altric Acid Solutiorut, intermittently'during the study. For example, in a test of. bulk addition of nitric acid to IRN-78 resins loaded with picolinic acid, the resin batch would be dewatered after each of three 3ml acid additions. In the experi-
- mental set-up for these studies, a Buchner. funnel was . set on a side-arm Ehrlenmeyer flask attached to a pump. .This arrangement facilitated addition of the nitric acid and potassium permanganate solutions to- the resin batches, measurement - of temperature changes in the resin during, and it allowed color change or precipitates in the resins and in the eluates to be
. seen_ easily.
A very large number of individual experiments have been carried out, sixty-two to date. These correspond to: e anion resin types IRN-78 and 10NAC A-365, each " loaded" with 50% and 100% theoretical full loadings of picolinic acid and also each loaded to 100% theoretical full loadings of EDTA; the cation resin IRN-77 was given a 100% theoretical full loading of Fe+ 2 from ferrous sul fate solution. e dropwise (total of 150 to 200 drops in three stages) and bulkwise (three separate 3ml) additions of 3M nitric acid and -0. 4M pota ssium permanganate individually and then sequentially, i.e., nitric acid foll owed by potassium ; i permanganate. 309
Table 2 Nitric Acid and Potassium Permanganate Additions to Oven-Dried and Vacuum-Aspiration Dewatered IRN.78 and 10NAC A-365 Resins loaded with Picollnic Acid "E f fects" Observed Added Vapor / 6T color Fumes (*C) Change Precipitate (s) Resin Chemical Ov en-D ried None 15 None White IRN-78 Nitric Acid (NA) Potassium Perman. Some O None White IRN.78 ganate (PP) NA None 9 Cream. White 10NAC A.365 light brown PP Some O None Reddish 10NAC A-365 Vacuun. Aspirated NA some 3 Orange- None IRN.78 licht yellow Little O Light. None IRN.78 PP dark NA Little 5 None White 10NAC A.365 10NAC A-365 PP Little 0 Yellow. None brown e original oven-drying or vacuum-aspiration dewatering of the loaded resins; and e vacuum-a spi ration between additions of nitric acid and potassium permanganate. Observations made in each of these experiments included: (1) resin bed temperature; (2 ) resin slurry and eluate pH; (3) resin slurry and eluate color; (4) precipitate color and quantity, and (5) presence of vapors or fumes. An abbreviated table of results is given for 100% picolinic acid-loaded resins, Table 2. The largest temperature changes (most heat genera-tion) were observed for nitric acid addition to initially oven-dried IRN-78 and 10NAC A-365 resins. Under these conditions, a white precipitate was also observed in the resin batch. Small er temperature changes, less precipitate and more resin bed color changes were observed for initially vacuum-aspirated resins. Potassium permanganate addition lead, in general, to resin color changes, to small amounts of vapor / fumes, and to little heat generation. Similar results were obtained for the EDTA-Loaded resins and for Fe+2 loaded IRN-77 resins, i.e. , the largest temperature changes occurred with initially oven-dried resins and when nitric acid was added. White precipitates were observed for all initially oven-dried, picolinic acid-loaded (50% and 100% theoretical loadings) IRN-78 and IONAC A-365 resins. For the initially vacuum-aspirated batches, only the 10NAC A-365 picolinic acid-loaded resins gave a precipitate when nitric acid was added. For EDTA-loaded resins, a different effect was observed: all of the initially vacuum-aspirated resins showed white precipitates while only one 310
l type of the oven-dried samples did, namely: 10NAC A-365 (50% and 100% theoretical EDTA loadings) when both nitric acid and potassium permanganate- , had been added. No precipitates were observed under any conditions for the I Fe+ 2 loaded IRN-77 resins. Slight resin color changes were observed in a number of cases across the whole spectrum of sample types. The eluate and resin slurry pH values were, as expected (given the addition of nitric acid), quite low: eluates, pH 2.8 to <1; and resin slurries, pH 3.3 to <1. Control tests were run on regenerated resins (0H- form for anion resins IRN-78 and 10NAC A-365, H+ form for cation resin IRN-77). It was hoped that the magnitude of the heat generation contribution could be found from resin hydration and/or neutralization. The results of these tests are given in Table 3. Oven-dried and vacuum-aspirated IRN-78, 10NAC A-365 and IRN-77 regenerated resins were separately subjected to addition of: (1)de-ionized water; (2) nitric acid; and (3) potassium permanganate. The results indicated that little heat of hydration is involved while neutralization heat may be significant. No precipitates were observed, however, on nitric acid addition. This is to be expected since the " product" of the exchange l (and the neutralization prodLcts, simultaneously) is H 20. j From this, it is believed that the precipitates observed on addition of nitric acid to picolinic acid-loaded (or EDTA-loaded) resins were solid picolinic acid (or solid EDTA, respectively). These precipitates will be analyzed to cofirm this. Some of the heat evolved in these systems would have been neutralization heat. Potassium permanganate addition to the regenerated control resin batches had very little effect. Summary of Resin Degradation Study IRN-78, IONAC A-365, and IRN-77 organic ion exchange resin moisture l contents vary significantly depending on the counter ion " loading." For these resins the EDTA, picolinic acid and Fe+ 2 " loaded" forms, respec-tively, had moisture contents lower than the regenerated, OH- and H+
" loaded" forms. Heat- and gas- generating reactions have occurred with two anion resins used, IRN-78 and IONAC A-365; color changes and precipitates were also observed. The resins were originally in the OH- form and a potassium permanganate and nitric acid were oxidizing solutions used to produce the reactions. The extent / vigor of the reaction is very highly dependent on the degree of dewatering of the resins and (probably linked to f
> this) on the method of solution addition (dropwise or in bulk). The heat i generation may be due, in part, to the heat of neutralization (acid addition ; f ! to hydroxide-form resins) [Brumfield and Kempf, 1988]. Ferrous ion loaded ! cation resins (IRN-77) showed little reactivity toward nitric acid and l potassium permanganate. 311
}% l
- s j 5
! Tabic 3 Control' Test Results ,
ji: _ _ . - - Delonized . . ' Potassium Resin- ~ Water , Nitric Acid .Permanganate 4 I
.IRN-78 (00)d None' Heat. Fumes Murky Eluate IRN-78' (VA)b . None. Heat, Color ; Precipitate im Change -(brown-black)-
p 10NAC A-365 (0D) . Heat Heat 10NAC A-365 (VA) None Heat IRN-77 (00) Heat.
!RN.77 (VA) None a 00 = Oven-dried-b VA = Vacuum aspiration dewatered lI I
- 2. 2 Compatibility of Container Materials with Decontamination Wastes -
'{ Task 2).
- This task' was " initiated to evaluate the compatibility of a range of container materials 'with- a simulated decontamination resin waste. The materials include Ferralium 255 (a duplex ' stainless steel), TiCode-12 -(a dilute ~ titanium alloy), Types 304 and 316 stainless steel, carbon steel,. and v'. high-density polyethylene. The carbon . steel' coupons 'were added after- the
' t first irradiation cycle when some of the original specimens were deemed surplus and removed to. provide space. Thus, the carbon steel specimens were .
exposed to resins which had been pre-irradiated to approximately 5 x 10 7 t-rad. The: resin decontamination waste chosen for this task simulates a LOMI
. process waste. The reagents used in this process promote rapid dissolution of surface oxides by changing the oxidation state of the metal ions, e.g.,
Fe(III) to Fe(II). .By definition, LOMI reagents contain 1) a reducing metal ion and 2) a chelating ligand [Bradbury, 1982]. The vanadous - picolinate / formate system is one such reagent which has been successfully applied to full scale reactor decontamination. Because of its . superior decontamination capability and the relative non-aggressiveness of the medium, it is one of
- i. the most important reagents for - present decontamination. The simulated L
LOMI ; resin waste used in this study consists of two volumes of IONAC A-365 anion. resin to one volume of IRN-77 cation resin. The IONAC A-365 is loaded with both picolinate and _ formate. ions whereas the IRN-77 is always in the
- as-received H+ form. The initial moisture content of the mixed bed resin was 47.3 percent by weight. Full details of the resin preparation procedure are'given elsewhere [ Adams and Soo, 1988].
.To check how corrosion is influenced by gamma irradiation (which is present in most' types of low level waste) and by the presence of organic reagents on the resin, four types of corrosion test were initiated:
312 -I 1 i __1______.______
~
Ww (4 a a , i a) corrosion- in imixed-bed . resins. with the anion-component loaded with- picolinate / formate; cation resin in the H+ form; q J
.:b) corrosion -in as-received L mixed-bed -. resins (i.e.,
anion ~ resin in the hydroxide; cation resin in the H+ form;- c) similar to (a) 4but in the presence of a gamma field - of about.1 x 10 rad /h; and ]
. . l d) similar to (b) but in the presence of a gamma field of about 2 x 10" rad /h.
..The four resin beds -were contained in glass vessel s measuring 7.0 cm ID.x 30.5 cm in height. Metallic specimens were placed horizontally in two layers, one resting on. the flat base of the vessel and covered. by resin., and another near the middle of the resin bed where specimens were
~
. contacted'on.both sides by resin.
The . high density polyethylene (Marl'ex CL-100) specimens . were made .from . strips measuring 10.2'x 1.25 x 0.32 cm. They were bent into a "U-bend" con-figuration by bending themL and fastening the two. ends with steel nuts and : b ol t s. 'In .the molding 'of the drum from which the specimens .were. cut, one side:of the drum becomes oxidized by air. When the oxidized material is.on the ~ outer surface .of a U-bend specimen, ' cracks are formed because of the' lower ductility. When the non-oxidized material is on the outer bend surface, no cracking is present. Crack propagation during -testing was studied for samples with both oxidized (cracked) surfaces and non-oxidized
-(uncracked) surfaces on the U-bend specimens. .The polyethylene specimens were placed between the two metallic specimen layers with the apex of each
. U-bend facing upward.
..Th e resin / container material irradiation systems were mechanically sealed so_thatigas generation could be monitored continuously. In the case of the unirradiated controls, the glass vessels were sealed with a " Para-film" plastic sheet.
Gas Generation During Irradiation During the first week of irradiation, the pressures in the irradiated systems containing simulated LOMI resin wastes and the as-received unloaded - resins showed a pressure drop of about 20 percent, after which the pressure began.to increase at a linear rate. The -initial pressure drop is caused by thef -scavenging of oxygen in the original air environment by the resin beads. It :is well-known that. gamma-induced oxidation of polymeric materials Analysis of gases ' throughout the
~
can reduce oxygen levels to low values. irradiation cycles shows that oxygen levels drop to less than 1 percent of the total pressure. The pressure increase is caused by hydrogen and carbon monoxide generation. For the simulated LOMI resin wastes, the H 2/C0 ratio
-is much larger than that for the control resins.
313
Corrosion Analysis At the end of the second examination of the irradiated container materials, a dose of I x 10" rad had been accumulated. No corrosion was noted for the Ferralium and the TiCode-12 which remained bright and shiny. The relevant unirradiated controls were similarly unaffected. Only the austenitic stainless steel and carbon steel showed evidence of attack. Figure 5 shows Type 304 stainless steel coupons which have been irradiated to 1 x 10 rad 8 in the presence of simulated LOMI decontamination waste resins. Two of the specimens were from the top layer in the resin column and the other two were in the lower layer. The spot-type localized attack is usually more pronounced fo r specimens in the bottom layer (see Specimen 13). Thi s is likely to be connected with the higher levels of moisture on resins near the bottom of the column. It was found that additional corrosion spots had been initiated since the last examination which was for an accumulated dose of 5 x 10 rad. Each spot was caused by 7 contact with an individual cation resin bead. Since the resins in contact with the steel were an orange-brown col or, it is speculated that the corrosion mechanism involves the replacement of H+ on the IRN-77 resin by Fe2 + from the stainless steel. A film of moisture at the contact point between the resin and the stainless steel facilitates ion exchange. This liquid is typically very acidic based on work by [Swyler and Weiss, 1981] who found that the pH approached 1.0 for a gamma dose of 10 9 rad. Figure 6 shows Type 304 stainless steel control specimens which were exposed for 412 d to LOMI-resins in the absence of irradiation. The specimens were shiny, with little evidence of corrosion. Type 304 stainless steel irradiated to 1 x 108 rad in the presence of as-received mixed-bed resins (i.e., non-LOMI waste) also showed spot corrosion similar to that described above. This would be expected since the resin component causing attack appears to be the IRN-77 cation resin which is in the H+ form for all four test conditions. Data for Type 316 stainless steel show basically similar corrosion effects to Type 304 However, the amount of corrosion in the former is signi ficantly less as would be expected based on its higher nickel and chromium contents. Carbon steel showed very marked attack for all test conditions. Figures 7 and 8, for example, show specimens exposed to LOMI-type resins for irradiated (5 x 107 rad) and unirradiated conditions, respectively. The depth of attack at the resin contact points is far greater than for Type 304 stainless steel. The most severe attack was for a specimen on the bottom layer of samples which had been irradiated. Apparently, extra moisture and irradiation enhance attack. 1 Attack was also noted for carbon steel exposed to non-LOMI 7 (as-received) mixed bed resins and irradiated to 5 x 10 rad. The attack appeared to be less regular than that shown in Figures 7 and 8, but it was severe in regions of the specimen surface where it was present. 314
E l .:; [.. ~ l . '][',[, (.. . >y -- .
;.p.;
p.... ' . .:_, ?... . '.;./,.
. .) .,. f .." . '
TOP
,l . .
r : . ,w, ' . : ,,. .; . . ' LAYER g,'._;:.*.'.'
> - [. :.. .. '._: -'. T :lg r . . .
' ' *'~
{',;.. .-. :. . . ', al..*j,:_:(:rq, -
. .u-a 1. - }. _ : .
, 7 -.- w T . - ' v,'
4e . ,[ f *. * .*. I *# ' ',
- e. .
y * ~,: .
*, .= "
BOTTOM
, e '... c L;.af
. + . .
b1* ,; , LAYER i
' 4 .
'i' j , .i , .
- l,t j , ya '
. '. i'L; ,
. *, 3 .. .,g,.
Figure 5 Effect of Gamma Irradiation (108 rad) on the Corrosion of l Type 304 Stainless Steel in the Presence of Mixed-Bed Ion-Exchange Resins. Mag. 2.5X. l mme - . _n - ER 1 ; _ c ~ -, -m E r !
. BOTTOM w t LAYER h1 E
6 _m Figure 6 Corrosion of Type 304 Stainless Steel After 412 Days Exposure to Mixed-Bed Ion-Exchange Resins Loaded with Simulated LOMI Reagent. Mag. 2.5X. 315
J I
* 'g ll
.1 m ..
Figure 7 Severe Local Corrosion on Carbon Steel After Irradiation to 5 x 107 rad in the Presence of Mixed-Bed Ion-Exchange Resins Loaded with Simulated L0til Reagent. Specimen was in the Lower Layer of Samples. Mag. 4X.
$';W
- M$,j }-
??((b,'
4: iv ...$ Q; ~::k,~M
'z'
, . r
.yl{rO%..a ';,,
w vi ' . 'j a ~.n; p
.(- r
?# ffRf$' ,.g
~~
Q ,
'T
~
,, wee b;,' i
?h
. h"h .
- : g .
- g. A.
3 - . 2,3 , t' - . . . , - '. ,f ., ' l Figure 8 Local Corrosion of Carbon Steel After 208 days' Exposure to Mixed-Bed Ion-Exchange Resins Loaded with Simulated LOMI Reagent. flag. 4X. 316 l l
, [ '{
l High-density polyethylene U-bend sgecimens were examined after irradia-tion doses of about 5 x 107 and 1 x 10 rad. The appropriate unirradiated ! control s . were examined al so.- The examination . involved carefully studying changes in ' crack patterns at the stressed : apexes of the bent specimens. to check for ' crack initiation and - propagation of cracks that were initially present as a result of the bending. All specimens were in contact with
- LOMI-type resins, as described above. The resul.ts are summarized below:
a) Samples with non-oxidized material on the apex did not show any crack initiation, either for the irradiated or the unirradiated state. b) Samples with oxidized material on the apex also show no additional cracking after an irradiation to 1_x 10s rad over a period of 412 d. c)' Only non-irradiated ' oxidized samples showed. crack growth during test. This lack of. crack. propagation, in irradiated . systems is associated with the rapid . loss of oxygen noted earlier. Degradation of many polymers is associated with a synergistic effect between irradiation and oxygen
;[Gillan and Clough, 1981]. When oxygen is unable to react with . polymer i chains which have undergone irradiation-induced scission, the ductility losses normally expected in oxygen-containing environments become -small or negligible. Brittle crack propagation is then greatly restricted.
p Summary of Container Corrosion- Study The corrosion of carbon steel and austenitic 3tainless steel in mixed bed resins is enhanced by ~ gamma irradiation. Hoe ver, cracking in -high density polyethylene is essentially eliminated because of the rapid removal of oxygen from the environment by~ gamma-induced oxidathn of the large resin mass. Ferralium-255 and TiCode-12 are not attacked by the resins, even for
. gamma doses up to 108 rad. ;
- 4. REFERENCES !
Adams, J. W. , and P. Soo, "The Impact of LWR Decontamination on Solidi fi-cation, Waste Disposal and Associated Occupational Exposure," NUREG/CR-3444, Vol. 5, Brookhaven National Laboratory, June,1988. l Bradbury, D., et al, " Decontamination Systems of BWR's and PWR's Based on LOMI Reagents," in Decontamination of Nucl ear Facil i tie s , International
' Joint Topical . Meeting, ANS-CNA, Vol. 2,1982.
Brumfield, K. and Kempf, C. R., " A Study of the Behavior of Ion Exchange Resins in the Presence of Potassium Permanganate Md Nitric Acid," Informal Report,' WM-3246-5, Brookhaven National Laboratory,1988. 317 i
--__-.u.m__- - - - _ _ . - - _ _ - -a_-__------._-----_---_----------_----,._,-,,-------,-___.a- -_-__x_
y REFERENCES (cont.) Gillen,-K. T. , and R. L. Clough, " Occurrence and Implications of Radiation Dosa-Rate Effects for Material Aging Studies," Radiation Phys. Chem., 18, 679, 1981. , Kempf, C. R. , Milian, L. , Soo, P. , Adams, J. , and Brumfield, K. , "Decontami-nrtion Impacts ori Sol idi fication and Waste Di spo sal ," Quarterly Progress Report, WM-3246-7, April-June 1988, Brookhaven National Laboratory,1988. Moody, G. and Thomas, J., "The Stability of Ion Exchange Resins," Laboratory Practice Volume 21 (9), 1972. , 1 Peters, D., Hayes , J. , and Hief tje, G. , Chemical Separation and Measure- I ments. . Sauders Golden Sunburst, Philadelphia, PA,1974. Swyl er, K. J. , and A. J. Weiss, "Criaracterization of TMI-Type Wastes and Solid' Products," Quarterly Progress Report, January - March 1981, NUREG/CR-2193, Vol.1, Brookhaven National Laboratory,1981. 1 l i 318
1 i ESF SYSTEM FISSION PRODUCT RETENTION EFFECTIVENESS FIN No. 2444 CONTRACTOR: Pacific Northwest Laboratory o PROJECT MANAGER: W. K.'Winegardner PRINCIPAL INVESTIGATORS: L. D. Kannberg and P. C. Owczarski ) i ABSTRACT FY 1988 efforts focussed ~on engineering-scale tests to obtain data for the development and. validation of the computer code ICEDF. This code was developed at the Pacific Northwest Laboratory (PNL) to estimate the extent of fission product retention'in the ice compartments of pressurized water reactor (PWR) ice condenser containment systems. The tests involved the use of a facility capable of passing a mixture of steam-air based aerosols through a vertical test section containing the equivalent of four full-scale ice basket columns typical of those at PWR plants. Initial tests have concentrated on flows having. low steam content or low bulk flow rates. System and test section aerosol particle attenuation efficiencies. greater than 95 and 75%, respectively, have been measured with both-ice and steam present. However, penetration of the. test. section approached 100% upon depletion of the ice inventory. Results i: of testing conducted to date' indicate that aerosol behavior is strongly affected by thermohydraulic conditions. A region of strong vertical thermal stratifi-cation develops in'the horizontal diffuser immediately upstream of the test section upon mixing of the warm inlet flows with the cold air developed by the ice columns. Recirculating flow cells are established in the vertical test section.as warm gases flow up open channels between the ice columns on one side and down similar channels on the other side of the test section. The initial tests have also revealed the need to. reassess the analytical procedures that have been developed for estimating the extent of particle retention. OBJECTIVE The objective of this research is to develop validated analytical models for use in estimating fission product retention effectiveness of light water reactor (LWR) engineered safety feature (ESF) systems. Program planning is directed toward reducing the highest priority uncertainties in severe accident / source term phenomena. Candidate ESF systems include spray, suppression pool, containment cooler, and containment and auxiliary air cleaning systems as well as the ice compartments of ice condenser containment systems. In addition to the development of analytical procedures, the work involves identifying, planning, and conducting experiments needed to validate models. Also included 1 are plans to provide guidelines for system design and operating and maintenance requirements, as well as ongoing efforts to identify information gaps and develop information that will not only identify the most important systems but will permit these systems to be emphasized in future regulatory processes. 1 319
-___.--___--_____ _ _ O
FY 1988 SCOPE FY 1988 efforts focussed on the performance of tests using an engineering-scale facility based on the design of the ice compartments of PWR ice condenser containment systems. Scope of the work included investigations of aerosol particle behavior in the presence of ice including evaluation of the effects of condensing steam and of varying flow velocity and particle size, density, and concentration. Efforts also included tests to better define the convective flow fields associated with the low flow rate aerosol tests. Both velocity measurements and smoke observations were made during a series of special non-aerosol, no-steam tests. The purpose of the above tests is to obtain data that will aid in the development and validation of the particle retention models of the computer code ICEDF, part of the formulation of a systematic, mechanistic approach to estimating the release of radionuclides from nuclear power plants under severe accident conditions. In addition to testing, FY 1988 work included a significant data reduction effort. A number of test facility modifications were also completed including the installation of addi-tional sampling stations to provide a more accurate spatial representation of aerosol mass flow rate.
SUMMARY
OF RESEARCH PROGRESS Nine test runs have been completed using the engineering-scale facility that was constructed to obtain data for the development and validation of the ICEDF code. A summary description of the test facility was presented in the i p: evious annual report (Materials Engineering Branch 1988) and by Kannberg, et al. (1988). In review, the facility test section has a square cross section and contains the equivalent of four full-height [48-ft (14.6-m)] ice basket l columns. The cross section has one whole basket column and four each of half-and quarter-section baskets. The inlet stream to the test section is composed r each ofsteamandhotairmixedwithvarioussolidparticlesasrequiredfg/gec individual test. A total inlet gas flow rate range of 0.03 to 0.38 m corresponds to prototypic plant flow rates of approximately 16 to 200 m /sec. A number of testing capabilities have been developed:
- generation of uniformly-mixed steady-state streams of heated air, super-heated steam, and various aerosols
= introduction of gas / aerosol mixtures into a representative full-height ice condenser test section a measurement of gas temperature, steam content, and aerosol concentration and size distribution at the entrance, exit, and various locations within the ice condenser test section a measurement of the test section gas temperatures and rate of liquid drain-ing from the test section.
! Six of the nine tests that have been conducted are associated with the l statistically designed matrix shown in Table 1. As indicated in this table, three of the " matrix" tests were conducted in FY 1988. FY 1988 tests also 320
I c 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0
- a s r e F T
n
.o 4 s mei 0 4 4 0 0 4 4 0 0 2 0 0 0 2 r al t 2 0 0 9 9 0 0 9 9 4 6 9 9 4 0 e eoc 0 0 0 0 t tMa 0 0 0 0 0 0 0 0 0 0 0 e S r m F a
r a P m w a o
. e l , 0 3 0 0 0 0 0 0 3 0 0 3 8 2 0
r F es 5 3 5 3 8 3 5 8 3 2 2 3 t t/ 1 0 1 0 3 0 1 3 0 2 2 0 3 2 0 S e a3 mR m 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 t u e l l o n V I n i o , t e c r e u x S t a 0 0 0 5 5 0 0 0 0 5 5 5 5 5 0 i t rC 6 6 6 0 0 0 0 0 0 2 2 0 0 2 6 r s e* 1 1 2 2 2 2 1 1 1 1 1 t e p a T m M e T t s e s T y t l t e ai k
. i l s 1 ri 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 a eb - - - - - - - - b E t u al e L Mo ec B S l i A b uy T l t s op c , sm i . ne t c i 8 s n 7 7 = 8 i o 3 3 5 = r/
r C3 4 0 4 0 2 4 0e1 e m 0 7 0 0 0 8 0 7 0 0 0 0 0 0 7 1 t 0
- ,e/
t s/ c sg 7 0 0 0 7 1 7 0 0 4 0 0 7 4 0 sm0 a a 2 2 ,t a1 r M eei a l kdf h b s o C uan l b as e , o ia7 t y sed 8 a t ced/ l i c =ime1 u sc 5 0 0 0 0 0 5 0 5 0 5 0 0 0 0 t 0 1l se/ c n/ 2 2 2 2 l sl 0 i eg 4 2 2 2 2 2 4 2 4 2 4 u a p1 t D : f mm r y oe a t = ccc - P i i i l 1 msi at s n 0 0 0 0 2 0 8 0 5 0 0 0 2 0 0 b ns 2 0 0 u: yed D, m 0 0 7 0 2 0 9 0 2 0 0 0, l ndt e A 2 2 0 2 4 2 4 2 1 6 2 2 4 6 2 ooo t M 1 1 1 1 1 1 1 sird e t eel M l canp aa im e ) i r=l o rf rC t) ) 0 ) a4 5 1 2 e D e c( ( ( e( t eMd = t s . i l 2 3 4 c i5
- sacMn eMI AU*
eo 1 2 3 4 5 6 7 8 9 0 1 1 p1 1 1 1 t TN 1 o o e
* *
- R N N M"
included. sample probe calibrations that provided additional particle attenuation
- data as.well as separate' investigations to better define flow patterns and velocity profiles. Information developed from the test data is presented below in terms of particle characterization and retention and thermohydraulic effects on particle ~ behavior.
Particle Characterization Aerosol test procedures-include' generation and characterization. Aerosols L .are, generated by' dispersing' powders in an energy mill or. atomizing liquid l solutions.in an ultrasonic nozzle. Aerosol generation procedures are driven , by. aerosol solubility, concentration, and particle size distribution require-u. ments. Aerosol characteristics are measured to provide information about' test' conditions, particle size distribution, and aerosol mass concentration-and mass flow rate'at the test system inlet, outlet, and within the. ice con-
. denser test section. These data provide measurements for determining the extent of system and compartment particle attenuation. Tables 2 and 3 list the aerosol source material and generation method for tests completed and planned. Results of particle size distribution measurements are listed in Table 4. Most samples indicated the aerosol to be' log-normally distributed; thus, .the measured mass-median aerodynamic diameter (MMAD) and geometric standard deviation (GSD) of the aerosols were used to describe the aerosols.
Samples were typically obtained twice during tests, at about 25% and 75% test completion periods. Samples were usually obtained at the inlet (Station 6A) and the bottom (Stations SA and/or 68) and the top (Stations 1A and/or 1B) of the ice columns. (The location of aerosol sampling stations is shown in Figure 1.) With the' exception of Test 9-7, tests with' steam showed significant par-the ice condenser. The flow rate in ticle growth as the aerosol passed Test 9-7 was very' slow (abcat 0.06 m up/sec), particularly after most of the steam had been condensed in the diffuser and the lower test section. The resulting increased residence time of particles in the ice condenser is specu-lated to have been sufficient to favor deposition of large particles, thus ! decreasing the MMAD present at the top of the section. Tests without steam provided no indication of particle growth. Particle Retention As suggested above, the purpose of the tests is to obtain data for the development and validation of numerical aerosol deposition models. The emphasis of aerosol sampling and characterization has been to determine the decontamina-tion (DF), the ratio of particle mass flow rate entering to that leaving various regions of the test facility, including the vertical test section. DF values developed from data taken during " matrix" tests are listed in Table 3 for the inlet section (between Stations 6A and 68 or 5A), the vertical test section (Station 6B or 5A to 1A or 18), and the overall test system (Sta-tions 6A to 0). Review of Table 2 will reveal that DF is substantially l increased during tests when steam is a major component of the inlet gas. l furthermore, a large portion of the DF occurs in the inlet diffuser, especially for tests with a large fraction of steam in the inlet gas. Test section DF 322 i i
1; TABLE-2. Aerosol Generation Techniques and Materials for ICEDF Tests I Test No.- Generator Primary / Backup Material 1~ EM CsI/- 2 EM SiO2 3 'AN Fluorescein /SiO2 4 EM SiO2/ Fluorescein l , 5- EM Nacl / kcl 6 EM KC1/ Nacl /Uranine 7 EM ZnS 8 EM SiO2/ Fluorescein i 9 AN CsI/- 10 EM kcl H AN ZnS I Material Formula Material Formula Primary Secondary Alumina Al 203 Calcium carbonate' CaC03 l l Cesium iodide Csl Cesium hydroxide Cs0H Potassium chloride kcl Magnesium oxide Mg0 Silica SiO2 Tellurium dioxide Te02 Sodium chloride Nacl Tin Sn Titanium dioxide TiO2 Zinc chloride ZnC1 Zinc sulfide Zns Fluorescein C20N1205 1 i EM = Energy mill AN = Atomizing nozzle i Underlined tests are those completed to date (note that in Table 1, Test 15 ) has same inlet aerosol conditions as Test 2) 323 ,
t F s D 5 3 o 1 A i . y t N e 2 9 3 8 u S ~ 3 l 1 2 '3 a
~- v e
g s a n 3 r o 8 1 8 4 e i 5 v t t n F 8 8 1 2 a a s 8 v e e r e T ioc D t 8 i i 1 8
+ +
2 A N 1 1 e r s e f b S 1 2 3
. O 1 4 e l
o n . ng o a s i r o 4 t r e 5 0 3 1 4 ia de A 8 8 8 v v t F 1 1 i e r s n e D i 1 i d e o l
+ A s t i n N 2 d b 4 4 8 8 r o s t c
I 2 1 a r. e l e 4 5 1 1 d n e
~_
T o s o S t a ise s u t D 8 8 4 3
- g r e
s e S C - 2 1 s c n f f 3 2 2 2 c a n A T 1 D / / S D
/
3
/
2
/
1
/
8 ~ 9 / 8 i i r r i d
. i a P iA p r t 8 5 9 t a n S . S 4 1 e v i u I m
o d d D , e e gt e r D 7 7 2 1 8 3 r d S C - 2 3 3 2 2 2
= is u e
r 5 D // S D A p
/
1
/
7
/
4
/
8
/
6
/
7 D S L 5 . cc
- o u ta P e 8 9 2 3 3 3 C
s 7 g S 81 s , n r o d n n i a , e i a l e D S 8 7 3 8 t t e a 4 t t M D // C - 3 3 2 2 m t a s
- e 8 s A / / / /
p i 1 S D N 5 s P A 8 7 8 3 d g n o
- l a
t 1
> 9 5 3 3 ici n
lp tss le a b m e r i t a e T e n, 5 n a t i le n
- e 8
2 8 8 3 8 2 3 d y s n e s id ni 2 3 3 o s d I o j 1 1 8 r u e n C a m 1 i 1 i 1 w o n - r / i e io t C o s t g 8 8 8 8 2 9 a r e s n e 5 8 b C a e 1 5 7 1 7 8 5 7 n a . W c 1 a v d s 1 1 i d t e et f e a l l n o
~
l m a ie a e t s t s t y i r l O 2 l S S s a a r n s r t e C K i S C n n l a e i e a a E Z Z C
=
w s e a t W lo w ar m u D f A A s O u e - 8 c S l n 1 s n n c e a o g W i , e io a 8 8 4 n t r a t 4 7 8 o l a o
. r s c o t f e a a 8 8 8 8 0 8 i t s s 3 v e r ~ ~ ~ u o n A t F b r o S e g n i E ir a i t L t lp ! v e, s r B e o A e u r i fo d a s s e
- g t T a a C e 5 r
c r r
- 1 9 5 4 3 5 ir d A t i
e v p e 9 8 2 8 2 4 s n 1 n lu A e 1 1 1 1 e a e d a r T e lc 4 dn ic a i s f h y n ti e s o r r A f B u c s s s a e, i a u n r C t i p ig t o ev e a s 1 1 7 4 8 8 d = F h T e R / g 3 3 8 3 8 1 2 5 1 5 1 5 8 n o D e i t a l' a a r se a c S e t 8 8 8 8 8 8 P S S t e l
) o v F A e ne n A 8 . l s s. b e o n ion lue eu la tw d i be stn l
l a o 9 1 2 2 3 t a a a e 7 8 7 8 8 a v v v i t c it i I c t a F e 8 8 8 8 8 8 S d d Dm n a ( e e t e I r g g e m t F t a a y e - s
- e r r t s e e t s a ln v v o y e 2
I A A N S W M5 3 4 5 8 7
- - - - - - ) ) ) ) ) )
_ 2 8 7 1 9 a b c d e f 1 1 1 ( ( ( ( ( ( _ wN
u f-r
?
'il Particle Size' Distribution for. Comp 1eted Tests at Inlet, Bottom',
~
- TABLE'4. ~
and Top ofl Test.Section. , q
.' Particle Size: MMAD(a) [pm]'(GSD)(b)-
Y Ice Columns
~ Test _ Ice - Steam- Inlet Bottom' Top
' 10-14 Yes. Yes' NM(c) ND(d)' '6.0.(2;6)>
~
l NM 3.8-(2.9) 9.0 (2.4) 15i2 !No- . No NM - 6.2' (2.8) 6.0-(2.7) NM 6.1 (2.6) 6.5 (2.6): 2-3 Yes No! >15 .(-) 12.1-(4.5)' 5.7.(2.7)-
>15 '(-) 7.3~ (3.0) 4.8 (3.1) 10-4 Yes Yes 7.9 (3.4) 2.9 (2.4). 9.3 (2.6).
11.6.(3.9) 1.9 (4.1) 9.0 (2.2) 7-5 'Yes No 5.4 (3.7) 3.8 (2.1)- ND 6.8 (4.0) 3.8 (2.2) 4.0 (2.1) 6 Yes- . Yes 3.3 (2.1) ND ~9 (-) 3.9 (2.5) 3.5- (2.8) ND 9-7L 'Yes- Yes 3.3 (2.7) 3.8 (2.3) 1.6 (2.3) 3.3 (2.4) 3.6' (2.3) ND a MMAD - mass. median aerodynamic diameter b GSD geometric standard deviation c NM - not measured d 'ND - not determined approached unity (complete particle penetration) as the ice inventory was
' depleted by st'eam. In addition to the " matrix" tests, sample 3 robe calibrations were conducted with monodisperse particles, in the' absence of )oth ice and steam. Test.section DF values developed from information taken during these calibrations are presented in Table 5. In combination with " matrix" tests, results of.the calibration tests indicate that there is very little retention in the vertical test section upon depletion of the ice inventory.
g i 325
, w , ' w-1;n ; i =re i
Station 0 - -!
; _ _lce Loading
.j To Scrubtier +c=t"*%, -___l EDoors- ~
Station 1B - [ ' o , i-"
-07 : t ..
j
, Platform e Bypass- Quad Station 1 A.
-Level b / \ 004 . t. (1 A3 & 1 A4 on East Side)L
] ".,.,t f 21 -y , . -g.
t
. .' .y J
Aerosol
~. - Station 28 :f Inlet: 0 t > Hot Air ands p,.
o-Steam inlet- \ i D~,- .
- Station 2 A '
, ,, m.- , ... _ _
~'"
l i 21 '
- 00 : Station 3B
- g. ,
-Probe Panel E(W e'
'i; ' '
O *-M, Station Station 48 3A-Mixing Chamber-+ I ae-Station 4 A b: __
; . ,, Station 58 -
0 Windows
/' , , Test Section i
. :. y' U
'g -Quad Station SA 2 -
(SA3 & SA4 on East Side) [ . T1 1 ~ T2 T3 T4 - y Station 6A W. ! *..i
'l g' yj 021 0
- Station 6B
.:a l1 # I';
*; S { w
.,t . . . . ;':
l 20'-O" Dia. " 'l FIGURE 1. Elevation View (West Side) of Test Facility Showing Locations of i' Aerosol Sampling Stations 326
i 4 E' (
. TABLE Si Summary'of DF for Monodisperse Aerosols 0ver the Length 'of the Ice Baskets from Station 68 to 18. (System flow rate = 0.15 m3/sec, no
. 'ce,-no.
i steam) Aerodynamic Particle. Size, Number of Range of. Data',- , um Measurements , DF _ aDF ! 1
'4.6- 2 1.017 *0.001; h(2 11.4 6 1.107 *0.016 15.0 7 1.155 *0.106 To' establish an independent measure of accuracy for the tests, an aerosol-material mass balance was conducted for Test 10-4. The aerosol masses asso-L > ciated with generation, mixing, deposition in the test section, meltwater and condensed steam (measured together), diffuser, and exhaust scrubber were deter-mined and compared with the estimated DF determined from the aerosol samples.
' A~ material mass balance was achieved accounting for 93% + <13% of the material -
introduced into the test. system via the generator. A DF of 14 1 2 was computed based on the mas's balance between the inlet and outlet sampling stations. This is similar to and lends validity to the DF of 13 to 17 (varying with time during the test) determined-from the' aerosol samples. Thermohydraulic Effects on Particle Retention / Measurement
'Results of testing' conducted to date indicate that particle behavior is strongly affected by thermohydraulic conditions in the test facility. More
.specifically, as a result of the mixing (during low flow conditions) of the
- steam and/or hot air with the cold air supplied from the ice columns, a region of strong:. vertical thermal stratification develops in the horizontal diffuser installed immediately upstream of the test section. Recirculating flow cells are established in the vertical test section as warm gases flow up the channels between the ice baskets on one side'and down similar channels on the other side of the test section (see Kannberg et al. 1988 for additional details concerning diffuser and test section temperature and velocity profiles). The j;
I measured ~ presence of downflow in 'one or two of the flow channels.under low l~ inlet flow rate conditions necessitated the installation of additional aerosol sampling stations--particularly the establishment of " quad" sampling stations
, at Levels 1 and 5 (near the top and near the bottom of the ice columns, respec-
'tively, see Figure 1). Quad stations differ from the other stations located along the ice condenser because they include four sampling nozzles, one in each open flow channel; other stations have only one sampling nozzle located in.the southwest flow channel.
Selected details from two FY 1988 tests, Tests 7-5 and 9-7, are presented in the remainder of this section. Observations of significant temperature differences between various flow channels and downflow in the south flow chan-nels near the bottom of the ice condenser for low flow rate cases suggested 327 _ _ - _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _O
that test section aerosol distribution might not be uniform, at least at lower i elevations. Figure 2 shows aerosol concentrations in each flow channel measured during Test 9-7 at quad stations near the bottom (Station 5A) and the top , (Station 1A) of the ice condenser. Greater aerosol concentrations were measured ! in the north flow channels at Station 5A; however, near the top of the ice i condenser (Station 1A), the concentrations were nearly uniform between the four flow channels. This is consistent with earlier observations of gas velo-city profiles and orientations (although for somewhat different conditions), , and, to a large degree consistent with measured temperature differences. An ' important result of these data, although perhaps only applicable to tests .i with low gas flow rates, is that measurements of DF determined between locations , of the test section other than the lower region of the ice condenser may be performed using aerosol data from only a single flow channel. Using aerosol mass flow rate data from the lower portion of the ice condenser to calculate DF requires that aerosol mass concentration data include all flow channels 4 and flow profiles be defined. ' Aerosol mass flow rate data for Tests 7-5 and 9-7 are provided in Figure 3 and 4 as functions of time and location within the test system. Of importance l is the substantial decrease in aerosol mass flow rate that occurred between the inlet and the bottom of the ice columns (between Stations 6A and 4B or 58,respectively). This DF is speculated to be caused by settling and possibly diffusiophoresis. Substantial particle growth, enhancing deposition by sett-ling, is suspected due to steam condensation. In addition, the presence of ' cold reverse flow in these locations provides an unknown degree of mixing and, possibly, increased particle residence time. As no aerosol sampling was per formed in the inlet diffuser, specific aerosol behavior can only be specu-lated based on inlet and outlet conditions, on temperature profile measurements indicating the presence of reverse flow, and on measured decreases in both bulk gas temperature and steam mole fraction. 0.020 0.004
$E $ #
[ g 0.015 - g 0.003 - g Level 5A j - Level 1 A { ! g SA 1(Sw) 5 _ E 1 A 1 (sw) E 0,010 - , 3 SA-2(NW) 0.002 - ( _,. . g 1A 2 (NW) b ' D SA-3 (SE) d 8 0 1 A 3 (SE) y
~
O 5^-4 (NE) O 1 A-4 (NE) If 2 0.005 - g 2 2 0.001 - . .
< E
~
@ gj : s 0.000 O.000 22.5 57.5 22.5 57.5 Time, min Time, min FIGURE 2. Aerosol Concentrations (mg/m3 ) in Various Flow Channels for Two Time Intervals at Levels 5A and 1A for Test 9-7 328
r 7, . L 3.- 1000 .
. m :-
G E Test 7-5
- Sta 6A E- ; Sta 4B
.h 100 -
- 1. --
; Sta 1B h.- h- -
{ Q 3 3.-- Sta0 2_. g m E e-10 .
. i i i 0 20 40 60 80 100 Time, min FIGURE 3. Aerosol Mass Flow Rate Through the Test Facility as Measured at-Various' Sampling Stations During Test 7-5 1;
s. m O
.1 - " + i Test 9-7 g
h a.-- Sta 6A 3: , Sta5A
+ *
{ .01 1 a Sta 1 A M ; Sta1B y a
+
3 -- Sta0 o .001 : m ' 8 e
.0001 . . i e 0 20 40 60 80 100 Time, min FIGURE 4. Aerosol Mass Flow Rate Through the Test Facility as Measured at Various Sampling Locations during Test 9-7 329 l
I Aerosol mass flow rates at other locations of the test system are also 'I shown in Figures 3 and 4, including average values measured at Quad Stations j
.5A and 1A during Test 9-7. Neglecting Station 6B, which is located within (and may have been influenced by) the area of probable flow reversals, a general decrease in aerosol mass flow rate is seen with increasing elevation in the ice condenser during both tests [the cause of a measured increase in aerosol mass flow rate between the top of the ice condenser (Station 18) and the test , . system outlet (0) during Test.9-7 is not known].
FUTURE RESEARCH PLANS ; Future research plans focus on the development and validation of the SPARC and ICEDF computer codes. As indicated earlier, the latter code was l developed at PNL to estimate the extent of fission product retention in the 1 ice compartments of PWR ice condenser containment systems (0wczarski, Schreck and Winegardner 1985). The former, SPARC code, was developed at PNL to estimate , attenuation in boiling water reactor (BWR) suppression pools (0wczarski, Postma and Schreck 1985). Validation efforts associated with the SPARC code will involve comparison of' code calculations with existing data from aerosol particle pool scrubbing tests sponsored by the Electric Power Research Institute (EPRI) at Battelle Columbus Laboratories (BCL) and at PNL. Data from the former EPRI-BCL tests, ! pool scrubbing investigations using a vertical downcomer vent geometry, supple-ments information that is available for other large vent geometries, speci- , fically horizontal vent arrangements (Materials Engineering Branch 1988). Data from the latter EPRI-PNL filtered containment study are the first from tests using multiple, closely spaced orifices. Initial model data comparisons were completed in late FY 1985 using SPARC code calculations and data from EPRI-BCL single orifice tests (Chemical Engineering Branch 1986). Calculated-values for particle retention were lower than those measured during the experi-ments, often by orders of magnitude. Underprediction of the extent of scrubbing was partially attributed to inadequate modeling of entrance effects, speci-fically the lack of a code subroutine describing internal circulation inside forming bubbles. Use of the information concerning large vent / suppression ( pool systems coupled with the expanded data base should result in significantly l improved models for the various exit geometries. Heat as well as mass transfer will be considered. To date it has been assumed that the gas instantaneously attains thermal equilibrium with pool water in the immediate vicinity of the entry point, irrespective of vent geometry. Data from the EPRI-PNL tests should provide valuable insights concerning interactions of bubble swarms from multiple orifice vent configurations. It is also hoped to expand the SPARC code validation effort to include data from the planned Italian test program SPARTA. Planned vent systems for the latter Ente Nazionale Energie Alternative (ENEA) tests include a full-scale x-quencher arrangement. 330
l I l Future research plans also include the development of more sophisticated particle growth models and algorithms that can be incorporated into the ICEDF as well as the SPARC code. Current models are based on a steady-state approach , using the van't Hoff equation for soluble particles and an independent unsteady- l state integration of the Mason equation for growth in supersaturated atmo- l spheres (Fletcher 1962, Byers 1985). Because Csl and Cs0H are expected to be ! the major soluble (as well as hygroscopic) components of nuclear accident aerosols, a single unsteady-state integration of Mason's equation for each particle size group is needed recognizing the vapor pressure of the particle's solution. The present algorithm for integrating Mason's equation may not be appropriate. The recent release of data concerning the vapor pressures of cesium solutions should facilitate this effort. Nine of the individual tests of the planned test matrix (Table 1) to obtain data for the development and validation of the ICEDF code remain to be conducted. FY 1989 plans include the completion of four to six additional tests include those with the highest volume flow rates and steam mole fractions. Performance of these tests will require the leasing of a larger boiler. The initial tests have revealed the need to reassess the analytical procedures that have been developed for estimating the extent of particle retention in the ice compartments of PWR ice condenser containment systems. In the initial version of the code, estimates are based on a material balance developed assum-ing either perfect mixing or unidirectional flow. However, the test conditions, i.e., the mixing, using low flow rates of hot air and steam with the cold air developed in the ice columns, have resulted in complex convective flow fields. It is now planned to use the COBRA-NC computer code in an effort to provide calculations, and possibly even subroutines, that can be incorporated into the ICEDF code to provide improved representations of temperature distributions and air and water vapor flow patterns (Wheeler et al. 1986). REFERENCES i Byers, H. R. 1965. Elements of Cloud Physics. The University of Chicago Press, Chicago, Illinois, i Chemical Engineering Branch. 1986. Compilation of Contract Research for the I Chemical Engineering Branch, Division of Engineering Technology, Annual Report for FY 1985. NUREG-1215, U.S. Nuclear Regulatory Commission, Washington, D.C. Fletcher, N. H. 1962. The Physics of Rainclouds. Cambridge University Press, London, England. 1 i Kannberg, L. D., M. W. Ligotke, B. A. Ross, E. J. Eschbach and J. M. Bates. l 1988. Particle Capture in an Experimental Ice Condenser at Low Flow Rates. PNL-SA-16138, Pacific Northwest Laboratory, Richland, Washington. 331 i
- - _ _ - _ _ - -- _ a
Materials Engineering Branch. 1988. Compilation of Contract Research for the i Materials Engineering Branch, Division-of Engineering, Annual Report for FY-1987. NUREG-0975, Vol. 6, U.S. Nuclear Regulatory Commission, Washington, D.C. Owczarski, P. S., A. K. Postma and R. I. Schreck. 1985. Technical Bases and User's Manual for the Prototype of a Suppression Pool Aerosol Removal Code (SPARC). NUREG/CR-3317, PNL-4742, U.S. Nuclear Regulatory Commission,
.; Washington, D.C.
Owczarski, P. C., R. I. Schreck and W. K. Winegardner. 1985. ICEDF: A Code for Aerosol Capture in Ice Compartments.. NUREG/CR-4130, PNL-5379,.U.S. Nuclear Regulatory Commission,. Washington, D.C. Wheeler, C. L., M. J.'Thurgood, T. E. Guidotti, and D. E. DeBellis. 1986.. COBRA-NC: A Thermal Hydraulics Code for Transient Analysis of Nuclear Reactor
, Components. NUREG/CR-3262, PNL-5515, U.S. Nuclear Regulatory Commission, i Washington, D.C.-
l I 332 l l l
y ? , g Contract
Title:
- Radionuclides Source Term Measurements for Decommission
, Assessments.(B2880-7) ~y ? Contractor:- Pacific-Northwest Laboratory T Richland,-WA 99352-Principal Investigator: David E. Robertson Abstract: 1 L This paper summarizes work conducted' during 'FY 1988 to provide detailed radiological characterization of reactor decommissioning materials for enhancement:of the technology and safety associated with the decommissioning of. nuclear power stations. During FY 1988, radiological analyses of.
decommissioning ~ materials from the Shippingport Atomic Power Station have continued-to document the concentrations, distributions, and inventory of
; residual radionuclides.in plant systems. These measurements have.shown that cobalt-60 is the-dominant radionuclides, and that there is a. virtual absence of.significant amounts of fission products and transuranic radionuclides in
- the decommissioning waste. ' Assessments have shown that e'ssentially all of
!the piping, components, and other waste materials (excluding pressure vessel internal' components) could be disposed.of as Class-A type waste. . In related studies of the radionuclides distributions in highly neutron activated metal components associated with spent fuel assembly hardware from three commercial-nuclear ~ power stations, it was shown that the nickel-63, nickel-59, and niobium-94 concentrations were frequently over the Class-C. limit. .It was discovered that niobium-93m, a 13.6 year halflife activation product, was present in concentrations over 1800 times higher than previous calculations.
had predicted._ During this work it was also possible to conduct separate detailed predictive. calculations of the concentrations of iron-55, cobalt-60, nickel-59,. nickel-63, and niobium-94.in the same samples of spent fuel assembly hardware.. A comparison of measured versus calculated radionuclides concentrations were_ generally in good agreement (within 10 to 50 percent).in the fueled regions of the assemblies,- but the' differences became auch greater
.near the ends of the assemblies. The discrepancies were attributed to uncertainties in the ca_1culational' methods. These comparisons are providing an assessment of the accuracy of predictive calculational techniques and identifying where the significant errors are introduced. During FY 1988 it was possible to conduct measurements of the radionuclides concentrations in t- several specimens of the pressure vessel from the decommissioned Gundremmigen
. reactor. These measurements showed that the Gundremmigen pressure vessel itself (excluding the internal hardware) contained below-Class-A levels of activation products. A comparison of measured versus predicted radionuclides
[ . concentrations in the pressure vessel at the sample locations were in good agreement.
. Objective:
The objective of this project is to provide an up-to-date regulatory assessment of the radiological factors, criteria and problem areas associated with the technology and safety pertaining to reactor decommissioning and a 333 I
r L related waste disposal. 'This is being accomplished through a measurements and appraisal program focused in the following key areas: Radiological characterization during Shippingport Station
! decommissioning
- Radiological characterization of greater-than Class-C wastes (highly activated reactor internal materials)
Evaluation of the accuracy of predictive activation codes and methods Assessment of decommissioning waste disposal options FY 1988 Scope: I This study comprises two main research areas associated with reactor decommissioning: 1) providing a detailed radiological characterization and ' assessment from the actual complete decommissioning of Shippingport Atomic Power Station, and 2) conducting a detailed radiological assessment of the ; highly neutron-activated metal components associated with reactor internals i and spent fuel assembly hardware. l The scope of work outlined for FY 1988 involved mainly an intensive radiological analysis program which encompasses three main research areas:
- 1) radionuclides source term measurements for decommissioning assessments utilizing components and materials removed from Shippingport Atomic Power Station during its dismantlement and disposition, 2) radionuclides characterization and waste classification of highly radioactive neutron-activated metal components removed from spent fuel assembly hardware and pressure vessels from commercial nuclear power stations, and 3) empirical testing of computer codes for calculating radionuclides inventories of highly neutron-activated metal components from within pressure vessels.
The radionuclides source-term characterization of dismantled Shippingport components include measurements of all long-lived radionuclides associated with the primary and secondary coolant systems, radwaste systems, and fuel pool systems. In addition, acquisition of activated stainless steel and Zircaloy materials from the fuel assembly hardware from Core 3 has been , accomplished. Measurements of all 10CFR61 radionuclides are being made for ' each of these materials and compared with predicted radionuclides concentrations calculated by state-of-the-art computer codes. The data base of radiological measurements obtainable through the Shippingport Station Decommissioning Project technology transfer program will be appraised to provide a complete assessment of the radiological conditions associated with the entire station, including other slightly contaminated materials such as concrete surfaces and equipment. The radionuclides characterization of neutron-activated highly radioactive metal components removed from the pressure vessels of commercial nuclear power plants will provide an assessment of the radionuclides concentrations and waste classification of radioactive materials which may exceed Class-C 334
Y limits for disposal. Well-documented materials associated with fuel su3 port structures and other highly activated reactor internal components have 3eer, ; acquired from commercial nuclear power stations for conducting complete 10CFR61 radionuclides analysis. The empirical measurements will be compared with calculated radionuclides concentrations in these materials to test the i state-of-the-art ability to predict radionuclides levels in activated metals, j l The ultimate utilization of these measurements will be to gain a better understanding of the radionuclides source terms in nuclear power stations so that more effective assessments of the technology, safety, and costs for decommissioning can be made.
SUMMARY
OF RESEARCH PROGRESS The above objectives are being accomplished through an intensive sampling and analysis program. This program is currently in full-phase and involves radiochemical analyses in both hot-cell and laboratory facilities. Specimens 4 of surface-contaminated and neutron-activated com3onents from Shippingport < Station, from three types of spent fuel assembly lardware, and pressure vessel samples from Gundremmigen KRB-A reactor are being subjected to extensive radiochemical analyses. These materials will provide the basis for ! evaluating the radiological safety and waste disposal options associated with reactor decommissioning. EXPERIMENTAL Shippingport Station Radiological Characterization from a radiological standpoint, the decommissioning operations at ; Shippingport Station have been extremely successful and have provided an i optimistic and positive projection for the ultimate decommissioning of commercial reactor stations. One of the most significant observations at ShippingportStationwasthefactthatessentiallyallofthgresidual radionuclides were neutron activation products dominated by Co. No significant concentrations of fission products or transuranic radionuclides were associated with the residual activity. This would be representative of the commercial nuclear power stations which have experienced little or no fuel cladfing gilurgs during,'their operations. Althoug Co, the their activation products Fe, Ni, Ni, and Nb were present with the combined concentrations associated with the radioactive residues in piping, plant components, and other waste materials (excluding the pressure vessel internals) never exceeded the 10CFR61 Class-A waste limit. Although the Shippingport Station is a DOE facility and not subject to the regulations contained in 10CFR61, the ramifications of the residual radioactivity levels in decommissioning wastes are of significance. First, it suggests that commercial stations having similar residual radionuclides inventories and distributions can expect to dispose of essentially most radioactive decommissioning materials and components (except reactor pressure vessel ' internals) as Class-A waste. Secondly, this will greatly simplify the disposal methods and the dismantling options during decommissioning. 335 _ __ ____________________________-___-________O
e
-Another vanguard operation at Shippingport Station was the methodology developed by DOE and its subcontractors for characterization, packaging, shipment, and disposal of the reactor pressure vessel and internal components as an LSA,' Type-B package conforming to Department of Transportation and Nuclear Regulatory. Commission regulations. The physical, chemical, and radiological characterization conducted by PNL of the radioactive corrosion film contained on the inside surfaces of the reactor pressure vessel and internal components showed that this material was extremely cohesive and would not be released under a variety of hypothetical severe accident conditions during transportation to the disposal facility at Hanford, Washington.
Other important radiological " lessons learned" during the decommissioning of l Shippingport Station as they apply to commercial stations are being assessed I and will be presented in the final report for this project. Radiological Characterization of Activated Metal Components During the past year this work has involved the radiological characterization of activated metal components from three commercial fuel assemblies, and characterization of steel specimens from the Gundremmigen reactor pressure vessel. Particular emphasis has been in measuring and assessing the significance of the long-lived activation product radionuclides specified in 10CFR61 (see Tables 1, 2, and 3). This work has shown that the relatively highnickelandniobiumcontentgfIngnel,andthenickelcontentof stainless steel has resulted in Ni, Ni, and g Nb concentrations in some I fuel assembly hardware components being ever the Class-C limit (see Tables 4 and 5). This would require that these components be disposed of in a high < 1evel waste repository. It was discovered in this work that the concentrations of
- Nb, a 13.6 year i half-life activation product were present in the activated metal specimens at levels over these 1800 are times To the in best of our j thehigher thanmeasurements previous calculations,Nb activated knowledge, first actual of metals. This radionuclides decays by emission of a 30-key gamma-ray which is essentially all converted, and the predominant external radiation is due to the 16-key Nb x-rays. This radionuclides has not even been considered in 10CFR61, and its long-term environmental significance will need to be assessed. i During the radiological characterization of the fuel assembly hardware it was possible to conduct separate detailed predictive calculations of radionuclides ,
concentrations in the same mageriah A gompagison of ghe measured versus ! calculated concentrations of Fe, Co, Ni, Ni, and 'Nb in the fuel gssembliegshowedquitegoodagreementinmostcases. Examples ofassembly the hardware Ni, and Nb comparisons for the Westinghouse assembly are shown in Figures 1, 2, and 3. The agreement between measured versus calculated values for these radionuclides in hardware from the fueled region of the assemblies were generally on the order of 10 to 50 percent, and never exceeded about a factor 336
0 5 9 8 1 1 9 6 4 - E E t E E) E
) E) E) )
- 7) 9) 7 7 2) 3 7
7 1 8 6 3 b 1 6 3 1 M - o 0 0 0 4 ! o 0 ! - t t 2 9 t t ! ! 9 9 9 8 O 4 0 7 ) n o 3 0 7 9
'. 1 4 0 7
( 4 ( 1 ( 6 ( 1 ( 4 ( 6( 2( it a t S 2 2 2 2 2 2 2 2 2 2 h 4 3 c - - - - - E E E a E) E) E E) E) E E) E) ) ) ) E) B e 7 2 1) 1 4 7) 9 8 0 8 0 3 0 1 0 2 0
, 0 0 0 0 0 0 0 t c 0 o 0 0 0 0 0 0 0 0 0 n 0 0 ! : t 6 t 2 !
i 7 7 2 1 3 5 8 3 4 3 0 1 o 3 6 3 3 0 4 8 0 0 0 2 3 P ) 1 ( 7 8 3 1 ( a 7( 1 ( 1 ( 1 ( 4 ( ( 8( 8( ( ( ( ( l s l l a a i t r e 2 2 2 2 3 3 3 2 2 2 e m 3 4 t E E E E a 4 E E) E) E] E) E) E) E) ) ) W / ) 3 4 4 2
)
4 8 8 8 8 4 4 3) 1 0 0 0 0 0 0 0 0 0 0 e C i 0 0 r ( q o 0 0 3 0 0 0 0 0 o 0 0 0 a n 6 ! 2 ! t t ! w 5 6 1 9 0 9 9 4 5 6 d o 0 1 9 5 4 8 9 8 7 7 8 7 5 0 r it 2 3 H a a r 2 ( 3 ( 3 ( 1 ( 6 ( 8 ( 8 ( 7 ( ( 8 ( 3( ( t y n e lb c m n 4 4 4 4 4 5 5 e o 5 6 5 5 4 s C - - - E E E E E E A s E) E E) ) ) E) E) ) E) ) ) 1
)
1 1 1 3 4 4 3 3) 1 1 0 1 0 l i 0 0 0 0 0 0 0 0 0 0 e g ot u 9 0 o 0 0 0 ot c t 0 0 0 0 F 5 t 7 6 5 5 8 0 8 6 1 2 5 0 0 1 6 3 5 5 5 2 7 3 3 4 . t b n 3 2 3 1 1 1 e 1 ( 1( 1 ( 7 ( 2 ( ( 3( ( ( ( ( ( p S nN) e 4 s 2 2 2 2 2 2 2 2 5)'n. u 3 3 2 2 o - - - E E E E pn h E) E E) E) E) E) E) E) ) g ) 2 1 5 5
)
5
)
5
)
5 2 4 4 nb.( n 2 4 0 0 N r e 0 0 0 0 0 0 0 0 0 0 (e3 r9 it 0 0 0 0 0 0 0 0 0 o 0 0 s 5 5 ! ! ! ! ! ! t 2 4 5 n e 4 7 7 9 1 3 7 5 2 4 5 1 o W 8 6 8 2 3 5 2 4 9 1 7 2 ni n 5( 6 6 6( 3( 4 4 iot c i 1( 3( 2 ( 1 ( 5( ( ( ( ( t a s ce ar n e o re 3 3 3 3 3 3 h it 5 4 4 3 - - nt a r E E E E E o E E) E) ) ) E) E) ) ) r y t n ) 8) 8 9 4 4 1 3) 6 8 6 3 3 3 6 9 tb u 3 - 1 2 3 3 e b 0 4 E E ed c ( 3 5 1 0 2 2 ! 0 0 ne n n 3 0 2 ! ! l. ! : c o 4
- 9 9
t 0 2 6 0 2 ! 2 6 t u sd C 5 7 3 7 3 C 5 5 7 9 5 < 5 2 a o e 3 3 2 f r 5 8 3 5 1 6 3( ( ( ( ( p id ( ( ( ( ( e y l c 1.htl t u 8
/yn n g g 8b a io g n n / n 9 g 0di d
a g i n n) i m 3 2 1 itt i t t 1 em co R n t t o 7 6 5 4 f ud
# # # 8 i n i t tt # 8 #
f fi oce or o rp f i i t f o d d d d d d d erp I i i i i i d d E t a s d d( b i i r g r g irg r g g r r g r g n e n e) t p a s L c n n n m d eg so B A l o wd on d e e r) ri n r e r e r e r e r e c r e c e c r m o) t p ot t t e isib w T d ep et c c c c t o p po ps a a a a a a a t o r%3% a pa p p p p p p p ot ob loo et v( uc s s s s s s s b( b( h5 9 c ht s i . . d eb F N o l l l t ss la l l le e le e ii ir len e e e n d n n n n n n et t o o o o o o o e o c c c c c c c t na t c n n n n n S 5 cee a nI S S S I n I n I I I I I S 5 emm M ? reel rl oee c t t yn n
. a ee o crr N l ea a 0 2 6 5 4 3 2 I 1 DPP 1 1 9 8 7 le - - -
W W W W W W W W W p P W W ))) m a (abc(( S dN l
* ? ? " " ""?
MM MM%%MEMW 2 R. *w* =. N. N. O. Q n. O. 2 ?w* "wT"w? f.
'~
S2 .
- g. . g ~v .E. .? .2 2g .. S a S .S.
g.
. ~ ~
' " v "I u* *.,
~ ~ . , ~
n -". w w ..e w w . w w w e
#88%
g . . ~ . . . -e y g aaa a a aaaw a O RRR E G 2 2* RR 0; 9 "! *: " 1 . " 1 i
. [" 2 2" S ! ?. ! I t S I E SE E l t R M N 3 G *O M f $5$ 5 $ $5$5 t S~ . e e .e .
5 $ Y v F $ 2 n..$Y SS ! ~.2 ! ! I E I .$. m. , $n.$ $. Ei S g S m n=n s a = - m e
." 5$5 5 $ $$53 0
* ; . . m S
...... ~~~
y '; aaaT
---w aaaaaaaaaa ---. -----
.e.
E I3 o e C. ". A. R S. G. G. G. Z. G. S. B. 2 2 S2 2 S **
. ** . ,Y S. E. O. g 2 T -R 3. ~g 2 . R. . G. . .
30 2 2 2 E Y 5 E t. O E O O O O O E 5
- S g 1
- a "
I= ? ? ? ? ? ? * *
- 7 ? 1
; E E2% % W EEEEWEEE s
% b x g4- G. E. G. 2 :. S. R. 3 0 : M.
- E. G. =.
_ k 7.
- g T.*mems ~ *. " S g Eg Ii#.W e S gg
.2 2* . *
=
- e . . . ". . . . #. ". E O*. OES*. ECS555OE af M o
c e ==.e m e m 5.: eaaaaw 9 =a a= a= a= ay w aa 8. m
;; gg2: crc1 R==aRg=~. .fy. 5
-ge .~9 -: -: 1 9 " ~ o., M t
, S ER
. Sy23SRSRES E ?. SS S S S R EG N e
* * ~
T '! E 5,o
$5i$555555555$ 0?F"5
@ M m N . et M M M M M M N se c aaaaaaaaaaaaaa De v.
---8.
.,2 8 9 E. E. E. E. G. E. E. a. G. S. G. *E to a : E E ?, ?, S EES S S S S S ?.
3 CC DG2 RG0 3 : E E G E"I E 3 si5a55555s5555 15l
.E ,- ,,. ... n n 2B aaaaaaaaaaaaaa *goS E
, _o o, G n.R n.R E.E. ~o R. E8 ~. 7R
. . . o. RRR~ ~. ~. 3 ." W g g .,
M Sh hh 8 $hh$ L e
- f . n. a. e. e.
. . .. E. m. =. =. e. =. m. .-n.
v w bw BOO .w b U bw NU s =i C N O
- k. O c .4
*=
Y N N
-8 5i ? ~" F. 'tQS C'
g g "*3
'i .2 O 2 2
- b. Ff. sg E ge-b' 2 6
'g,j g -% ;l gE 1
n22c _h WE , ,,,,, ,
, t*st 33t t _ 3 p-
.a .::
6 6 2 k b b b k k tv .E. E E 36 v v v v S .m -*' . y b O v. N n
. N N v. ee - e n .=
R 2 2 R R R . R22 2: n 691 w i 2 - - ', E g : ~%
< O 0 &&&a- ~
T.
- t i 2 2 2 . . ~
E E E E E.EE E In g nacccnn=n_fnn 2;___ d
-au
'f' "f 'f'
=
m m . o . %CCEtt g 4ru eeeeeeeeee"eae
? ? ? T *
?? ? ?.*
e
?? ? $"2:2
. v.
338
l i ill ll] 1 6 5 3 3 3 3 3 3 3 E E E I )
- 7) E) El E) E] E) ) E] 5 8 4 4 2 l 1 2 2 1 2 4 3 - 3 0 0 0 0 0 0 0 - -
lh E E E t l 9 ot 0: o o 0 0 0 10 7 3 5 2 1 3 2 1 4 4 6 1 tt 4 2 0 t 1 3 2 7 2 3 3 2 2 2 1 2 4 1 4 1 1 1 1 1 ( ( ( ( 1( ( ( ( ( 7
- 7. 5 5
5 6 7 E E E E
) ) E) E) E) )
c 6 2 5] 9 2 4 5) 2 6 1 7 2 1 5
) - -
n (g 01 o o t t 0 2
- - - - - 0 2
0 2 0 2 io 3 9 9 4 0 4 8 9 3 6 0 5 6 8 4 8 t a 2 1 3 2 1 6 4 t ( ( ( ( ( ( ( S r e p 0 1 9 7 9 9 9 8- 9 o 1 E E E E - o 1 ) ) E) E) } E) C ( 6 3 9 3 8 9) 5 9 l s M n y 2 1 1 2 6 et e t 9 9 5 1 l t 9
- - -- 2 t
4 t 8 6 t 3 l 7 t 9 o 4 t a 6 8 9 0 1 8 9 0 i r 7( 3( 5 5 ( 9 ( 9 ( 7 ( 6 ( e ( t i s l 4 4 3 4 3 3 3 4 4 4 2 3 e r E) E) E) E) E) E E) E E) E E E
)
a I I 3 6 5 1 1 0 1 0
]
7 8 2 0 7) 0 2 0 3) 0 7] 0 7 0 w I 0 0 0 7 me8 l d a 0 ot 0 0 0 8 0 0 0 s 0 r t 6 2 2 2 2 1 2 t t 2 a e 5 1 8 2 8 0 8 2 8 4 3 H m 9 4 8 4 6 2 6 3 6 2 2 0 e 5 5 1 2 1 9 1 7 4 2 6 7 ly f i ( ( ( ( ( ( ( ( ( ( ( ( b C m - b e s e 4 4 3 5 4 5 4 4- 5 - 5 3 3
- 5 c E E E E 5 s i ) E1 E) ) E) E) E) E) ) E] f) ) 2 A t 5 0 4 7 2 9 O 0 2 1 6 3 1 a , 9 9 0 2 1 1 2 5 3 7 1 l
r 3 !. o e t 3 0 ot ! 0 ot 0 0 o ! 0 o 0 t u n 6 2 t 2 2 t t 2 t 2 y F e 1 6 8 0 7 0 9 t 6 3 4 5 8 a c 0 8 4 1 0 1 3 9 0 6 2 t n c o 8 8 9 3 1 8 1 1 7 2 7 1 e n C ( ( ( ( ( ( ( ( ( ( ( ( d e p a t s e 6- 6 5 7 6 7 7 6 7 T 5 5 b icr E E E E E E E) E) E
, y
) E) E) E) ) ) ) ) b t , 5 8 8 4 6 1) 1 6 1 8 3 3) .
c e 9 5 3 5 c 6 ot e 6 3 0 0 1 3 3 l 1 1 o 0 0 8 2 1 l 1 et 4 Md. e m l E t t 1 2 t 2 t 2 2 t I
)
e 8 0 3 7 7 6 1 5 5 6 5 4 l. 6 0 7 7 0 5 4 3 4 1 1 1 l 3 le a 3 5 4 2 1 8 9 1 5 2 8 1 ( ( e f r ( ( ( ( ( ( ( ( ( ( ( e .. e n n S e Z 3 3 4 3 3 3 1 3 3 l 2 G - - - - - - - - - - - E E E E E E) n E1 E) ) E) E) E) E) )) i l 8 0 8 3 8 3 5 1) 8) 6 7) 9 0 1 t 'a . 7 0 3 0 5 0 0 1 2 0 0 0 s e c .n aa( n 5 02 c t0 0 0 0 0 0 0 0 02 0 e a i o 5 3 4 9 7 2 2 4 1 4 2 9 5 1 7 1 8 1 3 5 2 2 2 6 4 2 5 5 1 5 3 7 2 2 ryS t a 8 2 8 2 1 3 3 2 2 1 1 r l( ( ( ( ( ( ( ( ( ( ( ( t nn t n . moo eii e 4 4 5 4 4 4 4 4 4 ntt cc c - - - - - - - - - taa n E E) E E E E) E E E 4 see o ) 5 ) 3)
) ) ) )
0 3 arr C ., 0 3 2 3 E
- 5 9 5 4 9 3 7 3 2 2 3 7 E E f ee e 4 5 3 8 ehh 5 0 0 0 0 0 o o 0 9 htt d 1 2 9 1 2 2 t t 2 2 4 5 i
0 5 < 3 8 9 4 8 2 1 < < 2
.t yy l 1 1 5 5 4 1 6 0 0 8 yb b c 2 2 /b u 1
( 1 ( 3 ( 1 ( 1 ( ( 3 ( ( 1 ( 1 dd n /dee o 5ecc i d ) fuao cun a t g odoo R a )m gi n g erpp orr ( tp o nt n) ss 9g a a 3 n e nnii t t tais itn dsii E i o e a rp t 7 8 6 5 8 4 3 2 1otcte
# 8b i i ei b5 L t lp 'sl o ff f e B a ta d d d d d d d# o l 5 2
A c n o nu r eri r i ir i ri r i irg nonz d dz T o a g g g g gn eo eo e3 ? 1 c t h of g g it i i t n s n sf r r r r r r rt m m i . .. o re ano ec e c c ec ce c eca eta er ta or or deba fnS b p pp a a a a a a at te te o a p xo Sp So p Sp S Sp Sp S( ps on on tsss T U Et S( S( iii d ettt tnnn y y y y y y y ceee l a l e o o o o o o emmm reee t e l l a a a a a l l l ta l a rlll r n oeee e o cr c r c r c r cr c r c r c t c i i i i i i i 5 5 ttt a 5 s n 5 5 ynnn M 5 i i Z Z Z Z Z Z Z aeee crrr eaaa e DPPP i 9 8 7 5 3 l n 1 1 1 1 1 I 9 7 5 3 1 2 m E E a E E E E E E E E E E G E S G E E G G G 6 G G G do l l I ! i . i prj( Ill I,hI[l Lllll<
l E h-
- g. ,
1 q 1 V Table 4 . Average Concentrations-of 10CFR61' Radionuclides in-
." -Spent fuel Assembly Hardware Components-y Average Concentration (C'I/m3)
, Material 163Ni 59Ni 94Nb
. Westinghouse . .
h1 Upper end fittings- 55-304 1.34E4 5.83E1 - 8.69E-3 Bottom end fittings 55-304 2.61E4 9.80E1 ~1.07E-1
~S pacer grids inconel-718 5.63E5 2.13E3 .4.77E2 Upper holddown spring . Inconel-718- 2.30E4 1.00E2 6.28E-0 Combustion Enoineerino Upper holddown & flow plates- 55-304 7.09E2 3.24E1 1.90E-2 Bottom end fitting & retention plate $5-304' '8.78E4 3.09E2 8.13E-2 Spacer. grids Zircaloy-4 3.02El- 1.21E-1 1.46E-1 Upper holddown spring -Inconel-625 3.40E4 1.17E2 2.42E1 ,
Bottom spacer grid Inconel-625 3.83E5 -1.39E3 1.92E2-General Electric ' Upper handle & tie plate $5-304 6.66E3 3.43E1 1.84E Upper expansion spring Inconel-X750 8.05E4 4.01E2 5.06E-0
= Spacer grids .
Zircaloy-4 3.45E2 4.87E0 6.83E-2 Bottom end fitting 55-304 3.48E4 '3.67E2 3.39E-1 10CFR61' Class C Limit 7.0E3 2.2E2 2.0E-1
. TABLE 5'. Ratio of Measured Radionuclides Concentrations in Spent Fuel
' Assembly liardware to Their 10CFR61 Class C Limit Ratio: Measured Concentration / Class C limit Material 63Ni 59Ni 94Nb Westinghouse-Upper end fittings $5-304 1.91 0.27 0.043 Bottom end fittings $5-304 3.73 0.45 0.54 Spacer grids Inconel-718 80.4 9.68 2390 Upper holddown spring inconel-718 3.29 0.45 31.4
)
Combustion Engineering Upper holddown & flow plates 55-304 0.10 0.15 0.095 .a Bottom end fitting & retention plate SS-304 12.5 1.40 0.41 .1 Spacer grids Zircaloy 4 0.0043 5.5E-4 0.73 l
. Upper holddown spring inconel-625 4.9 0.53 121 Bottom spacer grid inconel 625 54.7 6.3 960 i i
-J General Electric !
Upper hanole & tie plate
$5-304 0.95 0.16 0.092 Upper expansion spring inconel-X750 11.5 1.8 25.3 k Spacer grids Zircaloy-4 0.049 0.022 0.34 l Bottom end fitting 55-304 4.97 1.67 1.7 i 340 i
l
]
___ _ i ]
Westinghouse Fuel ~ A'ssembly, 14x14 Ratio: Predicted /.fq Samale Measured .P r_e_dic t e c) , Cl/g Co 3 [I gp Num aer Cl/g Co Measured )
- 8. 7 8 E- 1 /;;_ly"dY
' \ W10 - (1.04io.13)E0 0.844 1 jkrN l
s 1.02 E 0 gll 'g. W12 - (1.0910.11)E0 0.936 3.04 E0 ::: :: : : W9 - (8.72i0.87)E0 0.349 '
~ . ~ ,
9.9 9 E 0 -
. W8 - (8.85i0.89)E0 1.13
.s H Hil l l H 1 H i-Hi 6.65E1 U$[{ W7 - (4.57io.46)E1 ' 1.46
_ IlDIuns ! 8.4 7 E 1 8 8 8 ' ' ' ' ' ' ' ' ' ' ' ', ', ' % W6 - (6.4 910.6 5)E i 1.31 i _ Illlms lH ' ' " 1.50 l 1.01 E 2 W5 - (6.73fD.67)E1 amu'11((.p'. _ IIIms 1.00 E 2 "!"!' - W4 - (6.73i0.67)E1 1.49 im'"u'8"rr'itIi
- 111ramii 9.0 3 E1 3 l!HH!!! ~ W3 - (6.18io.62)E1 1.46 ,
M,!"I
~
=~
2.57E1 ~sw b Ny 41l l l l H i l l l l lit t 111 W2 - (2.75i0.28)E1 0.935 ! 2.70E1 .. """].[... q.,.
. ~ ..
y W11 - (1.11i0.34)E1 2.43 4 1.74 E1 -- I W1 - (1.24io.12)E1 1.40 5 15 / FIGURE 1 "Co Specific Activities in Spent fuel Assembly liardware i 341 1
Westinghouse Fuel Assembly,14x14 Flatlo: Predicted Sam ale Measured Predicted I
) ,[ x Cl/g-Ni Num aer Cl/g-Ni ' Msa5Ured i --
- aI'E 0.0553 2.9 3 E-4 _.
'T W10 - (5.30io.06)E-3 3.3 8 E-4 1.19 E-3 gg W12 - (4.25i0.13)E-3 W9 - (3.35i0.10)E-2 0.0795 0.0355 iiisiisiissigissi
~ - ,
3.19 E- 3 :::: g .. . _ W8 - (3.70io.11)E-2 0.0862 ans un
*tlH H ilHI H i-H 1 5.5 7 E-2 { W7 - (1.4 010.04)E-1 0.398 IMI 8' 0.438 7.7 6 E-2 .
, - , ' WG - (1.77io.05)E-1
' ' ' ' ' ' ' ' ' '. 'E, .
_JRmois
- 8. 3 9 E-2 ' " ' ' " '. ' W5 - (1.73io.05)E-1 0.485 h
m-[RIDDiG 8.3 3 E-2 ' ' ' ' " ' " ' ", ' ! H, ' ' W4 - (1.70io.05)E-1 0.490
~llIrauis 7.5 3 E-2 "!""!U' - W3 - (1.81i0.05)E-1 0.416
- r _
2.18 E- 2 ~ ..s t m o N ' W2 - (7.5910.22)E-2 0.207
<ullilflitalin lin 9.64 E-3 W11 - (4.01io.12)E-2 0.240
..(-.<
]-[
7
/
4.2 8 E-3 --- -' WI - (3.59io.11)E-2 0.119
, = =i/
FIGURE 2 Ni Specific Activities in Spent Fuel Assembly liardware 342
Westinghouse Fuel Assembly,14x14 s
,. Flatio: ' Predicted Sam le Measured Eredicted i,,,,wyp Cl/g-N b - Number. Cl/g Nb Measurea 1.11 E-5
[kN k~ \
^p..k W10 - (1.64io.26)E-5 ' O.677 1.2 9 E-5 @j))J1djjj7 W12.- (1.24i0.26)E-5 1.04 4.19 E- 5 e11illitililltill W9 - (7.20i7.20)E-5 0.58.
1.4 3 E-4 x 7 --1 W8 - (1.59io.27)E-4 0.899
~_ IllIDJ et i+1111liiI i1644
- 1. 3 7 E- 3 f W7 -
1.91 E-3 Inuns onnissinaisis = W 6' - 2.0 6 E-3 ilinIHin si10 - W5. - (2.61io.70)E-3 0.789 2.0 0 E-3 spinluusHHi
- W4 -
c
.1. 8 5 E- 3 -
W3 - EMHuna._ ni,uilitil,nli
/
5.2 8 E-4 :- W2 - (9.0111.67)E-4 0.586 2_ 3.57E-4 ms. _ h 41I l l ll1111111 R t i ll W11 - (2.20i2.20)E-5 16
' 965-4
.. 1EIt
#- i
. 2 ..
* - n . 210.23)E-4 1.48
, = #
flGURE 3 "Hb Specific Activities in Spent fuel Assembly liardware 343
1 l 1 of two. As the neutron flux and energy spectrum drops rapidly between the-fueled region and the end fittings of the assemblies, the uncertainties in the calculational methods become much larger and'large differences in measuredversuscalculatedactivitiegwereotgerved. The largest discrepancies were observed for the Ni'and Ni. activities at the end fittings of the fuel assemblies. Since no adequate isotopic cross section u data exist for the stable parent nickel isotopes, elemental cross section-data were used, and this may have introduced relatively large uncertainties in the calculated results. The radionuclides measurements of the Gundremmigen pressure vessel steel were in very g Table 6) good . agreement The average with a blind comparison calculated-to-measured ratio forof calculgted Fe, gtivities Co, and Ni (see
-were 1.3, 1.9, and 0.69, respectively. The concentrations of the radionuclides were all below Class-A limits, indicating that the entire pressure vessel (not including internals) could have been disposed of as Class-A waste in a low level waste shallow land burial facility. ,
i The measurements and calculational methods utilized in this work have lent confidence to calculational methods for predicting radionuclides inventories in activated metals, and have identified certain problem areas where better cross section data or calculational methodology are needed. I FUTURE RESEARCH PLANS It is anticipated that the radionuclides source-term characterization work for this program will continue through FY 1990. Using the technology transfer data developed for Shippingport Station by DOE and its subcontractors, together with the supplementary radionuclides measurements conducted during this program, we will perform generic assessments of the various alternatives for waste disposal and evaluate optimum methods for packaging, shipping, and disposal of commercial decommissioning wastes in accordance with 10CFR61. We will recommend sampling and analytical methodologies which will enable us to determine if whole components such as steam generators and pressurizer can be disposed of as Class-A waste, and if entire pressure vessels can be disposed of as Class-C waste. Methods for removal, packaging and disposal of contaminated concrete will be evaluated. The data and information gained through evaluation of waste handling methods at Shippingport, as well as waste classification measurements for highly activated reactor internal _ components from commercial stations, will provide valuable information ' relating to waste handling and disposal during future decommissioning of commercial nuclear power stations. Wewillmakeaconcertedefforttoperformactualmeasuregentsgfsefected long-live Be, C1, Mo, j "Zr, and (radionuclides 'Ag. in radioactive materials, e.g.,We will also be assessing the adel radiochemical and analytical techniques for determining those difficult-to- ' measure radionuclides which have the greatest potential for analytical errors. ! i 344 i
_ u m# s a 2 0 1 1 a S >
- e -
oM i / td t ae it a l e 2 u l 9 6 3 - f c p3 7 6 1 6 o . l m# a a 1 0 1 1 se C S > nr ou i s t s ae ) rr A tP ( n ) 4 en l #) 6 8 5 2 ce e T - - - 1 ng e e7 E E E - oi t l 6 7 9 2 E C s p 6 0 4 5 m - m0 d e d a a( 2 3 1 3 er e/ S . td ti an aC 7 l u l( 7 uG u 9 1 c c l n l n , ai ao 3 C Ci 3 s t f) 6 8 5 2 1
.t a T - - - 1 y
sc r e1 E E E - vu t l 4 3 3 9 E r d n p 5 5 2 4 a d o e m0 u er c a( 4 7 3 4 n rP n S a u o J sn C f ao o ei Mt e a t f v ) oil A a t e ( d nce d) el 4 f) 6 8 6 2 n oat w s S re T - - - 1 o i n ue e7 E E E - st 6 2 7 5 E d rol l p t are as 3 3 2 1 u pt s e m0 mus oee Mo
/ S a( 1 4 9 3 h
s CNV yi l C r l( o a t c c i n a mo e 6 ei 3 r ht #) 6 7 5 2- o E ca T - - 1 t L or e1 4 E E E 3 4 1 E B it l A d n p 5 1 9 8 d T ae m0 e R c a( 2 1 2 2 t n S < c e o r C r o c y e a d c i e l D c o i e b u ) n C N F N k o 0 3 5 4 { i 6 6 5 9 d a}}