ML070580514

Relief Request IR-2-38, Structural Integrity Assessment Methodology for Brazed Joints

ML070580514
Person / Time
Site:	Millstone
Issue date:	02/28/2007
From:	Chernoff H NRC/NRR/ADRO/DORL/LPLI-2
To:	Christian D Dominion Nuclear Connecticut
Miller G, NRR/DORL, 415-2481
References
IR-2-38, TAC MC8893
Download: ML070580514 (19)

Text

February 28, 2007 Mr. David A. Christian Sr. Vice President and Chief Nuclear Officer Dominion Nuclear Connecticut, Inc.

Innsbrook Technical Center 5000 Dominion Boulevard Glen Allen, VA 23060-6711

SUBJECT:

MILLSTONE POWER STATION, UNIT NO. 3 - RELIEF REQUEST IR-2-38, STRUCTURAL INTEGRITY ASSESSMENT METHODOLOGY FOR BRAZED JOINTS (TAC NO. MC8893)

Dear Mr. Christian:

By letter dated June 9, 2005, as supplemented by letters dated September 14, 2006 and January 2, 2007, Dominion Nuclear Connecticut, Inc. (DNC) requested approval of a structural integrity assessment methodology for degraded brazed joints in the Millstone Power Station, Unit No. 3 (MPS3) service water system as an alternative to the repair and replacement requirements in American Society of Mechanical Engineers, Boiler and Pressure Vessel Code (ASME Code),Section XI.

The results of the Nuclear Regulatory Commission (NRC) staffs review indicate that DNCs performance of an ASME Code repair or replacement of the degraded brazed joints would result in hardship without a compensating increase in the level of quality and safety.

Therefore, DNCs request for relief is authorized for the remainder of the second 10-year inservice inspection interval for MPS3 pursuant to Title 10 of the Code of Federal Regulations Section 55a(a)(3)(ii)) on the basis that the proposed brazed joint assessment methodology in Relief Request IR-2-38, as an alternative to ASME Code repair or replacement, is acceptable because it provides reasonable assurance of structural integrity of the degraded brazed joints.

The details of the NRC staffs review are contained in the enclosed Safety Evaluation.

D. Christian If you have any questions, please contact your NRC Project Manager, Victor Nerses at 301-415-1484.

Sincerely,

/ra/

Harold K. Chernoff, Chief Plant Licensing Branch I-2 Division of Operating Reactor Licensing Office of Nuclear Reactor Regulation Docket No. 50-423

Enclosure:

As stated cc w/encl: See next page

D. Christian If you have any questions, please contact your NRC Project Manager, Victor Nerses at 301-415-1484.

Sincerely,

/ra/

Harold K. Chernoff, Chief Plant Licensing Branch I-2 Division of Operating Reactor Licensing Office of Nuclear Reactor Regulation Docket No. 50-423

Enclosure:

As stated cc w/encl: See next page DISTRIBUTION:

PUBLIC LPL1-2 Reading File Rids NrrOd RidsNrrDorlLpl12 RidsNrrLACRaynor RidsNrrPMVNerses RidsOgcRp RidsAcrsAcnwMailCenter RidsSecyMailCenter RidsRgn1MailCenter ADAMS Accession Number: ML070580514 OFFICE LPL1-1/PE LPL1-2/PM LPL1-2/LA CPNB/BC OGC NLO LPL1-2/BC NAME JHughey EMiller CRaynor TChan SHamrick HChernoff DATE 2/28/2007 2/28/2007 2/16/2007 2/28/2007 2/28/2007 2/28/2007 OFFICIAL RECORD COPY

SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELIEF REQUEST IR-2-38 FOR SECOND 10-YEAR INSERVICE INSPECTION INTERVAL DOMINION NUCLEAR CONNECTICUT, INC.

MILLSTONE POWER STATION, UNIT 3 DOCKET NUMBER 50-423

1.0 INTRODUCTION

By letter dated June 9, 2005, as supplemented by letters dated September 14, 2006 and January 2, 2007, Dominion Nuclear Connecticut, Inc. (the licensee) submitted Relief Request (RR) IR-2-38 for Nuclear Regulatory Commission (NRC or the Commission) approval.

Specifically, the licensee proposed a structural integrity assessment methodology for degraded brazed joints in the service water system as an alternative to the repair and replacement requirements in American Society of Mechanical Engineers Boiler and Pressure Vessel Code (ASME Code),Section XI. The subject RR will be used for inservice inspections (ISIs) for the remainder of the second 10-year interval at Millstone Power Station, Unit 3 (MPS3). The second 10-year ISI interval started on April 23, 1999, and is expected to end on October 23, 2008.

2.0 REGULATORY REQUIREMENTS Title 10 of the Code of Federal Regulations (10 CFR) Section 50.55a(g) specifies that ISI of nuclear power plant components shall be performed in accordance with the requirements of the ASME Code,Section XI, except where specific written relief has been granted by the Commission pursuant to 10 CFR 50.55a(g)(6)(i). Section 50.55a(a)(3) of 10 CFR states that alternatives to the requirements of paragraph (g) may be used, when authorized by the NRC, if (i) the proposed alternatives would provide an acceptable level of quality and safety, or (ii) compliance with the specified requirements would result in hardship or unusual difficulty without a compensating increase in the level of quality and safety.

Pursuant to 10 CFR 50.55a(g)(4), ASME Code Class 1, 2, and 3 components (including supports) shall meet the requirements, except the design and access provisions and preservice examination requirements, set forth in the ASME Code,Section XI, "Rules for Inservice Inspection of Nuclear Power Plant Components," to the extent practical within the limitations of design, geometry, and materials of construction of the components. The regulations require that ISI of components and system pressure tests conducted during the first 10-year interval and subsequent intervals comply with the requirements in the latest edition and addenda of

Section XI of the ASME Code, incorporated by reference in 10 CFR 50.55a(b), 12 months prior to the start of the 120-month interval, subject to the limitations and modifications listed therein.

The information provided by the licensee in support of the request has been evaluated by the NRC staff and the bases for disposition are documented below.

3.0 TECHNICAL EVALUATION

3.1 Licensees Evaluation 3.1.1 Components for Which Relief is Requested All ASME Code Class 3 service water system piping with brazed joints. The nominal pipe size is 3 inches and smaller. The piping has a design pressure of 150 psig or less, and a design temperature of 150 degrees Fahrenheit or less.

3.1.2 Applicable ASME Code Edition and Addenda The Code of record for the second 10-year ISI program and its evaluation at MPS3 is the ASME Code,Section XI, 1989 Edition with no Addenda. However, the ASME Code,Section XI, 1998 Edition with no Addenda has been approved for Section XI Repair/Replacement program activities. The original construction code is ASME Code,Section III, 1971 Edition with Summer 1973 Addenda.

3.1.3 Applicable ASME Code Requirement When leakage is discovered during normal plant operation, the degraded piping component must be repaired or replaced in accordance with ASME Code,Section XI, IWA-4000.

3.1.4 Licensees Proposed Alternative and Basis for Use In lieu of immediately performing an ASME Code repair, the licensee proposed to perform a supplemental ultrasonic test (UT) examination of the degraded brazed joint to assess the extent of the brazed bond. The UT results will be compared with brazed joint bond levels that are required for structural integrity of the specific piping under consideration and also account for the design basis loadings applicable to the condition.

The brazed joint will be examined by UT using a straight beam technique that monitors the relative strengths of signals returned from the internal diameter (ID) of the pipe and the fitting.

This technique was derived from, and is consistent with, the technique standardized by the U.S.

Navy for use on brazed shipboard piping. The examination technique has been documented in an MPS3 procedure. The procedure requires preparation of the examination surface to obtain satisfactory sound transmission, and use of calibrated equipment and approved couplants. The joint circumference is marked at a number of locations such that the marks are spaced no greater than 1 inch apart. Only Level II or III certified technicians who are familiar with brazed joint geometry and trained in interpreting brazed joint signal response characteristics may perform the UT examination or review the brazed bond readings.

The licensee stated that the proposed methodology to evaluate the structural integrity of the brazed joint includes an assessment with the following considerations:

(1) Assessments of system performance and indirect effects on other nearby equipment.

(2) Adjustment of bond readings to account for uncertainties.

(3) A review of design basis stress analyses of the piping to determine required joint strength.

(4) A comparison of the adjusted bond readings with the prequalified bond levels that have been shown empirically by physical testing to assure structural integrity.

As a prerequisite to the structural integrity assessment, knowledgeable engineering personnel will assess the effect of the leak on the system and other nearby equipment. The actual leak rate will be estimated and compared to service water system margins for loss or diversion of flow.

In addition, a walkdown will be performed to identify any nearby equipment that may be affected by the leak. If required, a drip collection device or spray shield will be installed and maintained for the duration that the leak continues.

If the average measured bond reading is equal to or greater than 60 percent, then no further structural integrity assessment of the brazed joint is required since the bond strength is expected to exceed the piping strength. This acceptance threshold is the same as the acceptance criteria in the U. S. Navy Standard that has been used for critical shipboard piping systems rated 300 psig and greater. This 60-percent threshold criterion is further confirmed by mechanical testing performed by MPS3 which is described in Attachment D of the licensees submittal dated June 9, 2005. The testing results have shown that if true bond in the joint exceeds 30 percent then the piping collapse load occurs before any bond failure. There is no brazed bond failure because the piping deforms plastically to relieve the imposed load.

If the average measured bond reading is less than 60 percent, further assessment of the brazed joint is required. The assessment consists of a review of the construction code qualification stress analysis for calculations and comparison of adjusted bonds to required bonds which are briefly described below.

The construction code qualification stress analysis of record is reviewed to determine design basis loadings at the subject brazed joint. Loads on the brazed joint include maximum operating pressure, deadweight, safe shutdown earthquake (SSE) and any transient dynamic loads that have been defined for the piping. The stress intensification factor (SIF) is not considered in the summation of nominal stresses used for assessment.

The load imposed on the brazed joint is calculated from the following equation and is expressed in terms of equivalent pipe stress (Seq) :

Seq = S(lp) + S(dl) +S(sse) + S(dyn) (1)

S(lp) = longitudinal pressure stress S(dl) = deadload stress S(sse) = safe shutdown earthquake (SSE) seismic stress S(dyn) = dynamic stress (if defined)

To account for the UT uncertainties in bond readings, the average measured bond reading or individual bond reading above 10 percent is adjusted downward on a sliding scale by the following formula:

badj = 100 x ( reading -10) / (100-10) units of percent (2) where badj is the adjusted average measured bond reading or individual bond reading. Using the above formula, all bond readings 10 percent or less are assumed to be zero.

For bond readings that are significantly non-uniform around the circumference of the brazed joint, an adjustment of the measured bond for an effective (lower) bond is computed-based on the equivalent moment of the adjusted bond areas.

The allowable loading (Smax(badj)) at a brazed joint for an equivalent bond level is calculated from the following equation:

Smax(badj) = (/4)((D2 x L(ins) x max x badj)/(Zpipe) (3)

D = pipe outside diameter L(ins) = insert depth of fitting socket excluding any insert groove Zpipe = piping section modulus max = 7.5 ksi (maximum braze shear stress) badj = adjusted effective bond

Smax(badj) is the safe loading level that the joint is allowed under the proposed methodology.

If the equivalent pipe stress (Seq, applied stress) multiplied by a safety factor of 1.5 is less than the allowable pipe stress at the braze joint, i.e.,

1.5 Seq < Smax(badj) (4) then the brazed joint is concluded to have adequate structural integrity for continued service.

The safety margin of 1.5 as noted in Equation (4) is consistent to that required in ASME Code Case N-513-1. An example of a structural assessment performed for a hypothetical leaking brazed joint is included in Attachment C to Enclosure 1 in the licensees submittal dated June 9, 2005.

The licensee has sponsored mechanical tests at an independent testing facility to demonstrate the correlation between reduced bond levels and joint strength. The results are shown in Attachment A of Enclosure 1 in the licensees submittal dated June 9, 2005.

The maximum braze shear stress (max) in Equation (3) is assumed to be 7.5 ksi. This is supported by the mechanical test results and the brazing procedure qualification test results provided in the licensees submittal.

The proposed assessment methodology also requires periodic monitoring of leakage to assure that the assumptions of the assessment remain valid. This is in addition to the monitoring conducted during normal daily plant operator rounds. The monitoring will be by visual observation of the appearance of the joint and its leak rate. UT will also be performed to reconfirm the percentage of bonding. The frequency of the monitoring of leak rate and percentage of bonding will be approximately once every three months, not to exceed 120 days between observations.

If the joint does not have adequate bond by this assessment, the licensee may perform repair or replacement of the joint, or temporary non-ASME Code repairs subject to NRC review and approval consistent with NRC Inspection Manual, Part 9900: Technical Guidance, "Operability Determinations and Functionality Assessments For Resolution of Degraded or Nonconforming Conditions Adverse to Quality or Safety", for the resolution of degraded and nonconforming conditions.

3.1.5 Duration Of The Relief Request The proposed relief request will be used at MPS3 for the remainder of the second 10-year ISI interval which started on April 23, 1999, and is expected to end on October 23, 2008.

3.2 NRC Staff Evaluation The licensees proposed structural integrity assessment methodology will allow the degraded brazed joint to remain in service in an unrepaired state for a limited period of time provided that the structural integrity of the degraded brazed joint is assured by UT examination and/or analytical evaluation and the effects of leakage are appropriately assessed and mitigated to ensure the functionality of the affected system. There are no ASME Code provisions to

address the evaluation of the degraded brazed joint for continued operation. ASME Code Case N-513-1, Evaluation Criteria for Temporary Acceptance of Flaws in Moderate energy Class 2 or 3 Piping,Section XI, Division 1," is not applicable to degraded brazed joints because the degradation is due to defects in braze bonding between piping and fittings. Therefore, the observed leaking brazed joints must be repaired. The licensee stated that the plant must be shut down in order to perform the ASME Code repair, because certain safety-related systems or components will not be available during the repair, which is a violation of the Technical Specification requirements. The NRC staff finds that the need to shut down the plant for implementing an ASME Code repair of the degraded braze joint would result in hardship to the licensee without a compensating increase in the level of quality and safety when the structural integrity of the degraded joint and the system functionality are assured by appropriate evaluation. A plant shutdown will also unnecessarily cycle plant components, which is not desirable in maintaining the structural integrity of the safety-related components. Therefore, the staff finds that it is acceptable to propose an alternative disposition to degraded brazed joints provided that the licensee can demonstrate the acceptability of the alternative.

3.2.1 Description of a Brazed Joint The typical configuration of a brazed joint is shown in Figure 1 of the licensees supplement dated January 2, 2007. The typical piping materials are Cu-Ni alloy (SB-466) or Nickel alloy (SB-165) and the fittings and valves are made of cast bronze (SB-61 and SB-62). The brazing filler material is typically a silver alloy (SFA 5.8, Bag-1, Bag-1a or Bag-7).

A typical brazed joint fitting has a deep socket for inserting the pipe. Although it appears similar to a socket welded joint, the fabrication and structural behavior are quite different. Whereas the socket weld achieves its joint strength by a filler weld, resulting in fusion of similar material between the pipe and the outer face of the fitting, the braze achieves its strength by surface bonding of the outside of the pipe to the inside of the fitting socket using a dissimilar metal braze filler of silver alloy. The resulting braze filler metal is very thin (approximately 1 to 5 mils).

The load transfer between pipe and fitting is thus primarily by shear through the braze filler. It is noted that there is no inherent stress concentration factor like that normally applicable to socket welds because there is no significant pipe wall bending induced by the shear load transfer over a length that is several wall thicknesses long.

The length of lap (the length of the pipe inside the joint), the shear strength of the brazing alloy, and the average percentage of the brazed bonding are the principal factors determining the strength of brazed joints. The shear strength may be calculated by multiplying the width by the length of lap, by the percentages of bond area, and by taking into consideration the shear strength of the alloy used.

Since the piping loads causing longitudinal stress in the pipe are all transferred by shear stress through the brazed bond, the shear stress in the brazed bond is directly related to longitudinal pipe stress divided by a factor equal to the overlap ratio. Thus, for a fully-bonded brazed joint, the shear stress is about one-fourth of the piping longitudinal stress. If the bond is only 50 percent of maximum, then the bond shear stress will be about half the piping longitudinal stress.

Given that piping and brazing filler metals have similar strength, a brazed joint has more than enough residual strength to tolerate moderate bond imperfections. Consequently, the joint is not the weak link in the piping assembly.

Consistent with this inherent over-design of brazed joints, the Construction Codes, such asSection III of the ASME Code and ANSI B31.1, require only visual inspection of the resulting bond. The ASME Code does not require surface examinations such as by liquid penetrant or volumetric examinations for brazed joints.

The licensee stated that a degraded brazed joint with weepage is the result of imperfections in the braze materials and not the result of a service-induced flaw in the pipe or fitting pressure boundary. The pressure-retaining boundary still retains its structural integrity. Although the shear load transfer between the pipe and fitting is clearly a pressure boundary function, the brazing material is designed to function more as a sealant between the connected components.

The licensee stated further that imperfections in the sealant function of the braze material are permissible, provided its load transfer function retains adequate margin. There is no direct degradation of the pressure boundary. In addition, the characterization of braze imperfections is very different from the planar flaws or loss of wall thickness that are addressed in ASME Code,Section III, IWA-3000.

The staff finds that the licensee has appropriately designed the brazed joints in accordance with the Construction Codes (i.e., the ASME Code Section III and ANSI B31.1).

3.2.2 Effect of Leakage As a prerequisite to the proposed structural integrity assessment, the effect of the leak on the system and other nearby equipment must be assessed. To prevent the leakage from affecting nearby safety-related equipment, the licensee stated that a drip collection device or spray shield may be installed, if required, and maintained for the duration that the leak continues.

Typical leakage from the brazed joint is very small in terms of drops per minute because the clearance between the pipe and fitting is normally about 0.005 inches. The licensee calculated the worst case leak at a 3-inch braze joint at 100 psig is about 6 gpm. This leak rate is small in comparison to service water pump capacity and would not affect the functionality of the system.

In a response to the staffs request for additional information dated September 14, 2006, the licensee calculated an upper-bound leakage for the same 3-inch pipe/fitting brazed joint assuming the piping is separated from the joint due to complete loss of the bonding. The leak rate is estimated to be about 699 gallons/minute. With this leak rate, downstream cooling to the affected components would be lost. Therefore, the staffs evaluation is focused on the maintenance of the structural integrity of the degraded brazed joint.

3.2.3 UT Examination Procedure for Brazed Joints The licensee stated that the UT examination procedure using a straight beam longitudinal wave technique was derived from the techniques standardized by the U. S. Navy for use on brazed shipboard piping. The examination technique has been documented in an MPS3 procedure (MP-UT-45). Because of certain publication restrictions regarding its use, the staff did not have

an opportunity to review the subject U.S. Navy document. The licensee stated that the UT examination procedure as described in MP-UT-45 has been independently validated and qualified for use at MPS3. In a response to the staffs request for additional information dated September 14, 2006, the licensee described in detail a trial demonstration using the subject UT procedure. Five UT operators including three qualified level II or III, and two with previous U.S.

Navy experience participated in the demonstration trial. They performed round robin tests on six brazed joint samples. These joint samples, consisting of two 2-inch tees, two 2-inch couplings and two 3-inch elbows had been previously installed at MPS3 but were removed as part of plant modifications. After completion of the UT testing, each brazed joint sample was mechanically cross-sectioned 3 times and examined to measure the actual percentage of bond at each section. The average percentage bond on mechanical sections correlate well with the UT percentage bond with an average percentage bond on mechanical sections 10 percent higher than the UT percentage bond. Therefore, the results of percentage bond measured by UT are conservative in comparison with the percentage bond destructively measured. There are variations in UT measured percentage bond on each sample among various examiners, particularly, on samples with low percent bond. To account for UT uncertainties, the procedure requires the results of UT measurements to be adjusted by 10 percent on a sliding scale using Equation (2) as shown in Section 3.1.4 of this Safety Evaluation (SE). The 10-percent adjustment is supported by the UT results reported in the trial demonstration.

In a response to the staffs request for additional information dated September 14, 2006, the licensee provided information regarding the personnel qualification requirements which are the prerequisites for performing the UT examination of brazed joints. Only certified Level lI and Level III UT examiners may independently perform, interpret, evaluate and report examination results. The UT examiners must meet the initial qualification requirements by successfully measuring the percentage bond on six test specimens. In addition there are requirements to maintain the proficiency every six months, and re-qualification every three years.

Based on an evaluation of the information provided by the licensee, the staff has determined that the UT examination performed on the braze joints at MPS3 using the technique and procedure as described in the licensees submittal will provide a reasonable estimation of the bonding level at the brazed joint.

3.2.4 Bonding Assessments The bonding between the socket fitting and piping may not be uniformly distributed around the circumference of the pipe due to fabrication or inservice degradation. To assess the joint bonding, the licensee measures the bonding of a brazed joint by UT in 18-degree increments around the circumference of the joint. The measurement will result in 20 bond readings of various percentage (100 percent being the full bonding). The average of the 20 bond readings is calculated. If the average bonding is less than 60 percent, the measured bonding are adjusted (reduced) for UT measurement uncertainties for further analysis. All bond readings at 10 percent and below are conservatively assumed to have zero bonding and readings above 10 percent are reduced by 10 percent (see Equation 2 in Section 3.1.4 of this SE). For bond readings that are significantly non-uniform around the circumference of the pipe, the bonding is adjusted further based on the equivalent moment of the adjusted bond areas to consider the offset to the principal axes of the brazed joint.

As stated above, the acceptance criterion of the 60-percent average bond reading or more is the same as the acceptance criterion used in U.S. Navy ships. The U.S. Navy criterion is applicable to piping systems rated at 300 psig and greater. The MPS3 acceptance criterion is applied to piping systems less than 150 psig, which is more stringent and conservative than the 300 psig used by the U.S. Navy.

In addition, the licensees test results showed that the 60-percent bonding will develop the full bending strength load of the piping, even when the bending is extended beyond required design levels. That is, the joint with 60-percent bonding has a bond strength that exceeds piping strength. The licensees tests also showed that the brazed joint is stronger than the pipe when the bonding exceeds 30 percent. The pipe will deform plastically to relieve the imposed load, and this occurs at loads greater than the maximum load permitted by the licensing basis analysis of the piping. The licensee introduced conservatism in its methodology by reducing the measured bonding to be used in its acceptance by analysis. The adjustment of bonding is to correlate the data from actual piping samples and to account for uncertainties in bond readings from UT.

The staff finds that the method of estimating bonding in the joint is acceptable because the licensee will use demonstrated UT techniques to measure the bonding, will consider the uncertainty in the UT measurement, and will conservatively reduce the measured bonding by 10 percent for further evaluation when the average bonding is less than 60 percent.

3.2.5 Analytical Model and Method The stress or load at any point of a brazed joint is proportional to its distance from the bending axis of the brazed joint. Therefore, the strength of a brazed joint is the integration of the strength of each bond area times its distance from the neutral axis. On the basis of this concept, the licensee derived the proposed allowable stress equation (Equation 3 in Section 3.1.4 of this SE) using the first principle of shear stress. The equation is based on several factors such as the maximum shear stress the bonding in the joint can hold, the effective bonding in the joint, section modulus of the joint, and the cross sectional area of the pipe. The applied stresses on the joint from normal operation and faulted conditions are then compared to the allowable stress to determine the acceptability of the joint as shown in Equation 4. The staffs concerns about the proposed analytical model, as represented by Equations 3 and 4, are the material property used for the brazed filler material and how the applied stresses are obtained. These two concerns are discussed below.

The ASME Code does not define allowable mechanical or material properties for brazed filler material. Also, for Class 2 and 3 piping such as service water system piping, ASME Code,Section III does not require certified material test reports. The fabricated sample brazed joints were fabricated from materials taken from station stock and are, therefore, representative of actual joints in service. The licensee stated that the failure of a brazed joint occurs at the interface between the fitting and the pipe. Therefore, the mechanical property of the brazed material is one of the parameters that would affect the strength of the joint.

In the original submittal dated June 9, 2005, the licensee assumed a brazed joint shear strength (max) of 5 ksi. In the January 2, 2007 letter, the licensee revised max from 5 ksi to 7.5 ksi based

on additional tensile test data taken from brazed joint qualification tests. In the qualification tests, each of the tested joints achieved a collapse load that would support a 7.5 ksi braze shear strength. The tests are discussed further in this SE.

In the June 9, 2005, submittal, the licensee proposed an acceptance criterion requiring that the applied stresses (without a safety margin) be less than the allowable stress. The staff suggested to the licensee that a safety margin of 1.5 should be applied to the applied stresses to be consistent with Code Case N-513-1 of ASME Code,Section III. By letter dated January 2, 2007, the licensee revised the brazed joint evaluation procedure in the original request and applied a safety margin of 1.5 to the applied stresses as shown in Equation 4.

The licensee reviewed pipe stress analysis of record to determine design basis loadings at the subject brazed joints. Pressure, deadweight, and SSE loadings are included as part of applied loading in the evaluation. The licensee obtained the applied stresses at the nodal point of the joint from the output of the pipe stress analysis which was calculated based on NB-3000 of ASME Code Section III. The applied stresses were then reduced to nominal stresses by eliminating the stress intensification factors that are required by ASME Code,Section III. The staff asked the licensee why the stress intensification factors in the applied stresses were eliminated. By letter dated September 14, 2006, the licensee responded that the theoretical and testing bases for the proposed alternative were derived from applied forces and moments.

The testing applied a load in a three-point bending configuration resulting in an easily calculated moment at the joint. As a convenience for evaluation purposes, these are converted to equivalent nominal pipe stress. However, the strength correlation to a braze bond is based upon empirical analysis of the load testing. Local stress concentration effects at the joint were inherent in the tests on actual brazed joint fittings. Therefore, the stress intensification factors as required for ASME Code,Section III, stress analysis of the joint does not enter into the strength correlation.

The licensee stated further that when existing stress analysis of piping is used as input to the evaluation, it can either access the detailed piping loads that are available as computerized output, or use the summarized pipe stress output that included the effects of the detailed piping loads. By removing the stress intensification factors from the stresses results, the actual joint loading, in terms of nominal stress, can be compared directly to joint strength, also in terms of nominal stress. The staff agrees with the licensee on the removal of the stress intensification factor in the comparison of the applied stresses to the joint strength.

The staff finds that the proposed analytical model is acceptable because (1) the model was developed based on the first principle of shear strength of the joint, (2) the material property used for the brazed joint is supported by test data, (3) the appropriate applied stresses with a safety margin of 1.5 are used in comparison to the allowable shear stress, and (4) the analytical model has been verified by the mechanical and qualification tests.

3.2.6 Mechanical Tests The maximum shear stress of the brazed material used in the calculation of the allowable shear stress (Equation 3) at a brazed joint requires validation from mechanical tests because the ASME Code does not specify the material properties of the brazed filler materials. The licensee

performed 3-point bending tests to demonstrate the shear strength of the joint. The licensee also demonstrated shear strength of the brazed joint by tensile tests as part of brazed joint qualification. The two types of tests are discussed below.

In the June 9, 2005, submittal, the licensee presented data from 3-point bending tests to demonstrate that the analytical methodology uses conservative shear strength to qualify the brazed joints. Three types of specimens were used in the three-point bending tests: fabricated joints with a controlled average bond, fabricated joints that had disbondment on a contiguous arc-segment of the joint, and field sample piping joints.

3.2.6.1 Fabricated Joints with a Controlled Average Bond Level By a combination of machining and use of insert-groove type fittings, the licensee fabricated a series of test joints with equivalent bond levels of 12, 30, 40 and 60 percent. The samples were fabricated for 2-inch and 3-inch joints for a total of 24 samples. The test results showed that all joints with 30 percent or higher bond achieved full piping collapse strength with no failure of the bond. One of the 40-percent bond joints had indications of bond failure when the test load is above the piping collapse load. The 12-percent bond level joints experienced bond failure before reaching piping collapse load, but still withstood a minimum of 37 percent of the piping collapse load.

3.2.6.2 Fabricated Joints with Disbondment on a Contiguous Arc-Segment of the Joint These specimens were intended to explore the effect of having a significantly non-uniform distribution of bond area around the circumference of the joint. The licensee fabricated six samples with disbondment segment angles of 36, 48, 72, 90, 108, and 126 degrees. The average bond levels for these ranged from 65 percent to 90 percent. The test results showed that from 36 through 72 degrees of segment disbondment, the specimens all achieved full piping collapse load. The specimens with 90 through 126 degree disbondment exhibited progressively lower collapse load. At 126 degrees disbondment, the specimens achieved about 60 percent of the piping collapse load.

3.2.6.3 Field Sample Joints The licensee obtained joints removed from the plant after about 20 years of service and screened by UT. Joints with lowest measured bond levels were selected for testing. The licensee tested a total of 9 field joints. The test results showed that the field samples showed considerable variation in collapse load. All specimens in this group also achieved their test collapse load at a load above the braze shear strength.

The above tests showed that if the bond exceeds 30 percent, the pipe collapse load occurs (i.e., the pipe fails) before any bond failure. The pipe will deform plastically to relieve the imposed load, and this occurs at loads greater than the maximum load permitted by the licensing basis analysis of the piping.

The staff asked the licensee to provide a technical basis for the three-point bending tests, including uncertainties and limitations. In the September 14, 2006, letter, the licensee stated

that the three-point bending tests were conducted because the most significant design loads experienced by the joints are bending due to deadweight and seismic loads. Testing in torsion or direct pullout would have required a complicated test fixture and the torsional and pullout loads are not the most severe when there is any non-uniformity of the bond. The staff agrees with this observation - that when the bonding is not uniformly distributed in the joint, the bending load, not torsional or pullout load, would be significant in degrading the joint.

The licensee stated that uncertainty on the loads and moments applied to the joint are reduced with the three-point bending testing fixture as compared to a test machine capable of imposing a very large load for direct pullout testing. The testing load cell is calibrated and the accuracy of the moment arm is known to within a fraction of an inch. The licensee concluded that accuracy of test loading for the three-point bending is reasonably adequate.

The test collapse load was derived from the load-deflection curve. The collapse load is defined in ASME Code,Section III, Appendix II, Section II-1430. The bond failure is defined as a discontinuity or knee in the load-deflection curve.

In the three-point bending testing, the licensee did not differentiate between local and total bond failure. Under progressive loading the initial bond failure is expected to be local failure, and additional loading results in additional bond failure. After the initial bond failure, all tests were continued up to a defection limit to determine an ultimate load capability. However, the joint is considered to have failed at the initial indication of bond failure even though the joint still has additional strength. After full deflection, the piping had ovalized and some joints were distorted, but the joints were not severed.

The staff finds that the three-point bending tests have demonstrated the structural integrity of the brazed joints with 60 percent of the bonding.

3.2.6.4 Brazed Joints Qualification Tests In addition to the above three-point bending tests, the licensee also provided tensile test data of the brazed joints from existing ASME Brazing Procedure Qualification Records by supplemental letter dated January 2, 2007. Brazing Procedure Qualification Tests were performed in accordance with ASME Code,Section IX. In order to pass the tensile test, the brazed specimen must have a tensile strength that is not less than the specified minimum tensile strength of the weaker of the two base metals being joined.

Three types of specimens were tested: (1) 3-inch Monel (nickel-copper alloy, P-110) pipe connected to copper-nickel alloy (P-107) fitting with pre-placed Bag-1a insert ring reduced section (a 3-inch wide and 12-inch long section of the pipe was tested); (2) 3/4-inch P-107 pipe connected to P-110 fitting with pre-placed Bag-7 insert ring; and (3) 3/4-inch P-107 pipe connected to carbon steel (P-101) fitting.

The results showed that in all but two of the reported tensile tests, the specimens failed in the base material which means that the pipe failed before the brazed joint. Therefore, most results could not provide an ultimate shear strength for the brazed joint. The reported values can only demonstrate that the brazed joint was capable of carrying at least the reported shear stress

without failure. The ultimate shear stress of the brazed joint could be much higher than the reported values. In the two joints where failure occurred in the braze joint, the ultimate shear strength of the braze was 15.8 ksi. Values of the other 10 specimens range from 10.0 ksi to 18.0 ksi. These shear strength values do not take into account any loss of shear area due to voids, inclusions or other flaws in the bonding specimens, which typically exceed 10 percent and may include up to 25 percent of the braze area and are still acceptable to ASME IX criteria.

The licensee stated that the indicated ultimate shear strength of these brazed joints is thus greater than 15.0 ksi. As a conservative measure, the licensee applied a margin of 2 on the joint shear strength which results in an allowable shear stress value of 7.5 ksi for evaluation of the structural integrity of the brazed joints as shown in Equation 3 above.

The staff finds that the licensee has demonstrated by the three-point bending and tensile testing that (1) degraded joints with 60-percent bonding would have sufficient shear strength to maintain their structural integrity, and (2) the assumption of the maximum allowable shear strength of 7.5 ksi for the brazed material is acceptable because it provides reasonable degree of conservatism in the calculation of the allowable stress at the brazed joint.

3.2.7 Monitoring The licensees proposed assessment methodology requires periodic monitoring of the degraded brazed joints to assure that the assumptions of the assessment remain valid. However, the licensees initial proposed monitoring consists only of the visual observation of the appearance of the joint and its leak rate at a frequency approximately once every three months, not exceeding 120 days between observations. The staff considers that by monitoring only the leak rate of the degraded joint, it will not provide adequate assurance for the structural integrity of the brazed joint. This is based on the concern that the change in leak rate of a small leak may not be sensitive enough to fully reflect the change of the bonding condition. Furthermore, there is no data to show the relationship between the leak rate and the percentage of bonding.

Therefore, the staff recommended that UT should be performed periodically to assure that there is no change in the level of bonding. In a response to the staffs RAI dated January 2, 2007, the licensee agreed to require a periodic UT of the affected brazed joint at least once every three months to reconfirm the percentage of bonding level used in the evaluation of brazed joint structural integrity. The staff finds that the licensees proposed UT and visual examination once every three months are sufficient to monitor the conditions of degraded brazed joints and, therefore, are acceptable.

3.2.8 Augmented Examination The staff finds that the guidance provided in the licensees submittal dated June 9, 2005, Section 5.6 Augmented Examination, is not consistent with that provided in ASME Code Case N-513-1. The proposed guidance allows the exemption of the previously-examined joints from re-examination. This could be non-conservative, since the joints may have been examined a long time ago or were examined using a technique that was not capable of identifying the degraded condition. In a response to the staffs request for additional information dated September 14, 2006, the licensee agreed to implement augmented examination consistent with ASME Code Case N-513-1.

3.2.9 Schedule for Code Repair The staff had concerns regarding the licensees initial proposal to apply the assessment methodology to leaking joints that were detected during the scheduled leak test. The staff was also concerned that the methodology would allow the deferral of Code repairs beyond the next refueling outage. In a response to the staffs request for additional information the licensee modified their proposal to limit the application of the proposed assessment methodology only to leaking joints detected during normal plant operation, and that Code repairs will be performed at the earliest of the following:

(1) Next scheduled shutdown of sufficient duration to complete repairs, or a scheduled shutdown greater than 30 days, (2) Next refueling outage, (3) Time at which flaw/leak size is predicted to exceed the flaw/leak size accepted by evaluations, or (4) Leaks discovered during plant shutdown.

The staff finds that the licensees proposed schedule to perform ASME Code repair as stated above is acceptable because the proposed schedule will require implementation of an ASME Code repair as early as possible without incurring an unnecessary plant shut down. In addition, this schedule will not compromise the level of quality and safety in plant operation since reasonable assurance of the structural integrity of the degraded brazed joint is provided.

3.2.10 Summary Based on the staffs evaluation of the licensees proposed structural integrity assessment methodology as discussed above, the staff finds that the proposed assessment methodology is acceptable because it will provide reasonable assurance that the structural integrity of the degraded brazed joint will be maintained prior to the performance of the ASME Code repair of the degraded components. The staff also finds that the performance of an ASME Code repair would result in hardship to the licensee because it would require a plant shutdown, with no compensating increase in the level of quality and safety.

4.0 CONCLUSION

Based on the above review, the staff concludes that performance of an ASME Code repair or replacement of the degraded brazed joints would result in hardship without a compensating increase in the level of quality and safety. The staff also concludes that the proposed brazed joint assessment methodology in Relief Request IR-2-38, as an alternative to ASME Code repair or replacement, is acceptable because it provides reasonable assurance of structural integrity of the degraded brazed joints. Therefore, pursuant to 10 CFR 50.55a(a)(3)(ii), the proposed alternative is authorized for MPS3 for the remainder of the second 10-year ISI interval.

All other ASME Code,Section XI, requirements for which relief was not specifically requested and authorized herein by the NRC staff remain applicable, including third party review by the Authorized Nuclear Inservice Inspector.

Principal Contributors: B. Koo J. Tsao Date: February 28, 2007

Millstone Power Station, Unit No. 3 cc:

Lillilan M. Cuoco, Esquire Mr. Joseph Roy, Senior Counsel Director of Operations Dominion Resources Services, Inc. Massachusetts Municipal Wholesale Building 475, 5th Floor Electric Company Rope Ferry Road Moody Street Waterford, CT 06385 P.O. Box 426 Ludlow, MA 01056 Edward L. Wilds, Jr., Ph.D.

Director, Division of Radiation Mr. J. Alan Price Department of Environmental Protection Site Vice President 79 Elm Street Dominion Nuclear Connecticut, Inc.

Hartford, CT 06106-5127 Building 475, 5th Floor Rope Ferry Road Regional Administrator, Region I Waterford, CT 06385 U.S. Nuclear Regulatory Commission 475 Allendale Road Mr. Chris Funderburk King of Prussia, PA 19406 Director, Nuclear Licensing and Operations Support First Selectmen Dominion Resources Services, Inc.

Town of Waterford 5000 Dominion Boulevard 15 Rope Ferry Road Glen Allen, VA 23060-6711 Waterford, CT 06385 Mr. David W. Dodson Mr. J. W. "Bill" Sheehan Licensing Supervisor Co-Chair NEAC Dominion Nuclear Connecticut, Inc.

19 Laurel Crest Drive Building 475, 5th Floor Waterford, CT 06385 Rope Ferry Road Waterford, CT 06385 Mr. Evan W. Woollacott Co-Chair Nuclear Energy Advisory Council 128 Terry's Plain Road Simsbury, CT 06070 Senior Resident Inspector Millstone Power Station c/o U.S. Nuclear Regulatory Commission P. O. Box 513 Niantic, CT 06357 Ms. Nancy Burton 147 Cross Highway Redding Ridge, CT 00870