ML20004B230
| ML20004B230 | |
| Person / Time | |
|---|---|
| Site: | Oconee, Maine Yankee, Rancho Seco  |
| Issue date: | 05/15/1981 |
| From: | BABCOCK & WILCOX CO. |
| To: | |
| Shared Package | |
| ML16148A424 | List: |
| References | |
| RTR-NUREG-0737, RTR-NUREG-737, RTR-NUREG-CR-1872, TASK-1.C.1, TASK-2.K.2.13, TASK-TM 77-1125756, IEB-79-05C, IEB-79-27, IEB-79-5C, NUDOCS 8105270423 | |
| Download: ML20004B230 (15) | |
Text
{{#Wiki_filter:.-- 1 O LETTER REPORT ON REACTOR VESSEL SRITTLE FRACTURE CONCERNS IN B&W CPERATING PLANTS l Document Identifie-77-1125756 May 15, 1981 l l Prepared by: Sabcock & Wilcox Company for the Owners Group of Babcock and Wilcox 177 Fuel Assembly NSS Systems 4% 8'105270
s se hf Table of Contents 'b PAGE I. A8STRACT 1 II. GENERAL 1 1 C A. Reactor Vessel Brittle Fracture during Design Basis LOCA 1 W B. Reactor Vessel Brittle Fracture during Small Break LOCA 2 1 Small Break LOCA - Specific 3 i j 1. Generic Assumpticns 3 2. Bounding Assumptions (No Vent Valve Flow Mixing) 4 3. Mixing Assumptions (Vent Valve Mixing) 4 4 j 4 Operating Vessels Fluences (EFPY) 5 5. Generic Analysis Conclusions 6 III. NON-LOCA OVERC00 LING EVENTS 7 J A. Comparisons of Overcooling Event to SSLOCA Analysis 7 m IV. CONCERNS EXPRESSED IN BASDEKAS' LETTER TO UDALL (a/10/B1) 8 3 A. " Overcooling Transient Cools Vessel to 150*F" 8 l B., Vessel fracture......cause core meltdown 8 C. Rancho Seco Transient, 3/20/78 9 D. Maine Yankee Vessel Fluence 9 j E. Operator Instructions 10 v. OTHER ACTIONS 11 Ti 6 A. Completed 11 1. ICS/NNI & EP4 Systems Upgrades 11. j 2. Revised SBLOCA Operating Guidelines 11 3. Recommendation to Maintain BWST Temperatures Greater 11 Than Technical Specification Minimums a 4. ATOG Considerations of Problem 11 d' 5. Owners Group document (BAW-1511P) on Reactor Vessel 11 Materials m i B. Currently Underway 11 i 1. Owners Group Reactor Vessel Materials Program 11 3 2. Reactor Versel Material Surveillance Programs in 11 Accordance with Appendix H of 10CFR50 np C. Imaediate Future Plans 12 1. Plant Specific Evaluations 12 j 2. 2-0 Vessel Heat Conduction Evaluations 12 3. NON-LOCA events Evaluations 12 !3
5 7-ri Table of Contents (Continued) y PAGE -3 0. Long. Term Plans 12 1. Consideration of Thermal Mix Test (discussion with 12 qj .EPRI,CREARE,etc.) 2. Consideration of Enhanced Inservice Inspectic.) 12 Technicuas i 9 3. Evaluation of In-place Reactor Vessel Thermal. Annealing - 12 4. Investigation of Improved Dosimetry and Fluence 12 Calculations I VI.
SUMMARY
- JUSTIFICATION FOR CONTINUED OPERATION 12 l
VII. ATTACHMENTS f Tabie 1 'A-1 p Primary System Response During Overcooling Transients 's Figure 1 . A-2 Allowable and Actual Pressures vs. Time for Rancho Seco. 40F BWST. Mixing References A-3 i J .fb
- 1 h4 i
l 1 .m LJ l t l I pt 4 2 - -~ 4 i. 4 g --ty =yv- .-w w-w e gw, w,y--e--- yy y. -r-y,p y, w,*g-y.pyr-- c--m --3-y,- 7 -,>, > 9 9m--,. g g -, ~.,, 7 o-, y
J it.9 Reactor Vessel Brittle Fracture I. Abstract: J This letter report summarizes the evaluations made to date regarding possible brittle fracture of B&W operating plant reacto. vessels during 53 transients that result in severe overcooling with potential i repressurization of the reactor vessel. It was prepared in response to an NRC request during a March 31, 1981 meeting between the NRC and various industry groups. The basis for concluding that there is no immediate 1 brittle fracture concern (into 1983) for S&W operating units resulting fecm thermal shocking of the reactor vessel during small break LCCA transients is presented. A comparison of the small break LOCA event with other ( overcooling events is made to demonstrate the small break analysis bounds the overcooling transient. Lonc term plans to resolve the concern are s unnari zed. ~ j II. General: A. Reactor Vessel Brittle Fracture during Design Basis LOCA -a Babcock & Wilcox evaluated the capability of its pressurized water reactor vessels to withstand thermal shock caused by the double-ended ruoture of a 36-inch-diameter hot 'eg pipe as early as 1969.(1) At that time, the hot leg rupture was ascertained to represent the most severe LOCA condition (i.e. from the standpoint of a brittle fracture 2, failure). Based on this early analysis of the hot leg rupture. it was concluded that "The reactor vessel will not lose its integrity due to i crack propagation as a result of thermal shock caused by actuaticn of [" the ECCS following a LOCA even if this transient occurs at the end of ao 40 years of irradiation and the vessel wall contains a flaw of critical i size"., 6 - i.
si? B. Reactor Vessel Brittle Fracture during Small Break LOCA As a result of the TMI-2 transient, new operating guidelines were issued which included operation of the HPI system in a once-thru b[ cooling mode as a means of core cooling until the plant could,be cooled and depressurized and then placed on the decay heat system. This mode of operation raised new questions concerning the thermal shocking of the reactor vessel due to the cold HPI flow being injected into the vessel with no RCS flow. Because of these new considerations and in response to N L' REG-0737( 2 ), analyses were performed in 1980 for the small is break LOCA transients with extended loss of feedwater. Reports documenting these analyses were submitted to the NRC by the Licensees in January. 1981.(3),(4) Recently, the issue has been raised by the NRC as to whether or not the small break loss-of-coolant transient with extende.d total loss ~~ of feedwater indeed represents the worst overcooling transient which should be considered with regards to reactor vessel brittle fracture. This report addresses this concern and concludes that the small break LOCA trarsient (as analyzed in BAW-1648) is the limiting transient for the B&W NSSS designs. This limiting event is, therefore, treated in scme dewail in the following section, followed by sections discussing the Non-LOCA events, other activities (engoing and planned) related to the brittle fracture concern and finally a summary presenting justification for continued p'. ant operation, s . i ? [ l.
~~. O Small Break LOCA - Specific The small break LOCA transient with extended loss of feedwater has ) been thoroughly analyzed with regard to reactor vessel brittle fracture (3),(4) (A description of the transient scenario is provided in Section 1 of Reference 3.) The analyses envelope all of the B&W operating units, (i.e.. worst-case inputs are combined). Some of the salient conservative assumptions used in these generic analyses are as follows: 1. All feedwater is lost for an extended period of time. 2. All reactor coolant flow is lost for an extended period of time. k 3. Core flow into the downcomer is assumed to pass through four vent-e Ed valves rather than the eight valves existing on all but one plant. This reduces the amount of warm water entering the downcomer. 4 A hypothetical maximum HPI flow capacity is assumed over the entire RCS pressure range analyzed. No single plant can achieve this hypothetical J capacity over the entire pressure range. This assumption affects all the analyses, including those which assume operator action to throttle HPI, since the initial reactor vessel cooldown prior to achieving 0 100 F subcooled conditions at the core outlet is maximized, resulting in increased thermal stress during the transient. 5. A worst-case HPI fluid temperature of a0 F was assumed. 0 6. Linear elastic fracture mechanics (LEFM) methods were used in the m brittle fracture analysis. No credit was taken for warm prestressing. 7. Materials information was taken from Regulatory Guide 1.99. 8. Reactor vessel most limiting welds were assumed to be located directly beneath the cold leg inlet no::les. i b 9. Reactor vessel cooldown was calculated based on a one-dimensional neat conduction analysis.
- 10. Mixing in the cold leg piping was not modeled.
3 ~
L-The major uncertainty associated with the coalyses is the degree of heatup of the high pressure injection water due to j - Upstream mixing in the cold leg piping - Heating by the reactor vessel walls c3 - HPI pump energy (minimal) - Heating by the cold leg piping (minimal) - Mixing with ver.t valve fluid The last item, the preheating of the incoming HPI by mixing with vent valve fluid, represents the most significant contributor to reducing the brittle fracture concern. In order to evaluate tr.e thermal shock concern, various thermal r' hydraulic assumptions were made. The major thermal hydraulic assumptions were: 1. Bounding Assumptions Analyses were performed assuning no heatup of HPI due to any c: the ~ above effects. When natural circulation was assumed to be inbibited at m approximately 10 min, into the transient, the downcomer fluid temperature at reacter vessel wall was ramped to the BWST temperature (400 0 F or 90 F) in approximately 60 seconds. This case is essentially a zero mixing case after 10 minutes into the transient. ~ 2. Mix Assumotions Analyses were also performed assuming HPI fluid enters the downcomer, rixes with the warmer vent valve flow, which is assumed to be circumferentially distributed, and then streams down the reactor vessel wall. This is believed to be a more realistic assumption since some degree of HPI mixing and heatup is excected. 2-. 4_ e-une
J J N-n:. F Also, the reactor vessel fluences were obtained frem the Effective Full' Power Years (EFPY) determined frem core follow and the methodology as i outlined in BAW-1511P which was submitted to the NRC on March 12. 1981. } This document reoresents a significant effort as part of the B&W Owners i n 1,.. Group since 1976.(9) (> The EFPY on 3&W operating plants as of 4/27/81 is as follows: 1 i Rtncho Seco~ 3.45 EFPY' Oconee I 4.90 EFPY J Cconee II 4.36 EFpY Oconee III 4.21 EFPY g3 m Crystal River III 2.19 EFPY i TMI-1 3.52 EFPY i-Davis Besse-I 1.25 EFPY ~ ) Arkansas Nuclear One 3.91 EFPY Unit 1 ~ ,1, j no e SAW-1511P also contairs information on Quality Assurance of Reactor Vessel weld properties. This includes weld number, vessel in wnicn located, type of filler wire, type of weldment and various other surveillance caosule measured and predicted information. 4 The analyses in SAW-1648 assumed operator action to throttle high pressure injection such that core outlet conditions would be maintained i less than 1000F succooled. Appropriate revisions to the Small 3reak [ ;p Cperating Guidelines have been issued to the affected Utilities. n l l addition, S&W has recommended to the coerating plants that 3WST i temperatures be maintained greater than the Technical Specification minimum of 40*F. c 3 I ..-.-wy w-,e.. .w- ....,,...s.- -.,%.w w ~. _,,,w ..,.,nr cy, e. -3.. ---.7%, .-,,.,9..y-gg. g
.4 The conservative bounding assumptions were used in the 1980 generic anclyses(3),(4) with the intent being to define the extent of the brittle fracture problem. With these conservatisms, the following j conclusions resulted from the analyses: 1. Rancho Seco and Oconee I reactor vessels represen'. the most,and the w second-most limiting operating S&W units respectively at this point in time. The limiting weit., as analyzed, with respect to brittle fracture in these reactor vessels are longitudinal welds. These vessels have limiting longitudinal welds near the cold leg nozzles. Hence, the analysis of these operating vessels currently bounds all others. 2. Using the conservative bounding thermal-hydraulic assumption (thermal hydraulic cssumption #1 on page 4) plus combining worst case inouts in s the generic analyses showed no immediate brittle fracture concern exists for the operating plants. The analyses show that operator action to throttle HPI flow will preclude brittle fracture. 3 3. Using the more realistic mix assumption (thermal-hydraulic assumption
- 2 on page 4) indicates the most limiting reacter vessel has more than
~ cne additional effective full power year beyond the present Lounding analysis (i.e. into 1983) before any concern is approached, even considering worst-case SWST temperatures. This is illustrated in Figure 1, which shows allowable and actual pressures during the transient for the generic analysis using Rancho Seco weld material properties at 4.8 EFPY, assuming worst-case 40*F BWST water.(3) j The actual Rancho Seco EFPY as of April 27, 1981 was 3.45 EFPY. , M i i I .f. f f e 1
Wh Therefore. given operator action to throttle HPI there is no immediate brittle fracture ccncern for B&W operating units resulting from thermal shocking of the reactor vessel during small break LOCA transients. b III Non-LOCA Overcooling Events i3. NUREG-0737, Item II.K.2.13, recuired that small break LOCA with extended loss of feedwater events be analyzed for reactor vessel brittle f r ac ture. Recently, the ACRS and the NRC have excressed the concern that perhaps other transients, such as steam line breaks. which have the potential for overcooling and subsequent system reoressurization, may be more limiting transients with respect to the reactor vessel brittle fracture concern. 9' As a result of the NRC's request in 1975 (Reference 5), our " position lir regarding these repressurization events has been that operator action to mitigate system repressurization (by throttling HPI and utilizing atmospheric dump or turbine bypass valves) is adequate to keep reactor y coolant pressure and temperature within technical specification limits over the service life of the reactor vessel.(0) ~ Table 1 compares primary system response during various overcooling events. As can be seen, the small break LOCA cases (case 1 and 2) already considered in 3AW-1648 result in more overcooling (to aporoximately 900F downcomer temperature) of the reactor vessel than unmitigated large steam line breaks.(7) Also, case 1, Table 1. clearly bounds all overcooling transients presented in Table 1 (with respect to the m temperature transient). Based on these considerations, plus reliance uoan the operator to mitigate the repressurization. the previous SBLOCA analyses are limiting, with respect to the brittle fracture concern. Assessment of r' the non-LOCA overcooling events (including subsequent repressurization) has confirmed this for operation into 1983. : M
( / 4 cs IV. Concerns Excressed in Basdekas' letter to Udall 4/10/91 i The Basdekas' letter of 4/10/81 has been reviewed and clarifications of several items for B&W designed plants are provided below. The quoted sq sentences have been extracted from the letter 08 A. "Such transients can cause the reactor vessel to cool-Jown to about 150*F in about 15 minutes, while the ECCS repressurizes it to about 2400 PSI." In response to the IE Bulletin 79-05C. and as indicated in Section III, a large steam line break was analyzed. The analysis assumed both OTSG's blowdown, no Main Steam Isolation Valve (MSIV) closure and e .'f Emergency Feedwater at full capacity. The results indicate a minimum Reactor Coolant System (RCS) temperature of 230*F will be reached approximately 14 minutes into the transient.(7) Operator actions to throttle HPI flow will prevent repressurizaticn of the RCS to 2400 j ?SIG. B. "A reactor vessel fracture is one of the most serious accidents a reactor may experience. Depending on its location and mode, it is almost certain that it will cause a core meltdown with all its oublic health and safety ranifications, on which, I an sure, I need not elaborate for you." n It is very unlikely that a reactor vessel fracture, at a location and mode which results in a core meltdown. will occur. This is demonstrated by the positive margins resulting from analyses previously II performed.(1.3.4) _g. r em s u
5. C. "This is suppceted by analyses' performed for the NRC. indicating that the overcooling transient that took place at Rancho Seco en March '20. 1978 would have caused such a vessel to rupture, had it been in operation for about 10 FPYE." j We are not aware of the information that Mr. 3asdekas has, but the ? Q Rancho Seco vessel on March 20, 1978 had only 1.55 EFPY of irradiation and therefore appreciable margin for Brittle Fracture at that time. In an analysis prepared for the NRC by Oak Ridge National Laboratory (ORNL-to Mr. Milton -Vagins (NRC) dated March 3,1981) a different analysis (Warm Prestressing) than that the one used in SAW-1648 indicates that the Rancho Seco vessel has a useful Full-Power Life greater than la EFPY. e 9 0. "Furthermore. a recent discovery of a discrepancy existing between the estimated vs. the measured values of neutron fluence for the Maine Yankee reactor vessel indicates a generic problem that makes things worse. The re.sults of dosimetry measurements indicate the actual ~ neutron fluence to be some 2.3 times higher than that estimate,d in the Maine Yankee Final Analysis Report." The fluence discrepancy at Main Yankee was apparently due to lack of azimuthal flux variation in their calculational medel and/or the use of cycle 1 extracolated data. Azimuthal variaticns in a B&W reactor are en the ordar of a f acter of 2 from maximtm to minimum. Core escace fiux i e is ger.erally lower during cycle 1 (ccmoared to subsequent cycles), and, therefore, ex-core fluences would be low. The fluence analysis crocecure used at S&W accounts for azimuthal flux variation by using the two-diner.sianal transport code 00T to model reactor and surveillance ,Lr capsules. and credicted fluences for extrapolated burnups are based en core escace flux from fuel management studies (P0Q criticality calculations) of future fuel cycles. B&W has always used the two-dimensional moceling approi.5 whereas the initial Maine Yankee data were from a cr a-dimensional model. -g-
s. ti'.i The B&W procedure has been used to calculate the fluence exposure of Cdpsules from five 177 FA reactors, four after cycle 1 and one after cycle 2 Comparisons to measured activities from capsule contained dosimeters have been +1_5%. All calculated data are subsecuently -{ normalized to dosimeter measurements before pressure vessel fluence is a determined. These data are documented in SAW reports that are sent to the appropriate utility after each capsule is analyzed. The B&W procedure was benchmarked when B&W participated in the " Blind Test" phase of the LWR Pressure vessel Surveillance Oosimetry Program, an on-going study of surveillance analysis procedures that is operated by HEDL and ORNL for the NRC. B&W calculated fast flux as documented in n NUREG/CR-1872, " Reactor Calculation Benchmarks - PCA Blind Test Results," January 1981, was within 10% of experimentally derived values at the simulated T/4 pressure vessel location in two experimental configurations. The " Blind Test" results are being documented in a J NUREG report, but data are not identified with respect to participant. E. "Moreover, as you may recall, one of the measures ordered by the NRC af ter the TMI-2 accident was to have all reactor operators not turn off the ECCS once it had been initiated." Revised Small-Break LOCA Coerating Guidelines have been issued to affected Utilities by B&W. The guidelines provide operator instruction i en when to throttle the HPI flow to prevent repressurization. 1 I 4 ~,,.,.w r-- -,3-- -1 e
~ T'. V. Other Actions - The thermal shock concern hcs been addressed and programs have been either completed, currently underway or planned to assure safe operation. 4 A. Comoleted tt 1. ICS/NNI upgrades per IE Bulletin 79-27 and associated Commission U orders. 2. EFW Systems Upgrades 3. Revised Small-Break LOCA Operating Guidelines regarding thermal shock have been issued to affected utilities. These guidelines are intended to: - Enhance understanding. - Provide operator instruction for HPI throttling on succooling when n in the HPI cooling mode with no RCS flow. - Emphasize re-establishing RC Loop Flow. 4. Abnormal Transient Operating Guidelines ( ATOG) procedure under development to address item I.C.1.in NUREG-0737 include censideration j of the brittle fracture concern. ~ S. B&W has recommended that Utilities maintain SWST temperatures higher c. than Technical Specification minimums. 6. BAW-1511P (reference 9) has been completed as part of an Owners Group program en Reactor Vessel materials. B. Currently Underway 1. The Owners Group reactor vessel materials program is geared toward demonstrating adequate structural integrity of the reactor vessel j throughout plant design life. Efforts currently underway include: l a determination of fracture toughness properties which are l expected to demonstrate higher resistance to fracture than EI i current industry predictions based cn Charpy V notch specimens. 4 - the development of less conservative fracture analysis procedures, which include elastic-plastic technicues. 2. Reactor Vessel Material Surveillance Programs in accordance with I Appendix H of ICCFRSO. l ,_y., -_..~,..._..
M -a C. Immediate Future Plans f 1. Plant specific evaluations to address the conservatisms associated with generic analyses are being investigated ? 2. More sophisticated vessel cooldown calculations are _ being considered to reduce the conservatisms associated with the one-dimensional heat ~ conduction analysis previously employed. 3. Consideration of analysis for Non-LOCA events. D. Long Term Plans 1. Discussions are in progress with EPRI regarding possible testing to obtain a better understanding of the thermal-hydraulic mixino phenomena associated with these overCooiing transients. 2. CREARE, Inc. and other consultants have been contacted and involved in discussions concerning the thermal-hydraulic mixing ascects of the problems. 3. The investigation of enhanced inservice inspections methods with the objective being the reliable detection of smalle flaw sizes. 4. The evaluation of the in-place reactor vessel thermal annealing to ~ recover some of the material properties lost through neutron c. irradiation. 5. The investigation of imoroved dosimetry and ficci.ee calculations. VI Summary - Justification for Continued Oceration As a result of the NRC's recuest of March 31, 1981 to put the reactor vessel brittle fracture issue in perspective, the following r. ave been concluded-A. Assessment of overcooling events indicates that the small break LOCA event as analyzed is bounding. B. Generic analyses (including mixing) of the small break LOCA events show v no immediate problem (into 1983) given operator action. C. Revised operator guidelines have been issued. Immediate acerator action is not required. Required operator action is straightforward. D. Efforts are underway to resolve the long term issue. =,.. ,-.-w . i
i e'. er Table 1 Primary System Resoonse During Overcoolina Transients i Minimum Cooldown Downccmer 4 c. Case Description Rate Temoerature Comment ,3 BAW1648(ggunding 1 460F in 90F No temperature Analysis / 60 Seconds recovery (4600 / min) No recressurization* F (90F SWST) 2 BAW 1648(Mix 445F in An alysi s 3) 40 minutes 90F No temperature (11.10 / min) recovery F i No reoressurization* (4CF SWST) 3 Unmitigated Large 320F in 230F Temperature recovers A I Steam L ga 10 Minutes System repressurizes** Rupture I (320 / min) F 4 Rancho Seco Rapid 310F in 285F Some temperature Cooldown Ig 60 Minutes recovery of3/20/781gjdent (5.2*F/ min) Stable pressure between 1400 and 2100 psig Assuming operator action a Can be mitigated by operator action 2 dB 'l I e I r: 4 B A-1
i s. 3 3 Ls t ';! E g; $ s.. 1 figure i ALLOWABL E AND ACTUAL PRESSURES VS TIME, 0.023-fl 2 PRESSURIZER BREAK WlIH OPERATOR ACTION, RAllCil0 SEC0, 40f BWST, MIX 2 NOTE: BASED ON GENERIC ANALYSES AND ASSOCIATED 500 CONSERVAllSMS DOCUMENIEI) IN REFERENCE 3. 2000 ^ i J. O
- u 1500 d
L' 1000 4.8 EfPY ALLOWABLE PflESSURE
- 1/1/83 BASED ON 100';
CAPACITY FROM 4-21-81 s 1 500 1 ACIUAL TitANSl[NI PRESSURE f r I i 1 2 3 Time (hrs) 5 t
!=. - 0 References (1) Analysis of the Structural Integrity of a Reactor Vessel Subjected to i Thermal Shock, BAW-10018, Sabcock & Wilcox, Lynchburg, Virginia, May 1969. Transmittal letter, J. H. MacMillan (S&W) to Dr. P. A. Morris ( AEC). Dated May 22, 1969. d (2) NUREG-0737, Item II.K.2.13 (3) Thermal-Mechanical Report - Effect of HPI on Vessel Integrity for Small Break LOCA Event with Extended Loss of Feedwater, SAW-1648. Sabcock & Wilcox, Lynchburg, Virginia, Novemoer 1980. (a) Reacto. Vessel Brittle Fracture Analysis During Small Break LOCA Events with Extended Loss of Feedwater, BAW-1628, Sabcock & Wilcox, Lynchburg, Virginia. December 1980. (5) F. Schroeder (NRR) to K. E. Suhrke -(B&W), Letter dated June -10,1975. (6) K. E. Suhrke (B&W) to F. Schroeder (NRR), Letter dated August 12, 1975. (7) J. J. Mattimoe (SMUD) to R. H. Engelken (NRC), October 24, 1979. (8) Ccmmittee Report on Rancho Seco Unit 1 Transient of March 20, 1978, Dated June 19, 1978, by Sacramento Municipal Utility District. (9) Irradiation-Induced Reduction in Charpy Upper-Shelf Energy of Reactor Vessel Welds, BAW-1511P (Proprietary), Babcock & Wilcox, Lynchburg, Virginia, October 1980, Transmittal Letter, J. H. Taylor (B&W) to r-J. S. Serggen (NRC), Dated March 12, 1981. as A-3 l a.o}}