ML20197D980
| ML20197D980 | |
| Person / Time | |
|---|---|
| Site: | Maine Yankee |
| Issue date: | 05/12/1986 |
| From: | Randazza J Maine Yankee |
| To: | Eisenhut D Office of Nuclear Reactor Regulation |
| References | |
| 7560L-SDE, MN-86-64, NUDOCS 8605150129 | |
| Download: ML20197D980 (11) | |
Text
- ------------
MAME HARHEE AmMMPOWER00MPARUe
,oougr,Ja?n"eo"als May 12, 1986 (207) e23-3521 HN-86-64 Director of huclear Reactor Regulation United States Nuclear Regulatory Commission Washington, D. C.
20555 Attention:
Mr. Darrell B. Eisenhut, Acting Director
References:
(a)
License No. DPR-36 (Docket No. 50-309)
(b) USNRC Letter to MYAPCo dated April 23, 1986.
Safety Significance of Reactor Coolant Pump Shaft Integrity Issue
Subject:
Maine Yankee Reactor Coolant Pump Shaft Integrity Gentlemen:
Your letter, Reference (b), summarized the recent findings of failure and cracking in reactor coolant pump (RCP) shafts manufactured by Byron Jackson Company. You requested that, pursuant to 10 CFR 50.54(f), Maine Yankee provide additional information on the integrity of its RCP shafts to justify continued plant operation and our plans and schedules for inspecting the shafts and other structural components of the RCPs.
This letter provides you with the required justification for the continued operation of Maine Yankee. Justification for continued operation is based upon several factors.
1.
Our assessment of a postulated RCP shaft failure indicates that it would not result in safety consequences more severe than a seized pump rotor which is a license basis event.
This is consistent with the Crystal River experience.
2.
Pump shaft failure probability at Maine Yankee is not more likely because of the Crystal River experience.
There are several significant design differences between the design of the Maine Yankee and Crystal River pumps.
3.
The examination of the pump shaft from the first RCP disassembled at Davis Besse, which is more similar to the Crystal River design than J
to the Maine Yankee design, has indicated no circumferential cracking i
in the areas of the shaft previously identified by U-T examination i
prior to pump disassembly.
From our review of the shaft failure information from Crystal River, differences in RCP designs, and examination of the first Davis Besse pump shaft, the circumferential j
shaft cracking found at Crystal River and the subsequent shaft failure may not be generic to all Byron Jackson RCPs having ASTM A461 GR660 shaft material. He also believe that the impeller bolting and impeller drive pin cracking problem seen at other plants is less likely at Maine Yr kee because of design differences.
7560L-SDE e605150129 e60512 bg PDR ADOCK 0500 9
i P
j
MAINE YANKEE ATOMIC POWER COMPANY United States Nuclear Regulatory Commission Page Two Attention: Mr. Darrell B. Eisenhut, Acting Director MN-85-64 Notwithstanding our justification for continued operation, we have conducted a special instruction program for our operators appropriate to the event including simulator training.
Furthermore, we are continuing to evaluate all new information on this subject as it becomes available.
Maine Yankee has retained Babcock and Hilcox (B&H) to provide technical assistance in this evaluation. Maine Yankee's engineers assigned to this project are in frequent contact with cognizant individuals at the affected utilities, B&H, and Byron Jackson.
At the present time, we intend to continue to monitor the results of the ongoing investigations regarding the causes of RCP shaft cracking and factor this information into our plans for inspecting Maine Yankee's RCP shafts during the next refueling outage.
Maine Yankee's specific responses to your request for additional information are contained in Attachment A.
If you desire, we would be available to discuss this information and our assessment of the integrity of Maine Yankee's RCP shafts with members of your staff at your convenience.
Very truly yours, MAINE YANKEE ATOMIC P0HER COMPANY John B. Randazza Executive Vice President JBR/bjp
Attachment:
(Attachment A) cc: Mr. Ashok C. Thadani Dr. Thomas E. Hurley Mr. Cornelius F. Holden Mr. Pat Sears STATE OF MAINE Then personally appeared before me, John B. Randazza, who being duly sworn did state that he is Executive Vice President of Maine Yankee Atomic Power Company, that he is duly authorized to execute and file the foregoing request in the name and on behalf of Maine Yankee Atomic Power Company, and that the statements tr.erein are true to the best of his knowledge and belief.
t h /0 k
a, Notary.Public 56%-SM y C0ujj,,
,,pg33,333,
MAINE YANKEE ATOMIC POWER COMPANY ATTACHMENT A MAINE YANKEE ATOMIC POWER COMPANY RESPONSE TO NRC INFORMATION REQUEST ON REACTOR COOLANT PUMP SHAFT INTEGRITY 1.
NRC RE00EST
[ Provide] a description of the design and operational history of the Maine Yankee RCPs, which are different from the design and/or operation of the Crystal River, Unit 3, and Davis Besse 1 RCPs.
MAINE YANKEE RESPONSE A review of the design and operating history of the Maine Yankee RCPs has been performed to provide the following summary of differences with respect to the Crystal River and Davis Besse RCPs.
Shaft Cracking j
The design of the Maine Yankee RCP shafts and hydrostatic bearing connection has unique differences which make it unlikely that the failures seen at Crystal River would occur at Maine Yankee.
These differences include the following:
Maine Yankee RCP shafts do not have the split ring groove machined in the pump shaft beneath the hydrostatic bearing journal as does the Crystal River shafts. This is significant, since this groove i
constitutes a stress riser and the circumferential cracking and shaft failure of the "A" RCP shaft at Crystal River initiated in this area. According to Babcock and Hilcox, Davis Besse also does not have this split ring groove under the hydrostatic bearing journal and examination of the Davis Besse shaft following pump disassembly has shown no cracking of the pump shaft in this area or any other area under the hydrostatic bearing journal sleeve.
The circumferential crack found in the "B" RCP shaft at Crystal River appears to be in the heat affect zone of a weld between the top of the hydrostatic journal connection and pump shaft. The existence for the weld in this region cannot be explained. Welding ASTM A461 GR660 material in the heat treated (hardened) condition is not recommended nor was it included by the manufacturer in his design.
The Maine Yankee RCP pump shaft design does not specify, or indicate, welding in this area. According to Babcock and Hilcox, the pump shaft examined after pump disassembly at Davis Besse does not have welding in this area.
It is also our understanding that the other RCPs at Crystal River do not have such welding and circumferential cracking of the other shafts have not been observed in this region.
7560L-SDE m-
_. _.. -... ~ _. - -.
MAINE YANKEE ATOMIC POWER COMPANY IE Information Notice No. 86-19 stated that the licensee attributes the shaft failure on Pump A at Crystal River to." residual fabrication stresses coupled with thermal stresses from cool seal water injection." Since Davis Besse did not find circumferential cracking following pump disassembly, it is difficult to conclude how much the i
relatively cool seal injection water entering the pump case in the area of the hydrostatic bearing at Crystal River could have contributed to the circumferential cracking problems. Since this type of cracking was not observed at Davis Besse on the disassembled pump, the cracking problems observed at Crystal River may be related 4
to the specific shaft geometry (split ring groove) and may not be a generic problem to all pumps having seal injection.
l Comparison of the drawings for the Crystal River and Maine Yankee RCPs indicate that the length of the Maine Yankee throttle bushing region of the pump cover is different than that at Crystal River and 4
l Davis Besse.
The length of the throttle bushing in the Crystal River and Davis Besse pumps is shorter than in the Maine Yankee pump design and the length of the drilled hole heat exchanger,,in the Crystal River and Davis Besse design, extends approximately the entire length of the throttle bushing. The length of the bushing in the Maine 3
j Yankee pump design appears to extend approximately an additional four i
and one-half inches below the drilled hole heat exchanger before j
entering the pump casing. He feel that this is significant since the i
seal injection water will gain heat in this area before entering the j
pump case and therefore, should mitigate the chilling effect on the shaft in the area of the throttle bushing outlet.
Examination of the ACME thread region of the pump shafts at Crystal River has shown shallow axial cracks in the lower end of the ACME i
threads and a " craze" pattern below the threads. The ACME thread i
region is above the hydrostatic bearing journal of the pump shaft within the area of the throttle bushing. This type of indication has been observed on other pumps manufactured by Byron Jackson with seal 1
i injection.
It is our understanding that this type of cracking is caused by thermal stresses which do not exist at a depth sufficient i
to generate a critical crack size and thus are believed to be self j
arresting based on experiences in General Electric BHRs having Byron Jackson recirculation pumps. The circumferential cracking observed 1
in the "A" RCP at Crystal River does not appear to eminate from this i
region, therefore it does not appear that the axial cracking is i
linked with the observed circumferential cracking and shaft failure.
Inspection of a pump shaft taken out of the "C" RCP at Maine Yankee in 1973 after approximately eight months of operation has shown negligible axial cracking on the lower ACME threads. He
{
feel that the longer throttle bushing arrangement at Maine Yankee as i
described above may have minimized axial cracking in this area. Very fine axial cracks have been found on,the very top of the ACME threads i
within approximately a 70 degree arc of the 360 degree circumference of the thread. The cracks appear to be very shallow and difficult to identify. A " craze" pattern was not observed below the ACME 1
threads.
Based on experience in General Electric plants, different degrees of cracking in this region are not unusual and does not appear to have specific bearing on the circumferential type shaft j
cracking seen at Crystal River.
l 7560L-SDE l
i
MAINE YANKEE ATOMIC POWER COMPONV
)
Imoeller Boltino and Drive Pins In accordance with discussions with Byron Jackson, the RCP design at Maine Yankee attaches the impeller to the pump shaft with eight 1 1/4" - 8 UN diameter A461 GR660 cap screws torqued to 600 ft-lbs located on a 9 1/2 inch bolt circle.
In addition to the bolting there are four 1 3/4 inch l
diameter A461 GR660 drive pins located on a 101/4 inch bolt circle used to transfer the torque from the pump shaft to the impellar. The RCP designs at Crystal River and Davis Besse differ from that at Maine Yankee in that the impeller is fastened to the pump shaft with four 1 1/2" - 8 UN diameter cap screws on a 8 5/8 inch bolt circle and torqued to 300 ft-lbs. The drive pins are the same as at Maine Yankee with the exception that they are located on a 9 1/2 inch bolt circle.
Because of the increased number of fastners and design preload we do not expect the performance of the bolted joint to be the same at Maine Yankee as at Crystal River and Davis Besse. The design preload of the Crystal River and Davis Besse designs is approximately 27 percent of the design preload in the Maine Yankee shaft impeller connection. Therefore, if low cycle fatigue is the root cause of cap screw failure at Crystal River or Davis Besse, the additional preload margin provides substantial assurance that the same problem does not exist at Maine Yankee.
Ooeretion History He have identified no difference in the Maine Yankee RCP operation which would apply to the problems identified at Crystal River, with the exception that Maine Yankee's RCPs develop a lower discharge pressure and have approximately twice as many hours of operation. During our operating history of the Maine Yankee pumps, we have not cbserved operational changes that would indicate shaft degradation.
2.
NRC REOUEST
[ Provide] the results of any analysis performed subsequent to those done for the FSAR which would address the consequences of a locked rotor or broken shaft event during plant operation.
MAINE YANKEE RESPONSE Maine Yankee's original Final Safety Analysis Report (FSAR), which formed the basis for the operating license, contained an evaluation of the consequences of a seized rotor on a single Reactor Coolant Pump (RCP).
The results reported in the original FSAR indicated that less than 7.51. of the fuel rods would experience Departure from Nucleate Boiling (DNB) with acceptable offsite dose consequences.
l 7560L-SDE J
r
/
MAINE YONKEE ATOMIC POWER COMPANY The seized rotor analysis was updated in References 1 and 2 for the Maine Yankee power uprate to 2630 MHt and to account for operation with positive moderator temperature coefficients.
The updated analyses, as in the original design basis, assumed that offsite power remains available.
Each fuel cycle, the seized rotor analysis is updated to demonstrate that offsite doses are within 10CFR100 limits fo the reloaded core. The most recent analysis, supporting Cycle 9 operation, was reported in Reference 3.
A conservative estimate for Cycle 9 indicates less than 7.5% of the fuel rods would experience DNB and that offsite doses would be within the limits of 10CFR100.
In response to Reference 4. Maine Yankee conservatively evaluated the consequences of a seized rotor coincident with loss of offsite power for Cycle 8.
The results of the evaluation, provided in Reference 5, conservatively demonstrated that less than 8% of the fuel rods would experience DNB and that offsite doses would be within the limits of 10CFR100.
He have also performed an assessment of these licensing analyses in order to provide a realistic assessment of the consequences of a Reactor Coolant Pump shaft breaking during full power operation. A summary of our assessment follows:
As noted atove, the current licencing enalysis indicated less than 7.5% of the fuel rods would experience DNB and that the offsite doses would be within the limits of 10CFR100. Table 1 provides a list of conservative assumptions used in our licensing analysis of a locked RCP rotor. The most important of these assumptions, with respect to offsite consequences, are 1) fuel failure is assumed to occur if DNB is predicted, 2) core power distributions are based on a design shape, 3) most positive moderator temperature coefficient is assumed, and 4) primary to secondary steam generator tube leakage is assumed to be present.
Departure from nucleate boiling (DNB) indicates a degraded heat transfer condition and does not correspond.to fuel rod failure.
Time at temperature would be a more representative indication of possible failure. Since the seized rotor event is very~short in duration fuel failure is significantly overpredicted by the licensing methods.
Core power distributions assumed in the licensing analysis are representative of operation at the edge of the symmetric offset operation band and include axial shapes from adverst yenon transients.
Power distributions more typical-of norm l plant operation would rasult in significant improvements in DNB performance.
l 7560L-SDE L
MAINE VONKEE ATOMIC POWER COMPANV The moderator temperature coefficient assumed in the licensing analysis was the most positive projected at any time in core life. A positive moderator coefficient is actually present only at the beginning of core life while at low power.
The seized rotor event is not limiting from low power. The existing negative temperature coefficient would aid in reducing power during the flow coast down and would lead'to improved DNB performance.
The maximum technical allowable primary to secondary steam generator tube leakage is assumed in the licensing analysis. Maine Yankee does not have a steam generator tube leak and has not had one in thirteen
, years of operation.
Therefore, our evaluation of the consequences of a seized rotor event at Maine Yankee using realistic assumptions would result in minimal, if any, fuel; failure and negligible offsite dose consequences.
We also considered the possibility of a loss of offsite power coincident with a broken RCP shaft.
In our opinion the probability of a loss of offsite power due to random events occurring precisely at the time of a RCP shaft break is extremely remote. A broken RCP shaft would not cause a loss of offsite power since a broken shaft would merely cause the motor to go to zero load amps.
A loss of offsite power coincident with the reacter trip and turbine trip is also unlikely since it has never happened at Maine Yankee in over thirteen years of operation. Nevertheless, the unit trip could result in an electrical system disturbance which could be postulated to cause a loss of offsite power.
For this reason we evaluated the effects of a loss of offsite power coincident with a unit trip following a postulated RCP shaft break. He concluded for the reasons stated above that an analysis using realistic assumptions would show little if any offsite consequences.
3.
NRC RE00EST
.Considering the higher probability than previously envisioned of a postulated RCP shaft-failure, describe any actions you have implemented or have planned such as operator review and associated training concerning the specific events at' Crystal River, Unit 3, and Davis Besse I and monitoring plant parameters such as primary to secondary reactor coolant leakage.
MAINE YANKEE RESPONSE IE Information Notice 86-19, the March 17, 1986, Inside NRC article and a memo from the Manager, Operations Department, which discussed expected indications of a RCP shaft. failure, were discussed by Plant Shift Superintendents with all licensed operators.
7560L-SDE
M AINE Y ANKEE ATOMIC POWER COMPANV As a result of the events at the Crystal River Plant, we modified our current Licensed Operator Requalification Simulator Training to include a seized RCP rotor as the initiating event for a functional recovery from a subcriticality challenge. The training session included a discussion of the following major points using I&E Notice 86-19 as a guide:
alarms received by Crystal River / Davis Besse operators different accident scenarios possible (including LOCA) use of our E0Ps during broken shaft or locked rotor accidents The Maine Yankee Licensed Operator Qualification and Requalification courses include classroom and simulator training on diagnosing sheared RCP impeller and seized RCP rotor events from control board indications.
I 1
)
i 7560L-SDE
M AINE YANKEE ATOMIC POWER COMPANY REFERENCES 1)
P. A. Bergeron, et al, " Justification for 2630 MHt Operation of the Maine Yankee Atomic Power Station", YAEC-ll32, July 1977.
2)
P. J. Guimond, et al, " Justification for Operation of the Maine Yankee Atomic Power Station with a Positive Moderator Temperature Coefficient",
YAEC-1148, April 1978.
3)
" Maine Yankee Cycle 9 Core Performance Analysis", YAEC-1479, April 1985.
l t
)
7560L-SDE
MAINE YANKEE ATOMIC POWER CCMPANY TABLE 1.
CONSERVATIVE ASSUMPTIONS UTILIZED IN THE MY LOCKED ROTOR LOSS OF FLOH ANALYSIS 1.
Maximum allowable reactor power plus a 2% uncertainty is assumed.
2.
Minimum allowable Reactor Coolant System pressure including a 25 psi uncertainty is assumed.
3.
Maximum allowable core inlet temperature plus a 4*F uncertainty is assumed.
4.
RCS flow in the affected loop instantaneously goes to zero.
5.
A maximum allowable moderator temperature coefficient of
+0.5X10-4 delta p/*F is assumed.
6.
The BOC Doppler coefficient is used including a 25% calculational uncertainty.
7.
A conservatively low, bounding value for CEA reactivity insertion worth is utilized. Minimum Tech. Spec. allowed BOC worth is used.
8.
A maximum CEA insertion time of 3 seconds is assumed.
9.
The rate of CEA reactivity insertion is skewed to represent a bottom-peaked axial core power distribution.
- 10. Coolant thermal conditions and MDNBR were calculated using power distributions which are conservative with respect to the worst power distribution limited by the symmetric offset LCO band.
~
- 11. Complete fuel rod failure is assumed for any rods which undergo a DNBR of less than 1.20 (YAEC-1 correlation).
- 12. The maximum allowable primary-to-secondary steam generator tube leakage is present.
- 13. All steam generator tube leakage is occurring in only one steam generator.
- 14. A worst-case single failure of a stuck-open steam generator Code safety valve is assumed.
- 15. The stuck-open Code safety valve is on the steam generator having a maximum allowable primary-to-secondary tube leakage.
- 16. The Code safety valve remains stuck-open for 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> while the plant cools down.
- 17. The affected steam generator is assumed to boil dry, so no credit is taken for partitioning factors.
7560L-SDE
M AINE YANMEE ATOMIC POWER COMPANY l
TABLE 1 (Continued)
- 18. The maximum allowable steam generator tube leak rate continues throughout the 8-hour plant cooldown process.
- 19. The maximum allowable iodine inventory is assumed to exist in the reactor coolant prior to the event.
(ie:
Tech. Spec. limit on iodine spiking, 60 uCi/gm D.E. I-131.)
- 20. Noble gas releases are based on assuming the maximum allowable reactor coolant inventory exists prior to the event.
(ie:
Tech. Spec. limit of I
100/E uCi/gm.)
- 21. Fuel rod gap inventory analysis assumes the gap activity equals 10% of the total core activity.
l I
l l
7560L-SDE
_