ML18092A484

Safety Evaluation for Replacement Interval of Undervoltage Trip Attachment & Reliability Calculation of DB-50 Reactor Trip Breaker.

ML18092A484
Person / Time
Site:	Salem
Issue date:	11/21/1984
From:	Public Service Enterprise Group
To:
Shared Package
ML18092A483	List: ML18092A482 ML18092A484
References
S-C-R120-CSE-28, S-C-R120-CSE-287, NUDOCS 8502150019
Download: ML18092A484 (5)

Text

... *' ~ *' '* " "' ,> >' .>. -h ,,, '" * >> **. -~ ,*> *' ~ <> ., ;. '-. I\ *'< ,; *' ,* *'* , N '* ' ~. " " ."" .~**

- Page. 1 of 5 S-C-Rl20-CSE-287 Date 11/21/84 Public Service Electric and Gas Company P.O. Box 236 Hancocks Bridge, New Jersey 08038 TITLE: SAFETY EVALUATION FOR THE REPLACEMENT INTERVAL OF THE UVTA AND RELIABILITY CALCULATION OF THE DB-50 RE.ACTOR TRIP B.RE,AKER

~

1.0 PURPOSE To document and justify the replacement interval of the Undervoltage Trip Attachment and the reliability calculations for the Westinghouse DB-50 reactor trip breakers.

2.0 SCOPE This document applies to the Westinghouse DB-50 reactor trip breakers used at Salem Generating Station.

3.0 REFERENCES

3.1 Westinghouse summary of Test Results NS-EPR-2824, Letter from E. P. Rahe (Westinghouse) to H. Denton (USNRC),

October 1,, 1983.

3.2 IEEE Standard 352, 1975 Guide for General Principles of Reliability Analysis o.f Nuclear Power Generating Station.

Protection Systems *

. 3.3 Operational verification Program Reactor Trip Breakers, Letter from R~ A. Uderitz (PSE&G) to D. G. Eisenhut (USNRC), May 31, 1983.

3.4 Letter from s. Varga (USNRC) to R. A. Uderitz (PSE&G), July 26, 1983.

3.5 WCAP 10426, Reliability Estimates of the UVTA's in Reactor Protection System Trip Breakers. (January 1984) 3.6 Franklin Research Center Review Operational verification Program for Westinghouse DB-50 Reactor Trip Breakers, Letter from E. A. Liden (PSE&G) to s. A. Varga (NRC) dated September 16, 1983.

4.0 BACKGROUND

- PSE&G committed to a test program to determine the life cycle and replacement interval for the UVTA's and to verify the adequacy of the maintenance and surveillance programs used on the DB-50 reactor trip breakers. Reference 3.3 identified the Westinghouse Breaker Test Program as the test program to fulfill the PSE&G commitment. The NRC requested that PSE&G respond to RTB/l_J_ _ ---*------ ---- ----,

,-,-8502150019 850207  ;

PDR ADOCK 05000272

, **p - PDR EDD-7 FORM 1 REV 0 .lOSEPT.81*

95-2168 (100M) 12-~

2 of 5 S-C-Rl20-CSE-287 Date: 11/21/84 issues raised by Franklin Research Center - NRC consultant in Reference 3.4. In addition this letter requested that PSE&G address the confidence interval, sources of bias, and factors that influence service life.

The Westinghouse test program comprised two separate phases.

Phase I of the testing utilized five UVTA 1 s. Four devices were actually tested. The fifth device was maintained as a control sample for comparison purposes at the completion of the test.

In Phase I A, two UVTA 1 s were cycled a total of 2500 times with periodic lubrication and inspection every 200 cycles. In addition, the Phase I A test included 600 operations utilizing the shunt trip attachment. The other two devices were subjected to 2500 UVTA cycles without periodic lubrication, during Phase I B.

In Phase I A, one UVTA test sample performed without failure to trip the total test run of 2500 UVTA and 600 shunt trip cycles.

The other UVTA, after the 200 cycle lubrication and inspection, exhibited several intermittent misoperations. Physical and visual examination indicated that the lubricant had not properly carried into the intended points on the UVTA. Both UVTA's were removed from the test breakers thoroughly lubricated and the test was continued utilizing a similar lubrication procedure throughout the rest of the test. Both devices performed for the remainder of the test without failure.

The two test samples from Phase I A and the control sample were then disassembled and subjected to a metallurgical examination to establish the significance of the service wear on the performance of the UVTA during service life. The wear at three different locations on the attachments was determined to be the most significant. These locations were: (1) the push rod to pin joint, (2) the bottom surface of the latch where it makes contact with the flat leaf spring and (3) the leading edge of the guide notch of the latch where it makes impact sliding contact with the latch pin. An additional wear scar was identified during this examination and was part of the basis for instituting Phase II testing.

In Phase I B one test sample performed its trip function for 571 cycles before it failed to latch and the second test sample performed its trip function for 1700 cycles before it malfunctioned.

Phase II of the testing was conducted to evaluate a design modification that would capture the joint pin and to evaluate a new lubricant application technique. The design modification prevents the lateral pin motion which was identified during the metallurgical examination in Phase I A.

RTB/l 2 EDD-7 FORM 1 REV 0 10SEPT81

3 of 5 S-C-Rl20-CSE-287 Date: 11/21/84 Phase II testing was similar to Phase I testing which included inspection and lubrication every 200 cycles for two UVTA's, (Phase II A) and only initial lubrication and inspection on two UVTA's (Phase II B). During Phase II A, both UVTA test devices performed 2500 UVTA and 600 shunt trips without malfunction.

During Phase II B, one device performed 2500 UVTA cycles without trip malfunction while the other device performed appproximately 2390 cycles before it malfunctioned. The metallurgical examination of these devices indicated similar wear patterns as those found during Phase I with the exception of the casing wear which was removed by the captive pin modification. As a result of the Phase II testing, it was determined that the captive pin modification did not adversely affect UVTA performance and that proper device lubrication could be conducted with the UVTA installed on the breaker.

In addition to the testing, Westinghouse performed a statistical evaluation of actual plant operating history covering the use of the UVTA in DB-50 reactor trip switchgear applications for the Westinghouse Owners Group. This study included 26 operating plants, some of which performed maintenance on the breakers and some which have not performed maintenance. The total UVTA population cycles were 9984 operations and of these operations 19 failures were reported. Based upon this data we estimate the reliability to be .9981.

5.0 DISCUSSION PSE&G is confident that the testing performed provides valid data with respect to the installed reactor trip breakers at the Salem Generating Stations. The Phase II A testing described above tested approximately 12.5% of a typical UVTA production batch (16 UVTAs) utilizing identical lubrication procedures and intervals actually used by PSE&G. The samples tested were selected from a manufacturing program which requires 100%

dimensional inspection of ten critical parts and a post assembly acceptance test of twenty-five successful operations at the manufacturing facility. This manufacturing program is in accordance with the PSE&G purchasing specification for the UVTAs. In addition the main breakers at both Salem Units will be or have been replaced with new breakers similar to those tested.

RTB/l 3 EDD-7 FORM 1 REV 0 10SEPT81

S-C-Rl20-CSE-287 Date: 11/21/84 ea 4 of 5 In Reference 3.6, one of the Franklin Research Center's comments was that each time the breaker is operated from any source (UVTA, shunt trip, or manual trip) the UVTA is either partially or fully cycled and should be counted as a UVTA cycle. During Phase II A testing, each breaker was subjected to 2500 UVTA operations and 600 shunt trip operations without failure for a total of 3100 breaker operations. From IEEE Standard 352-1975 (Reference 3.2), a binomial distribution is the distribution of the number of successes in 11 n 11 independent Bernoulli trials in which the probability of success at each trial is P. The estimate of the parameter P is;

/\o p = number of successes total number of trials Since there were zero (0) failures during Phase II testing one (1) failure will be assumed in order to calculate a meaningful value of the parameter P, and therefore

/"-

p = 6199 6200, assuming (1) failure

/\.

p = .9998387 Also from Reference 3.2 an approximate confidence interval for P is given by

~+ c~ (1~1~ ~ (1~;1 2 2 Z d/2 P < -; + Z1- <>'f2 Where z cl.( 2 = -1. 96 Zl- o{/2 ~ +l. 96 n = 6200 trials From the above, the estimated confidence interval on P is

.9995226 < P < 1.0; at a 95% confidence interval It should be noted here that the calculated reliability from Phase II A test data also supports the plant evaluation study performed by the Westinghouse Owners Group.

The present replacement interval used by PSE&G for the UVTA is each refueling outage. This interval corresponds to a conservative 400 breaker cycles for an 18 mon~h refueling RTB/l 4 EDD-7 FORM 1 REV 0 10SEPT81

5 of 5 S-C-Rl20-CSE-287 Date: 11/21/84 cycle. The Westinghouse test program tested the UVTA to twice the Westinghouse service life of 1250 cycles. PSE&G chooses to continue the present replacement interval of each refueling outage. With this information, the reliability or the probability of zero (0) failures in "N" trials is given in Reference 3.2 as Rn = pn Rn = (.9998387)400 Rn = .9375125 Also, since the reactor trip breakers are redundant components the reliability or the probability of zero (O) failures to drop the control rods utilizing the 400 cycle replacement interval is

.9960953.

6.0 CONCLUSION

When lubricated and maintained every 200 breaker operations in a similar manner to that used in the Westinghouse Phase II A testing program, the DB-50 reactor trip breaker will exhibit an approximate reliability of .9998387. The 95% confidence interval for this reliability value indicates that the actual reliability is somewhere between .9995226 and 1.0. PSE&G chooses to replace the UVTAs at each refueling outage which will equate to approximately a 400 cycle service life, which is much more conservative than the manufacturer's recommended replacement interval of 1250. The resultant reliability of zero (0) failures during any 400 cycle life is .9375125. Since the reactor trip breakers are redundant, the reliability of the reactor trip breakers to drop the control rods into the reactor with zero (0) failures during the UVTA service life at Salem will be .9960953.

PSE&G has also modified the reactor trip circuit by installing an auto shunt trip feature. The auto shunt trip feature operates the shunt trip attachment on the main reactor trip breakers whenever a solid state protection system trip signal is generated or the console manual reactor trip switches are operated. With this design feature both the UVT and shunt tripping devices are actuated simultaneously whenever a reactor trip is demanded. The above testing described actuated only one of the two tripping devices with each trial, therefore the above reliability values are minimum reliability values since breaker operation is no longer dependent upon the successful operation of just one of the tripping devices.

RTB/l 5 EDD-7 FORM 1 REV 0 10SEPT81