ML20205R668

Safety Evaluation Supporting Continued Use of Static O-Ring Differential Pressure Switches

ML20205R668
Person / Time
Site:	LaSalle
Issue date:	04/01/1987
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20205R644	List: ML20205R639 ML20205R668
References
NUDOCS 8704060409
Download: ML20205R668 (13)

Text

_ _ _ _ - _ _ _ _ - _ _ _ _ _ _ _ - _ _ .

b UNITED STATES O ^

o NUCLEAR REGULATORY COMMISSION h

j WASHINGTON, D. C. 20555 e

SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION SUPPORTING CONTINUED USE OF SOR INC. DIFFERENTIAL PRESSURE SWITCHES AT LA SALLE COUNTY STATION, UNITS

• AND 2 COMMONWEALTP EDIS0N COMPANY DOCKET NOS. 50-373/374

1.0 INTRODUCTION

On June 1, 1986, La Salle County Station Unit 2 was operating at 95 percent power when a feedwater transient occurred which caused reactor vessel level to drop to the level 3 trip setpoint. One of four level 3 trip channels actuated resulting in a half scram. The operators were able to restore level to the normal operating range and continue plant operation. Subsequent investigation determined that the reactor vessel level had dropped below the level 3 trip setpoint with one out of four channels actuating, which did not result in a reactor scram. Commonwealth Edison Company (licensee) declared an " alert," shut the plant down, and notified the NRC. At the time of the Unit 2 event, La Salle County Station Unit I was at cold shutdown for a scheduled refueling outage.

Further investigation at Unit 2 determined that the event was due to a failure of three of the reactor vessel level 3 differential pressure (dp) switches to trip at their set level. A special level drop test was per-formed to check the actual setpoints of each level 3 switch while at oper-ating temperature and pressure. With the reactor at 950 psig, the reactor water was slowly lowered out of the normal operating range down to the trip point of each level 3 switch. The results were erratic with the switches tripping at levels between 3.9 and 10.2 inches versus their i calibrated setpoint of 13.5 inches and a limiting condition of operation of 11.0 inches. Similar anomalous behavior of the same type of switch used to actuate the emergency core cooling systems, primary containment isolation system, and other engineered safety feature systems was also revealed.

8704060409 070401 3 DR ADOCK 0500 J

~, _ .- _

The initial investigation identified two collateral primary causes for the above described nonconservative behavior of the Static-0-Ring (SOR) switches:

1. Rusted cross shaft bearing, and

2. Deformation of 0-ring seals Subsequently, after restart of La Salle Units 1 and 2, a third cause was identified:

3. Switch diaphragm failure.

The first cause, " Rusted Cross Shaft Bearing," resulted in the functional failure of one of the two SOR switches used to initiate the automatic depressurization system. Initial calibration of the switch showed that the setpoint (13.5 inches) had shifted 25 inches. A second calibration, a few hours later, also showed setpoint shift. Based on these calibration findings, the licensee concluded that rusted cross shaft bearings will cause a nonconservative shift of the calibration setpoint. In order to detect future setpoint shifts, the licensee implemented an augmented surveillance program for Unit 2 and a similar program.for Unit 1. The augmented surveillance program is discussed in References 2 and 8.

The second cause, " Deformation of 0-Ring Seals," was identified by the SOR switch manufacturer. When the 0-ring seals are subjected to operational pressures over a period of time, they change shape such that a greater frictional force is exerted in opposition to the rotation of the cross shaft. Thus, the switch trip setpoint shifts with the change of pressure from the "0" psig calibration level to the operational pressure level (1000 psig for most applications). The licensee has determined through testing by the switch manufacturer that maximum shift i

1

l l

1 of the calibrated setpoint occurs during the first 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period of operation and that negligible shifts occur beyond this time. Because the trip setpoint shift occurs at operating pressures of approximately 1000 psig over a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period, switch failure was not detected through the use of existing procedures (i.e., initial calibration, preoperational testing, periodic functional testing, and periodic calibrations) which are performed at atmospheric oressure.

The third cause, " Switch Diaphragm Failure," was identified during a December 9,1986 functional test. One SOR switch failed to maintain a differential pressure across its high and low pressure sensing ports.

An investigation found a total of four switches (inclusive of the December 9,1986 discovery) had failed because differential pressure could not be maintained. In all four cases, subsequent inspection of the switch's Kapton diaphragm revealed holes or tears which were apparent-ly large enough to allow process fluid to pass from one side to the other.

In its submittals of July 18, 21, 23, and 24, 1986, the licensee proposed a plan for testing and monitoring operation of the switches to support restart of La Salle Unit 2. The staff evaluated this plan and approved restart and short term operation of Unit 2 (Reference 9). Following this staff approval, the licensee submitted its final report on its investigation of SOR switches for Unit 2 (Reference 8) and by letters dated August 29, September 8, 10, 12, and 15, 1986, the licensee proposed a plan (similar to that approved on Unit 2) for restart of Unit 1. Based on the Unit 2 cvaluation, the staff approved restart and similar short term operation of Unit 1 (References 13 and 18).

For restart ar.d short term operation, the licensee was required to evaluate (and submit for staff review) test results acquired through its periodic and increased surveillance calibration test program and data from the 50R Inc. test program in order to justify continued use of

Static-0-Ring differential pressure switches in safety applications.

In addition, the licensee was required to submit its long term ccrrective action plan for S0R switches by January 1,1987. In accordance with these requirements, the licensee (by letters dated December 22, 1986 and January 15 and 21, 1987 and a January 22, 1987 meeting) presented its evaluation of test results to justify continued operation of Units 1 and 2 and their long term action plan for compensating for erratic dp switch operation for staff review and approval.

The long term action plan proposed by the licensee is summarized as follows:

a. Monthly calibration will be performed on the reactor water level 3 switches.

b. Quarterly calibration will be performed on the remaining SOR differential pressure switches with increased frequency if the calibration setpoint exceeds a predefined action limit.

c. Reactor water level switches will be replaced in both Units 1 and 2 with an analog trip system during the third and fourth refueling outage for Unit 1 and the second and third refuel.ing outage for Unit l 2.

l l

d. Monthly functional or calibration tests will be performed on all SOR differential pressure switches.

l l

e. If the setpoint of a switch during calibration exceeds its.

predefined rejection limit, the associated instrument channel will be imediately declared inoperable and action will be taken in accordance with the applicable La Salle Technical Specifications.

f. If the setpoint of a switch during two consecutive calibrations exceeds a predefined action limit, the switch will be replaced.

g. Any switch that is replaced due to items e and f above will be disassembled and inspected,

h. Technical Sspecification changes are being considered for the level 3, level 8, and main steam line flow applications.

The licensee's evaluation of test results acquired from its periodic and increased surveillance calibration test program and data from SOR Inc.

test program are surrrnarized below. Based on these results, the licensee has concluded that continued operation of La Salle Units 1 and 2 with SOR differential pressure switches in various safety functions is justified because:

a. Since Unit 2 startup, the switches have been performing satisfactorily. Only three out of 117 switches have exceeded their predefined action limits and only one switch has failed.

b. SOR long term test program and calibration data show that drift is not a problem and the program verifies the adequacy of the new calibration intervals.

c. Calibration data show the switches to be performing within the values j found in the studies and testing programs. Switches with drift and l repeatability characteristics that exceed pre-defined limits are I readily identified and replaced, when necessary.

d. The failure of a switch is detected quite readily with the revised surveillance intervals.

e. Successful demonstration of adequate performance through system performance tests under various pressure conditions (reactor level drop tests and emergency core cooling system minimum flow loop tests) has verified that the settings chosen for the switches are within and conservative to the existing Technical Specification allowable values.

1

~ - - . - . - . . . .-

6-

f. La Salle's maintenance personnel have been familiarized with requirements for observation of the switch performance behavior during surveillances so that they can more readily identify switches which exceed predefined limits or those which require an unusually large number of adjustments to achieve a stable trip setting.

g. Operating personnel have been familiarized with how potential switch behavior may affect system operation.

h. Rusty bearings are not a common mode problem. Corroded bearings have been found in only one out of 18 failed switches which were dis-assembled and inspected,

i. Diaphragm failures (in which dp fails to be maintained across the switch's high and low pressure ports) are readily detectable during functional or calibration testi.ng, pud

j. The diaphraguT faifure rate at la Salle is comparable to or lower than failure rates of other dp devices in use in nuclear service applications,

k. Since diaphragm failures have not been shown to be related to the frequency of operation of the system or to time in service, diaphragm failures are considered to be random in occurrence.

The staff's evaluation of the licensee's long term action plan and justification for continued operation of La Salle Units 1 and 2 while using SOR dp switches is described below in Section 2.0 of this report.

2.0 EVALUATION As reported in Section 2.0 of the safety evaluation for La Salle Unit 2 (Reference 9), the staff has found the licensee's effort to determine the

o l

root cause of tne anomalous SOR switch behavior (described in Reference

2) to be sufficiently comprehensive. The discussion provided on 0-ring defonnation is clear and reasonable; however, the licensee's assessment l

of switch failures due to rusted cross shaft bearings may not fully explain the observed switch failures. Therefore, the staff required the j licenseetoevaluate(andsubmitforstaffreview)testresultsacquired through a periodic and increased surveillance calibration testing program 4 and from the SOR Inc. long term testing of the S0R dp switch. The licensee by letters dated December 22, 1986, January 15, and March 10 and 12, 1987, and at a meeting on January 22, 1987, provided, for staff review, an evaluation of test results in support of continued use of SOR switches in safety applications. In addition, the licensee submitted its long term action plan for staff review.

One of the causes for anomalous switch behavior (discussed in Section 1.0

of this report) was " Rusted Cross Shaft Bearing." To determine this

! cause, the switch was disassembled and inspected by the manufacturer.

This inspection (Appendix H of Reference 8) found foreign material, troisture, and corrosion in the switch. It was suspected that water

! leakage around 0-ring seals could be a common problem among all SOR-j differential pressure switches. As part of the licensee's subsequent investigation into this problem, the manufacturer disassembled and inspected 12 additional switches from Units 1 and 2 which had failed to meet the licensee's screening criteria for static pressure shift, i repeatability and setpoint drift. These inspections (Appendix I of Reference 8 and letter dated October 21,1986) found similar foreign

! material in 9 of the 12 switches but found no rust or moisture indicative of water leakage around 0-ring seals. In addition, the 0-ring cross shaft seals in each of these switches successfully passed a halogen leak test at 1000 psig. Based on these investigative results, the staff concluded that the subject; failure (rusted cross shaft bearings) is not common to other SOR switches and that quarterly calibration intervals,

--. ._ , -_. ~._ - _ ._ __. _ _ _ . - _ ___ __ . _ . _ _ _ _ . . . . , _ , __. _

when supplemented by actions described below, will assure an acceptable level of switch perfomance. The licensee has proposed (reference 22) and the staff has accepted the following supplemental actions to be taken by the licensee for continued operation of La Salle Units 1 and 2:

1. Level 3 SOR switches shall be calibrated monthly.

2. As a minimum, all remaining SOR dp switches shall be calibrated on a quarterly basis except as follows:

a. If a switch (as found) trip setpoint equals or exceeds the action limit, the calibration period for that switch shall be reduced to an interval of one month.

b. Each time a switch (as found) trip setpoint is less than the action litit, the calibration interval for that switch can be increased by one month intervals, up to a maximum calibration j interval of three months.

c. If a switch is replaced, the calibration period for the newly I

installed switch shall be one month. After one month, the calibration interval for the replaced switch shall be in accordance with 2.a or 2.b above,

d. Following the implementation of this sumille ce program, the  !

starting calibration interval for each >wity' s to be based on the last calibration results and in accordance with 2.a or 2.b above.

3. If a switch (as found) trip setpoint exceeds the action limit for i two consecutive calibrations, it shall be calibrated every 14 days until it is replaced. It shall be replaced as soon as practical.

9-t

4. If a switch (as found) trip setpoint exceeds the reject limit, the associated instrument channel shall be declared inoperable until the switch is replaced, appropriate action shall be taken in accordance with the plant Technical Specifications, and the licensee shall

- report the incident in accordance with the requirements of 10 CFR 50.72 and 10 CFR 50.73.

5. Any switch that is replaced in accordance with 3 or 4 above shall be disassembled and inspected. The inspection results shall be sent to the NRC in a docketed formal transmittal.

The staff and the licensee believe that the above supplemental actions are needed because calibration test results, accumulated during short term operation, indicate (for a few switches) nonconservative erratic behavior of the switch to trip at its calibrated trip setpoint and because the contribution that foreign material (found during switch disassembly and inspection) may have made the cbserved erratic behavior has not been fully explained. The staff is concerned that this foreign material and erratic behavior may cause the switch trip setpoint to drift beyond allowable limits. The licensee has agreed to irrplement the above supple-mental actions (References 21 and 22).

i In addition, another cause for anomalous switch behavior (discussed in Section 1.0 of this report) was " Switch Diaphragm Failure." Investigations (Reference 20) into this cause found a total of four switches that had failed because of holes or tears in the switch's Kapton diaphragm. Also, more detailed subsequent inspection for holes in other 50R dp switches (removed from service for other reasons) identified three additional switch diaphragms I with holes. Microscopic analysis revealed that some holes appeared to have l been caused by sharp-edged metallic and nonmetallic particles found deposited 1

! on the exposed diaphragm surfaces. To date, neither the source of the particles '

nor the mechanism for puncturing the diaphragm has been determined.

1 L_:

8 l Based on the licensee's evaluation results summarized in Section 1.0 of this report, the staff agrees that: (1) the failure rate for SOR dp switches for any cause including diaphragm failure is no worse than other dp devices; (2) diaphragm failures are random in occurrence; and (3) diaphragm failures that are catastrophic in nature (a diaphragm with a large hole such that dp fails to be maintained across the switch's high and low pressure ports) are readily detectable during functional or cali-bration testing. For detecting small holes in the switch's diaphragm, the licensee has proposed (reference 22) and the staff has accepted a plan to supplement the calibration by the following described action (in addition to the supplemental actions described above), which should pro-vide reasonable assurance that the diaphraam failure will be detected during calibration test:

The general maintenance (GM) procedure shall be revised or a new r#. procedure shall be developed to include instructions and actions for determining the condition of SOR dp switch diaphragms while per-forming the periodic calibration trip setpoint checks. As a minimum this GM procedure shall require the technician to apply an operational differential pressure to the switch and monitor the dp for a sufficient period of time to evaluate the potential for diaphragm internal leakage.

This evaluation will be accomplished by observing the dp and the ability of the calibration unit to maintain nomal water level in the unit's re-servoirs. If abnormal switch performance is observed during this evalu-ation, the technician shall notify his or her supervisor for resolution of switch performance. Appropriate actions will be taken in accordance with the resolution determined. The licensee has also agreed to imple-ment this supplemental action (References 21 and 22).

i l

s In addition, by letter dated March 10, 1987 (Reference 22), the licensee added Tables A-2 and A-3 to its proposed long term action plan and

! justification for continued use of S0R dp switches in safety applications. The added tables define the criteria by which the calibration setpoint, action limit, or reject limit are to be determined for each SOR dp switch and the basis by which they may be changed in the ,

future. The staff believes that the criteria provides reasonable assurance that SOR dp switches will perform their designed safety function within Technical Specification operating limits.

3.0 CONCLUSION

Based on the licensee's submittals, the staff concludes that the long term plan and justification for continued use of SOR dp switches in safety applications provides an adequate basis for operation of La Salle Units 1 and 2 and is acceptable.

1 i

i

+

I I

i

n .-

3 TABLE OF REFERENCES l

1. Letter from Region III to Commonwealth Edison dated June 19, 1986 regarding confinnatory action to resolve problem prior to startup.

2. July 18, 1986, letter from M. S. Turbak to H. R. Denton transmitting the draft " Report of Investigation of SOR Differential Pressure Switches."

3. Letter to Denton from Turbak, dated July 21, 1986, regarding the validity of 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> static pressure tests.

4. Letter to Denton from Turbak, dated July 23, 1986, transmitting daily and two week test data.

5. Letter to Denton from Turbak, dated July 24, 1986, transmitting the Executive Sumary.

6. Letter to Denton from Turbak, Response to I&E Bulletin 86-02, dated July 25, 1986. *

7. Letter to Denton from Allen, dated August 1, 1986, transmitting comitment to supplemental actions.

8. Letter to Denton from Allen, dated August 8,1986, transmitting a final

" Report of Investigation of Static "0" Rina Differential Pressure Switches, La Salle County Station" dated August 1, 1986.

9. Letter to Commonwealth Edison Company from Keppler of HRC, dated August 7,1986, transmitting the safety evaluation by the Office of Nuclear Reactor Regulation supporting the restart of La Salle County Station Unit 2.

10. Letter to Denton from Allen, dated August 29, 1986, transmitting information regarding "50R" differential pressure switches for La Salle County Station Unit 1.

11. Letters to Denton from Allen, dated September 8 and 10,1986, transmitting information regarding "SOR" differential switches for La Salle Unit 1.

12. Letters to Denton from Allen, dated September 12 and 15,1986, transmitting information on plans for evaluating SOR differential pressure switches for Unit I which did not pass La Salle screening criteria.

_- ,. . ... , m- . m_- -._

n*

l

13. Letter to Commonwealth Edison Company from J. Keppler dated September 16, 1986, authorizing restart and short term operation of La Salle Unit 1.

14. Letters to Denton from Allen, dated October 8 and 24, 1966, transmitting information and commitments on the 950 psig reactor vessel level drop test.

15. Letters to Denton from Allen, dated October 14, 1986, transmitting S0R, Inc. long term test data.

16. Letter to Denton from Allen, dated October 21, 1986, transmitting results of the disassembly and inspection of SOR differential pressure switches.

17. Letter to A. Bournia from Allen, dated' November 25, 1986, transmitting SOR switch calibration i

18. Letter to Commonwealth Edison Company from J. Keppler, dated December 24, 1986, permitting operation of Unit 1 beyond January 3,1987, until the NRC has completed its evaluation of the corrective action plan.

19. Letter to Denton from B. L. Thomas, dated December 22, 1986, transmitting long term action plan for SOR switches and justification for continued operation.

20. Letter to Denton from Farrar dated January 15, 1987, transmitting investigation of the failure of SOR dp switch due to diaphragm failure.

21. February 4,1987, summary of meeting on January 22, 1987, to discuss CECO's long term resolution of SOR switches.

22. Letters to the NRC document control desk from Allen, dated March 10 and 12, 1987, transmitting revised surveillance requirements based on agreement with NRC staff.

L