ML18101A880

LER 95-015-00:on 950711,failed to Complete Documentation of EDG TS Surveillance.Caused by Lack of Procedural Clarity Re Method of Timing EDG Start & Standby Performance.Developed Special Surveillance Testing Procedures

ML18101A880
Person / Time
Site:	Salem
Issue date:	08/09/1995
From:	Mortarulo M Public Service Enterprise Group
To:
Shared Package
ML18101A879	List: ML18101A878 ML18101A880
References
LER-95-015-01, LER-95-15-1, NUDOCS 9508150171
Download: ML18101A880 (6)

Text

• -**e-NRCFORM 366 U.S. NUCLEAR REGULATORY COMMISSION APPROVED BY OMB NO. 3150-0104 (4-95) EXPIRES 04/30/98 ESTIMATED BURDEN PER RESPONSE TO COMPLY WITH THIS MANDATORY INFORMATION COLLECTION REQUEST: 50.0 HRS.

REPORTED LESSONS LEARNED ARE INCORPORATED INTO THE

.LICENSEE EVENT REPORT (LER) LICENSING PROCESS ANO FED BACK TO INDUSTRY. FORWARD COMMENTS REGARDING BURDEN ESTIMATE TO THE INFORMATION ANO RECORDS MANAGEMENT BRANCH (T~ F33), U.S. NUCLEAR (See reverse for required number of REGULATORY COMMISSION, WASHINGTON, DC 20555--0001, ANO TO THE PAPERWORK REDUCTION PROJECT (315CH>104), OFFICE OF digits/characters for each block) MANAGEMENT ANO BUDGET, WASHINGTON, DC 20503.

FACILITY NAllE (1) DOCKET NUllBER (2) PAGE (3)

Salem - Unit 1 05000272 1 OF6 TITLE (4)

Incomplete Documentation of Emergency Diesel Generator TS Surveillance EVENT DATE (5) LER NUMBER (6) REPORT DATE (7) OTHER FACILITIES INVOLVED (8)

FACILITY NAME DOCKET NUMBER MONTH DAY YEAR YEAR I SEQUENTIAL I REVISION MONTH DAY YEAR NUMBER NUMBER Salem - Unit 2 05000311 07 11 95 95 -- 015 -- *oo 08 9 95 FACILITY NAME DOCKET NUMBER 05000 OPERATING THIS REPORT IS SUBMITTED PURSUANT TO THE REQUIREMENTS OF 10 CFR §: (Check one or more) (11)

MODE (9) 5 20.2201(b) 20.2203(a)(2)(v) x 50. 73(a)(2)(i)(B) 50. 73(a)(2)(viii)

POWER 20.2203(a)(1) 20.2203(a)(3)(i) 50. 73(a)(2)(ii) 50. 73(a)(2)(x)

LEVEL (10) 000 20.2203(a)(2)(i) 20.2203(a)(3)(ii) 50. 73(a)(2)(iii) 73.71 20.2203(a)(2)(ii) 20.2203(a)(4) 50. 73(a)(2)(iv) OTHER 20.2203(a)(2)(iii) 50.36(c)(1) 50. 73(a)(2)(v) Spec~ in Abstract below or in C Form 366A 20.2203(a)(2)(iv) 50.36(c)(2) 50. 73(a)(2)(vii)

LICENSEE CONTACT FOR THIS LER (12)

NAME TELEPHONE NUMBER (Include Area Code)

Mike Mortarulo (609) 339 - 2741 COMPLETE ONE LINE FOR EACH COMPONENT FAILURE DESCRIBED IN THIS REPORT (13)

CAUSE SYSTEM COMPONENT MANUFACTURER REPORTABLE TO NPRDS tr~~~~j~~~j~f~~r CAUSE SYSTEM COMPONENT MANUFACTURER REPORTABLE TONPRDS SUPPLEMENTAL REPORT EXPECTED (14)

IYES (If yes, complete EXPECTED SUBMISSION DATE).

IXINO EXPECTED SUBMISSION DATE (15)

MONTH DAY YEAR ABSTRACT (Limit to 1400 spaces, i.e., approximately 15 single-spaced typewritten lines) (16)

On 7/9/95, during a routine monthly Surveillance Test (ST) of the lB Emergency Diesel Generator (EDG), it failed to meet the 13 second (frequency) start time acceptance criterion. The cause was due to an overly conservative Acceptance Value (AV) . The event led to an investigation which found, on 7 /11/95, that the 18 month ST contained incomplete documentation of technical specification ST requirement (4.8.1.1.2.d.4) which states, "The generator ... frequency shall be . .. 60 +/- 1.2 Hz within 13 seconds after the auto-start signal and shall be maintained within these limits during this test." The EDG starting frequency response (unloaded) is such that the frequency overshoots the upper band allowable of 61.2 Hz for a few seconds prior to settling within the acceptance band. In the past, operators timed the EDG start until the frequency reached the lower bandwidth AV(58.8 Hz) . The operators would then ensure that the frequency stayed within the bandwidth following the 1 to 2 second overshoot.

As a result, no documentation could be verified to account for the few seconds between the time where the frequency reached 58.8 Hz and when it settled within the acceptance band. Subsequent testing, using revised acceptance criterion, found that the frequency does settle within the acceptance band in less than 13 seconds. Procedures were modified to specify the correct timing methodology.

9508150171 950809

• • .,.,,.,NRC FORM 366 (4-95) PDR ADOCK 05000272 S PDR

NRC FORM 366A U.S. NUCLEAR REGULATORY COMMISSION (4-95)

LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION FACILITY NAhllE (1) DOCKET NUMBER (2) LER NUMBER (6) PAGE (3)

YEAR ISEQUENTIAL NUMBER IREVISION NUMBER Salem - Unit 1 05000272 95 -- 015 - 00 02 OF 06 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

Plant and System Identification:

Westinghouse - Pressurized Water Reactor Energy Industry Identification System (EIIS) codes and component function identifier codes appear in the text as {XX/XX}.

Identification of Occurrence Failure to correctly time the Emergency Diesel Generator (EDG) {VJ} start resulted in incomplete documentation of the EDG up-to-rated frequency response start time. This "incomplete documentation" time corresponds to a few seconds (approx. 1-2 seconds) between the time that the EDG reaches the lower bandwidth Acceptance Value (AV) (58. 8 Hz), continues to overshoot the upper bandwidth AV (61.2 Hz), and recovers back to 61.2 Hz. Therefore, full compliance with Technical Specification (TS) Surveillance Test (ST) requirement of 4.8.1.1.2.d.4 could not be objectively verified based on available documentation.

Event Date: July 11, 1995 Report Date: August 9, 1995 This report was initiated by Incident Report No. 95-1102.

Conditions Prior to Occurrence:

Both Units were in a self-imposed extended shutdown.

Mode 5 Reactor Power % Unit Load Mwe Description of Occurrence On 7/11/95, while investigating a timing methodology discrepancy in the 18 month (unloaded) EDG ST, the proper documentation of the EDG's voltage and frequency responses per the requirements of TS 4.8.1.1.2.d.4 could not be verified. This surveillance requires:

"Verifying that on an ESF actuation test signal without the loss of offsite power the D/G starts on the auto start signal and operates on standby for ~

5 minutes. The generator voltage and frequency shall be ~3950 and ~ 4580 volts and 60 +/- 1.2 Hz within 13 seconds after the auto-start signal and shall be maintained within these limits auring this test."

The literal interpretation of the above requirement is as follows: the first phase of the test corresponds to the first 13 seconds following the EDG auto start signal. The second phase starts from t=13 sec. until t=5 minutes. The frequency characteristics of the EDG when starting (unloaded) is such that the frequency overshoots the upper AV (61.2 Hz) of the acceptance band for a few seconds prior to recovering within the acceptance band.

NRC FORM 366A (4-95)

lllRC FORM 366A U.S. NUCLEAR REGULATORY COMMISSION (4-95)

LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION FACILITY NAME (1) DOCKET NUMBER (2) LER NUMBER (6) PAGE (3)

YEAR ISEQUENTIAL NUMBER IREVISION NUMBER Salem - Unit 1 05000272 95 -- 015 - 00 03 OF 06 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

Description of Occurrence (Cont'd)

The issue stemmed from the method chosen to time each phase of the EDG STs. In the past, operators timed the EDG from t=O (auto start signal) until the frequency reached the lower AV of the band (58.8 Hz), in lieu of accounting for the overshoot time. According to the historical ST documentation, and discussions with testing personnel, the few seconds where the frequency overshoots the upper AV were disregarded because it was considered a normal (unloaded) EDG starting characteristic. The TS requirement to "operate on standby for ~ 5 minutes" was interpreted by the operators to start at the time that the frequency settles to within the upper limit (61.2 Hz). Consequently, no documentation exists to verify satisfactory performance of the EDG between the time it reached 58.8 Hz and the time that the frequency settled to 61.2 Hz.

Analysis of Occurrence On 7/10/95, at approximately 0200, during a routine monthly ST, the lB EDG start time (i.e., "up to rated frequency") was noted to exceed the ST acceptance criterion and the associated TS allowable value of 13 seconds. The EDG was declared "inoperable".

After interviewing Operations personnel, it was determined that the method of testing the start time, during this test, was interpreted slightly different than that which was routinely performed in the past. Previous testing terminated the stop-watch timed start evolution when the EDG frequency first entered the acceptable frequency band (58.8 Hz). This latest test was terminated when the frequency returned into the acceptable band (i.e., below the loaded EDG 60.4 Hz acceptance criterion) following a common starting phenomenon, referred to as "frequency overshoot". Overshoot is defined as the time that the frequency is observed to be above the upper AV (61.2 Hz) of the acceptance band to the time that it re-enters the acceptance band. At Salem 1 and 2, this phenomenon is observed to reach a maximum value of approximately 64 Hz. As the governor and speed control react, the frequency returns to the setpoint value (60 +/- .2 Hz.), within a very short time span (1 - 2 seconds).

Engineering & Licensing personnel reviewed the verbiage, literal interpretation, and intent of the monthly ST, and in conjunction with input from various industry sources, determined that it was acceptable to time test the diesel start (unloaded) frequency to the point where. it first enters the acceptance band. Licensing position memorandum LR-I95359 was issued dated 7/10/95, documenting the acceptability of this method, in accordance with the design/licensing basis for the plant, the approved Technical Specification, and other regulatory criterion.

In following up to clarify the above issues, both procedurally and within the Tech Specs, it was further discovered that the 18 month Tech Spec contained unique ST criterion. Specifically, the frequency after start must be within the NRC FORM 366A (4-95)

NRC FORM 366A U.S. NUCLEAR REGULATORY COMMISSION (4-95)

LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION FACILITY NAME (1) DOCKET NUMBER (2) LER NUMBER (6) PAGE (3)

YEAR ISEQUENTIAL NUMBER IREVISION NUMBER Salem - Unit 1 05000272 95 -- 015 -- 00 04 OF 06 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

Analysis of Occurrence (Cont'd) acceptable band within 13 seconds, and also "must be maintained within this band for the duration of the test."

On 7/11/95, following extensive discussion with industry, vendor, and licensing contacts, it was concluded that due to the overshoot described above, and the fact that the timing for the 18 month ST was also terminated when the frequency first entered the acceptance range, documented evidence of compliance with the verbatim criterion of the 18 month ST could not be firmly established. All three Unit #1 and Unit #2 EDG's were declared inoperable at approximately 17:25 hours, in accordance with TS section 4.0.3. The four hour NRC notification occurred on 7/12/95 at 00:21. Both Units were in mode 5 at the time; containment integrity was re-established as required.

On 7 /12/95, a special ST Procedure (TSl/2. OP-ST. DG-0022, 23, 24 (Q)) was developed, approved, and issued, to reconstite the indeterminate test data for all six EDG' s. These surveillances were run successfully overnight, and the EDG' s were subsequently returned to operable status, by approximately 05: 00 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> on 7/13/95.

Apparent Cause of Occurrence The cause code classification "D", "Defective Procedure" (per NUREG 1022) is attributed to this event. The root cause of this problem was determined to be lack of procedural clarity relative to the method of timing the EDG start and standby performance requirements.

A final ancillary issue was uncovered during the TS & ST review which played a contributing role in this LER. In October, 1994, the Design Basis calculation for steady state EDG loading was revised to incorporate and* evaluate, among other issues, the impact caused by postulated long term frequency excursions (i.e. governor or speed control drifting/failure, causing higher frequency without overspeed).

This calculation concluded that during design basis accident conditions, the DG's could fully support their loads up to a steady state frequency of 60.4 Hz.

To demonstrate the machine's settings within this range, this value was added as a ST acceptance criterion in December, 1994. However, this value (60.4) was inappropriately imposed as an acceptance criterion for the unloaded starting tests. The steady state loaded test criterion was properly applied to the test procedure.

During starting, attaining the 60.4 Hz value following the frequency overshoot requires a longer duration than the required 61. 2 Hz.) . This incorrect, and overly conservative acceptance criterion, contributed to the initial EDG ST start time testing failure on 7/9/95 and resulted in the identification of the NRC FORM 366A (4-95)

~RC FORM 366A U.S. NUCLEAR REGULATORY COMMISSION (4-95)

LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION FACILITY NAME (1) DOCKET NUMBER (2) LER NUMBER (6) PAGE (3)

Salem - Unit 1 05000272 95 -- 015 -- 00 05 OF 06 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

Apparent Cause of Occurrence (Cont'd) incomplete documentation of the 18 month (loaded) tests.

Although not causally related to this LER (indeterminate 18 month Tech Spec Testing), this information is included to clarify all initiating and contributing events. Because the initial lB EDG test failure was caused by overly conservative acceptance criterion, no TS required Special Report is necessary.

Prior Similar Occurrence There are no previous similar events associated with the method of timing the EDG "up to rated frequency" start acceptance criterion.

Safety Significance This condition is reportable pursuant to 10CFR50. 73 (a) (2) (i) (B) due to the failure to properly satisfy the Technical Specification Surveillance Testing requirements. There is no safety significance related to this event. This condition was identified to document that compliance with the verbatim criterion of the 18 month TS surveillance 4. 8 .1.1. 2. d. 4 could not be assured without reconstituting the existing indeterminate test data. Subsequent testing showed that the performance of the six EDG' s met the TS 18 mos. (unloaded) "up-to-rated-frequency" testing requirements. Frequency overshoot during EDG starting is a widely understood and accepted phenomenon, conunonly associated with the starting response time of the EDG electro-hydraulic governor control system.

Although the frequency overshoot lasts approximately 1 - 2 seconds, there is no delay in the EDG' s ability to accept load, which occurs inunediately upon reaching rated speed and voltage. If the load was sequenced during this time-frame, there would be an improvement (i.e., dampening of the overshoot) in the frequency response due to the applied load.

Corrective Actions

l. A special ST Procedure (TSl/2.0P-ST.DG-0022,23,24 (Q)) was developed, approved, and issued, to inunediately reconstitute the indeterminate test data for all six EDG's. These surveillances were run successfully and the EDG's were subsequently returned to operable status, by approximately 05:00 hours on 7/13/95.

2. Licensing basis (Tech Spec, UFSAR. etc.) criteria related to the EDG's, along with its associated implementing procedures (e.g., STs), is being reviewed as an integral step in the Salem Station System Restart Readiness Program.

Inconsistencies found as a result of this review will be evaluated and corrected, as necessary, prior to the restart of each Salem Unit.

NRC FORM 366A (4-95)

NRC FORM 366A U.S. NUCLEAR REGULATORY COMMISSION

• (4-95)

LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION FACILITY NAME (1) DOCKET NUMBER (2) LER NUMBER (6) PAGE (3)

YEAR ISEQUENTIAL NUMBER IREVISION NUMBER Salem Unit 1 05000272 95 -- 015 -- 00 06 OF 06 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

Corrective Actions (Cont'd) 1

3. A revision request (R11433) has been initiated to revise appropriate EDG ST procedures relative to clarifying the timing methodology. These procedures will be revised in the near future and prior to their future use.

4. A review of other departmental procedures was conducted to determine if inadequate procedures exist relative to timing methods. In particular, the review considered those procedures which fulfilled TS related ST requirements. This review was not limited to those processes which use stop watches. The review considered the timing methodology and when to start and stop the timing sequence for key procedures governing the Operations, Mechanical, ISI, I&C, Electrical, and Chemistry organizations. No procedural inadequacies were identified relative to timing methods.

NRC FORM 366A (4-95)