NOC-AE-000036, Special Rept:On 971111,valid Failure of Sdg 21 Occurred. Caused by Failed Governor Speed Setting Potentiometer.Sdg 21 Governor Speed Setting Potentiometer Replaced

Special Rept:On 971111,valid Failure of Sdg 21 Occurred. Caused by Failed Governor Speed Setting Potentiometer.Sdg 21 Governor Speed Setting Potentiometer Replaced

ML20203G050
Person / Time
Site:	South Texas
Issue date:	12/11/1997
From:	Masse R HOUSTON LIGHTING & POWER CO.
To:	NRC
References
NOC-AE-000036, NOC-AE-36, NUDOCS 9712180021
Download: ML20203G050 (5)

Text

- - - - _ - - - ..

4 N r Opera pony  ;

niqczm WY

&ed T pIsfohnHrGmeratsyStdtw PC Aw1Af Mh6 mea ream //fAJ m

December 11,1997 NOC-AE-000035 File No.: G02 10CFR50 U. S. Nuclear Regulatory Commission Attention: Document Control Desk.

Washington, DC 20555-0001 South Texas Project Unit 2 Docket No. STN 50-499 Special Report Regarding a Valid Failure of Standby Diesel Generator 21 on November 11.1997 and a Non-Valid Failure on November 14.1931 Pursuant to South Texas Project Technical Specifications 4.8.1.1.3 and 6.9.2, the South Texas Project submits the attached Special Report regarding a valid failure of Standby Die.sel Generator 21 on November 11,1997 and a non-valid failure on November 14,1997.

If you should have any questions on this matter, please contact either Mr. S. M. ricad at (512) 972 7136 or me at (512) 972-7988.

.f j .f >u R. E. Masse Plant Manager, .

Unit 2 DNB/ s '

Ie% e 9712100021 971211 'll!!!!

• • aHllil111!!"ll . l e .

PDR ADOCK 05000499 8 PDR _

Attaciunent: Special Report Regarding a Valid Failure of Standby Diesel benerator 21 on November 11,1997 and a Non-Valid Failure on November 14,1997

. w w . m m s. m mm

NOC-AE-000036 File No.: G02 1

.- '+ Page 2 Ellis W. Merschoff Jon C. Wood Regional Administrator, Region IV Matthews & Branscomb U. S. Nuclear Regulatory Commission One Alamo Center 611 Ryan Plaza Drive, Suite 400 106 S. St. Mary's Street, Suite 700 Arlington, TX 76011 8064 San Antonio,TX 78205-3692 Thomas W. Alexion Institute of Nuclear Power Project Manager, Mail Code 13113 Operations - Records Center U. S. Nuclear Regulatory Commission 700 Galleria Parkway Washington, DC 20555 0001 Atlanta, GA 30339-5957 David P. Loveless Richard A. Ratliff Sr. Resident irspector Bureau of Radiation Control c/o U. S. Nuclear Regulatory Comm. Texas Depanment of Health P. O. Box 910 1100 West 49th Street Bay City, TX 77404-0910 Austin, TX 78756-3189 J. R. Newman, Esquire C. R. Crisp /R. L.Balcom Morgan, Lewis & Bockius llouston Lighting & Power Co.

1800 M Street, N.W. P. O. Box 1700 Washington, DC 20036 5869 Ilouston, TX 77251 M. T. Ilardt/W C. Gunrt Centrai Power and Light Company City Public Service ATTN: G. E. Vaughn/C. A. Johnson P. O. Box 1771 P. O. Box 289, Mail Code: N5012 San Antonio,TX 78296 Wadsworth,TX 77483 1

J. C. Lanier/A. Ramirez U. S. Nuclear Regulatory Commission .

City of Austin Attention: Document Control Desk Electric Utii'ty Department

. Washington, D.C 20555-0001 1 721 Barton Springs Road l Austin,TX 787(M i

l 1

i

Attachment NOC-AE-000036 Page 1 of 3 Special Report Regarding a Valid Failure of Standby Diesel Generator 21 onNovember 11.1997 and n Non Valid Failure on November 14.1997 Descriptio i OfIIvent On November 11,1997, Standby Diesel Generator 21 (SDO 21) was started for its monthly surveillance. At the 25% load plateau during the loadirg sequence, load began fluctuating without operator action. Due to the potential for Standby Diesel Generator overload, SDO 21 was secured to prevent damage.

Troubleshooting found that the governor speed setting potentiometer (a trimming potentiometer in the govemor control circuitry) was not functioning properly. The potentiometer was replaced.

Successful post maintenance test runs and a successful surveillance test were performed on November 14,1997. During the cooldown cycle following the surveillance test, generator voltage did not decay to zero as expected and smoke was noted coming from the generator panel.

SDG 21 was secured, inspection and troubleshooting found that the K1 (generator field) relay main coil had failed.

The K1 relay failure was traced to a cover hold down clip which was not in its normal position.

With this clip out of position, the K1 relay was physically prevented from closing to short out the generator field. When SDG 21 was placed in cooldown, the Kl relay main coil was energized through the 4X1 (master run) relay. Since the K1 relay was ut.able to close, the main coil remained energized until it failed. The K1 relay failure resulted in the failure of the associated 4XI relay contact.

The K1 relay and 4XI relay were replaced. A successful post maintenance test run and a successful surveillance run were performed on November 15,1997.

SDO 21 was inoperable for approximately one hundred and thirty-one hours.

Cause Of Event The cause of the problems witu SDO 21 load control was the failed governor speed setting potentiometer. A review of work history did not identify previous failures of this potentiometer.

The cause of the K1 relay failure has been attributed to inadvertent personnel contact with the K1 relay cover clip, which moved the clip from its normal position. The K1 relay is located in a normally closed panel. The K1 relay cover and hold-down clips face outward for accessibility.

This orientation makes it possible for the hold-down clips to be inadvertently moved if personnel unknowingly brush against the reiay. The panel was open during the testing / maintenance and personnel were working within the panel. The Kl relay functioned properly for the post-maintenance tests, but not for the surveillance test, which establishes the time period in which the

, +m,mm, w mm

Attachment NOC-AE 000036 Page 2 of 3 hold-down clip was moved from its normal position. 'Ihere was no work being performed on the K1 relay during this time period, which indicates that the holddown clip was moved from its normal position by inadvertent contact.

Annlysis Of Event 1he governor speed setting potentiometer is part of the governor speed reference circuit in both the test and emergency modes ofoperation. The load fluctuations noted during test mode operation in the November 11,1997 surveillance nm would manifest themselves as speed citanges during emergency mode operation. These changes may have been enough to force engine speed outside the required 58.8 to 61.2 liz ficquency bnnd, or could have increased the speed settireg to the point where the mechanical governor would have taken over. Either condition would render the Standby Diesel Generator inoperable, and ceuld potentially prevent the Standby Diesel Generator from fulfilling its design function. Since the ability of the Standby Diesel Generator to perform its design function is in doubt, the malfunction of the governor speed setting potentiometer has been classified as a Valid Failure of SDG 21.

Including thi. m -1 Failure, there have been two Valid Failures in the last 20 Valid Tests, two Valid Failures in the last 50 Valid Tests, c"i two Valid Failures in the last 100 Valid Tests.

Since two Valid Failures in the past 20 Valid Tests exceeds the Technical Specification trigger value, testing frequency has been increased from monthly to at least weekly.

The K1 relay failure and 4XI relay contact failure did not affect the ability of the Standby Diesel Generator to perform its emergency function. The K1 relay failed in the open position, which would allow the generator to build and maintain voltage normally. With the hold-down clip mispositioned, the K1 relay could open but not close, so the K1 relay could only fail in the conservative position given the mispositioned hold-down clip as the cause. The failed contact in the 4XI relay opened the circuit to the K1 relay main coil, isolating the control power from the fault. No breakers tripped as a result of this failure, so control power was available. These isolated and conservative failures did not affect the Standby Diesel Generator's ability to perform its design function. The surveillance test being performed passed the testing criteria and could be considered a Valid Test. Ilowever, given that the K1 relay failure caused the operator to secure the Standby Diesel Generator on the possibility doing so might prevent damage to the Standby Diesel Generator, it is conservative and prudent to classify this event as a Non-Valid Failure.

The K1 relay failure has no impact on testing frequency since it was classified as a Non-Valid Failure.

Corrective Actinas

1. The SDG 21 governor speed setting potentiometer was replaced.

2. SDG 22 and SDG 23 were run with load applied to ensure no common mode failure.

. em,mam. ~ m u.m

Attachment NOC AE.000036 Page 3 of 3

3. The SDO 21 K1 relay was replaced.

4. The SDO 214XI and 4X2 relays were replaced (the 4X2 relay is redundant to the 4XI relay and has a contact in series with the failed contact in the 4XI relay in the exciter shutdown circuit).

5. All Standby Diesel Generators were walked down to verify that the K1 relay cover retaining clips were properly positioned.

6. SDO 21 passed appropriate pot.t maintenance testing.

7. Actions to preveat pioblem recurrence are being developed and are being tracked in the corrective action program.

Additional Information The Standby Diesel Generator is a Cooper Bessemer model KSV 20 engine rated at 5500 KW.

The govemor is a Woodward model 2301.

The govemor speed setting potentiometer is a 500 ohm Spectrol model 534.

The K1 relay is a Telemechanique model A143F12D-X3.

The 4XI relay is an Allen Bradley model 700-DC R440-Zl.

. we , _ mow. = in mm

._. .