ML20198S662

Special Rept:On 971009,during D21 EDG Monthly Surveillance Test,Engine Speed Was Cycling Unusually,Load Oscillations Observed on Control Room Indicators & Flame Emanated from Exhaust Stack.Caused by EDG Internal Forces.Repaired EDG

ML20198S662
Person / Time
Site:	Limerick
Issue date:	11/10/1997
From:	Boyce R PECO ENERGY CO., (FORMERLY PHILADELPHIA ELECTRIC
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 9711140087
Download: ML20198S662 (4)

Text

_

_4 n.6.et w.s.r..-

- Plant Manager Limonck oeneratmg Station

.. PECO ENERGY sco. ";,va = l g ,

Po Dox 2300 Sanar p.PA 19464 0920 21s 3271200 Ext. 2000 -

T.S. 6.9.2 November 10,1997 '

Docket No. 50-353 License No. NPF-85

- U.S. Nuclear Regulatory Commission D Attri. ;)ocument Control Desk Washington,DC 20555

SUBJECT:

Limerick Generating Station, Unit 2 Special Report for a Valid Emergency Diesel Generator Test Failure

REFERENCE:

Technical Specifications Sections 4.8.1.1.3 and 6.9.2 This Special Report is submitted pursuant to the requirements of Technical Specifications (TS)

Section 6.9.2 as required by TS Surveillance Requirement 4.8.1.1.3. TS Surveillance Requirement 4.8,1.1.3 requires reporting of all diesel generator failures, valid or non-valid, within 30 days. The report is required to include the information recommended in Regulatory Position C.3.b of Regulatog Guide (RG) 1.108, " Periodic Testing of Diesel Generator Un:ts as Onsite Electric Power Systems at Nuclear Power Plants," Revision 1, August 1977.

On 10/09/97, the D21 Emergency Diesel Generator (EDG) was being run for a monthly surveillance test, which iricluded post-maintenance testing (PMT) for generator vibration balancing acdvities. At approximately 23:00 hours, abo's one half hour into a loaded test run, the engine speed was heard cycling unusually, load oscil.ations were observed on control room indicators, and a flame was observed emanating from the exhaust stack. The local crankcase high pressure alarm annunciated. The engine was secured and an ALERT was declared in accordance with plant procedures for a fire that could affect the operability of a Class IE system.

The ALERT was terminated at 03:05 hours on 10/10/97 when the remainder of Unit 2's Class 1 E power supply subsystems were determined operable and unaffected by the D21 EDG failure.

This special report addresses only the D21 EDG issues. Notification of the Emergency 7

Declaration was addressed sey,arately in accordance with 10CFR50.72(a)(1)(i) and Limerick's i Emergency Plan.

pg 9711140087 971110 PDR ADOCK 05000353

L l'

. November 10.1997 Pcge 2 of 4 Internal inspection of the engine revealed a failure in the #4 cylinder that involved damage to the lower piston, connecting rod, connecting rod to crankshan bearing, crankpin, and cylinder liner.

Hardware failure analysis was performed by PECO Nuclear personnel, with technical assis%nce from the EDG vendor, it was determined that the physical cause of the D21 catt. strophic failure was the failure of the #4 cylinder lower piston connecting rod joint with the crankshan. Refer to the attached figure for clarification as necessary. The joint failure began when the pre-stress on the opposite control side connecting rod bolt was relieved, allowing the connecting rod nut to back off. This occurred at some point following the last five year overhaul (completed September 1994). The nut then backed ofTsufficiently to allow thejoint to begin to open. As the joint was pulled ann, the joint bearing began to degrade, impacting the lubrication of the bearing and piston. The bearing was eventually extruded from the joint, the unlubricated piston lodged in the cylinder wall, the connecting rod joint was pulled apart, and the connecting rod was bent by the crankshaft.

Through formal root cause analys.s, it was concluded that the most likely cause of the joint failure was a combination of EDG intemal forces and reversal of the lower connecting rod bearing halves during the last five-year overhaul completed in September 1994. These bearings have been installed with the halves reversed on three prior occasions in PECO Nuclear diesels, but have not led to failure. The most probable scenario is that the reversed bearing, made of aluminum, eventually yielded, allowing the joint to loosen over time Reverst ' ' earing installation would be caused by a combination ofless than adequate: human penormance, procedural guidance, and immediate supervision.

Three other possible causes, which are considered less likely but could not be eliminated, are:

failure to properly torque the connecting nut and bolt assembly, foreign material between the mating surfaces of the joint, and bearing dimensions out of specification.

Corrective measures takan involved immediate stions in response to the failure and generic implications, repair of the EDG, and longer term actions to address the causal factors, immediate actions included the following. Upon indications of a significant problem, the D21 EDG was secured. All of the remaining Unit 2 Class 1E subsystems and buses were inspected. The other tlree Unit 2 EDGs were tested (satisfactorily) within the next 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Following the initial internal inspection of D21 and preliminag analysis, the other 7 EDGs (both units) were visually inspected to verify the presence of the lower connecting rod castle nut cotter pins. No missing cc,tter pins were identified.

Repair of the EDG was accomplished via a complete five-year overhaul, and included the following. The damaged #4 cylinder assembly was removed and replaced in its entirety. The #4 lower crankpin journal, damaged by the impact forces of the broken connecting rod during the event, was machined and non-destructively examined. The lower crankshaft was further inspected by performing various runout checks to assure that it was not distorted. Lower crankshaft main bearings near the failure (4 through 6) were inspected along with other normally

. November 10,1997 Pcge 3 of 4 heavily loaded main bearings (7,10, and 13). Main bearings 4 and 5 were replaced. All connecting rod bearings were inspected on both upper and lower crankshafts and all bolting retorqued and repit.ned. All upper and lower crankshaft main bearing bolts were retorqued and repinned. Adjoining upper piston insert assemblies were pulled (#4,5,6, and 7) to assure no damage occurred from the injection ofjacket water inte the #4 cylinder. All upper piston floating bushingr were verifica foi proper clearance. The lube oil system was completely drained and all debris (associated with the failure) was removed. The vertical drive and turbochargers were disassembled and inspected. The exhaust system was inspected. A complete set of post-maintenance and performance tests were performed. No abaormalities were obse ved.

Longer term actions to address the causal factors for this event include the following.

. A series ofinspections are planned on each EDG, which include inspecting all lower crankshaft connecting rod bearings for orientation, presence of castle nut cotter pins, and sample torque checks. These inspections are 50% complete with no deficiencies identified.

s An assessment of the EDG overhaul procedures regarding human factors.

. An assessrnent of the work environment during an EDG overhaul, related to the possible impact on human performance.

. Inspection of a sample of new connecting rod bearings to determine the consistency ofdimensions.

The D21 EDG was inoperable from 23:50 on 10/09/97 until 04:23 on 10/31/97, for a total of 21 days,4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, and 33 minutes.

This event constitutes a valid test failure, and represents the 4* valid test failure in the last 20 valid tests for D21. The current surveillance test interval for D21 is once per eve;y 7 days, in accordance with Limerick's Technical Specifications. Limerick is not committed to the surveillance interval specified by Regulatory Position C.2.d. in Regulatory Guide 1.108 Revision

1. This event is the 5* valid test failure in the last 100 attempts for the unit.

If you have any questions, please do not hesitate to contact Mr. Timothy A. Moore at (610) 718-3400.

' truly yours, I

c> f .

TST cc- H. Miller, Administ.ator Region I, USNRC A. Burritt, USNRC Senior Resident Inspector, LGS l

. November 10,1997 Page 4 of 4

' '~

FUEL C "l-*

@JE_CT s p, __

4EEd -

COMBUSTIONi di C00uNG yCYUNDEn g 'p. n

• - ' . j lN -

_$. , JI j, -

s 1.hja._

fM Aliku.t~

/ usT l'  !. -m l' I!4 qM

~

PISTON OiY m f; ,- . g.- p ,e

_l,ij  ; . g;a-

.m M

e O

e sum *

- _ __ _ _ _ _ _ _- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _