ML20012F059

Special Rept:On 891120-1230,valid Test Failures of Div I Diesel Generator.Current Valid Failure Count for Diesel Generator Is Three.Diesel Generator Continues to Be Tested on Weekly Basis Per Tech Spec Table 4.8.1.1.2.1

ML20012F059
Person / Time
Site:	Clinton
Issue date:	03/29/1990
From:	Spangenburg F ILLINOIS POWER CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
U-601632, NUDOCS 9004090424
Download: ML20012F059 (4)

Text

-

(

U.601632 1A7 90(03 29).LP 8E.110e til/NotR POWER COMPANY y cteos Pata 57AtsoN. P.o. to 98. et mos ittisois ont March 29, 1990 f

L Docket No, 50 461 U.S. Nuclear Regulatory Commission

! Document Control Desk Washington, D.C. 20555 i

Subject:

Special Report: Supplemental Report on Test Failures of the Division I Dianel Cenerator at Clinton Power Station (CPS)

Dwar Str:

L Sy SPECIAL REPORTS dated December 20, 1989 and January 11, 1990, Illinois Power (IP) reported valid test failures of the Division I diesel generator (Dw'1.A) as a result of slow starts on November 20, 1989, December 11, 1989 and December 27, 1989. Additionally,.by SPECIAL REPORT dated January 29, 1990, IP reported a valid test failure of DG1A as a result of a failure to start on December 30, 1989. The January 11, 1990 and January 29, 1990 SPECIAL REPORTS identified that an Action Plan i

had been developed to identify the root cause and resolve those factors which may have contributed to the slow starts of DG1A. Additionally, the January 29, 1990 SPECIAL REPORT identified that a failure analysis would be performed on the K19 control relay failure which had caused the-DCIA failure to start on December 30, 1989. This supplemental SPECIAL REPORT provides the available results of this Action Plan and K19 control relay failure analvsis.

DCIA Action Plan Detailed assessments of DG1A commenced following the slow start on November 20, 1989. The initial investigation relied upon instruisentation which could be installed with DCIA in service. This initial investigation was previously discussed in the January 29, 1990 SPECIAL REPORT. As described in that SPECIAL REPORT, corrective action for the December 27, 1989 slow start led to the replacement of the Woodward Covernor actuator on the 12. cylinder engine (DG1A utilites

B tanden 12. cylinder and 16. cylinder diesel engines). The decision to replace the 12. cylinder engine governor was based upon the observed characteristics of the response of DG1A during starting sequences.

Instrumentation installed on the DGIA control system circuitry per the Action Plan had not revealed any specific cause far the slow starts outside of the governor itself. Since plant operating conditions

. precluded installing full instrumentation per the Action Plan prior to the governor replacement, the functional characteristics of the removed (

governor could not be fully assessed prior to its removal.

[$[

S

$$ 1 Id M

'/o

l i

As identified in the January 11, 1990 SPECIAL REPOkT, prior to i replacement of the 12. cylinder engine governor on December 27, 1990, the governors of DGIA were different models. Both sodels were approved for use on the installed diesels and the governor manufacturer had stated ,

that the tvo models were interchangeable. Replacement of the governor '

on the 12. cylinder engine (EGB.13P model 9903 266) resulted in restoring t

consistency in the models for the two engines. The currently installed i governors are both EGB.13P model 9903 265.  ;

1 Purther specific testing of DGIA under the Action Plau was i accomplished on February 23, 1990. Test instrumentation was installed (

to allow monitoring of 24 additional points (10 on the air start system,  ;

4 on the agine governor, 6 on the fuel system, and 4 on the  !

start / con;tol circuitry). In addition to using this test  !

instrumentation, existing pressure gauges, the installed General l Electric Transient Analysis Recorder (GETARs), and observations by l support personnel were utilized, During the four start attempts on i February 23, 1990, all external characteristier. of the DG1A response j were observed to be normal. The governor actuators responded to the  ;

start signals properly, as did the air start actors and fuel pumps.  !

DGIA start times ranged between 9.4 and 8.9 seconds, with the first  ;

start being the slowest. Start times of the successive three starts  !

vere all within the 9.1 to B.9 second range that has been characteristic i of DG)A starts subsequent to the 12.cylindor engine governor i replacement.  ;

The measured pressure and flow characteristics of the air start system indicate that the existing air start system design is adequate.  !

As discussed in the January 29, 1990 SPECIAL REPORT, modifications to i the air start piping to improve air start motor performance had '

previously been recommended for evaluation. A recom w.dation to proceed I with this improvement is being reassessed based upon the data obtained  ;

under the Action Plan.  :

Similarly, fuel system characteristics were also assessed as being  !

adequate. Two modifications, one to provide continuous operation of the i fuel priming pumps during diesel generator operation (Field Alteration  :

DGP023), and one to raise the setpoint of the fuel oil day tank level  ;

switches to reduce the likelihood of air entrainment during transfer 1 pump testing (Field Alteration D0F002) are being pursued as reliability l enhancements. No immediate corrective actions were determined to be i required for maintaining fuel system operability. ,

i A central feature of the investigation under the Action Plan has i been assessment of governor operation, rollowing the 12.c)1inder  !

governor replacement on December 27, 1989, the relative responses of the r 12 cylinder and 16. cylinder engine governors of DG1A were assessed.  !

Using stop watches, the duration of the

• full fuel" output stroke of each governor was simultaneously measured. The times for both governors l were between 7 and 8 seconds for the duration of the maximum fuel demand >

signal. The times for DG1A to reach rated speed during these tests were between 8.9 and 9.1 seconds.

(2)  :

L

0

(  ; ,

v .

i During the add!tional testing conducted on Dbrtary 23, 1990, a >

more detailed asm.essment. of the response of the g6vernors was conducted.  :

{k. '

These tests showed that the magnitude and duration of the input.and output pressure signals of each governor booster servomotor were very  ;

similar, and that the governor internal actuating, fluid pressure built  !

up properly and was maintained throughout the diesel generator '

acceleration to rated speed. Conettrrently, observers stationed at the ,

governors evaluated fuel rack motion. All observation results were favorable, i

As previously noted, the governor on the 12 cylinder engine .

(Woodward ECB.13P, model 9903 266 S/N 2113014) was replaced before an i An. situ assessment of the starting response of this device could be ,

conducted. Since it is essential to understand what impact operation of this governor had on the DG1A start time, the governor was returned to j Woodward for functional testing. The following discussion details the ,

results of this functional testing. ,

Testing at Woodward (wirnessed by an IP engineer) was planned to recreate the conditions surrounding an engine start from standby conditions following a dormant period. However. limitations in the t testing facilities did :not allow the questionable governor (model 9903 266) to ho tested in tandem with an ECB.13P model 9903 265 governor.

Therefore, the response characteristics of operating the model 9903 266 governor alone were evaluated. This was done using a test stand with an acceleration rate s/9t to match that of the CPS DCIA.  :

Ten test starts were conducted. The first test start was expected to be the most representative of the CPS conditions since the slow i starts had only occurred following some dormant tito. During this first test start, the governor did display a lack of response. However, this i test was interrupted by a Woodward Covernor engineer who felt that the governor was not responding because of internal air pockets. The ,

subsequent nine test results were very consistent. During these tests, L the governor setpoints were annessed as being correct. In all tests, '

the initial response (zero RPr( to 200 300 RPM) of the model 9903 266 ,

governor was observably slower than that known to occur during starts j using a 9903 265 model governor. The impact of the slower response of r the 9903 266 model actuator is not yet fully understood. However, i further testing of the removed governor at Woodward is planned.

M19 fontrol Relav Failure Analysis The January 29, 1990 SPECIAL REPORT identified that the relay [

manufacturer would perform a failure analysis of the K19 control relay.

The results of this analysis are presented below.  ;

As identified in the January 29, 1990 SPECIAL REPORT, the DCIA '

failure to start was caused by a failure of the 2D 2E contact pair of the K19 relay. Failure analysis of the K19 relay (Potter and Brumfield Company type MDR, model number 5095) revealed high contact resistance ,

across the 2D 2E contacts. The small load that this relay contact pair -

carries (22 mA) has been evaluated as being insufficient to ensure ,

, (3)  ;

1

v~~~~ -

.o'o y [i , ' "

4 o

' l / g fc .

l' a o .

F electrical cleaning of the contactor surfaces following extended usage.

This led to circuit interruption, resulting in the DGIA failure to F start.

! This relay failure is not considered to be related to the observed slow st.ert, problem. No further corrective actions are being considered for immediate septementation, however, relay operation is being monitored during routine diesel generator tests. The service history of the K19 relay at CPS provides assurance that the recently installed relay will provide several years of reliable service under normal operating conditions. An appropriate frequency for periodic replacement of this relay is currently being evaluated.

G2nclM19na rollowing replacement of the governor on the 12. cylinder engine on December 27, 1989, no subsequent start attempts have resulted in start times in excess of the 12 second limit of the Technical Specifications, except for the one start attempt that failed due to the K19 control relay malfunction. (As identified above, the December 30, 1989 failure is considered to be independent of the slow start problem.) The objective data strongly suggests that replacement of the 12. cylinder engine governor with a model 9903 265 under the Action Plan has resolved the DG1A slow start problem. Although, as discussed above, the root cause has not been conclusively identified, implementation of the Action Plan has confirmed the effectiveness of corrective actions and effected a thorough investigation.

Further testing of the removed governor at Woodward during April, 1990 is planned. Scheduling of specific dates for this additional testing is dependant upon development of a test method which will more fully assess the performance of this governor under simulated operational conditions. Until this additional testing can be completed, the remaining spare model 9903 266 governor at CPS has been placed on hold to prevent it from being used as a replacement. The results of this additional testing vill be provided in an additional supplemental report by May 31, 1990.

Additionally, the Action plan testing and analysis results have been reviewed for applicability to the Division 2 (DG1B) and 3 (DG10) diesel generator units. The DGIA slow start problem has not occurred on either of these two units, DG1B has always had the same model governors (EGB.13P model 9903 265) on both engines and has never displayed a tendency to start slowly. DG1C is a single. engine unit and utilizes a L

completely different type of governor (Woodward model UG.8). DGIC has never experienced a slow start. The reliability enhancements being l considered for the DGIA air start system and fuel system are likewise l being considered for DG18.

The K19 relay problem does not impact DG10, since DG1C does not utilize this type of relay in any application. DG1B does use a Potter 6 l Brumfield MDR 5095 relay in an application identical to that used in IXil A. This DG1B relay is scheduled to be replaced under MVR D08247 l during the current plant outage (PO 3).

l l

(4)

4 %

Additional Information As of March 26, 1990, a total of 13 successful, successive, valid tests of DG1A have been completed since the last valid failure on December 30, 1989. The current valid failure count for DG1A is 3 in the last 20 valid tests performed and 8 in the last 100 valid testa performed (only 81 valid tests 1. ave been performed for DG1A since receipt of the operating license). Therefore, DG1A continues to be tested on a weekly basis in accordance with Technical Specification Table 4.8.1.1.2 1.

Sincerely yours, 9 %qb T.A.Spabgenbbrg,III Manager Licensing an Safety DAS/can ec: NRC Clinton Licensing project Manager NRC Resident Office Illinois Department of Nuclear Safety i

(5)

_ _ - _ _ - - _ _ . _ _ _ _. _ ._ __ - __ _ - _ _ _ _ - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ -