ML20248B944

Special Rept PIR 2-M89-0227:on 890829,diesel Generator 2B Was Run at Greater than 4,000 Kw for Approx 80 Minutes. Caused by Variable Start Trends & Metal Diaphragm Right Bank Wear.Action Plan Developed to Help Prevent Recurrence

ML20248B944
Person / Time
Site:	Mcguire
Issue date:	09/29/1989
From:	Tucker H DUKE POWER CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
PIR-2-M89-0227, PIR-2-M89-227, NUDOCS 8910030334
Download: ML20248B944 (6)

Text

n _. _

., - ~,

. DuxE POWER GOMPAhT P.O. DOX 33189 -

CHARLOTTE, N.C. 28242 HALH. TUCKER reterniowe vuos russinuar (7o4) 073-4534 wool RAS FSODMMON September 29, 1989-U.S. Nuclear Regulatory Commission Document Control Desk

.' Washington, DC- 20555

Subject:

McGuire Nuclear Station, Unit 2 Docket No. 50-370 Diesel Generator Special Report Gentlemen:

Pursuant to Technical Specification (T.S.) 6.9.2:as specified by T.S. 4.8.1.1.3,

~

find' attached a special report concerning Diesel Generator 2B.

If you have any questions,'please contact S.E. LeRoy at (704) 373-6233.

Very truly yours, c.4 h4 -

Hal B. Tucker SEL/461 Attachment xc: Mr. S.D. Ebneter, Administrator Mr. P.K. Van Doorn U.S. Nuclear Regulatory Commission NRC Senior Resident Inspector Region II McGuire Nuclear Station 3 101 Marietta St., NW, Suite 2900 i Atlanta, Georgia 30323 -

Mr. Darl Hood, Project Manager Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20555 INPO Records Center i Suite 1500 I 1100 Circle 75 Parkway Atlanta, Georgia 30339 h

i (-

8910030334 890929 PDR I

S ADOCK 05000370 PDC

Document Control Desk

, Docket No. 50-370 September 29, 1989 Attachment Duke Power Company McGuire Nuclear Station Diesel Generator 2B Special Report ~

(PIR 2>M89-0227)

On August 29, 1989 at 1522, Diesel Generator (DG) 2B was started (start attempt No. 730) per the normal operating procedure (OP/2/A/6350/02 Diesel Generator) and run at greater than 4000 KW for approximately eighty (80) minutes as part of PT/2/A/4200/09A, ESF Actuation Periodic Test. This start attempt is considered a valid failure because the time to achieve 4160 volts was greater than the Technical Specification limit of 11 seconds. The slow start, 11.24 seconds, was a concern but did not affect the ESF test results so a decision was made to complete the ESF test and develop a plan of action for trouble shooting the slow start. At 1719 and 1842, the DG was started (start attempt Nos. 731 and 732 respectively) for the ESF test. Trouble shooting had already begun for the start attempt no.

730 failure; therefore, start attempt No. 731 is considered an invalid test and start attemp No. 732 is considered an invalid failure. A meeting was conducted with representatives from Operations, Instrument and Electrical (IAE), and Mechanical Maintenance (MNT) to discuss potential causes for the intermittent slow starts, and to develop a plan of action. Start trends indicated much more variation following the EOC5 refueling outage. Technical support staff at Cooper Bessemer were consulted and based on their recommendation, an inspection was performed on both starting air distributors. A metal diaphragm was replaced in cach distributor. The right bank displayed some wear and the left bank appeared to be acceptable but was replaced as a precaution. Other actions taken were as follows:

. The starting air distributor oil misters were checked for leaks (none found);

. The starting air manifold was checked for leaks (none found);

. The fuel racks were greased and exercised;

. The fuel pumps were lubricated;

. The fuel oil, starting air purifier, and starting air distributor filters were changed;

. Each starting air valve was pressurized and exercised;

. Laboratory analysis was performed on the fuel (no problems found);

. A gauge was installed on the fuel oil header to verify steady pressure was maintained by the booster pump (no problems found);

. The solenoid valves were checked for proper operation (no problems found); and, j

. The starting air tanks were checked for water (none found).

On August 30, 1989 at 2039, a start attempt (No. 733) was made. Start times were )

satisfactory but the DG oscillated excessively at idle. The mechanical governor  !

was checked and the unit was stable. This start attempt is classified as an invalid test because the DG was not run for one hour at greater than 2000 KW.

At 2048, start attempt No. 734 was made. The start times were satisfactory but again, idle speed oscillation of approximately +/- 0.2 HZ occurred.

l

.i

_ _ __ _ . j

j Document Control Desk

, Docket No. 50-370 Diesel Generator 2B Special Report +

September 29, 1989 Attachment ,

IAE adjusted the EGA control box stability pot one quarter torn clockwise and the-

~

unit stabilized. This start attempt was also classified an invalid test for the same reason as start attempt No. 733.

At 2208, start attempt No. 735 was made. DG stability at idle was good and the start time was 10.8 seconds. At 2214 and 2219, several starts attempts (Nos. 736 and 737, respectively) were made to determine if another slow start would occur.

The start times were 10.3.and 10.5 seconds, respectively.

On August 31, 1989 at 0021, start attempt No.738 was made to perform a Benchmark test required because the electric governor centrols were adjusted during start attempt No 734. This test provides a load change transient greater than any experienced during an ESF test. The Benchmark test results were satisfactory indicating the governor would respond properly during an accident condition. This start attempt is classified as an invalid failure for the same reason as start attempt No. 733 above.

At 0200, another trouble shooting start attempt (No. 739) was made. The DG required 11.028 seconds to achieve 4160 volts. The calibration of the speed switches was checked and found to be satisfactory at the 50% setpoint which would not affect starting operation because it picks up auxiliary relays that drop out starting air and flash the generator field. However, the 95% and 97% setpoints on switch 51A1 were found to be high. These setpoints were lowered to the correct setpoint. The 97% setpoint being high caused the sequencer to kick out of the accelerated sequence mode during the ESF test but has no bearing on the slow start time problem. Neither the 95% nor the 97% setpoints being high are thought to have caused the slow starts. The engine speed sensor was then checked. The sensor is a magnetic pickup that receives a pulse each time a camshaft gear tooth f passes. These pulses provide a signal to the local speed indication and the speed 1 switches. The gap between the sensor and the gear was thought to have been set at l 0.015 inches; however, a 0.015 inches gap gauge would not slide into the gap. The sensor was removed and it was discovered that it had at some point been rubbed by the camshaft gear and damaged. The sensor was replaced and the gap was set at 0.020 inches per procedure (the factory concurred that this setpoint was acceptable). This problem is though* to have caused the slow starts and intermittent start time data noted since the refueling outage. Several starts were attempted to check start times. At 2102, the DG was started (start attempt No. 740) and declared operable following this valid success.

On September 5, 1989, instrumentation was connected to the DG and start attempt Nos. 741 through 743 were made to provide data on speed switch operation, voltage build-up, fuel pressure, frequency, and signal from the speed sensor to the speed switches. The data indicated all of these inputs were normal and stable. The ]

-start times were very consistent. Future start times will be carefully monitored to note any increase in variation among the data. These start attempts were classified as invalid tests for the same reason as start attempt No. 733 above.

'I I

Document Cont rol Desk Docket No. 50-370 Diesel Generator 2B Special Report September 29, 1989 Attachment On September 7, 1989 at 0918 and 1024, start attempt Nos. 744 and 745 were made for normal surveillance testing. DG start times were both 10.6 seconds to achieve 4160 volts. These starts were classified as valid successes.

Based on the consistent start time data obtained since the replacement of the enRi ne speed sensor, the device causing the problem appears to be isolated, (see attached graphs).

This problem was difficult to trorble shoot because the indications were different from those in past situations where the speed sensor was found damaged. This was the fifth speed sensor known to be damaged in a similar fashion, though in other cases, the damage was more severe. The sensor on DG 2A, as reported in letter of August 14, 1989, was also discovered to be damaged. Apparently its r. a,e also occurr- during the refueling outage. The root cause of the sensor damist .s still not well understood. It has to either be a movement of the camshaft geer toward the sensor or vice-versa. The mounting of the sensor appears to be very secure as evidenced by the difficulty encountered removing the sensor. A check of j the run out on the camsha~t gear indicated that it was only 0.002 inches. I Therefore, a clearance of 0.015 inches should be sufficient to account for any growth during operation. An action plan has been developed to give a higher degree of confidence that the problem will not return.

As indicated in my letter of August 14, 1989 concerning DG 2A, the mechanical periodic tests that guide the sequence of work on the DG during the outage will be changed to include a step to have IAE check the gap and the mounting of the sensor just before the engine is run for the first time following the outage. An order already placed for four new sensors will be expedited. When these new sensors are received, work requests will be issued to have IAE to check the sensors on all  !

four DGs. The DG 2B sensor will be checked first since it has just been replaced '

and the gap was widened to 0.020 inches,. Any indication of wear on the tip of the sensor will cause a reanalyzes of the situation. The sensor will be reset to have

, the gap in the range of 0.020 inches to 0.030 inches. Proper operation of the speed switches will then be monitored via start times and timing of DG events.

'The output signal of the sensor will be monitored for proper amplitude and ,

stability. Once the DG 2B data is obtained, the other DGs will be checked at the  !

earliest opportunity.

There have been two (2) valid failures in the past twenty (20) valid tests, and six (6) valid failures in the last one hundred (100) valid tests. The Technical Specification surveillance frequency is weekly, i

i l

l l

y . _

0 5

7 e

- I

, i

' 8

,J 4

1 1 4

'9 _

1 0 ._

4Jd a 3 m

^

i 7 _

2::

m _

r::

L:: e

I C::

E P

h S:e  :

t 0

, 1 lf : 7 t:

D h ::

T: #1 O E  :

# R GES -

P E

B M

/ %  :

U D5 9  :

N T

R 8OT  :

i 0

9 6

A T

S 4

2E M  :

b A:9 I

T .

1

,J I

I Y  : - i

0 .

,7

6 _

=

2

0

5 6

8 5 6 1 5 '0 5 9 5. -

• 1 1

0 1

9 8 7 1 -

1 r

aOZoOJW I lfilll \! l

0 5

' 7 -

,l; e

8 m .

f.  ;

3n 1 o gip O l

i t la

. 0 P:

i 3 7 -.

t

i

+

1

1. p' g - 6 0

1 WW"swg 7

C o c

GV A s 7 m

s 't e a s

/0 6 kAf I '

i s

u u

D14 =

= ' w r

O '/" e BT E 0 9 w s i

2MT I 6

M\

M)M t

w e

i l M m r

SWG f e e i 0

7 6

i E( C ..

E 9 _

. $ i 4e T -

w

• _

a 0 5

- - - - - - 6

.3 5 2 5 1 5 0 5. '

1 1 1 '

1 2 1 0 9 1 1 1 t

Oomi f

' l l,IlIlfI