3F0287-17, Ro:On 870121,during Performance of Surveillance Procedure SP-110,CRD a Ac Breaker Failed to Trip.Caused by Malfunction of Undervoltage Trip Coil Device.Breaker Replaced & Successfully Tested

Ro:On 870121,during Performance of Surveillance Procedure SP-110,CRD a Ac Breaker Failed to Trip.Caused by Malfunction of Undervoltage Trip Coil Device.Breaker Replaced & Successfully Tested

ML20211P540
Person / Time
Site:	Crystal River
Issue date:	02/23/1987
From:	Eric Simpson FLORIDA POWER CORP.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
3F0287-17, 3F287-17, GL-83-28, NUDOCS 8703020383
Download: ML20211P540 (8)

Text

GL 83-28 je oj e

ee.

Florida Power C OQ PO R ATIO N February 23, 1987 3F0287-17 Document Control Desk U.S. Nuclear Regulatory Commission Washington, DC 20555

Subject:

Crystal River Unit 3 Docket No. 50-302 Operating License No. DPR-72 Generic Letter 83-28 Supplemental Information

Dear Sir:

The attached report describes an incident which occurred at Crystal River Unit 3 on January 21, 1987. This information is provided voluntarily by Florida Power Corporation.

Sincerely, E. C. Simpscn Director, Nuclear Operations Engineering and Licensing AEF/feb Attachment xc: Dr. J. Nelson Grace Regional Administrator, Region II Mr. T. F. Stetka Senior Resident Inspector

~

0703020383 070223 PDH ADOCK 05000302 f S PDH ,

l 1 GENERAL OFFICE: 3201 Thirty-fourth Street South

• P.O. Box 14042

• St. Petersburg, Florida 33733 * (813) 866 5151 A Florida Progress Company

~

Frbruary 23, 1987

. 3F0287-17 Pagn 1 cf 4 CRD "A" BREAKER FAILURE EVENT DESCRIPTION On January 21, 1987, at 2200 during performance of Surveillance Procedure SP-110, RPS Functional Testing, the "A" Control Rod Drive (CRD) A.C. breaker failed to trip as required. The breaker (CRDM #10) had tripped when given a trip signal moments earlier, yet failed to trip when given only an undervoltage (UV) trip signal. The failure was attributed to the failure of the undervoltage trip coil device (see Figure 1). The UV trip device was found with the trip paddle jammed on top of the UV trip device armature. The mounting bracket for the entire UV trip assembly was loosened in order to free the trip paddle. The "A" channel was the last channel tested in SP-13 3. All remaining in-service breakers had already satisfactorily responded to the surveillance test.

The UV trip device was removed from the breaker. The entire .CRDM #10 was removed and replaced with CRDM AC-1A. The UV device for this breaker was taken from breaker DC-1G; inspected; tested per PM-118, AC and DC Breakers -

Control Rod Drive System; and installed into CRDM AC-1A. Surveillance procedure SP-110 was then completed satisfactorily for the "A" RPS channel .

The CRDM #10 breaker and its failed UV device were readied for further examination and testing by site personnel and General Electric (G.E.)

representatives during the week of January 26, 1987. In addi tion, all remaining in-service breakers (which had already successfully passed SP-110) were visually examined for a jammed UV trip paddle. All of the remaining in-service breakers were satisfactory.

CAUSE The cause of the breaker failing to trip was apparently due to the malfunction of the UV trip device. Several possible causes of the failure of the breaker to trip on a UV trip signal have been postulated. Each of the possible causes identified are listed with additional clarification as necessary.

1. Loosely mounted UV trip device:

This was proven to be a viable condition for allowing the trip paddle to become jammed on top of the trip armature. Subsequent testing of a spare breaker UV assembly resulted in the trip paddle being able to be jammed on top of the trip armature. In this test case it was due to the trip assembly being loosely mounted (i.e., the 2 mounting nuts, item #16 on Figure 2, were loosened). In the case of the failed breaker CRDM #10, this cause was discounted based on discussions with the electricians who dislodged the trip paddle from the armature. They said the mounting nuts were tight, and in fact had to be loosened in order to free the trip paddle.

2. Excessive UV device rivet to armature gap:

This problcm had been documented in IAE Notice 85-58 as a result of problems encountered at Rancho Seco. At the time of issuance of this Notice, FPC reviewed appropriate procedures and found thcm acceptable to detect this type of problem. Procedure PM-118 requires measuring the

- February 23, 1987 3F0287-17 Page 2 of 4 roller rivet to armature gap. The procedural ' acceptance criteria is between 0.002 and 0.009 inch (G.E. criteria is .001 .010 inch). The last time the subject UV device was checked, during PM-118 performance in November 1986, the gap was within tolerance. Af ter its removal from CRDM

1. 10 and being subjected to several inspections, tests, manual movement of the armature, etc., the roller rivet to armature gap was found to be .016 inch. It is not known if the change in the roller rivet to armature gap occurred before or after the CRDM #10 failure.

3. Manufacturing / construction defect:

It_ is possible a construction-type defect caused this event. Since the vendor maintains all of the drawings and specifications for the individual components which comprise the UV trip device, the failed UV device and entire CRDM #10 breaker are being sent to G.E for further testing and measurements. Until the G.E. report is completed, it will not be known if a construction or manufacturing error was involved in the component failure.

4. CRDM #10 breaker frame slightly bent:

On January 26, 1987, while performing his preliminary inspection of CRDM

1. 10, one of the G.E. field engineers observed the breaker back-frame was bent slightly inward. This was the first time this condition was observed. Previous test documentation was reviewed to determine if the bent frame had ever been observed, and it had not been observed. There is the possibility the bent frame affected the operation of the UV trip device. However, the failed UV trip device was installed in two spare breakers wi th straight frames and the jammed trip paddle condition wasrecreated in both cases.

SAFETY CONSIDERATIONS The significance of the UV trip device failure to trip the breaker was minimized by the fact the shunt trip device worked to trip the breaker, as was shown in the test performed on January 21, 1987. In other words, if this breaker had been called upon to trip while in-service, it would have tripped.

Therefore, this event did not have direct Technical Specifications implications since the operability of the breaker was not impaired.

However, several other utilities have experienced problems with G.E. reactor trip breakers; and specifically, UV trip assemblies on reactor trip breakers.

Due to possible generic concerns, this event and the subsequent failure analysis have received high priority from FPC, the on-site NRC inspectors, and General Electric. Further information is expected to come from G.E. once they have completed their analysis.

CORRECTIVE ACTIONS Upon discovery of the malfunction of the CRDM #10 trip assembly, the remaining in-service reactor trip breakers were visually inspected to ensure this condition did not exist elsewhere. All remaining in-service breakers were

]

F:bruary 23, 1987 3F0287-17 Page 3 of 4 satisfactory. The entire breaker CRDM #10 was replaced by CRDM AC-1A, and this breaker was successfully tested per SP-110. The trip device from CRDM

1. 10 was removed in order to perform measurements. These measurements were mainly distance measurements in an attempt to determine if there was an excessive amount of side-to-side motion of the trip device armature. Al so ,

the location on the armature pad which appeared to have the most contact marks from the contact with the trip paddle was recorded. This point was at the top of the pad and slightly off-center (refer to Figure 3).

The UV trip device was mounted back in the CRDM #10 breaker. The jamming of the trip paddle on top of the armature was recreated with relative ease. With the UV armature held in the energized position, the trip shaf t was moved such that the trip paddle was engaged. Moving the annature to the de-energized position resulted in the as-found condition when the breaker failed to trip.

The UV device was removed and the roller rivet to armature gap set at 0.005 inch. Again, the event was recreated, although a slight prying motion was required on the trip shaf t. A new UV device was installed in CRDM #10 and again the condition was recreated. A significantly greater amount of force was required, most likely more force than would normally be seen during breaker testing and operation, to jam the paddle.

The failed UV device, with the roller rivet to armature gap set at 0.005 inch, was then tested in 2 other breakers to see if the armature and trip paddle could become jammed as in CRDM #10. In both cases, the trip paddle could be jammed up on top of the armature by rotating the trip shaf t. In addition, the UV devices normally found in these 2 breakers were also tested in the same manner, and neither could be made to recreate the jammed condition. These results tend to lessen the significance of the fact the breaker frame was bent, although it is possible that when the breaker frame was bent. the UV device was also jarred. This also minimizes the significance of the roller rivet to armature gap being out of tolerance during the failure, since it was demonstrated the UV device could fail with this gap in tolerance.

A summary of the major correctiv, actions and conclusions is as follows:

1. The failed CRDM #10 was removed and replaced.

2. The UV trip device for the replacement breaker was successfully tested per PM-118 prior to placing the breaker in service.

3. The replacement breaker was satisfactorily tested per SP-110.

4. The physical condition and measurements of the UV device from the failed breaker were recorded to aid G.E. in their analysis.

5. Contact marks on the UV device armature pad indicate the pad and the trip paddle are slightly misaligned.

6. The CRDM #10 breaker frame was bent, although testing of the failed UV device in spare breakers with straight frames produced the jammed trip paddle condition found in the CRDM #10 breaker.

A

, .. February 23, 1987 3F0287-17 Page 4 of 4

.7. The - roller rivet to armature gap was within specifications the last time it was tested per PM-118 in November 1986, yet it' was out of tolerance when measured by the G.E. field engineer on January 26, 1987.

Other than the above actions, FPC is awaiting the- final report from G.E.

before implementing any further actions.

MANUFACTURER DATA CRDM #10 breaker:

Type: G.E. Model AK-2A-25-1 S/N: 256A8930-38 I

t I'

i l

/

FIGURE 1

'Y 3

~ Y) 4, T ~'

<> + D  %

0 Y

0 '

g ,

i OS

[-

5

\

UV TRIP DEVICE gEACTOR TRlY

)

1 FIGURE 2 i

1 f

1 to O

g. ,

1. Moucti g Se r e, V, *

2. Trzme 58 3. Arma:ae r 4. Sprtag

5. W e g g 2 :7 6. Adfusting Sef,,

X '

hy 7 LocW Nut

/ ij ;6 8. Bushing 3 ,

.. " , - 9. Clamp T '

q' ~.; ' 10. Masset 4  ; V e - iS II* Screws o M i2. Coti

N 72rfE5 13. Rnet AM3

• - 14. Adfu. sting Screw

15. Lock.g Wire 6 -- - Mgg

r g ,

y '3

16. Mounting Nut

17. Mechanum Frame l g' 18. Trip Paddle Cla=ps 7

$l -

i 12 19. Trip Paddle qu 20. Adfusting Screw -

g- l l l d

= g l "

4 l l l= ~

w I > l 4 .\

+

+ba2.] -

o N

-9 UV TRIP DEVICE

u i

FIGURE 3 CONTACT AREA

, m t '

i .' , ,

'\ , , . .

L TOP VIEW AWME MD ,

I r

8 '

\kh.\ \

\, . . ,,

UV TRIP PADOLE

\k L ..

W

\s AR}fA'1URE PAD AR}tATURE 1