ML20215L543

Special Rept:On 870603,during Normal Startup of Reactor,One Linear Percent Channel Failed to Indicate Power Level on Panel Meter.Caused by Malfunction of Test Switch.Maint Practice Changed to Include Cleaning of Test Switch

ML20215L543
Person / Time
Site:	University of Iowa
Issue date:	06/17/1987
From:	Hendrickson R IOWA STATE UNIV., AMES, IA
To:	NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III)
References
NUDOCS 8706260089
Download: ML20215L543 (2)

Text

9

• Department of Nuclear Engineering 261 Sweeney Hall IOWA STATE ^=# io- 5*"

UNIVERSITY T* phone 515 2m TRIORITY ROUTIf:G Docket No. 50-116 ~

g_ >

.1 _

June 17, 1987 _ <

I g Zf~~i T~

(/1 C J }j-4 f1LE //

U. S. Nuclear Regulatory Commission Office of Inspection and Enforcement - Region III 799 Roosevelt Road Glen Ellyn, IL 60137 Re: Malfunction of a Reactor Safety Channel The enclosed "Special Report" was written to describe an interesting failure and is submitted for your information. The event described is classified as an " unplanned shutdown", but the malfunction that caused the shutdown deserves some discussion.

As described in the report, the failure was traced to a manually operated switch used to test the function of a scram circuit in a one-of-two safety system prior to startup. The switch failed to complete the switchover back to the operating position and the scram circuit was deprived of the true signal.

Questions may be directed to me at 515-294-6422.

Sincerely, 0-Wdbbh .

Richard A. Hendrickson Reactor Manager

/rpa Enclosure C: U. S. Nuclear Regulatory Commission Washington, D. C. 20555 h

Attn: Document Control Desk k\s gpggg S

g6

.JUN 191987.

k

I Docket No: 50-116 Special Report Malfunction of a Safety Channel Test Switch Description of the Event During a normal startup of the Iowa State UTR-10 on June 3, 1987, the pre-critical tests of the high power scram circuit for each of the two redun-dant linear percent power channels were completely satisfactory. However, as reactor power was increased above 50%, the senior reactor operator noted one linear percent power channel was not indicating power icvel on the panel meter. (The scale on the analog meter indicating power for this channel is not graduated until 20%.) The SRO ordered the reactor shut down, and he initiated the investigation to discover the cause.

Cause of the Malfunction One essential task performed by the operate-test switch is the disconnec-tion of the true signal and substitution of the test signal when it is in the test position. While in the test position, the test switch logic prohibits normal operation until the switch is in the operate position.

There is no way to know that the true signal is available to the scram circuit, after completing the test, until the power level is high enough to be observed on the linear percent power meter.

After the reactor was shut down, the SRO performed the precritical high power scram test again, and it was successful. This showed that the panel meter was operable. The SRO, with the concurrence of the Reactor Manager, applied ordinary contact cleaner to the switch and observed normal opera-tion during subsequent reactor operation later the same day.

Conclusion:

When the test switch was returned to the operate position after the precritical test of the high power scram function, one of the many contacts of the switch did not make proper contact and the true signal was not available to the panel meter or the scram circuit.

Plan to Prevent Recurrence The design and construction of the safety channels are not questioned; they have given flawless service for almost 17 years.

Maintenance practice will be changed to include switch contact cleaning at the time of the annual calibration of the safety channels.