ML20207L083

Safety Evaluation Re Systematic Evaluation of Ipsar Topic VII-1.A, Isolation of Reactor Protection Sys from Non-Safety Sys. Issue Resolved

ML20207L083
Person / Time
Site:	Oyster Creek
Issue date:	10/11/1988
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20207L077	List: ML20207L074 ML20207L083
References
TASK-07-01.A, TASK-7-1.A, TASK-RR NUDOCS 8810170170
Download: ML20207L083 (3)

Text

. _ . _ _ _ _ _ _ _ _ _

p na

1. vy'o *

g UNITED STATES l' o NUCLEAR REGULATORY COMMISSION 5 E W ASHINGTON. D. C. 20555 t c SAFETY EVALUATION BY THE OFFICE OF NUCLEAR POWER REGULATION RELATED TO THE SYSTEMATIC _ EVALUATION IPSAR TOPIC VII-1.A (IPSAR SECTION 4.2],}

"ISOLATION OF REACTOR PROTECTION SYSTEM FROM NON-SAFETY SYSTEMS GPU NUCLEAR CORPORATION OYSTER CREEK NUCLEAR GENERATING STATION l DOCKET NO. 50-219 l

1.0 INTRODUCTION

i By letters dated August 3,198*, July 8,1985, April 4.1986 and August 19, 1968 (Reference 1-4), GPU huclear^ transmitted the results of their evaluation concernir.g SEP topic,VII-1.A and their responses to NRC request for additional inf omation.

(

2.0 DISCUS $10N .

SEP topic YII-1.A concerns the electrical isolation of the reactor protection system (RPS) from non-safety systems to assure that the safety systees will function as required. At Oyster Creek there were three areas which did not appear to have qualified electrical isolation. The first was the direct path between the non-safety computer and the RPS. The second is the connection i between the nuclear flux monitoring system interinediate range monitors (1stMs),

1 average power range monitors (APRMs) and the computer. The third item was the electrical isolation between the IRMs and APRMs and the non-safety process recorders for these systems.

In response to the staff questions concerning the isolation between the RPS and the non safety computer, the licensee stated '(Reference 2) that they provided relay (coil to contacts) isolation. The computer tie-ins are accessed on separate contacts of the particular relays associated with RPS logic. This method provides isolation which will prevent a fault within the computer from propagating to the RPS.

G810170170 881011 PDR P ADOCK 05000219 PDC

l l

The second item concerned the connection between the nuclear flux monitoring and the computer. Intheirresponse(Reference 4)tostaffquestionsthe '

licensee provided details of the isolation provided which included as it's primary element a Burr-Brown isolation ar.plifier. The staff has previously  ;

reviewed and accepted Burr-Brown isolation devices for other applications.

The third item concerns the neutrr.,n flux menitoring and their process recorders.

During the integrated evaluation of the subject SEP topic, the NRC staff requested GPUN to perform a failure mode and effects analysis (FNEA) to evaluate the need for isolation devices between the nuclear flux monitoring system and their process recorders. The staff previously agreed that a properly performed FNEA can demonstrate that comon mode electrical faults do not exist. The licensee has submitted the FMEA (Reference 1) and their responses to subsequent staff ouestions (References 2-4). .

As described in the FMEA, at Oyster Creek a failure at one IRM/APRM process recorder can affect at most one APRM (or IRM when in the startup mode), and will affect only one sub channel of the RPS. The process recorders are actually two separate, independent measuring and recording systems within each recorder cocplete with separate internal power supplies though they are supplied by a comon power source cable. Though the recorder has input filters, choppers and transfonners which would reduce a fault in the recorder from propagating bhck to the IRM/APRM sub channel, they have not been shown to be qualified isolation devices. The RPS logie is wired in a one out of two taken twice configuration with one process recorder for each of the sub channels. Each sub channel is further subdivided with two separate inputs. A fault in one recorder section would affect only one-half of one sub channel with the worst case fault being one which propsgates through both recorder sections and both APRM/IRM inputs and therefore disables one sub channel. Due to the RPS logic, a failure of a single sub channel will not compromise RPS logic capability to provide a reactor scram.

I l

l l

.~

3.0 CONCLUSION

The staff has concluded that SEP topic VII-1A has been satisfied at Oyster Creek. For the first item the staff accepts relay (coil to contact) isolation to provide isolation between the RP! and the non-safety computer. For the electrical isolation between the nuclear flux monitoring .nd the computer, the staff finds the isolation amplifier to be acceptable. For the last item, isolation between the nuclear flux monitoring and the process recorders, the staff finds the recorders to be unacceptable as isolation devices, however the FMEA demonstrated that the logic configuration of the RPS and the internal component separation within each recorder as described above would not inhibit the RPS from performing its safety function, and therefore, the staff finds that additional electrical isolation is not required for this interface.

4.0 REFERENCES

l

1. Letter dated August 3, 1984 Fiedler(GPU)toCrutchfield(NRC) Failure Modes and Affects Analysis

2. Letter dated July 8, 1985 Fiedler(GPU)'.oZwolinski(NRC)Responseto Questions

3. Letter dated April 4, 1986 Fiedler(GPU)toZwo11nski(NRC)Responseto Questions

4. Letter dated August 19, 1988 Nagat(GPU)toDramerick(NRC) Additional Information -

Principal Contributor: J. Stewart Dated:

e

.