ML19290E688

From kanterella
Jump to navigation Jump to search
Forwards Analysis & Evaluation of Plant Incident. General Recommendation Categories Include Coping W/Various Loss of Instrumentation & Control Function Combinations & Improving Instrument Indication Reliability
ML19290E688
Person / Time
Site: Crystal River Duke Energy icon.png
Issue date: 03/11/1980
From: Wilkinson E, Zebroski E
AFFILIATION NOT ASSIGNED, ELECTRIC POWER RESEARCH INSTITUTE
To: Harold Denton
Office of Nuclear Reactor Regulation
Shared Package
ML19290E689 List:
References
NUDOCS 8003140386
Download: ML19290E688 (2)


Text

, . -

NUCLE'AR SAFETY lg D 3

ANALYSIS 1820 Water Place CENTER Atlanta, GA 30339 g (404) 953-3600 Neketre utthtv industry En the '

~

EJectric iwer ,- ,_v -

wg - - ,s

[nstitute March 11, 1980

=n= p- aid Demn Nuclear Regulatory Commission Washington, D. C.

Subject:

Report on Crystal River Unit-3 Incident of February 26, 1980, by NSAC and INPO

Dear Dr. Denton:

In accord with discussions with Mr. William Lee of Duke Power and Mr. Andrew Hines of Florida Power, there is transmitted herewith our report on the referenced incident.

The recommendations represent the best independent judgements of NSAC and INPO staff in the time available. They do not, and are not intended to represent a concensus of utility or contractors' views. Factual inputs - as distinct from recom-mendations - were solicited from all B&W owner utilities and from B&W.

Our recommendations are set forth in the attached report and are grouped in order of priority. The general categories include:

1. Those which will allow the operator to cope with various combinations of loss of instrumentation and control functions. This may involve changes in (A) equipment and control systems to give clear indications of functions which are lost or unreliable; (B) procedures and training to assure positive and safe manual response by the operator in the event that competent instruments are unavailable.
2. Those which Lmprove reliability of instrument indications to accomodate the range of faults which have been observed.

This may involve either or both (A) power supply Dmprove-ments to insure greater independence and redundancy, (B) independent instrumentation similar to the " system state vector" indications proposed in NUREG 0585 to provide essential shutdown informati^n. Functional tests may be desirable on a periodic basis. 7 3412 HdivewAvenue &O 0' Ibst Offre Bm 10412 PabAltaCA94303 PQ 9L9 85-2am /o 8003yao3 %

1 f

. 4 .

Letter to Denton '

March 11, 1980 Page Two

3. Those which require analysis of the response of the control system to various faults: from input or output signals or input or output power supplies. Design changes may be required to limit any improper responses which might occur under fault conditions.

We believe that the operators coped with the Crystal River-3 incident adequately and maintained ample margins to protect public health and safety. However, as noted above, it is desirable for all plants to review procedures and training to ensure that orderly manual shut down can occur for the full range of the types of faults which have occurred. We are of the opinion that if, by a combination of equipment and system upgrading and op ator training, this ability to shut down is demonstrable, then Item and Item 3 above can be paced so as to insure the time required to attain high quality in the design, analysis, and implementation of changes which may be appropriate.

We are expecially sensitive to the need for a deliberate approach on such changes, As emphasized by the ACRS recently - a change which appears to improve one function must be exhaustively analyzed and tested before implementation is required so that it does not increase risk of having a malfunction occur in another system. We see again that among the general causes of the Crystal River event was the lack of dissemination of information about events of a similar nature which occurred earlier.

We hope that the attached report will help our mutual objective of making the cumulative learning process work better.

SQ d}f E. P. Wilkinson E. L. b oski cc: W. Lee F. W. Lewis B. Lee F. Culler