Operability Testing of Relief and Safety Relief Valves (Generic Letter 79-27)

GL79027 

                               UNITED STATES 
                       NUCLEAR REGULATORY COMMISSION 
                          WASHINGTON, D. C. 20555 

                               July 16, 1979 

ALL BOILING WATER REACTOR LICENSEES 
(Except Big Rock Point and Humboldt Bay) 

Gentlemen: 

In December 1977, we sent letters to the majority of licensees who operate 
Boiling Water Reactors (BWR) regarding the relief and safety-relief that are
installed in the reactor coolant system and/or the automatic 
depressurization system. The letters requested licensees to propose 
Technical Specification changes to incorporate additional Surveillance 
Requirements for these valves. Model Technical Specifications were included 
for guidance in preparing plant specific requirements. The principal feature 
of the new requirements was a variable frequency test schedule for 
operability testing of relief and safety-relief valves. 

Some licensees objected to this feature on the basis that increased testing 
of the Target Rock safety-relief valves could significantly degrade valve 
reliability because such testing could aggravate pilot valve leakage thereby
increasing the likelihood of future malfunctions. This objection was not 
voiced by all licensees and was not shared by the NRC staff at the time. 
However, we did believe that further consideration of this view, within the 
context of overall reactor safety, was warranted. 

We have since made an independent study of BWR pressure relief system 
failures. The results of the study have been published in NUREG-0462, 
"Technical Report on Operating Experience with BWR Pressure Relief Systems",
dated July 1978. A copy is enclosed for your convenience. Based on the 
findings of this report and further information obtained from General 
Electric, in our meeting of March 30, 1979, we have concluded that 
implementation of a requirement for increased surveillance testing would not
be the most effective way of assuring safety-relief valve reliability. 

Consequently, unless you supply information to the contrary, we do not plan 
to act on any proposed Technical Specification changes you may have 
submitted in response to our December 1977 request. However, due to the 
potential effects of safety-relief valve malfunctions, the NRC staff 
continues to believe that licensees should make all reasonable efforts to 
increase the reliability of these valves and to reduce the frequency of 
inadvertent actuation and subsequent failure to reseat properly. This 
general matter is further discussed in the NUREG-0560 staff report 
concerning the Three Mile Island Unit 2 accident, wherein the pressurizer 
power operated relief valve failed to close during a feedwater transient and 
resulted in  a small break LOCA. Staff review of other operating events 
indicates a significant frequency of such valve failures leading to small 
break LOCA events. Accordingly, reliability goals are currently being 
developed by the staff for safety and relief valves which are part of the 
reactor, coolant pressure boundary, consistent with the recommendations of 
NUREG-0560. 
.

                                  - 2 -                      July 16, 1979 

General Electric has made recommendations to licensees that we believe would
substantially reduce the likelihood of future failures of BWR safety-relief 
valves. General Electric's recommendations consist of a short term and a 
long term program. Basically, the short term program consists of an 
intensified maintenance program and minor modifications to the valve 
assembly which will enable the simmer margin of the valve to be increased to 
about 120 psi. General Electric has provided operating data that indicate 
the malfunction of valves having a simmer margin of about 100 psi is 
appreciably less than those with smaller simmer margins. The long term 
program consists of replacing the original three stage pilot operated 
actuator with a redesigned two stage pilot operated actuator. It is our 
understanding that the newly designed actuator has, by tests, demonstrated 
improved reliability due to the elimination of the bellows and its reduced 
sensitivity to pilot valve leakage. 

The NRC plans to continue to monitor the performance of safety/relief 
valves, and the status and effectiveness of actions to improve their 
reliability over the long term. To apprise us of the current situation at 
your plant(s), as well as your plans for future actions, we request that you 
provide responses to the items identified in the enclosure within 60 days. 
If your plant design does not utilize Target Rock safety/relief valves, so 
indicate within 60 days; and disregard the enclosure. 

If you have any questions, or care to discuss this matter, please contact 
us. 

This request for generic information was approved by GAO under clearance 
number B-180225 (S79014); this clearance expires June 30, 1980 

                                   Sincerely, 


                                 Brian K. Grimes, Acting Assistant Director 
                                   for Systems Engineering 
                                 Division of Operating Reactors 

Enclosures:
1.   Request for Information
2.   NUREG-0462, dated July 1978
.

                                ENCLOSURE 

                          REQUEST FOR INFORMATION 

                     TARGET ROCK SAFETY/RELIEF VALVES 

1.   What is the status of each of the Target Rock safety/relief valves at 
     your plant(s); i.e.: 

     a.   Are they in their original design configuration? 

     b.   What is the existing simmer margin? 

     c.   What modifications have you implemented to improve reliability? 

     d.   On what date were these modifications made? 

2.   What maintenance and testing do you routinely perform on these valves 
     and how often is it performed? 

3.   What additional modifications and/or maintenance do you plan to 
     implement in the future? 

4.   On what date will the modification(s) and/or maintenance in item 3 be 
     implemented?