In August 1992, the US Nuclear Regulatory Commission (NRC) began a maj
or initiative to develop requirements for containment testing that are
less prescriptive and more performance oriented and risk based than c
urrent requirements. This action was a result of public comments and s
everal studies that concluded that the economic burden of certain pres
ent containment testing requirements was not commensurate with their s
afety benefits. The rule-making included considering relaxing the allo
wable containment leakage rate, increasing the interval for the contai
nment integrated leak rate test, and establishing intervals for the co
ntainment local leak rate tests on the basis of the performance of con
tainment isolation valves and penetrations. A study has been conducted
to provide technical information to establish the preformance criteri
on for containment tests, i.e. the allowable leakage rate, commensurat
e with its significance to total public risk. For the study, the resul
ts used were from a comprehensive study conducted by the NRC (NUREG-11
50, 'Severe accident risks: an assessment for five U.S. nuclear power
plants') to examine the sensitivity of containment leakage to public r
isk. Risk was found to be insensitive to a containment leakage rate up
to levels of about 100% volume per day for certain types of containme
nts. Probabilistic risk assessment methods have also been developed to
establish risk-based intervals for containment tests on the basis of
experience. Evaluations show that increasing the interval for the inte
grated containment leakage test from three times to once every 10 year
s would have an insignificant impact on public risk. Analyses of opera
tional experience data for local leak rate tests show that performance
-based testing (valves and penetrations that preform well are tested l
ess frequently) is feasible with a marginal impact on safety. These te
chnical studies have been used to develop efficient (cost-effective) r
equirements for containment tests.