Risk-Informed Testing Program for Emergency Diesel Generators

Abstract

Risk-informed justification and optimization of surveillance test intervals of standby safety systems based on risk is possible only with sophisticated component models having failure probabilities (unavailabilities) dependent on the intervals. Furthermore, it is necessary to carry out detailed failure cause analysis to determine what fraction of failures may be caused by stresses associated with the startup-demands themselves rather than stresses between the tests while the system is on standby. Failures and defects (critical and incipient failures) also have to be clearly separated in the analysis and in the models. This paper illustrates the issues and solutions with an example applied to emergency diesel generators (EDG) at a nuclear power plant. Extension of test intervals is justified with reduced burden when it has a minor impact on the core damage risk.