Proceedings of the
The 33rd European Safety and Reliability Conference (ESREL 2023)
3 – 8 September 2023, Southampton, UK
PSA Model Approaches for Asymmetries in the Electrical Power Supply for Nuclear Power Plants
1Safety Analysis Department, Gesellschaft für Anlagen- und Reaktorsicherheit (GRS) gGmbH.
2Plant Operation Department, Gesellschaft für Anlagen-und Reaktorsicherheit (GRS) gGmbH.
ABSTRACT
In the more recent past, several events reported from nuclear power plants (NPPs) indicated that failures caused by asymmetries in the electric power supply can trigger correlated failures of redundant trains of safety related systems. Since the risks from these incidents had been rarely investigated for German NPPs, an existing RiskSpectrum® Level 1 probabilistic safety analysis (PSA) model of a German pressurised water reactor (PWR) has been extended by GRS to include independent and correlated failures caused by asymmetries.
Five different approaches have been developed to model the failures. Four approaches take correlated failures into account. Two of them use the common cause failure groups of RiskSpectrum. The other two are more complex and need different steps with different computer programs, e.g., for external parameter sampling. The evaluation of the five approaches and the sensitivity study of modelling assumptions led to the following results and conclusions. The core damage frequency (CDF) caused by correlated failures resulting from asymmetries is clearly higher than the overall CDF from plant internal initiating events if no instrumentation and control (I&C) equipment to detect and protect against asymmetries is installed. Hence, the consideration of correlated failures caused by asymmetries in PSA models seems to be relevant. But only one scenario of asymmetries has been considered and more scenarios are necessary to realistically reflect their effects.
Keywords: Asymmetry, Phase fault, Common cause failure, Correlated failure, Electrical power supply, External hazard, Model development, Nuclear power plant, PSA model.