Consideration of human actions in combination with the probabilistic dynamics method Monte Carlo dynamic event tree

Abstract

The various accident sequences to be considered in the framework of a probabilistic safety analysis (PSA) for a nuclear power plant derive from interactions between the physical process, technical system functions, operator performance and stochastic influences along the time axis. Probabilistic dynamics methods have been developed to adequately account for these interactions. They can potentially cover the spectrum of event sequences which may actually evolve and achieve a realistic probabilistic safety assessment. The probabilistic dynamics method MCDET is a combination of Monte Carlo simulation and the discrete dynamic event tree (DDET) method. It was implemented as a module which can operate in tandem with any deterministic code simulating the system and process dynamics. MCDET was supplemented by a so-called Crew-Module which permits to simulate the dynamics of human actions depending on but also acting on the system and process dynamics as modelled in the deterministic code and on stochastic influences as modelled in MCDET. This paper presents the Crew-Module and gives an overview of the results which may be obtained from its combination with MCDET and a deterministic dynamics code.