Optimization method to branch-and-bound large SBO state spaces under dynamic probabilistic risk assessment via use of LENDIT scales and S2R2 sets

Abstract

Traditional probabilistic risk assessment (PRA) methods have been developed to evaluate risk associated with complex systems; however, PRA methods lack the capability to evaluate complex dynamic systems. In these systems, time and energy scales associated with transient events may vary as a function of transition times and energies to arrive at a different physical state. Dynamic PRA (DPRA) methods provide a more rigorous analysis of complex dynamic systems. Unfortunately DPRA methods introduce issues associated with combinatorial explosion of states. In order to address this combinatorial complexity, a branch-and-bound optimization technique is applied to the DPRA formalism to control the combinatorial state explosion. In addition, a new characteristic scaling metric (LENDIT – length, energy, number, distribution, information and time) is proposed as linear constraints that are used to guide the branch-and-bound algorithm to limit the number of possible states to be analyzed. The LENDIT characterization is divided into four groups or sets – ‘state, system, resource and response’ (S2R2) – describing reactor operations (normal and off-normal). In this paper we introduce the branch-and-bound DPRA approach and the application of LENDIT scales and S2R2 sets to a station blackout (SBO) transient.