ABSTRACT
Risk-informed decision making is becoming common practice e.g. in nuclear, chemical and transportation industries. Typical subjects for decision making are ranking or prioritizing Systems, Structures, Components or Human actions (SSCH) for enhanced maintenance, quality assurance, safety categorization, periodic testing programs or other such Preventive Safety-Assurance Activities (PSAA), and optimizing these programs. Another area of applications is control of plant state or system configurations when failures occur in operation, orwhenplanning plant outages or component maintenances. A necessary tool in this decision making is a high quality plant-specific probabilistic safety assessment (PSA) and the capability to obtain accurate quantitative risk results and importance measures. Computerized fault-tree and event-tree software tools have been used in this process extensively. However, there are signs that some of the techniques used may not be sufficiently accurate, or that full potential of the tools may not be used (Epstein & Rauzy 2005). This paper addresses several of these issues.
