Mutual integration of classical and dynamic PRA

Dynamic Probabilistic Risk Assessment (PRA) methods couple stochastic methods (i.e., sampling methods) with system simulators (e.g., RELAP5-3D and MELCOR) to determine the risk associated to complex systems such as nuclear power plants. Compared to Classical PRA methods they can evaluate with higher resolution the safety impacts of timing and sequencing of events on the accident progression without the need to introduce overly conservative modeling assumptions and success criteria. When comparing Classical and Dynamic PRA approaches two considerations might arise from the analysis. First, as part of a Dynamic PRA, it is not uncommon that some components of the system under consideration might not require a complex and computationally expensive simulation model due to their intrinsic characteristics (e.g., no time or physics dependency). From a modeling point of view, such components could be included in the analysis by employing simpler Classical PRA models such as Event Trees (ETs) or Fault Trees (FTs). Second, studies have indicated that timing/sequencing of events might impact accident progression modeled in an ET. Typically, only a limited number of aspects of Classical PRA might require simulation enhancements due to the intrinsic dynamic characteristics of the considered system. This paper addresses both of these issues by presenting several methods and algorithms that can be employed to link Classical PRA models into a Dynamic PRA (i.e., first issue) and to integrate Dynamic PRA results into Classical PRA (i.e., second issue). These algorithms have been developed within the RAVEN statistical framework and applied to a few test cases.