On the need for dynamic methodologies in risk and reliability studies

Abstract

Abstract Whenever there is complex interaction between the system components and process variables, the impact of uncertainties in the process physics on the predicted Top Event properties may be very difficult to account for without failure modeling techniques based on quantitative, time-dependent plant models (dynamic methodologies). New quantitative justification is given as to why dynamic methodologies may be needed in risk and reliability studies by demonstrating that these impacts can be large. The process under consideration is the feed-bleed cooling of BRWs following a small-break loss-of-coolant-accident (SBLOCA) through the intermittent operation of the high pressure core spray system (HPCS) and one safety relief valve (SRV). The dynamic methodology models the process by regarding pressure and water level evolution in the reactor vessel (RV) following the SBLOCA as transitions between the computational cells in the process variable space. It is shown that the predicted probability of demand for the low pressure core injection system during one hour following the SBLOCA due to the malfunction of the SRV/HPCS system can differ by more than one order of magnitude depending on: (a) the SBLOCA break size, and (b) whether the stochasticity in the HPCS and SRV demand rates arising from their interaction through the process variables is accounted for or not.