Parameter uncertainty modeling of safety instrumented systems

In this study, a novel safety integrity level (SIL) determination methodology of safety instrumented systems (SISs) with parameter uncertainty is proposed by combining multistage dynamic Bayesian networks (DBNs) and Monte Carlo simulation. A multistage DBN model for SIL determination with multiple redundant cells is established. The models of function inspection test interval and function inspection test stages are alternately connected to form the multistage DBNs. The redundant cells can have different M out of N voting system architectures. An automatic modeling of conditional probability between nodes is developed. The SIL determination of SISs with parameter uncertainty is constructed by using the multistage DBNs and Monte Carlo simulation. A high-pressure SIS in the export of oil wellplatform is adopted to demonstrate the application of the proposed approach. The SIL and availability of the SIS and its subsystems are obtained. The influence of single subsystem on the SIL and availability of the SIS is studied. The influence of single redundant element on the SIL and availability of the subsystem is analyzed. A user-friendly SIL determination software with parameter uncertainty is developed on MATLAB graphical user interface.