Optimization of RAMS plus C for safety instrumented system design with diverse redundancy

This paper addresses the optimization of safety instrumented systems design based on a RAMS+C approach using a multi-objective genetic algorithm. The design includes optimization of safety and reliability measures against lifecycle cost, paying special attention to the effect that different diverse redundancy schemes have on common cause failure and thus how they affect the overall system objectives. The safety integrity requirements and modelling detail indicated by the standard IEC 61508 are addressed. The problem is approached as one of redundancy allocation with diversity and design variations. In order to include safety integrity, the Average Probability of Failure on Demand is one objective to optimize. This is complemented with the Spurious Trip Rate (safe failure) and Lifecycle Cost as additional objectives. For demonstration of the overall method, a practical example from chemical industry is used: a protection against high pressure and temperature for a chemical reactor. Three different technologies are available for use per subsystem to implement diversity, at the same time also modifying the diagnostic coverage.