Optimal Safety Control System Design

Abstract

This paper is focused on the optimal design of Safety Instrumented Systems. A combinatorial optimisation problem is formulated that involves as design variables the architecture and type of components of the three subsystems (sensors, logic solvers, final elements) and their test intervals. The life cycle cost is minimized subject to safety and spurious trip constraints. An evolutionary technique based on Genetic Algorithms is presented to solve the problem. Application results are reported for case studies.