Grid–sampling optimisation of safety systems

Safety systems are usually the last line of defence against the occurrence of a potentially hazardous event.Failure of a safety system on a potentially hazardous industrial system or process may have severe consequences. For a system whose failure could result in fatality it could be accepted that a merely adequate level of system unavailability is not sufficient. The aim should be to produce the optimal performance attainable within the constraints imposed on resources. This paper investigates a design optimisation scheme that is appropriate for safety systems. The methodology presented in this paper adopts the latest improvements to the fault tree analysis technique, the binary decision diagram approach, to analyse the individual system designs. The grid-sampling optimisation technique is used to generate the final design specification with the constraints incorporated. To demonstrate the practicality of the method it has been applied to a High Integrity Protection System. In all there are 42,831,360 combinations of twelve design variables. There are three constraints imposed on the system in terms of cost, mean down time, and spurious trip frequency.