System failure minimisation using automated design optimisation

2012-01-23T13:51:36Z (GMT) by D. Astapenko L.M. Bartlett
Safety systems are designed to prevent the occurrence and future development of hazardous situations. Consequences of the failure of a safety system varies from minor inconvenience and cost to personal injury, significant economic loss and death. The operation of a safety system can be improved by either introducing better performing components or by increasing the number of redundant components. At the same time, such design alterations can influence how available resources are utilized. The focus of this paper is to introduce a generic optimisation method for constructing an optimal design case for any safety system, with the aim of maximising its likelihood of functioning on demand and at the same time ensuring optimal usage of available resources. The analysed optimisation problem is represented as the constrained single objective problem. The implemented optimisation method employs Fault Tree Analysis (FTA) to represent system failure causes and Binary Decision Diagrams (BDDs) to quantify its failure probability. A Single Objective Genetic Algorithm (SOGA) has been chosen as the optimisation technique. The methodology is illustrated with the optimisation of a High Integrity Protection System (HIPS) design. The constraints imposed are on system dormant failure probability, cost and maintenance down time. Results of the application, with the generic implications of the analysis, are discussed.