Stochastic programming in dynamic reliability

PSA results are expected to be used to improve the design, the operating procedures or the safety policy of a system. It then appears logical to combine a risk analysis with an optimisation procedure, aiming at minimising a cost function to determine the most appropriate strategy. This idea is applied in this work to dynamic systems, i.e. installations for which the probabilistic risk analysis must be conducted using dynamic reliability methods. As several system parameters can be uncertain, the cost function turns out to be a random variable. Stochastic programming techniques are convenient for such calculations. The whole optimisation procedure is presented, and a first application of the algorithm on a HNO3 supply system is provided.