Fault tree analysis is usually used in the reliability evaluation of industrial systems. However, problems may be encountered in terms of accuracy and efficiency when complex systems are studied. To overcome such disadvantages, the BDD (binary decision diagram) has been developed. The efficiency of this method depends on the choice of a variable ordering scheme. Indeed the choice of a variable ordering scheme has a significant effect on the resulting BDD size. This article thus proposes a NN (neural network) methodology suited for the fault tree analysis. In order to optimise the NN size, a heuristic association and ordering approach is proposed.
展开▼