Novel method of fuzzy failure probability analysis on the basis of subset sampling

For fuzzy failure probability evaluation of structure with fuzzy failure and fuzzy safety state, a novel method is presented on the basis of subset sampling. Firstly, the fuzzy failure region is separated as some subsets, in which the value of performance function keeps a constant approximately and membership of the performance function to the fuzzy failure state keeps a constant approximately as well. Secondly, Markov Chain based Monte Carlo method is employed to calculate the random failure probability of each subset. Lastly, the products of the random failure probability and die corresponding approximate constant membership value at each subset to the fuzzy failure state are added to obtain the fuzzy failure probability. Comparing to Monte Carlo method, the efficiency of the presented method is improved greatly while the precision of the fuzzy failure probability evaluation is acceptable.