Reliability and performance of multi-state systems with propagated failures having selective effect

The paper presents an algorithm for evaluating performance distribution of complex series-parallel multi-state systems with common cause failures caused by propagation of failures in system elements. The failure propagation can have a selective effect, which means that the failures originated from different elements can cause failures of different subsets of system elements. The suggested algorithm is based on the universal generating function approach and a generalized reliability block diagram method (recursive aggregation of pairs of elements and their replacement by an equivalent one). The performance distribution evaluation procedure is repeated for each combination of common cause failures. Illustrative examples are provided.