An adapted component-connection method for building SBDD encoding a dynamic fault tree

Abstract

Dynamic fault trees (DFTs) are a commonly used tool to analyze the reliability of systems with sequential failure behaviors. A sum of disjoint product (SDP)-based analysis methods are widely accepted as efficient approaches for a DFT, such as dynamic binary decision trees (DBDTs) and sequential binary decision diagrams (SBDDs). However, for a large DFT, the process of obtaining the structure function is error-prone and very time-consuming. In contrast, SBDD built on the improved ite algorithm does not rely on the structure function but is limited to a DFT whose dynamic gates are located at the bottom. This method requires predefined variable ordering for basic events, which greatly influences the computational efficiency, and the caching operation is a problem. In this paper, an enhanced component-connection–based method is proposed to build SBDD encoding a DFT. New component-connection rules are developed to deal with dependent variables having repeated basic events, and several heuristic connection strategies are also developed to reduce the size of the final calculable terms. The proposed method is straightforward and easily implemented. To demonstrate the applications and merits of our method, several case studies are carried out, and the results show the reasonability and effectiveness.