Research on the method of testability modeling based on Bayesian Networks

Abstract

Due to the complexity of the fault mechanism and the testing method, the research on the Design for Testability (DFT) for mechanical system becomes difficult. In this paper, according to characteristics of mechanical system, it puts forward using the combination of the Bayesian Networks (BN) and the dependency matrix for testability modeling. The method gives full play to BN's probabilistic reasoning ability, improves the traditional testability modeling method and solves the problems that the traditional testability modeling method just basing on qualitative analysis and considering the condition of insufficiency. It is helpful to develop the testability analysis quickly and accurately and optimize test strategy for the mechanical system.