One fast fault isolation algorithm based on Bayesian network

The contribution of this paper is to prove the equivalence of the conditional distribution item and the item of MYT (Mason Young Tracy) decomposition (a decomposition method of the T² statistic) in fault isolation and root cause identification. The experiment results verified its equivalence. Fault isolation is done by using the dependency between variables in Bayesian network, and using this technique has no limitation on variables´ statistical distribution. Comparing with the traditional MYT decomposition, the computational work of the MYT-based conditional probability items reduces dramatically when variables follow normal distributions in Bayesian network, which leads to a fast fault isolation.