Abstract
This paper presents a new fault diagnosis strategy for the event-driven controlled systems such as Programmable Logic Controller. First, the controlled plant is considered to be divided into some subsystems and the time interval between succesive two events in each subsystem is modeled as a randam variable. Then, the fault diagnosis algorithm is proposed by using the Bayesian Network (BN), which can represent the causal relationship between subsystems. Furthermore, the design algorithm for the structure of the BN is developed using the information on the control law and the controlled plant. As a result, large scale diagnosis problem can be addressed with less computational burden. Finally, the usefulness of the proposed strategy is verified through some experimental results of an automatic transfer line.
