Distributed Identification of Discrete-Event Systems with the Aim of Fault Detection

Abstract

Recently, different Discrete-Event Systems (DES) identification techniques with the aim of fault detection have been proposed in the literature. In general, a monolithic model is identified using the observed input and output binary signals exchanged between the plant and the controller. However, if the system is composed of several subsystems with concurrent behavior, the number of states and transitions of the monolithic model may not converge even observing a large number of signal vectors. In these cases, it is necessary to use a distributed identification technique, where models for the fault-free behavior of the subsystems are obtained, and then these models are run in parallel for fault detection. In this work, we present a method for identifying concurrent subsystems based only on the observation of the input and output controller signals. A virtual plant, simulated using a 3D simulation software and controlled by a programmable logic controller, is used to illustrate the proposed method and to compare the efficiencies of the fault detection methods using the subsystem models and the monolithic model.