Fault detection and diagnosis in manufacturing systems: a behavioral model approach

Abstract

An approach to online fault detection and diagnosis in automated manufacturing systems with discrete controls and sensing is described. The approach is based on the concept of behavioural models of the individual system components. These models, which can be developed while the system is being designed, characterize the responses of the devices in the system to arbitrary input signals over the range of acceptable operating conditions. The expected flow of signals through the system, from control inputs to sensor outputs, is captured in the behavioural model dynamics. This model provides the basis for online fault detection by generating expected system response signals which are compared online, in real-time, to the actual sensor signals from the system. Fault diagnosis is accomplished by maintaining a current set of operational assumptions which identify the system components which could cause deviations from the expected behavior.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>