Fault diagnosis of machines via parameter estimation and knowledge processing—Tutorial paper

Abstract

For further improvement of the reliability and safety of machines the automatic early detection and localization of faults is of high interest. The conventional approach is to monitor some important variables like temperature, pressure, vibration and to generate alarms if certain limits are exceeded. However, developing internal faults are in this way only detected at a rather late stage. By applying static and dynamic process models and common process input and output measurements the inherent relationships and redundancies can be used to detect faults earlier and to localize them better. Changes in process and signal parameters are very well suited for fault detection. The paper describes a general methodology for machines and other processes by using few measurements, dynamic process and signal models and parameter estimation to generate analytical symptoms. Heuristic symptoms observed by the operator or based on, e.g. the process history are another source for fault diagnosis. Both kinds of symptoms are weighted by confidence measures and are fed into a knowledge based fault diagnosis procedure. The diagnosis is performed by forward and backward chaining according to causal fault-symptom trees. The considered machines consist of a motor, a drive chain and a working process or load. They may be electrical motor or combustion engine driven pumps, fans or machine tools with gear or belt drive chains. As one example, experimental results are shown for a DC motor powered feed drive of a machine tool. A summary of practical results with other machines is also given.