Fault prediction oVers a new perspective to diagnostic engineering systems by covering a wider area of the diagnostic task. This paper presents the fault prediction and diagnosis process of a knowledge-based diagnostic system that is able to predict and diagnose faults in hydraulic systems. Expert systems technology is used cooperatively with dynamic modelling information and on-line sensor information in a suitable environment for the interaction of symbolic and numerical data. This system has been successfully implemented and tested on an actual electrohydraulic system.

Increased interest in problem-solving decisions in the area of engineering lea& to an increased demandfor expert system applications. This paper presents an expert system that detects faults in hydraulic systems by simulating the reasoning ofthe expert, using as guide the hydraulic circuit ofthe system on the screen and extensive graphical information. A graphical user inteflace, appropriate explanations by using hypertext and user modelling have been implemented in order to produce a system that overcomes the usual interface problems of expert systems in engineering. This system has been developed in close collaboration with the company Automation Systems SA (Mannesmann Rexroth -Greece) and is currently used to detect faults in hydraulic presses, a fact thatpresents evidence of its usefilness.

The development of on-line fault detection methods for drive and control systems is of main importance for the modern production technology. Modelling information improves the reliability of the diagnostic method when it is involved in a fault detection system. In this paper, the use of modelling information for the fault detection of hydraulic driven machines as well as for the compensation of incipient faults is presented. For this purpose a suitable implementation environment was developed that allows the on line interaction of real time data and simulation results and makes possible their direct effect to the actual system.

No abstract

No abstract