Vizintin J.

This paper presents a design of a complex intelligent expert system for failure diagnostics of rotating machinery. The key idea is to synergistically combine machine lubrication fluid analysis and machine vibration signals analysis into an intelligent hybrid system which is able to perform on-line rotating machine monitoring as well as to detect and recognize different kinds of faults of such machines. The first part of the paper presents a scheme of proposed system. It is divided into four modules, which perform specific tasks: the signal processing module, the feature management module, the fault classification module and the fault database module. The second part of the paper explains the purpose of these four modules and presents an overview of recently published methods, which seem to be applicable in the proposed failure diagnostics system. A modular approach was selected because it allows an implementation of the system into a particular rotating machine with a minimal number of adjustments necessary to be made The proposed system is being developed in cooperation with a Slovenian steel production plant and a Slovenian power plant. However, we are interested to widen our cooperation to other interested partners worldwide.