A real-time expert system for the condition monitoring of hydraulic control systems in a hot steel strip finishing mill

Abstract

This paper describes the evolution of an expert system for the condition monitoring of hydraulic control systems in a hot steel strip finishing mill. This intelligent management information system integrates real-time LabVIEW data acquisition with the Gensym G2 expert system, running on a Windows NT client-server computer network. The software architecture for building the predictive maintenance application is based on a conceptually new signal-conditioning technique for detecting vibration in roll stack assemblies. The design of the graphic user interfaces incorporate plant data models, where real-time feature extraction algorithms classify the failure mode and effects relating to the operational control of the mill. The use of object-oriented software methodologies for data analysis and system modelling are discussed including the identification of condition-monitoring variables and the fault classification procedures.