Automated diagnosis of motor-operated valves using motor power signature analysis

The development of nonintrusive diagnostics methods makes it possible to identify malfunctions in plant components during normal plant operation and to avoid catastrophic failures and associated costs. If a plant component is driven by an electric motor, then the measurement of electrical variables, generally provides enough information to establish the operability of the component. Electrical measurements have been used by Duke Power Company and Oak Ridge National Laboratories, among others. The purpose of this research project is to develop an automated diagnostics system to identify and isolate operational malfunctions in motor-operated valves (MOVs). The traditional approach uses motor current signature analysis for valve monitoring. In this research project the motor power signature analysis (MPSA) was developed as an alternative and a more sensitive technique for MOV diagnosis. A PC-based expert system was developed for integrating MOV motor power pattern analysis, a rule-based expert system for anomaly detection and classification, a fuzzy logic decision module, a data interface between the expert system and an existing data acquisition system, and a user interface with on-line help. The diagnostic system was than tested using data from a power utility and laboratory valve test data acquired at The University of Tennessee. The system was demonstrated for on-line MOV anomaly diagnosis. The present system has proved to be very effective in detecting MOV malfunctions with minimum user input.