An expert system for AC induction motor vibration diagnosis

Abstract

AC induction motor is generally used in most manufacturing plants. Its breakdown can cause the manufacturing to a halted. To prevent it from breakdown, monitoring its status seems inevitable. Vibration analysis is one of the techniques that can be used to detect motor malfunction. However, this technique relies heavily on human's experience, which is scarce in Thailand. Furthermore, human experts may not available when needed. Hence, there is a need for the development of an expert system called ESMVD (an Expert System for AC Induction Motor Vibration Diagnosis) for diagnosis the failure of AC induction motor from the monitored vibration signal. The range of induction motor in this study is 10-1000 kW. The knowledge base related vibration analysis is collected from various available sources, e.g. literature, manual, human experts, Internet. The backward chaining is used as the searching strategy in ESMVD. After testing with the real data, it is clear that ESMVD can give satisfactory results. It can tell the possible source of vibration signal, both on time domain and frequency domain. Moreover, it can show status of motor and gives recommendations for further action to be taken. However, the limitation of the ESMVD is that it does not cover all the possible source of vibration signal. That is vibration on AC induction motor from its load side is excluded. To implement the ESMVD in particular manufacturing plant, further enhancement in knowledge base has to be done to increase diagnosis range to cover the load side.