Fault Diagnosis of Electric Motors Using Vibration Condition Monitoring Technique

Tee, Deng Yeong (2020) Fault Diagnosis of Electric Motors Using Vibration Condition Monitoring Technique. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Tee Deng Yeong Full Text.pdf Restricted to Registered users only Download (2MB)

Abstract

Electric motors are widely used in the industry due to its high reliability and easy access to electricity and power supply. Hence, it is very important to prevent faults in electric motors to prevent failure so that financial damage and safety concerns can be reduced and eliminated. Motor faults are categorized into mechanical faults and electrical faults and all these faults can cause failures in electric motors. Predictive maintenance such as vibration condition monitoring is widely used to monitor the real time condition of electric motors as well as to detect mechanical faults in electric motors while motor current signature analysis is used to detect the electrical faults in electric motors. Vibration condition monitoring techniques are proven to be effective in detecting mechanical faults in electric motors but its application in detecting electrical faults in electric motors is still not recognized. Hence, this project uses vibration condition monitoring technique to detect both electrical and mechanical faults in electric motor so that only one technique is required in the predictive maintenance of electric motors. Besides a more comprehensive classification of electric motor faults using 3 axis vibration analysis at 3 different locations on motor is also performed. It is observed that at different locations and at different axis, the response is different. The response at Y-axis is observed to be able to able to differentiate the mechanical fault of broken cooling fan from the mechanical fault of unbalanced load, which is not differentiable by the other 2 axis. Meanwhile, front location allows the loose front frame fault and shaft wear fault to be further differentiated as the front location is able to provide different vibration response for these 2 faults, while the main body provides the same response without differentiating these 2 faults. Hence, the classification of the mechanical faults are performed according to the Y-axis at the front location as the Y-axis at the front location shows different response in all the mechanical faults. The project observes that the unbalanced load fault will cause a new peak to be produced at 24 Hz, the loose front frame fault will cause multiple peaks from 472 Hz until the maximum settings of 1024 Hz, the shaft wear fault will cause multiple peaks from 400 Hz until 672 Hz and lastly, broken cooling fan will produce two new peaks, one at 24 Hz and another one at 552 Hz. Hence, this allow the faults to be further classified which simplify repairing work and allow maintenance to be carried out more effectively. For the electrical fault, when the overcurrent fault is produced by increasing the current above the rated current, it is observed that the maximum peak at 260 Hz, the response when no fault is produced, shifted to higher or lower frequencies at all 3 axis, without consistency. The only consistent observation is that the maximum peak is at different frequency across the 3 axis when the overcurrent fault is produced, which allow us to deduce that fault is present.