Fault detection and diagnosis in induction motors using motor current signature analysis.

Abstract

The study assessed the effect of internal induction motor faults on motor current signature. The internal faults under study included bearing faults, stator faults and rotor faults which together constitute about 90 % of all induction motor internal faults. The case study was done in Hima Cement plant Kasese district in Uganda. The study objectives were to model and simulate the identification and diagnosis of stator, rotor and bearing faults in induction motors using the MCSA technique, develop discrete wavelet transforms for motor current signal pre-processing and develop artificial neural networks for motor fault classification. The study employed experimental research design through triangulation using both qualitative and quantitative approaches. Data sample representing 168 motors was obtained and the motors were selected using simple random sampling technique and the tool used to measure and record data was the Power Quality Analyzer model PQA 824. Data was exported to matlab where a program of DWT was written to pre-process the data and get out coefficients which are a representation of signal change from time domain to a more revealing frequency domain. Still using matlab, a neural network was designed and trained on this data so as to aid in fault classification. Finally new data sets were applied to the neural network and it was able to classify the data properly. With the results obtained, the study concludes that bearing faults, stator faults and rotor faults can be detected and classified during the course of motor operation before catastrophic failures happen. The study recommends that stakeholders especially plant owners should focus at adopting these new technologies that ensure faults are detected early enough so as to aid continuous performance. It is also recommended that operators need to cooperate with the maintenance team fully so that whichever state that appears on the display is reported the maintenance team. This helps them in planning their work.