The Research For Fault Detection And Diagnosis Of Induction Motors Based On Wavelets And Independent Component Analysis

Induction motors are widely used due to their operation convenience,simple structure and low cost.They have become an indispensable part of industrial production nowdays.The continuous and stable operation is a must determined by the process characteristics of induction motors.At the same time,in order to improve induction motors reliability and protect them from accident breakdown,it becomes more and more important to monitor the condition.Concerning the fault diagnosis detection technologies of induction motors,the key issues are to analyse and extract the fault signals,which are directly related to the veracity of fault diagnosis and reliability of trouble prediction.There are many methods to deal with fault signals,such as neural network,fuzzy neural network,wavelet transform and so on.In this paper,a wavelet packet transform-based method combined with Independent Component Analysis(ICA) is proposed for fault detection and diagnosis of the induction motors.An acceptable signal to noise ratio(SNR) is very important to analyse the signal effectively.Wavelets are used to denoise the current signals of the induction motors.The signal denoising method by using the wavelet threshold restores the current signals to higher SNR,and then the denoised signals are analyzed by the method of Independent Component Analysis(ICA).The experiment results demonstrate that the proposed method can be well applicable for detecting fault or potential fault of induction motors.