A comparative study dedicated to rotor failure detection in induction motors using MCSA, DWT, and EMD techniques

This paper presents the detailed detection of broken rotor bar faults in squirrel cage induction motors (SCIMs). This study used three diagnostic techniques. Early faults detection allows us to avoid catastrophic damage. In this work, we have exploited the stator current signal by three recent methods. The first technique uses the fast Fourier transform (FFT) which is often called motor current signature analysis (MCSA or MCSA-FFT). According to this technique, we carefully checked the spectral content of the stator current. In addition, we have clearly noticed to new harmonics that indicate the broken rotor bar (BRB) faults exists. The second technique is based on the discrete wavelet transform (DWT); this technique is widely used in the diagnosis field of rotating machinery. In order to detect the BRB faults in SCIMs, we have exploited this method by three important indicators. One of them is based on the mean square error (MSE) of each detail coefficient. In this study, we applied a new indicator (MSE) for the BRB fault detection based on DWT. The last method is empirical mode decomposition (EMD) to perform the current signature analysis in order to decompose the motor current signal into intrinsic mode functions (IMFs). It is currently competing with several methods such as: MCSA, DWT, etc. So, it is possible to detect BRBs through the evaluation of the different IMF levels for both conditions, healthy and faulty state of SCIM. An experimental test for different conditions: at no load or at load operation, healthy or faulty state of the induction motor has been performed. So, experimental results using three methods showed a detailed comparison between them in order to achieve a judicious decision on the broken rotor bars detection. Finally, we have confirmed the proposals ideas in this subject in order to detect the broken rotor bar faults