Induction motor broken bar fault detection based on MCSA, MSCSA and PCA: A comparative study

The early failure detection of induction motors is considered very important in order to ensure their stability and high performance. Thus, condition monitoring of these motors are essential to ensure that. However, the effectiveness of the fault diagnosis depends on the quality of the fault features selection that is used in the adopted method. In this way, this paper will present a comparative study of three methods based on the electrical signal analysis. The first one is the most known method, motor current signature analysis (MCSA). Another method based on the analysis of spectral current, is the motor square current signature analysis (MSCSA), will also be used. Finally a third method based on principal component analysis (PCA) is also used. The performance of the methods is tested under different fault and load conditions. Experimental results are presented in order to compare the performance between the three methods.