Various and multilevel of coiflet discrete wavelet transform and quadratic discriminant analysis for classification misalignment on three phase induction motor

Abstract

For a long time ago, induction motor has been used in various industry due to strong construction, high efficiency, and cheap maintenance. Induction motor needs to be maintained regularly so that it can operate for a long time. Based on studied, bearing faults result in failure of 42% -50% of all motor failures. One of the causes of bearing failure is misalignment when installing induction motor. This research proposes classification misalignment in induction using coiflet discrete wavelet transform and Quadratic Discriminant Analysis. Simulation of motor condition is introduced in this research as normal operation and two misalignment variations. And then, various type of coiflet discrete wavelet transform in first level until third level is used to extract motor vibration signal into high frequency signal. Then, three types of signal extraction, namely sum, range and energy level, will be used for input to Linear and Quadratic Discriminant Analysis. Linear and Quadratic Discriminant Analysis will analysis signal extraction and classify them into normal operation and two misalignment condition. The results show that first level of coiflet discrete wavelet transform is the best level for classification misalignment on induction motor, both using the Linear Discriminant Analysis and Quadratic Discriminant Analysis methods. The accuracy obtained from the two methods is the same or almost the same.