Combination Of Quadratic Discriminant Analysis And Daubechis Wavelet For Classification Level Of Misalignment On Induction Motor

Nowadays induction motor is widely used in industrial process because of strong construction, inexpensive maintenance, and high efficiency at rated speed and torque. Various environmental conditions make damage to the induction motor step by step. Misalignment on induction motor is one of the damage that often occurs. This paper is focused to simulate various of Daubechis discrete wavelet transform for classification level of misalignment on induction motor. Level of misalignment in this paper is introduced as normal operation, lmm of misalignment of motor and 1. 5mm of misalignment of motor. For this task, Daubechis 1-5 in first level until third level is used to extract vibration signal of induction motor. Then, feature extraction of high frequency signal, like as sum of signal, energy of signal and range of signal, is calculated. And then, quadratic discriminant analysis is used to classify level of motor misalignment. As comparison linear discriminant analysis will also be used in the classification process. The results show that Daubechis 3 at first level, Daubechis 4 at first and second level, and Daubechis 5 at first level are able to perfectly classification level of misalignment on induction motor. The quadratic discriminant analysis is more accurate than linear discriminant analysis which is only able to classify perfectly misalignment with Daubechis 3 at first level.