Determining motor's quality types using principal component analysis on current waveforms

Abstract

This study proposes a Principal Component Analysis (PCA) method to analyze motor's current waveforms for determining the motor's quality types. The proposed method which consists of data training algorithm and motor's quality types decision algorithm. In the data training algorithm, the input signals are selected from the sample motors with known motor's quality type. It carries out three major processing stages: (i) the preprocessing stage for enlarging motor's current waveforms' amplitude and eliminating noises; (ii) the qualitative features stage for qualitative feature selection of a motor's current waveform; and (iii) the PCA procedure to obtain the projected space and projected coefficient. This projected space and projected coefficient are then used for motor's quality type decision. In the motor's quality type decision algorithm, the input signals are selected from the test motor with unknown quality type. In the experiment, the classified results are 92.80%, 96.83%, 99.91%, and 99.80% for statistical indices Se, PPV, Sp, and NPV. The total classification accuracy was approximately 99.73%.