Power Quality Disturbance Classification Based on Wavelet Transform and Support Vector Machine

This paper presents an effective approach for classification of power quality (PQ) disturbances based on wavelet transform (WT) and support vector machine (SVM). Wavelet transform was applied to disturbance signal in order to obtain decomposition coefficients at six levels that represents signal in time and frequency domain. Eight statistical methods were used to extract features that characterize each disturbance signal. Forward sequential feature selection was then applied to the feature vector to identify the most discriminative features. SVM method was used for designing the classifier which is trained with the data simulated in MATLAB. High classification accuracy, reliability and robustness of the proposed classifier were confirmed on the testing data in noisy and noiseless environment.