Classification of power quality events using support vector machine and S-Transform

Abstract

Classification of power quality events (PQE) to enhance the power quality is a vital problem in end users. In this article a novel method to classify PQE with random white noise of zero mean based on wavelet energy change and Support Vector Machine (SVM) is presented. Here PQE waveforms are disintegrated into 10 layers by db4-wavelet with multi-resolution. Energy Changes (EC) of every level between PQE waveforms and standard voltage waveforms is drawn out as eigenvectors. Principal Component Analysis (PCA) is implemented to decrease the dimensions of eigenvectors and gives the main structure of the matrix, which creates new feature vectors and these vectors separated into two sets, namely training set and testing set. The method of cross-validation is adopted for the training set to identify the optimum parameters adaptively and build the training model also the testing set is replaced into the training model for testing. In conclusion the suggested method accuracy is compared with S-Transform (ST) based PQE classification to prove the accuracy of classification. The classification accuracy of SVM is great and having strong ability to resist noise, speedy classification of PQE.