Application of LS-SVM in Classification of Power Quality Disturbances

A new method classifying power quality disturbances based on least square support vector machine(LS-SVM) and wavelet packet decomposition is presented. Normal voltage and several power quality disturbances(voltage swell,voltage sag,voltage interruption,transient disturbance,transient oscillation,harmonics and voltage flicker) are decomposed by wavelet packet,and the standard deviation of the wavelet packet coefficients of each end node are extracted as eigenvectors. Adaptive optimizing algorithm is used to optimize LS-SVM. The disturbances are classified using LS-SVM based on optimized parameters and minimum output coding. Compared with the classification of BP neural network,the method is able to overcome the shortage of long training time,easy to fall in local minimum and has more fast training speed and higher classification accuracy percentage. It also performs well when the samples are fewer. The simulation verifies its validity to classify power quality disturbances.