Classification method of dynamic power quality disturbances based on SVM

The SVM(Support Vector Machine)method is introduced to classification of power qua-lity disturbances.The concerned disturbances,including voltage sags,swells,interruptions,switching transients,flickers and harmonics,are modeled with Matlab.The features of sample waves are extracted by Fourier transform and wavelet transform to form the training and testing samples.The steps of disturbance classification using LIBSVM are described.The influence factors are analyzed according to the classification result.The trained support vector classifier is tested and validated effective.When using C-SVC and RBF kernel,parameters can be adjusted to achieve the optimal effect,while the maximal classification ratio reaches 96.67 %.