Power Quality Disturbance Identification Based on Statistical Measurement in Wavelet Domain

Power quality disturbance has become an important issue in power system over the past several years. Identify the disturbance effectively is an urgent problem needs to be solved. The paper proposes a power quality disturbance classification method based on wavelet transform and SVM. The statistical values including energy, mean, variation, skewness, kurtosis are extracted as feature vectors. Also a novel distance variance which calclulates the probability distance between the disturbance signal and standard signal is explored as complementary feature. The classifier uses SVM to train and test disturbance signals. Simulation experiment results shows the proposed method has an excellent performance of classification accuracy even in the presence of noise.