基于小波变换和二叉树SVM的电能质量扰动分类

lassification of power quality (PQ) disturbances has two key processes: feature extraction and classifier design. A new method based on wavelet transform and Support Vector Machine with Binary Tree Architecture (SVM-BTA) is presented for power quality disturbances classification. At first, by use of alternative transients program (ATP) eight typical and two complex disturbance signals are generated as sample set. Then wavelet transform is applied to obtain multiple features, including the number of modulo maximum value, signal energy and the standard deviation of wavelet coefficients in special frequency band. This can reduce the data size and reflect the local feature of disturbance signals better. At last, the SVM-BTA multi-classifier is trained and tested by the sample set. The test results show that the proposed method can effectively classify ten disturbance signals with high classification accuracy and short training time.