The Classification of Multiple Power Quality Disturbances Based on Dynamic Event Tree and Support Vector Machine

Abstract

It’s difficult to classify multiple disturbances with the single disturbance classification method. This paper describes multiple disturbances and designs the classifier of multiple power quality (PQ) disturbances; the process of power quality disturbance classification can be divided into two-stage, feature extraction, and classification. This paper extracts features of disturbances with dq Transform, Wavelet Packet Transform (WPT) and S-Transform (ST), and combines them to reflect the characteristics of disturbances better. The design of binary tree Support Vector Machine (BT-SVM) with the concept of the class distance of the clustering analysis makes classifications intelligently. And dynamic event tree is proposed to make classifications of multiple disturbances. By these methods, disturbances can be classified fast and accurately. The results of simulation show that the classification method in this paper is able to classify multiple disturbances effectively.