Identification of Power Quality Disturbance Sources using Gradient Boosting Decision Tree

This paper proposed a new method based on statistical feature extraction and gradient boosting decision tree (GBDT) to recognize the power quality disturbance sources. Statistical calculation is adopted to extract the features of power quality disturbance sources, which has the advantage of small calculation. GBDT is proposed to apply in the recognition of power quality disturbance sources. First, according to the inherent characteristics of high-speed railway, ordinary railway, wind farm and photovoltaic plant, the proposed method uses statistical calculations to extract features which are the input of GBDT. Then, GBDT is applied to classify the power quality disturbance sources. Experiment results show that the proposed method can classify power quality disturbance sources accurately. Compared with other classification methods, GBDT has better recognition performance.