Prediction of EV Charging Load Combination Based on Grey Correlation Analysis

Abstract:
The analysis and prediction of EV charging load state are directly related to subsequent charging planning and power consumption management, so it is of great significance to carry out EV load analysis and research. Firstly, grey correlation analysis was used to calculate the similarity between the predicted date and the historical date. Thus, the best similarity date was obtained by orderly arrangement, and the multi-correlation date scene set was established. Then, based on the similar day data, a prediction model was established using a support vector machine algorithm. Considering the prediction deficiency of the single model, a prediction model based on the time series algorithm was established to predict EV charging load according to the data of the first few days of the predicted day. The combined prediction method based on weight is proposed to comprehensively consider the complementary advantages of the two prediction models to improve the accuracy and stability of load prediction. Finally, the actual load of electric vehicles in Chongqing is predicted to verify the correctness of the proposed method.