Study on orderly charging strategy of EV with load forecasting

The development and popularization of electric vehicles (EVs) is of great significance to environmental protection, energy saving and emission reduction. With the wide popularization of EV, the EV's disorderly charging brings the security hidden trouble to the grid. Firstly, according to the safe operation of power grid and the charging requirements of EVs, an optimal scheduling model based on grid loss is established, then, the optimal scheduling model is transformed by second-order cone relaxation technology. Secondly, because the orderly charging schedule of EV is based on accurate charging load forecasting, this paper based on LSTM-XGBoost dynamic combination forecasting, the dynamic combination model of LSTM and XGBoost is optimized by using Bayesian optimization method, and more accurate charging load forecasting results are obtained. Finally, the accuracy of the prediction method and the effectiveness of the optimal scheduling strategy are verified by the charging data of the EV in the actual area.

Automated summary

The development and popularization of electric vehicles (EVs) are crucial for environmental protection, energy saving, and emission reduction. However, the disorderly charging of EVs poses security concerns to the grid. In this study, an optimal scheduling model based on grid loss is proposed and transformed using second-order cone relaxation technology. Additionally, the combination of LSTM and XGBoost is optimized for accurate charging load forecasting. The effectiveness of the optimal scheduling strategy is verified using actual EV charging data in the area.