High-precision Prediction Method of Electric Vehicle Trading Power based on Neural Network

Introduction: Electric vehicles have become a development trend due to their good environmental protection and energy saving, etc. Prediction of electric vehicle charging volume can help relevant departments optimize power supply, service, and construction. Methods: In this paper, the Support Vector Machine (SVM) model and the combined Long Short Term Memory (LSTM) and Support Vector Regression (SVR) prediction model are constructed for the prediction of charging capacity, and simulated by actual trading power data in Hubei Province. Results: The results show that the prediction effect of the two methods is good, and the LSTMSVR algorithm is judged to have better performance and less error in predicting the fluctuation of transaction power. Conclusion: LSTM-SVR can be used as a charging prediction method to provide a reference basis for the power control strategy of electric vehicles charging management platform, which is conducive to the healthy development of electric vehicles industry.