SSA-SVR-Based Prediction Model of Charging Load for Electric Vehicles

Abstract

The reliable and secure operation of power grids can be efficiently supported by the charging load prediction of electric vehicles (EVs). To address the problem of insufficient accuracy of existing charging load prediction models, a technique for predicting charging load for EVs using the sparrow search algorithm-support vector regression (SSA-SVR) is proposed. First, the daily travel patterns of space and time of EV users are analyzed. Therefore, EV charging load data is obtained by Monte Carlo simulation. Finally, a support vector regression (SVR)-based model for predicting EV charging load is established and the sparrow search algorithm (SSA) is further used to find the optimal kernel function factor and penalty factor of SVR to achieve the optimized prediction effect. The simulation experiments show that, compared with the backpropagation (BP) neural network, SVR methods and PSO-SVR methods, the proposed prediction model can enhance the prediction accuracy of the charging load of EVs.