Charging load forecasting of electric vehicle charging station based on support vector regression

In allusion to the problem that electric vehicle(EV) charging time and state of charge(SOC) randomness leads to the traditional application of EV charging load characteristic forecasting method low accuracy problem, applying support vector regression(SVR), a charging load forecasting model based on historical load is proposed. The proposed model considers various kinds of factors which could influence the load, including the historical data of charging load, the number of EVs, the number of normal working charging pile, weather information, week properties, holiday properties and other information, in addition, the model corrects the false data before the establishment of the training sample set, which effectively improves the precision of forecasting. The effectiveness and correctness are validated by numerical example of an EV charging and switching station.