Study of Short-term Load Forecasting Based on PSO-SVR

Abstract

Support vector machine (SVM) algorithm was used to predict the energy load of an energy station in Shanghai, including hourly and half-day load two parts. The input parameters were the outdoor temperature, the air relative humidity, the temperature difference between the supply and return water and the flow of the supply water at the moment, adding the load at the last moment. The output parameter was the load at the moment. The results showed that the predicted value was in good agreement with the measured value. The average absolute errors of all results were less than 1.5, and the average relative errors were less than 3.