Research On Short-Term Load Forecasting Based On Load Decomposition And Combination Model

Power system load forecasting is one of the important tasks to ensure the safe and economic operation of power grid,and short-term load forecasting is becoming increasingly important with the continuous development of electricity market.Meanwhile,the introduction of price mechanism has put forward higher requirements for short-term load forecasting.Therefore,it is necessary to study and improve the method of short-term load forecasting,thus to meet the requirements of safe and economic operation of power grid.The paper focuses on the forecasting of summer load,which has the most obvious fluctuation.A short-term load forecasting method based on load decomposition and combination model is proposed in this paper to forecast the summer load.First of all,based on the analysis of load characteristics,the method of load decomposition is adopted in the paper,it divides summer load into basic load and weather sensitive load.And the correlation analysis method is used to analyze the correlation between meteorological factors and weather sensitive load in summer of Xuzhou.Secondly,the basic principle of Support Vector Regression(SVR)model is introduced and applied to short-term load forecasting.In the forecasting process,the weather factors,day type and other factors are mapped and used as the input of the model.The grid model and Particle Swarm Optimization(PSO)algorithm are used to optimize the parameters,and the case analysis shows that the PSO-SVR model is much better than grid model in the application of short-term load forecasting.Then,the comparison with BP neural network algorithm lays the foundation for the application of PSO-SVR model in weather sensitive load forecasting.Furthermore,the short-term load forecasting method based on load decomposition and combination model is put forward in this paper.In this model,the grey theory is adopted to forecast the basic load,and the Fourier transform is introduced to correct the residuals.Then,the PSO-SVR model is used to forecast weather sensitive load,in addition,K-means clustering algorithm is introduced to select the input samples in the process of weather sensitive load forecasting,thus to strengthen the regularity of sample data.Finally,based on the historical load data and historical meteorological data in Xuzhou,the short-term load forecasting is carried out on working day and rest day using the forecasting model proposed in this paper.Through the case comparisons with the load forecasting method based on K-means clustering and PSO-SVR,the method based on load decomposition and PSO-SVR,it is proved that the model presented in this paper can get better forecasting results,it also verifies the feasibility and effectiveness of the proposed method.