A Study Of Load Forecasting Taking Account Of Local Characteristics

According to the importance and need of load forecasting, the paper studies on short-term load forecasting and its model by analyzing the local load of TieLing area. For the short-term load forecasting, according to regularity of load variation, the load curve is divided into three parts. Each part owns its unique model by the combined-forecasting method, in which the combined weights are obtained by an ANN. An ANN based nonlinear combined forecasting is used by making a comparison among the single. linear and nonlinear combined forecasting methods . The daily forecasting is gained. Meanwhile this paper also studies on load forecasting error caused by small-scale hydroelectric plants and thermal plants in regional network. After analyzing on composition and characteristics, the model of daily load forecasting is revised and the old software is developed. The short-term load forecasting is achieved by time series analysis and combined forecasting. It can be concluded that the accuracy of advanced load forecasting has enhanced to some degree compared with the old software by testing on scene data. A software for daily load forecasting has been tested in fields successfiully and put into operation.