As for the load forecasting on the Power Markets , the components of electric loads and their characteristics are analysed in detail. The loads of the JiLin Power Markets are calculated as an example . The models for the middle-term and long-term load forecastings are proposed . Based on these models , the yearly , monthly and daily load forecastings for the JiLin Power Markets are given by using the time-series method and the combined-forcasting method .For the middle-term load forecasting , the seasonal forecasting method is used because of the long forecasting duration and the periodical variation of loads . The yearly and the monthly load forecastings are obtained . And , for revising the loads , a progressive reasoning method is proposed .For the short-term load forcasting , according to the periodicity and the regularity of load variation ,the load cure is divided into three parts . Each part has its unique model by the combined-forecasting method , in which the combined weights are obtained by an ANN . An ANN based nonliner combined forecasting is used by making a comparison among the single , linear and nonlinear combined forecasting methods . The daily forecasting is gained.A software for yearly , monthly and daily load forecastings hasbeen tested in felds successfully and put into operation.
|