Load Forecasting Basing On Models With Multi-Indices

Electric power, as an important kind of second energy, palys a key role in economic growth and society development. For healthy economic growth, the electric power supply is a necessary factor. Therfore, it is significant to make study on load forecasting to ensure enough power supply. In the study of load forecasting models, forecasting model with multi-indices is a main subject to put focus on the relationship between load and indices. However, there is few studty on that kind fof model in the dometic litratures which is lacked of the modeling strategy and the methods. Thus, in this paper, we proposed three kinds of load forecasting models with multi-indices for the short term load forecasting and mid-term load forecasting within Chongqing, and the long term load forecasting with China. With the study in literatures and the uniquty of each case, soft comupting is introduced for every model.Firstly, a review is given for the study on short term load forecasting (STLF) with multi-indices. Until now, there is little study on the complete effect of weather indices which are important for STLF. Therefore, with study in the literatures we proposed a modle with multi-weather indices for STLF in Chongiqng. And a hybrid of soft computing methods like rough set and articifial neural network are introduced to make analysis of the relationship between load and indices. The case study shows the improvement of accuracy of forecasting results, and the comparision shows the effectiveness of proposed model and hybrid method.Secondly, for mid-term load forecasting (MTLF), we proposed a model with economy indices and uncertainty indices, because they are important in MTLF while there is little study on them in literatures. Especialy, there is no study on the effect of them to MTLF by considering both of them. The modeling strategy is from the literatures and the unquity of case of Chongqing from 2001-2007. With rough set, the importance of each index to load is obtained, and then the forecasting is performed. Finally, the case study indicates the effectiveness of proposed model and hybrid method.Finally, with the review on literatures of long term load forecasting (LTLF), we proposed a model with an economcy index -GDP- for the lack of study on LTLF. According to the literatures, we discussed the feasibility of modeling strategy, and we analysize the relationship between GDP and nanual load output, importand export. With the analysis, the forecasting is performed accordting to the expected GDP in the 11th five-year plan of China. Basing on the forecasting results, we give out recommendations for the effect of load to economic growth as well as the discussion of feasibily of them.