Mid-long Term Power Load Forecasting Based On Grey Theory

Load forecasting is an important content of power system planning and also the basis ofpower safe and economical running. Mid-long term power load forecasting takes 5 or 10years as an unit for forecasting and is used to work out a plan for expanding power system,provide reliable reference for confirming the future power supplying sites, power constructionscale, power industries layout, and for balancing power grid funds and human resources in thelocal area. Because the mid-long term power load forecasting is affected by many uncertainfactors, up to now, no one model can obtain the satisfying forecasting results under differentconditions of time and areas. It is necessary to analyze the local load change, think about thepractical situation, and to choose proper method.This thesis firstly introduces the importance of power load forecasting, gives a briefanalysis on characteristics of different power load and makes a summary of power loadforecasting. After thorough research into grey load forecasting mechanism, the localization ofgrey forecasting model is found and the improved method is derived. Correction to GM(1,1)with absolute error grey model is presented. After pre-treatment on historical data, theapplicable range and precision of grey forecasting model is enhanced. Equally dimensionalnew information grey model forecasting method not only overcomes the shortcoming that themathematical model is changeless in simple grey forecasting method, but also makes use ofits advantage of high precision in short term load forecasting. So it satisfies the request for themid-long term load forecasting. Period equal division series method changes parameters indifferentiate equation into time variables, so the shortcoming of constant parameters inGM(1,1) is removed, therefore, GM(1,1) is more applicable for the mid-long term loadforecasting. The method is used for forecasting power load of Anshan area, the results showsthat the improved GM(1,1) method is valid.