Research And Application Of Mid-long Term Power Load Forecasting

Mid-long term load forecasting of power system is an important content in the researchof power system planning and running, it is the premise to ensure reliable and economicoperation of the power system, basis and also the power grid planning and construction. Withthe development of the national economy under the influence of the electric power industry ismore and more big, the correct forecasting of power load becomes more and more important,so the rationality greatly power industry investment, network layout and operation depends onthe precision of load forecast.Load forecasting is affected by many uncertain factors in different degree, differentmodels have their own applicable scope, it is difficult to apply in all cases. So far, also cannotfind a prediction method can obtain satisfactory predictive results in any case, so in themedium and long term load forecasting, to learn how to combine the local actual situation,choose a different prediction method, the prediction results are compared to each other andcheck, determine the final prediction value.Forecasting of power load forecasting total belong to strategic planning, it is the overallload area as the research object, to provide basic data for network planning, load forecastingresult has very important significance on the city power grid planning, but also provide theimportant basis for determining the power points of the city.This paper studied a variety of different medium and long-term load forecasting theoryand its application, and briefly introduces the purpose and significance of long term loadforecasting of power system, the development status and Prospect of domestic and abroad,and then systematically expounded the theory of forecasting methods, and combining withpractical situation of medium and long-term load forecasting, deeply analysis of the specificimplementation methods for different prediction theory is applied in the power system,through the simulation experiment on the medium and long term load of different predictionmethods, including single model, the improved model, a comprehensive model and combinedmodel, the prediction results of comparative analysis of various methods for forecastcombination, or using several methods, or for overall segmentation part of each part of the model, application of appropriate, through the combined application of series, the forecastingaccuracy of maximum, minimum error.