The Research And Application Of The Annual Peak Load Probabilistic Analysis And Forecasting

Annual peak load is affected by multifarious factors, specially by local circumstance. At present, the methods of annual peak load forecasting are simple and low precision, and furthermore, none of these methods is considering of local circumstance. This thesis presents a probability method to forecast annual peak load based on the research of the divided load model, and a practical system has been successfully formed according to the actual circumstance of Henan province.Firstly, those factors affecting on annual peak load are divided into two part including tendency and randomicity, based on which, the divided load model is improved. At the same time, this thesis present the nature random part of power load, in order to study the random part of power load more accurately.Secondly, based on the characteristic datum extracted from the datum of daily peak load, the probability model of the nature random part of power load is established; the grey GM(1,1) model is improved to forecast the basis part of power load; after the relation model is established on the basis of the researching the relationship between the climate part of power load and climate factors, the probability model of the climate part of power load is established combined with the tentative probability model of temperature.Finally, the three parts of annual peak load are united to make a uniform probability method. From the results of simulation, it is found that the new method proposed in this thesis is efficient. A practical software has been formed, whose functions and usage are expatiated in the fourth part of this thesis.