Power Load Forecast Method Research

Power load forecasting is a main work of the power sector,an accurate power load forecasting of power system and the operation of the production planning has a great influence,is beneficial to promote the safe,stable and economic operation of power system,is necessary to ensure the power system safe and economic operation of one of the means.This paper first introduces the background and research significance of load forecast,as well as the meaning and classification of the power load forecasting,in different periods with different types of load forecasting of practical significance and research methods,including short-term power load forecasting are the analysis of power system security and the important basis of power dispatching work.This paper introduces the characteristic of load forecast,and influence the composition of the load forecast accuracy,and through the actual data curve vividly the periodic and trend and holidays load characteristics and influence factors such as temperature,humidity were analyzed aiming at all kinds of main methods of forecasting power load analysis,and introduces the principle of all kinds of forecasting methods,the merits and demerits of each method and the applicable conditions are compared.Then the key to the traditional method of time series forecasting method and modern forecasting method of artificial neural network to the basic ideas and prediction process were studied.Last March 2012 in a certain area of power system load,for example,using time series method of exponential smoothing method and the stationary time series method and BP neural network method to establish mathematical model for prediction of short-term projections for the load data of a week,and make comparative analysis,the prediction results obtained considering the effect of the meteorological factors model for improving the accuracy of the short-term forecasting has very important role,has a certain practical significance.