The Research Of Short-term Power Load Forecasting Model Based On Time Series

Short-term power load forecasting is an important task of electric power dispatching system, accurate effectively predict can provide scientific basis to the study of power system programming problem, electric system economic operation and auto dispatch.Power load is a changing sequence with time, therefore time series prediction is an important method of power load forecast. However, it wasn't sensitive to only think the factor of time, not think weather conditions, Special holidays.so the effect of predict is not ideal. At the same time series forecasting demands sample series is a stationary sequence, and power load sequence is often not smooth, thereby establishing a time-series model is difficult.This paper studies the principe of the common method, application range, and their advantages and disadvantages of the short-term power load forecasting, proposing a time prodict method based on attributes, based on analysing large amount of historical power load. series prediction method based on the attribute classification. include the following steps:collect power load historical data,the influencing factors based on historical load data analying electric loading, which define properties classification rules, Collect the attribute of forecasting task and decompose forecasting task, According to the characteristics of forecasting task screening prediction sample, Analysing correlation of sample, Identify times sequence model. estimate model parameters, establish Scientific Prediction model; calculate results with model;analyze the predict result.power load forecasting program based on properties classification think the time,weather,holiday. properties classification,task decomposition,sample screening mechanism provide smooth time sequence, make up the defect of the power load time sequence forecasting. Experimental results prove that time sequence prodict program based on properties classification is a efficient and ideal short power load forecasting program.