Study On Short-term Load Forecasting Method Of Power System

Short-term load forecasting of power system is an important work in management parts of power system planning,power utilization and power scheduling,furthermore,short-term load forecasting of power system has important guidance function in fields of controlling the amount of power generation and arranging the unit maintenance,so,improving the forecasting accuracy of short-term load forecasting of power system has important influence in aspects of safe and stable operation of power system,security economic dispatch and power system planning.Based on the above mentioned,study on short-term load forecasting of power system is of great significance.In this paper,the method of short-term load of power system was studied,and the method of improving the accuracy of forecasting was discussed.Firstly,the characteristics and main influencing factors of short-term load were studied,and the conclusion that short-term load has the characteristics of periodicity,continuity and volatility was obtained through the analysis of short-term load characteristics;The conclusion that short-term load is mainly affected by time quantumt and meteorological factor(holidays,emperature and rainfall)was obtained by analyzing the relationship between short-term load and different factors.Secondly,the historical load data were processed by wavelet de-noising method,but the soft and hard threshold method exists some shortage,in order to overcome these defects,the weighted average method was used to estimate the wavelet coefficients.Then the optimal wavelet basis function and the decomposition layers were found out through trial and error,thus the original load data was repaired.Again,the number of factors that affecting short-term load was reduced through the principal component analysis and the gray correlation degree method,and a set of factors that affect the degree of higher was extracted.Through principal component analysis,on the one hand,the amount of data can be reduced and redundant information can be eliminated under the premise of retaining almost information.On the other hand,the weight coefficient of differentindex can be obtained through principal component analysis method.Thus,the weight coefficient of each factor was more scientific when calculating the relevancy in the application of gray correlation degree method,avoiding the artificial defects to determine weight coefficient.Finally,the short-term load was predicted based on support vector machine with the optimization of biogeography.Three parameters of the support vector machine: the penalty factor C,the insensitive loss function ? and the parameters of the radial basis function ? were optimized by using the biogeography optimization algorithm,and the predict results that optimized by using the biogeography optimization algorithm were compared with the predict results that before optimized,which fully reflected the good results obtained by the biogeography optimization algorithm.