Medium Term Power Load Forecasting Of Regional Distribution Network Based On BP Neural Network

Power load forecasting is one of the most important work of power system workers,which plays an important role in the regional power system.Accurate load forecasting can be more reasonable to arrange the operation state of the grid generation unit in the target area.On the basis of realizing the safe and stable operation of the power grid,reducing the waste of electric energy and reducing the consumption of the total cost.Regional distribution network of the mid-term electric load forecasting the main significance is to provide a relative reasonable reference value,reach the optimal allocation of limited resources in each cycle of investment,electric power planning department daily work for the contact work is an important part of the power grid planning department of distribution network resource allocation and the construction direction.After previous years of power load forecast research and the practical forecast method of that there are a lot of,traditional methods are mainly regression analysis method,the unit consumption method,and the modern intelligent method of grey system theory,expert system and artificial neural network,under different conditions have their respective advantages and disadvantages and has been developing continuously.The regional power system load forecast has accuracy,time characteristics,the trend of electricity data due to a series of factors from the local economic,environmental and climate impact,showing a nonlinear.BP neural network for nonlinear and fuzzy internal rules of data processing has a good adaptability.This paper mainly discusses the BP neural network application in the middle term load forecasting,the research of distribution network area instance object for the area of the old city in Qingyuan City,through the first three months of electricity consumption to predict the next month the whole society electricity consumption,through six months of actual consumption data and predicted values for comparison and verification.In this paper,the MATLAB software is used to design BP neural network and the simulation calculation is carried out.From 2009 to the first quarter of 2016,Qingyuan urban area as the basis of the actual electricity consumption data database,the use of momentum BP algorithm to predict and compare with the actual value of the error.The practical calculation results show that the accuracy of the BP neural network in the medium term power load forecasting of the regional distribution network is generally satisfied with the application requirements.Then,combined with the classification and LM algorithm is used to optimize the BP neural network to further improve the precision and error in practice from the point of view,although the monthly power load forecasting data and actual value still has certain deviation,but from the average number of six months still reaches a higher accuracy,overall in the region is with practical application value.