The Short Term Power Load Forecasting Based On County Dispatching System

County power system load forecasting is a important part of the construction of electricity transmit and transforming, it is the basis of ensuring grid safely and economicly run and realizing grid scientific management and dispatching,it is not only the part of energy management system, but also is the contents which grid run commercially need for in the future.This paper introduced electric power dispatching automation system developped by the research team, and then summarized these present methods in county power system load forecasting, analyzed county power system load's feature, and designed the whole plan .This plan including: data preprocessing, feature parameter extraction, load forecasting,and so on.In the data preprocessing, through methods such as mathematical statistics deviate from the calculation of the rate of load capacity to carry out abnormal data pretreatment, and then, the raw data sequence become reasonable. In the process of feature parameter extraction ,using wavelet analysis method has the advantage of amplification of a part of function of neural networks to extract the characteristic parameters. In the load forecasting algorithm,BP(back propagation) neural network which is the most mature algorithm at present was selected.On this basis, I researched various algorithms, and analysised emphatically the problem building neural network model,given the modeling method is more applicable to construct a three-tier BP Neural network, determining parameters used in BP network model. Through analysising the shortcomings of the basic BP neural network, its improvement algorithm called OBP(optimization back propagation) was proposed.Finally, examples of the simulation experiment to compare results of two plans come to the conclusion: In this paper,the short-term load forecasting method based on neural network is feasible, and also is reliable.