Short-Term Load Forecasting On Qingpu Country

Short-term power load forecasting (STLF) is an important component in the daily safe and economic operation of the electric utility. It is important to find out effective forecasting arithmetic to decrease the dependence of artificial experience and increase forecasting precision. Electric load has both regularity and randomicity. Load of next period has close relation to historical load, current operation status, meteorologic factor of forecasting period and date type, in which there are a lot of linear and non-linear relations.The peculiarities of the load supplied by the Qingpu Power Supply Company are studied in details. This paper prefer a back propagation(BP) algorithm of Artificial Neural Network(ANN), which is adept in fitting non-linear mapping. The algorithm trains the network through Levenberg-Marquardt(L-M) rule, it can fit initial data well and bring less error between forecasting output and actual value. So it can satisfy the requirement of forecasting precision. The computation result of short-term load forecasting on Qingpu country shows that this model has perfect precision and adaptability.