Short-term Load Forecasting Based On BP Neural Network In Power System

Load forecasting of power system is one of the most important contents to realize power system materialization management, and short-term load forecasting is the most widely used method in power system load forecasting. However, the traditional power load forecasting method is not suitable for forecasting load in some areas, and some methods can not reflect the future load trend accurately, as well as the load forecasting procedures is complicated which can not really realize computerized.Analyzing classification and characteristics of the power load, and summarizing factors which impact the load forecasting, BP neural network algorithm is improved by increasing momentum term and adding steepness factor in activation function. The problems existed in training BP neural network, such as the converging slowly, falling into paralysis and local minimum easily are solved by this approach. Then Linyi power system load is forecasted by using modified BP neural network algorithm.The concept of human comfort index is introduced, which transforms humidity, temperature, wind and some other factors into indicator value when building the BP network load forecasting model. This approach can reflect influences of several meteorological factors on the load curve properly and avoid overstaffing network caused by nodes. In addition, the network prediction accuracy can be improved considering meteorological factors.A large number of samples which cover load value of96points in24-hour as much as possible are used to train the network and the trained network weights are saved which can enhance the versatility of the trained network, and allow users easy access recently.Finally, the short-term load forecasting system platform the power grid dispatching automation system iES-600and iPAS network analysis software developed by Integrated Electronic Systems Lab Co., Ltd are analyzed. And the process of BP neural network load forecasting is programmed by VC++6.0, then flow charts are painted and illustrated,which have reflected the computerized prediction method.