Short-Term Load Forecasting In Shenyang Based On Artificial Neural Network

Short-Term Load Forecasting (STLF) is one of the most important contents of running and dispatching power system. It is a very important aspect of power system to ensure operating safely economically and achieve scientific management in the power system. And it is one part of energy management system (EMS) as well as a necessary content of the electricity market place operation management.In this paper, the constitutes and characteristics of the electrical load are analysed, and the difference of several present STLF methods is introduced. At the same time, the present STLF method and it's short points in ShenYang are pointed out. By studying the data and characteristics of the load in ShenYang, analyzing the factors which effect the precision of the load forecasting, and making the use of the relationship of load and temperature, a BP network is constructed. In data processing,this paper checks and modifies the bad-data by method of cluster analysis. Besides, a decision tree classifier(CART) is used to eliminate the unnecessary variables, and reduces the number of parameters to be adjusted. The load data are divided by three day-types,and at the same time, the neural networks models for weekday,weekend and holiday establish independently. In order to enhance the training speed and the precision of neural networks,the daily load curve is divided and proper input variables are chose in each model. Besides, the divided-line optimizing method is applied to make the neural networks converge more quickly. The network was trained with historical electric power load data,and the result of the short-term electric power load forecast is successful. And the precision has risen to a higher level compared with the result forecasted by present methods in ShenYang,which proves the validity of the method in this paper.At the same time, a load forecasting program has been made for Shen Yang. The function of this program include: the data processing, the choice of input variables, the treating of the ANN model, the renewing of the model, load forecasting and so on. It's made by C language, and adapts the fact of Shen Yang.