Research on Daily Load-forecasting for Power Grid Based on ANN
|Keywords||power system short-term load forecasting BP neural network GNNM(1,1)|
Load forecasting is the basis of power system safety and economical operation. Improving safety and economy of power system operation, and power quality, relies on accurate load forecasting in power system. So, accuracy of dispatch department short-term load forecasting is paid more attention to. In order to improve accuracy of short-term load forecasting, reduce work load of dispatcher, and get rid of personal experience dependence, it is significant of exploring a way of more scientific and programmed load forecasting, and making research on model and method of Wenzhou power system load forecasting.Research on theory of short-term load forecasting enriches continuously after many years development. Artificial neural network is adopted as the main research method for load forecasting in this paper, based on analysis about advantages and disadvantages of different forecasting theory, and actual condition of Wenzhou power system.Influential factors of load are always the key for short-term load forecasting. Influential factors of load are deeply researched and model is built with artificial neural network, according to composition and variation characteristics of Wenzhou power system load. For workday and weekend, optimization BP neural network model is used according to day type. Both meteorological factors direct input and temperature humidity input are adopted for comparative analysis. For special holidays, grey forecasting method is used for building grey neural network model, with advantages of less load data, no considering load distribution and load variation trend, high short-term forecasting accuracy, and easy operation and test. Accuracy and speed of above forecasting model is proved, through analysis of example forecasting.There are many hydropower plants in Wenzhou, and these plants play important roles in load adjustment of power grid. The output power of local generator set is needed to forecast in order to provide accurate plan of purchasing energy from network in dispatching operation. To improve generation benefit of hydropower plants and carry out the spirit of impartiality and openness of power dispatch, the evaluation method for generating capacity and willingness of hydropower plants is presented in this paper, which is proved feasibility through examples.