Study On The Method Of Short-Term Load Forecasting Of Grid

Short-term load forecasting is basis of power system for safe and economical operation. Power and load should keep balance automatically. Because power can not be stored economically in large-scale, so the load variation lives in an active position and the power variation of generations live in a passive position, therefore, power of generation should be changed basing on the load variation. In order to get the arrangements of generations and operation mode of grid, the research of short-term load forecasting theory and algorithm have become an important work of dispatch center of power company.As grid covers a larger geographic area, short-term load forecasting of grid should taking into account the characteristics of its own meteorological factors. Meteorological characteristics of grid can not be described by a series of meteorological data. This paper proposes a method of short-term load forecasting basing on the characteristics of grid. In this method, the grid is divided into some subnet, so the differences of meteorological characteristics in grid can be reflected by meteorological characteristics of each subnet.The load of each subnet can be forecasted according to their meteorological characteristics. The load of grid can be summed together by the load of each subnet. As the line losses and other factors, the load of grid is often higher than the sum of subnet, so a correction is used to revising the sum. At last, in order to improve the accuracy of short-term load forecasting, we can choose forecasting method in each subnet. The forecasting method can be different in subnet.