Short-term Load Forecasting Based On Divisional Hourly Weather Forecasting With Applications

The precision of short-term load forecasting is one of important index assessed by State Grid Company under the circumstance of power marketing. There are two shortage in previous load forecasting with weather factors, one is only general everyday weather factors are used just like maximum temperature , the other is that only provincial weather factors are used. In this paper, much more detailed hourly and divisional weather factors are full adopted in short-term load forecasting to improve the precision. Hereinafter three major works are included:1. By comparing the relationship between load and weather to find the most related weather factor to load, giving support to forecast modeling.2. Discussing the principle of division, Considering the two way of forecasting in whole or in part, we know that some index in power system could be added directly, but others can't, neither does the weather factors obviously, thereout three kinds of strategy for load forecasting are put forward, and one of them that load forecasting based on total load and complex weather factors is recommended.3. ameliorating two forecasting model based on hourly weather factors, which are the models of the similar day multi-elemental regression and the similar day radial basis function, the process and procedure of 96 load points forecasting modeling are introduced, meanwhile the results gotten from two models are also compared and analyzed. On this base the application software is designed and put into use in Henan Power Grid Company and shows better outcome.