Short-term Power Load Forecasting System In Inner Mongolia Based On Extreme Learning Machine

With the rapid development of economy and society demand for electricity continues to grow,China's power grid construction efforts have gradually increased.Because power load forecasting involves a wide range of impact factors and uncertainties.It always a diffculty for power network planning departments to consider.Research on a comprehensive consideration,the prediction accuracy of load forecasting method,providing a reference for scientific and rational planning and distribution grid companies,has extremely important theoretical and practical value for the long-term development of power grid construction and grid companies.Based on the gray relational analysis,this paper constructs the index system for short-term power load forecasting in Inner Mongolia.Because the grid load may be affected by multiple indicators,so a reasonable index system is the fundamental prediction.Through the analysis of the meaning of the evaluation index,we can select some indexes which are closely related to the load of the power grid.Through the common gray relational analysis method and the correlation degree analysis method,the load value and history of Inner Mongolia during the "Eleventh Five-Year" period The data are analyzed.Finally,according to the above analysis results,the short-term forecasting index system of Inner Mongolia power grid is put forward.Based on the GM(1,1)model,the forecast of the next month of the forecast of the province and its subordinate cities in Inner Mongolia was forecasted.In this paper,the simulation of the popular forecasting algorithm in recent years is proved by the GM(1,1)model,which is more suitable for predicting the load forecasting index of the power grid.The Because the load of the grid is affected by many indexes,the load forecasting has great uncertainty.It is difficult to accurately distinguish the relationship between the index and the load by the conventional mathematical analysis method.Therefore,this paper adopts the method of limit learning machine to establish the short-Forecasting model.Through the simulation experiment of the load of the province and its subordinate cities in 2012,the accuracy of prediction is greatly improved compared with the original method.On the basis of the above theoretical research,this paper designs and realizes the short-term power load forecasting software system in Inner Mongolia.The system combines the theoretical research and practical application,the Inner Mongolia Electric Power Company's short-term power load forecast work provides an effective support.