The Research Of Short-term Load Forecasting Based On Geographic Information System

With the development of energy Internet,the access of a variety of new energy sources and the real-time change of user's load information make challenges to the accuracy and timeliness of load forecasting in power network.So,a critical question arises: How to predict short-term load precisely.Otherwise,geographic information system(GIS)plays an important role in planning and decision making in power system.In order to highlight the advantages of geographic information system in the power grid,GIS is used for short-term load forecasting and based on that,a comprehensive study is conducted in this paper.First of all,combined with the development of short-term load forecasting and the c onstitution and application of the GIS,the development prospect of short term load forec asting based on geographic information is expounded in this paper.Aiming at how to improve the accuracy of short-term load forecasting,grid partition and merging based on GIS is studied firstly.The load of a certain area or several areas is impl emented a new grid division according to the "load application" and "load density".On th e basis of the above analysis,the load is implemented a multi-level grid partition.Then a simple clustering method based on the load variation is combined with the method of spe ctral manifold clustering method is employed to cluster the influence factors of a given lo ad to realize the combination of the grid load and simplify the process of load forecasting.Finally,the validity of the meshing and merging model is verified by the results of the s hort-term load forecasting model.In order to increase the prediction of load forecasting,the algorithm of BP-RBF neural network is improved in this paper.Genetic algorithm is taken as the function to change input weight in BP neural network,and then the disadvantages of local optimization and slow computation speed are overcome.The effectiveness of improved algorithm is verified by example analysis.Finally,GUI interface design is used to build short-term load forecasting platform with GIS.The goal,principle and system structure of platform design are introduced first,and then the prediction process is demonstrated,which make prediction results more intuitive.