Research On Spatial Load Forecasting Method Of Urban Power Grid

With the development of society and economy,the energy crisis is becoming more and more serious.The renewable energy,such as photovoltaic and wind power generation,is heavily connected to the distribution network.The flexible control ability and the active coordination mechanism of the distribution network will increase.Load forecasting plays an important role in distribution network planning,and its accuracy is directly related to the advantages and disadvantages of distribution network planning.Load forecasting includes total load forecasting and spatial load forecasting.This paper focuses on the characteristics and applications of two spatial load forecasting methods(land simulation and load density index).It lays a foundation for the comparison between load density index method and land use simulation method.At the same time,it also analyzes the relevant factors of the spatial load forecasting of distribution network,and analyzes the key points and difficulties in the project.In view of the present load density index method,the experience method and other methods are usually adopted,and the problem of low accuracy exists.A spatial load forecasting method based on genetic algorithm and load density index is proposed in this paper.Firstly,fuzzy clustering algorithm is applied to cluster different power consumption characteristics into several grades,and a load density index system for spatial load forecasting is established.Then the training samples of SVM are determined to provide appropriate training data for the later prediction model.Finally,based on support vector machine and genetic algorithm,the density of the predicted area is obtained.And the load can be obtained according to the area area.The example shows that this method can improve the adaptability and accuracy of prediction models.In view of the inaccuracy of land use simulation method for land use types,In this paper,the spatial load forecasting of genetic algorithm and fuzzy reasoning is proposed,and fuzzy knowledge base is established by using fuzzy logic system to combine the actual mode of spatial load forecasting in distribution network.Fuzzy logic technology is used to fuzzily,inferring and clarification of raw data and related information.The score of each cell is obtained.Considering the fact that fuzzy sets and fuzzy rules used in fuzzy reasoning need to be constantly adjusted.Therefore,genetic algorithm is introduced to train fuzzy rules,which achieves good results.Finally,the detailed steps of spatial load forecasting of distribution network based on GIS are given.First,the power supply area is divided into a number of small and uniform plots.Then,the land use is allocated by genetic algorithm,and the land use decision is obtained.Finally,the prediction value of the cell load is calculated.Through the calculation and analysis of the above two methods,Most of the land use in the country is determined by the government.Spatial load forecasting based on load density index is more applicable.For foreign or domestic parts,land use is determined by market.The accuracy of spatial load forecasting based on land use simulation method is even higher.