Research On Improvement Of Grid Division And Load Forecast In Urban Distribution Network Grid Planning

The grid planning practice of the distribution network in many regions in China shows that the reliability and quality of power supply in the grid planning area have been significantly improved,the line load problem has been solved,the internal users in the area have shortened the installation and connection process,and obtained electricity.The index is improved.However,there are also some problems in the practice of grid planning in China:on the one hand,China’s grid division has not formed a unified standard or method,and the result of grid division is highly dependent on the subjectivity of planners,and comprehensive consideration needs to be taken during the division process.Area,load demand,geographic information,feeder lines,tie lines,urban planning and other indicators,the division process requires a large number of repeated calculations of indicators,the planning efficiency is low;on the other hand,the emergence of grid planning puts forward new methods for load forecasting requirements,the grid planning process often needs to obtain the total amount information and spatial distribution information of the saturated annual load forecast of the planned area.The traditional total load forecasting method can no longer meet the needs of grid planning for the load spatial information of the dialogue area,and the existing space In the process of grid planning and application of the load forecasting method,there is no clear standard method for selecting both rate and load density indicators.Most regions adopt national common reference values,and the wide range of values cannot achieve good forecasting accuracy;besides,domestic New factors such as new energy vehicles,solar energy,and electricity price policy reform put forward higher requirements for the scientific nature of the grid,and at the same time put forward demands for the time distribution characteristic information of load forecasting.Based on the problem orientation,this paper first analyzes the current situation and problems of grid planning,and conducts improvement research on the existing problems.In terms of grid division optimization,the power supply grid division problem is regarded as an optimization problem of classifying power supply units and continuously searching for optimization,and related optimization algorithms are analyzed.Finally,genetic algorithm is selected to optimize the grid division problem,and through analysis Grid division practice,division guidelines and division steps,select main factors such as the number of feeders crossing the power supply grid in the power supply area,the number of tie lines crossing the power supply grid,and the distribution of distribution transformers as the optimization objects,and use genetic algorithm to establish the grid Divide the optimization model.Numerical examples prove that the model can overcome the dependence of the grid division results on the planner’s subjective experience,improve the efficiency of grid division,and provide a meaningful reference plan for the final grid division.In terms of load forecasting model research in grid planning,this paper analyzes the characteristics of grids and establishes a spatial load forecasting model based on the superposition of plot load curves.The model uses K-means clustering algorithm to cluster the daily load of users in the planning area.Obtain typical user daily load curves,and at the same time,for regional user types complex or unknown plots can be classified according to their historical daily load;in view of the difficulty in determining the saturated annual load density index,collect user load density information in mature areas of the local city,The non-parametric kernel density estimation method is used to analyze the data information,and finally various land load density indicators that better reflect the actual local conditions can be obtained;according to the load curve characteristics and load density indicators,the general spatial forecasting method can be used to obtain the load forecast of each block Finally,according to the load forecast curve of each block,using the idea of load curve superposition,the spatial distribution information and time distribution information of the load forecast of the power supply grid and the entire power supply area can be obtained.The calculation example shows that using the idea of load curve superposition to carry out load forecasting in the planning area can effectively solve the problem of simultaneous selection of spatial load forecasting.The load density index obtained by kernel density estimation can better reflect the actual local situation.Combining the solar energy and solar energy in the calculation example The consideration of electric vehicle factors also shows that the time distribution information of load forecast in the planning area has important reference value for the planning area.