Research On Theoretical Line Loss Optimization Based On Clustering Partition

Grid power loss(abbreviated as network loss)is one of the important methods for assessing the operating level of power supply companies.Theoretical line loss optimization calculation and analysis are important methods for formulating loss reduction plans and improving line loss management and grid operation levels.At present,the power system is becoming larger and larger,and at the same time,the requirements for calculation and analysis speed are becoming higher and higher.Therefore,from the perspective of reducing the calculation scale of the power grid and increasing the speed of calculation,a theoretical line loss based on clustering zoning is proposed.The optimization method is as follows:Firstly,the mathematical definition,working principle and clustering process of the clustering algorithm are analyzed in detail.Several representative clustering methods and the characteristics of each clustering algorithm are analyzed and summarized.The research focuses on different application backgrounds.The principle of grid division and the requirements for grid partitioning,and the analysis of several commonly used grid partition methods and partition principles and purposes;analysis of the application of fuzzy clustering algorithm in grid subarea,two types based on fuzzy clustering The grid partition method was introduced.Secondly,aiming at the problem of parallel optimization of theoretical line loss,from the point of view of line loss calculation,the formula for calculating the sensitivity of the line loss rate to the node power is deduced.The concept of gravitational distance is defined by the physics concept of gravitation to define the concept of gravitational distance to represent the coupling relationship between nodes.With the improved fuzzy clustering algorithm,the grid is partitioned,and the optimal number of partitions is used to determine the optimal number of partitions.A method to improve the fuzzy clustering network partition method with the aim of parallel optimization of theoretical line loss is proposed.Finally,based on the grid partition,the multi-agent technology is combined with the parallel computing method,and it is used for the parallel optimization of theoretical line losses,and a theoretical line loss parallel optimization computing framework based on multi-agent technology is proposed.The detailed design of the management coordination relationship and data interaction mode between the general agent and each regional agent,as well as the functions of the modules contained in the agents at each layer.Combined with the basic principles of the optimization calculation method,the data exchange and coordination mechanism in the calculation process between the regional agents is designed.By dividing the large-scale grid system area by applying the partition method,the optimization problem of the line loss of the whole network is decomposed into the parallel optimization of multiple sub-networks.With the premise of ensuring the calculation accuracy,the overall calculation speed is improved and the calculation time is reduced.