State Estimation Of Power Grid Division Based On Innovation Graph

With the continuous development and renewal of power network technology,the scale and structure of power network are increasing day by day.The exponentially increasing number of substations and huge measurement information have created new tests for the real-time,accuracy and rapidity of power system state estimation.The innovation graph state estimation has the characteristics of fast identification speed and low redundancy.In most cases,abnormal events in the network can be accurately identified.However,as the network scale increases,the effect of the innovation graph state estimation sometimes changes affected.Therefore,in order to further improve the recognition ability of the innovation graph state estimation,based on the existing theory,this paper conducts a divisional study on the innovation graph state estimation.The necessity of state estimation for the partition of the innovation graph is pointed out and verified by a numerical example.Improve the partition theory of the innovation graph,classify the existing partitioning methods into two types: the partition with the node as the boundary and the partition with the branch as the boundary,and two partition methods are given to deal with the innovation of the boundary node.When the network has topology errors,bad data and sudden changes in generator power,the performance characteristics of abnormal events after partitioning are analyzed.The partition state estimation of the innovation graph is applied to the actual provincial power grid to identify the topological errors of the network.Aiming at the problem of incorrect topological state identification when the innovation graph state estimation is applied to the actual power grid,two partitioning methods are used to calculate respectively to verify the effectiveness of the innovation graph partition state estimation.The multi-level partition identification method of innovation graph of generator mutation is studied.According to the voltage level of the line,the network is divided into two levels.The backbone network with a higher voltage level is the first level,and the low-voltage side network with a lower voltage level is the second level.According to the geographical location of the substation,the second level network is divided into multiple levels.Partitions.Search for power mutation nodes in the primary backbone network,find the mutation path,continue to identify the power mutation in the secondary low-voltage network according to the mutation path search results of the backbone network,correct the continuous branch of the branch where the mutation path is located,calculate the new information,and exclude The impact of sudden power changes.The partition identification when the bad data and the sudden change of generator power exist at the same time is studied.The bad data is set in different positions of the partition,and the influence of the bad data on the identification of sudden change in generator power is analyzed.According to the characteristics of bad data in the partition,eliminate the influence of bad data.