Abnormal State And Weak Point Identification Method Based On Random Matrix Theory

The continuous enlargement of the power system,the further increase of generator ca-pacity,the large-scale investment of energy sources and new types of power electronic devices have made the power grid structure more and more complex,and the consequences of abnor-mal power system operations have become more and more serious.The research on the safe and stable operation of power systems is vital.At present,the analysis of power system secu-rity and stability is mainly based on the model method.The analysis results depend heavily on the accuracy of the modeling of the system,and can not fully utilize the large amount of oper-ating data generated during the operation of the power grid.As a mathematical tool that can effectively analyze complex systems,the random matrix theory has good performance in an-alyzing and processing big data.It has been introduced into the field of power systems in recent years and has broad development prospects.Based on the summarization,induction and analysis of domestic and foreign literature,an operation evaluation method is proposed considering the characteristics of WAMS data and random matrix method.A WAMS measurement model in the form of a random matrix is es-tablished and the disturbance of the system is identified based on the random matrix theory.The system operation is judged and the correlation of the system state is analyzed.The weak nodes of the power system were identified to conduct differentiated protection.The work of the paper mainly includes the following aspects:1.A high-dimensional random matrix model of WAMS is constructed.In order to realize better description of big data in power system based on the matrix theory,a data processing process for grid data and a method for constructing random matrix are proposed.Based on the sliding time window method,the operation status of the grid is analyzed,and historical data and real-time data of the grid operation are fully utilized.It is not necessary to model the power grid and can analyze system behavior from the perspective of data.2.A detection method of abnormal operation state of power grid based on random matrix theory is proposed.In order to realize the condition monitoring that takes into account the dynamic changes of the power grid,spectral distribution analysis based on Ring law,MP law,and linear eigenvalue statistics in random matrices is used to detect weak disturbance events and abnormal status in the power grid;severity of disturbances are detected by Trace detection and spectral distribution;correlation analysis of the power system operating status is proposed based on the grid factor expansion matrix to identify the disturbance source that caused the grid disturbance.This method avoids the simplification assumption of the grid and can reflect actual operation of the system.which is fast and can provide a new means of the grid operation status analysis based on WAMS data.3.A method for identification of weak nodes in power grids based on random matrix the-ory is proposed.In order to overcome the limitation of the model method in the weak nodes identification,voltage data is used to construct a matrix,quantifying and analyzing the data fluctuation characteristics based on the MP law and linear eigenvalue statistics,and construct-ing a power grid vulnerability assessment model based on the entropy theory,a node vulnera-bility evaluation index to rank the weak nodes in the power grid is present.This method does not require physical modeling of the system and has improved accuracy over existing methods.New means of identification of weak nodes based on data-driven is provided.