Application Research Of Random Matrix Theory In Power System Situation Awareness

Power system is a typical big data system.Data mining method provides an effective tool for power system situation awareness.Especially with the construction of the ubiquitous power internet of things,the big data analysis of the power system is paid more attention.Starting from the power big data mining,this paper aims to use the random matrix theory to study the power system big data modeling,disturbance location,weak link identification and voltage stability online monitoring to promote the safe and stable operation of the power system.The power system big data modeling method is summarized.The power system can use the measurement data to construct a random matrix,and use the translation time window method to realize real-time data processing.In order to meet different analysis needs,the augmented random matrix construction method of expanding data sources,changing weights and correlativity research is adopted.The branch index which can reflect causality is introduced into the random matrix,and the improved mean spectral radius of fusion node and branch information is proposed and its effectiveness is analyzed.The power system disturbance localization strategy based on improved mean spectral radius is presented.The power system disturbance localization algorithm is realized from the two aspects of constructing distance index and K-means clustering,which can quickly locate the disturbance location.Based on the eigenvalue probability density function,the weak link judgment index of power system is given.The fusion node information and judgment index propose a weak node identification strategy for power system.By introducing branch index into random matrix and combining node and branch information,a weak branch identification strategy for power system is proposed.The effectiveness of the proposed strategy is demonstrated by simulation.The voltage stability situation of the system is evaluated by monitoring the mean spectral radius curve.The determination principles of initial mean spectral radius value,the voltage stability warning mean spectral radius value and the voltage critical collapse mean spectral radius value are given.Combined with the weak link identification method based on random matrix theory,a voltage stability monitoring strategy based on random matrix theory is proposed,which can effectively monitor and evaluate the voltage stability of power system.This paper focuses on introducing data reflecting causality into big data analysis of power system.Some proposed analysis strategies based on random matrix theory can be used in the analysis of power system disturbance location,weak link identification and voltage stability online monitoring.After further improvement,it can provide effective tools for power system situation awareness and early warning,with certain theoretical significance and engineering application value.