Research On The Evaluation Of Power Grid Static Voltage Stability Based On The Theory Of Situation Awareness

Static voltage stability has always been a concern in power system research.With the continuous application of artificial intelligence and large data technology in power system,the use of artificial intelligence algorithm to analyze the static voltage stability of power system has become a hot spot.In addition,in order to better grasp the real-time operation status and future development trend of the power system,situational awareness theory in the power system research is also ongoing.The key of situational awareness is to collect grid operation data,real-time state analysis and future state pre-diction,which is suitable for real-time static voltage stability analysis and prediction of power system.Based on this,situational awareness theory is applied to evaluating the static voltage stability of power grid in this paper,and evaluates and analyses the static voltage stability of power grid from two levels of real-time and future time respectively.The main contents of this paper are as follows:(1)Aiming at the real-time operation state of power grid,a static voltage stability assessment method based on stochastic matrix theory is studied.Data mining technol-ogy and random matrix theory are used to evaluate the real-time static voltage stability of power grid.Firstly,the correlation analysis of operation data such as voltage ampli-tude,voltage phase angle,active and reactive power input at the head and end of branch is carried out.The random matrix model is constructed based on the node voltage am-plitude data,and the real-time static voltage stability of power grid is evaluated accord-ing to the mean spectral radius index.The simulation results show that this method can evaluate the static voltage stability of power system in real time.(2)On this basis,a comprehensive evaluation method of weak nodes in power grid based on the entropy weight method is proposed.The weak nodes in power system are identified by combining data model and physical model.On the one hand,from the point of view of data-driven,the random matrix theory is applied to the identification of weak nodes in power grid.The random matrix model is constructed based on the voltage amplitude data of nodes,and the evaluation index of spectral radius entropy is proposed by combining the mean spectral radius with entropy theory.On the other hand,considering the physical characteristics of the power grid,singular value entropy index is used to evaluate weak nodes.Spectral radius entropy and singular value entropy re-flect the effect of node load disturbance on node voltage amplitude from the perspective of data characteristics and physical characteristics,respectively.Considering the above two indicators,the entropy weight method is used to determine the weight value of the indicators,so as to calculate the comprehensive evaluation index of the weak nodes in the power grid,and identify the weak nodes in the power grid.The validity of the pro-posed method is verified by the analysis of simulation examples.(3)A static voltage stability prediction method based on online sequential extreme learning machine is proposed for future operation state of power system.Using online sequential extreme learning machine method,the data prediction model of power grid is established to predict the voltage amplitude of nodes,and then the change of node voltage in the future is obtained.According to voltage stability index,the static voltage stability in the future operation state of power grid is evaluated and predicted.The va-lidity and validity of the proposed method are verified by simulation analysis of an example.This method can provide some theoretical guidance for power system dis-patchers.