Static Voltage Stability Boundary Feature Extraction Driven By Information And Its Application

With the continuous expansion of the scale of power grid interconnection and the rapid development of new energy grid connection,the safe and stable operation of the power grid is seriously threatened,which makes the static operation close to the border and even lose stability,resulting in major blackouts.Therefore,it is of great practical significance to extract the feature of static voltage stability boundary for clarifying the behavior of power system in the stable boundary and ensuring the static stable operation.At the same time,with the continuous development of big data and artificial intelligence methods in the field of power grid application research,the "dual channel" research paradigm represented by causal analysis and statistic analysis has been widely recognized,and the theoretical analysis framework of information driven power system has been proposed.Therefore,under the guidance of the "dual channel"research paradigm,this dissertation carries out the information driven static voltage stability boundary feature extraction work,and accordingly develops the grid intelligent control system,the main research contents are as follows:(1)A representation model of the geometric feature of the static voltage stability boundary is proposed.The fast and accurate construction of the stability region boundary can be realized based on the topological characteristics of the power system stability boundary,taking into account the continuous change correlation of each boundary point,The model first uses the initial value of a stable boundary point provided by the continuous power flow method,and uses the continuous parameter tracking method to realize the migration generation from any initial value point to the whole stable boundary solution set,so as to realize the "point to surface"generation of boundary shape.The simulation results show that the proposed method can not only maintain the accuracy of boundary points,but also reduce the time cost of boundary construction.It has a certain engineering value and can effectively describe the geometric feature of the boundary.(2)A feature representation method of boundary topological distance is proposed in this dissertation based on reinforcement learning theory.Aiming at the shortcomings of existing methods in evaluating boundary distance,and combining with the current reinforcement learning technology,an agent with the ability of boundary detection interaction is constructed.Through the information of action decision track acquired by the agent in the complex high-dimensional parameter space,it can be evaluated shortest distance from the front operating point to the stable boundary,and then effectively evaluates the static voltage stability margin,and the knowledge extracted from the trajectory information of the agent behavior,which can provide the power grid auxiliary operation basis for the operation dispatcher;finally,the effectiveness of this method is verified by the simulation example,which can achieve the effective extraction of the boundary distance feature.(3)In view of the increasingly complex regulation and operation situation of the current power grid,combined with the current advanced big data processing technology and guided by the concept of information driven,the system architecture of the intelligent regulation and control platform of the power grid is designed.Taking an actual regional power grid as an example,the algorithm verification and visual development work are completed through the laboratory cluster environment,and the boundary feature extraction algorithm is completed in the intelligent power grid dispatch and control platform.