Power System Transient Stability Margin Assessment Using Steady-state Information

Transient stability analysis is one of the core tasks in power system online dynamic security assessment.It is of great significance to ensure the safe and stable operation of the power grid and prevent potential disturbances caused by destabilizing factors.The development of fast and accurate methodologies for online transient stability assessment(TSA)is one of the major research topics in the area of power system operation and control.Traditional transient stability analysis mainly adopts analytical methods,in which time-domain simulations are performed for an expected contingency set according to current operation mode.Large amount of online computation is required in traditional analysis,which brings about a heavy computational burden to the control system and can hardly meet the real-time requirements of online evaluation.In recent years,with powerful self-learning ability and high computing efficiency,artificial intelligence has become an important tool for fast transient stability assessment.For any TSA method with the use of artificial intelligence,most transient stability studies use dynamic characteristics as input variables and depend on post-disturbance trajectories to make judgement.When a judgement is made,the system already loses its stability,which is too late for preventive control and only emergency control measures can be taken.To overcome this disadvantage,some studies choose steady-state characteristics as input variables and thus the computational efficiency of online evaluation is largely improved.The major advantage of using steady-state information is the high computation speed,which makes it possible to identify the critical regions for dispatchers to monitor so that preventive control measures can be taken.For any TSA method with the use of steady-state information,accuracy of transient stability quantification assessment indices and choose which artificial intelligence method to establish the mapping relationship between input and output are its two key points.In this paper,these two problems are deeply studied and discussed.Main work of this paper is as follows:(1)The trajectory analysis method based on the potential energy conversion characteristics was proposed to assess transient stability.In this paper,a multi-swing based strategy that overcomes the false identification issues of previous methods is firstly illustrated to improve the accuracy of transient stability assessment method.(2)For an expected contingency set,a composite neural network is constructed and trained offline to establish the mapping relationship between steady-state characteristics and generator transient stability indices.It combines probabilistic neural network(PNN)with back propagation neural network and uses PNN to classify the sample data in advance,then uses BP to predict the security margin of different classifications.In order to improve the defects of PNN misclassification,a network is built previous network to improve the accuracy.(3)Using modeling simulation and data mining technology,the regression analysis method based on the operating data was proposed to reveal the relationship between steady-state information and transient stability indices,and to select the combination of characteristic variables with the greatest influence on stability.It can provides important reference for guiding prevention and control.The validity of the proposed methodology is demonstrated in the IEEE 39-bus system.The method proposed in this paper can quickly assess transient stability margin under the most serious fault only through steady-state information.It is of great significance for the security control of power system.