Online Monitoring Model Research Of Regional Power System Voltage Stability Based On Pattern Recognition

Power system stability research is in a position of growing significance, directly influencing the security stability of the power system. The nowaday challenging problem is the precise and real time evaluation of power system voltage stability condition.In order to resolve the problem, based on principls and modeling methods of pattern recognition and decision tree, considering the operational situation of the system, a combined traditional static voltage stability analysis method-pattern recognition approach has been proposed by the thesis. The principle, structure and its modeling methodology were illustrated and discussed. A decision tree based classification and regression model was built, which has the capability of real time evaluation of power system voltage stability. The model proposed in this thesis can be used to support the remedial action decision process.The main topics in this thesis are:(1) The nonlinearity of the state space of power system voltage stability was studied. The characteristic vector capable of describing the voltage stability status of power system was proposed. After the study of response characteristic of the state vector, a method of acquiring the state vector was found.Considering the nonlinearity of the system state space, an algorithm of acquiring the state vector recursively characterizing all the state space of system.(2) After the study of features of system state vector, the most suitable feature vector was choosen. Based on the feature vector, considering the nonlinearity relationship of votlage magnitude, voltage angle, current magnitude, current angle and the power system stabilty criterion of the operating point, with decision tree algorithm, a online monitoring model of regional power system voltage stbality status was established. The mapping relationship was achieved by combining the realitively easy accessible attributes and the voltage stability evaluation criterion. Using this method, the long tedious model-based computation process of the traditional evaluation is ignored. And the model can be used to predict the real time voltage stability of regional power system, leading to the online monitoring of the system security status.(3) The response characteristics of online monitoring model is studied during the dynamic simulation of a certain power syste. By carefully selecting the proper contingencies,during the preset fault dynamic process, the online monitoring model of power system votlage stability was used to detect the system instability. And the result is accurate and robust. The online monitoring model of power system voltage stability can be used on the online evaluation of long-term power system voltage stablity status.