Transient Stability Prediction Algorithm Research Based On Wide Area Measurement System

With the development of power network's interconnection, the problems of power system's stability become more and more severe. In the last few decades, the world's economy was heavily struck by the happening of large scale outage frequently. In this circumstance, a new real-time dynamic monitor system (Wide Area Measurement System, WAMS) is brought into use. WAMS that based on Synchronized Phasor Measurement Technology (SPMT) provides a new technology platform for the stability problems.Considering the time characteristic of transient disturb, transient stability prediction becomes very important to prevent the enlargement of faults. Based on the real-time generator angle data provided by WAMS, a novel time series data based transient stability prediction method is proposed to solve the on-line transient stability problems in the paper.Because there are a lot of generators in the modern interconnected power system, CCCOI-RM transform is used in order to reduce the calculation complexity in the paper. In this way, multi-machine system is aggregated to OMIB system. Then we judge the system's transient stability from the predicted angle curve with Extremum plus Threshold Double Criterion.According to the angel curve prediction method, Least Square Support Regression (LS-SVR) prediction algorithm is used in the paper. LS-SVR algorithm has excellent non-linear problem processing abilities by transform non-linear problems into high dimension space linear problems, and simplified the calculation by using kernel function. Compare to AR algorithm, LS-SVR algorithm has advantages in speed and accuracy.The method is fast and has no direct collection with the system detail model. Simulation results show that the proposed method yield satisfactory accuracy for transient stability prediction, and provide a new angle of view in transient stability prediction area.