Research On Transient Stability Assessment Based On Support Vector Machines

Power system transient stability assessment is one of the most important problems which relate power system secure-stable performance. The fuzzy theory and combination theory combined in dealing with TSA problem in this paper have been systematically studied in this paper, which includes: 10-machine-39 bus-New England system is simulated in this paper. First, Power system BPA software is used for transient stability simulation, thus the original data sets are formulated; SVM method is adopted in TSA, and the detailed algorithm flow is given. In this part, ROC curve used as a new method of SVM training parameter is proposed, and more assessment indexes are introduced, which conformed the results of the simulation; FSVM which apply a fuzzy membership to each input point of SVM and reformulating SVM into fuzzy SVM, is proposed for overcoming the limitation of SVM, that is SVM treats all the training points uniformly. So FSVM is proposed for dealing with TSA problem.Fisrt, K nearest method (K-NN) is applied to construct fuzzy membership. Then, the TSA results are given by C++ programming. The simulation results show that the accuaracy is improved; AdaBoost learning method is proposed for boosting the results of SVM. This method combines different SVMs, which overcomes the limitation of SVM. The simulation results are given and show that this method is the most effective in classification reliability; The above methods are compared. The results show that the improvement methods are effective in dealing with TSA problem.