Power System Dynamic Analysis by Support Vector Machine

The main purpose of this paper is to utilize the method of Support Vector Machine (SVM) to assess the power system dynamics to decide which one of the loads should be tripped after the fault occurs and then is isolated to make the system stable. Because once the fault occurs the protection relay is triggered to isolate the fault point but the system may be still unstable, therefore the backup protection relays in the system are then triggered, and this may cause the outage region wider and even results in system blackout. To avoid system blackout, the suitable load is to be tripped to make the system return to another stable operating point after the fault has been isolated. The suitable trip loads are the transient stable samples which are selected by SVM, and then the load is selected which has the lowest impact on system. This paper we employs different loading conditions to increase the number of training samples to promote the accuracy rate of SVM. The results show that the accuracy rate of the purpose method can reach 70.86%.