Online identification of coherent generators in power system by using SVM

With the deregulation of the electrical power system and competitive business environment, power system monitoring and control have been emerged as a potential research area. Operation and control of power system are based on several strategies under normal, critical and emergency states of a power system. Control actions can be initiated either at load end (load shedding) or at generator end (generation rescheduling). Both of these actions required the knowledge of the swing of the generators at different operating/contingency conditions. This paper presents an application of supervised learning model to detect coherent machines in a large power network. A method to find the real-time transient stability state and identification of the coherent generator groups by predicting the rotor angle values following a large disturbance through Support Vector Machines (SVMs) is proposed in this work. The proposed method enables to determine synchronism state of the individual machine in real-time. The validity of proposed method is investigated on IEEE-39 bus test system.