Using artificial neural network to assess transient stability of power systems

For the safe running of a power system, it is very important to assess, online, power system transient stability. The frequency domain and time domain calculation methods for stability assessment are time consuming, therefore, neither one is suitable for the purpose. ANNs have recently been proposed as an alternative method for solving certain traditional problems in power systems where conventional techniques have not achieved the desired speed, accuracy or efficiency. The application areas of ANNs are mainly regression, classification and combinatorial optimization. Using ANNs to assess power system transient stability is a classification problem. An ANN needs quite a long time for offline training, however, having been trained, it needs only a very short time to make a classification for a problem. Here, the authors describe how a multilayer feedforward ANN is suitable for the online transient stability assessment of power systems.<>