Power System Dynamic Security Assessment – Classical to Modern Approach

Dynamic security analysis is performed to ensure that a power system will survive any recognized contingency and will transfer to a new but acceptable steady-state condition. This article deals with one of the very important dynamic security indices – the critical clearing time. Some of the popular methods of calculating the critical clearing time (CCT) have been reviewed. A more recent method of determining CCT through artificial neural networks (ANN) has been presented. ANN approach of determining CCT is based on classical pattern recognition method. The CCT of a test power system has been estimated through two artificial neural networks – the back-propagation algorithm and the radial-basis function network. The neural networks were trained for a large number of simulated data obtained from numerical integration of the system dynamical equations. The trained networks were then tested with randomly selected data. It was observed that the radialbasis function network was more suited in terms of speed of computation and accuracy of prediction.