An investigation on the use of artificial neural networks for rapid transient stability simulation [of power systems]

Due to their massively parallel and highly interconnected architectures, the computational efficiency of artificial neural networks (ANN) is much higher than conventional computer simulation techniques. Consequently the feasibility of using ANN for emulating power system transient stability simulations is explored in this paper. In the proposed approach, the power system is broken down into its individual components, each of which is represented by a distinct ANN module. These individually trained modules are subsequently coupled together to form an integrated ANN-based power system simulator capable of emulating transient stability behaviour. Results obtained on a small 5-bus test system show that high quality simulation can be achieved using this approach