Framework of fast stability assessment of power system by neural network-based pattern recognition

The authors discuss the application of the artificial neural network-based pattern recognition method (ANN) to the fast stability assessment of power systems. The problem is very important in a modern energy control centre. The acceleration energy and the kinetic energy of key generators, and the voltage magnitude and phase angle of the key buses in power system are used as the input. A stability index as the output of the neural network provides an evaluation of the system security level, which provides potentially valuable guidelines to operators and leads to preventive control strategies. The input-output sets of typical operation conditions and disturbances are selected as training samples. The proposed approach is illustrated by a three-machine system. The preliminary result is satisfactory.<
>