General framework for neural network based real-time voltage stability assessment of electric power system

The need for real-time security assessment of electric power systems has increased due to open systems, an increase in the number of power wheeling transactions and environmental concerns. In this paper, special attention is focused on neural network generalisation in large-scale system modelling. Generalisation has been improved by operation points classification and a reduction of the number of neural network inputs. The results prove the capability of neural networks to model the most critical voltage stability margin in a large electric power system. The proposed approach is tested with an IEEE 118-bus test network. The generalisation and training time of a neural network model can be improved significantly using the proposed methods.