Voltage stability monitoring by different ANN architectures using PCA based feature selection

Voltage Stability is a challenging issue for secure and reliable operation of modern power systems. In this paper, a fast and efficient Artificial Neural Network (ANN) based approach with dimensionality reduction is proposed for online voltage stability monitoring of power systems. The dimension of the system data is reduced by selecting suitable training features for ANN using Principal Component Analysis (PCA). The performance comparison with different types of ANN architectures is also carried out for the proposed approach. Various voltage stability indices are used as indicator for voltage stability monitoring under varying operating conditions including N-1 contingency. The effectiveness of the proposed approach is demonstrated on IEEE 39 bus New England test system.