ICA based Artificial Neural Network model for voltage stability monitoring

In this paper, Artificial Neural Network with imperialist competitive algorithm (hybrid ANN-ICA) approach for online monitoring of long-term voltage instability has been proposed. The Imperialist Competitive Algorithm (ICA) has been used to improve the accuracy of ANN by tuning its meta-parameters such as input activation function, output activation function, number of hidden layer neurons and learning rate. The proposed approach uses the voltage phasor as the input vectors measured by phasor measurement units (PMUs) and outputs the Voltage Stability Margin Index (VSMI) vector. The performance evaluation using the proposed approach is carried out on New England 39-bus test system. The results of the proposed hybrid ANN-ICA approach for voltage stability monitoring is compared with ANN model on same data set.