FAST VOLTAGE ESTIMATION USING AN ARTIFICIAL NEURAL-NETWORK

Fast estimation of bus voltages is important for contingency analysis and security assessment of a power system. In this paper, an approach based on artificial neural networks (ANNs) is presented to estimate bus voltages in a very efficient manner. In the design of the ANN, a set of system variables which affect bus voltages most are first selected as the inputs to the ANN using an entropy function. A number of training patterns are then created to train the ANN. The resultant ANN is applied to estimate bus voltages following an outage event in a 30-bus system and in the Taiwan power system. Since accurate bus voltage predictions can be achieved very quickly by the proposed method, it is expected that the developed ANN can provide valuable information for operators in real-time system operation.