NEURAL NETWORK APPROACH FOR PROCESSING SUBSTATION ALARMS

Abstract

This paper studies the method of processing of alarms in an electric substation using artificial neural networks (ANN) as a tool. Whenever there is a fault in the power system, or there is a significant change in the network, alarms are issued to the substation operators, with each alarm being associated with a message. This is the driving force for research and experiments regarding innovations of more sophisticated methods of alarm processing. The method used in this study is artificial neural network. The network is a pattern recognition unit. Before using the network for alarm processing, the network is trained using back propagation algorithm to identify all possible combination of alarms received in a power system.