Planning of a neural network based integrated condition monitoring system for application in substations

With the rapidly changing technologies and new equipment such as gas insulated switchgear (GIS) and gapless surge arresters, new condition monitoring systems are required to assist looking into specific indicators relating to the equipment conditions. Integrated condition monitoring systems (ICMS) are required to study the multiple parameters and their effect on the equipment condition since plant aging and ultimate failure are usually a result of combined influence of more than one parameter. Artificial neural networks (ANN) have distinctive characteristics such as fault tolerance and analysis of complex networks at high speed and have already been proven to be superior than any other AI techniques for solving problems of diagnostic nature. Hence ANNs can be used for the diagnosis of the insulation problems and condition monitoring of equipment. This paper is intended to describe the planning and design approaches for an integrated condition monitoring system using ANN