Artificial neural network (ANN) application in dissolved gas analysis (DGA) methods for the detection of incipient faults in oil-filled power transformer

Abstract

Power transformer is one of the fundamental equipments in the power system. Transformer breakdown or damage may interrupt power distribution and transmission operation, as well as incur high repair cost. Thus, detection of incipient faults in power transformer is essential and it has become an interesting topic to study. This paper presents the application of artificial neural network (ANN) in detecting incipient faults in power transformers by using dissolved gas analysis (DGA) technique. DGA is a reliable technique to detect incipient faults as it provides wealth of information in analyzing transformer condition. For this project, ANN was developed to classify seven types of transformer condition based on three combustible gas ratios. The development involves constructing several ANN designs and selecting network with the best performance. The gas ratio are based on IEC 60599 (2007) standard while historical data were used in the training and testing processes. The selected ANN design yields a very satisfactory result where it can make a reliable classification of transformer condition with respect to combustible gas generated.