Soft computing techniques for Prediction and classification of Faults in power ransformers on Load using dissolved gas analysis

Abstract

Analysis of transformer oil for dissolved gas is a useful predictive newlineand maintenance tool for determining the residual life of a transformer newlineInsulating mineral oils used in transformers contain mixture of many newlinehydrocarbons These hydrocarbons decompose and liberate gases when newlinesubjected to thermal or electrical stress The components of solid insulation newlinealso take part in the formation of gases which dissolve in the oil The sources newlineof gas generation are transformer windings loose joints tap changers core newlinelaminations and bushings Excessive formation of a gas may indicate newlineinsulation deterioration An assessment of these gases both qualitatively and newlinequantitatively would help in diagnosing the faults in the transformer newlineThe Dissolved Gas Analysis DGA is carried out by extracting newlinenine key gas components namely hydrogen oxygen nitrogen methane newlinecarbon monoxide carbon dioxide ethane ethylene and acetylene from newlinesamples of oil taken from oil filled equipments newline newline