Evaluation and Interpretation of Dissolved Gas Analysis of Soybean-Based Natural Ester Insulating Liquid

Abstract

The electrical industry needs a better understanding in the topic related to the fault diagnosis, since these fluids are chemically different in comparison with conventional mineral oils, therefore, its performance in field changes, for example the generated gases. This paper presents an investigation related to the gas generation of a new soybean-based natural ester liquid, which includes a data compilation of distribution and power transformers diagnosis during factory tests by dissolved gas analysis (DGA) method. This work covers an evaluation of the generated gases for three common types of transformer faults (overheating, partial discharge and high energy discharges) at lab scale testing and their interpretation by using different diagnostic methods to predict the fault type. It is found that overheating problems in natural esters can be associated with ethylene, while acetylene and hydrogen are linked for high energy fault and partial discharge faults, respectively. On the other hand, it is confirmed that Duval triangle is the most reliable method to diagnose natural esters liquids. The authors consider that the DGA data presented in this work represent valuable information for the transformer industry.