Evaluación del método Pentágono de Duval para el diagnóstico del estado de transformadores de potencia

Abstract

Studies have been carried out on monitoring of gases formed inside a transformer due to internal faults or other factors that affect its useful life. The DGA analysis is the first predictive diagnostic technique, which consists of the separation of gases that are dissolved in oil, in order to identify the root of their breakdowns and avoid possible contingencies and unavailability in service. The study explains the methodology and the assessment process of the method called Duval Pentagon, presented at the end of 2014, as a new tool that proposes to provide more information about the transformer condition by relating in a single algorithm the most significant gases present in oil for fault analysis. A sample of 35 transformers were sampled that presented some type of damage caused by failures in their isolation or due to external factors. With this information from electrical tests and visual inspection of each transformer and the use of a Matlab program that incorporates the 7 main methods of interpretation of DGA analysis, it was possible to compare results, determine the reliability of the Pentagon by the number of correct diagnoses, and indicate the advantages and disadvantages by comparing it also with other methods.