Análise comparativa de métodos de interpretação de gases dissolvidos em óleo isolante de transformadores de potência

Abstract

Power transformer plays a crucial role in the energy supply chain and it’s going to dramatically escalate due to the increasing complexity of the electrical power system (EPS) and the society demand for energy. In view of this, the unscheduled shutdown of this equipment can result in major impacts on society. Thereby, maintenance is the most efficient way to predict and prevent such events, especially in its predictive approach. Like this, dissolved gas analysis (DGA) on oil of power transformers stands out as one of the most promising techniques of predictive maintenance. Given this scenario, this work seeks to carry out a comparative analysis of methods of DGA applied to power transformers. To this end, the main DGA methods are delimited in the standards, two of them are selected (Duval's Triangle 1 and Basic Gas Ratio) and then applied to a dataset, evaluating the respective performances through proposed parameters. From the results obtained, there is a close diagnosis accuracy rate between the two methods, around 50%, while the errors manifest themselves in different ways: wrong fault diagnoses for the first method and failure indefinitions for the second. Therefore, this result gives the general behavior of the delimited DGA methods with their respective diagnostic particularities.