Comparison of Artificial Intelligence Techniques and Conventional Techniques for Power Transformer Fault Diagnosis with Dissolved Gas Analysis (DGA): Challenges and Opportunities

In power systems, transformers have a key role in terms of electrical energy sustainability in transmission and distribution. Moreover, this crucial instrument of power systems must be operated at optimum performance in electric transmission lines and power distribution system as well as rational economic conditions. When usual operating conditions is became, mineral oil comes out on a slow and typical degradation. If on the contrary due to thermal or electrical stress, the degradation status accelarates. In this case, some gases (ethylene,acetylene,carbonmonoxide,carbondioxied, hydrogen, methane,ethane) are occured in the mineral oil. For dissolved gas analysis (DGA), custom traditional methods seems to identify and classify the initial faults. However, they are unable to notice the status in case of multiple electrical and thermal errors simultaneously. While the serious limitation of conventional DGA methods in terms of accuracy and consistency, AI methods gives well prediction. In the study, future challenges and opportunities are elaborately presented for researchers.