Artificial Intelligent Fault Diagnostic Method for Power Transformers using a New Classification System of Faults

Conventionally, the defects and faults in power transformers are classified in based on the phenomena like discharge and thermal problems, which makes it difficult that inspectors find the position of defect or fault parts in the power transformers. In this paper, a new diagnostic method for power transforms is proposed. The method presents a new classification system of defects and faults based on structures and parts in the power transformers and artificial intelligent algorithms. The proposed method uses totally 189 DGA data, certified through internal inspections by KEPCO and classified to 6 major categories (winding, Core, Clamp, Bushing, OLTC, and Oil) and 18 sub-categories. In the last of this paper, the diagnostic performance of the proposed method is verified by simulations.