Analytical Incremental Learning for Power Transformer Incipient Fault Diagnosis Based on Dissolved Gas Analysis

Power transformers are important assets for electrical network operation. The diagnosis of transformer condition is important for the continuity of electricity distribution to consumers. When be used for a significant amount of time, the transformer will be degraded in performance. The degradation leads to a fault caused by the appearance of dissolved gases in oil insulating material. Dissolved Gas Analysis (DGA) is a method of analyzing dissolved gases in insulating oil to diagnose the condition of the transformer. Many models have been developed in analyzing incipient fault in transformers, both conventional methods and artificial intelligence-based methods. However, these methods can give conflicting results and provide unstable and inconsistent analysis. In this study, DGA Analysis using Analytical Incremental Learning (AIL) will be explored. The AIL model is applied by using the IEC TC10 dataset with diagnosis accuracy at 91.82%.