Enhanced Transformer Fault Diagnosis Using SFRA, ANFIS-Based Classification, and Dielectric Impact Analysis of Bushing Insulation

Power transformers are crucial to the stability of the electrical grid, and their failure can result in severe financial losses and power disruptions. Faults such as short circuits, mechanical stresses, and insulation degradation result in axial and radial deformations, compromising both structural integrity and electrical performance. Sweep frequency response analysis (SFRA) is a widely used diagnostic tool for detecting these faults; however, its reliance on manual interpretation limits its accuracy and efficiency. This study integrates SFRA with an adaptive neuro-fuzzy inference system (ANFIS) to enhance fault classification by incorporating a bushing model to analyze moisture ingress and dielectric variations. Statistical features, such as skewness, kurtosis, and cross-correlation, are extracted to quantify fault severity, thereby improving classification accuracy. MATLAB-based simulations validate the impact of dielectric changes on SFRA response, demonstrating that bushing insulation has a significant influence on transformer reliability. Results confirm that ANFIS effectively differentiates winding faults from bushing-related insulation issues, offering a more comprehensive health assessment. By enabling early fault detection and predictive maintenance, this approach enhances transformer reliability and contributes to grid stability. Future work can explore deep learning techniques to further improve classification accuracy and automation in transformer diagnostics. The proposed SFRA–ANFIS framework achieves a classification accuracy of 98% for transformer winding faults and 92% for bushing-related dielectric and moisture-induced insulation faults, demonstrating reliable fault identification and severity assessment compared to conventional SFRA-based diagnostic approaches.

Cite this article as: P. S. Kumar, C. Kumar, V. S. Kumar and S. Chatterjee, “Enhanced transformer fault diagnosis using SFRA, ANFIS-based classification, and dielectric impact analysis of bushing insulation,” Electrica, 2026, 26, 0359, doi: 10.5152/electrica.2026.25359.