Modeling and predicting the thermal state of a transformer through an adaptive intelligent approach

The thermal condition of power transformers is a critical factor influencing insulation aging and overall operational reliability, especially under varying load and environmental conditions. Accurate forecasting of transformer temperature enables preventive maintenance and supports the development of intelligent monitoring systems within smart grid environments. In this study, an Adaptive Neuro-Fuzzy Inference System (ANFIS) model is developed to predict transformer temperature based on ambient temperature and load factor. The proposed ANFIS model is trained and validated using experimental field data and its performance is benchmarked against a classical linear regression model. The comparative evaluation demonstrates that the ANFIS model provides superior accuracy, achieving RMSE = 0.89 °C, MAE = 0.75 °C, and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mrow> <mml:mi>R</mml:mi> </mml:mrow> <mml:mrow> <mml:mn>2</mml:mn> </mml:mrow> </mml:msup> <mml:mo>=</mml:mo> </mml:math> 0.987, whereas the regression model yields RMSE = 1.95 °C, MAE = 1.63 °C, and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mrow> <mml:mi>R</mml:mi> </mml:mrow> <mml:mrow> <mml:mn>2</mml:mn> </mml:mrow> </mml:msup> <mml:mo>=</mml:mo> </mml:math> 0.871. These results indicate that ANFIS effectively captures nonlinear thermal dynamics that cannot be represented by linear approximation methods. The study highlights the suitability of the proposed approach for real-time transformer thermal monitoring and its potential integration into smart grid-based predictive maintenance and asset health management frameworks.