Power transformers play a vital role in power systems, from the generation of electrical energy through transmission and distribution to end consumers. They are often loaded near rated-load conditions during operation. This leads to increased losses and a rise in temperature in the power transformers. The thermal performance of power transformers is a critical factor influencing operational reliability and insulation life,
yet standard International Electro technical Commission-based models often assume a constant ambient temperature, limiting their predictive accuracy under real operating conditions. The objective of this research is to develop an improved International Electro technical Commission-based thermal model that incorporates time-varying ambient temperature effects during transformer-type testing. The research method includes exponential equations that are validated against stepped-load data from a 5 MVA transformer, where the ambient temperature ranged between 27°C and 32°C during the experimentation. A comparative analysis showed that the improved model reduced the mean absolute error for oil temperature rise from 3.15% to 2.04%, and for winding temperature rise from 2.30% to 1.88%. Similarly, root-mean-square error
and mean percentage error decreased by 30–40%, indicating a consistent improvement in predictive accuracy across metrics. These findings indicate that incorporating ambient temperature variability enables the model to more accurately reproduce both steady-state conditions and transient heating and cooling behavior, thereby aligning more closely with physical expectations. Inference from the results suggests that the improved model offers a scientifically rigorous yet computationally efficient alternative to more complex data-driven approaches. The novelty of this work lies in bridging the gap between oversimplified, constant-ambient International Electro technical Commission-based thermal models and computationally intensive methods, thereby providing a standards-consistent framework with enhanced realism. The outcomes have direct
implications for transformer manufacturers, enabling more reliable type testing; for utilities, enabling improved insulation life assessment and overload management; and for standards developers, recommending integration of dynamic ambient effects into standard guidelines.