Research on transformer temperature rise detection and optimization method based on multi-physical field coupling

To address the inefficiency of conventional temperature rise tests for dry-type transformers in box substations, this study proposes a closed-loop "Standard-Experiment-Simulation-Optimization" methodology. An electromagnetic‑thermal coupled model was built, calibrated with field test data. The simulation revealed a saturated exponential growth pattern of hotspot temperature rise, verified experimentally. Based on this, additional thermocouples were placed beyond standard positions, enabling earlier and more accurate hotspot measurement. A predictive model using the first‑five‑hour data was derived to estimate steady‑state temperature, rise value, and required test duration. This approach reduces testing time and provides a reusable reference for improving hotspot detection accuracy.