In this paper, standard loss of insulation life equations have been applied to step up transformers as part of the design of an intelligent relay that would calculate the time left before a possible trip (assuming that conditions remained constant) and should allow the operators to avoid tripping the transformer. The basic idea is that the transformer can withstand temperature in inverse proportion to time. The algorithm to derive the hot-spot temperature, the relationship of the hot-spot temperature to probe temperature, as well as the ability to provide operating guidelines for a derated transformer are shown. Revised transformer overload ratings as well as predictive protection can be obtained with an accurate thermal model of the limiting hottest spot temperatures. Based on the constant feedback from actual probe measurements, the thermal model is developed is detailed enough to consistently obtain very accurate results for various loads and seasons.