Real-Time Determination of Overall Heat Transfer Coefficient from the Seebeck Effect by Using Adaptive Learning-Rate Optimization

Abstract

It is challenging to determine the actual overall heat transfer coefficient under thermal conditions during processes. In a conventional approach, they are obtained as a constant with the empirical formula for the given conditions. In this study, the Adaptive moment estimation (Adam) technique is investigated for adaptive learning-rate optimization in the real-time determination of the overall heat transfer coefficient via the Seebeck effect in the thermoelectric modules. Two thermoelectric modules detect heat transfer as solid surfaces exposed to the outdoor air. The principle of energy balance and the Seebeck effect determine the overall heat transfer coefficients over time. The heating/cooling process of a copper plate is considered with exposure to the outdoor air. The overall heat transfer coefficient is determined with the proposed methodology over time. The temperature of the copper plate is numerically determined by the mathematical models with the obtained values of the overall heat transfer coefficient. It is confirmed that the calculated values of temperature are close to the measured values, with RMSE = 0.07 °C.