Adaptive optimal design of active thermally insulated windows using surrogate modeling

This paper explores the adaptive optimal design of Active Thermally Insulated (ATI) windows to significantly improve energy efficiency. The ATI window design uses thermostats to actively control thermoelectric (TE) units and fans to regulate the overall thermodynamic properties of the windows. The windows are used to maintain a comfortable indoor temperature. Since weather conditions vary with different geographical locations and with time, the thermodynamic properties of the windows should adapt accordingly. The electric power supplied to the TE units and the fans are dynamically controlled so as to provide an optimal operation under varying weather conditions. Optimization of the ATI window design is a multiobjective problem. The problem minimizes both the heat transferred through the window and electric power consumption. The heat transfer through the ATI windows is analyzed using FLUENT; and the optimization is performed using MATLAB. Since the computational expense of optimization for numerous weather conditions is excessive, the power supplies are optimized under a reasonably small number of weather conditions. Based on the optimal results obtained for these conditions, a surrogate model is developed to represent the optimal results in a wide range of weather conditions. The surrogate model is used to evaluate optimal power supplies with respect to different values of outside temperature, wind speed, and intensity of solar radiation. Thermometers, anemometers, and solar radiation sensors are used to sense these weather conditions. With the inputs from the sensors, the thermostats determine the operating conditions and calculate the corresponding optimal power supplies using the surrogate model. Since the ATI windows are operated with optimal power supplies, high energy efficiency is achieved.