Numerical Optimization of the Thermoelectric Cooling Devices

The present study proposes the unified numerical approach to the problem of optimum design of the thermoelectric devices for cooling electronic components. The method is illustrated with several examples which are based on the standard mathematical model of a single-stage thermoelectric cooler with constant material properties. The model takes into account the thermal resistances from the hot and cold sides of the TEC. Values of the main physical parameters governing the TEC performance (Zeebeck coefficient, electrical resistance and thermal conductance) are derived from the manufacturer catalog data on the maximum achievable temperature difference, and the corresponding electric current and voltage. The independent variables for the optimization search are the number of the thermoelectric coolers, the electric current and the cold side temperature of the TEC. The additional independent variables in other cases are the number of thermoelectric couples and the height-to area ratio of the thermoelectric pellet. The objective for the optimization search is the maximum of the total cooling rate or maximum of COP. In the present study, the problems of optimum design of thermoelectric cooling devices are solved using the so-called Multistart Adaptive Random Search (MARS) method [16].