Parameter optimization of thermoelectric modules using a genetic algorithm

Thermoelectric modules (TEM) are solid state components which are able to convert thermal to electric energy without moving parts and therefore are generally recognized as durable and reliable. However, the performance of TEM is strongly depending on the temperature difference driving the heat flux and causing thermo-mechanical stress within the module. This study introduces a multi-objective optimization procedure based on a genetic algorithm with which this conflict in objectives can be solved under realistic boundary conditions. Therefore a finite element model is presented and an optimization model is derived. The results of the optimization show that a significant improvement in electric power output and mechanical stability can be achieved by regarding all relevant design parameters in the module.