There is a growing interest in utilizing electrical power sources that recuperate part of the exergy destruction by conventional power systems, and thermoelectric generators (TEGs) are sound candidates. This paper reports a mathematical model – based on Onsager formalism that relates gradients (temperature, electric potential) to fluxes (heat flux, electrical current) – and its numerical, finite element (FEM) implementation that is used to simulate the underlying, multiphysics processes. In this study, an elemental TEG cell – the simplest system under investigation – is optimized for maximum efficiency. Then, it is packed into more complex, higher order ensembles that inherit its outlining features. This “growing” technique is derived from the constructal theory that explains shape and structure in systems of finite size (volume, resources), with internal flows (e.g., heat flux, currents), subject to specific internal and external constraints.

Keywords: Thermoelectric cells, numerical simulation, structural optimization