Topology Optimization of Segmented Thermoelectric Generators

The thermoelectric (TE) power output, \(f_P\), and conversion efficiency, \(f_{\eta }\), for segmented thermoelectric generators (TEGs) have been optimized by spatially distributing two TE materials (BiSbTe and Skutterudite) using a numerical gradient-based topology optimization approach. The material properties are temperature-dependent, and the segmented TEGs are designed for various heat transfer rates at the hot and cold reservoirs. The topology-optimized design solutions are characterized by spike-shaped features which enable the designs to operate in an intermediate state between the material phases. Important design parameters, such as the device dimensions, objective functions and heat transfer rates, are identified, investigated and discussed. Comparing the topology optimization approach with the classical segmentation approach, the performance improvements of \(f_P\) and \(f_{\eta }\) design problems depend on the heat transfer rates at the hot and the cold reservoirs, the objective function and the device dimensions. The largest performance improvements for the problems investigated are \(\approx \) 6%.