Optimal Design and Multi-Parameter Sensitivity Analysis of a Segmented Thermoelectric Generator

Abstract

Thermoelectric generators are high-profile energy conversion devices that can convert heat energy into electricity. In this study, a novel 1D resistance model was established to evaluate the performance of a segmented thermoelectric generator (STEG) with variable properties, and the genetic algorithm was adopted to optimize the performance. Then, influence factor analysis, multi-parameter optimization, and sensitivity analysis for an STEG couple were conducted. The results showed the great influence of geometric sizes on performance. Moreover, the optimal length ratio between the length of the high-temperature segment and the total leg length increased when the temperature difference (ΔT) was raised, but it remained unchanged as the convective heat transfer coefficient (h) changed. Furthermore, the ratio of the leg length to its cross-sectional area is affected by thermal conditions and the length ratio, while the cross-sectional area ratio between P- and N-type thermoelectric legs was not affected by the convective heat transfer coefficient. Under the conditions of ΔT = 300 K and h = 2000 W/m2K, the maximum power increased by 11.02%. Finally, the global sensitivity analysis found that material properties, especially the Seebeck coefficient, dominate the influence on optimal power. These results could contribute to the optimal design of STEGs.