A comprehensive review of thermoelectric generation optimization by statistical approach: Taguchi method, analysis of variance (ANOVA), and response surface methodology (RSM)

Wei-Hsin Chen,Manuel Carrera Uribe,Eilhann E. Kwon,Kun-Yi Andrew Lin,Young-Kwon Park,Lu Ding,Lip Huat Saw
DOI: https://doi.org/10.1016/j.rser.2022.112917
IF: 16.799
2022-11-01
Renewable and Sustainable Energy Reviews
Abstract:The thermoelectric generator (TEG) can directly convert heat to electricity. However, its efficiency is low, so optimizing TE systems to maximize output power is necessary. Many review papers have focused on this technology. However, there has not been a comprehensive review of TEG optimization by a statistical approach. This study reviews thermoelectric generator optimization by the Taguchi method, analysis of variance (ANOVA), and the response surface methodology (RSM) to identify the major optimization findings and tendencies for this technology. Three optimization paths are identified: operating conditions, geometrical configuration, and TE materials for thermoelectric generators (TEGs). Although there is no “one-size-fits-all” combination of characteristics that a TEG system should have, some tendencies based on the results of previous studies have been identified. The key parameters that show the most significant effect on the TEG system for each optimization path are the heat source temperature for the operating conditions and the TE leg height for the geometrical configuration. However, there are no distinctly recognized parameters for TE materials. Thus, these results show that optimizing the heat source conditions of a TEG system will yield the best possible results, and optimizing the TE leg height in the TE module would further improve the system. About 70% of the studies optimizing thermoelectric generators utilized the Taguchi method; thus, the Taguchi method remains the most popular statistical tool for TEG analysis. Finally, the perspectives and challenges of optimizing thermoelectric generators using statistical approaches are underlined.
energy & fuels,green & sustainable science & technology
What problem does this paper attempt to address?