Multicriteria optimization based comprehensive comparative analyses of single- and two-stage (series/parallel) thermoelectric generators including the influence of Thomson effect

Abstract

The thermodynamic investigations on the thermoelectric devices (TEs) discard the influence produced by the non-linear Thomson effect. It could direct the incomplete/partial modelling solutions laterally through some critical gaps in the performance evaluation of these devices. On the contrary, a suitable arrangement of several designing constraints for TEs is essential to improve their operating characteristics. In this context, the modeling of multi-element single- and two-stage thermoelectric generators based on the thermodynamic principles is done in MATLAB 9.2. The irreversibility due to Thomson influence along with Joule/Fourier effects are undertaken for the system modelling. The optimization of the generators is done in pursuance of obtaining the optimal values of four input parameters using two different evolutionary algorithms, viz., NSGA-II and MOEA/D. The optimum solutions from the Pareto front of two-/three- objective are found using different decision-making methods, viz., TOPSIS, Fuzzy, and LINMAP. It is observed that the proposed optimization yields trivial variances amongst ideal/obtained solutions, named as the deviation index, in comparison with the single/dual ones. In addition to this, sensitivity analysis is done to examine the impact of Thomson effect on the output power/thermal efficiency of the generators. The test results obtained through NSGA-II are in coherence with those of the data and figures reported in the available literature.