Multi-objective optimization of heat exchanger in an automotive exhaust thermoelectric generator

Currently, thermoelectric generator optimization mainly focuses on the transformation or improvement of a heat exchanger. This study aims to eliminate the contradiction between the thermal and pressure performances of a heat exchanger through the mathematical optimization of the fin distribution. To evaluate the thermal properties and pressure losses of a heat exchanger, five fin parameters and four optimization targets (i.e. average temperature, temperature gradient in longitudinal direction, temperature gradient in transverse direction and average static pressure drop from the inlet to the outlet of the heat exchanger) are proposed. An L-16 (4(5)) orthogonal array computational fluid dynamics simulation is employed, and the influence of the five fin parameters on thermal and pressure performances is explored. Results show that fin height has the most significant effect (34%) on the average temperature increase. The influence of fin interval distance on the four targets contradicts that of the fin height. Accordingly, the fin interval distance is the only parameter that is good for the pressure drop decrease. The fin angle shows a homogeneous effect on the four targets as well as interval distance, except for the significant decrease of longitudinal temperature difference. Four surrogate models are constructed via a third order response surface approach to extract the variation regularity between the five parameters and four targets. Finally, an archive-based micro genetic algorithm is used to obtain the optimal fin parameters. Although the horizontal temperature difference after optimization is reduced, the average temperature is improved from 222.46 degrees C to 226.4 degrees C, whereas the longitudinal temperature difference is decreased from 29.36 degrees C to 28.9 degrees C. Moreover, the pressure drop is decreased by approximately 20%, which may be significant for the global improvement of a thermoelectric generator system. (C) 2016 Elsevier Ltd. All rights reserved.