Optimization assisted CFD for using double porous cylinders on the performance improvement of TEG mounted 3D channels

Abstract

Performance of a thermoelectric generator (TEG) mounted into a channel is analyzed by using double porous cylinders in channel flow with finite element method. COBYLA (Constrained Optimization BY Linear Approximations) algorithm is used to obtain the optimum size and permeability of the porous objects to achieve the highest power. As compared to a non-object channel configuration, the installation of the porous object results in higher power generation. A lower permeability of the double cylinders increases the TEG power while 10.5% increment is obtained when lowest and highest permeability configurations are compared. When sizes of the porous objects are varying up to 19% rise in the TEG power is attained as compared to no-object channel case. The optimum set of parameter to achieve the highest power is obtained as ap1 = 0.232hz, ap2 = 0.75hz and Da = 10-6. At the optimum conditions, there is 22.5% rise of TEG power when hot side Reynolds number is increased from 200 to 600 while this value is 18% for the no-object case. At the highest nanoparticle loading, 20.5% higher power is obtained at the optimum case when compared to no-object channel case. The optimization assisted computational fluid dynamics (CFD) study of TEG installed channel flow with passive methods is very effective in achieving the best performance as compared to computationally expensive high fidelity multi-parametric CFD. As thermoelectric energy conversion and related devices are used in diverse energy systems technologies including solar-thermal applications, thermal management in different thermal systems, refrigeration and many others, the outcomes and computational methods will be useful for efficient design and optimization studies.