GPU-optimized LBM-MRT simulation of free convection and entropy generation of non-Newtonian power-law nanofluids in a porous enclosure at REV scale
Mashnoon Islam,Md Farhad Hasan,Sidhartha Bhowmick,Md. Kamrujjaman,Md. Mamun Molla
DOI: https://doi.org/10.1080/01430750.2022.2160811
2022-12-30
International Journal of Ambient Energy
Abstract:The present research investigates the heat transfer and entropy generation of a non-Newtonian power-law Cu-water nanofluid in a square porous enclosure at a representative elementary volume (REV) scale using the multiple-relaxation-time lattice Boltzmann method (MRT-LBM). The Graphics Processing Unit (GPU) accelerated the computations involved. The Brinkman-Forchheimer-extended Darcy model and the shear rate dependent power-law viscosity model were taken into account to characterise the non-Newtonian fluid flow in porous medium. Three types of model validations have been conducted, followed by the mandatory grid independence test, to demonstrate the precision and accuracy of the computational code. The numerical results were obtained by varying five key parameters namely, Darcy number ( Da = 10 −3 , 10 −2 , 10 −1 ); power-law index ( n = 0.6–1.0); Rayleigh number ( Ra = 10 4 , 10 5 , 10 6 ); porosity parameter ( ε = 0.6); volume fraction of nanoparticles ( φ = 0.0–0.05), and the Darcy Rayleigh number ( Ra m = 100, 1000). The numerically simulated outcomes were presented in terms of streamlines, isotherms, velocity, and temperature profiles as well as the Nusselt ( Nu ) numbers. Furthermore, a correlation among the parameters to determine the average Nu was developed using the Levenberg-Marquardt (LM) algorithm. The mathematical correlation was also subject to validations, and, in general, good correlation coefficients were obtained.