Multiple-relaxation-time lattice Boltzmann simulation of natural convection flow in a partitioned cavity using GPU computing

Md Jahidul Haque, Md. Mamun Molla, Nasrin Akhter, Suvash C. Saha
In this paper, we demonstrated the implementation of General Purpose Graphics Processing Unit (GPGPU) programming in Compute Unified Device Architecture (CUDA) C for the simulation of natural convection flow in a side-heated three-dimensional (3D) rectangular cavity with a partition. In the present lattice Boltzmann method (LBM) D3Q19 multiplerelaxation-time (MRT) and D3Q6 single relaxation-time (SRT) model are implemented for the simulation of fluid flow and temperature phenomena,
more » ... The parallel code is validated with the benchmark problem of a side heated cubic cavity. The results are presented by the temperature distribution in terms of isotherms, local and average Nusselt number and 3D view of iso-surface for the different Rayleigh number (Ra) and the Prandtl number fixed at Pr = 0.71. It is also observed that the present parallel implementation of the MRT-lattice Boltzmann simulation in GPU has a substantial computational efficiency rather than the sequential programming in central processing units (CPU).
doi:10.1063/1.5115862 fatcat:yynsn27s6ndtdhbjuyfbcccw2e