Computational Investigation of total-enthalpy based Conjugate heat transfer in particle-laden flows using Lattice Boltzmann Method

All the results have been obtained using the open source lattice Boltzmann framework WaLBerla developed by the chair of system simulations of the Friedrich Alexander Universität Erlangen-Nürnberg. To be able to reproduce the results pertaining to this paper, please follow these steps: unzip walberla.zip cd walberla mkdir build cd build cmake .. -DWALBERLA_BUILD_WITH_CODEGEN=ON (per default, the code is built for CPUs, if it should run on NVIDIA GPUs, add "-DWALBERLA_BUILD_WITH_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=60", if it should run on AMD GPUs, add "-DWALBERLA_BUILD_WITH_HIP=ON" instead) folder "catalyst_validation" refers to the raising catalyst particle validation folder "ardekani_validation" refers to the spherical particles in Couette flow validation and verification folder "specific_heats_validation" refers to the validation of sedimenting spherical particle with varying specific heats and conductivity ratios. For raising catalyst_validation, go to: cd apps/benchmarks/MaterialTransport/particles_total_enthalpy_cht/catalyst_validation For particles in Laminar Couetter Flow validation, go to: cd apps/benchmarks/MaterialTransport/particles_total_enthalpy_cht/ardekani_validation For sedimenting cold particle validation, go to: cd apps/benchmarks/MaterialTransport/particles_total_enthalpy_cht/specific_heats_validation Once in the desired folder, for compiling: make -j The general command to execute an mpi parallel simulation of walberla is: For catalyst_validation: mpirun -np $n ./catalyst_virtualfluid_cht_sc_3 catalystParticles.prm For ardekani_validation: mpirun -np $n ./executable ardekani.prm For specific_heats_validation: mpirun -np $n ./specificheats_virtualfluid_cht_sc1 specific_heats_becnhmark.prm The files attached are for the following results: velocity_vs_time_w1_sc3.txt correspond to the catalyst validation results for the velocity and displacement evolution velocity_vs_time_cp1.txt velocity_vs_time_cp2.txt velocity_vs_time_cp4.txt velocity_vs_time_cp8.txt correspond to the cold particle sedimentation validation case with different specific heats with the naming of the file and using TRT collision operator the files with "_srt" in their name are results of the SRT simulations. the files with "kr_10" "kr_100" and "kr_1000" correspond to the conductivity ratios 10, 100, 1000 The files ardekani_0.1_pr7.log : corresponds to volume fraction of 0.1 and Prandtl number of 0.7 and similarly for the other files. The performance results have been obtained on the LUMI cluster and the bash scripts to build and run the simulations on LUMI cluster on both the cpu and gpu partititions has been attached with the names: build_virtual_cht_cpu.sh : For building on the LUMI-C partition, can be executed as ./build_virtual_cht_cpu.sh build_virtual_cht_gpu.sh : For building on the LUMI-G partition, can be executed as ./build_virtual_cht_gpu.sh to RUN the scaling runs: 1) weak scaling: weakscaling_virtual_cht_cpu.sh and weakscaling_virtual_cht_gpu.sh for weak scaling on LUMI-C and LUMI-G respectively 2) Strong scaling: strongscaling_small_cpu.sh, strongscaling_medium_cpu.sh, strongscaling_large_cpu.sh on LUMI-C strongscaling_small_gpu.sh, strongscaling_medium_gpu.sh, strongscaling_large_gpu.sh on LUMI-G