Kernel Mixing Model for Turbulent Combustion and Transport PDF method.
This repo is for 0-D simple validation, for PaSR validation please refer to PaSR
- Fortran version:
- src_fortran/mixing_test.f90: main program for testing.
- src_fortran/mixing_model_kerm.f90: source code of
KerMwith differential diffusion supported. - src_fortran/mixing_model_iem.f90: source code of
IEM, used for comparison. - src_fortran/mixing_model_mcurl.f90: source code of
MC, used for comparison. - src_fortran/mixing_model_mcmg.f90: source code of
MCMGadopted from 1, which implements the mapping closure mixing model used for comparison. - src_fortran/emst.f and src_fortran/emst_subs.f: source code of
EMSTadopted from 2, used for comparison.
# build executable file (single precision float) make fortran # run simulation, faster than cpp version ./mix # show results python performance.py fortran
-
Comparison results of 10k particles for EMST and MCMG; 50k particles for other models, where the dash lines are for the DNS data from 3 (For
KerM, sigma_k=0.25 is used)
-
Performance of mixing models
C++ version
-
C++ Version:
- src_cpp/MixingModels.hpp: implementation of mixing models
- src_cpp/main.cpp: run simulation
- performance.py: show results
# build executable file make # run simulation, taking around 10 seconds for N=1000 ./mix # show results python performance.py
Comparison results of 4k particles for EMST and 100k particles for other models (
KerMsigma_k=0.25), where the dash lines are for the DNS data from 3
Performance of mixing models (
EMST-1Ddo not account aging properties and noIEMassisted, for originalEMSTimplementation, please refer to the Fortran version)
Python version
-
Python Version:
- src_python/models.py: implementation of mixing models
- src_python/mixing_py.py: run simulation
# run simulation, EMST taking around 6 seconds for N=400 python src_python/mixing_py.pyComparison results of 400 particles for EMST and 50k particles for other models (only uniform weighted samples are used), where the dash lines are for the DNS data from 3
Please refer to TheoryGuide.pdf
Footnotes
-
The original MCMG implementation comes from https://tnfworkshop.org/wp-content/uploads/2019/03/MM-INTAS.zip ↩
-
The original EMST implementation comes from https://tcg.mae.cornell.edu/emst/ ↩
-
A. Juneja, S.B. Pope, A DNS study of turbulent mixing of two passive scalars, Physiscs of Fluids 8 (1996) 2161. ↩ ↩2 ↩3
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent