| taylor-green-3d.tar |
02/2014: 3D version of Incompact3d for the Taylor-Green flow at Re=1600 (up to t=20) and for a resolution of 129^3 mesh nodes.
More information about this flow: M. Brachet, D. Meiron, S. Orszag, B. Nickel, R. Morf, U. Frisch, Small scale structure of the Taylor-Green vortex, J. Fluid Mech. 130 (1983), 411–452.
Plots for enstrophy and kinetic energy can be obtained from the output of the simulation:
nohup mpirun -np 16 incompact3d > OUTPUT.file &
grep ENSTROPHY OUTPUT.file > enstrophy.dat
grep KINETIC OUTPUT.file > kinetic.dat
Figure of the vorticity field at t=12 for isocontour 5 (generated with the code in the archive)
More information about this flow: M. Brachet, D. Meiron, S. Orszag, B. Nickel, R. Morf, U. Frisch, Small scale structure of the Taylor-Green vortex, J. Fluid Mech. 130 (1983), 411–452.
Plots for enstrophy and kinetic energy can be obtained from the output of the simulation:
nohup mpirun -np 16 incompact3d > OUTPUT.file &
grep ENSTROPHY OUTPUT.file > enstrophy.dat
grep KINETIC OUTPUT.file > kinetic.dat
Figure of the vorticity field at t=12 for isocontour 5 (generated with the code in the archive)
| incompact2d.tar |
05/2016: updated version uploaded (two typos were found in convdiff.f90)
04/2016: Incompact2d: 2D version of Incompact3d
For 2D simulations with a single core. The flow configuration is an array of 3 cylinders in a channel flow. Periodic BC are used in the streamwise direction.
-->To use inflow-outflow in the streamwise direction: 1) set up nclx=2 in incompact3d.prm and 2) set up nx=513, nxm=nx-1, mx=nx+1 in module_param.f90
-->To use inflow-outflow in the streamwise direction and periodic BC in vertical direction: 1) set up nclx=2 and ncly=0 in incompact3d.prm and 2) set up nx=513, ny=256 nxm=nx-1, nym=ny, mx=nx+1, my=ny+2 in module_param.f90
-->To use a refined mesh in the centre of the domain: set-up istret=1 in incompact3d.prm and ajust parameter beta (large beta=weak refinement, small beta=strong refinement), adjust your time step accordingly!
-->vtr files for Paraview are generating during the simulation with ux, uy, the vorticity and the pressure field.
-->average_stats.f90 can be used for the visualisations of the time averaged statistics (see STATISTIC subroutine in visu.f90).
-->the solid cylinders can be changed in body.f90.
-->the flow configuration can be changed but please note that not all possible configurations have been tested.
-->please read the user guide before using the code.
Top left figure generated with the code in the archive; top left figure generated with inflow-outflow in the streamwise direction; bottom left figure generated with inflow-outflow in the streamwise direction and periodic BC in the vertical direction; bottom right figure generated with inflow-outflow in the streamwise direction and periodic BC in the vertical direction and a stretching in the vertical direction with istret=1 and beta=2.
04/2016: Incompact2d: 2D version of Incompact3d
For 2D simulations with a single core. The flow configuration is an array of 3 cylinders in a channel flow. Periodic BC are used in the streamwise direction.
-->To use inflow-outflow in the streamwise direction: 1) set up nclx=2 in incompact3d.prm and 2) set up nx=513, nxm=nx-1, mx=nx+1 in module_param.f90
-->To use inflow-outflow in the streamwise direction and periodic BC in vertical direction: 1) set up nclx=2 and ncly=0 in incompact3d.prm and 2) set up nx=513, ny=256 nxm=nx-1, nym=ny, mx=nx+1, my=ny+2 in module_param.f90
-->To use a refined mesh in the centre of the domain: set-up istret=1 in incompact3d.prm and ajust parameter beta (large beta=weak refinement, small beta=strong refinement), adjust your time step accordingly!
-->vtr files for Paraview are generating during the simulation with ux, uy, the vorticity and the pressure field.
-->average_stats.f90 can be used for the visualisations of the time averaged statistics (see STATISTIC subroutine in visu.f90).
-->the solid cylinders can be changed in body.f90.
-->the flow configuration can be changed but please note that not all possible configurations have been tested.
-->please read the user guide before using the code.
Top left figure generated with the code in the archive; top left figure generated with inflow-outflow in the streamwise direction; bottom left figure generated with inflow-outflow in the streamwise direction and periodic BC in the vertical direction; bottom right figure generated with inflow-outflow in the streamwise direction and periodic BC in the vertical direction and a stretching in the vertical direction with istret=1 and beta=2.
| guia_do_usuÁrio_incompact3d_v2.pdf |
04/2016: Latest version of the user guide in Portuguese
| cylinder.tar |
10/2015: This is an archive of Incompact3d in a wake around a cylinder configuration with an explicit time advancement with Re=300. It is based on our new customized Immersed Boundary Method using Lagrange reconstructions for the flow inside the solid regions (see Gautier R., Laizet S. & Lamballais E., 2014, A DNS study of jet control with microjets using an alternating direction forcing strategy, Int. J. of Computational Fluid Dynamics, 28, 393--410).
It is of course possible to change the geometry of the object by replacing the subroutine "cylinder" in the file "genepsi3d.f90" with your own subroutine.
To get a turbulent state around the cylinder:
--> First, compile and execute the file "genepsi3d.f90" ("gfortran -O3 genepsi3d.f90" and then "./a.out")
--> Then run incompact3d (with ilag=1 in the file "incompact3d.prm").
After about 20,000 time steps you should get a fully turbulent state as in the picture (obtained using paraview_incompact3d.f90).
Note: this version does not work yet in double precision but we are working on it!
It is of course possible to change the geometry of the object by replacing the subroutine "cylinder" in the file "genepsi3d.f90" with your own subroutine.
To get a turbulent state around the cylinder:
--> First, compile and execute the file "genepsi3d.f90" ("gfortran -O3 genepsi3d.f90" and then "./a.out")
--> Then run incompact3d (with ilag=1 in the file "incompact3d.prm").
After about 20,000 time steps you should get a fully turbulent state as in the picture (obtained using paraview_incompact3d.f90).
Note: this version does not work yet in double precision but we are working on it!
| user_guide_incompact3d_v2.pdf |
10/2015: Latest version of the user guide
| paraview_incompact3d.f90 |
08/2015: This Fortran file can be used for visualizations using Paraview. It can directly read the binary files generated by Incompact3d in the subroutine visu.f90. No need to write vtr files anymore!
| channel.tar |
10/2015: This is an archive of Incompact3d in a channel flow configuration with an explicit time advancement. The configuration is the one from Laizet S. & Lamballais E., High-order compact schemes for incompressible flows: a simple and efficient method with the quasi-spectral accuracy, J. Comp. Phys., vol 228-15, pp 5989-6015, 2009.
To get a turbulent state:
-->5000 iterations with rotation (to be activated in convdiff.f90, line 253)
-->10000 iterations with no rotation to get a fully turbulent channel flow as seen in the picture (obtained using paraview_incompact3d.f90).
To get a turbulent state:
-->5000 iterations with rotation (to be activated in convdiff.f90, line 253)
-->10000 iterations with no rotation to get a fully turbulent channel flow as seen in the picture (obtained using paraview_incompact3d.f90).
| user_guide_incompact3d_v1-1.pdf |
11/2013: Basic user guide of Incompact3d. Old version.