Results for case Asymmetric plane diffuser

Code: Code_Saturne

Version: 2.0-beta2

Authors: J .Uribe

Method and Numerical Options

The %$k-\varepsilon$% model is used with scalable wall function (IDEUCH=2).

Centred scheme is used for the velocities (BLENCV = 1) and upwind for the turbulent variables (BLENCV=0).

Both an unsteady algorithm with a constant time step (IDTVAR=0) and the steady algorithm have been used (IDTVAR=-1).

Models

%$k-\varepsilon$%, SSG, %$\varphi-f$%, SST

Mesh

The mesh is refined near the walls to have y+=1 and solve the viscous sub-layer. Since the mesh has no colours, the boundary conditions are imposed using geometrical values.

The mesh format is the standard CGNS.

Description of the results files

The files contain measures at the following planes to match the experiments by Buice and Eaton.
Section code x/H U / Ub uu / Ub2 uv / Ub2 vv / Ub2
-6 -5.87 x x x x
03 2.59 x x x x
06 5.98 x x x x
13 12.75     x x
14 13.56 x x    
16 16.14     x x
17 16.93 x x    
19 19.53     x x
20 20.32 x x    
23 22.91     x x
24 23.71 x x    
26 26.3     x x
27 27.09 x x    
29 29.69     x x
30 30.48 x x    
33 33.07     x x
34 33.87 x x    
40 39.85 x x x x
47 46.62 x x x x
53 53.39 x x x x
60 60.17 x x x x
67 66.94 x x x x
74 73.71 x x x x

There is a ASCII file with the results at each location with the following columns:

y/h U/U_b uu/U_b^2 uv/U_b^2 vv/U_b^2 x/h

and two files with the non-dimensional coefficients, one for each wall, with:

x/h Cp Cf y+ U/U_b

The profiles are calculated using the subroutine usproj.f90

Boundary conditions

The inlet boundary conditions are taken from a precursor 1D channel flow simulation with the corresponding turbulence model. The mesh for the channel flow is 1D and matches exactly the inlet of the diffuser so no interpolation is needed.

The walls are treated as no-slip but wall function is used for %$k-\varepsilon$%

The sides are set to symmetry.

Reference Publications

Additional information (links, pictures, etc.)

figures 1-5 show the results of the current version compared to the previous. All profiles are in good agreement showing almost no difference between the selected models.

The non-dimensional friction and pressure coefficients can be seen in figures 6 and 7 for all turbulence models studied. The SST model give a too early separation, almost at the start of the diffuser, whereas the %$k-\varepsilon$% and SSG models predict almost no separation. The %$\varphi-f$% models yields a reasonable good agreement for the separation and reattachment points but with a slightly higher magnitude than the experiments. For the pressure coefficient, the SSG model produces the closest results to the experiments.

Figures 8 to 10 show profiles at three different sections along the diffuser: %$x/h=-5,20,27$%. Although all models represent the inlet velocity relatively well, the SSG is the only one that can reproduce the anisotropy. In the middle of the recirculation zone (%$\/h=20$%) it can be seen that even though the SSG model predicts correct levels for the stresses, ti has an incorrect behaviour of the velocity near the wall. This has been observed in previous ERCOFTAC workshop with some Second Moment Closures that do not resolve the viscous sublayer. The %$\varphi-f$% model gives overall better velocity prediction but since is it a eddy viscosity model, the stress levels are not as accurate as the SSG.

cf_inc_20_31_prev.png

  • Figure 1. Friction coefficient for versions 2 and 1.4:

cp_inc_20_31_prev.png

  • Figure 2. Pressure coefficient for versions 2 and 1.4:

KEPS_2vs1.4profil_14.png

  • Figure 3. Profiles at x= 14 for KEPS versions 2 and 1.4:

KEPS_2vs1.4profil_27.png

  • Figure 4. Profiles at x=27 for KEPS versions 2 and 1.4:

KEPS_2vs1.4profil_34.png

  • Figure 5. Profiles at x=34 for KEPS versions 2 and 1.4:

cf_inc_20_50_60_31.png

  • Figure 6. Friction coefficient for versions 2 and 1.4:

cp_inc_20_50_60_31.png

  • Figure 7. Pressure coefficient for versions 2 and 1.4:

PHI_SST_SSG_profil_-6.png

  • Figure 8.profiles at x=-5 for Phi, SST and SSG:

PHI_SST_SSG_profil_20.png

  • Figure 9. Profiles at x=20 for Phi, SST and SSG:

PHI_SST_SSG_profil_27.png

  • Figure 10.Profiles at x=27 for Phi, SST and SSG:



Current Tags:
Saturne_Validation_2.01Add my vote for this tag create new tag
, view all tags
Topic attachments
I Attachment Action Size Date Who Comment
pngpng KEPS_2vs1.4profil_-6.png manage 14.0 K 2009-10-02 - 15:07 JuanUribe profiles at x=-5 for KEPS versions 2 and 1.4
pngpng KEPS_2vs1.4profil_03.png manage 14.6 K 2009-10-02 - 15:08 JuanUribe profiles at x=3 for KEPS versions 2 and 1.4
pngpng KEPS_2vs1.4profil_14.png manage 14.3 K 2009-10-02 - 16:12 JuanUribe profiles at x= 14 for KEPS versions 2 and 1.4
pngpng KEPS_2vs1.4profil_20.png manage 14.7 K 2009-10-02 - 15:09 JuanUribe profiles at x=20 for KEPS versions 2 and 1.4
pngpng KEPS_2vs1.4profil_27.png manage 14.4 K 2009-10-02 - 16:12 JuanUribe profiles at x=27 for KEPS versions 2 and 1.4
pngpng KEPS_2vs1.4profil_34.png manage 14.3 K 2009-10-02 - 16:11 JuanUribe profiles at x=34 for KEPS versions 2 and 1.4
pngpng KEPS_2vs1.4profil_47.png manage 15.2 K 2009-10-02 - 15:10 JuanUribe profiles at x=47 for KEPS versions 2 and 1.4
pngpng KEPS_2vs1.4profil_60.png manage 15.0 K 2009-10-02 - 16:13 JuanUribe profiles at x=60 for KEPS versions 2 and 1.4
pngpng PHI_SST_SSG_profil_-6.png manage 16.2 K 2009-10-02 - 16:13 JuanUribe profiles at x=-5 for Phi, SST and SSG
pngpng PHI_SST_SSG_profil_03.png manage 15.9 K 2009-10-02 - 16:14 JuanUribe profiles at x=3 for Phi, SST and SSG
pngpng PHI_SST_SSG_profil_14.png manage 16.5 K 2009-10-02 - 16:15 JuanUribe profiles at x=14 for Phi, SST and SSG
pngpng PHI_SST_SSG_profil_20.png manage 17.0 K 2009-10-02 - 16:16 JuanUribe profiles at x=20 for Phi, SST and SSG
pngpng PHI_SST_SSG_profil_27.png manage 17.0 K 2009-10-02 - 16:17 JuanUribe profiles at x=27 for Phi, SST and SSG
pngpng PHI_SST_SSG_profil_34.png manage 17.1 K 2009-10-02 - 16:17 JuanUribe profiles at x=34 for Phi, SST and SSG
pngpng PHI_SST_SSG_profil_47.png manage 17.9 K 2009-10-02 - 16:18 JuanUribe profiles at x=47 for Phi, SST and SSG
pngpng PHI_SST_SSG_profil_60.png manage 17.6 K 2009-10-02 - 16:19 JuanUribe profiles at x=60 for Phi, SST and SSG
pngpng cf_inc_20_31_prev.png manage 11.1 K 2010-09-20 - 13:45 JuanUribe Figure
pngpng cf_inc_20_50_60_31.png manage 11.4 K 2010-09-20 - 13:45 JuanUribe Figure
pngpng cp_inc_20_31_prev.png manage 11.7 K 2010-09-20 - 13:45 JuanUribe Figure
pngpng cp_inc_20_50_60_31.png manage 12.1 K 2010-09-20 - 13:45 JuanUribe Figure
elsegz diffuser_v2.0b2.tar.gz manage 95665.7 K 2009-11-30 - 17:38 JuanUribe Case study file for v2.0.0beta2
Topic revision: r100 - 2019-04-29 - 09:57:28 - ConstantinosKatsamis
 

Computational Fluid Dynamics and Turbulence Mechanics
@ the University of Manchester
Copyright & by the contributing authors. Unless noted otherwise, all material on this web site is the property of the contributing authors.