drag value not stable even within small geometry changes

krzychu111 · November 26, 2018, 09:04

Hi,
I'm trying to perform calculations on set of axisimmetric geometries which differ only by definition of foci points of ellipse. Foci y dimension is in range -0.25m to 0.15m with step 0.001m so there are 401 cases. Pictures for -0.25, 0 and 0.15 are in attachments.
Regardless using small step differences between geometries and different case settings - diffs between neighbor geometries are quite big and unstable (see graph in attachments).

What I tried so far:

k Epsilon with wall functions and y+ between 30 and 100 - calculated quite fast but unstable results
k Epsilon with y+ under 5 and with wall functions- not much stable
k Epsilon without wall functions (to increase cd calc stability) y+ around 50 - single case calculated by 40mins and unstable results
k Epsilon realizable, without wall functions and y+ around 50 - still unstable results
k Epsilon, without wall functions and y+ around 10 - calculated only one case which took 8h but it will take a month to calculate 400 cases. And I'm still far away from y+ being around 1

Can anyone here help me please with finding out what is making my simulation so complex? And also the source of these big differences in drag result between neighbor geometries.

Enclosed zip with case and gmsh file.

pete20r2 · November 26, 2018, 17:42

Can you show your mesh and a checkMesh report?

krzychu111 · November 26, 2018, 18:02

@pete20r2 thanks for your reply

Enclosed pictures of the mesh.

I was trying to make mesh with quadrilateral elements - unfortunately quad mesh is unavailable in gmsh when boundary of area is not regular.

Current mesh meshCheck:

Code:

/*---------------------------------------------------------------------------*\
| =========                 |                                                 |
| \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox           |
|  \\    /   O peration     | Version:  4.0                                   |
|   \\  /    A nd           | Web:      www.OpenFOAM.org                      |
|    \\/     M anipulation  |                                                 |
\*---------------------------------------------------------------------------*/
Build  : 4.0
Exec   : checkMesh
Date   : Nov 26 2018
Time   : 23:48:38
Host   : "PC"
PID    : 3236
Case   : /home/krzy/Desktop/ost/proba_21/case
nProcs : 1
fileModificationChecking : Monitoring run-time modified files using timeStampMaster
allowSystemOperations : Allowing user-supplied system call operations

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create polyMesh for time = 0

Time = 0

Mesh stats
    points:           145430
    internal points:  0
    faces:            486872
    internal faces:   203377
    cells:            138066
    faces per cell:   4.99941
    boundary patches: 7
    point zones:      0
    face zones:       0
    cell zones:       1

Overall number of cells of each type:
    hexahedra:     0
    prisms:        137891
    wedges:        0
    pyramids:      94
    tet wedges:    0
    tetrahedra:    81
    polyhedra:     0

Checking topology...
    Boundary definition OK.
    Cell to face addressing OK.
    Point usage OK.
    Upper triangular ordering OK.
    Face vertices OK.
    Number of regions: 1 (OK).

Checking patch topology for multiply connected surfaces...
    Patch               Faces    Points   Surface topology                  
    wedge0              138066   72756    ok (non-closed singly connected)  
    front               2710     5421     ok (non-closed singly connected)  
    inlet               11       23       ok (non-closed singly connected)  
    wedge1              138066   72756    ok (non-closed singly connected)  
    back                4577     9156     ok (non-closed singly connected)  
    outlet              42       86       ok (non-closed singly connected)  
    tunnel              23       48       ok (non-closed singly connected)  

Checking geometry...
    Overall domain bounding box (-3.25 0 -0.0981436) (2.5 2.24786 0.0981436)
    Mesh has 2 geometric (non-empty/wedge) directions (1 1 0)
    Mesh has 3 solution (non-empty) directions (1 1 1)
    Wedge wedge0 with angle 2.5 degrees
    Wedge wedge1 with angle 2.5 degrees
    All edges aligned with or perpendicular to non-empty directions.
    Boundary openness (-4.66596e-19 7.00356e-16 -9.34168e-16) OK.
    Max cell openness = 3.16784e-16 OK.
    Max aspect ratio = 3.22455 OK.
    Minimum face area = 8.85554e-09. Maximum face area = 0.0588718.  Face area magnitudes OK.
    Min volume = 1.59329e-12. Max volume = 0.00602443.  Total volume = 1.2402.  Cell volumes OK.
    Mesh non-orthogonality Max: 33.888 average: 7.3683
    Non-orthogonality check OK.
    Face pyramids OK.
    Max skewness = 0.559861 OK.
    Coupled point location match (average 0) OK.

Mesh OK.

End

Older mesh (less cells):

Code:

/*---------------------------------------------------------------------------*\
| =========                 |                                                 |
| \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox           |
|  \\    /   O peration     | Version:  4.0                                   |
|   \\  /    A nd           | Web:      www.OpenFOAM.org                      |
|    \\/     M anipulation  |                                                 |
\*---------------------------------------------------------------------------*/
Build  : 4.0
Exec   : checkMesh
Date   : Nov 26 2018
Time   : 23:50:18
Host   : "PC"
PID    : 3273
Case   : /home/krzy/Desktop/ost/proba_20_run/case
nProcs : 1
fileModificationChecking : Monitoring run-time modified files using timeStampMaster
allowSystemOperations : Allowing user-supplied system call operations

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create polyMesh for time = 0

Time = 0

Mesh stats
    points:           88141
    internal points:  0
    faces:            295069
    internal faces:   123250
    cells:            83679
    faces per cell:   4.99909
    boundary patches: 7
    point zones:      0
    face zones:       0
    cell zones:       1

Overall number of cells of each type:
    hexahedra:     0
    prisms:        83511
    wedges:        0
    pyramids:      92
    tet wedges:    0
    tetrahedra:    76
    polyhedra:     0

Checking topology...
    Boundary definition OK.
    Cell to face addressing OK.
    Point usage OK.
    Upper triangular ordering OK.
    Face vertices OK.
    Number of regions: 1 (OK).

Checking patch topology for multiply connected surfaces...
    Patch               Faces    Points   Surface topology                  
    wedge0              83679    44109    ok (non-closed singly connected)  
    front               1615     3231     ok (non-closed singly connected)  
    inlet               11       23       ok (non-closed singly connected)  
    wedge1              83679    44109    ok (non-closed singly connected)  
    back                2773     5548     ok (non-closed singly connected)  
    outlet              39       80       ok (non-closed singly connected)  
    tunnel              23       48       ok (non-closed singly connected)  

Checking geometry...
    Overall domain bounding box (-3.25 0 -0.0981436) (2.5 2.24786 0.0981436)
    Mesh has 2 geometric (non-empty/wedge) directions (1 1 0)
    Mesh has 3 solution (non-empty) directions (1 1 1)
    Wedge wedge0 with angle 2.5 degrees
    Wedge wedge1 with angle 2.5 degrees
    All edges aligned with or perpendicular to non-empty directions.
    Boundary openness (-1.84021e-18 -8.1401e-16 -2.90374e-15) OK.
    Max cell openness = 2.87837e-16 OK.
    Max aspect ratio = 3.0378 OK.
    Minimum face area = 1.90437e-08. Maximum face area = 0.0569313.  Face area magnitudes OK.
    Min volume = 5.14846e-12. Max volume = 0.00637896.  Total volume = 1.2402.  Cell volumes OK.
    Mesh non-orthogonality Max: 34.55 average: 7.4025
    Non-orthogonality check OK.
    Face pyramids OK.
    Max skewness = 0.560643 OK.
    Coupled point location match (average 0) OK.

Mesh OK.

End

pete20r2 · November 26, 2018, 19:08

Aw man, you really don't want tets in the near wall region. Especially in low Re turbulence models. I'd recommend you remesh to using quads, you'll need to break your region down into more zones and mesh them separately. If you can mesh with a program that can generate a quad boundary layer that would be way better, you can use polyDualMesh to convert tets to poly if they are Delaunay but only for far field mesh, not wall bounded.

krzychu111 · November 27, 2018, 09:07

Ok, so I will try to create hex mesh by using SnappyHexMesh.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Calculation of lift and drag coefficients on airfoil	CoolHersheys	OpenFOAM Post-Processing	5	September 27, 2021 06:04
Unrealistic value for drag coefficient	tw463	Main CFD Forum	13	May 10, 2020 15:59
[ICEM] O-grid around wing geometry	Anorky	ANSYS Meshing & Geometry	15	November 27, 2017 13:43
Drag coefficient_cylinder	B.Stas	FloEFD, FloWorks & FloTHERM	1	August 26, 2014 07:27
CFD analysis of geometry with very small inlets	tarkesdora	Main CFD Forum	4	July 2, 2014 13:48

November 26, 2018, 17:42		#2
pete20r2 Senior Member Peter Baskovich Join Date: Jul 2014 Posts: 127 Rep Power: 11	Can you show your mesh and a checkMesh report?

November 26, 2018, 19:08		#4
pete20r2 Senior Member Peter Baskovich Join Date: Jul 2014 Posts: 127 Rep Power: 11	Aw man, you really don't want tets in the near wall region. Especially in low Re turbulence models. I'd recommend you remesh to using quads, you'll need to break your region down into more zones and mesh them separately. If you can mesh with a program that can generate a quad boundary layer that would be way better, you can use polyDualMesh to convert tets to poly if they are Delaunay but only for far field mesh, not wall bounded.

November 27, 2018, 09:07		#5
krzychu111 New Member Join Date: Oct 2018 Posts: 19 Rep Power: 7	Ok, so I will try to create hex mesh by using SnappyHexMesh.