CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > OpenFOAM > OpenFOAM Running, Solving & CFD

Non convergence for a cyclic Geometry/values for Spalart-Allmaras model

Register Blogs Members List Search Today's Posts Mark Forums Read

Like Tree4Likes
  • 1 Post By RodriguezFatz
  • 1 Post By RodriguezFatz
  • 1 Post By RodriguezFatz
  • 1 Post By RodriguezFatz

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   October 9, 2015, 17:51
Red face Non convergence for a cyclic Geometry/values for Spalart-Allmaras model
  #1
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
I am trying to simulate 3D air flow inside the vanes of a closed ventilated rotor.
I am using SRFSimpleFoam/Mixer as my sample case, and was able to run a successful initial simulation with my cyclic geometry, with slight modifications in the files from the tutorial case.
However,when I tried to check the convergence for my initial variables, p,Urel,epsilon, and omega, I hardly see any convergence in the results.

Initially I checked the mesh, and mesh seems fine. Skewness0.79) and non-orthogonality Max:56.2308, avergae:16.9368

Now, I as am trying different settings in the system/fvSchemesand system/fvSchems, I plan to take it to to next step by moving to Spalart-Allmaras model(Using as template:http://rapidof.com/centrifugal-pump-simulation/). After looking through a lot of forums, I was able to find the variables like nu,nut and nuTilda, however, I still cant find the values for DnuTildaEff,DepsilonEff,DkEff and nuEff. ... I have looked through a lot of tutorials, to find any files I could use as a template(I plan to not create the input file from scratch , since i am new, and I done know what class types should I use for my boundary layers).....

Does anybody by any chance know, where can I find a sample tutorial files with those values, or if you could forsee any other issue that could cause divergence?



--------------------------------------------------------------------------------
I am bit new to OpenFoam, so I might have missed some documentation
arpitpatel is offline   Reply With Quote

Old   October 12, 2015, 10:00
Default
  #2
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
I found the tutorials for turbulence models in the $FOAM_SRC folder.
However, they only contain the .H and .C files. Is there any way i could figure out what is the class types of my input variables. Is there anywhere, where I can get a simple turbulent 1D or 2 equation tutorial models?

And the second question about improving the convergence of my model still remains.
arpitpatel is offline   Reply With Quote

Old   October 12, 2015, 12:41
Default
  #3
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
Solved the first issue
Only the second issue about imporving the covergence remains
I am playing around with fvSchemes currently, is there any other thing that could be wrong
arpitpatel is offline   Reply With Quote

Old   October 13, 2015, 02:45
Default
  #4
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,297
Rep Power: 26
RodriguezFatz will become famous soon enough
Arpit, I don't know if that is suitable for your problem, but I had a problem with cyclic as well. I was trying to get a simple periodic pipe or channel flow running and could not get any convergence, no matter what I set in fvSchemes and fvSolution. This was driving me insane. I figured out that I can do exactly what I want with "mappedPatch" boundaries, namely directing the outflow of one boundary to the inflow of a second boundary. For the case "piece of pipe" with periodic boundaries, I modified the "boundary" file in the "constant\polyMesh" directory:

Code:
/*--------------------------------*- C++ -*----------------------------------*\
| =========                 |                                                 |
| \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox           |
|  \\    /   O peration     | Version:  2.3.x                                 |
|   \\  /    A nd           | Web:      www.OpenFOAM.org                      |
|    \\/     M anipulation  |                                                 |
\*---------------------------------------------------------------------------*/
FoamFile
{
    version     2.0;
    format      ascii;
    class       polyBoundaryMesh;
    location    "constant/polyMesh";
    object      boundary;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

3
(
    PER_PIPE_INLET
    {
        type            mappedPatch;
        samplePatch    PER_PIPE_OUTLET;
        sampleMode    nearestPatchFace;
        nFaces          3060;
        startFace       170068;
        offsetMode uniform;
        offset (0.048 0 0);

    }
    PER_PIPE_OUTLET
    {
        type            patch;
        nFaces          3060;
        startFace       173128;
    }
    PER_PIPE_WALL
    {
        type            wall;
        inGroups        1(wall);
        nFaces          2584;
        startFace       176188;
    }
)

// ************************************************************************* //
"offset" is the length of my pipe and of course you need to use all the numbers (nFaces...) of your own boundary file.
In your \0 boundary files you must set mapped as type:
Code:
/*--------------------------------*- C++ -*----------------------------------*\
| =========                 |                                                 |
| \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox           |
|  \\    /   O peration     | Version:  2.2.1                                 |
|   \\  /    A nd           | Web:      www.OpenFOAM.org                      |
|    \\/     M anipulation  |                                                 |
\*---------------------------------------------------------------------------*/
FoamFile
{
    version     2.0;
    format      ascii;
    class       volScalarField;
    object      k;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //



dimensions      [0 2 -2 0 0 0 0];

internalField   uniform 2;

boundaryField
{
    PER_PIPE_INLET
    {
    type            mapped;
    value           uniform 1e-12;
    //interpolationScheme cell;
    setAverage false;
    average 1e-12;
    }

    PER_PIPE_OUTLET
    {
         type inletOutlet;
        inletValue uniform 1e-12;
        value uniform 1e-12;
    }

    PER_PIPE_WALL
    {
        type            kqRWallFunction;
        value           uniform 1e-12;
    }
}


// ************************************************************************* //
I hope this works.
Cheers, Philipp
arpitpatel likes this.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   October 15, 2015, 12:12
Default
  #5
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
Hi Philipp

Thank you, taking your time out to review my posts.
I am currently facing a bigger problem than convergence. Although , my residual is 0.002 > (desired based on forum recommendations= e-4),I am not too worried about it.
However, my results from the openfoam are completely off, and I feel I have inputted wrong boundary condition somewhere .
If you could please take some more time, and review my case, where I went, I would really appreciate it.
I tried your settings with the type mappedFace(Settings 2), but nothing has changed.

Please follow the below link, which describes my case files and results
https://www.dropbox.com/sh/js5zmei0e...EhRP0OwDa?dl=0

MY CASE
I am trying to do a CFD on a car rotor, especially the air flow inside the vanes of a ventilated rotor. I am rotating the rotor anticlockwise, such that it acts as a pump, and pushes the air out.
I am trying to model it as a stationary rotor, and moving fluid around it.
There is no forced flow, everything is induced by the angular velocity of the rotor.

Thanks a lot for the help .... , unfortunately , I am still stuck
arpitpatel is offline   Reply With Quote

Old   October 16, 2015, 11:09
Default
  #6
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
Attached is an excel sheetBoundary conditions.xls detailing my boundary conditions.
I have tried changing the direction of the rotation(angular velocity) in SRFProperties, but my flow is still flowing inwards as compared to outwards(out of the vane like an impeller).

Last edited by arpitpatel; October 16, 2015 at 14:51.
arpitpatel is offline   Reply With Quote

Old   October 19, 2015, 04:02
Default
  #7
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,297
Rep Power: 26
RodriguezFatz will become famous soon enough
Arpit when I look into the "settings_1" fvSchemes I see very bad schemes for the first run.
Try this:
Code:
divSchemes
{
    default          bounded Gauss upwind;
    div((nuEff*dev(T(grad(Urel))))) Gauss linear;
}
laplacianSchemes
{
    default         Gauss linear uncorrected;
}
The log file also shows that you interrupted the calculation at pretty high residuals. You should let it run unti the residuals don't drop any longer.
arpitpatel likes this.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   October 19, 2015, 14:53
Default
  #8
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
Hi Philipp
I changed my fvSchemes settings based on what you recommended and also increased my number of iterations until my residuals were mostly steady, however my results are still not correct.
I am modelling this blade as an impeller blade, and no matter what what rotation I specify my velocity vectors are pointing inwards(towards the center) as compared to outwards(as it should).
You can see a few pictures of my results in "Settings1_Results " at the same link
I think there is something else wrong, which I cant figure out.

I feel my setup is somewhere wrong, and does not capture the necessary physics. I dont how,or what is wrong
arpitpatel is offline   Reply With Quote

Old   October 19, 2015, 15:29
Default
  #9
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
Hey guys
I am really stuck on simulation and would appreciate any help or suggestions you could provide.
I am trying a CFD simulation of a ventilated car rotor. I am trying to run it as am impeller , such that the flow of the air is outwards(away from the center). However , my simulation results indicates the direction of velocity vectors to be inwards(I tried reversing the direction of my rotation, but the simulated flow is still inwards). I am trying to use the frozen rotor approach where,my cyclic geometry is stationary and there is air all around it(which is moving based on a rotating frame of reference) , which .Thus, I believe there is something wrong with setting up my problem.My original post can be found here

http://www.cfd-online.com/Forums/ope...tml#post569114

Link to my result and setup folder is-https://www.dropbox.com/sh/js5zmei0e703ptm/AAAwMqsDmycfVFMwEhRP0OwDa?dl=0

Thanks for all the help you could provide.
Arpit
arpitpatel is offline   Reply With Quote

Old   October 20, 2015, 05:25
Default
  #10
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,297
Rep Power: 26
RodriguezFatz will become famous soon enough
Arpit, your numerical settings are not good.
You stop at t=0.3 and don't see what happens after that. You should (at the beginning) always set an incredibly high endTime, such as 1e9 and let the simulation run until all residuals settled down.
Then you will see, that your fvSchemes are not able to obtain a convergent solution. Try these settings for a safe run:
Code:
divSchemes
{
    default         bounded Gauss upwind;
    div((nuEff*dev(T(grad(Urel))))) Gauss linear;
}

laplacianSchemes
{
  default         Gauss linear uncorrected;
}
I think your boundary conditions are wrong. Look at the Urel b.c. "wall_0" is a wall and inlet_0 is an inlet and they have both exactly the same b.c. I don't know enough about this solver but for me that doesn't make any sense.
arpitpatel likes this.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   October 22, 2015, 16:01
Default
  #11
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
Thanks Phillip
I changed my boundary conditions and increased my number of iterations and solution worked. At least the cyclic geometry is working as an impeller(throwing air away from the center) as it is supposed to.
Thanks a lot.

However, if you happen to know how to create streamlines from inlet(a boundary patch) as compared to a point, let me know.
I followed up this link - http://www.cfd-online.com/Forums/ope...seed-type.html , but am having trouble doing the step number one. Since my inlet is a curved plane, it is hard for me to specify an accurately enough. Currently i am using a point in middle of my inlet as a point source seed and playing around with seed radius to cover my geometry. My current streamline file is as shown(i wished I could specify the direction of the streamline flow and change it from horizontal to along the curved geometry)

Run4.jpg

Thanks a lot, btw for your responses
arpitpatel is offline   Reply With Quote

Old   October 23, 2015, 03:55
Default
  #12
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,297
Rep Power: 26
RodriguezFatz will become famous soon enough
I never did this, sorry.
arpitpatel likes this.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   October 23, 2015, 11:46
Default
  #13
New Member
 
Arpit
Join Date: Oct 2014
Posts: 11
Rep Power: 11
arpitpatel is on a distinguished road
Cool.
Hopefully, somebody else might be might be able to help me out.
Thanks a lot, though
arpitpatel is offline   Reply With Quote

Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Why Menter's SST model low-Re issue has not been seriously investigated? vkrastev OpenFOAM 58 January 8, 2018 16:20
Spalarat - Allmaras turbulence model saisanthoshm88 Main CFD Forum 1 June 16, 2014 17:33
Problem: Convergence in unsteady EBU combustion model Defend STAR-CCM+ 7 April 22, 2014 15:48
about Subgrid-scale model impecca OpenFOAM Running, Solving & CFD 4 December 20, 2013 11:36
convergence in particle transport model Jose Tinoco CFX 0 March 3, 2003 04:10


All times are GMT -4. The time now is 01:42.