[grandgo]

hello francois,

thanks a lot for you support, i will try what you've suggested!

may i ask you for another favour? could you please take a look into my post "Adding a new member function in src/../turbulenceModel" ?

you could help me there big time.....

best regards
grandgo

[N. A.]

Hey Folks,

I am wondering if in OF1.7, the way we extact mean values is by specifying following in controlDict file:
U
{
mean on;
prime2Mean on;
base time;
}

I want to confirm if this is:
Umean = time average of U at each cell ....(1)

or
Umean = time average of (U+ Uprime); ....(2)
where Uprime is the sub-grid scale part.

I want to get a average value using 2nd option to compare with time-averaged experimental data.

Can comeone confirm what is OF1.7 doing with this mean function in controldict?

If it is not doing 2nd option, how can I extract the mean using the second equation.

Thank you,
Nir

[eugene]

Usgs or Uprime as you call it is an isotropic vector, so averaging it would result in a mean field of zero.

[sandy]

Originally Posted by panda60

Dear Fransje,

I am sorry, I haven't explain clearly.
I run the case in the directory "tutorials/........./incompressible/les/pitzDirectMapped".
I have tried 3 kinds of inflow conditions:
1. directMapped (nearestCell).
2. directMapped (nearestFace).
3. directMappedVelocityFlux (nearestFace).

I use the solver "pisoFoam".The following is the detailed information.

Hi,

Could you explain detailed, what else are the differences between directMappedVelocityFlux and directMapped, nearestCell and nearestFace, setAverage=true and setAverage=false ? In addition, offset is just a distance but not a coordinate value from the inlet plane, right?

[sandy]

Hi Tabor,

In your paper(Inlet conditions for LES using mapping and feedback control), you did not refer to "directMapped" method. Could you explain what are differences between directMapped with directMappedVelocityFlux ?

[sahm]

Hi Everyone,
I was reading about LES turbulentInlet boundary condition. It looks like this boundary condition does not have ability to read profile for U_mean and also U_rms, which can be used as a fluctuation scales. But it uses flat profile for both.
Besides, I know there's a timeVaryingMappedFixedValue field that reads profiles for boundary conditions.
My question is is there any boundary condition that reads profiles for mean velocity and their fluctuation scale (U_rms) and uses them to make a turbulent inlet for U? and If there's none, how can I make one? I know some C++ and some basics of openfoam, but it gets hard when it comes to openFOAM source code, as it has complicated structure.

[vitvit]

Hello,
I want to run a combustion simulation with swirling inlet conditions.
How do I define the initial pertubations?
As I understand it,directMapped is not suited for this purpose.
I know that Dr.Tabor worked on this subject with his Phd student (now Doctor).
Did someone implemented this method?
What should I do?
Did someone used Synthesised turbulence methods in swirl flows? (I'd rather to use the Mapped method, but beggars can't be choosers...)

Thank You!

[valikiut]

I want to simulate dual fuel diesel engine with OpenFOAM. can I simulate this with dieselEngineFoam? If not, which solver I must use?
If I can use dieselEngineFoam, can I use (CH4) mass fraction in initial condition such as O2 and N2?

Thanks.
Best Regards,

[ehsan]

Hello

I had the problem of under resolved solution of my LES simulation of cavitating hydrofoil flow. Is there any remedy for overcoming under resolved solutions? Is it filtering?

Thanks

[gascortado]

Hi everyone. I am really desperate trying to find a way out of this problem using directmappedpatch. Please take a look. Any suggestion will be appreciated. Thanks

I'm trying to model a fluid structure interaction using LES. I generate the turbulence at the inlet implementing a recycling technique (similar to that of Lund et all) using directMapped BC. Now, my mesh has to be able to deform, however when I run my case I get:

Create time

Create dynamic mesh for time = 0

Selecting dynamicFvMesh dynamicMotionSolverFvMesh
Selecting motion solver: laplaceFaceDecomposition


--> FOAM FATAL ERROR:
Unknown faceTetPolyPatchFaceDecomp type directMappedPatch. Valid faceTetPolyPatchFaceDecomp types are :

7
(
wall
processor
wedge
patch
empty
ggi
symmetryPlane
)


From function faceTetPolyPatchFaceDecomp::New(const polyPatch&, const tetPolyBoundaryMeshFaceDecomp&)
in file tetPolyMeshFaceDecomp/tetPolyPatches/faceTetPolyPatch/newFaceTetPolyPatchFaceDecomp.C at line 59.

FOAM exiting

Does this mean that deformable meshes cannot be used with DirectMapped Bc's? Do I have to set three different domains and link them together? (one recycling domain, one fluid domain with deformable mesh and one solid domain). Is there a way to use directmappedpatch with deformable meshes???

I will appreciate any help or suggestions. Thanks

[hjasak]

There is no reason why a direct mapped condition would be a problem: this is to do with the way constrained patches are created on fields. Some minor coding required, ie. you need a simple tet-fem wrapper for the direct mapped patch and fields.

Hrv

[gascortado]

I really, really appreciate you reply.

I'm glad to hear that directMappedPatch should not be a problem, however, would you be kind enough to be a little more specify about

you need a simple tet-fem wrapper for the direct mapped patch and fields

Pardon my C++ ignorance in this case but, what do you mean by "tet-fem wrapper"? And in which file should I look into? Maybe into newFaceTetPolyPatchFaceDecomp.C?

Again I will be very grateful if you can give me any other hint. Thanks again

[gascortado]

@ Bruno,

Here I uploaded my solver and my case. The meshes are greatly simplified in order to make the folder small enough to be uploadable. The case as it is generates the error mentioned above. Please let me know if are able to run it. Thanks a lot

The file my_PisoFSI is the solver and the other one is the case.

[wyldckat]

Greetings to all!

@Santiago: I'm not able to run you case because I get the following error when running changeDictionary in the "fluid" folder:

Create mesh for time = 0



--> FOAM FATAL ERROR: 
Did not find sample (0.5 0.075 0.095) on any processor of region region0

    From function directMappedPatchBase::findSamples(const pointField&, labelList&, labelList&, pointField&)
    in file directMapped/directMappedPolyPatch/directMappedPatchBase.C at line 373.

FOAM aborting

I got that error because I used the Allrun script you provided in the case "workstationMappedBruno".

Best regards,
Bruno

[gascortado]

Hi Bruno.

Thats strange. I run it here and i get

Create time

Create mesh for time = 0

Read dictionary changeDictionaryDict with replacements for dictionaries
1
(
boundary
)

Replacing entries in dictionary boundary
Special handling of boundary as polyMesh/boundary file.
Loaded dictionary boundary with entries
8
(
plateFixed
plateFluid
right
left
outlet
upperWall
lowerWall
inlet
)

Merging entries from
1
(
inlet
)

fieldDict:
{
plateFixed
{
type wall;
nFaces 5;
startFace 64;
}
plateFluid
{
type wall;
nFaces 5;
startFace 69;
}
right
{
type wall;
nFaces 12;
startFace 74;
}
left
{
type wall;
nFaces 12;
startFace 86;
}
outlet
{
type patch;
nFaces 9;
startFace 98;
}
upperWall
{
type patch;
nFaces 12;
startFace 107;
}
lowerWall
{
type wall;
nFaces 12;
startFace 119;
}
inlet
{
type directMappedPatch;
nFaces 9;
startFace 131;
offset ( 0.5 0 0 );
sampleRegion region0;
sampleMode nearestCell;
samplePatch none;
}
}

Writing modified fieldDict boundary

End

Please replace the changeDictionaryDict with the one attached. (un-tar first). This file must be placed in "worstationMappedBruno\fluid\system". I just changed the sample location to 0.4 instead of 0.5. At 0.5 the fluid mesh has a void and that maybe causing an error. Please let me know the outcome. Thanks a lot for trying

[wyldckat]

Hi Santiago,

OK, I've made some progress. Using the old 1.6-ext deb package from the CAE-Team on Ubuntu 10.04, along with the swak4Foam also available with it, I've managed to reproduce your error.

I've also started working on an independent add-on library(ies) (one for faces and another for cells, wrapped into a single source code folder), which I've attached the code for it.
Nonetheless, it's not building yet. It did build with the sample folder for the "timeVarying*" BC, but I picked a bad example, since this one uses a type of data structure different from the "directMapped*" one.

I'm attaching it already just in case someone else wants to step in and help!

The idea is that after the libraries are built and possibly working, they are added to the libs list at the case's "controlDict".

Best regards,
Bruno

[wyldckat]

Hi Santiago,

OK, I've made some advances in the research on how to wrap up "directMappedPatch" for tet, but don't have it working yet. It basically boils down to this:

1. Let's first look at the "wedge" patch. Attached is the previous base preparation, now adapted for reimplementing the "wedge" patch.
2. The "wedge" is split into 4 major components:
   1. Face wrapper for "wedgePolyPatch":

      tetPolyMeshFaceDecomp/tetPolyPatches/constraint/wedge

   2. Cell wrapper for "wedgePolyPatch":

      tetPolyMeshCellDecomp/tetPolyPatches/constraint/wedge

   3. Does the element connection:

      fields/elementPatchFields/constraint/wedge

   4. Does the point connection:

      fields/tetPolyPatchFields/constraint/wedge

3. "a" and "b" are easy to implement. It should be as simple as adapting for wrapping around "directMappedPolyPatch" ("meshTools/directMapped/directMappedPolyPatch") and replace "wedge" by "directMappedPatch" where needed.
   "a+b" handle the mesh side of business.
4. "c" and "d" possibly are going to have to be wrappers around:
   • the "genericPolyPatch" ("OpenFOAM/meshes/polyMesh/polyPatches/basic/generic")
   • or "directMappedFixedValue" ("finiteVolume/fields/fvPatchFields/derived/directMappedFixedValue")
   "c+d" seem to handle the field/matrix side of business, and seem more complicated. The reason is because this seems to be where the solution matrices are hooked up between the FVM and FEM portions of the code.

I'm itching to try and figure out how to finish connecting the dots, but that's all the time I have for today. I certainly hope I can scrape some time during this week, but only time will tell

edit: for using the latest 1.6-ext version, I had to change the "changeDictionaryDict" file on the fluid side from "(0.5 0.0 0.0)" to "(0.699 0.0 0.0)".

Best regards,
Bruno

[wyldckat]

Hi Santiago,

Sadly, I'm at the limit of my experience with OpenFOAM source code Attached is what I've got now.

I had to modify the case file "system/controlDict" by adding "libfaceDecompTetDirectMapped.so" to the libs list.

The error I get when running the solver is this:

--> FOAM FATAL ERROR: 
Attempt to cast type wall to type lduInterface

    From function refCast<To>(From&)
    in file /home/username/OpenFOAM/OpenFOAM-1.6-ext/src/OpenFOAM/lnInclude/typeInfo.H at line 115.

FOAM aborting

I had based myself on the "wedge" boundary condition for reference, but sort-of found out that "makeTypePointPatchFieldTypedefs" couldn't be used and had to use "makePointPatchFieldTypedefs".

The other issue was that I couldn't create a proper "directMappedElementPatchFields.C", but I don't know if this is what is missing and/or wrong...

Good luck! Best regards,
Bruno

[kanuk]

Hi Bruno,

Have you had any luck in implementing your direct mapped BC?

I am really trying to implement a cyclic BC with the laplaceFaceDecomposition solver in 1.6-ext. It appears you are on the right track - though I too have limited c++ experience. Regardless, any insight you or anyone else out there has about writing the wrapper, I'd appreciate. I'll post back if I have any luck...

Cheers,
Dave

[wyldckat]

Hi David,

Unfortunately I was not able to figure out how it works and how it can be done
I don't know if Santiago managed to figure out some other solution in the mean time.

Best regards,
Bruno