[JM27]

Hello,

I am running a simulation using compressibleInterFoam for laminar flow and I would like to calculate the wall-shear stress on the wall patches in my case.

I have successfully managed to implement the wallShearStress function, however after looking through the forum I feel confused whether the results I am getting are correct. A few of my doubts are below:

1) Some previous posts seem to claim that the wallShearStress function cannot be used for multiphase solvers like interFoam as it yields incorrect results

2) the components of wall shear stress to use. In my case the wall boundaries are aligned with the coordinate axis so I take the WSS component parallel to this boundary. Is this correct?

3) My results seem to be underestimating the wall shear stress on the wall patches, but I have checked the units of wall shear stress and these are kg/(ms^2) which is equivalent to Pascal, so I do not think it is a matter of multiplying by density. Despite the dimensions for WSS are in Pa, after looking through the forum I am thinking that perhaps this utility does not work for multiphase solvers?

Is there any other way I can calculate the wall shear stress in my simulations?

Any comments or suggestions would be greatly appreciated. I have already looked through existing threads but I wasn't able to find someone using a compressible multiphase solver like myself.

p.s. I am using OpenFOAM 5

Please help

[JM27]

Hello again,

I'm really struggling to find answers to these questions online, so if someone can share some knowledge it would be greatly appreciated

[HPE]

1) `wallShearStress` function object does not know whether the solver is multiphase or not. It nominally works with multiphase flows, though there might be a bug in the worst case.

2) yes.

3) Though I am not sure `wallShearStress` can work for `laminar` flows wherein there is no Reynolds stress tensor, which is required in wall shear stress computation, and in `wallShearStress` function object implementation. `wallShearStress` only computes the inner product between `R` and `n`, Reynolds stress tensor and wall normal vector. The shear-stress symmetric tensor field is retrieved from the turbulence model, which seems not to be available in your case.

[JM27]

Hi HPE,

Thanks for your replies.

1) I was hoping someone who previously used this function would confirm whether there is a bug or otherwise, hence my post in this forum.

2) Great!

3) With regards to number 3 and the turbulence model. In the log file I have:

Reading g

Reading hRef
Caculating field g.h

Selecting turbulence model type laminar
Selecting laminar stress model Stokes

so it looks, at least, like the solver is looking up a dummy model for laminar flow to compute the shear-stress symmetric tensor you mentioned.

[JM27]

Hello again,

Unfortunately I am still having trouble in comparing my WSS results. After looking extensively through this forum I have some doubts with regards to the usage of the wallShearStress function (many posts claim this in fact computes wall traction rather than shear stress at the wall e.g. here Measuring wall shear stress in bend pipe).

Furthermore, the wallShearStress function only computes the vector directly on the boundary, whereas I would like to compare this 'shear stress' value at various distances from the wall. So, essentially, I would compute the shear stress for the internalField also not just on boundaries.

I've read about the possibility of using grad(U) to do this, but:

1) I am still unclear as to how I would compute the actual shear stress from grad(U)

2) I'm not sure how to interpret the tensor (9 components) that this prints in each time directory of the solver

Can someone please share their experience on how to do this?

[HPE]

>> So, essentially, I would compute the shear stress for the internalField also not just on boundaries

The Foundation OpenFOAM has shearStress function object. Might help?

[JM27]

Hi HPE,

Thanks for the tip. The shearStress functionObject looks appropriate, however, it seems its only available in the development version of OpenFOAM? https://cpp.openfoam.org/dev/dir_bae...61920a91f.html

I only have accesss to v5 and v6 and sadly that does not seem to be available for those versions. https://cpp.openfoam.org/v5/dir_bae0...61920a91f.html

~~

Since my last post I've made a bit of progress on using grad(U) and understanding the tensor components, so I'm thinking if I compute the shear stress as:

tau = - mu * grad(U).yx();

[no need to multiply by rho since solver is compressible]

should give me the shear stress ?

where grad(U).yx() is equivalent dU_x/dy to match the theoretical definition of the wall shear stress in textbooks.

~

I've also come across this definition on the forum:

volTensorField stressTensor = - mu * (fvc::grad(U) + (fvc::grad(U)).T());

but it's not yet clear to me why the transpose of grad(U) needs to be added to grad(U) as well?

~~

Not sure whether you can help on this front but it looks like my computation will need to be a bit more fundamental since I don't have access to the shearStress functionObject. It's also good to make sure I understand what the computation does exactly