[sita]

Hi Arash,

As I said, the Oldroyd-B model is not suitable for flows with an elongational component. If you take a look at the equations for the stress, you'll see why. I don't know what your flow geometry looks like, but unless it's something like a straight, constant cross-section channel without anything in it, chances are that your flow contains some elongational component.

Even if you haven't seen any articles that use a FENE model to model a Boger fluid (I did a quick search on "boger fluid fene-cr, and got about 130,000 results, so perhaps you should look again), why not give it a try? Most applications that you'll come across use the FENE-P model, but this shows shear-thinning behaviour, and so isn't suitable to model a Boger fluid. The FENE-CR model was designed to show constant viscosity, which is what you're looking for.

The main reason why FENE models do work for elongational flow, and the Oldroyd-B model doesn't, is in the FE-part of the name: Finitely Extensible. That is, the FENE model limits the extension of the polymer chains, by using a more physically realistic equation for the spring force.

Cheers,
Sita

[arashfluid]

Hi Sita,
Thanks for your complete answer. I'll try it.

[arashfluid]

Originally Posted by sita

Hi Arash,

As I said, the Oldroyd-B model is not suitable for flows with an elongational component. If you take a look at the equations for the stress, you'll see why. I don't know what your flow geometry looks like, but unless it's something like a straight, constant cross-section channel without anything in it, chances are that your flow contains some elongational component.

Even if you haven't seen any articles that use a FENE model to model a Boger fluid (I did a quick search on "boger fluid fene-cr, and got about 130,000 results, so perhaps you should look again), why not give it a try? Most applications that you'll come across use the FENE-P model, but this shows shear-thinning behaviour, and so isn't suitable to model a Boger fluid. The FENE-CR model was designed to show constant viscosity, which is what you're looking for.

The main reason why FENE models do work for elongational flow, and the Oldroyd-B model doesn't, is in the FE-part of the name: Finitely Extensible. That is, the FENE model limits the extension of the polymer chains, by using a more physically realistic equation for the spring force.

Cheers,
Sita

Hi sita
I ran rheoFoam solver with FENE-CRLog viscoelastic model, but the pressure still didn't converge well. Can you explain the parameter L2? As L2 tends to zero, the FENE-CR model is reduced to the Oldroyd-B. How is the value of L2 calculated?

[sita]

Hi Arash,

To be honest, I have no experience with rheoFoam, I used Jovani Favero's viscoelasticFluidFoam for all my viscoelastic simulations. As far as I remember, finding acceptable parameter values was a bit of a trial-and-error process. To simulate a PEG-PEO Boger fluid, I used rho = 1040 kg/m3, eta_s = 0.06 Pa.s, eta_p = 0.8 Pa.s, lambda = 0.0042 s (longest Zimm relaxation time), and L2 = 25. L2 (as in L-squared) is the finite extensibility parameter.

For more information on the choice of parameters, and background information on the models, you can take a look at, for example:

S. Drost. Extrusion instability in an aramid fibre spinning process. PhD thesis, Delft University of Technology, 2015 (pp. 106-107)

M. Herrchen and H.C. Öttinger. A detailed comparison of various FENE dumbbell methods. Journal of Non-Newtonian Fluid Mechanics, 68:17-42, 1997

Hope this helps,
Sita

[arashfluid]

Originally Posted by sita

Hi Arash,

To be honest, I have no experience with rheoFoam, I used Jovani Favero's viscoelasticFluidFoam for all my viscoelastic simulations. As far as I remember, finding acceptable parameter values was a bit of a trial-and-error process. To simulate a PEG-PEO Boger fluid, I used rho = 1040 kg/m3, eta_s = 0.06 Pa.s, eta_p = 0.8 Pa.s, lambda = 0.0042 s (longest Zimm relaxation time), and L2 = 25. L2 (as in L-squared) is the finite extensibility parameter.

For more information on the choice of parameters, and background information on the models, you can take a look at, for example:

S. Drost. Extrusion instability in an aramid fibre spinning process. PhD thesis, Delft University of Technology, 2015 (pp. 106-107)

M. Herrchen and H.C. Öttinger. A detailed comparison of various FENE dumbbell methods. Journal of Non-Newtonian Fluid Mechanics, 68:17-42, 1997

Hope this helps,
Sita

Hi Sita,
Thank you for taking the time.

[NG1803]

Does anyone have a tarball of the full directory of a viscoelastic example? I am having issues trying to figure out the logic for what to change to include something like oldroyd-b.

I can't seem to understand the logic of the headers or find a list of what files need to be altered to include the pointers.

Thanks for any help! Or if some kind soul wants to provide a list of files to alter given that I have extend installed and it appears to be ready to work.

[sita]

Hi Nick,

Are you looking for an example for viscoelasticFluidFoam, or rheoFoam, or a non-Newtonian example in OpenFOAM-7, or something different altogether?

I think I could dig up an example for viscoelasticFluidFoam, please let me know if you'd be interested in that.

Cheers,
Sita

[NG1803]

An example of viscoelasticFluidFoam would be a big help. Doesn’t really matter the momentum equations. I just can’t sort out the logic somehow...

[sita]

Hi Nick,

Attached you'll find a zipped example. It was adapted from on of the viscoelasticFluidFoam tutorials, so I'm also sending a link to Jovani Favero's thesis, which gives more information on the solvers and examples.

Hope this helps,
Sita

[NG1803]

Thank you!

[nskelly]

Hi,

Does anybody know how I would go about doing some LES simulations with the viscoelasticFluidFoam solver?

I think I would have to combine pisoFoam and viscoelasticFluidFoam solvers together, but I am a bit unsure how to do this.

Any help would be greatly appreciated.

[nskelly]

Can I use this solver to perform DNS of a viscoelastic fluid?

Or would I need to modify the solver further?

[mnobrega]

This solver can be used to perform DNS of viscoelastic fluids flow. On the available tutorials you can find several examples.

[nskelly]

Thanks mnobrega.

Are there any high Re cases available that you know of?

I am interested in turbulent viscoelastic flows.

[mnobrega]

I'm not aware of any solver with viscoelastic turbulent models in OpenFOAM main distributions.

[nskelly]

Precisely, I have not much success finding turbulence models that can handle viscoelastic fluids.

This is why I am curious whether DNS is possible using viscoelasticFluidFoam.

[mnobrega]

Notice there are many related publications about that, but as far as I know, nothing is implemented in OpenFOAM yet. Are you planning to do it?

[nskelly]

Yes there are many publications on DNS of viscoelastic fluids, many of which actually use FE over FV. There is very limited publications using OF, and those that are available are only concerned with very low Re.

I am running some tests at the moment to see the feasibility of performing DNS using the viscoelastic solver, and hopefully come out with some preliminary results.

[goswami]

Hello guys,

I am trying to run viscoelastic flow, FENE-P simulations for pipe flow at Re=5000, using DNS. I solved the pressure convergence issue by making a viscoelasticPimpleFoam solver, to go through many inner loops to converge pressure in long pipe. But the issue with foam-extend is that its too buggy. Every solver is designed to run with just few selected schemes, and changes like implementing transitional periodicity is very tough. Also at this Reynolds number the validity of solver is not tested in literature. The examples given in the tutorials are also limited and not as varied as the ESI or openfoam.org versions of OpenFOAMs.

Are any of you guys aware of literature or any references related to viscoelastic pipe flows at this Reynolds number of has anyone tried to do something similar?

Also, has anyone tried and succeeded in implementing a transitional periodicity in pipe flow in foam-extend? with momentum source at inlet?

[hjasak]

Starting a post with an insult isn’t good. Foam-extend is not buggy and I put a lot of effort into doing things right. Perhaps the problem is somewhere else.

For a periodic channel, use pressure jump boundary conditions: this is a much more stable way of dealing with the flow. You change one boundary condition in one file.

Regarding validation and Re numbers, kindly read - and quote - the papers! We make sure the cases used in papers are publicly available and all results are reporoducable.

Hrvoje Jasak