[mpml]

Hi,

First, I would like to thank FOAM developers for this wonderful code. I am looking at the transport models you defined for non-Newtonian fluids, namely, the CrossPowerLaw and the BirdCarreau. I find that both models have 4 parameters in them, nu0, nuInf, m, and n.

According to what I know, the PowerLaw for polymer is expressed as "eta=m*gamma^(n-1)", where eta is the viscosity(unit: Pa.s), gamma is the shear rate(1/s).

In addition, CrossCarreau model is expressed as "(eta-eta0)/(eta-etaInf)=(1+(lambda*gamma)^a)^((n-1)/a)", where eta0 and etaInf are the zero and infinite shear viscosity(Pa.s), gamma is the shear rate(1/s), lambda a time constant, a=2 is a coeffecient, and n is the same as in Powerlaw.

I believe the above models have been transformed into some other forms, so could you explain the meaning of nu0, nuInf, m, and n in OpenFOAM, as opposed to the standard definition?

Also, I find that the sample program for nonNewtonianIcoFoam has a very low material viscosity, do you think it will be a problem if I increase the viscosity to the order of 10^3-10^6?

Thanks very much.

[henry]

To see how these viscosity models are implemented in OpenFOAM take a look at the source code for them in

OpenFOAM-1.1/src/transportModels/incompressible

and you will be able to work out the correspondence of the parameters used with the form that you have them in. If you are not happy with the way these functions are defined in OpenFOAM you can easily reimplement them or any other similar function according to you needs.

Yes you may very well have problems if you increase the viscosity that much. What I recommend is to remove the momentum predictor step and run with as many PISO correctors as is necessary to obtain adequate convergence. With this approach I have been able to run at Reynolds numbers as low as 1e-5.

[mpml]

I found it, thanks a lot.

One more question, has anyone considered simulating polymer injection molding with OpenFOAM? The governing equations for this process are complicated but fine solutions can be extremely valuable, both academically and commercially.

In my case, I am a Ph.D. student of advanced polymer processing technology (microcellular foaming). Injection molding has been simulated with FEM based on the Hele-Shaw model but there the flow front behavior is ignored. I am quite interested in developing a code for research based on FVM, which may track the behavior of flow front and bubble formation. Is there any similar project going on at OpenFOAM?

It would be great if someone can comment on OpenFOAM's potential for injection molding.

[henry]

I have done polymer flow but without free-surface or solidification.

[billy]

Hi Jing Wang,

I am very interested in modelling injection moulding using OpenFOAM. I have done some preliminary tests but nothing specific yet.

[mpml]

Actually I've only studied OpenFOAM for 1 month, but it is quite impressive that OpenFOAM is well structured and easily extensible (the best open source CFD code I've ever known). There is a famous injection molding simulation software called Moldflow, but nobody knows how it's implemented and it is not powerful enough for my research. My plan is to first spend a few weeks researching the potential of FVM technique for injection molding and polymer extrusion.

Billy, can you explain a little bit more about your preliminary test? Did you write a solver for molding simulation? What are your references(papers, books...)? Thanks a lot.

[billy]

I have worked with interFoam because I would like to simulate two phases: polymer and air. I am interested in studying flow defects such as jetting, surface defects and air bubbles. I have been developing my own code to do this using FVM.

I agree with you. OpenFOAM is very impressive and few open sources have such good tutorials and documentation as OpenFOAM. Also this forum is another way of people interacting with their research. I also have been looking at OpenFlower which looks like another good effort to develop open source CFD.

[yeblod]

Anyone still interested in a Hele-Shaw polymer flow model? I'm trying to evaluate interest an open source 2.5d application.

[n_sathya]

Hi I'm interested in using this capability.
please let me know how i can help
i'm trying to use this for 3d injection molding simulation

[chris_sev]

Hallo,

I got problem with nonNewtonianIcoFoam. When I run simulation in parallel (3 nodes with 2 processor on each) intermediate transient results are not beeing written.
Actually they are written but only on my master slave (processor0 and processor1), other processor folder on other machines contains only constant and 0 folder.
As I said simulation runs without any problem. But I don't see transient results.

I don't have such problems running other solvers.

So I would like to find out if nonNewtonianIcoFoam is suittable to run in parallel?

thanks for any answer!!!