CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Injection of Lagrangian particles (http://www.cfd-online.com/Forums/openfoam-solving/82306-injection-lagrangian-particles.html)

alberto November 23, 2010 03:28

Injection of Lagrangian particles
 
Hello,

I have a doubt about the Lagrangian classes. I would like to simulate the injection of two jets of particles using the O'Rourke collision model.

My case is extremely simple: particles are injected from two patches at constant velocity and constant diameter, and they can collide.

It seems collision models is only available using the spray sub-models, which adds the complication of using injectors (there does not seem to be an injector which uses a patch, at least if my understanding is correct).

Could you give some hint to set this case up?

Thanks

sradl November 23, 2010 19:33

Hey Alberto!

I did some LPT using injections in a region (e.g., a sphere), but not in a domain. I imagine one could write a routine do inject particles near an inlet.

I'd be interested to work on an implementation of the MP-PIC approach of O'Rourke into twoPhaseEulerFoam. We have started a project on that here recently, and would love to share code of LPT (coupled to LES, using an eddy interaction model)

cheers
Stefan

alberto November 23, 2010 22:38

Hi Stefan,

thanks for your answer. I guess definedInjector can do that. My case is a pure Lagrangian simulation without fluid, so I do not really need an inlet, just an area next to a wall from where particles come out. I was trying to find the easiest way to do this, since injectors want a whole bunch of information I do not really need! :-)

About MP-PIC, you probably need a more robust solver for dense flows than twoPhaseEulerFoam, since the key in some of the recent implementations of MP-PIC is enforcing the particle packing by means of the Eulerian equations, to evaluate the acceleration term responsible of preventing packing implicitly.
I saw you joined my project on extend, so you know I am working on that ;-)

What kind of applications would you like to do with MP-PIC (I saw you are in Prof Sundaresan's group, so we might have a chance to meet sometime).

Best,

AMahrla November 24, 2010 05:04

Hi Alberto!

There is a PatchInjectionModel in OpenFOAM..
It's used in the solver rhoPisoTwinParcelFoam with KinematicParcel class and can be found in src/lagrangian/intermediate/submodels/Kinematic/InjectionModel/PatchInjection..

Hope this helps!

Best,

Astrid

sradl November 24, 2010 09:57

Hey Alberto,

thanks for the useful input. Yeah, I'm aware of the stability issues associated with MP-PIC - Snider is working on that for more than 10 years...

I think the best way is to have a stable and two-fluid solver, and knowing how to stabilize the particle pressure calculation near the packing limit. I'd be glad to contribute to this development in the twoPhaseEulerPimpleFoam code.

Are there already implementations of MP-PIC in OF available to have a quick look at that?

My goal is compare MP-PIC with E-E based on filtered simulations (similar to the work of Igci et al.), and then go to CFB.

cheers
Stefan

alberto November 24, 2010 11:16

Quote:

Originally Posted by sradl (Post 284615)
Hey Alberto,

thanks for the useful input. Yeah, I'm aware of the stability issues associated with MP-PIC - Snider is working on that for more than 10 years...

Yes, I do not have a hands-on view (never coded anything related to MP-PIC), but I think their issues are related to the exchange between Euler-Euler and Euler-Lagrange framework, which forces them to use quite strong relaxation. A similar problem was discussed at ICMF last year by a "famous commercial code vendor", which is implementing MP-PIC under another name ;)

Quote:

I think the best way is to have a stable and two-fluid solver, and knowing how to stabilize the particle pressure calculation near the packing limit. I'd be glad to contribute to this development in the twoPhaseEulerPimpleFoam code.
Thanks :)

Quote:

Are there already implementations of MP-PIC in OF available to have a quick look at that?
Not that I know. I remember seeing some thread talking about MP-PIC on this forum, but it was about intentions of implementing it.

Quote:

My goal is compare MP-PIC with E-E based on filtered simulations (similar to the work of Igci et al.), and then go to CFB.
That's interesting!

Best,

alberto November 24, 2010 11:17

Quote:

Originally Posted by AMahrla (Post 284569)
Hi Alberto!

There is a PatchInjectionModel in OpenFOAM..
It's used in the solver rhoPisoTwinParcelFoam with KinematicParcel class and can be found in src/lagrangian/intermediate/submodels/Kinematic/InjectionModel/PatchInjection..

Thanks Mahrla :)


All times are GMT -4. The time now is 01:47.