[chrisadeniyi]

Hi everyone,
for my dissertation, i'm basically trying to model the API filter test using star ccm. I am unsure of how to model the filter process. I have tried using porous baffle interface and creating three separate pipes with the middle one as a porous medium.
Is star ccm the best software to do this? as i also have Ansys. i do not know how to use it yet but im learning to. Any help, tips will be much appreciated.

[me3840]

You will have to be more descriptive, I don't know what the API filter test is.

[steve lee]

What's your goal of simulation? you have to explain more detail.

I think porous baffle function is the same in CCM+ and Fluent.
It means software is not important.

[chrisadeniyi]

Thanks guys for your reply.

Basically, I am trying to model the creation of filter cake. So I’ve built a 3-part cylinder making one the inlet where fluid will be injected, the second part is where I want the filtration to happen and the third part will be the filtrate. But I don’t know how to make the second part (middle part) into a filter or porous medium which will trap the particles and show filter cake build up.

[me3840]

This sounds like a very complex analysis. I haven't done something like that before, probably the closest thing would be a packed bed reactor.

The way I see it you have two options. One is to make a mathematical model representing a passive scalar which "freezes" under some kind of momentum or position condition which then acts as a porous source term once frozen. The other method is to use DEM or lagrangian to explicitly model the particles, and you will have to come up with some mathematics to determine when they bond together.

This is not an easy application, you should really look into journal papers to find out how others have done this problem.

[steve lee]

In CCM+, you should define filter's specification for using porous media.
Porous media doesn't mean just material with hole.
You should know why and when use porous media.
Porous media usually used in heat exchanger (when you know about heat gain or loss in the porous region) or filter (when you know about pressure drop or Cf value).

I have no information to solve other kind of porous media problem, but as i know ccm+ can't help you

If you want real physics, you have to do direct modeling (make real porous cake with material specification) with lagrangian particles.. but it's really heavy

[steve lee]

If i were you, I will make just small part of whole model of porous cake.
(If you have much computation power, the model can be bigger but don't need too big model. But you should modeling real size of holes)

and simulate your first designed model with lagrangian particle(not use porous media), and than simulate again with second model(change CAD). Even though your model is just part of whole model, you can judge which one is better.

[fluid23]

If you are trying to model the actual filtration process you will have to do a very small section of media. The trick is setting up particle properties and physics models to capture all the different filtration modes (imp action, interception, sieving, diffusion and electrostatic). This requires resolving individual fibers or pores in your media, which is very computationally expensive. I doubt you could do this on an industrial problem without throwing some very serious power at it. I think you would also probably have to throw a DNS solver in the mix... I've never actually attempted it and I work for a major filter OEM. Its just not a practical exercise. That's not to say it's never been done, but I doubt its been done on the scale you describe.

The porous medium is good for setting pressure drop and thermal properties, but is sounds like that wouldn't get the job done for you. You could at least approximate the change in pressure drop over time by setting up a time dependent field function for the porous resistance values.