CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Perforated plate (

franzdrs June 2, 2009 12:18

Perforated plate
I want to simulate a perforated plate with the porous media model. If the perforated plate is 1mm thick, is it important to make de geometry 1mm thick? or is it unimportant? I am guessing it is unimportant because the relevant information is in the loss coefficients already. I am right?
Also, is it possible to use the porous jump model or not? In the Userīs guide they mention perforated plates in the Porous media model, not in the Porous jump, which they say is siutalbe for thin membranes. Cant a perforated plate (1mm thick) be regarded also as a thin membrane, or is it not thin enough?

pd.: I am also wondering: in case the porous media model is the way to go, and the plate thickness must be the real 1mm (or lets say 2mm), can the grid of this zone be 1 element of thickness (in the flow direction) or should it be several?

Please if somebody has experience with this, I would appreciate some comments.

mettler June 3, 2009 08:36

what exactly are you trying to model? But, you can model a perforated plate as porous media. There has been a lot of research done on heat transfer in porous media and perforated plates. I did my whole dissertation on it. I used the thermal non-equilibrium model. Your heat transfer will vary based on the thickness of the plate. Also, if you have a stack of perforated plates, separated by a spacer plate (Matrix heat exchanger) how you have to the plates organized will affect the heat transfer.
You are going to have to find the pressure drop thru your perforated plate and relate it to the porous media.

franzdrs June 3, 2009 16:36

thanks for your answer. Actually what I want to simulate is comparatively easy. I want to simulate the air distribution in a food dryer. no heat transfer simulation. this dryer has a perforated plate on which the product to be dried lies. So for the perforated plate, since it is a thin sheet (about 1-3 mm), I dont know if in the geometry I should make the volume 1-3 mm thick or it does not matter. Again, no heat transfer. So, if I had the required coefficients of the plate for the porous media model, is it important to make the plate geometry with the actual 1-3 mm thickness? and should the grid for the plate have several grid elements in its thickness or is it ok to make it 1 element thick?
what about the porous jump model for the plate? can it be used?
thanks again

mettler June 3, 2009 16:42

is air flowing thru this plate? The porous jump model is just a pressure drop - it is simulating the pressure drop that would occur thru the porous media. That is why you need to know the pressure drop.

franzdrs June 3, 2009 18:19

yes, the air flows through the plate. I want to use a perforated plate (or more than one) also to improve the air distribution after a (not so) gradual expansion.
So in this case, with no heat transfer modeling, the porous jump would do it, wouldnt it?
What about the mesh for a perforated plate modeled with the Porous Media model? a thickness of one grid element is ok?`
thanks a lot

mettler June 3, 2009 19:36

what you need to do is look for a paper written by Sparrow in 1986 (I think) is old. He will show you how to model air flowing thru a perforated plate. I used his paper to model my perforated plate in Fluent with very good results.. I am guessing you just want to model the air flow thru the perforated plate.

franzdrs June 5, 2009 06:59

Hi Andrew,
yes, I only want to simulate the flow through the plate. I will look for that paper. Thanks for the tip.
I have one more question: Letīs say I have the data of pressure drop vs air speed for the perforated plate. I guess that from these data i can calculate the coefficients for the porous media model in the direction normal to the plate. what about the other two directions? are the coefficients for these directions zero?

mettler June 5, 2009 08:36

When I used the porous media model in Fluent a few years ago it was necessary to have a graph of the pressure drop versus velocity. You can generate it yourself if you have the data. From that you can curve fit a line and get a second order equation for the line. From that equation you can get the coefficients. I will have to dig up the paper that I used for that. That is how I did it anyway. I am sure there are other ways. Also, a lot of journal papers dealing with heat transfer in porous media will give you value for the coefficients. I checked mine against these values.

Air flow parallel to the plate? That is a different kind of problem since most of the air will not flow thru the holes, since it is going to go the path of least resistance. Your pressure drop will be affected by the holes, but nothing like you have when you are forcing air to flow thru the perforated plate.

TypeSpeed January 26, 2010 18:30

Hello mettler,

I am simulating flow through a porous plate. I am very interested in the paper by Sparrow you mentioned. I searched the database (Gee, He did publish a lot of paper), but I can not find this paper. Can you give the title of this paper and the journal's name?

Thank you very much!

mettler January 26, 2010 18:53

email me on here and I will send you the info. I think it was published way back in the 80s, so you are probably going to have to have your library track it down via inter library loan. If I can find the pdf I will send it to you.

TypeSpeed January 26, 2010 19:07

Thanks, Mettler.
I just send you a private message. I am looking forward to receiving your reply.

TypeSpeed January 26, 2010 21:01


Thank you very much for your information. I really appreciate it.

Best wishes,

mrjonezz May 3, 2010 10:25

Hi everyone,

I am currently modelling a perforated plate myself. It is 1.5mm thick and the holes are 8mm of diameter. Shortest distance between the holes centers is 12mm. The Pattern is hexagonal. Kinda like:

.o o o o
o o o o o
.o o o o

I hope that made sense (ingore the dots, they are just place holders to get the Os in position).

What I did so far is thet I simulated a small portion of the perforated plate (about 9 holes) as a periodic case with different flow angles (e.g. 0° == perpendicular to thew plate; 90° == parallel to the plate) to get an idea on how great the losses are. Also I wanted to get some info on how the flow angle changes over the plate. How can I use this information to calculate the necessary coefficents for the axial and transverse directions in a porosity?

Thanks for any help!


I would also be interested in this paper (Sparrow 1986) that you mentioned. Could you let me know the full title of it please?

mettler May 3, 2010 11:23

you don't need to model more than one hole - as long as they are organized on your plate and the wind flow is perpendicular to the plate surface. You can model only one hole. I verified the lamp-black oil technique that showed the shear stresses on the plate resulted in a repeating flow pattern around each of the holes. It formed a hexagon around each hole. So, you only have to model on hexagon with a hole in the middle. I model it in Fluent and my results were very close to Sparrow's - he did a naphthalene sublimation to measure the convective coefficient. If you pm your email I will send the Sparrow paper and some others.

will321321 March 29, 2015 21:32

Hello mettler,

I am also trying to model porous media can you please send me a copy of sparrow's paper? Ill pm you my email in awhile.

Thank you.

mettler August 28, 2015 00:33

before trying porous media, I would use the lampblack technique on a piece of your perforated sheet. That will let you see if there is any sort of pattern that would let you model it - possibly only modeling one hole. That is what I did and Sparrow did.

As for modeling it as porous media that would be tough. You need to know some of the variables used in the porous media model. The only way you would be able to do that is to measure the pressure drop across the plate at different velocities and then back out the variables (permeability and inertia) after curve fitting the data. You could try that and see if it matches up with experimental data.

All times are GMT -4. The time now is 01:13.