Simulating a fence with perforated plate

Project Support
1

I trying trying to simulate (Nageli henge measuremenets by jjansman | SimScale , simulation “Kort, =17.5, z0=0.01m, Fineness=7.5”) a fence (as used in an article, Nageli, 1953: Nägeli, Werner, Untersuchungen über die Windverhältnisse im Bereich von Schilfrohrwänden, Mitt. Schweiz. Anst. Forstl. Versuchw., 29, 213–266, 1953 ).

The fence (at y=0) is a box ( 25m wide, 2.2m height, 40cm thick) which I gave a perforated plate. The ABL wind comes from negative y to positive.

In Porosity & Porous Media | Advanced Concepts | SimScale I read for ‘Direction’:

Flow direction: With this input, the user can specify the principal flow direction within the medium. This direction is based on global coordinates.

So I expected that I have to give ‘Direction’ (x,y,z) of the perforated plate as (0,1,0) as that is the general wind direction. If I do that the wind behind the fence is almost independent of the ‘Free area Ratio’. See Runs ‘ap-12-010’ <[aerodynamic porosity]-[value]-[xyz]> and ‘ap-8.5-010’. Also the results are quite different from the experimental data of Nageli.

If I use (x,y,z) as (1,1,0) ‘Direction’ in perforated plate; the simulation (Run: ‘‘ap-8.5-110’’) is very close to the experiment data of Nageli.

It looks I don’t understand the meaning of ‘Direction’ in perforated plate. Can you help me?

All the best.

P.S. For ‘Free area Ratio’=0.085, different Direction configuration (000,001,010,011,100,101,110,111) are evaluated, see that Simulation “Kort, =17.5, z0=0.01m, Fineness=7.5”

3

Interestingly, the Expected profile can be deducted from the Found profile, by using:
rankxyzExpected=mod(-rankxyzFound,8).

The data to support this, can be found in the earlier mentioned URL: Simuleren van wind bij Nageli artikel

This must be some coordination transformation… Why this (coordination) conversion (by intend or otherwise) is there? So I still don’t understand…

4

Any feedback from someone. I find it so strange that I need to provide a different flow direction then I would expect?! I must interpret something wrong.
PLEASE HELP!

8

Hi, here are some thoughts that come to mind:

  1. The perforated plate model is a porous media model intended for perforated plates (as the name suggests). This is an approximation that should be used when you don’t have experimental data related to the perforated plate that you are dealing with (i.e. using experimental data will always be more accurate). I wouldn’t say that a fence classifies as a perforated plate
  2. When defining a porous media, you will sometimes be able to define a flow “direction”. From a practical perspective the fluid flowing in this “direction” will undergo less delta pressure than in the other 2 “blocked” directions. Note that fluid can still flow in the blocked directions, it will just need to overcome a much larger delta P
  3. As you might expect from the previous points, a porous media will cause large pressure gradients. To be able to capture these gradients accurately you have to refine the porous media properly (at least 5 cells, as per the documentation). At the moment you have around 1 cell over the thickness of the porous media, which will throw off all gradients.

I hope this helps.

9

Thanks Ricardo, I will workon this.
So the mesh I use over the thickness of the media is certainly not enough.

But I find it still strange that I need to define such a strange direction (xyz=110) (while the wind direciton is mainly in the +y axis:(010) ).
Anyway, I will first work on improving the mesh!