Simulation of orifice flow meter
 

Hello,

I tried to validate the simulation of a standard orifice flow meter. The main goal is to investigate and check the discharge coefficient. Based on the theoretical calculation and measurement, the discharge coefficient should be about 0.6 in the investigated Re number region.
http://www.engineeringtoolbox.com/or...uri-d_590.html
Based on the result of my numerical analysis, this value is 0.77. I think, this difference is high, but I don’t understand its reason. I used unsteady analysis, SST turbulence model and very fine structured hexa mesh. The y plus value is correct. I used symmetry boundaries, so a quarter tube (90 degrees) was calculated instead of the full (360 deg) model. Maybe is it the problem? Or any other thing?

Regards
Roland

...sorry, I made a mistake: the simulation is steady

Few things:

Are you pulling P out in same manner as a Tap would?

Did you adjust your P locations to match each of the 3 listed in that article? Does the cd vary across each of those locations? I dont think they listed which location was used for the cd that is shown (or are you using different information for your experimental data) so it has some uncertainty in it?

Did you make sure your BCs provide a fully developed flow into and eventually away from the orifice?

Hello Singer,

sorry for my late answer.
I dont know the details of measurements in the article. But I think that the discharge coefficient should be correct based on the geometry and the used Re number region.

Based on my university studies, the discharge coefficient should be ~0.6 in case a standard orifice flow meter. This value is higher in my simulation. I think that the separated zones are smaller in the simulation than in the reality. I would like to know its reason.

Regards
Roland

An additional question to this problem: how can be determinated the discharge coefficient from CFX result?
I determinated it based on the streamlines.
http://imageshack.us/photo/my-images/22/13826578.png/]http://img22.imageshack.us/img22/9907/13826578.png[/URL]
There is a cross section from "D geom", and a smaller one from the "D flow". Discharge coefficient has been calculated as: A flow/A geom.
What do you think about this? Is this method correct?

Regards
Roland