CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   FLUENT not even close to experiment (http://www.cfd-online.com/Forums/fluent/98393-fluent-not-even-close-experiment.html)

khoopes March 9, 2012 13:55

FLUENT not even close to experiment
 
I have a question regarding FLUENT and getting it to model the proper physics. To begin, I am not that familiar with CFD, I am just learning, so pardon my ignorance. The basics of the problem is that for a research project we are designing a screen to be put in front of a jet engine to create a velocity distortion. The idea that we are playing with right now is creating a hexagonal grid (think honeycomb pattern) with varying porosity and airfoil shaped walls. The idea is that the velocity will vary with porosity. We have a reduced order prediction model that we are working on and would like to use CFD as well. We have a geometry generation system that can output our 3d models and then we send these off to a 3d printer to make an actual screen.

A few weeks ago we printed a screen with a simple linear porosity, and put it in our wind tunnel. Just as we thought, we got a linear velocity profile downstream of the screen. A pretty substantial variation too. When normalized to velocity one side of the screen was seeing 100% velocity (as it was the highest) and the other side 25% in a nice and linear as expected.
The problem is that when I try to model a 2d section of this screen (the same section that we traversed with our probe, I get no velocity gradient at all!
I am using a the steady, pressure based solver with ideal gas for compressibility. Our inlet speed is about 44 m/s.

For the inlet boundary condition I have a pressure inlet with a pressure that will have the same inlet velocity as our wind tunnel. The outlet is a pressure outlet with 0 gauge pressure. The airfoils are walls with a prescribed roughness that I approximated using the tolerances of the 3d printer. I am using the transitional sst as some boundary layer codes that we have run tell us that we are transitioning to a turbulent boundary layer on some of our airfoils.

I am just at a loss as to what else to change? I know the pictures are pretty rough, but it should give you an idea of the geometry I am looking at. I got the geometry straight from the same program we used to generate the geometry for the 3d printer. The total height is .15m and the length is .9m with the airfoils .3m from the inlet. I keep thinking that I should see some sort of velocity gradient downstream like we saw in our wind tunnel? I am getting none at all. The velocity is pretty much uniform.

To generate the grid I used the ANSYS workbench mesher. I used the y+ calculator on CFD-online to try to get a y+ of 0.25. the nearest point to the wall on each airfoil is at 1e-6m with 100 points going out to the freestream at a growth rate of 1.04. Each airfoil has 1000 points along the perimeter. The freestreem mesh has a prescribed element size of 1e-4m. The mesh may be my problem, again I am pretty ignorant about most of this stuff.
Is there something that I am missing that is making it not produce a velocity gradient? If there is anything else that you would need to help out let me know.

http://dl.dropbox.com/u/386423/full%20geometry.png


All times are GMT -4. The time now is 03:15.