CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Drafttube Simulation Problem (http://www.cfd-online.com/Forums/cfx/19209-drafttube-simulation-problem.html)

Begineer October 7, 2002 04:28

Drafttube Simulation Problem
 
Hi,

I'm working on the drafttube simulation for Francis turbine. Can anyone help me with the following problems? Thanks for your advice.

1. I am applying the logarithmic velocity profile as the inlet boundary condition. Since the formulas (velocity) for viscous sublayer and log region are different, I may have to use the if.then.else algorithm. Can anyone advise me on how to create a user subroutine for this problem? Is there a simpler way to do this?

2. I have merged the surfaces before generating the mesh to get rid of the small angles. However, the warning messages still appear while creating the suface mesh. I have used the merged surface to generate the surface mesh. Can anyone tell if the message refer to the original surface or the merged surface? Can I ignore this message? Any solution I can apply without redrawing my model again?

3. How can I check the no of mesh elements in my model? Can I limit the number to a certain amount?

Begineer October 7, 2002 15:20

Re: Drafttube Simulation Problem
 
Can I create an expression with If..then..else command in the expression editor? If not, how can I do that?

Mike October 7, 2002 23:44

Re: Drafttube Simulation Problem
 
Hi, 1. I think you'll need to use a User Subroutine. I'd suggest using the one provided for the Flow Through a Butterfly Value Tutorial as a template - maybe complete this tutorial if you want. 2. If you manually create the mesh in CFX-Build through the mesh forms then the message will refer to the original, not the merged mesh. If you can write out a Definition file Ok then that should be fine. You might want to check the mesh quailty by loading the Def file into CFX-Post before solving - there are some mesh quality checking tools in Post. When you say you have merged the surfaces, you shouldn't in general merge surfaces that lie in different planes since you will loose accurate geometry representation. 3. You should also be able to get the mesh size in Post. It is also given at the start of the OUT file when you start the solver. You can't predict the exact size in advance, so getting a feel for reducing/increasing the mesh length scale is the best way. Hope this helps, Mike

Begineer October 8, 2002 03:07

Re: Drafttube Simulation Problem
 
Thanks for your valuable inputs, Mike. There's no need to use IF..then..else algorithm in Tutorial problem as the power law velocity profile is used in this case. However, the Reynolds no in my problem is around 5*10^7 (Q=150m^3/s and R=2.033m), which may be too large to get an accurate solution using power law. That's the reason I'm thinking of using log profile. I didn't get from this tutorial of how the build-in Fortran program can be prepared (e.g. Pipeinlet_valve.F in tutorial 9).

On the way, if you think the power law can still be used, what value of n do you think I should put in? I can't find a reference to determine n at Re = 5*10^7.

The followings are the boundary conditions (velocity) that I'm hoping to use:

Viscous Sub layer: If Y+<12, then U+=Y+ Log region: else U+=2.5ln(Y+) + 5.2

Does anyone know how to write this in user-subroutine (I mean how the file like Pipeinlet_valve.F is written)? Is there any CFX5 statement/format that I have to put into the fortran program? How's that look like?

Again...Thanks for your time and comments.

Jon October 8, 2002 11:31

Re: Drafttube Simulation Problem
 
You can use the step(x) function as an elseif function.

i.e var = f1(whatever)*step(g1)+f2(whatever)*step(g2)



All times are GMT -4. The time now is 15:50.