CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Turbine: Transient Simulation Problem (http://www.cfd-online.com/Forums/cfx/94891-turbine-transient-simulation-problem.html)

John C. November 29, 2011 07:30

Turbine: Transient Simulation Problem
 
Hi,


I have a little Problem with the evaluation of an transient Simulation in ANSYS CFX.


The basic Conditions are the following:


2 Stator Passages, 3 Rotor Passages and 3 Diffusor Passages. => Turbine


Stator: Stationary (No Rotation)
Rotor: Rotating (-10617 rev/min)
Diffusor: Rotating (-10617 rev/min)


Interface Stator to Rotor:

  • Frame Change/Mixing Model set to Transient Rotor/Stator
  • Pitch Change: Automatic


Interface Rotor to Diffusor:
  • Frame Change: None

I did the first simulation with an non-rotating Diffusor and no Problem did come up. So I changed the condition, because i will need it for further simulations.


My Problem:


The POST Processor is showing that the Rotor is moving with 3.3 degree per 100 Iterations, but the Diffusor is moving 30 degree per 100 Iterations. I thought they should move equally, because the have the same angular velocity.


Could somebody please help me?

ghorrocks November 29, 2011 17:15

Have you read the documentation on rotating frames of reference, pitch angles and bets practises guides, and done the tutorials for rotating machinery?

These seems like a simple setup problem.

John C. November 30, 2011 05:52

Hello ghorrocks,


actually I have done an similar simulation just some months ago, so I am not that new to this kind of simulation. But there I did not look at the “Angular Shift for Transient Rotating Domains”. Here I can choose from “Automatic”, “Always Rotate” and “Never Rotate”. (No helpful explanation in the Documentation) During my last simulation I have used “Never Rotate”, but this time I have changed it just to see what would happen. And believe me, I am not that happy about having it changed...


“Always Rotate” gives the result that I explained yesterday. And when I choose “Automatic” the Rotor moves again with 3.3 degree per 100 Iterations and the Diffusor stands still.


I have looked up the Problem in the documentation but there is nothing that would really help me so I thought somebody here had an similar problem and could help me.


The interesting part for me is that the data seems right (Here I am not completely sure), but the visual exposition is wrong (Here I am sure!). 3.3 degree is to slow for the Rotor, it should be higher and of course equal to the Diffusor.


I have done some considerations yesterday and on the one hand I think it might just be an problem with the visual exposition and does not effect the data, or on the other hand it is a Problem with the calibration at the Interfaces between Rotor to Stator and Rotor to Diffusor. But at the last Idea the Documentation of Ansys tells me that there should not be an Frame Change between two domains which have the same angular velocity.

John C. December 5, 2011 05:14

So it seems that nobody had an similar Problem.

Well, I kinda hoped that my Problem would at least start a little Discussion, but I guess that won't happen.

If I work out an Answer for the Problem I will let you know.

D.B December 5, 2011 06:06

In which Software are you meshing ? One way to check if there is an interface issue is to combine both the rotor and diffuser meshes in ur meshing software, then you wont have to define seperate domains for each of em. See if that works.

ghorrocks December 5, 2011 06:21

It still sounds like a simulation setup problem to me. These are not very "interesting" problems as you just need to set it up correctly and it will work fine. That is why you did not get any discussion on it.

D.B December 5, 2011 06:23

Could you copy paste or attach your out file here, that would help in finding out the problem ( if any) in the simulation setup.

John C. December 5, 2011 09:31

For the Meshing I used Turbogrid. Friday I merged the domain Rotor and Diff so that I only had one domain. That did work for my simulation, the data was not effected and i think it was just a visual exposition.

Only Problem was that I had to change my geometry, because ANSYS had some problems in the Preprocessor.

Thanks for the help!


All times are GMT -4. The time now is 22:19.