CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   3D Rotating Channel (

Curtin University April 6, 2012 01:03

3D Rotating Channel
Hi all,

I'm currently trying to model a rotating channel using FLUENT 13.0. The channel is a certain distance from the origin in the y-dxn and rotating about the z-axis. I'm using an ASME paper 'Heat Transfer in Rotating Passages with Smooth Walls' by Wagner et al 1991, to validate my results before moving on to conducting parametric analysis.

I've tried using a simple UDF to incorporate coriolis and centrifugal body force source terms. I've also tried using a moving reference frame (MRF) using 'cell zone conditions'. However, my results (particularly the local Nu plots) are nothing like those presented in the Wagner et al paper.

Has anyone here successfully modelled rotating channels/passges before? Any suggestions on the solver settings or wall boundary conditions that need to be used?

Currently, I've set a constant rotational velocity under MRF. I've also set a moving wall boundary condition with a 0 rad/sec relative rotational velocity to adjecant cell zone.

Your help is much appriciated!


Zigainer April 9, 2012 13:45


my calculations focus on a rotating impingement configuration.
A MRF works fine for me. What kind of inlet boundary condition do you use? Make sure your direction is not “absolute”, but "relative to adjusted cell".
Do you use a temperature and velocity profile? Which turbulence intensity?
Have you done a calculation with no rotation? Just to see if these results fit with other papers. Maybe there is a problem with the mesh or your choice of turbulence model.

Curtin University April 9, 2012 22:44

Hi Zigainer,

The boundary conditions I have used are as follows:
* Velocity-inlet bonundary condition - constant magnitude & set 'relative to adjacent cell zone'. Further, a turbulence intensity of 5% has been defined.
* Pressure-outlet with a gauge pressure of 0 Pa. Turbulence intensity is 1%.
* All wall boundary conditons have been set to a 'moving wall' (rotational) b.c with a relative to adjacent cell zone velocity of 0 rad/s. Further, the walls have a constant temperature (higher than the fluid temperature).

What is interesting (or rather frustrating) is that the stationary channel model seems to match reasonably well with those of experimental values. I've even done a rib-roughened stationary channel and I'm getting good trends. However, the smooth rotating channel is not even showing the trend.

The turbulence model I'm using is a k-e realisable. SIMPLE mode has been set for P-V coupling. Futher, LSCB & BFW discretisations methods have been used for gradient & pressure respectively, whilst SOU has been used for momentum, KE, dissipation rate & energy.

I've heard suggestions of using an 'incompressibel ideal gas' for the density instead of boussinesq approximation. However, I'm yet to see if that solve the problem....

What sort of solver settings did you use when modelling the rotating impingmen configuration i.e. turbulence model, discretisation methods etc...?

If you think any of the solver settings I've mentioned above may be invalid for this application, please let me know.


Zigainer April 10, 2012 03:24


all of your settings seem to be all right.

I am using the "ideal gas law" for density and "piecewise-linear" for all the other properties.
My discretization methods are: PRESTO! for pressure (recommended by the manual) and SOU for all the other.
So maybe give the "ideal gas law" a try.

All times are GMT -4. The time now is 15:48.