How do I get mean velocities in cylindrical coords in unsteady statistics?

Pfalcon · August 12, 2009, 07:15

Hi,

I'm performing URANS calculations (Fluent6.3) in a rotating turbine passage. I need to obtain the time mean velocities in the moving domain in cylindrical coordinates. I can only get the Mean X-Y-Z-Velocities in cartesian coordinates. Unfortunately this is meaningless because as the domain physically rotates and the coordinate system does not. A simple coordinate transformation does not help either because the mean is the average of all the different angular positions.
Does anyone know a solution to this?
Another thing that would interest me is the period over which the instantanious data is averaged. I can specify the frequency of timesteps to use but how many timesteps from the past are taken into account in the averaging process?

Many thanks for helping me out ...

Pfalcon · August 14, 2009, 13:45

Ah, at least I found a little note in the FLUENT manual saying that the time mean velocities in the unsteady statistics are meaningless for moving meshes. Although it would be pretty simple for FLUENT to convert the XYZ velocities into intrinsic coordinates (polar coordinates in rotating meshes) and use these to obtain the time mean velocities.

I can't believe people are not interested in the time mean velocities in moving meshes. Turbomachinery guys should be! Or is there some easy way of getting this that I don' realise? ...

August 12, 2009, 07:15	How do I get mean velocities in cylindrical coords in unsteady statistics?	#1
Pfalcon New Member Join Date: Aug 2009 Posts: 9 Rep Power: 16	Hi, I'm performing URANS calculations (Fluent6.3) in a rotating turbine passage. I need to obtain the time mean velocities in the moving domain in cylindrical coordinates. I can only get the Mean X-Y-Z-Velocities in cartesian coordinates. Unfortunately this is meaningless because as the domain physically rotates and the coordinate system does not. A simple coordinate transformation does not help either because the mean is the average of all the different angular positions. Does anyone know a solution to this? Another thing that would interest me is the period over which the instantanious data is averaged. I can specify the frequency of timesteps to use but how many timesteps from the past are taken into account in the averaging process? Many thanks for helping me out ...

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Macros for unsteady statistics	Abdel Dehbi	FLUENT	0	August 30, 2008 10:31
unsteady statistics in UDM	A. Dehbi	FLUENT	2	August 27, 2008 06:12
unsteady statistics for UDF	Luis	FLUENT	0	April 16, 2008 15:45
unsteady statistics	shefali	FLUENT	0	August 12, 2006 20:13
mean boundary values for unsteady statistics	Francois	FLUENT	0	March 21, 2005 10:28

August 14, 2009, 13:45		#2
Pfalcon New Member Join Date: Aug 2009 Posts: 9 Rep Power: 16	Ah, at least I found a little note in the FLUENT manual saying that the time mean velocities in the unsteady statistics are meaningless for moving meshes. Although it would be pretty simple for FLUENT to convert the XYZ velocities into intrinsic coordinates (polar coordinates in rotating meshes) and use these to obtain the time mean velocities. I can't believe people are not interested in the time mean velocities in moving meshes. Turbomachinery guys should be! Or is there some easy way of getting this that I don' realise? ...