Coordination Frame Problem - strange

Luk_Fiz · July 30, 2010, 04:00

Hi,
Since I am preparing parts to bigger simulation I am checking small things which I have never performed in CFX before. This is a strange problem related to coordination frames.
My problem is multi-domain with interfaces. One of them is a simple domain, which is (during time dependent simulation) twisted and squeezed accordingly to some prescribed functions, in time. These functions are related to "Total Mesh Displacement X", "theta" and other CFX functions. After careful geometrical considerations I managed to describe the movement accurately and everything is ok. And now:
In the other part of my entire arrangement is a second domain of exactly the same size and initial shape. This domain is shifted and turned comparing to the first one however it performs the same movement and deformations during the simulation.
And here is the problem:
Because of similarity of domains and motions I wanted to use the same functions to describe the movement, but prepare second coordinate frame that way, that this second domain is in the same position relatively to second coords as the first on in the first (global) coords. In this purpose I have created second coords, set up second domain to them but... It does not work!! These formulas creates some strange movement of second domain and it looks like it was working in the old one coords.

Can somebode be that kind and describe how to overcome this. In shortcoming I want to know how to force CFX expression to hold in local coordinate system.

Luk

To be more specific: assume that we move the domain and the local coords the same way (the domain is always in the same position relatively to local coords):
1- If I translate relatively to coord 0 - everything is ok
2- If I rotate relatively to coord 0 - it fails

Luk_Fiz · July 30, 2010, 06:42

Ok,
After few trials I have found that the problem was related to Total Mesh Displacement.
I decided to modify my functions describing domain squeezing in such a way to not to use any build-in CFX functions. During this I have found that while "r" and "theta" seems to be local (I mean x and y that are used in them are local x and y) than Total Mesh Displacement x uses global x. I was not aware of and I cannot find in documentation is it proper on not.

Sorry for answering my own questions but before I got it I was working two days on this problem.
Luk

stumpy · July 30, 2010, 09:20

That sounds correct. "x", "y" and "z" are the only variables that are transformed into local coordinate frames. There is a note in the doc about this:
"A variable or CEL expression that is used in the specification of some quantity is evaluated with x, y, z treated as being local, and all other non-scalar variables (such as u, v, w) treated as being global. Non-scalar CEL functions (such as torque, force) are treated as being local. The quantity that was specified by the variable or CEL expression, once evaluated, will be treated as being local, if applicable (that is, if it is a vector or tensor quantity, or component thereof)."

July 30, 2010, 04:00	Coordination Frame Problem - strange	#1
Luk_Fiz Member Lukasz Join Date: Mar 2009 Posts: 67 Rep Power: 17	Hi, Since I am preparing parts to bigger simulation I am checking small things which I have never performed in CFX before. This is a strange problem related to coordination frames. My problem is multi-domain with interfaces. One of them is a simple domain, which is (during time dependent simulation) twisted and squeezed accordingly to some prescribed functions, in time. These functions are related to "Total Mesh Displacement X", "theta" and other CFX functions. After careful geometrical considerations I managed to describe the movement accurately and everything is ok. And now: In the other part of my entire arrangement is a second domain of exactly the same size and initial shape. This domain is shifted and turned comparing to the first one however it performs the same movement and deformations during the simulation. And here is the problem: Because of similarity of domains and motions I wanted to use the same functions to describe the movement, but prepare second coordinate frame that way, that this second domain is in the same position relatively to second coords as the first on in the first (global) coords. In this purpose I have created second coords, set up second domain to them but... It does not work!! These formulas creates some strange movement of second domain and it looks like it was working in the old one coords. Can somebode be that kind and describe how to overcome this. In shortcoming I want to know how to force CFX expression to hold in local coordinate system. Luk To be more specific: assume that we move the domain and the local coords the same way (the domain is always in the same position relatively to local coords): 1- If I translate relatively to coord 0 - everything is ok 2- If I rotate relatively to coord 0 - it fails Last edited by Luk_Fiz; July 30, 2010 at 04:16.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Urgent Help !!!Strange reverse flow in naca0012 problem	parvez2683	FLUENT	5	May 20, 2013 10:25
question about governing equation in CFX using rotating/non rotating reference frame	rystokes	CFX	0	January 12, 2010 07:14
case-loading problem	EllenW	FLUENT	0	September 14, 2009 18:35
CHT strange problem	MatejK	CFX	3	September 1, 2005 06:32
strange problem	SS	FLUENT	4	March 25, 2005 13:04

July 30, 2010, 06:42		#2
Luk_Fiz Member Lukasz Join Date: Mar 2009 Posts: 67 Rep Power: 17	Ok, After few trials I have found that the problem was related to Total Mesh Displacement. I decided to modify my functions describing domain squeezing in such a way to not to use any build-in CFX functions. During this I have found that while "r" and "theta" seems to be local (I mean x and y that are used in them are local x and y) than Total Mesh Displacement x uses global x. I was not aware of and I cannot find in documentation is it proper on not. Sorry for answering my own questions but before I got it I was working two days on this problem. Luk

July 30, 2010, 09:20		#3
stumpy Senior Member Join Date: Apr 2009 Posts: 531 Rep Power: 21	That sounds correct. "x", "y" and "z" are the only variables that are transformed into local coordinate frames. There is a note in the doc about this: "A variable or CEL expression that is used in the specification of some quantity is evaluated with x, y, z treated as being local, and all other non-scalar variables (such as u, v, w) treated as being global. Non-scalar CEL functions (such as torque, force) are treated as being local. The quantity that was specified by the variable or CEL expression, once evaluated, will be treated as being local, if applicable (that is, if it is a vector or tensor quantity, or component thereof)."