how to use UDF with cylinderical coordinate system in fluent

aestas · January 4, 2015, 20:27

Hi my friends,
i am working on plasma distribution problems, and it is more convenient for me to written the set of equations and boundary conditions in cylindrical component form, but as far as i knew, fluent only has Cartesian coordinate system

so how should i do to put the cylindrical component form into UDF and use them in fluent? I am desperate for your help~

Sun · January 5, 2015, 03:59

Hi,
Can you provide more details on the nature of equations?
Meanwhile, have a look at this doc, a nice summary of field variables/functions in Fluent. You might be able to find most of the access functions you need.
hope it helps.
cheers!

aestas · January 5, 2015, 06:57

Thanks for your replay, my friend.

it is in the paper titled "2D expansion of the low-density interelectrode vacuum arc plasma jet in an axial magnetic field",with a 2D MHD model, but the energy equations are not considered.The pdf is too large to upload.~

i wonder whether the equation set 3 can be calculated in fluent.

And you can see, the self magnetic field Bθ is more simple to express in cylinder system~

Quote:

Originally Posted by Sun

Hi,
Can you provide more details on the nature of equations?
Meanwhile, have a look at this doc, a nice summary of field variables/functions in Fluent. You might be able to find most of the access functions you need.
hope it helps.
cheers!

Sun · January 5, 2015, 08:39

Peter,
Unfortunately, I don't have access to the full paper. But Looking at Eqs. 3.1 to 3.6, you need to convert the Cartesian velocities into axial, radial, and tangential ones. I've googled around a bit and stumbled upon these slides.
On slide 34 and 35, you'll find a way to convert the velocities. However, I am not quite sure if you can use the very same approach for other terms in the equations. It might be a bit tricky.
cheers

aestas · January 5, 2015, 09:37

Thanks my friend, i got what you mean by the slides, use Cartesian in fluent while use cylinder system in UDF.
It is a bit tricky to me, but i will try two ways, to take the same approach in the slides, or just convert the basic equation set 2 into Cartesian system, and the tangential component has to be divided.
BTW, the original auther of the paper didnot use fluent to calculate, insead they program themselves

aestas · January 5, 2015, 10:00

Hi, Sun,
I think if it's a 3D model in Cylinder Coordinate, it can't directly be used in fluent, because the momentum components are written as Vx,Vy,Vz in fluent.As a consequence, the source terms in the momentum equation can't directly be used from Cylinder to Cartesian.

But my problem is a 2D axisymmetric model, the z and r component in the figure just equals the x and y component in Cartesian, the only difference is the tangential component vs y compenent. I think this may be done by the approach you told me. I would give it a try

Sun · January 5, 2015, 11:49

Alright great and thanks for the update. I guess that method might actually work for the 2D case.
good luck.

January 4, 2015, 20:27	how to use UDF with cylinderical coordinate system in fluent	#1
aestas Member Peter Aestas Join Date: Dec 2013 Posts: 64 Rep Power: 12	Hi my friends, i am working on plasma distribution problems, and it is more convenient for me to written the set of equations and boundary conditions in cylindrical component form, but as far as i knew, fluent only has Cartesian coordinate system so how should i do to put the cylindrical component form into UDF and use them in fluent? I am desperate for your help~

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
problem in using parallel process in fluent 14	aydinkabir88	FLUENT	1	July 10, 2013 02:00
fluent udf problem: write specific data for every iteration in a file.	nnvoro	Fluent UDF and Scheme Programming	1	May 27, 2013 15:26
Change to local Coordinate system in Fluent	Constantine	FLUENT	3	November 12, 2010 01:05
Compiling UDF in Fluent 12	jsm	FLUENT	1	July 22, 2009 07:40
On Coordinate system used in Fluent	varghese	FLUENT	9	March 22, 2002 11:16

January 5, 2015, 03:59		#2
Sun Senior Member Join Date: Nov 2010 Posts: 103 Rep Power: 15	Hi, Can you provide more details on the nature of equations? Meanwhile, have a look at this doc, a nice summary of field variables/functions in Fluent. You might be able to find most of the access functions you need. hope it helps. cheers!

January 5, 2015, 08:39		#4
Sun Senior Member Join Date: Nov 2010 Posts: 103 Rep Power: 15	Peter, Unfortunately, I don't have access to the full paper. But Looking at Eqs. 3.1 to 3.6, you need to convert the Cartesian velocities into axial, radial, and tangential ones. I've googled around a bit and stumbled upon these slides. On slide 34 and 35, you'll find a way to convert the velocities. However, I am not quite sure if you can use the very same approach for other terms in the equations. It might be a bit tricky. cheers

January 5, 2015, 09:37		#5
aestas Member Peter Aestas Join Date: Dec 2013 Posts: 64 Rep Power: 12	Thanks my friend, i got what you mean by the slides, use Cartesian in fluent while use cylinder system in UDF. It is a bit tricky to me, but i will try two ways, to take the same approach in the slides, or just convert the basic equation set 2 into Cartesian system, and the tangential component has to be divided. BTW, the original auther of the paper didnot use fluent to calculate, insead they program themselves

January 5, 2015, 10:00		#6
aestas Member Peter Aestas Join Date: Dec 2013 Posts: 64 Rep Power: 12	Hi, Sun, I think if it's a 3D model in Cylinder Coordinate, it can't directly be used in fluent, because the momentum components are written as Vx,Vy,Vz in fluent.As a consequence, the source terms in the momentum equation can't directly be used from Cylinder to Cartesian. But my problem is a 2D axisymmetric model, the z and r component in the figure just equals the x and y component in Cartesian, the only difference is the tangential component vs y compenent. I think this may be done by the approach you told me. I would give it a try

January 5, 2015, 11:49		#7
Sun Senior Member Join Date: Nov 2010 Posts: 103 Rep Power: 15	Alright great and thanks for the update. I guess that method might actually work for the 2D case. good luck.