Hi

As a simple example I have started with modeling laminar flow through a pipe

I have created a mesh in ICEM CFD in cartesian coordinates(I presume ICEM allows you to create geometries in cartesian coordinates only)

After exporting it to CFX, How should I specify the velocities and other details in cylindrical coordinates?

Also let me know if ICEM permits creation of geometry in cylindrical coordinates.

Thanks

VSB

After exporting it to CFX, How should I specify the velocities and other details in cylindrical coordinates?

Create a new cylincrical CS in Pre.

CFX actually gives you cylindrical coordinates (r and theta) about the Z axis of any coordinate system. If you pipe is aligned with the Z axis, you'll already have access to these.

If all you want is the ability to specify velocities in cylindrical coordinates, just pick this option in the boundary condition GUI. It will ask you for the primary axis, so you may still need to create a coordinate system if your pipe is not aligned with a global axis, but I suspect it is.

What exactly do you need the cylindrical coordinate system for?

Regards, Robin

Hi Robin and Joe,

I would like to know one thing.

If I model a pipe flow problem using cylindrical coordinates and cartesian coordinates, will there be a difference in the solution values obtained in both the techniques?

I have been getting contrasting views from different people.

I need your help.

Thanks

VSB

Dear VSB,

ANSYS CFX only works with cartesian coordinates.. You can input data in cylindrical coordinates, and they will be converted to their cartesian components.

Most 3D codes are in cartesian coordinates, and 2D codes offer an option for axisymmetric cases.

What is your main concern? mesh discretization of the curvilinear terms? The physics of the flow does not know anything about the mesh, or reference frame selected..

Opaque

[oewnce932c]

Hi
Do ANSYS CFX still accept only Cartesian coordinate system?
Trying to add cylindrical system, for some output data..but couldn't find it out;(

[mohammad]

Dear all,

It might be very simple question and I have sought for the answer in previous posts...but!!!

My question is how to define a Cylindrical coordinate system; either in Pre- or Post-???

Tnx a lot

[icem_god]

Do ANSYS CFX still accept only Cartesian coordinate system?
Trying to add cylindrical system, for some output data..but couldn't find it out;(

the question is still actual in 2019...

[Opaque]

May I ask what is your goal?

Every coordinate frame in ANSYS CFX provides cylindrical variables such as radial distance, angular position and axial distance.

Will that work for you?

[icem_god]

my goal is to set up the inlet pressure-temperature radial and circumferential irregularity of axial compressor and it`s in cylindrical CS. I`ve tried to use initial profile data with R and theta components, new CS with Z-axis as rotating, but it seems like it`s not a cylindrical CS

[Opaque]

We may need more details on how you set it up:

1 - What do you mean by ".. with Z-axis as rotating, .." ?

2 - did you use a local coordinate frame at the boundary, or just the global frame ?

3 - is the boundary condition in the rotating frame, or in the stationary frame ?

4 - which variables did you use from the CEL repertory for cylindrical coordinates ?

[icem_god]

Originally Posted by Opaque

We may need more details on how you set it up:

1 - What do you mean by ".. with Z-axis as rotating, .." ?

2 - did you use a local coordinate frame at the boundary, or just the global frame ?

3 - is the boundary condition in the rotating frame, or in the stationary frame ?

4 - which variables did you use from the CEL repertory for cylindrical coordinates ?

2. Yes, i`m using my new CS at the boundary and UserFunctions-Profile Data.
3. BC is in the stationary frame.

1/4. My model has X-rotating axis. I`ve added another CS with Z-rotating axis.
If i don`t have circumferential irregularity I use Initialize profile data with R/Pressure/Temperature data - and it "appears" in cylindrical CS. But when I`m trying to add Theta to this data (R/Theta/Pressure/Temperature) it "becomes" XYZ.

[Opaque]

Perhaps it would be better to show the expression you are using, in ANSYS CFX there are several cylindrical coordinate triplets

1 -> (r, theta, z) also known as (Distance from local z axis, Angle around local z axis, Z Coordinate)

2 -> (Radius, Theta, Axial Distance), also known as (raxis,taxis,aaxis)

There are subtle differences between them, and though they can mixed in the same expression it would be really confusing to understand the expression later on.

Are you familiar with these coordinates and their subtleties? Which of the two triplets are you using?

Have you used the CFX-Pre boundary plot feature to see how your expression behaves when using any of the two triplets?

[icem_god]

I`m using CFX built in option Tools-Initialize Profile Data
Then selecting a csv-file with R/Theta/Pressure/Temperature like this

[Name]
Data Field

[Spatial Fields]
R, theta

[Data]
R [ m ], theta [degree], Pressure [ Pa ], Temperature [ K ]
0.2 , 0, 95000 , 300
0.2 , 15, 110000 , 315
...

After that i use profile boundary setup on my domain boundary and my Pressures and Temperatures of this boundary becomes Data Field.Pressure(r, theta) and Data Field.Temperature(r, theta)

[icem_god]

The main question is parametres distribution

I have two cases:

First case is Radius-Pressure-Temperature distribution

[Name]
Data Field

[Spatial Fields]
R

[Data]
R [ m ], Pressure [ Pa ], Temperature [ K ]
0.15 , 95000 , 300
0.2 , 110000 , 315
...

In this case I see classic cylindrical irregularity on the plot. Bottom image

Second case is Radius-Theta-Pressure-Temperature distribution

[Name]
Data Field

[Spatial Fields]
R, Theta

[Data]
R [ m ], Theta [degree], Pressure [ Pa ], Temperature [ K ]
0.15 , 0, 95000 , 300
0.2 , 0, 110000 , 315
0.15 , 15, 95000 , 300
0.2 , 15, 110000 , 315
0.15 , -15, 95000 , 300
0.2 , -15, 110000 , 315
...

I supposed to get a similar result. And you can see its not true. Upper image
The result looks like an 'Cartesian' distribution

[icem_god]

[Opaque]

Great information, thank you.

You are using the (r, theta, z) triplet; therefore, you must be aware "r" is measured respect to the "z axis" used at the boundary.

Could you tell which coordinate frame is used at the boundary? Coord 0 or an user defined coordinate frame?

In any case, "r" is measured respect to the z-axis of that frame regardless of rotation axis (in case of a rotating component), i.e.

r = sqrt (x^2 + y^2)

You can play with a simple expression, and verify what each variable means. For example, set your pressure to

Relative Pressure = r * 1 [Pa m^-1]

and look at your contours, they better be radial, correct? or

Relative Pressure = raxis * 1 [Pa m^-1]

Similarly for the "theta", it is measured from the x-axis towards y-axis around the z-axis.

Hope the above helps