Field output transformation doesn't work?

[potf]

Hi!

My model is as follow:

3D solid, revolved, C3D8 elements, material: steel.


I am trying to transform the field output of my model into a cylindrical coordinate system with CAE.

Therefore I create a news CSYS in the visualisation module and in "result options" I choose transformation -> user-defined.

This seems to work fine with the stresses, however the strains are not converted correctly: the model and load are axisymmetric, so the strain in theta direction (S22 in the new CSYS) should be constant, but it isn't.

As this didn't work I tried out a lot of other ways to transform my results: with *orientation in **part I got the same results.
The only way that works is to use *orientation in material, but I don't like this solution because my material is steel and therefore isotropic. Why should I use material orientation then!?

Does anyone know why the transformation in the visualisation module doesn't work as I want? (Or what I am doing wrong?)

Thanks alot in advance,

Gael

 

[corus]

I can't answer this but if the geometry is 3D solid axi-symmetic then why not use a 2D axi-symmetric model? That would solve the problem.

corus

 

[Tamlin]

Hello Gael,

the transformation in Abaqus/Viewer converts correctly.

When your model and load is axisymmetric, then use a
2D-axiasymmetric model as corus advised.

Tamlin

 

[potf]

Hi,

The model (a contacting part) and the load are not axisymmetric in the following steps, which are the important steps, that's why I didn't use the 2D axisymmetric elements.

Unfortunately I don't think I can use the "2D axisymmetric with non axisymmetric load" elements (CAXA) here either. The documentation for these elements is not really helpful but from what I understand it doesn't seem possible.


I checked the stresses and strains in the first (axisymmetric) step to verify that the conversion was correct, and that's how noticed something went wrong.

Yes I guess the conversion works correcty in CAE, but could someone please tell me what I am doing wrong, or how it should be done step by step so I can verify what I do?

thanks again,

Gael

 

[Drej]

If your loads/mesh/geometry/etc. are axisymmetric, then your transformation should show that your RADIAL (r) stresses/strains etc. are constant (concentric at least) as well as the hoop(theta) stresses/strains. Are they..? You might expect some non-uniformity if your mesh is non-uniform, but as you say it's revolved (swept?) then all your output should show signs of uniformity. You mention contact being involved, is this affecting the strain? What shape is your model? Have you rotated your results into the user defined cylindrical system? Can you list results and check to see the coordinate system that these are listed in? What type of cylindrical system have you specified (r, theta, z) or (r, z, theta) etc...?

 

[potf]

As I had this problem, I tried out the transformation on a very simplified model (which has actualy nothing in common with my original model) for testing purpose of the transformation function:

It is a simple C3D8 cylinder under axisymmetric load, with axisymmetric mesh, nothing else. (no contact)

As you say Drej, after the transformation into a cylindrical (r,theta,z) CSYS (origin in the center), I expect the following results:
concentric r,theta and z stresses/strains.

What I get:
-concentric r,theta and z stresses.
-a concentric z strain.
So far so good. But I also get
-non concentric r and theta strains. (with differences ~50% between min and max. value)

Is it a definition problem, an error or am I expecting wrong results?

In the Abaqus/CAE visualisation module you can use the Symbol plot to plot the tensor direct components. The components are as I expected after the transformation (1 is radial etc.)

I would really appreciate it if someone could try out the result transformation on a simple model as I did, and tell me the result.
Thanks alot,

Gael