Pb with YieldCriteria Filter

ParaView Support
1

Hi,

I have an XDMF file containing a model with an attribute named Static_Stress, which includes tensors in the Tensor6 format.

It appears to be displayed correctly in Paraview:

image
image505×602 23.9 KB

When I try the YieldCriteria filter, I don’t see any available Point Tensors in the YieldCriteria Filter Panel:

image
image510×432 15.6 KB

Does anyone know how to use this filter and if I need to prepare my data in any way?

You can find below the data files used.
StressTest.h5 (197.6 KB)
StressTest.xdmf (1.1 KB)

I’m using Paraview 5.12.1 under Windows 10.

Best regards,

LB

2

Hello @Loic,

I think this filter show only 6 components point data (probably because some tensor are symmetric).

FYI @Francois_Mazen

3

Hello @Lucas,

My data contains originally symmetric tensors stored in XDMF Tensor6 format with 6 components per tensor.

The XDMF reader converts them into generic tensors with 9 components per tensor.

Is there a way to convert them back to a symmetric tensor format, compatible with the YieldCriteria filter?

Regards,

4

Hello @Loic,

You can do that with a programmable filter.

In the Request Data you can copy paste this python script (troncate a 9 components tensor to 6 and name this new data array Stress6):
programmable_filter.py (674 Bytes)

and don’t forget to enable the Copy Arrays property of this filter.

Here is what I have:

image
image1129×680 103 KB

Let me know if it works, also I want to open an issue regarding this, the Yield Criteria filter should be able to use a 9 components, can I link you shared data ?

5

Thanks @Lucas for your feedback,

If I have to write a script, I still prefer to write a filter to calculate criteria like YieldCritria.
This would allow me to add my own criteria if needed.

I would still prefer to use the filter provided by Paraview.
So I would find it logical to open an issue on this problem.
Don’t hesitate to link this test case to illustrate the problem.

6

Thanks, it makes sense, I open the issue in VTK as the root cause of this issue is the vtk filter itself: https://gitlab.kitware.com/vtk/vtk/-/issues/19614