Contact test example result not correct

[dzheng26]

Description I copied the example from ./ContactStructuralMechanicsApplication/tests/ALM_frictional_contact_test_3D and enabled the output to vtk. Then I plotted the deformation. The deformation looks not right.

Scope test example results

To Reproduce Run the example

Expected behavior The example is with two blocks. The top block is sliding from left to right in contact with the bottom block. The expected result is the lower part of the top block been bushed backward (right) due to friction. The calculated result does not show that. The original example is with frictional coefficient 0.1. When I changed it to 0.5 or 0.9, the deformation was even worse and the simulation did not finish to the end.

Environment

• OS: fedora 7.7
• Branch: master
• Python 3.7

[AlejandroCornejo]

Hi @dzheng-conti ! Can you share with me the input files that you are using to run it in my pc?

[dzheng26]

@AlejandroCornejo , Here is the test examples I did.

friction_bug.tar.gz

[loumalouomega]

Hi @dzheng-conti I have checked, and you are right. There is an unexpected behaviour which had not noticed before. We are working on it.

[loumalouomega]

Greetings @dzheng-conti, the example has been updated, and now it gives consisting behaviour (see gif). There were two problems:

• The examples were not updated, and the friction was imposed in a deprecated manner, using a property instead of the contact process. This was giving already some strange behaviour. The reason this was changed is that the friction coefficient is not a property of a material, but the interaction between two surfaces. This change was some month ago, and I forgot to update the examples.
• There is an implementation detail that is not automatic enough. For now I have corrected manually the value operator_threshold, in the example. I have as a pending task to add a procedure to automatically calculate this value instead of using a constant value. This parameter is used to do a simplistic calculation of the slip* in some cases where a zero division may occur, the slip determines also the direction of the "frictional force", therefore a bad estimation of the slip may lead to a wrong value of the "frictional force". Until now the value considered was not giving serious issues, now I will consider adding this automatic procedure in a near future.

*NOTE: Instead of considering the objective formulation for the slip calculation detailed in my thesis, a non-objective calculation is considered. In theory for small displacements and problems not dealing with rigid movements the values should coincide, but of course in practice is a bit more troublesome.

Sorry for the inconvenience, best regards

Vicente Mataix Ferrándiz

[dzheng26]

@loumalouomega Thanks for the quick response. Could you tell me where I can find the changed example? I cannot find it in the master.

[loumalouomega]

@dzheng-conti I have updated the example in the example repository: https://github.com/KratosMultiphysics/Examples/tree/master/contact_structural_mechanics/validation/press_fit/source/3D

I am currently checking something in the repository and I will update the code and tests accordingly.

Best regards,

Vicente Mataix Ferrándiz

[rubenzorrilla]

@loumalouomega is this still an issue?

[loumalouomega]

The issue is still there but the "fix" in the example is already applied. I need a lot of time to actually test and check the fix.