Learning Forum

[Andre Conde Vazquez]

Hello, I am a university researcher and an ANSYS Mechanical 2022 R2 user. I came up with three doubts regarding the use of connections/contacts and I wonder if anyone here can help me with any of them.

 

My first doubt arises when solving the stresses in a Static Structural simulation of linear elastic 3D solids with two surfaces that maintain contact. If this contact is Bonded type, will the same results be produced as if those two surfaces maintain a Rough type contact and also maintain a No-Separation type contact (both contacts at the same time)?

 

My second question is based on the following methodology:

(i) Calculate through a simulation the stresses and deformations of a 3D linear elastic solid subjected to an external force

(ii) Export the Vector Principal Stress to a file

(iii) Solve a second simulation with the same solid with no external force applied, but importing the stresses from the file as initial stresses

(iv) The deformation obtained will be the opposite of that of the first simulation

This methodology works fine except when my solid includes a contact surface that is not Bonded. Are the contact surface equations incompatible with this methodology?

 

The third doubt is based on the same methodology previously described. In it, the stresses obtained at the end are almost zero (because there are no loads applied to the solid). But if after step (i) and before step (iii) I change the type of some of the contact surfaces, then the stresses obtained at the end are no longer close to zero. Are these stresses the difference between the original stresses and those that would exist if these forces were applied to the new geometry?

 

I hope I have explained my doubts clearly.

[Sahil Sura]

Hello Andre Conde Vazquez,

Answer for the Q1-
So if you consider the functionality of the contacts,
Bonded contact is a contact that won’t let the contact and target bodies separate and slide with respect to one another.
If seen for No Separation contact, the contact face can slide with respect to the target but cannot be separated,
while for Rough, it can separate in a normal direction but cannot slide in a tangential direction (as the coefficient of friction is assumed to be infinite)

Your case is to compare bonded with the other two combined (No Separation and Rough) so as to get a similar kind of behavior.

If you consider it theoretically, it might be possible to get similar results. But this will lead to over-defining a contact which might not help the problem to converge!
Practically, with the computational considerations, if a problem can be solved with a linear contact (Bonded), using a rough contact which is a non-linear type, would increase the computational expense.
So when using a combination of contacts, you might need to deal with some convergence issues.
 

Answer for the Q2-

On solving a well-defined model, you can insert requested results for post-processing of the results. These can be pre-defined ones or you can define the results as per criteria using the ‘User Defined Results’ object.
To export the result to a file (.txt or .xls), you can simply right-click on the result and export it to the respective file format.
Exporting Results (ansys.com)

Now if you want to import the data from a file to the new system, the ‘External Data’ system from the workbench homepage is useful. You may want to select how the data is arranged, for what it must be scoped to, etc.
For more details please refer to the following link-
External Data Import (ansys.com)

For more details on contacts please refer to the –
Contact (ansys.com)
Contact Formulation Theory (ansys.com)

The above references would also help you get familiar with the functionality of each contact type, which might serve you helpful in addressing the changes in forces observed.
 Hope this helps!
Thanks,
Sahil
https://forum.ansys.com/forums/topic/how-to-access-the-ansys-online-help/
https://forum.ansys.com/forums/topic/guidelines-for-posting-on-ansys-learning-forum/#latest

 

[Andre Conde Vazquez]

Hello Sahil Sura,

Thanks for answering my questions

Regarding Q1 I am willing to accept more computational expense and the risk of possible convergence problems if the results are going to be the same as with the Bonded contact.

Regarding Q2 I am familiar with the methods for exporting and importing results with ANSYS and I have verified that the methodology I have mentioned works correctly when there are no contact surfaces (in this video you can see an example www.youtube.com/watch?v=aAnSv7r7EnM). My question was if this methodology is compatible with non Bonded contact surfaces.