Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Spherical Indentation

Please login with a confirmed email address before reporting spam

Hi,

Im trying to model the indentation of a spherical probe into a nearly incompressible material (Hertzian Contact). The model seems to get a stress concentration at a single point rather than a distribution and it also does not converge for all indentation depths. I have tried refining the mesh to improve convergence but no results.

Comsol support provided me with a file called Cylinder Roller Contact in which they do almost exactly what I'm doing except in 2D geometry for a cylinder rather than a sphere. I have tried taking their working model and replacing the point load that they use with a parametric displacement and the model immediately fails to work.

Edit: Ignore the first "spherical_ind2" file as that one does not have the contact pairs set up.



5 Replies Last Post Jul 16, 2013, 9:21 a.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 8, 2010, 5:29 p.m. EDT
Hi Daniel,
I was not able to open your second *.mph file to check some details.
Just some hints:
perhaps it is trivial but did you defined - beside the contact pairs- also a load in direction of the contact destination ?
Don't mistake the "contact pressure" with load !
If you used really a load, did you used also face load instead of point load ?

Regards
Ender
Hi Daniel, I was not able to open your second *.mph file to check some details. Just some hints: perhaps it is trivial but did you defined - beside the contact pairs- also a load in direction of the contact destination ? Don't mistake the "contact pressure" with load ! If you used really a load, did you used also face load instead of point load ? Regards Ender

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 9, 2010, 6:46 p.m. EDT
When I define displacement, I do not define a load for the model because that is what I want comsol to solve for.

In the end I am looking for reaction force vs displacement data.

I managed to get the model to somewhat work for using the static solver and just one single displacement except that it fails to converge at 100% ( I don't know why). If I try to parametrically solve for multiple displacements, the model simply doesn't work.
When I define displacement, I do not define a load for the model because that is what I want comsol to solve for. In the end I am looking for reaction force vs displacement data. I managed to get the model to somewhat work for using the static solver and just one single displacement except that it fails to converge at 100% ( I don't know why). If I try to parametrically solve for multiple displacements, the model simply doesn't work.


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 10, 2010, 3:21 p.m. EDT
Hi

I'm not sure it my following comment has its full importance, but I would have made the "contact pair" shorter and not ending up one perpendicular to the other, i.e. cut the sphere by a line at perhaps 60°.

Then I would try a parametric displacement ramp starting at some small "eps" to have an initial contact.

I only have V4.0a with me, but I cannot run your model, for some reason the translation makes it "separate" and not compress the sample ?

Good luck
Ivar

Hi I'm not sure it my following comment has its full importance, but I would have made the "contact pair" shorter and not ending up one perpendicular to the other, i.e. cut the sphere by a line at perhaps 60°. Then I would try a parametric displacement ramp starting at some small "eps" to have an initial contact. I only have V4.0a with me, but I cannot run your model, for some reason the translation makes it "separate" and not compress the sample ? Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 12, 2010, 10:33 a.m. EDT

You were right about having to start at some small indentation to avoid convergence issues. I also had to increase the max # of augmented Lagrangian iterations and now my model finally works.

Thanks for your help!

Edit: I saved the file and closed it. When I reopened it and attempted to run the simulation again, it did not converge. Why is this if the settings should not have changed?
You were right about having to start at some small indentation to avoid convergence issues. I also had to increase the max # of augmented Lagrangian iterations and now my model finally works. Thanks for your help! Edit: I saved the file and closed it. When I reopened it and attempted to run the simulation again, it did not converge. Why is this if the settings should not have changed?

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 16, 2013, 9:21 a.m. EDT
Hello,

I am also looking for Hertzian contact file...If you have it.please share it with me..i will be very thankful to you...
Hello, I am also looking for Hertzian contact file...If you have it.please share it with me..i will be very thankful to you...

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.