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.

Contact problem in solid mechanic: sover don't convert

Please login with a confirmed email address before reporting spam

Hello!

I want to simulate a screw in a boneblock which is bonded with a force vertical to the screw axis (see picture). The two blocks are a union with differend materials. There are used 3 different materials: the screw, the cortikalis (smaler block) and the cancellous bone (bigger block). In Finish I've form an assembly, activated Create pairs and the Pair type is Contact pair. I 've create a contact pair between the srew surface an the bone in Definitions. In boundary conditions I've assign the Contact Pair and the load in the screw head. The bone surfaces are fixed Constraints. A free tetrahedral mesh is used.

I used the stationary solver, solved in the direct way with MUMPS and fully Coupled. But the Solver don't convert.

Maybe the solver is wrong for this probem or I've forgot something. Can you please help me?
What is the best solver for my contact problem and what I did wrong?

Thank you for your help
Ariane


6 Replies Last Post May 13, 2013, 10:04 a.m. EDT
Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 6, 2013, 10:54 a.m. EDT
Hi Ariane,

the red cross at the Materials node indicates that material assignments are missing. You have quite a delicate geometry there. I would recommend to test the boundary conditions in a simpler setup and in 2D axial and then get back to 3D.

Cheers
Edgar

--
Edgar J. Kaiser
www.emphys.com
Hi Ariane, the red cross at the Materials node indicates that material assignments are missing. You have quite a delicate geometry there. I would recommend to test the boundary conditions in a simpler setup and in 2D axial and then get back to 3D. Cheers Edgar -- Edgar J. Kaiser http://www.emphys.com

Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 6, 2013, 11:31 a.m. EDT
Hi,

Some quick additional hints:

1. Something is wrong with the solver sequence. A contact problem should have a segregated solver. Regenerate the solver sequence.

2. Check the scaling of the contact pressure under Dependent variables so that it is relevant for your physics.

3. Assembly problems like this are often singular in the initial state before contact is established. If this is the case, please refer to the documentation and Knowledgebase about how to avoid that.

Regards,
Henrik
Hi, Some quick additional hints: 1. Something is wrong with the solver sequence. A contact problem should have a segregated solver. Regenerate the solver sequence. 2. Check the scaling of the contact pressure under Dependent variables so that it is relevant for your physics. 3. Assembly problems like this are often singular in the initial state before contact is established. If this is the case, please refer to the documentation and Knowledgebase about how to avoid that. Regards, Henrik

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 6, 2013, 2:17 p.m. EDT
Thank you for the quick answer, it is helpful.

The material data are the poisson ratio and the Young's modulus, the density is missing (because I don't know it), but in a simpler model (a cylinder insead of the screw) the problem was solved.

Is MUMPS a good solver for this problem? I'm not sure, because I used Comsol since 3 weeks.

Thank you,
Ariane

Thank you for the quick answer, it is helpful. The material data are the poisson ratio and the Young's modulus, the density is missing (because I don't know it), but in a simpler model (a cylinder insead of the screw) the problem was solved. Is MUMPS a good solver for this problem? I'm not sure, because I used Comsol since 3 weeks. Thank you, Ariane

Josh Thomas Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 6, 2013, 5:03 p.m. EDT
Ariane -

MUMPS is just fine for this problem. You could also use PARDISO. More important is that you've selected the Segregated Solver rather than the Fully Coupled as suggested by Henrik. The default solver sequence should generate with the Segregated Solver if you are not sure how to set this up yourself.

No need to add the density if you are solving the Stationary form of the Solid Mechanics equations.

Best regards,
Josh Thomas
AltaSim Technologies
Ariane - MUMPS is just fine for this problem. You could also use PARDISO. More important is that you've selected the Segregated Solver rather than the Fully Coupled as suggested by Henrik. The default solver sequence should generate with the Segregated Solver if you are not sure how to set this up yourself. No need to add the density if you are solving the Stationary form of the Solid Mechanics equations. Best regards, Josh Thomas AltaSim Technologies

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 13, 2013, 1:52 a.m. EDT
Hello!

Thank you for the help. I tried all you said (changed the scaling factor, used the segregated solver, set a small initial contact pressure). But when I use the default settings with the segregated solver, there are two steps a segregated (for the displacement) and a lumped (for the contact). With this default settings the cylinder don't flexure but becoming thicker. If I use only the segregated step for the displacement and the contact, the cylinder flexure. In both settings the convergence plot is not so good.

What should I do?
I thought it is an easy problem but I'm wrong. Maybe I can mail somebody my model to check the settings?

Thank you for your help.
Ariane
Hello! Thank you for the help. I tried all you said (changed the scaling factor, used the segregated solver, set a small initial contact pressure). But when I use the default settings with the segregated solver, there are two steps a segregated (for the displacement) and a lumped (for the contact). With this default settings the cylinder don't flexure but becoming thicker. If I use only the segregated step for the displacement and the contact, the cylinder flexure. In both settings the convergence plot is not so good. What should I do? I thought it is an easy problem but I'm wrong. Maybe I can mail somebody my model to check the settings? Thank you for your help. Ariane

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 13, 2013, 10:04 a.m. EDT
Hi Ariane,

Based on the figure you posted that is a tough contact problem to solve! In addition to the suggestions mentioned above I would look at the convergence log closely and also monitor the non-converged solution during analysis (via the “Results While Solving” section in the Solver window). That may provide clues on what is causing the convergence problems.

Nagi Elabbasi
Veryst Engineering
Hi Ariane, Based on the figure you posted that is a tough contact problem to solve! In addition to the suggestions mentioned above I would look at the convergence log closely and also monitor the non-converged solution during analysis (via the “Results While Solving” section in the Solver window). That may provide clues on what is causing the convergence problems. Nagi Elabbasi Veryst Engineering

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.