Interference-fit analysis

After trying to run your case, I see that actually the first increment fails to converges because the relative residual (=the force "mis" equilibrium divided by the total applied force) is too high; but the absolute value of the residue is very low which means that the system is in equilibirum. To converge, by default, Aster requires that both absolute and relative residuals are below a certain value. This convergence problem is typical of cases where you don't have external load applied to the system and thus, relative residual is always something like 1/0=inf. and thuis does not converge.
So in your case, you should ask Aster to consider only the absolute residual (in unit force, so Newton) in the convergence check. So try to use the following settings which will enable ONLY global absolute residual magnitude verification.

CONVERGENCE=_F(RESI_GLOB_MAXI=1e-4,
ITER_GLOB_MAXI=60,),

Then your problem should converge.

In the following file, I also have increased the number of iterations for contact and the number of increments, but it should not be necessary... This proble has been running successfully on the Aster 10.1 OpenMP included in Salome-Meca 2010.

[file]<br /><br />Post edited by: Administrator, at: 2010/09/01 19:14