Re: [Getfem-users] Strange behaviour of add_penalized_contact_with_rigid_obstacle_brick method

[Renard Yves]

Dear David,

Your program seems to me correct. Are you sure that the Newton algorithm 
converged ?

A standard value for the penalty parameter is E/h where E is Young's 
modulus and h the mesh size. Is you penalty parameter of this order ?

Best regards,

Yves


Le 17/01/2022 à 19:03, David Danan a écrit :
> Dear Getfem users,
>
> i wanted to try the add_penalized_contact_with_rigid_obstacle_brick 
> method a while ago and the results seems consistent with the other 
> methods i had used so far, therefore i was satisfied.
>
> However, on my test case, i found out that the contact condition seems 
> to be ignored for some reason.
>
> I was able to reproduce the issue using an already existing mesh in 
> Getfem and the test example demo_static_contact.py file, please find 
> enclosed the python script associated.
>
> If i use enforceContactWith="Penalization" or 
> enforceContactWith="Multipliers" for E=1, everything is ok and the 
> solutions exported are "roughly" the same
> However, for E=1e6, while the solution obtained by using multipliers 
> gives relatively reasonable results, that's not the case for the 
> penalization (see picture enclosed).
> I tried to increase the penalization coefficient but it hardly had any 
> impact on the solution.
>
> Did i miss something?
>
> Thank you in advance.
> Kind regards,
> David.