Code_Aster contact analysis - train rail

Hi kees,

I have attached the latest version of my attempts. I made a few refinements to the mesh and change how I was declaring the fixation of the sleepers and this has made an improvement and cleared alot of the errors...

I believe what is left to be directly associated to either the fact the wheel has to move in order to contact the rail (the 5mm clearance) or the contact state itself.

Hope you get a chance to take a look, enjoy the sun! It is wet here (Suffolk, UK)

Thanks,

Kev

Hi

Has anyone had a chance to have a look at this problem, after many hours of trying, and pulling out of hair, I am stuck!

I have been reading the DEFI_CONTACT and STAT_NON_LINE documentation and think I need to add CONTACT_INIT='OUI' as there is a clearance between the wheel and rail, but eficas does not recognise the keyword in DEFI_CONTACT? I have been unable to access the V11 docs on code_aster website in English so have been referring to the V12 docs, not sure if this makes a difference?

Also, will it help if I convert the mesh to quadratic within code_aster (then to convert to QUAD9) as convert to quadratic from Salome brings up more mesh errors?

Still best case is the one posted in previous post.

Many Thanks,

Kev

Hi,

I posted on the Salome pre-pro section as I think the errors are due to the clearance between the rail and wheel resulting in an unsolvable problem...

I think this was right, and I need to add some 'spring' elements on the wheel. I also think I need to tie the sleepers to the rail (further experimentation in linear on just the rail).

Does it sound reasonable to create a discreet node above, below and on all 4 sides of the rail, the connect them all with 'edges' to the wheel?