Some problems in first attempt with Abaqus

[Paraboloide]

Hello everybody. Is my first time with Abaqus, and I´m trying to do nonlinear analysis that involves nonlinear material and model change.

I want to model 2 principal beam: one is a brigde deck (30 m. lenght), the other is a railway track (630 m.). The object is analize the track-bridge interaction. This interaction, depends of nonlinear material behavior between bridge and track in longitudinal direction (bilinear relation force-displacement).

The magnitude of the behavior, depends if the train is over the track or not (different values of the force). I modelize it with a nonlinear spring.

The principal loads are: uniform thermal load (50ºC - track; 35ºC - deck bridge), vertical uniform distributed load (80 kN/m - train weight) and horizontal uniform distributed load (20 kN/m - brake). The uniforms loads, are in function of the train position.

The order of the loads: thermal loads (without train), vertical train load and horizontal train load. So, I have to apply "model change" to change the behavior of the spring before the train loads, an after that.

The problem (finally) is that Abaqus don´t converge. I have imposed a substep size of 1E-09, but don´t solve it.

 

[corus]

I don't see why this is non-linear unless you expect the stresses to exceed yield significantly. Run it as a linear model and run the thermal loads in one step and your mechanical loads in the second step. You can combine them later in Viewer. Why you need model change because the loads change, I don't know. The reason it may not converge may be because you have a very high stress at a singularity. Running a linear model will show you that.

Tata

 

[Fabi0]

Corus, what do you mean with combining them in the viewer? I've a similar problem and maybe could be helpful for me.

Thx in advance

Fabio

 

[Paraboloide]

Thanks for the answers .

Corus, this is nonlinear because the material behavior between the track and the deck has a bilinear relation force-displacement, and depends if the train is over the rail or not.

Is something like that:

force
|
| _______________________ (with train)
| /
| /
| /
| /
| / ....................... (without train)
| / .
| / .
| /.
|/_____________________________________________ displc.

Again, thanks a lot for answering. I solved it this morning

 

[corus]

FabiO, It's sometimes useful to separate thermal loads from mechanical loads in order to determine dominating effects, or to assess the stresses with different design stress limits. To combine the separate steps or load cases just create a new step in Viewer under the tools/create field output./From fields.






Tata

 

[Paraboloide]

Again, I try to increase the difficult of the problem, but Abaqus don´t solve it.

My question is: if my first step is the boundary conditions definition and remove of certains elements ONLY, why large and strange displacements happens in several nodes?

NOTES:
---> The STEP definition:
*STEP
*STATIC
*MODEL CHANGE, TYPE=ELEMENT, REMOVE
BalastoCViaDcha02, BalastoCViaDcha03, BalastoCViaIzqda03, BalastoCViaIzqda04
*BOUNDARY
40001, 2
40081, 2
1, 2
2, 2
3, 2
4, 2
....
*END STEP

---> The STRANGE results:
***WARNING: SOLVER PROBLEM. NUMERICAL SINGULARITY WHEN PROCESSING NODE 30023
D.O.F. 1 RATIO = 6.06968E+09.


***WARNING: SOLVER PROBLEM. NUMERICAL SINGULARITY WHEN PROCESSING NODE 20017
D.O.F. 1 RATIO = 6.74370E+09.

....

Can it be, that I have overconstraints?