Abaqus Force Displacement Curve Not Smooth 2

[Struct123ure]

Hi, I am pretty new to Abaqus and have been studying and practicing it intensely.
For the past month I have been trying to get a Force v. Displacement curve for a reinforced concrete beam. The force response I was seeing was not making sense so I simplified it to a homogenous single material concrete beam(no reinforcement to start). I am performing the analysis using the explicit solver because I want to ultimately do dynamic once I build up my knowledge and verify the mdoel is working correctly.
After a month of work and learning till the curve looked like below attached, with very sharp peaks up and down.

I had a breakthrough yesterday where I switched the “Amplitude” from “Type: Tabular” to “Type: Smooth Step” which fixed the beginning of the curve making it look proper, see below:

My issue is that after the peak (concrete cracks and elements deleted) it should be a flatline on the bottom, since it has no more capacity.
But there are random peaks after concrete cracks(cicled in red) and I don’t understand why?

I have attached my CAE model file just in case.

 

[yassou]

Hello Struct123ure
I did not run your model, But i think it's related to your contact definition.
Maybe when your model experience Crack/Element deletion, Between your components (Rigid and Deformable) some sort of gap and discrepancy happened.
Check is there any gap or deviation happened between nodes that have contact with rigid body.
With best regards.
yassou.

 

[FEA way]

Explicit is really tricky when it comes to the smoothness of the response. It's not only prone to aliasing (especially when it comes to accelerations and reaction forces) but also to peaks resulting from various highly dynamic phenomena such as local chattering in contact or sudden snaps. Take a closer look at the deformed shape and contour plots in critical areas. And check all the relevant global energies, it's crucial for the verification of Abaqus/Explicit results. If your problem can be considered quasi-static (the use of the smooth-step amplitude would confirm that), try solving it using dynamic implicit quasi-static procedure.

 

[Struct123ure]

Hi FEA way,

I think your aliasing comment was bang on. I increased the loading (of 5mm deflection) from a 5 second Time Period to a 50 second Time Period and got the graph below which after peak load looses capacity and flatlines at 0kN (or close to zero being 1.1*10^-6).

On a separate note is interesting that in the beginning of the graph went from curved to a straight-line. Not sure why at this point but will have to think about it.