Problem with what should be a simple 2D flexural test!

[robmalkin]

Hello everyone,

Long time reader, first time poster. I have had a look in the search and done my best to look through the masses of information in the guide but am still having an issue.

I am simulating a simple 4 point bend/flexural test using Abaqus Standard. Please see Image1 for the whole model. The beam is 3mm in thickness and is divided into a few sections. The top part of the beam is 2.75mm thick and is modelled as an anisotropic solid. The bottom of the beam is divided into two parts with a gap between the ends (this is to simulate a composite beam with a cut ply on the tensile surface). Between the gap is a small block given neat resin properties. Image2 should show this more carefully!

This all sounds kind of complicated so...
Image 1 - Shows the beam and the location of the rollers. In the initial step and load step the bottom rollers are fully constrained. In the load step I apply a displacement to both top rollers to bend the beam. Left and right motion is contrained via a BC.

Image 2 - This shows the bottom middle section of the beam. I have added labels to show the part which is modelled as a beam and the part which is neat resin. The red lines show the location of the cohesive elements.

Some details of the model...

Beam + resin elements: CPE8
Cohesive elements:COH2D4

I have given the cohesive elements a 1E-5 value for the viscocity.

In the step I have turned on Nlgeom. I have tried the automatic stabilization, but to be honest am not too confident with what I am doing!

Anyway, the model fails to solve, I have reduced the minimum step to 0.001 and inceased attempts to 20, but till no joy.

Im sorry if I am missing some basic info here, but I am a bit new to this.

Any ideas? I have had it run to the stage where the bottom section almost seperates from the top, and think maybe I have done something silly with the stabilization or maybe the mesh.

Would really appreciate help on this!

Thanks,

Rob Malkin

 

[robmalkin]

And Image2

 

[rstupplebeen]

1. Posting your model might help describe the second image more.

2. Instead of cohesive properties why not model this as 1 surface that is cut into various regions with different material properties? Are you trying to model crack propagation?

3. Model only 1 half of the beam since there is symmetry.

4. You probably do not need stabilization if the supports are initially in contact.

I hope this helps.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/

 

[robmalkin]

Hi Rob,

1- I am away from my desk until monday, but will post the model when I get in.

2- I am trying to model a crack yes, I am keen to know where the stresses around the resin pocket are and how the crack forms. Forgive the stupidity of the next statement... I was under the impression that to model a crack you would need to make two parts. For example, if you were doing a simple mode I opening, you would need to make two parts (such as top and bottom) a cohesive interlayer and give the cohesive interlayer cohesive properties (cohesive elements). Or have I just been making my live much more difficult by making many parts when I could just make one part with many partitions?

3- The model is symetrical at the moment, so I could model half, but I will be making it unsymetrical at a later stage.

4- Good to know.

Thank you for getting back to me soon, I really appreciate the help. The biggest problem is that I have been self-teaching myself from a few PhD students at work. And bad habits get passed on!

I will get back with more info on monday. But for now, its friday evening and I should enjoy it!!


Thanks again,

Rob

 

[robmalkin]

Hey everyone...

So I could not leave it, and here I am on a sunny Saturday.

I think I may have found the problem. I made a simple error with my selection of master and slave surfaces. I feel like a bit of a dummy.

I hope that now it should run ok.

I will let you know....

Rob