[ers127]

I am a new Abaqus user and working on a 4-point bending problem, but with a bone instead of a beam. I have read through the Abaqus manual, dozens of websites, and watched dozens of videos to configure my analysis; however, I believe my contact interactions aren't quite right, but I'm not sure why.

I have written a Matlab code to mesh and configure the assembly, so when I open the input file in Abaqus, the assembly opens as an orphan mesh. I calculated the positions of the rollers such that they should be resting on the bone. When I check the contact pairs in Abaqus, it indicates 3 of the 4 rollers are intersecting the bone, and one is offset. I have checked my calculations, and the positions I calculated should work, as I copied the formula and applied to each roller. The assembly is configured in Matlab because I cannot accurately calculate 2 of 3 bone coordinates corresponding to the rollers' position.


To address the intersection/overclosure, I enact the option to Adjust interactions to remove overclosure in the Interaction Module. This seems to work for the top left and bottom left rollers, but not the right rollers, as I receive the warning:
"For contact pair (assembly_cp-2-slave-assembly_cp-2-trmaster), adjustment was specified but no node was adjusted more than the adjustment distance = 2.22000e-16. and For contact pair (assembly_cp-4-brmaster-assembly_cp-4-slave), adjustment was specified but no node was adjusted more than the adjustment distance = 2.22000e-16." which leads to another warning: There are 2 unconnected regions in the model. Then, the analysis completes, but not quite correctly. When I open the output file in the Visualization module, I see there is some deformation, but since the right rollers aren't contacting the bone, these results are not valid.

I have tried several other techniques, including general contact with contact initialization for strain-free adjustment, which resulted in an additional warning: "Displacement increment for contact is too big," ultimately leading to analysis failure and error: "Too many attempts made for this model."

Many videos I watched involved partitioning the beam and adding a datum point, which allows a coincident point position constraint. However, I am not able to partition the model since it has an orphan mesh, and I cannot add datum points because I cannot accurately calculate the coordinates of the contact points on the femur. I have attempted to estimate contact points to enact a position constraint, and also attempted to add a position constraint between the roller and a plane which passes through the contact point in hopes of sectioning the femur and extracting the femur contact coordinates, but Abaqus does not allow such a constraint.

Perhaps creating a set for the top rollers and another set for the bottom rollers or defining a different position for the reference nodes will work - I will try these next.

I understand this is a lot of information - please let me know if I can clarify anything, or if scheduling a meeting would be beneficial! I have been troubleshooting this around the clock for about 5 days now, and have tried everything I can think of, so any help would be greatly appreciated.

I have attached some files as well. Due to size restrictions, I'm not able to upload the .inp or .cae files. Any help with this is also appreciated! Please let me know if there is any other information I can provide.

Matlab Assembly.jpg

Abaqus Warnings.jpg

[random_ran]

It is unusual for people using 4-point bending technique to study a
bone. We civil people using this method to test beams in most of the
case.

> I have written a Matlab code to mesh and configure the assembly

Abaqus has a module to mesh the geometry, so you could just input the
geometric information to Abaqus. Then, applied the load on the node or
a reference line.


> "Displacement increment for contact is too big," ultimately leading
> to analysis failure and error: "Too many attempts made for this
> model."

Then, using the smaller time step to calculate it.


If you are new to Abaqus, then maybe try LS-dyna, which is more
specialized in the contact simulation.