Appllying bolt pretension load in Abaqus Explicit Solver

[Tinni1]

Hello,

Using a dynamic explicit solver in Abaqus, I am performing a post-buckling quasi-static analysis of thin-walled (2 mm thick) cold-formed steel C sections connected to a U section at both ends.

As per my model geometry, the U-section is connected to hot rolled steel loading plates (12 mm thick) at both ends with M8 bolts, 4 numbers on each side.

I am applying the bold pretension load using two different files.

In the first file I performed a static analysis using bolt load command.

In the second file I imorted the deformed geometry of the first analysis file as imperfections.

My queries are:

1. Is the method that I employed correct?
2. Could you please direct me to any references of chapters of ABAQUS manual about this?

Many thanks in advance!

 

[SWComposites]

Why are you bothering with bolt pretension? Or even bothering with modelling the fasteners at all?
Thin gage structure in a post-buckling state should not be critical in the bolted joint areas.

 

[Tinni1]

Yes, I agree.

I modeled it for only one benchmark analysis, for checking.

The analysis provides correct results when bolt pretension is not modeled.

With the pretension, I am getting an artificial increase in the column capacity and also an incorrect failure mechanism that does not match my test results.

That's why I want to check if the pretension analysis method , I am following correct or incorrect.

 

[SWComposites]

I would start by closely examining the displacement results from both models.

 

[FEA way]

What elements are you using to model the bolts ? What happens to them after import ?

 

[Tinni1]

Hi, thanks for your response.

I am using solid elements of 1mm x1mmx 1 mm mesh size.

I am attaching two pictures:
1. The picture of the bolt after adding the bolt load (7 N).
2. The picture of the bolt after the post-buckling analysis where imported geometry from the bolt load file was incorporated as imperfections.

There is no much visible difference between the two. However, the post-buckling analysis gave me a 16% higher column capacity than testing.
Although the slope of the load-axial displacement curve is the same with and witout bolt load.

 

[SWComposites]

Instead of showing us this ridiculous bolt mesh, please show the column buckling mode and displacements for the two models.

 

[Tinni1]

My answer was against the previous question about bolt meshing only

 

[FEA way]

If done correctly, the approach with pre-loading bolts in Abaqus/Standard and importing the model to Abaqus/Explicit is reasonable. But it makes sense mostly for models focusing on joints themselves. If it's just a part of a large structure, it might be better to model the bolts in a more simplified way using 1D elements. Connectors are really good for that and you will be able to apply pre-tension more directly.

 

[Tinni1]

That's great! thanks for your response. I will try using connectors