Pin - Joint connection - modelling with linear gap elements 1

[GDeGreve]

Dear,

i'm currently involved in a project for obtaining my masters degree in civil engineering in which i should investigate the stress concentration factor (Kt) around the pinhole of a lug. Thanks to a lot of tutorials of MSC and additional exercices i already managed to establish a fairly acceptable FEA model of the problem. However ...

- currently i'm modelling the contact pressure of the bolt as a sinusoïdal pressure varying between -90° ... 90° where 0° is the direction in which we apply the load. I'm able to generate fringe results, plot XY graphs and export the data back to Excel for calculating Kt results.

The results (hoopstresses in function of theta, in which theta defines the position on contact curve) are not yet accurate. We think the main reason for this inaccuracy is the inability to model the bolt and the contact between the deformable bolt and the lug. In order to tackle this problem i want to give a shot towards modelling gap elements. I was thinking of modelling the contact with the linear gap elements in PATRAN.
The main problem i stumple upon is the fact that i cannot find a clear instruction list or information on how to correctle model these gap elements. I manage to define them between 2 nodes but for some reason NASTRAN will not solve.

What i did so far:

- model the gap elements with the utility "linear gaps"
- create analysis subcase
- selected the linear gaps file and imported as bulk data

My question now can someone please provide us with a clear instruction on how to get the gap elements working? On the net i already found quite a lot of scientific research papers where the author applied these elements, but no one is appearently eager to explain the modelling steps I even tried to implement the 3 beams with gap elements exercise of MSC strategy (model gap elements as BAR2).

Many thanks for your input,

Günther

 

[ESPcomposites]

- You have to be a bit careful since this forum is for primarily for practicing engineers and not school projects. But that is OK, I may be able to answer a few things.

- The cosine distribution is not accurate compared to what? Compared to experimental data (i.e. Peterson's) or compared to other contact solutions? Were the other solutions "neat fits" (perfect fits) or did they have a clearance? I have modeled this problem extensively so I am pretty familiar. The cosine distribution will not exactly match the nonlinear contact solution. Neither will exactly match Peterson's solution.

- I have not modeled it with gap elements, but I would think can just put a gap between the nodes of the hole and nodes of the fastener, at each node location around the hole. While you can use gap elements, why not just use a general contact solution? The gap elements have some good uses, but I am not sure this is one of them.

- With gap elements, you will only get the solution for a neat fit. If you do a general nonlinear contact solution, you can see the effect of hole clearance. Of course, then the Kt will be a function the load level.



Brian

http://www.espcomposites.com

 

[Rhei]

Hi,
what do you mean when you say: " I manage to define them between 2 nodes but for some reason NASTRAN will not solve" ?
Do you get a FATAL error?

Also, have a look here:

http://www.6dof.com/index.php?optio...ZUFszrskT2+9RBAAAAwMAALiVrAAIBwAAeAEAAC22FAA=