Shell vs. CBend vs. CBar

[m1esk]

I'm trying to model a bend of rectangular hollow-section tube. Typically, I have used Cbars to model and output the elforce and then apply various factors (stress concentration due to bend; extra shear from torsion etc.) to calculate the "applied stress" on the tube.

As a simple test case, I have setup a 90deg bend test case, separately, with CBend and CBar elements. CQuad model is used as a benchmark.

The bent tube is being pulled apart by a load. However, the stiffness of the 1D elements is almost double of the shell's. Also, CBar and CBend results are practically identical.

1. Is this the limitation of 1D vs 2D elements?

2. What would be the better approach in modelling thin-walled rectangular tubes with bend, on assembly-level model without using CQuads?

3. Would the bend stress concentration and torsional shear have been modelled within the 1D elements already?


Thank you all.

 

[dktoao]

It would make sense to me that the smaller your wall thickness to diameter ratio is the worse your results are going to be. 1D elements would not account for walls warping and changing shape under load. Like how with 1D beam theory as a guide I could design a pretty stiff column with just aluminum foil. Just my 2 cents.