K joint Ansys

erbometro
Member • Jul 3, 2011

K joint Ansys

Hey
I am trying to use Ansys to model a K joint of tubular section.
So my problem is, I want to check the strenght of K joint with tubular section by FEM analysis ( Ansys). So the point is how to create the model and how to mesh it.
I already know what is the strenght of joint by using some formulas given by CIDECT. When I want to model an example given in this book for a correct K joint.
I always found an equivalent stress higher than the yield stress of the material. That means that the joint is not well designed, whereas this joint is correct for CIDECT.
When I decrease the size of element, the stress is decreasing that seems really normal, But when I refine the mesh around the connection part of the joint. The stress at this part increases.
Any idea what is my mistake to model this K Joint?
Really appreciate if someone can give me a hand.
Charles

Replies

Howdy guest!

Dear guest, you must be logged-in to participate on CrazyEngineers. We would love to have you as a member of our community. Consider creating an account or login.

Replies

Capt Spark

Member • Jul 5, 2011

that may not be a mistake. It is not possible to comment without seeing any result plots, but there are usually stress concentrations at joints if you model it in detail
If you make the model finer, then sometimes it captures the effect of the concentration better which results in higher stress.
Can you post a few stress / mesh pics?
It is a good idea to cross check nodal solution (PLNSOL) and elemental solution (PLESOL) in ansys to find whether the results have converged. They shouldnt be much different

Are you sure? This action cannot be undone.
Cancel
erbometro

Member • Jul 7, 2011

Sorry for delay,
I use Ansys workbench, I am not very familiar with apdl command. I tried to insert some command in Ansys mechanical but I got no chart in trade. So I don't really know. I use converge function to see convergence. But I guess this one make vary number of element and get result of the stress. So that doesn't really help me to check if i am at th right solution.
Moreover the stress seems to be right located but always too big compared to standart.
I can send you some picture as you want.
Charles

Are you sure? This action cannot be undone.
Cancel
Aashish Joshi

Member • Jul 8, 2011

It would be great if you could post a pic of the section over here, both the solid model and the FE model.

Are you sure? This action cannot be undone.
Cancel
Capt Spark

Member • Jul 8, 2011

what I wanted to mean was that if you have a coarse mesh, it cannot be always presenting the correct geometry. usually it misses the correct radius and models something greater than it, resulting less stress intensity factor (refer the attached figure).
But if you continue refining the mesh, at some point the stress should not change for any more refinement.

Are you sure? This action cannot be undone.
Cancel
Aashish Joshi

Member • Jul 8, 2011

Refining the mesh is one option. Depending upon the geometry, you can also go in for a higher order element!

Which element type are you using anyway?

Are you sure? This action cannot be undone.
Cancel
erbometro

Member • Jul 10, 2011

As I a m using ansys workbench, I used Solid186 but I tried to model with shell181 by creating surface from solid. I got similar results.
I attached some pictures. But in this case, more I increase the mesh, more stress decreased always over the yiel strenght of used steel.
Fine mesh:

Coarse mesh:

Are you sure? This action cannot be undone.
Cancel
erbometro

Member • Jul 10, 2011

In fact still can manage to post some picture, internet is really lagging and that doesn't work. But I can send by mail if you are interrested.
send me email at #-Link-Snipped-#, I will send you picture of K Joint

Are you sure? This action cannot be undone.
Cancel