ASD CONCRETE 3D TRansient Analysis Mesh sensitivity

Sylwester · Post by **Sylwester** » Mon Aug 26, 2024 7:46 pm

Thank you very much.
I have checked this solution and it also provides a result independent of the mesh density.
But why does the stiffness of E have to be so much higher if we are treating this part of the plate as elastic?
Why 100x more and not 10x or 1000x?

STKO Team · Post by **STKO Team** » Tue Aug 27, 2024 7:51 am

But why does the stiffness of E have to be so much higher if we are treating this part of the plate as elastic?
Why 100x more and not 10x or 1000x?

Tips:
Ask this to the new ChatBot we implemented in the forum:

I have a slab modeled as a shell. Its Young's modulus is E. I want to make a portion of that slab rigid. What alternatives to rigidLink do I have?

Here's the rationale behind it:

I want to make a portion of a deformable slab (with Young's modulus E) rigid
I should use kinematic constraints (rigidLink in this case) to enforce correctly the constraint
Alternatively I can impose the constraint approximately using a penalty approach
In the penalty approach I "penalize" the deformability of the stiff part using a stiffer material Estiff = alpha*E
In theory, to obtain the correct result alpha should be INFINITY, but you can't due to numerical issues.
Assume you are fine with a 1% error, you can make alpha = 100
Likewise, if you want an error of 10%, alpha = 10

Sylwester · Post by **Sylwester** » Wed Jan 15, 2025 8:05 pm

Dear STKO,
I wanted to return to the topic of the sensitivity of the result to the density of the MES mesh.
In my attached model, I have done all the instructions in this topic, regarding elastic shear and the connection of the slab to the column also as elastic.

: Displacement.png (37.84 KiB) Viewed 2472 times

I performed the anlysis depending on the density of the mesh. And the smaller the mesh, my deflections differed. Now I don't know which result can be good? I noticed in the results that the value of stresses in the steel (over the columns) does not vary depending on the mesh. The same is the case with the damage d- of the compression concrete. The big difference I noticed is that the smaller the mesh, the damage d+ has a different range. But should it affect the result that much?