For my master thesis I am comparing the structural efficiency of slabs with stress-aligned ribs vs. traditional slabs with straight, orthogonal ribs (more info on my previous post).
Originally I expected that a slab with stress-aligned ribs would be stronger than a slab with orthogonal ribs, but my first test showed the opposite — the variant with stress-aligned ribs has more displacement than the variant with orthogonal ribs:
Thus, I checked if the following parameters could be causing an inaccurate simulation:
Influence of the ribs. Both slabs have ribs with equal height and cross-sections, and a similar spacing; also the total area of the ribs’ surface is similar in both slabs (5.10 m² for the orthogonal ribs, and 5.29 m² for the stress-aligned ribs). This was done in order to make the shape of the ribs the only influencing factor for the slab stiffness.
Mesh resolution. The images below show that, when the mesh resolution of a slab is increased, the displacements increase as well:
Thus, to avoid the mesh resolution from affecting the accuracy of the simulation, in both the stress-aligned ribs variant and the orthogonal ribs variant I used the same value (0.05) for the “Mesh Resolution” parameter of the “Mesh Breps” component (this component is used to make the vertices of both slabs’ meshes coincide with the vertices of the ribs’ meshes, so that they can be connected with line joints).
Here’s the mesh of the slab with orthogonal ribs, with the average area of its faces:
Do you think that these meshes have equivalent resolutions, and thus I am doing a fair comparison of their displacements? In my opinion both slabs have almost identical mesh resolutions so I think this is not affecting the accuracy of the simulation. The same happens with the meshes of their ribs.
And lastly, which other parameters I should check besides the influence of the ribs and the mesh resolution, in order to have a fair benchmarking between both variants and thus prevent any inaccurate results in the Karamba simulation?
Your ribs dont seem to be connected to each other. This changes the structural behavior significantly.
I would also first try to understand why this would be the case and under which circumstances, to understand what the theoretical outcome should be (probably, it has sth to do with the fact, that principal stress lines, do not have any shear stresses). The principal stress lines are those of a flat slab. What happens if the original flat slab becomes thin relative to the ribs, or just as thin as your ribs? (… not an answer, I am just thinking out loud)
Also: Would suggest reducing your GH file to the minimum components necessary, so that others can help more easily.
What do you mean with “convergence analysis”? Does it mean to increase the mesh resolutions until the results become stable and stop being influenced by the mesh resolution?
Yes, I will work on connecting the ribs with line joints to see the effect of this.
Indeed, since principal stress lines have no shear stresses, in my opinion this means that ribs that are placed along them will not experience torsion and thus can become very thin. These two slides I prepared for a university assignment explain this idea better:
Because of this, I would have expected the slab with stress-aligned ribs to have less displacement and less “crumpling” on its ribs (due to torsion) than the slab with orthogonal ribs. But the opposite happened instead:
Exactly what do you mean with this? In my case I used a thickness of 0.50 cm for the slabs and 0.25 cm for the ribs, all with a steel material.
Sorry for that; the model was much more complex initially and I already simplified it a little bit, but I will make it much more simpler for the next time.
In any case, your intersecting elements must be composed of independent geometric elements, (independent lines connected to nodes, independent surfaces with joining lines between them) for the connection of the elements to be effective in Karamba.
