Elmer Discussion Forum

The attached xml was able to recreate the example and run to the correct solution without having to edit the sif.
Although the sif options were not the same as Mika's sif the resulting answer was the same.

Kevin

Wow, well done! I spent quite a bit of time trying to learn the script for the shell solver but there were still some areas of confusion for myself, an Elmer beginner. This will definitely help connect the dots.

I'm going to piggyback on the initial post. Another portion of the project involves replicating deflection data for 4-point bending and ring on ring testing using FEM analysis. I'm brainstorming how to create the mesh in order to apply boundary constraints for use with the shell solver. Thus far my Elmer analysis has entailed placing constraints/forces on edge boundaries in 2D meshes (SMITC, shell solver) or using surface boundaries in 3D meshes (linear and non-linear equation sets). Is it possible to place a boundary within a rectangular plate on a 2D mesh that will allow a load application? Attached a crude image that may help explain. Thanks.

I added internal edge elements which Elmer accepted without complaint. However I could not do an edge load in SMITC nor ShellSolver. I then made an internal row of elements a separate body so I could load only that row. Then ShellSolver calculated 4 point bend, you could approach a line load by making the strip of element smaller in size. Kevin

Mika's suggestion to use the shell solver works well. I have used the files attached to replicate the deflection solution. Does anyone know an easy way to add "Calculate stresses = True" to the shell solver?

Or does this require the use of two solvers at the same time?

I have tried but not found anyway to get shellsolver to output stresses. I looked at the code and it appears to calculate them, but the postprocessing output is not there.

Unfortunately the stresses and strains are not yet generated via postprocessing. Some additional coding would be needed to add this functionality.

-- Mika

Can the shell solver be used to define the deformed shape, followed up by second solver (elastic plates for example) to calculate stress and strain using the deformed shape?

I have tried several attempts at this method - but as of yet been unsuccessful. It seems like a valid work around.

Using the linear elastic as the second solver has also been unsuccessful, likely because of the difficulty in setting the plate thickness. Example below,...

I have tried that and it did not work for me.

Kevin,

In regards to the examples you posted, (internal edge element addition and internal row of elements as a separate body) how did you go about manipulating the original mesh to include these additions? Did you use a mesh generation software or are these additions possible in Elmer?