Uniform Pressure on Shell

Helo,

I have a shell (Spherical Cap) and I want to obtain the buckling modes under uniform pressure (in a radial and not vertical direction)

In the Load component I use:

Type of Load: Mesh load

Orientation: Local to mesh

but in the Generation option if I tick Line Loads it crashes

and if i tick Point Loads I get an error saying I should do ThII manually.

Any ideas?

Thanks!

Attachments:

Shell-Uniform Pressure.gh, 38 KB

Replies to This Discussion

Permalink Reply by Robert Vier on July 14, 2015 at 11:32am

Do you mean 'crash' - Rhino crashes? or the component turns red?

For the THII component you could try setting the maximum number of iterations higher, or the tolerance higher so it calculates..

Best

Robert

Permalink Reply by marinako on July 14, 2015 at 11:43am

Hi Robert,

I mean Rhino crashes.

Just tried with 1000 iterations for Local to Mesh & Line loads and Rhino crashed again.

Also, is my methodology correct? Setting this parameters to Load Component in order to mimic uniform radial pressure?

I compare Karamba to Analytical solutions for Spherical Cap Shells under uniform loading and trying to understand why the results vary widely.

Thanks!

Permalink Reply by Robert Vier on July 14, 2015 at 3:48pm

Hi,

you have

A) fed a nurbs-surface into the 'MeshToShell' component. Grasshopper performs the conversion with some standard settings, but this meshing step needs to be controlled for finite element analysis - benchmarking >> bad condition of a face falsifies, resolution-dependence ...

B) if you use the karamba- or gh-mesher it still gives you bad results as your sphere has its nurbs-edge running through your cap. rotate the sphere 90° around x before and you are getting a nice mesh.

C) your supports are not defined just around its edge which i guess the benchmark would require?

D) you defined wood as the material, and there are some non-benchmark defaults for that i guess. also i am not sure but i think there have been some issues about the computability of shell element's materials with low shear modulus, and therefore the one of wood was raised. but you have to ask clemens for that.

nevertheless you can define your own material-properties with the resp. component

for me now, it calculates the first 5 buckling modes

good luck!

best

rob

Attachments:

ShellUniformPressure.gh, 36 KB

Permalink Reply by marinako on July 14, 2015 at 6:16pm

Hi Rob,

Many thanks for looking into this and your comments they are very helpful.

I have to use a low shear modulus as I am looking into the equivalnce of grid-shell and continuum in the context of timber shells. I understand how this can be a problem.

However, in your definition, and using steel as a material this time, if in the load component I change the orientation from 'global' to 'local to mesh' Rhino crashes again.

So my last question is: if I need the load to be uniformly distributed pressure (left diagram q1) rather than uniformly distributed load at the z direction only (right diagram q2), is it ok to use in the orientation parameter 'global' or it will give me wrong results?

I did a small experiment using the semi-circular arch and I get different results

Thanks for your input, really appreciated :)

Permalink Reply by marinako on July 14, 2015 at 7:49pm

The Load I get from the above definition is the one below which doesn't look like uniform pressure, thus my concern for falsified results.

Permalink Reply by Robert Vier on July 15, 2015 at 4:20am

Indeed it crashes if you change this, this should not happen in any case. It is caused by a negative value of the local vector, it will be fixed soon.

But if you change the vector to positive, everything does as it should.