Load distribution on arched structure using claddings or FEs

[minoas96]

Hello to everyone!

I came across some difficulties regarding the distribution of loads to the elements of the stucture, using one time claddings and one other  FE meshing.

The material of the whole structure is steel S355

See the structure on the png "Structure"

The structure is a roof with a glass toping

Regarding the Claddings :

Because of the fact that my structure has arched members  both on plan and elevation views, I'm aware that claddings won't distribute the load correctly as long as they are not planar. So I have divided each cladding contour in triangles and the load distribution seems like in the png "1- claddings" (seems really odd).

Deformation because of the snow load is about 13cm

Regarding the FEs :

I converted the divide claddings (in triangles) into panels, set the panel calculation model as shell type of FE and the trasnfer of loads as simplified two - way. For meshing I used coons with the advanced parameters that are advised in multiple other forum discussions and the load distribution seems like in the png "1-FE meshing" ( I'm assuming the green colour means that no load in transfered)

Deformation because of the snow load is about 3 cm

Does anyone know which way is best and more realistic to distribute the load in such a structure and how to overcome the load not being transfered? mainly in the FE meshing procedure.

Thank you in advance for your time and sorry for the long post!

[Krzysztof_Wasik]

Hi @minoas96 

Claddings are designed to transfer loads to linear elements. If there are no linear elements, bars defined within cladding or on its edge loads are not transferred (there are no objects which can transfer load).

I would recommend using meshed /shell panels with analytical load transfer.

[minoas96]

 

Hi and thanks for your response!

I've already converted the claddings into panels and the result of the mesh is the one in the png.

Its seems pretty accurate but when I open the load distribution view it shows that the load is not being transfered in the whole surface (as shown in my first post) .

I will try to create the panels from scratch and will inform you for the outcome.

Thanks again!

[minoas96]

Hi again!

So I created some panels between some of the bars of the structure, with the advised meshing options , but it seems really strange and not consistent as shown in the png below:

But when I mesh the same panels but divided into small triangles ( a whole panel between bars, is divided in smaller triangle panels ) , it creates a more continuous flow of the mesh as shown in the second png below :

Can I accept the second solution as more realistic? If anyone has any idea please inform me!

Thank you in advance!

[Krzysztof_Wasik]

Hi @minoas96 

Second screenshot looks better

[minoas96]

The load distribution regions now in the view settings, has any power here for the meshing panels? Or it only has to do with the distribution via claddings?

Thank you for your responses!

[Rafacascudo · Accepted solution]

load distribution areas are only shown when simplified load distribution method is select for meshed panels on calculation model dialog , or in case it is a cladding

[minoas96]

Ok got it!

I would like to ask another thing if I'm not being wearing.

So, now that I have chosen the analytical meshed panels for the distribution of the loads to the bars, this means that the whole plate has a stifness which I don't want that to function like a concrete slab for example , as my structure is a roof and will be covered with glass.

I created a new material with the properties of the glass and chose settings like this :

 

 

 Are these settings close to the real conditions that i'm trying to create?

If you need the rtd file I'd be happy to upload it!

Thank you so much for your response!

[Rafacascudo · Accepted solution]

For the orthotropic panel option ,use small stiffness coeficients like 0,01 so the glass stiffness will be like a jelly.

Another option would be using a regular homogeneous panel , and create a jelly new material( tools/job preferences/material) ,inputing a very low value for "E" .

[minoas96]

With the reduction of the coefficients t,he deformation of the bars are more realistic (about 20 cm in SLS ) but in the buckling analysis the critical coefficient reduced dramatically :

 Regarding your second proposal , I already have created a new material with these properties (glass physical properties)  :

 Would that E be appropriate? Do you have any proposition for the reduction of E? 

Thank you for your time!

[Rafacascudo · Accepted solution]

You need to know what part of the structure is buckling for each mode checking its deformed shape .If it is the glass panel buckling, you can ignore as its stiffness is fake.

For the glass material "E" property , you can also input a small value so it behaves like a jelly  in a way that the panel acts like it is pinned on its edges

[minoas96]

Indeed , the deformation concerns the glass panel even on the 500th mode!

Thank you so much for your responses , you really helped me out a lot!