Best practice for modeling thin wall square vessels

Famous_Mortimer

I am working on a thin walled (10 gauge = .134" thk) square vessel and am looking for any thoughts from others on the best element to model this with. I am looking at it in a state where it is full of produce. Thanks - Mortimer

prop_design

I would use the ccx shell elements. they will expand to a solid. it's a weird formulation. but it seems to work pretty good. start by importing a surface mesh of your model. the choice of which surface can take trial and error. typically you use the midplane. but because ccx expands the shell to a solid it can cause errors. then you have to try the interior and exterior to see which will work. their are settings to control how the offset will work out too. once you get the hang of it, it's not too horrible. but it would be nice if there was a true shell formulation. it would make it a lot easier. it also has a negative impact on the model size. if your surfaces are flat then you can use the mecway shell element. that is a true shell so you won't have issues like with ccx. however, it pretty much has to be a flat surface to use the mecway shell formulation. so as far as i know the mecway shell is really a plate element. victor will know more.

Victor

Additionally, I think you would need to use CCX and nonlinear analysis for this because it's likely to be a nonlinear problem with stress stiffening.

badbunny

Just a few thoughts, some of which I'm sure you've already had...
1) There is a hydrostatic pressure loading for the product fill dead gravity load (assuming the product can 'flow' to fill out the tank), but in addition to this there will be a dynamic load from the product flowing into the tank, as well as potential sloshing from the fill or subsequent tank movement. Also if it is a sealed tank, can it see over pressure?
2) If its a square tank and static loading (and the supports are also symmetrical) consider using quarter symmetry
3) Be mindful of the stiffness of the tank supports and how the supports attach to the tank (ie be careful that you don't influence the behaviour of the tank with unrepresentative boundary conditions)
4) Same as (3) but for any stiffening structures on the tank
5) As Victor said this will be non-linear
6) Are there any thermal loads? (the thin tank wall will change temperature quicker than its stiffeners/supports so if your product is hot/cold consider thermal expansion/contraction loading)

Pete.