analysis of a circular plate of 72m dia and 7 mm thickness using ansys wb 4

[hari sankar]

i need to analyse a circular membrane of 72m dia and 7mm thick using shell element in ansys workbench. the edges are fixed. I am getting very large deflections at the bottom. i have turned on the large displacement option 'on' in analysis settings. am not able to validate the result with roarks formula. what am i doing wrong?

 

[jhardy1]

For 7 mm thickness and 72 metre diameter, I would expect to see "very large deflections", so that much makes sense. (What sort of structure is this - presumably it's not a steel plate?)

Does Roark's have formulas for "very large deflections" in membranes? I thought Roark's mainly provides classical solutions based on linear elastic theory, which would not apply for a structure with these proportions.

http://julianh72.blogspot.com

 

[hari sankar]

yes, roarks provide solutions based on linear elastic theory. roarks consider deflections greater than half the plate thickness as large deflection, and certainly the deflections which i get will me far higher than that. Am not familiar with other theories dealing with very large deflections. i have used the material as carbon steel.
Actually my objective is to analyze an internal floating roof of a storage tank. Since i dont have high performance computer, i will have to rely on shell elements for modelling the entire structure. so i thought it will be better to start from simple plate structures before going in to the complex deck.
Then i found that, even the deflection of the circular plate due to its self weight with clamped edges is so large.

 

[karachun]

Maybe this clamped plate even can not carry self weight, it will broke IRL. Look at stress values.

 

[hari sankar]

sure, will look at that. thanks

 

[jhardy1]

If it's a floating roof, surely clamped edges would be an inappropriate boundary condition? You will need to model the various applied loads and the bouyancy support cases as self-equilibrating load sets to get anything resembling a sensible simulation.

http://julianh72.blogspot.com

 

[hari sankar]

so this analysis cannot be done using static analysis using sensible assumptions right? will have to go to transient right?

 

[ThomasH]

Hi
You can't user linear static analysis, it has to be non-linear. Then you can include the large deflections and the yielding in the material.
Static vs transient depends on the loading and if it varies in time. Is it a transient load?

Thomas

 

[rb1957]

How "flat" is this roof ? (will it deflect out-of-plane by more than it's thickness due to weight ? ... then it's no longer "flat".

What sort of loads are applied ? in-plane ?, out-of-plane ?

I think you're asking rather a lot of this plate ... 72m dia, 7mm thk. I'd be looking to add stiffeners to give it some "proper" bending stiffness, and then maybe reduce the thickness of the plate.

A fixed boundary condition would normally be thought unconservative, compared to a a simply supported assumption (unless you have designed a significantly stiff boundary attmt.

another day in paradise, or is paradise one day closer ?

 

[Shu Jiang PEng SE]

do not forget select element type with "membrane" function in your case.

 

[hari sankar]

thanks everyone, got a lot of quality inputs. will change the boundary conditions and will let you know the results.

 

[IFRs]

If it's floating there much different load condition(s) than if it is standing.
Some floaters had a cone built into the deck, those would not be flat when floated.
Floaters have buoyancy around the edge and some also in the middle.
Un-symmetric loads are not uncommon.
Depending on the pontoon design, the deck may not all be in contact with the liquid.
The single deck is connected to a rim of some strength and stiffness that typically varies.
Most decks are lap welded and have leg, column and other penetrations.
If it is on its legs the deck is more of a membrane with initial deflections.