Buckling of Flat Plate with Step Change in Thickness

[aerohead56]

Are there any referencable methods for determining how much step change in thickness needs to occur in a plate to provide at least simple support for buckling? We are assuming a 50% change in thickness forms a support. I've heard of an old BDM method that said 20%, but we can't use that reference.

Thanks,

Ilex Brandenberger PE

 

[40818]

Bae did a load of testing back in the days and produced stress office report 51 which semi-empirical and was based upon test data.
Simple support may not be achieved, and if you dont then your buckling allowable drops accordingly. Its not a simple case of having a pad-up of certain thickness, as its the relation between the support stiffeners and thepanel geometry and thickness. If you have a thin panel, then you dont need much in the way of stiffeners, conversely, a thick panel would need massive girders to give the simple support condition. There is an ESDu (cant remember off the top of my head which one)which if you work backwards gives you the required stiffener inertia to achieve simple support.
Have a look bsck into some diagonal tension work by Kuhn in the Naca's to get started.

 

[aerohead56]

The panel in question is ramped up into a pad-up over a spar, so I assume that there is plenty of moment capability in the spar. I have a copy of SOR51 handy, so I'll take a look at that. If you can remember the ESDU number that would be greatly appreciated.

 

[40818]

Damn, sorry.
I forgot to look it up today for you.
However, i recall that the latest airbus implementation of SOR51 incorporating other improvements, a pad-up step of double the panel thickness is used as a differentiator in the analysis. Will try to remember the esud on monday.

 

[40818]

Damn again.
Sorry, so busy at work that i keep forgetting about this.
In my alchohol filled stupor presently, i believe the esdu is 02.03.02.
You can work out the required second moment of area value to obtain simple support. Using this in conjunction with SOR51 ( which uses half the panel width as effective in the suport equations).
Then if you have adequate support your K is as Simple Support.
However, if you dont meet the SS required level, you move down the graphs to get a reduced K which is applicable to the actual I level. Hope that makes sense?