Buckling of Composite Curved plate under Compressive Load 2

[jpenche]

Hi,

I am studying the influence of several parameters in the critical buckling load but I Have some problems.

My working procedure is the following:

1-I use a theoretical method to calculate the load
2-I make a FEM to compare with this method( I usually use Nastran and Abaqus)
3-I see the error between both

My problem appears when I try to study the influence of the Radius.

I am using the ESDU 94007 to calcule the critical load and Nastran and Abaqus for the FEM and the errors that I obtain are quite big while error between Nastran and Abaqus are almost 0%. For exaple, in a 50in x 8 in plate with a radius of 15in I get a error=60%

The boundary conditions that I use are:
-Simply suported plate (image Attached)


I only have seen that the more the radius increase the less the error is. I looked for something about limits (Radius, a/b...)of the ESDU but I didn't find anything.

Does someone know which the problem can be?
Do you know about other manuals similar to ESDU that I could use to compare its results?

Thank you in advance

 

[SWComposites]

Which result is higher, ESDU or FEM?

The ESDU formula may have an empirical factor to account for imperfections. The smaller the radius the more sensitive a panel or cylinder under axial load is to initial imperfections. See NASA SP-8007.

See also buckling equations in:

Whitney, Structural Analysis of Laminated Composites

Kollar and Springer, Mechanics of Composite Structures

 

[ESPcomposites]

This program will also solve your problem. You can use it to compare your results.

http://www.espcomposites.com/software/eStability.html

Brian

http://www.espcomposites.com

 

[jpenche]

First of all, Thank you both for your quick answer. I have just finished my degree and I feel a bit lost in this new world.

The results of the ESDU are, in all cases, higher.

As for the books, I think I have not them but I will try to find because they would help me.

Anyway, thank you a lot.

Brian, I read the thread called "Buckling Composite Curved Panel" in which you give some advices to rybose and there I saw your website. I used it and I have to congratulate you. I think it is so good and the results are similar to my FEM results, so it discard the FEM error.
On the other hand my plates have thikness between 0.06 in to 0.24 and I would need Pro version.How can I get it?

I am going to start reading the Nasa document.

Thank you again.

 

[jpenche]

I forgot to mention that the ESDU provides a large number of buckle half waves ("m") comparing with the FEM.

For example, while the FEM gives m=3 the ESDU gives m=21 in a plate of R=30in


Thank you

 

[SWComposites]

ESDU is higher?? are you sure you are comparing results for the same boundary conditions?

 

[jpenche]

Yes,
it is quite strange, I know.
As for the boundary conditions, I first thought that the problem was there but at ESDU report only appears "Boundary conditions:Simply supported" and then while I was doing the analysis I saw that the error is very low when the radius becomes higher.For example is almost zero when radius =500.
I don't know which kind of simply supported boundary conditions uses the ESDU
In addition Brian's program gives the same results as my FEM.

I don't know which the problem can be

 

[SWComposites]

I would suspect then that ESDU is 4 sides SS; check the equations against a textbook.

 

[jpenche]

You were right!!

I read again the ESDU report and I saw that the boundary conditions follow the equations:

u ? cos (?x ) sin (?y)
v ? sin (?x) cos (?y)
w ? sin (?x ) sin (?y)

where

lambda=Pi*m/a
eta=Pi*n/b

and y,x are the coordinates of the points of the edges.

So finally,

I do not want this boundary conditions so I will use other theories that you have suggested like the kollar(I have not found the Whitney).

Anyway you have helped me a lot.

I will follow your thread and comments to learn more.

Thank you very much

 

[jpenche]

Hi again,

I upload an image with the boundary conditions of the ESDU, maybe someone wants them.