HyperSizer Support Forum

Engineering & Analysis Methods => Panel Buckling => Topic started by: cmcho on November 10, 2008, 08:01:11 AM

Title: About maximum Y-span length in panel buckling analysis
Post by: cmcho on November 10, 2008, 08:01:11 AM
I have been working on the unstiffened panel.
It was strange when I found the thickness of the panel which is a part of upper skin under +3g load is designed to have almost  the same thickness as that of the corresponding panel of the lower skin under -1.5g load by the panel buckling criteria.( I assume the thickness of the upper panel is twice than that of the lower skin)  Both panels have curvatures. Applied load in upper skin panel is definitely larger than that in lower skin panel.

The reason I found is the maximum Y-span length in the buckling tab. These are the data  I have.

Upper skin panel:
   Nx,Ny,Nxy=-707, -8, -35
   X-span=1073, Y-span=340 (126 max), Radius of Y-curvature= 1021
   Design results: t=5.64, MS=0.35 by Panel Buckling (Discrete Design Vaule used)

Lower skin panel:
   Nx,Ny,Nxy=-358, -6, -3
   X-span=1070, Y-span=325 ( 317 max), Radius of Y-curvature= 2560
   Design results: t=5.64, MS=0.31 by Panel Buckling (Discrete Design Vaule used)

At this point, my questions are
  1) Why and how is the Y-span limited by different max values?
  2) Is the buckling load calculated with the max Y-span values?
  3) How do I take actions if I would like to evaluate the buckling load with the actual value of Y-span length?

Thanks in advance.
 
Title: Re: About maximum Y-span length in panel buckling analysis
Post by: Phil on November 10, 2008, 09:03:36 AM
HyperSizer assumes that a buckling wave on a cylindrical panel cannot go past 1/2 the circumference of the panel.  The axial or uncurved direction is assumed to be in the X direction and the curved direction is assumed to be in the Y direction.  Therefore the Y span (that is the longest assumed buckling wave length) is assumed to be 1/2 the circumference or

y_max = pi * radius of curvature

In the two examples you described, I can see by the max y span that these two panels have two different radius of curvatures.    For the upper skin, the radius of curvature appears to be:

radius = y_max / pi = 126 / pi = 40.1

for the lower skin, the radius appears to be

radius = 317 / pi = 100

Because the radius of the upper skin is so much lower than that of the lower panel, it is much less likely to buckle, so that explains the difference in panel strength that you are seeing.
 
Title: Re: About maximum Y-span length in panel buckling analysis
Post by: cmcho on November 10, 2008, 07:04:36 PM
Then, does it mean the Hypersizer can not evaluate the buckling of the curved panel with the Y-span length bigger than the maximum value?
What is the way in the Hypersizer for this kind of panel ?

For example,
Hypersizer imports my FE model (upper skin) as follows:

   Nx,Ny,Nxy=-707, -8, -35
   X-span=1073, Y-span=340, Radius of Y-Curvature=1021

In this case, Hypersizer shows the max value of Y-span as 126 and it does buckling analysis for the panel 1073 x 126 with radius of Y-curvature, which is not actual size, 1073 x 340. This is not correct or conservative. When I select the fully cylindrical panel option or flat panel in the buckling tab, the max value disappears and buckling analysis is performed for the panel 1073 x 340 of full cylinder or flat panel. This is not correct either because the curvature effect is different. Is there any way to reflect the panel buckling as it is?

You can see the appearance of the limit value in the figures with and without option of full cylinder.
Title: Re: About maximum Y-span length in panel buckling analysis
Post by: Phil on November 12, 2008, 12:57:16 AM
After looking closer at your problem, I discovered that the MAX value shown on the buckling tab is not correct.  It is showing you the max Y span in INCHES rather than in your native units, which I believe are millimeters.   For your dimensions, r = 1021 mm, the maximum y buckling length is actually pi * 1021 = 3207 mm.  If you convert 3207 mm to inches, you get  126 inches, which is the value being displayed on the form as the max value. 

If you put in any value of y-span less than 3207 mm, HyperSizer will properly evaluate buckling for your panel.

Another clue that you can see.  Try entering a y span of 5000 mm and then press Save.  You should see the letters for the max value turn red.  This indicates to you that you have entered a y-span greater than the allowed maximum of 3207 mm.  Notice that when you enter a y-span of 340 mm, the letters are not red, indicating that this y-span value is allowed.

This is strictly a display problem where the max value is not converted to your native units before displaying on the screen, but I can assure you that HyperSizer is correctly applying your dimensions of 1073x340 when doing the panel buckling check.

This issue will be added to our list of fixes for a future update.
Title: Re: About maximum Y-span length in panel buckling analysis
Post by: cmcho on November 13, 2008, 09:51:21 AM
Thanks, Phil.

I got it.

It was hard to understand  "half the circumference". This means the curved panel can not go pass the half of the full cylinder. That is, arc angle of the cylndrical panel should be less than 180 deg.
Title: Re: About maximum Y-span length in panel buckling analysis
Post by: Phil on November 13, 2008, 11:04:53 AM
It was hard to understand  "half the circumference". This means the curved panel can not go pass the half of the full cylinder. That is, arc angle of the cylndrical panel should be less than 180 deg.

Yes, that is correct.