Theoretical neutral fibre of the material 2

[soljarag]

Hey,

I'm currently using Solidworks to flatten a compound shape using the plugin BlankWorks. It will only flatten solids that are less than 15mm thick, so I have to use surfaces for anything thicker. In an email they explained:

"the user should build the input surface as closes as possible to the theoretical neutral fibre of the material (rather than just the middle). Our software assumes that the input surface is at the neutral fibre."

So I was wondering if anyone knows how to calculate where the neutral fibre is on a steel plate

Thanks!

 

[RPstress]

The neutral fibre is almost certainly the neutral axis which is the centreline of the plate, the plate being homogenous and isotropic. In small deflection bending theory material on the neutral plane does not strain when the plate is subject to bending.

 

[Felipe31500]

RPStress,

Then, how do you define the neutral fiber for a non homogeneous material, e.g. for a bi-material?

Felipe.

 

[soljarag]

Thanks for the reply guys. I forgot to show you the shape I need to form

http://img526.imageshack.us/img526/7113/shotdg5.jpg

they can be up to 200" dia and 6" thick


so when I go to build the surface in solidworks (since surfaces have no thickness) I need to figure out where in the 6" is the neutral fibre


Since the bend dia. of the shape is much larger than the thickness, I assume it has to be pretty close to the middle, but I want to make sure

 

[soljarag]

I forgot to mention, in that screenshot, the model is a solid model, not a surface model

 

[MikeHalloran]

To a first approximation, the neutral surface is halfway through the thickness of a homogeneous material. A closer approximation used in small radius sheet bending puts it at 0.3 to 0.4 x T, depending on the hardness.

... but it depends on how you make the part, too.

E.g., I think for stretch forming, the neutral surface doesn't really exist, or could be considered to be somewhere inside the inside radius, since all the material is in tension while forming.

Do you have a clue how the parts will be formed? The old- timer in charge of that operation can probably tell you what you need to know, at least indirectly. I.e., he may not understand the question you ask, but if you can get him talking about what he does, you can extract what you need from your notes. Pay attention.






Mike Halloran
Pembroke Pines, FL, USA

 

[limct]

distance from the inner surface to the neutral plane, let say X, is a variable factor depending largely upon the ratio of the stock thickness T to the bend radius R, the optimum value may be considered to be

R<2T X=0.33T
R=2T to 4T X=0.4T
R>4T X=0.5T
Best regards,
ct


Best regards,
ct

 

[RPstress]

Felipe31500: for non-homogenous materials things get complicated. The treatment is essentially an extension of what is done to calculate how things like bi-metallic strips behave. Depending on the nature of the inhomogeneity there may still be a neutral plane or there may not: the neutral plane for bending about one axis may be different from that for bending about another.

For soljarag's problem (and I cannot view the image, that site is blocked to me by our net nanny) I assume he/she has a thick curved plate of some sort and wants to use small-deflection (linear) finite element analysis to investigate flattening it. If the IBR is less than about 10 times the thickness, then the neutral plane will indeed begin to shift significantly, as limct points out. However, if a real plate is flattened and is going plastic then the neutral axis position will vary as plasticity increases. This could be hard to allow for. However, if your radius is large compared with thickness it might not be a significant source of error.

Also, if the deflection to flatten the plate is more than a couple of times the plate thickness or about a fiftieth of the plate minimum dimension then linear analysis will not apply. (I was simply assuming the analysis is linear from the use of a plate-type analysis - if Solidworks is happy to tackle the problem then that's fine.) Also, all this sort of analysis will almost certainly only apply to bending deformations - if the part is being stretch formed then the analysis may not be appropriate.

The neutral axis shift for a curved surface can be approximated by a curved beam analysis such as presented in Roark, or use limct's rule of thumb, which doesn't look a mile out.

Depending on what you want to find out about the flattening, the analysis may or may not be useful - caution is advised if a real item is to be flattened using information calculated like this (such as max forces required or over-bend to prevent springback in the event that the part is being permanently deformed).

 

[soljarag]

Thanks for the reply guys

These shapes will be Hot or Cold pressed, using a die... they are pressed in a couple steps (they press one part, then move it, and press another section)

The most common metal they use is A-516 - Grade 70

The main information I want to find from Solidworks is an accurate blank shape,

Right now they cut out a really rough blank shape (using a 10 year old DOS program someone wrote) Then they form it, and go back and trim it to the shape.

 

[MikeHalloran]

Even with just one press hit, getting an accurate blank would be iffy, because of variations in the material. With more than one hit, you add positioning errors.

I think the best you can do is arrive at a 'less rough' blank, and I suspect the easiest way to do that is to etch the surface of a rough blank with small circles and lines, as are used for measuring strain, but additionally identify enough of the circles so you can identify where the edges of the trimmed finished part were in the blank.



Mike Halloran
Pembroke Pines, FL, USA

 

[jte]

soljarg-

So why didn't you just tell us you're trying to evaluate how hemispherical heads for pressure vessels are formed?

jt