Local Buckling of Tube

[rjoto]

Hi all

I've browsed this forum for a while, have found it very helpful, and recently decided that I'd like to get more involved.

I have a conveyor that I'm rolling things across. Recently some excessive forces were applied to the top of one of the rollers by another tube on top of it and I ended up with a large dimple in the top of one of rollers. It's an end supported galvanized steel tube (1.9" OD, 12 ga/.1084" thickness) that I have subjected to bending in the middle (by having another much larger tube on top of it). I'm familiar with bending and general column buckling but this seems to be some sort of local buckling and I've never come across that before. Can someone help me identify this failure mode and help me figure out how to calculate the strength of the tube?

I'm trying to figure out how much more strength I would get by going to solid rollers. A simple MOI analysis says that the solid roller should be about 2.6X stronger in bending.
pi*(1.9^4-1.683^4)/64=.246 for hollow
pi*(1.9^4)/64=.640 for solid

But I have a feeling that it should be higher than this because a solid roller shouldn't have any local buckling issues.

 

[rb1957]

"forces were applied to the top of one of the rollers by another tube on top of it" ... a contact problem ?

Quando Omni Flunkus Moritati

 

[rjoto]

Yes I suppose that's an element to it. But the fact that the conveyor is hollow, the tubes are perpendicular to each other, and the conveyor roller is not fully supported makes it much different than any of the contact situations that I'm familiar with or that I have formulas for in any of my books. So I guess even if I need to go down that route I still have no idea where to start with that. And I'm not convinced that that would be all of the solution.

 

[rb1957]

what i meant was i think it's more of a impact problem (the roller deformed the "thin" wall of the tube) than a buckling problem

Quando Omni Flunkus Moritati

 

[Compositepro]

I cannot tell you how to calculate anything, but what you describe is a very common failure mode for "thin-walled" tubes. The contact pressure tends to flatten the tube wall curvature at the contact point and, as a result, the resistance to buckling due to the flexural load goes down. I suppose you can calculate the following:

>the local wall buckling strength based on curvature of the wall (an effective tube diameter).
>the local wall curvature based on contact force.
>the compressive stress at the contact point due to flexural loading caused by the contact force.

 

[Tmoose]

http://pipeng.com/index.php/ts/itdmotdiam005b/

??

 

[GregLocock]

Can we se a photo of the dent? I'm guessing a diamond shape with creases.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[rjoto]

Here's a picture

 

[Tmoose]

Hi rjoto,

the dent in the picture appears to have a uniformly curved bottom. is that an illusion?
It also appears to have been "dented" so the tube is about half as "deep" as when it was round.

It was created when one (larger diameter) tube was forced against another at right angles, somewhat like this?

https://encrypted-tbn0.gstatic.com/...HOKToe3miwAZYCtvx4jq9s5i0VxpUiVAXCPTvkWRlRRFx

Was the load, and resulting denting, similar to what you are trying to design against in current service?

 

[rb1957]

i see the surface of the dent the same way ... as thought it's popped through ?

Quando Omni Flunkus Moritati

 

[rjoto]

Tmoose,

"the dent in the picture appears to have a uniformly curved bottom."
That's correct. Not an illusion.

"It also appears to have been "dented" so the tube is about half as "deep" as when it was round."
Significantly less height but not half (see picture)

"It was created when one (larger diameter) tube was forced against another at right angles, somewhat like this?"
Yes

"Was the load, and resulting denting, similar to what you are trying to design against in current service?"
Yes

 

[chicopee]

Can you provide a picture of the larger tube surface that you think created this dimple. The depth of the dimple does not seem consistent with its profile shown in the second picture for two tube surface area contacting each other to cause this problem.

 

[dhengr]

Rjoto:
It looks like the pipe that applied the load moved some distance downward, the depth of the dent, and then hit the solid parts on either side of this roller. At the same time, the full length of the roller (btwn. end supports) would have this load applied as a concentrated bending load and cause some roller defection. The second picture (the side view) may show some slight yielding of the roller in bending, but not much. Put a straight edge on the bottom surface of the roller to check this permanent set (deformation). Put a straight edge on the center seam of the roller, and you may see a little more apparent set, but some of this will likely be due to the buckling deformation right at the dent. These two conditions must have been compatible at the time of denting, but I doubt that the pipe was rolled across the conveyor. I’ll bet that loading pipe was dropped on the roller at the dented point, or the dent would be longer, along the axis of the roller, and maybe be less symmetrical along its length. In the case of the denting deformation, as rb1957 suggested, that’s a contact stress or Hertz stress problem, and there will be some deformation in bearing yielding and in localized buckling as the roller wall gets thinner. The beam bending problem is pretty straight forward, except the denting and buckling start to change the beam cross sectional shape and mechanical properties. Then throw in some indeterminate plastic deformation in the area of the dent. These two loading and stress conditions should be worked together to approach a compatible deflection/deformation condition. Your MOI calcs. are o.k., but note that you gain much more, per pound of material, by thickening the roller walls some amount. The middle half of the diameter doesn’t add as much as that thicker wall at a more favorable position on the cross section. Quit dropping pipes on your conveyor rollers and this problem will go away.

 

[rjoto]

chicopee & dhengr,

Sorry I can't post a picture of the larger tube because it's a highly proprietary process. The reason it deformed so much is because it's part of a custom build machine and there's extra force applied on top of the larger tube. I know exactly what's causing this and nothing was dropped. But it's part of the process so I need to find a way to accommodate those forces.

I realize that I would get more bang for my buck with thicker roller walls but the weight of the roller is negligible in comparison to the forces applied to it and there's a practical limit to what kind of tubes are commercially available. The MOI of a solid roller is still about 2x what it is for the thickest hollow roller/tube I can find.

I know that going with something solid is probably going to be my solution but I would still like to know how much stronger this will make it.

 

[rb1957]

this was intentional ??

if you make the roller tube a solid bar, what'll happen when the two substantial things bash together ??

please explain ...
1) the roller is in the right spot ?
2) the "tube" that hit the roller was doing some by design ??

this just doesn't sound right ...

Quando Omni Flunkus Moritati

 

[Compositepro]

The issue is not the MOI of the roller. It is not failing due to flex it is failing due to wall buckling. A thicker wall or solid roller will not buckle.

 

[chicopee]

Something else is at play to cause this formidable dimple and to me it does not appear that two tube surface areas contacting each other cause this problem; it looks as if there was a concentrated load part of the larger tube that caused this problem.

 

[rjoto]

"if you make the roller tube a solid bar, what'll happen when the two substantial things bash together ??"
That's part of why I want to calculate the strength of the hollow roller. So I can get an approximation of what kind of forces I'm dealing with.

"the roller is in the right spot ?"
Yes

"the "tube" that hit the roller was doing some by design ??"
Yes

I know this all sounds kind of weird, but it's hard to explain exactly what's going on without getting into proprietary information. I appreciate the help trying to look deeper into the situation to try to see if I'm asking the right questions, but really given the IP restrictions in this situation, all I really want to know is how to calculate the force that can be applied to a tube before it dents. If I haven't given enough information about the situation to figure that out, that's fine, but I just figured I'd ask.

 

[rb1957]

man, that's odd ... well, if you're going to be bashing things together then you have to make them stout, but that can make the problem worse ... if it's an enforced displacement. and i'd make lots of spare parts.

Quando Omni Flunkus Moritati

 

[tbuelna]

Sorry for being late to the discussion.

The roller in the picture does not seem to show a buckling failure. A buckling failure mid-span would look like a "kink" in the tube wall, and not like a deeply concave localized dimple. In fact, based on the description provided of the tube OD, wall thickness, span, etc, I can't even see how a foreign object passing between the large/small rollers would produce that sort of localized deformation in the small diameter roller without bending it far more. In order to produce that sort of localized plastic deformation in the tube wall, the opposite side of the tube would have to have been supported by something.

Consider it from this standpoint. Imagine you wanted to form a dimple in that 12ga. steel tube wall with the depth and corner radius shown in the picture. Do you think you could produce that dimple with the tube simply restrained at each end and unsupported in the center?

It took a huge amount of force to produce that dimple in the tube wall. In fact, there are formulas used by manufacturing engineers to calculate the amount of force required to produce that size and shape of dimple in a steel part. Making your roller from solid bar won't likely resolve the problem. Instead it will simply shift the stresses and failure point to some other location of the roller.