Twisting a 160x80x3mm Rectangular hollow tube.

[Jay_]

Dear Experts,
I rarely post in this forum, but we have a project that consists of twisting the above mentioned hollow tube for a 90 degrees.
I have read that people tend to heat the tube, or to put a pipe inside of it to prevent collapsing or even sand and to pressurize it a little and then to start twisting.
My problem is that we can't heat, or put a pipe inside (since its a rectangular shape) and the section is fairly large and requires a lot of torque applied.
Some people cut the four corner of the tube and then grind diagonally along each face of the tube and then twist and then weld the cut edges. but I'm not sure than I can do that sine it's an architectural project.
Any insights??


Detailing is a hobby,

 

[3DDave]

I thought there would be suppliers that offered this - Google is not helping me find them. What it suggests is that by the time the short faces yield the large faces buckle. At least in square tubes all sides are seeing the same load so the loads are uniform.

In thin wall aluminum tube they stuff the tube with plastic strips but I doubt that would help you.

I think the cheapest way is to develop the four faces as flat patterns and have them cut, twisted, and then weld the corners and grind to smooth it. Other than that incrementally twisting with an internal plug might work.

 

[robyengIT]

I agree with 3DDave and as much easier as longer the final tube

 

[HTURKAK]

The picture depicts square shape rather than rectangular.


You may try to twist open rectangular tube ..If the local bucklings which may develop at the end surfaces is not acceptable , I will suggest plug one end and fill with dry sand and compact with rodding then plug the other end. Then twist..

I have found this video for twisting of square tubing..

https://www.youtube.com/watch?v=46W3KE7mbFw

 

[LittleInch]

it's an architectural project

Nothing more needs to be said.

Tell him or her you can do it in cardboard or machine it / twist it from a solid bar or maybe 3D print it, but a rectangular steel tube (how long exactly) just isn't going to work, at least not evenly.

Could you do two 80 x 80 tubes then glue them / weld them together after bending?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[MIStructE_IRE]

Is this subject to any load? I don’t fancy the buckling calc on that!

 

[rb1957]

could you 3D print this ?

why can't you heat the tube ?

What material ? (Steel or Aluminium)

If you need it to be structural, hide a structural element (tube) within the decorative tube.

another day in paradise, or is paradise one day closer ?

 

[WARose]

If you can't do it.....you need to find a specialty steel fabricator. I've had them fab some weird stuff for me over the years.

 

[r13]

The length will have significant impact on the level of difficulty, wouldn't it? Rectangle will be tough though.

 

[dhengr]

Jay_:
Take a good look at some good Strength of Materials textbooks, particularly torsion on open and closed sections. Study how they warp, and the stresses caused by this twisting. Advanced Mechanics of Materials and Theory of Elasticity texts dive a little deeper on the subjects. You say a 6” x 3” x 1/8” struct. tube, but you don’t say anything about the material you are using, the length of the tube, how it is loaded, and the loads, or what it is being used for, etc. These might all be important to how you would approach this problem.

I would break form this tube, with three full sides and two half sides, with the split probably being on a 6” side (that is, 3” & 3” for that side). Now comes the tough part, how to grab the three full sides, at the ends of the tube, to twist it. You have an open section, and I suspect that mostly what you will see is a sliding action of the two half sides at their split joining edges. You want to impart a twisting action at the ends and your grabbing features must be able to allow some warping of each side too. When you have a little more than enough twist, tack weld the split seam on the one open side, and then fully weld that joint. I would like to apply some tension to the tube while I am twisting it. For concept only…, make two mandrels, 8” long and 5.75” x 2.75” to fit inside the tube 2-3” and be fillet welded to the end of the tube. Put the two mandrels into chucks in a lathe and twist the tube some angle past the desired twist, and release the torque to see what the final twist is. Cut a few inches of scrap off each end to remove the mandrels. For the open/split tube, use 1.5” fillet to the mandrel on each of the 3” sides. Use a couple large C clamps at each mandrel on the 6” sides, to allow the split longitudinal edge to move lengthwise during twisting. You should end up with much less torsional stress built into the final tube with the split tube approach, but the longitudinal welding might cause some distortion. You will waste a few inches of tube at each end to do this. You will be cutting off some yielded material, distortion, buckling, etc. You might want to stress relieve the tube after forming.

 

[JStephen]

Jay may BE the "specialty steel fabricator" in this case. It's not a common fabrication.
First thing, I would try to work with the heaviest wall I could to avoid the buckling issues. The 3mm is awfully light.
My idea is to put it in a press, with wedges under it and on top at one point, and wedges sloping the other way at another point just down from it. Press that assembly, slide it down, and repeat. You could vary the twist angle by varying the wedge slopes and by varying the spacing between the wedges. It's hard to visualize how the finished product would look. One advantage of that technique is you could try it on a short piece, and wouldn't have to screw up (<== note pun) a 40' piece testing how it worked. Hot forming would be an option.

You could cut rectangular plates and insert them in various ways to help hold the shape, if that was required. I'm not sure if sand would help there, as the shape could still skew (ie, shift from a rectangular cross-section to a rhombus cross-section). Putting rectangular plates in while forming, I'm not sure if they'd come back out or not.

Rectangular tubing is normally a high-yield-strength product, which complicates it. I don't know if there are any options to avoid that.

 

[HTURKAK]

Dear Jay ;

I know that the use of sand filling is an option to twist of circular and square tubes. I am not sure whether will work or not for thin rectangular tubes.

The following idea may seem crazy but may work.. I think it is worth to try...

- Plug one end,
- Prepare frozen sand plates thick. may be 10 mm which will be used for insert , ( prepare a mould and fill with sand and soak with water and freeze to -18 degr ),
- Fill the tube with frozen inserts with papers between them to avoid sticking of inserts ,
- Closed the other end ,
- Divide the total length e.g. 6 and twist each section 15 degr , to obtain uniform twist along the full length.

 

[rb1957]

I like JStephen's idea for plates inside the tube, regularly spaced say 1' apart. then how much twist per ft ? then twist each segment. The plates would need to be a very neat fit. Installing them too would be tricky … maybe have a central spline (threaded), maybe run a nut down the rod to the location for the plate. Maybe build the tube out of segments ? Weld them together, maybe a finish machine operation to make the surface smooth ?

I still like my idea of 3D printing too.

another day in paradise, or is paradise one day closer ?

 

[LittleInch]

I mentioned 3D as well, but we don't know length, any stresses/forces, but if you could do that then paint it silver....

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Jay_]

Thank you so much for the suggestions!!
Been so busy at work..ill update you tomorrow morning on the status. Many suggestions helped us. We are still making samples. In the meantime I’ll appreciate more feedbacks if there is any..
Material is carbon steel st37-2, length is 2000mm and with a quantities of 100.
I will reply to each suggestions tomorrow morning!

Detailing is a hobby,

 

[r13]

Twist a longest tube the shop can handle to minimize the torsional stress, and potential for buckling.

 

[Compositepro]

Make tooling from !/2" thick steel strips 8 ft long with rectangular (or square) holes milled in them to closely match your tubing. Slide these over your tube spaced about 12" apart and twist the tubing in progressive sections.

It might be easier to make the tooling by welding plates together.

 

[CANPRO]

If this isn't carrying any load aside from self weight, this might be an option:

Use a straight piece of tube or pipe with a series of brackets placed along length of the pipe to match the rotated section. Then you can use a thin skin plate fastened to the rotating brackets. If you cut the skin plates in the right shape (I don't think they would be simple rectangles) they should conform the brackets easily and with no bulging/buckling. I think this might end up being cheaper/easier and might give you more consistent results as well.

 

[WARose]

Jay may BE the "specialty steel fabricator" in this case.

Sounds like he needs to sub this one out....or invest in some new equipment.

Speaking of that.....they do make machines that can do this (at least with pipe):

https://www.youtube.com/watch?v=3agjHb9x9qM&t=18s

Looks like they make it square in spots.

 

[EdStainless]

These are fairly short that large to twist. They need to start with about 1' extra on each end, they will put solid plugs into the ends so that they can grab them, and then just twist. Filling is often the last choice as it is a PIA. Warming tubes is common, and it doesn't take much to help. Talk to a local steel fabricator.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed