Practical experience with thick plate rolling - developed length?

[kingnero]

Does anybody knows the difference between theory and practice when it comes to rolling of heavy wall section cilinders from thick plate?

See the picture below. Assume gradual, cold rolling in many, many passes. No cracks on the outside surface.
Would the length of the plate before rolling equal the length of the neutral axis of the plate after rolling, or equal the inside diameter of the cilinder (obviously after rolling), or somewhere in between these?
I know what theory says (neutral axis, with plastic deformation both out- and inside the NA), however this does not concur with reality on smaller scale tests.

Please share your opinions...

https://files.engineering.com/getfi...ff5-b68d-64819049003a&file=vb_thick_plate.jpg

 

[racookpe1978]

Sorry, boss. Everything I have rolled has been much thinner material, and has required two tangent lengths at both ends to pass into the 3-roller machine. So the final OD/wall/ID checks required the original material be trimmed off at both ends to eliminate the two tangents so the final OD would be correct. Thus, whatever the expansion was on the outside, whatever the contraction was at the inside, I'd never see it.

 

[racookpe1978]

Frankly, I'd specify the ID.
Roll it to make the ID perfect.
Tack it when the OD and ID are sat.
Tack the ground out weld prep at the joint.
Weld he*l out of the joint.
Grind the OD flush.
NDE. Inside and out.

 

[DriveMeNuts]

Mathematically, I would say the NA is in the centre until the inside surface reaches yield as material will remain elastic when compressed.
Above Yield, the calculation is compleely different where the NA is at the inside surface and the thickness reduces as the outside surface stretches.
I'm not sure how that works in practice.

 

[JStephen]

I've never seen anything on the topic, actually.
For single-pass rolling, I would think you could work it out theoretically. The catch is that you'd have to assume what the stress-strain curve is beyond the yield point, and then the results would change if that curve changed.

 

[r6155]

See Roark’s Formulas for Stress and Strain - Seventh Edition

TABLE 9.1 Formulas for curved beams subjected to bending in the plane of the curve.
e = distance from centroidal axis to neutral axis measured toward center of curvature

Regards
r6155

 

[racookpe1978]

R6155

True. But are the Roark's formula's valid for a highly stressed thick plate? That's theoretical, true.

That's why I recommended grinding the weld prep so the ID will come out correct.
Rolling it so both weld prep's touch at the the center of the plate.
Weld it. Fill the weld prep region to a surface higher than the ideal ID or ideal OD.
Grind the weld smooth.

 

[kingnero]

Thx for all your replies. I'll check out Roark's, it might help me in the right direction.
For what it's worth, this question has nothing to do with welding or weld preparation, I'm merely interested in how much the NA shifts place...

 

[GoguGigi]

Hello kingnero , just a heads up about rolling a sheet of metal in many,many passes.
In our workshop we tried at one point to roll a 40 something mm sheet of plate on a machine that was strong enough to roll a 24 mm sheet in 2 passes . We also thought to roll it in many passes. The problem we faced was that, due to the number of passes the sheet actually became thinner by about 3 mm so be careful of this .
In the end we used shell made of forged metal and we machined it to achieve internal and external diameter ( it was for the shell of a distribution chamber of a shell and tube heat exchanger ).

 

[kingnero]

OK - thanks for the warning. The reason I mentioned "many, many passes" was to clarify that this thought experiment discusses gradual deformation, to mitigate the risk of cracking on the outside surface.
But thanks for the heads up, I had not thought that this would be a concern.