How to measure diameter of pressure vessels?

[Curiousgeorge]

Hi all.
A company I work for makes cylindrical pressure vessels that are essentially a cold rolled piece of stainless steel that is then seam welded and threaded bosses TIG welded in the walls of the vessel (sometimes near the ends so can disfigure the shape). Then the vessels are put in a machine where the operator then adds a dished end to either side and uses a belt-like steel apparatus to tighten the vessel's ends to, theoretically, make the vessel and dished ends fit perfectly so the operator can then tack the dished ends to the vessel before a machine then welds each seam.
However, in reality, the vessel and dished ends rarely meet up so after the belt-like apparatus is tightened and the operator tacks around the dished end, there is a gap between the seam welded part of the vessel and the dished end (photos would be useful here). Then the operator uses a hammer to force the raised part in and then aims to fill the gap with tacks.

Obviously this is not ideal and in a perfect world, both ends would meet up each and every time. So I am trying to get to the bottom of this. I am thinking that the first thing I need to do is find out if there is a variance in diameter in the vessels after they have been rolled and the bosses tacked in.
Please can I ask what is the best way to do this? Some have suggested a tape measure but I'm not confident that the starting point and end point will always be through the centre of the vessel so are there any tools out there to measure the 'circularity' of a circle?

Thanks all, I'd love to solve this issue.

 

[merryman]

Internal micrometer?

 

[Pete.]

Hi all.
Some have suggested a tape measure but I'm not confident that the starting point and end point will always be through the centre of the vessel so are there any tools out there to measure the 'circularity' of a circle?

Thanks all, I'd love to solve this issue.

Cut a circle from sheet that is 10-15mm smaller than the rolled ID. Cut some hand-holds in it. Simply offering it up inside the tube will tell you visually if it's out of round.

 

[Burdekin]

Measure the circumference and divide by pi.

 

[8ob]

Sounds like in an ideal world the cylinders would be corrected with some type of former or mandrel.

Bob

 

[daleyd]

I’m struggling to understand that a company involved in the manufacture of welded pressure vessels cannot fathom how to check the diameter/circularity of a cylinder

 

[Curiousgeorge]

I’m struggling to understand that a company involved in the manufacture of welded pressure vessels cannot fathom how to check the diameter/circularity of a cylinder

They might well know....but they don’t, to my knowledge, do this. However, I’m keen to be proactive.

 

[Curiousgeorge]

Cut a circle from sheet that is 10-15mm smaller than the rolled ID. Cut some hand-holds in it. Simply offering it up inside the tube will tell you visually if it's out of round.

All the replies are helpful but I like this one!

 

[MattH]

PI Tape.

 

[daleyd]

Weld them up on a jig. If the “length” of the plate they use to form the cylinder is identical every time then the diameter should be identical as well, the only way it can vary is if the material is a different size, or if it’s being welded in a different way - but given it’s a pressure vessel they presumably have (and are following) a documented welding procedure.

 

[Curiousgeorge]

Weld them up on a jig. If the “length” of the plate they use to form the cylinder is identical every time then the diameter should be identical as well, the only way it can vary is if the material is a different size, or if it’s being welded in a different way - but given it’s a pressure vessel they presumably have (and are following) a documented welding procedure.

I am not 100% sure but I think the material might vary in size. This is part of the reason why I want to ‘prove’ it by showing that the diameter varies.
On that - is it diameter variance that I should concentrate on or the fact that the circle is not anywhere near a perfect circle.
I’m new to this...

 

[Pete.]

PI Tape.

I was going to suggest a pi tape but it only gives a true diameter reading if the part is actually round. Since out of round is one problem he has the pi tape might not be accurate. Imagine if one end were truly round but the other were slightly oblong. The pi tape would give similar results but the shape would be obviously wrong.

 

[daleyd]

Well, it’d be easy to measure before it gets welded (ie while it’s flat) so this would be a good place to start. If that size varies then you’ve got no hope.

Size and uniformity would both be of vital importance, if either (or both) is out I would think you would be urinating in a force 9 trying to get it to mate up to another piece. I would have thought a former or jig would be essential (maybe not essential but surely worthwhile?) for this type of work.

 

[Pete.]

Well, it’d be easy to measure before it gets welded (ie while it’s flat) so this would be a good place to start. If that size varies then you’ve got no hope.

Size and uniformity would both be of vital importance, if either (or both) is out I would think you would be urinating in a force 9 trying to get it to mate up to another piece. I would have thought a former or jig would be essential (maybe not essential but surely worthwhile?) for this type of work.

This is what I was thinking. A small hydraulic ram with curved plates on the end to force the rolled tube into roundness. A slightly over-sized rolled ring as a gauge. Really though if you ask someone to roll you a cylinder you should get a cylinder which is fairly round and square on the ends.

 

[Welderpaul]

If the bosses are causing the 'out of round' and therefore the problem in getting the dished ends to meet then can't the bosses be welded in after the dished ends are welded and therefore the tube stabilised?

 

[waddycall]

how big are these vessels?

 

[Curiousgeorge]

If the bosses are causing the 'out of round' and therefore the problem in getting the dished ends to meet then can't the bosses be welded in after the dished ends are welded and therefore the tube stabilised?

Damn it. I meant to say that some of the bosses are tacked in after it’s been rolled. It’s the welding of the interior piping before the dishes ends are added that causes the slight distortion. The rest of the bosses are added after the dished ends are fitted. Sorry for this.

 

[Curiousgeorge]

how big are these vessels?

On average 5 ft in length and 12 to 18 inches in diameter.

 

[gordon stephenson]

I would think that a tape measure will do, Just hold the hooked end stationary and move the tape around until you get the highest reading, then you know that you are across the ctr.

 

[Hitch]

Is the body laser cut, with holes in it by any chance?
Longitudinal seam welded and dressed out prior to putting on the ends?