Supporting short section of undersized tubing

[SPND]

My apologies for the somewhat vague description but I was having a rough time trying to give a concise description of my delima.

I've got a situation wherein I need to join a 316L tube to a machined 304L fitting. The fitting has a lower yield strength than the tube (the tube is slightly work harneded whereas the fitting bar stock is fully annealed) so it has a thicker wall section than the tube. The joint is a full penetration autogenous weld. I've got to reduce the fitting end dimensions to something which closely matches the tube dimensions so that the parts can be successfully welded, the wall section of the fitting will draw too much heat away if not modified. The only issue though is that the fitting cannot stand up to the external pressure if its geometry is close to that of the tubing. Is it possible that the reduced section of the fitting could be supported by the adjcent stronger tube and thicker fitting cross section for a short length 'x' and if so could someone suggest a possible method of analyzing (no FEA, manual method please)? To give you guys an idea of scale, the 316L tube is in the neighborhood of .285" ID, .035" wall and the fitting in the thicker section has a similar ID but a wall of .045". I need the reduced section of the fitting to be in the range of .05 to .08" long. I've tried looking through Roark's for an example approaching this scenario but have not found anything. Any suggestions? Thanks.

 

[SPND]

A bit of feedback would be really appreciated.

 

[gr2vessels]

SPND,
Sorry mate, but I think that there is a major discrepany between your hair splitting issue, using a kitchen cleaver and the code requirements. I am having trouble believing that this is a real fabrication issue, particularly with using autogenous welding procedure in your nuclear application. Any laser or even TIG welding procedure will end up in a disproportionatelly big weldment on the joint area of you capilary tubes, exceeding any code requirements for the given joint strength. I don't believe that you'll be able to control the autogenous welding of the components to the finesse you imply in your post. Perhaps that was the reason people wouldn't answer it.
However, in real life you don't need to do anything for that weld, because the simple square butted fitting and the tube end can be full pen welded and the covering run will provide the transition angle of additional material, in top of the pipe wall thickness, until reaching the fitting wall thickness. So you don't have to machine the 0.01" thickness difference, to taper the 0.045" down to 0.035", thus loosing significant strength of the joint...The only problem you'll be facing is the machining of the inside diameter, because any full penetration welding tends to shrink the ID, not to mention the weldment protruding inside the joint. I can se at least a 4.283% reduction of the assembly ID.
Cheers,
gr2vessels

 

[LSThill]

Please provide a photo or drawing

 

[SPND]

Thank you gr2 and LSThill

FYI, we have been doing these welds for the past 20 years with GTAW, and yes, we get a bit of ID sag. The welds typically have a HAZ about .050 to .075 wide and we have no problems getting a full pen weld that's extremely clean. The probelm I am facing is that we have been brow-beaten by our customer (one of the larger US reactor designers) into one of the thinnest walls for this given instrument size and while the component parts on either side are sufficiently strong the weld geometry my supervisor is proposing is not. GR2, I would like to know more about what you have to say regarding not reducing the fitting for the weld. I do not have much welding knowledge but I was questionning (in my head) my supervisor's perceived need to reduce the fitting dimesnions so much. And just so you guys know, I am not trying to "make this configuration pass", I just wanted to know if it was possible that a short section could in any way be acceptably supported by adjacent features. I will post some pics shortly, thank you very much for responding.

 

[SPND]

Here is my attempt at posting a pic.

The pic show a welded and unwelded fitting. Note that the welded sample has the tube hacked off very close to the weld, the actual part would have more than 40' of tube extending off the end, kinda hard to work with on my desk. The wall thickness of the fitting as-machined matches up perfectly to the 316L tube. When you consider the size of the parts and the actual (.080") length of the machined portion it's hard to imagine there'd be a problem but I have to prove it on paper. At the moment the minimum dimensions of the .080" long section fail by only several hundred PSI, the fitting and tube on either side is OK.

 

[gr2vessels]

Sorry, a bit late;- to me seems a simple logic. If you cannot machine down the fitting, then increase the pipe. I meant to say that you build up/deposit first some 0.01" thickness on the 0.08" length end of the tube 316L, to equal the 0.045" thickness of the fitting, then do the autogenous full pen weld. I got no idea how would you machine the weldment to taper it down to 0.035" thikness...
Cheers,
gr2vessels

 

[SPND]

Unfortunately the pipe dims are set in stone, the pipe is 45+ feet long and there's no latitude to change at this point (couple thousand feet of tubing in-stock). Looks like we're just going to make some weld samples and thake them up to test pressure (4000psi) to prove it works. Thanks for your input.

 

[unclesyd]

You mention 4000 psi, is this external or internal?.

I think you can work a taper by just cutting it a little short from the end. Remember that an autogenous weld can handle a little mismatch on the O.D.

Does your design criteria permit a socket weld or weld ring?

 

[SPND]

A socket has been considered in the past on similar applications. One problem we've run into is that the lip of the socket does not always get pulled into the melt and once the you miss it during the weld process it's really hard to get it on a rehit. One thing to keep in mind here is the smallness of the parts. If the tube is out by just a couple thou from your machined part it makes it that much harder to wick the lip of the socket into the weld, the geometries of these small parts is a real killer for reproducability. An amp or two can make all the difference in the world. A ring is not possible as the ID and OD are at their min/max already, no more material can be added.

The pressure is external, I'm rounding up from actual. We test at 3850, 1.5x's operating pressure (2500) plus 100 for a margin of safety. We are going to use 4000 just to round it out and give ourselves that much more "warm and fuzziness".

We are also thinking of doing just as has been suggested, to weld the fitting and the tube together without any adjustment to the fitting and see what happens. In the past (so I've been told) the thicker section will want to pull the arc over and getting full pen on the thinner side can be a challenge. I will post some pics of the results, may not happen until the new year though.

Thanks for the feedback.

 

[unclesyd]

If welding as is isn't practical for the reasons mentioned try cutting down just 0.005" instead of your stated 0.0010". At your diameters this is a lot of metal. i would still try a taper.
Also remember that with a socket type weld you don't have to break the outside edges of the socket down.

Have you looked at a mini orbital welding machine? It could make your welding process a lot simpler. In the words of a famous TV pitchman "Set it and Forget It".

You might want to look into renting a machine like the Arc Machine mini or one of the many other available.

http://www.arcmachines.com/prodMed/m9250/m9250.html

 

[SPND]

We have mini orbitals but we also have some in-house units we call "cubes", essentially a small lathe with a horizontally mounted torch and custom fixturing to hold the 40 foot plus tube. I have been told the cube set-up is more desirable as it is easier to get good concentricity between the machined fitting and the tube whereas the orbital requires specifically sized adaptor collets (which sometimes become "consumables" due to the proximity with the weld) and the orbital heads need to hang from the part which makes fixturing more onerous than our cube set-up.

It isn't that the welding process is the problem, in fact we've got that process down to an art (as the pic in my 4th post shows). This all stemmed from my query into finding a calc method to show that the short .080" section of undersized tubing would be a never-no-mind with regards to pressure bearing capability when I've got a section on either side more than capable to the task. I know in the real world there'll not be an issue and my 4000psi pressure test will confirm this, I was hoping to find a way of proving it on paper and thus negate the need to do a test.The conversation around changing the fitting end geometry was to negate the need for a pressure test and calculations - if the weldworks with an unmolested fitting then everything is hunky-dory, there would be no undersized wall at the weld and I think this is what I am going to try.

 

[unclesyd]

You might be able to glean a little information from this table. If I remember correctly that the collapsing pressure is very pressure application rate sensitive for SS.

There are also other formulas that use the material modulus or reduced modulus that is supposed to work better for small tubes. I seen these used in some subsea calculations.

http://www.finetubes.co.uk/pdfstore/Datasheet e118 Safe Tube Pressures 05 lores.pdf

 

[EdStainless]

Upset the end of the tube in order to make it thicker. Then do a straight butt weld. Any reduction in the wall of the fitting should be strictly forbidden.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube