Forming and using an Internal Omega Joint in tubes

[unclesyd]

First off please excuse the drawing, as you can see I'm not a draftsmen by any means. I have a
tendency to to think using this cad program 1992 version. it works quite well as I have
nearly 300 macros and programs to aid me. Not that all of the excuses are out of the way
here is my problem or quest.

This is an effort to reduce the thermal stress on 194 6" 0.81 wall x 6' long 304L in a
twin cone vacuum dryer that rotates to move the polymer through the tubes. The purpose of
this part of the process is to do a solid state polymerizations by driving off the last vestiges of
H20. tubes. The problem steams for the very rapid introduction of hot oil, 500°F, to shell side
idling at 70°F . Tube in the inlet area of the 100" dia shell have a temperature rise of 450°F in
about 5 seconds by FEA analysis. This effect is on about 20-30 tubes and it the cause of weld
failure. The analysis says that the highest stress 20%-30% over allowable on heat up, but
experience has shown that the tube to tube sheet weld cracks on cool down. This is mostly
an operational problem that production will not address due to production demands. The
failure mode is the weld fails in shear so various attempts have been made to get a better,
stronger welds and if successful the tube cracks and you end up with gobs of weld metal.
I came up with the idea many years ago of an internal expansion joint in each tube in an
attempt to uncouple the tube from the tube sheet The problem was there were no means to
form the tube without extreme work hardening and very high cost. I have now found a way
to form the the bead in the 6" tube with very little work hardening.
The posted sketches show one approach by using a a radius bead similar to an expansion
joint convolution to take some of the thermal expansion. In actuality the form of the bead is
nearly unlimited Crude analysis shows that it will work and I'm waiting on a detail analysis.
As this dryer is charged in the vertical position the expansion bead will be in bottom section
of the tube to take the initial thermal shock. As built the tubes are floating in the tube sheet
an are just seal weld. II can set the tubes in the tube sheet if it is in good enough shape and
weld.
II appreciate all comments good and bad as this is totally outside of the box.

 

[unclesyd]

Hope to get it right soon. here is the second sketch.
Is dwg the right format to upload?

 

[btrueblood]

Syd,

Probably better to post a jpeg for some, but your dwg's opened fine for me. Your omega joints should work to absorb some of the thermal expansion in axial direction, but is there any significant stress due to radial expansion of the tube at the tubesheet weld? Would a bellmouth joint at the tube sheet welds help in that case?

 

[unclesyd]

b]btrueblood[/b],
As posted the tubes are called out as a zero clearance between the tube and tube sheet. On some we cutout the crappy welds there was a perceptible gap between the two. I have looked at a bell mouth flush with the tube sheet face and the possibility of a single round bottom groove in the tube sheet. If the tube sheet is good enough shape I could expand the tube into the groove and get a leak tight seal even without weld. The problem is that some of the ligament have cracked and been repaired son this may maybe out of the question. It is also possible to form a bell mouth and form a i=external bead on the on the inside of the tub sheet//tube interface. This would make tube removal a little more difficult.
Thanks for calling out the proper nomenclature for the bead. As posted preliminary calculations show that it will work. The fact that the forming process unlike rolling imparts very little cold work to the part so just about any profile is possible.. In the following sketch are some possible variations though some would probably put too high of a tress on the tube sheet

 

[btrueblood]

Syd,

Hope you bring your expansion/forming tool/device to commercial fruition - I can think of a lot of applications where it could be used, esp. if cold work can be minimized as you say. Good luck!

 

[unclesyd]

btrueblood,
This project is like the toaster, i had to wait 20 years from the original idea to find equipment that could do the job..
It exiit now, though not in the configuration need, tooling. In reality I saw what the existing machine was doing was doing and asked the manufacturer could it be modified for our application and the answer was yes, no problem.
I too am looking at other possible used to usethe tube setting capabilities to set some tubes in a Therminol Vaporizer, now weeping . The tubes will be mechanically set snas roller. Looking at replacing tubes in air reheaters and air heaters, hopefully without welding.
Type in the following address;
"A-M-H.com" look at machines and go to #8516 and look at the bottom videos. By changing the die it will form the bead in 3-5 seconds. As I recall one can go to 10" and also set 10" tubes.
I will ask Dave to fix my title asothers maybe interested. `

 

[rmw]

Syd,

First off, on kind of a personal basis, it is nice to see you posting so much after a long dry spell. I've checked your "handle" from time to time and knew you were visiting the site regularly, but your post count for 2012 has really fallen off compared to your prior activity. You were missed - at least by me. I value what I have learned from you over the years. I've even made some comments in some posts where I proposed something (tube rolling/welding order) where I knew you favored the opposite approach baiting you to come on and present your side, but alas, I couldn't get you to bite. And it may have just been that you were posting somewhere I never lurk.......

Now to your question:

You use the term "seal weld" which to me when I have my Hx hat on is a different and much weaker weld than a full strength weld. I don't have my ASME manual here at home, but I am sure you have access to one to look at the weld bead difference between the two. Had you already thought of that?

While I am not a strong advocate of grooving in the thin wall and hard metal tubing I typically deal with, I know it has its place with the right materials. In this case, however, with the wall thickness you have, I think you would get minimal help from a groove, BUT that said, you are looking for all the help you can get, so any would help and even "minimal" in this case would go a long way toward taking stress off of the weld.

You said no idea was a bad one, so how about replacing the TS with one with more thickness so you could get a stronger mechanical joint (in addition to the full strength weld) and room for more grooves? If this cull has had all the ligament damage you cite, it might be a candidate for changeout. I think the presence of ligament damage itself is a message - a cry out for help.

The way I interpret your drawing (had to go to my work computer to open it - Voloview on my personal didn't like it - the swage you want to add would be from outside the tube to in, or do I have it backwards. I've seen the type of machinery the A-M-H outfit produces used in expansion and flexible metal hose manufacturing and in all cases the pressure was inside pushing the convolutions to a larger diameter. Clear me up on what is going on here. Which ever way it is, if you can get the machinery to produce it, and the convoluted zone doesn't raise issues with product hide out, etc., go for it - in addition to the best tube to TS joint you can get at the TS.

And, BTW, you will notice that I didn't even touch on recommending rolling then welding or welding then rolling because I knew you would go your own way on that one, so there - at least you know I didn't forget it.

Nice to see you on the boards again.

rmw

 

[unclesyd]

rmw,
Yes the bead, or Omega joint is formed from the outside by changing the die in one of their machines., the 8516. As this tube revolves iat about 1 rpm and the product is 1/8 ‘ square Nylon flake I presently can see no problem with the product hang up and their would be no loss of surface area. According to A-M-H they can form the bead, their term in 3-5 seconds near what would be the hot end of the tube. Additionally they have the tooling to allow me to set the ^‘ tube , either straight wall or flair the end. This dryer is a trunion mounted two cone all weld to gaiter. To change the tube one has to cut one end of the cone off ot remove and install new tubes.. I have looked several was to take the strain off the tubes with a flued or expansion joint in the shell and the possibility of flanging one on with alight flanges and hold taking the polymer weight loads that pass through the tube. To other head. Thee addition weight has to transfer through the trunion . The trunions are also giving trouble. One other thing that due to the design, and operating conditions the tubes have to act individually. As I posted the analysis says the highest stress is on heat up and in reality the tubes welds crack on cool down..
As you noted, correctly I’m a notoriously wrecker heat exchanger design, the end results are that mine work even though my approach is some what radical. at times. At the present I’m embroiled in a discussion about a problem of two heat exchangers that immediately fall off in performance after in stallion due to pulling of the shell side. Once removed they are impossible to clean because of the triangular pitch. I have proposed a bundle on square pitch, loss of 20 % surface area. The response to this hurt my feelings. The only problem I that I’ve have done this 30 years ago on 6 exchangers and increased the on stream lime from 3 weeks to over three years.
As I’ve stated before we used grooves roll the tubes and make a seal, autogenously if possible. Another thing we do in house is that we polish the external surface of the tube prior to insertion in the sheet

I'Appreciate everyone's comments post the sketch of how the bead will be formed.

 

[unclesyd]

-rmw,
Here is a rough sketetch of the proposed die to make the internal bead.
I will have to agree with you that we slightly overkill on some exchanger Our guidlines come from when we had 3 processes that eat exchangers, Nitric+Organics, Adiponitril Synthesis, and the afore mention EA exchangers. At one time we had approximately 70 exhangers installed or in variuos states of repair. Most of them were the external floating head variety. During the peroid that we removed Cr from the CTW we were losing an exchanger a week due to MIC. If you ever have this problem where MIC is after the skirt or bottom sheet just use a ring of lead packing, don't tell anyone. This will let you reuse most of the skirts tube sheets.
You might note that I've been burned, bad on a 2 Hastalloy 276 exhangers, $500,000 litterly down the tubes. The process people raised the Formic acid and Peroxide levle to where the process went to an oxidizo=ing state, the process ate the tubes up in the top tube shett 1 3/4" of the 2"in about 2 weeks.

 

[btrueblood]

Ah, that makes even more sense. Hard to pull a tube with an external bead out thru the tubesheet.

 

[MikeHalloran]

I don't think you'll get a bead formed as nicely as you drew it, absent some internal support for the part of the tube that you don't want to shrink. E.g. a segmented removable plug on both sides of the bead. ... and a _lot_ of high pressure lubricant like Never-Seez, because you'll be pulling material axially into the bead.

As shown, without internal support, the portion between the end and the bead will just wrinkle. The portion toward the bulk of the tube _may_ pull down into a vaguely V-section groove half, probably with a few wrinkles too.



Mike Halloran
Pembroke Pines, FL, USA

 

[unclesyd]

The people at A-M-H are quite confident that the bead will be smooth. The have machinery and know how to make expansion joints and understand what I'm after. The die is segment and all segment are move the same rate. If the made an external bead the bead is very smooth. As posted most any configuration will probably work, due to the small amount of expansion that needs o be taken up.
AS suggested above replacing the tube sheet with a thicker one would help in several ways, go to a thicker wall safend and very easily set the 6" tube in the sheet and then seal weld. Previously setting 6" tube was out of the question. one big manufacturers laughed at us many years ago when asked the question about setting 6" tube or even a 4" one.

I have made an internal bead on a lath using a 3 point roller, steady rest. But the work hardening was a problem as any beads made the way would have to be annealed to get the need spring rate.
I maybe in left field but this is the only new approach to redesign the dryer to accommodate our process in the 25 years or so of operating off the design condition.