welding from one side 1

[HgTX]

Need to do some fracture-critical welding with access only from one side of the weld once both plates are in place. Tensile stress across the weld, so can't use stay-in-place steel backing (which would be welded to one side of the joint before the plate is put into place, if this weren't a tension connection).

Thought about ceramic backing (and qualification per AWS D1.5 section 5.7.7), but is there a version of ceramic backing that doesn't require access to both sides of the joint in place? The only stuff I've seen involves a strip of tape that runs across the ceramic and tapes to both sides.

If not, AWS D1.5 section 5.7.7 allows qualification of PJP as CJP even without any backing at all, I think. There's a bend test & macroetch to demonstrate adequate fusion & penetration at the root, but wouldn't melt-through still be a fatigue concern? (Or is the no-backing case blocked by other provisions of D1.5 such as 4.6.9 and 4.7.6?)

Hg



Eng-Tips policies: faq731-376

 

[metengr]

HgTX;
I have a little trouble in following the terminology you reference in your post. Can you qualify the weld joint as CJP using a consumable insert with a single-vee weld prep?

 

[Goahead]

With an adequate Welding Procedure Specification melt through is not acceptable even with complete penetration. A consumable insert could be possibly considered for the root pass. Backing should not be necessary. Thickness of plates and joint configuration (bevel angle, root face, root opening) before welding were not specified.

http://www.welding-advisers.com/

 

[gtaw]

A purge on the root side is usually required with a consumable insert. You would also have to qualify the WPS with GTAW, at least for the root pass and possibly the second pass. You could then switch to a different process that is already prequalified since any additional welding would be considered to be with backing.

Al

 

[weldtek]

Depending on the joint and assembly configuration, you may be able to install a non fusing backup, if you can access the backside sufficiently to remove it.
PJP vs. CJP sounds like a design issue.

 

[HgTX]

Most of these are going to be T joints though some are butt joints. Plate thickness up to 2".

A simplified description of the structure would be built-up box (not big enough to put someone inside even if we could get them out after welding) with diaphragms. And welded field connections.

Once the joint is assembled, can't get to the back--not to put in a root pass, not to tape on ceramic, not to backgouge, not to remove backing. I didn't immediately see another way to design the structure to avoid this (I'm not the designer but I have some input), but I'll look again when I get my hands on the plans again tomorrow.

I am completely ignorant about consumable inserts. Rather than bug y'all with all kinds of stupid questions, where should I look to get some basic information so I know whether to recommend this option? (The fabricator is still responsible for knowing what they're doing; the fate of the welding is not in my hands.) Has anyone used a consumable insert on a D1.5 job?

Hg

Eng-Tips policies: faq731-376

 

[weldtek]

A-5.30 covers consumable inserts so that may be of some use.

 

[metengr]

thread178-25616

From your description you could flood the cavity with Argon as a back purge when welding the consumable insert.

 

[unclesyd]

Does anyone know how the Navy is doing there one sided welding?

There was a "Megastructures" Problem mention that the Navy has a new procedure for welding thick section from one side that claimed to do a better job that the older welding procedure. The programs description of the welding procedure left a lot to desired.

 

[HgTX]

Looks like consumable inserts are pretty much a pipe thing...

Best suggestion I've received so far is to leave the designer on the back side of the weld with a flashlight and a file. Except that the space is too small.

We're probably looking at a redesign in some areas, and where they can't do that, go back to the fused steel backing and design for a Category D or E detail...there's just simply no way in a bridge application to do a one-sided weld without some kind of notch on the back side, whether it's potential lack of full penetration, effects of melthrough, or the actual notch created by the steel backing.

Or get some Aethyopean bolt tighteners and teach them welding...

http://tinyurl.com/5p87r

(Paul Pendragon, Process Engineering magazine, 1973)

Hg


Eng-Tips policies: faq731-376

 

[pipewelder1999]

Here is an example of a consumable insert.

http://www.weldinginspectionsvcs.com/misc/consuminsert.html

Gerald Austin
Iuka, Mississippi

http://www.weldinginspectionsvcs.com

 

[unclesyd]

A lot about pipe but some about plates.

http://search.google.dot.gov/DOT/DO...=&refresh=false&proxystylesheet=DOT_Pages&oe=

 

[HgTX]

It's not so much the thickness but the one-sidedness that's the problem.

Hg


Eng-Tips policies: faq731-376

 

[metengr]

HgTX;
I may be looking at this too simplistically but based on your original post why can't you simply have a grinder secured to one end of a pole that can reach around the back side of the inaccessible weld joint and locally back-grind the full penetration weld to reduce stress concentration?

 

[pipewelder1999]

Open Root GTAW and SMAW can be performed provided welders and procedures are qualified. Open root GMAW could be performed but as a welder, I don't prefer it unless fitup is perfect.Is that NOT an option ?

It is more difficult on Tee joints than butt joints but still possible.

Gerald Austin
Iuka, Mississippi

http://www.weldinginspectionsvcs.com

 

[HgTX]

metengr--There's no access behind the weld once everything is in place except for possibly some small corner clips. (One example would be a diaphragm inside a box section; another would be a web plate closure for splicing two box sections.)

I think they can put access holes in some locations, at least enough to get an arm in to apply ceramic backing, and possibly enough to just go in there and grind. For example, on the box splice, they could weld the fascia side with access from inside the box and then do something bolted on the non-fascia side. Not that I've heard anything yet about them actually planning to add any access holes, but it's at least a possibility.

But I think there are still some welds we just won't be able to get access to.

pipewelder--the question is whether open root welding can be performed that doesn't effectively create a notch or similar effect on the back side of the weld. We're working within AWS D1.5 though I see some noted exceptions headed our way fast...

Hg

Eng-Tips policies: faq731-376

 

[weldtek]

HgTX,
Have you considered making up some mock up assemblies and trying out your options? It may be the best way to determine which track is the best to take.

 

[HgTX]

You mean like take a bunch of pieces of cardboard and try to find an assembly sequence that won't block access to the back side of any of the connections? No, I don't think the designers did that.

Hg

Eng-Tips policies: faq731-376

 

[weldtek]

Well that might help, but, actually I meant fitting up some small representative assy's, with different joint details you might consider as options, and actually welding them up, with at least the root pass. Then you can examine the backside to see the results. For example you might do one with a specific bevel, land and gap, and try another one with a a particular joint geometry and with a consumable insert, etc. Just a suggestion.

 

[HgTX]

The contractor (not in the picture yet) is going to have to demonstrate a lot of this stuff with mockups no matter what; the bridge code requires non-standard joints to be qualified.

Asking around among bridge people, though, no one seems to believe that there's any kind of one-sided welding that wouldn't be considered to be a Category D or E fatigue detail.

And the design consultant doesn't have a welding lab...

Do they make non-pipe consumable inserts? How does the backside of a typical such weld compare to the backside of one made against ceramic backing?

Hg

Eng-Tips policies: faq731-376