Welded Gusset Connection 1

[Scoper]

I'm using a slotted tube to gusset plate plate connection and the fabricator made the slots a bit wide - some of the slots are 3/4", which leaves 3/16" gaps on each side of the 3/8" gusset plate. The contractor is filling the gap with weld material, then adding the required filet over that, but, as I interpret AWS, he must oversize the filet by the size of the gap - i.e., a 5/16" weld must be upped to 1/2" (8/16") to achieve the proper effective size. It is my interpretation of AWS that this is too large a weld for 3/8" material, particularly when the filet goes on both sides of the gusset - only guidance I can find is a formula in AISC which states that the minimum plate thickness should be 5.16xD/Fy, but no specific code reference. Contractor's welder does not believe this is a problem. Any comments/suggestions would be appreciated.

 

[JAE]

Why can't you offset the tube to be tight to one side, using the original fillet on that side. Then, on the other side, insert a 3/8" bar, or 5/16" bar, to infill the slot. The bar could be extended about 1/4" to 3/8" beyond the face of the tube such that you can now install twin fillet welds, one from the tube to the bar, and one from the bar to the gussett.

Your tube would have a slight eccentricity so this would have to be reviewed, but it should work.

 

[Scoper]

Thank you - this sounds like a good solution for the joints where the connection has not yet been welded. I'm still wrestling with the joints where the contractor has already filled the area on each side of the gusset with weld material (majority of locations), as well as my interpretation of how AWS requirements apply to this condition. I may have to have them remove the welds that have already been made and have them start over, but don't want to if it's not necessary. Would appreciate any opinion on this. Thanks again!

 

[HgTX]

I like JAE's solution. Just make sure that you really do have one fillet weld from the tube to the bar, and another one from the bar to the gusset. They can't just stuff some filler in there and then do the 5/16" welds as if there were no gap.

Another possibility, which is kind of what they're doing when they fill the gap with weld metal, is to make it a CJP or PJP groove weld. If what they've done already fits one of the prequalified joint geometries in D1.1, you're in luck.

Hg

 

[VeryPicky]

I like siple, quick solutions like that one JAE offered.
Do yourself a favour though, clear it with the client and designer first. The insert bar might introduce new variables that were not considered by the design. I don't know if whatever you are building is a regulated item. If it is the inspection body might need to be involved too.
Good luck.

 

[metengr]

If I were the client's engineer, I am not sure I would accept this modified repair approach. I totally concur with the suggestion provided by JacekM. In my opinion, this is no different then "slugging" a weld, and could result in an inadequate fillet size or lack of fusion or IP defects.

Unfortunately, the only solution would be to weld build-up the oversized gap along one slotted edge and trim as required to achieve the proper weld joint clearance. The weld build-up would provide material to adequately tie into, and result in only one fillet weld of the correct size. Since this is a scew-up, somone has to make this thing right regardless of cost to repair.

 

[pmkPE]

My understanding of the minimum and maximum weld sizes is based on the need to avoid cracking in the weld at the throat (here it can not be seen)due to inadequate heat treatment. A thin material such as 3/16" plate welded to a very thick plate, say 1 1/2" will need heat treatment to insure that the dense thicker plate will not act as a heat sink and promote rapid cooling of the weld. Also in many cases the base metal thickness of the thinner member governs and the weld stresses can not exceed the allowable shear strength of the base material - this is to protect against lamelar tearing.

I don't believe from your description that the weld is oversized in the context that AWS is referring to as the extra weld is not on the tube being connected, it could be on the gusset however. All in all the fabricator is responsible for adequate pre-heating as well as inspection. If you have concerns about their procedures you can specify NDT testing of the weldment as well as the base metal to check that everything is acceptable.

If you have an opportunity please review the literature published by the Lincoln (Arc Welding) Electric Company - they have excellent materials that describe the various welding processes and the prices of their publications only include publishing costs, i.e. they are not looking to make a profit. They want to disseminate info on the current state of welding. Two of the books that I have bought from them are entitled "Metals and How to Weld Them" and "The Procedure Handbook of Arc Welding"; another Classic that I have of theirs is by Omer Blodgett "Design of Welded Structures". I know many structural and mechanical engineers that have the latter in their collections and they are all well worn with use.

 

[HgTX]

JacekM is right that this needs to be cleared with the designer (if that's someone other than Scoper's own self). But the solution JAE describes isn't exactly slugging, if there are separate 5/16" welds from pipe to strip and then strip to gusset. You'd need a strip at least 1" wide and probably a little more just to be sure.

One problem, though, is if you concentrate all the gap on one side, using a 3/8" strip, you introduce eccentricity into the connection. On the other hand, if you split the difference, all you have are 3/16" strips which are smaller than the originally specified 5/16" weld.

On the other hand, if the 5/16" weld was called for because that's the minimum size required by AWS because of heat issues (D1.1 Table 5.8), and not required for strength, then I'd go with the 3/16" strips if a 3/16" fillet will suffice for strength. You need to look into the fatigue category at the end of the strip and decide whether to just do longitudinal welds or wrap around the end; what did the original detail call for?

For the welds already made, take a look at Figures 3.3 and 3.4 in D1.1. Maybe one of those configurations applies.

If you don't want to go the strip route another possibility is to bevel the slot and make a real PJP or CJP instead of the fillets.

I'm not aware of any AWS restrictions on maximum fillet weld size. If you're worried about melt-through, use a low-heat-input procedure with multiple small passes to avoid a big molten blob all the way through the gusset thickness.

Hg

 

[Scoper]

Just wanted to say that I very much appreciate the input on this. I did speak with the fabricators welding engineer, who advised me that this situation was not that unusual. My main concern was the size of the filet on the 3/8" gusset plate, since, as indicated by pmkPE, it is welded both sides. Although I like the solutions suggested here since they limit the weld size, at this point I have not been able to find anything in either AWS or AISC that explicitly limits the weld size for this geometry. So, since this is his preferred solution, at this point I am having the contractor proceed with his fix in accordance with AWS D1.1, Section 5.22, 'Tolerance of Joint Dimensions', increasing the filet size as necessary to properly account for the gap width. I am also requiring random magnetic particle testing of some of the larger welds. Once again, I appreciate all of the input so far, and if anybody has further thoughts on this, am very interested. Thank you!

 

[sdz]

The requirement to increase weld size by the size of the root gap is to ensure the effective throat thickness is maintained. If the contractor welds the slot first and then applies a fillet over the top then this would probably satisfy that requirement. You would need to check the effective throat thickness at the most critical section.

Of course making him increase the weld size will encourage him to be more careful in the future.

 

[HgTX]

Welding up the slot any old way and then applying a fillet weld of the original size would be slugging as metengr pointed out on 9/22. To make this a valid D1.1 prequalified weld, they have two choices:

1. Get enough weld into the slot that they will be able to do a fillet, and then do a fillet with increased size. The weld deposited inside the groove wouldn't be counted for strenght purposes, just something to allow the fillet weld to have the proper profile.

2. Declare the whole thing a groove weld rather than a fillet weld and prep the joint accordingly.

Hg