AISC Section J2.2b - limitations on fillet welds

[Deener]

Hey guys,
I'm checking the strength of a fillet weld used to connect an end plate to HSS in a pin connection (HSS is diagonal cross bracing for a frame). I have uploaded a snapshot of the connection.
To increase weld length and to help with shear lag, we have added a stiffener around the perimeter of the HSS. All plates will be fillet welded on both sides. Summing the total length of weld, we exceed the requirement of staying below 100*weld size. This would mean we have to apply a reduction factor (referred to as ß by the AISC)to the effective weld length. In fact, we even exceed the next requirement of staying below 300*weld size so we are forced to use a weld length of 180*weld size. I don't believe this should be applied in our case because this is not really the end loaded fillet weld that the AISC mentions (i.e. weld direction is not purely parallel to the load). I'm wondering if others have an opinion on this? I'm also looking at reducing the end plate thickness. Seems a little odd that the plate is 1.5" thick and the HSS wall is only 1/2" thick. Perhaps this is driven by bearing of the bolts.
Thanks in advance for your comments.

 

[dhengr]

Deener:
I don’t have the last few AISC manuals, so I can’t comment on their exact formulas, numbers, directions or wording. But, I think you misunderstand their intent or the concept of shear lag. You should dig out some good Strength of Materials and Theory of Elasticity books and do some reading on shear lag, and when and how it applies. They are saying that after some length, removed from the force input, the weld becomes less effective and should be discounted. The weld stresses are highest at the immediate force input and within a few inches (some short length) of weld length. The weld stresses start to decrease as you move away from the load input, and as the load/force is distributed out into the member. In fact, some of the starting welds, near the load input, probably yield to bring the rest of the weld length into play.

Your detail is not a particularly good detail for this connection and load condition and your “stiffener around the perimeter of the HSS” really just complicates the fabrication and adds cost, rather than adding much to improving the stress distribution. Spend that same money slotting the HSS at its end, so you can slide the full width end pl. into the HSS. Then weld the length of that pl., 4 fillets and maybe a few inches inside if you can get in there. But, as usual with many OP’s you didn’t bother to show any loads, sizes, dimensions, etc., all important engineering design details and info., so its tough to know the magnitude of the problem. Finally, the really nasty condition in your detail is the fact that you are trying to move the entire bolt load on that pl. out and around reentrant corners on the pl. and into the tube. This is stress concentrations on top of stress concentrations, a serious problem.

 

[Shu Jiang PEng SE]

AISC J2.2b. The reduction factor is applied in the following situation: when direction of force is along the weld length and the single weld length is more than 100 times of weld thickness. You may not need to consider strength reduction.

 

[Deener]

@dhengr - I'm sorry that you're frustrated with my initial question. I'll try to clear a few things up as I appreciate feedback from knowledgeable members.
1) Shear lag - This occurs in connecting members when their entire cross sectional area is not recruited in the initial connection. Bolting a single leg of an angle is a great example of this. I have uploaded a drawing which is commonly used to convey what shear lag is.
2) I have never seen the change of stress along a longitudinally loaded weld (also applied to the change in load for bolts along a shear connection) referred to as shear lag so that's not what I'm referring to as shear lag.
3) shear lag in the connection which I have shown occurs because the plate is welded to only 2 of the four sides of the HSS. The effectiveness of the stiffener around the perimeter was not part of my original question. Although, I'm confident it recruits the entire cross sectional area of the HSS. It's basically like an end plate. Shear lag is not an issue in end plates.
4) My original question asks for some input on the effective length of the weld that connects the HSS to the bolting plate. It's my understanding that this effective length applied only to longitudinally loaded fillet welds. Accounting for the length of the weld transverse to the load (connecting stiffener around perimeter to the HSS) does not seem correct as that weld would be loaded equally if it were an end plate.
I appreciate your comment about slotting the HSS. I have changed this detail and am confident we can also get some fillet weld on the inside of the HSS and also diminish the stress concentration around the reentrant corners you mention. I have not asked you to evaluate the FOS on this connection so including my actual loads is completely irrelevant to my initial question. Are you wanting to QA the design? I would have to include a lot more information. IMHO it's unprofessional to be asking strangers to QA my designs so only my original question pertaining to the AISC code stands.
@Shu Jiang - thanks for the input. I agree with your suggestion.