Haunch Connection Details

[SPJ1]

Hi,

I have been recently been looking at haunch connection details from the Australian Steel Institute (formerly Australian Institute of Steel Construction) and I note that their typical detail (taken from their publication "Standardized Structural Connections" published 1984) has a 100mm section with "no weld" specified.

I realise this is an older publication but I was just wondering if this is still common practice and what the reasoning is?

Thanks in advance

 

[dhengr]

SPJ1:
It’s because they took a bucket full of welds up there and ran out just before they got to that 100mm length of weld. Alternatively, they couldn’t get into that tight space, btwn. flanges, to make a quality weld, so rather than leaving what looks like a bunch of chicken shit, full of defects, lack of fusion and stress raisers, etc., they want you to stop short of that tight corner/space, where you can still make a good weld termination.

 

[SPJ1]

dhengr:

I'd have to disagree, typically there is 20mm deep at the end of the haunch flange and back in '84 the primary welding method would have been with a stick so you'd easily be able to get an electrode within 25mm of the end, I was thinking there may be a reason to do with the mechanics

 

[CANPRO]

I think dhengr is correct. For a proper weld, the electrode should be at a 45deg angle to the corner. Looking at that detail, there is no way you could get a proper fillet weld for the full length of that web.

If you disagree, what do you suspect is the reason for stopping the weld short?

 

[BridgeSmith]

Looks to me like it's just a matter of it being physically impossible to get into that corner to weld it. It also seems that it's unnecessary. If it needed a weld at the end, laying a bead across the outside of the flanges would be far easier. There may be some concern about stress concentrations due to the proximity of two very stiff sections.

 

[JoshPlumSE]

Yes, in my opinion it's because it's too tough to get a weld in that narrow area.

For what it's worth, I prefer to haunch without a taper. Just take a WT and weld it to the bottom flange for a certain length. Easier to manufacture, easier to construct in the field. Doesn't look as nice though.

 

[SPJ1]

Thanks guys well I'll take the consensus as being that it was done for ease of construction.

Ironically it was brought to my attention by a fabricator who didn't think it was strong enough and said he had welded it anyway haha and wanted the detail revised

 

[Toby43]

Interestingly enough the British SCI Publication 060, they recommend to design that part of the weld for a length equal to the depth of the beam for 1/2 the flange force. The other 1/2 of the flange force should be designed to be taken by the Haunch tip Buttweld.

 

[Ron247]

I was not sure why they do not show a weld on the end of the flange. That could be welded. To me, I would rather cut the unwelded end off and weld the end. That whole crack driving force thing.

 

[BridgeSmith]

If the reason for detailing it that way is not fully understood, I think it's unwise to change it. It could have been simply for constructability, but it's also possible that it was done for another reason (such as avoiding a fatigue-prone connection due to stress concentrations). Fabricators are typically not qualified to make those decisions. The way our specs are written, our fabricators would not be authorized to make the change, either. We would probably reject the change if it was on the shop drawings, or reject the piece if it wasn't fabricated in accordance with the shops.

 

[BridgeSmith]

Ron247, a weld transverse to the direction of stress has a lower fatigue resistance than one parallel to the direction of stress. Depending on the repetitive loading it's subjected to, that could be an issue.

 

[Ron247]

Good point HotRod10.

 

[FLCraneBuilder]

1) - agreed that its impossible to acheive a quality weld
2) - why is the a flange on the bottom vertical plate anyway? Did someone calc that its actually needed? - or did it just "get thrown in" to the detail without thought
3) this is yet another example of why I hate codes - they are too often a "catch-all" that does not fit the true problem at hand - yet people rely on the codes to do the work - even if they dont fit right... but this is a topic of another thread