How Does This Joist Seat Work?

[KootK]

I'm in the process of reinforcing some open webbed steel joists. They are a style of joist where the chords are made from cold formed hat sections and the first tension diagonal is a flat bar. They are extremely common in my neck of the woods. Please refer to the attached sketch.

In attempting to reinforce the joist seats, I've come to realize that I don't actually understand how the original joist seats work. As you can imagine, that's a problem. In the attached sketch, I've proposed three models for how force transfer might work at the joist seats. My gut tells me that model #1 is the way to go but I'm really not sure.

Visually, when you look at one of these joist seats, it's very tempting to assume that the joist "hangs" from it's support by the flat strap diagonal which is extended up onto the joist support. This is analytically expedient as well as it means that the vertical shear in the joist never "lives" up in the joist top chord. My suspicion is that most engineers feel that this model is correct because we almost never see any seat reinforcement in joists manufactured this way.

Thought?

Kootenay Kid

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[BAretired]

Your sketch is very similar to a detail used by GWS (Great West Steel) with a couple of exceptions:
1. Depth of hat section was variable, but maximum depth was 67.9mm (2.67").
2. GWS always used an inverted hat section to form the joist shoe, rather than having the joist bearing directly on the end diagonal as you are showing. Their end diagonal was usually almost square in shape and its dimensions were selected to fit neatly into the inverted hat section to ensure a good weld.

The force in the end diagonal intersects with the center of gravity of the top chord a little inside the edge of the beam flange. The exact position depends on the slope of the end diagonal. If the top chord is adequate to resist the theoretical compression from the applied loads, and the weld between end diagonal and top chord (including gusset) are adequate to carry the tension, I don't see a problem with the seat. The precise stress distribution would be difficult to determine by hand calculation but perhaps could be estimated using finite element techniques.





BA

 

[KootK]

Thanks for the response BA. I was on site last week and snapped a photo of the joist seat in questions (attached). I'm sure that they are quite similar to what you've encountered. To try to get to the heart of my concern, perhaps you'll indulge me by giving me your thoughts on the following two questions:

1) Do you think that the joist seat assembly would work without the triangular gusset plate? I do not.

2) Do you think that the joist top chord near the seat needs to be able to handle a shear force approximately equal to the joist reaction? I do. And it's causing me considerable grief.

Thanks for your help.

KootK

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[dhengr]

KootK:
I think they look at their built-up seat as consisting of the top chord, the bottom bent bearing pl. and the small gusset pl. Then, the tension in the, rectangular bar, end diag. is transferred to the underside of the bent brg. pl. with two fillet welds 4 to 6" long, and stopping just short of the beam flg. and brg. area. I think that the gusset is an essential part of this seat arrangement, otherwise the tension in the end diag. would be transferred to the bot. of the top chord over a very short length of welds at the start of their attachment. The only thing that would work to prevent these two welds from unzipping from the top chord would be the bearing reaction right in the same area. I’ve seen what I think BA is describing and it consisted of an identical hat section, 5 or 6" long, and rolled 180̊ (a mirror image, end view, of the top chord), as the lower half of the seat. The end diag. is then brought up and bent to fit into the lower hat section, and welded, both sides. I always worried that this end diag. might rip out of the lower hat section, as that weld started to unzip. But, the jst. manuf’er. has tested these and they do seem to work. I just don’t like the way those loads are applied to the start of the two welds.

I think your det. A (mod. #1) is about right, and the gusset protects against my concern about the diag. weld unzipping as it attaches to the top chord. I think that a good share of the tension in the end diag. is taken right up onto the bearing as an axial load in that bottom bent pl., plus some shear in that pl. Of course, the gusset causes some of this transfer into the top chord too. As long as the entire seat arrangement remains compatible, the total transfer happens. I think your bearing reaction should be adjusted a bit, it is a trapezoid, not really a triangle. As the jst. deflects, the tip of the beam flg. will bend a little, and transfer more of the reaction back toward the beam web, to the extent that the stiffness of the whole seat detail allows/causes this. I also suspect that the horiz. shears and the vert. forces btwn. the gusset and the top chord are trapezoidal; starting at some value on the right edge of the gusset and increasing in value as you move to the left, toward the bearing. Try this thought process: The reaction would get to the beam with just the end diag. (the top chord removed), but there would be a significant torsional pull on the beam from this; the top chord provides a reaction (compression) against this lateral load (torsional pull), through the horiz. shear flow in the welds to the top chord. Also, the top chord does have some bending strength from the first regular diag. member out to the bearing area, although this is a small part of the total load transfer to the bearing. Cut the truss vertically (a vert. section), starting at the end of the bot chord, through the first regular diag. and up to your cut line on the top chord. What does this look like as a free body?

 

[BAretired]

1) The triangular gusset plate is essential to transfer some of the horizontal component of the end diagonal into the hat section and also to prevent unzipping the weld as dhengr suggested.

2) I think there will be a shear force in the webs of the hat section but I suspect it will not be more than about fifty percent of the joist reaction provided that the top chord and end diagonal intersect over the center of bearing. If they intersect outside the bearing area, in this case the beam flange, then the hat section must handle the full joist reaction. That does not appear to be the case in the photo you provided.


BA