Point Loads applied to steel roof diaphragm

[hardyworld]

I am retrofitting an old structure for a new purpose. Part of the proposed changes would replace an existing non-load-bearing masonry wall with a metal panel wall. This would induce loads too high on the existing columns for axial-bending interaction. Therefore, I proposed that additional 'columns' be placed at third points of the span to support the wall from wind forces applied to the metal panel wall.

This solution solves the axial-bending issue with the existing columns, but I don't know how to design the new 'columns' (which act as beams) to be attached to the existing roof diaphragm at the top. There is information on diaphragm design based on spread loads from the Steel Deck Institute, but I cannot find any help on how to design connections of point loads to a roof diaphragm. My thought was that I'd connect it to the nearest joist and then cross-brace it with parallel joists in the direction of the applied load until it spread the load onto enough of the joists/roof diaphragm to support the wind loads.

Does anyone have links on how this should be designed/detailed? Or are you aware of any documents from the Steel Deck Institute or other notable standards that lay out how this should be done? I feel like I'm missing some publication that is probably obvious.

 

[BAretired]

Does the new wall run parallel to the joists? If so you could install a pair of members from the top of each new column angling over to the first joist and let the joist transfer the load into the deck.

A sketch would help.

BA

 

[hardyworld]

Yes the new wall runs parallel with the joists. What you describe is what I want to do. But I have concerns that the lateral point load applied to the joist will be too high for the existing joist-roof connection. If I spread that load to multiple joists running parallel with the wall (by adding new members) I think it will adequately spread the load to the roof diaphragm. I would think that a design on how to spread the load into the diaphragm would have to be published somewhere.

 

[BAretired]

The Canadian code contains shear resistance values for spot welds. Screw manufacturers have ultimate shear resistance values for various gauge sheet metal.

I think it is a matter of making a reasonable assumption as to how many fasteners are engaged by the joist. If in doubt, err on the safe side.

You could add screw fasteners between the horizontal leg of an angle connected to the column into the underside of the deck flutes between the edge of roof and first joist in order to gain more capacity.

Or you could engage the second joist with additional braces.

You may find some useful information in literature from the Steel Deck Institute such as:

http://www.sdi.org/DDMO3/appendixVI.pdf

BA

 

[BAretired]

Another suggestion is to use a distribution plate, say 12" x 12" x 1/8" mounted with self drilling screws to the underside of the steel deck and welded or bolted to the angle strut. It could be welded to the first joist if desired.

How large a horizontal load do you have? What is the edge condition, i.e. will you have a beam over the new columns?

BA

 

[wannabeSE]

Instead of a adding columns, can steel studs be used to create a wall that spans vertically rather than horizontally (or have the metal wall panels span vertically rather than horizontally)? Then, the devil is in the detail of attaching the top of wall to the diaphram and allow the roof to deflect. How is the existing masonary wall connected to the diaphragm? Perhaps, it is possible to modify the existing slip connection to be used with the new wall.

Unless the wind forces are small, I would avoid transfering the lateral forces into the joist. This may create a somewhat tortured load path with weak axis bending in the joist. A combination of bridging and bracing with the top bridging attached to the diaphragn might be a better detail. The bridging would need a connection detail that provides continuity past the joists. Are the existing columns lateral columns? If not, then their connection to the diaphram should also be evaluated.

 

[BAretired]

Maybe a cross section through the wall would help.

BA