Any idea how this structural system works?

[Ron247]

I have a old 84’x120’x17’ tall masonry wall building with a metal roof deck on bar joists. The roof covering is tar and gravel. The building is probably at least 40 years old or more.

My Client wants to remove the exterior 12” hollow block wall that runs parallel to the bar joists. I went out and looked to see how the building was originally constructed. I am sure the interior frames are not rigid frames and not fixed base columns. I am a little puzzled by what I have seen so far and was wondering if anyone had any thoughts about how this system was supposed to work. Here is what I know:
• The building is a simple 4-side masonry box. The block walls parapet on 3 sides.
• The interior frames run the 120’ direction and the end of the roof beams set on the 84’ hollow block walls. There are no roof rods, x-bracing, portal frames, K-bracing etc present anywhere.
• There is no visible connection of the bar joists or the underside of the metal deck to the parapet block wall on the 2 walls that are parallel to the bar joists. I can see up the full depth of the 1.5” roof deck and see no connection from the wall to the roof panel or roof structure. After the 1.5” metal deck, I cannot see anything. 1" Insulation and tar block anymore view.
• The only portion of the roof that definitely connects to the block wall is the ends of the 2 roof girders. They set at the 3rd points of the 84’ block wall.
• The masonry walls have no bond beams on the walls parallel to the bar joists. The other 2 walls I am not sure about.
• The roof definitely is not concrete, it is tar and gravel.

Bear in mind, where I live, we have some of the lightest design loads and actual loads. For that reason, poor designs can still stand up because they rarely get tested. I think the highest wind we have had in years was around 58 mph in the old way of measuring wind. At that point in time, the older codes made this a 70 mph design. In the old code, 10 psf was the total wind to the MWFRS. So 58 mph would be about 7 psf. Also in my location, architects have been allowed to do entire building designs via the “allowed to do engineering incidental to architecture” clause. Our actual seismic is non-existent for all practical purposes. I am less than 100 miles from a seismic zone 0 in the old codes.

My original thoughts were this is a metal diaphragm roof and 4 masonry shear walls. But not seeing a physical connection of the roof to the sidewalls has me puzzled. I have never seen anyone just use the ends of the girders and the corner masonry wall connections as their only way of transferring shear for wind or seismic. Am I missing the presence of some other system or is this just a poor design that has withstood time? Is there a good way to connect the roof to the walls from above the metal deck that I do not know of?

 

[bob33]

For wind parallel to the 84' walls, the beams transfer the diaphragm load to the shear walls. The end joists are the drag struts. It is not clear how the 120' long chords work. Maybe the deck at the perimeter acts as the chord. Maybe the wall does. A very poor design but that is probably how it works.
I would also be concerned about the minimal lateral support for the tops of the 84' walls if there are no anchors.

 

[TpaRAF2]

Hi, Ron247

My impression this is basically a girder-beam-stringer system.
The girders are obvious in the first photo. The bar joists are likely simple-span, and serve as the beams. The roof deck acts as the stringers between the beams.

Since the wall parallel to the bar joists supports the end of the girder supporting the bar joists, you cannot remove it. My quick estimate is about 25% of the total roof load occurs at the end of the girder.

The deck is probably fastened to the bar joist top angles. This can be done with redheads, sheet metal screws or even tack welds. In turn, the ends of the bar joists rest in masonry pockets or are fastened to the steel girder. The end of the girder sits in another masonry pocket or blockout.

Since the bar joists support the deck, and the 2nd photo shows a joist really close to the wall, a structural connection between the deck and wall is probably not present. I figure they'd want to seal it though, to keep out gusting rains and winds. Should be some sort of attachment for that purpose.

I would also bet the hollow cells under the girder ends have been filled with concrete, basically making a pilaster.

Since you're the eyes on the ground here, you'll need to investigate these 'armchair quarterback' guesses.

I don't think you can take out those end walls. But we're engineers, so given enough money (to build it) something can be figured out! The compromise is to give them two 20' openings in the end wall, after installing a column and foundation to support the girder end.

Good luck with your project! RAF

 

[Ron247]

The deck is welded to the bar joist. You can occasionally see signs of a weld, but not often. Due to our light loading, there is less than 6 kips of shear for wind. The girder ends do appear to be the only shear transfer method. It is the only thing I see that actually connects to the wall other than the masonry corners that have many hairline cracks in the mortar joints at the corner. The hairline cracks do not appear to be vertical settlement related. Once you move away from the corners, the hairline cracks disappear.

TpaRAF2, they did grout one vertical cell near the girder. They missed the actual girder location by 12" but got close enough.

 

[gte447f]

Unless I am missing something (quite possible), this does seem to be poor design/construction, or at least it doesn't seem to match up to the theoretical load paths that we are taught as engineers. I share your concerns for lack of shear connection between roof diaphragm and 84 ft wall, and for lack of out of plane bracing for the top of the 84 ft wall. Also consider wind in the other direction (perpendicular to the 84 ft wall). How does that wind load get into the roof diaphragm and over to the 120 ft walls if there is no connection between the 84 ft wall and the roof diaphragm? I guess the only mechanism is 2-way plate bending of the 84 ft wall between the foundation, the 90 wall corners, and the restraint points at the ends of the girders.

 

[TpaRAF2]

Hi, Ron247

Is it possible there are 'U' block in the 84' wall at various courses? That would allow some reinforcement of that wall. The images do not show any lintels or really anything other than masonry courses.

The use of masonry 'ladders' (within the mortar) is also common, but I do not think it counts for adequate shear wall bracing. Maybe a bit of chipping into the mortar near the corner will help you find them if present.

If the filled cell is well constructed, that serves to divide the wall into two panels. Then, as gte447f suggests, the 2-way distribution is a bit more feasible. There should be signs of 'X' shaped cracking in the grout about the middle of this panel, where the wall can flex or deflect the most if gte's hunch is correct. Maybe 21' over and 9' up, or H/2 over and H/2 up (lazy 'Y' crack).

This still gets you no closer to figuring out what to do about dropping the wall. As 'crappy' as the fabrication appears, you might get more demo than you planned on. I think some supplemental construction would be needed to adequately frame the ends of the long walls and support the steel beam, then you can drop the wall (and build a better one).

As far as the roof goes, I would think the deck edge is cantilevered from that last bar joist. I am guessing it only extends 1 - 2' past the short wall. Since there is no connection, dropping the wall * should * not causes problems with the roof framing.

Best of luck on this! RAF

 

[Ron247]

Thanks all for the input. It seems the 2 girders are the only shear transfer mechanisms. Due to the light loading and lack of any seismic, this has worked for at least 40 years.

I have pretty much told them we can remove the wall but will have to construct a very stiff x-braced frame to replace the lateral ability of the wall. I cannot use any system that allows too much movement to occur.