Force Transfer Through Anchors 2

[Gopher13]

All,

I have an existing building where a new mechanical unit is being installed on the roof. The additional load causes the existing steel roof joists to be overstressed, so I will be adding new joists to share in taking this load. On one end, I am forced into bearing the new joist on a bent plate (or channel) that is adhesive anchored to the wall. How does the gravity load transferred to the anchors? Does each column of anchors take load based on the stiffness of that row similar to what happens with a rigid diaphragm? So in my drawing attached would all anchors share the load equally, or would the center two rows of anchors take most of the load?

 

[dik]

I would let the top anchor take the tension and distribute the shear equally between the two anchors.

Dik

 

[PengStruct]

If the angle or channel is stiff enough, I would say that the shear could be distributed evenly among all anchors. Hopefully you also have a fully grouted CMU wall. However, As dik mentions, you should account for tension as well as shear in the upper row of anchors.

 

[Nor Cal SE]

Both responses are good, but here is one additional thought. If your space on the wall below can sustain it, and if your required bearing width is not huge, you could consider a narrower angle that has only two columns with perhaps 3 anchors in each column, therefore 6 total (if that's enough for the load). You could then assume the top 2 anchors take all the tension, and the bottom 4 all the shear, for simplicity. This angle depth increase will serve to increase your moment arm and thereby reduce the tension.

A few other thoughts to consider:
1) Unless you have a door or window opening below this connection, you might have a very large distance to the nearest CMU edge, which will result in large capacity and you might not need many anchors. Shear capacity is primarily dependent on edge distance as opposed to embedment, at least for concrete; I haven't done much masonry anchorage lately, so you can check the code.
2) Depending on your distance from load 'P' to the wall, you might want to check capacity of the angle's horizontal leg in weak axis bending. It will likely be okay if your distance and load are not extreme.

 

[Lnewqban]

I have an existing building where a new mechanical unit is being installed on the roof. The additional load causes the existing steel roof joists to be overstressed, so I will be adding new joists to share in taking this load.

Consider/negotiate the location and orientation of the new roof equipment.
Moving the location away from the mid-span of the joists may be feasible (most codes mandate some safe distance (~10 feet minimum)from the edge of the roof for service/maintenance purposes).

If a commercial package unit over a roof curb, laying the longer side of the curb across the joists normally distributes the weight and the wind-induced turn over load over three joists.
If condensing units on a roof rack, the same principle may be used via distributing the legs of the rack.

"God will not look you over for medals, degrees or diplomas, but for scars." - Elbert Hubbard

 

[KootK]

I'd share the shear load equally among the anchors. The angle doesn't span far enough between anchors for angle flexure to enter into play in a meaningful way.

I would recommend either flipping the angle or fitting it with stiffeners. As you have it, a prying mechanism will exist that will amplify the tension in the upper anchors in a way that is difficult to predict.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Ron]

Without regard to how the load gets transferred, I would be concerned about transferring load from a mechanical device that likely also transmits vibration to an adhesive attachment. If the support is transferring vertical shear through the adhesive with no redundancy, I would probably not do it.

 

[Gopher13]

Thanks for all the responses!
As suggested, I have designed the top anchors for combined tension and shear and the bottom anchors for shear.
I don't want to rotate the angle 180 degrees because I don't want the contractor to have to have to mess around with removing the roofing and flashing on the parapet above. Plus, aren't prying forces present just as much with the angle rotated 180 degrees anyway?
Moving the mechanical unit to a different location on the roof doesn't really help either. It is large enough to require snow drifting.
Attached is the detail that I decided on. I did add stiffeners as suggested. I also added a vertical plate running parallel to the wall spanning between the stiffeners. It is a little Harlem Globetrotter style razzle dazzle (and perhaps unnecessary) sending the load from the horizontal leg of the bent plate into this plate, into the stiffeners, into the vertical leg of the bent plate, then finally into the anchors, but it provides me with a traceable load path ensuring all anchors share the shear load equally.

I do have a follow up question; On the detail I am showing the existing partially grouted cmu wall having the cores grouted at and around the connection. I have seen other engineers show this on details too. How is this done? Does the contractor knock out a face shell above and below, block out the core below, fill just that section with grout, and then repair the face shells, or do they just knock out the face shell above and fill the wall full height with grout?

 

[KootK]

Gorgeous detailing.

Plus, aren't prying forces present just as much with the angle rotated 180 degrees anyway?

It depends somewhat on your arrangement but, in general, no. With the vertical leg down you're prying will be greater and, probably more importantly, more difficult to predict. The prediction difficulties stem from the fact that, with the angle leg pointing down, it tends to press into the masonry whereas, with the angle let pointing up, it tends to pull away from the masonry. That said, I certainly understand your spatial limitation at a roof condition.

Does the contractor knock out a face shell above and below, block out the core below, fill just that section with grout, and then repair the face shells

I believe that it's this. In my market, this kind of masonry work tends to be pretty unreliable. Here, on a project of mine, I wouldn't count on it getting done unless I were able to have boots on the ground to verify that it did. Really, I'd prefer a hollow anchor solution (HY70) either way, even if it meant a larger bracket. I've verified grouted cores in the past, had them be deficient, and still wound up with a black hat on my head because I was slowing down the project and only called out the grout in two places rather than... I don't know... the forty or so places that it apparently would have taken to get the point across.

Some general, non-mission critical suggestions on your detailing:

1) 5/8" plate seems huge to me for the application. I don't know your loads but, if a 2.5" joist seat is going to survive, I'd think that 5/16" or 3/8" plate would work just fine too, particularly with the stiffeners. The 5/8" plate may cause you a couple of problems. Firstly, your bracket will have a weight on par with a compact car. Secondly, your bend radius on a 5/8" plate is going to be 2-3" which will ugly up your detail some.

2) Like you, I love your vertical plate from a mechanical load transfer perspective. It just feels like it's too much however. I'd eliminate it and span horizontally between the stiffeners. This will save on some fabrication, look a bit more "normal", and allow you to shift your upper bolt upwards some without screwing with installation tolerances.

3) Combining the two suggestions above, I see it like this:

- angle welded rather than bent.
- vertical angle leg 3/8" max
- stiffeners 3/8" max
- horizontal leg 1/2" or whatever is required for it to span between stiffeners.

4) In commercial situations, I never use an anchor smaller than 5/8". I just think that 1/2" anchors look puny, especially installed in 5/8" plate.

5) You may not have space to make your topside joist field weld. Some options:

- Overhead welding.
- Increase joist seat to make space.
- Remove roof and deck locally.
- Provide a bearing plate that gets bolted to your angle (my favorite).
- Somehow weld the bracket to the joist prior to installation.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Gopher13]

Kootk: Thank you for your input. I really appreciate it.

You are right. The plate doesn't need to be as thick as it is. However, being as large as it is, I thought it should be at least 1/2" thick, and for some reason, at the time, 5/8" felt good. But you make a great point. I will reduce it's thickness.

Also a great point on the bending radius. That is something I never would have thought of. I love your idea of a fabricated angle made from two plates welded together.

I have heard of other engineers not wanting to use anything smaller than 5/8" diameter anchor as well. However, per HILTI's manual, with a spacing of 8", there is an adjustment factor for shear of something like 0.67 for a 5/8" diameter anchor, whereas the adjustment is 1.0 for a 1/2" diameter anchor. So for all four of my interior anchor, assuming 5/8" diameter anchors, the allowable shear load is the table value multiplied by 0.67^3 multiplied by any other adjustment factor. This results in number considerably less than that of the 1/2" diameter anchor. Seems counterintuitive, and I don't know if I buy it, but I like to have traceable calcs.

Just as you warned, I am nervous about them grouting the cores correctly. I am providing two additional anchors (calcs. call for 6 and I am providing 8) as an additional safety factor. The building is a few buildings down from the building I work in so I will be able to keep a close eye on them. I also will have the opportunity to write the special inspection contract so I could require the special inspector to witness all grouting and anchor installations. Plus, I have never really understood how the HY70 system works (my fault - I should research that). To me, it seems as though the only capacity comes from the 1.25" of anchor through the outermost face shell and the rest of the anchor, mesh basket, and adhesive just kind of flops around?

I also like you idea of a bearing plate being bolted to the horizontal leg of the angle. I am going to use that.

Again, thanks a bunch for your responses!

 

[jayrod12]

However, per HILTI's manual, with a spacing of 8", there is an adjustment factor for shear of something like 0.67 for a 5/8" diameter anchor, whereas the adjustment is 1.0 for a 1/2" diameter anchor. So for all four of my interior anchor, assuming 5/8" diameter anchors, the allowable shear load is the table value multiplied by 0.67^3 multiplied by any other adjustment factor. This results in number considerably less than that of the 1/2" diameter anchor. Seems counterintuitive, and I don't know if I buy it, but I like to have traceable calcs.

I too have noticed this. Really for me I rationalized using 1/2" anchors in masonry applications because it's never the steel strength governing the design, so what do I care whether I meet my arbitrary minimum. And I have found that you can have less anchors when space is tight using 1/2" compared to 5/8" when "properly" applying the reduction factors.

For concrete applications, 5/8" minimum.

 

[KootK]

This results in number considerably less than that of the 1/2" diameter anchor. Seems counterintuitive, and I don't know if I buy it, but I like to have traceable calcs.

I design to the 1/2" anchors and then just toss the 5/8" anchors in there anyhow under my strong belief that adding 1/8" to the diameter couldn't possibly make anything worse. But then I have the luxury of mostly working in Canada where we, oddly, are usually not required to submit our calculations. I certainly agree with your stance on that front.

To me, it seems as though the only capacity comes from the 1.25" of anchor through the outermost face shell and the rest of the anchor, mesh basket, and adhesive just kind of flops around?

Yeah, you got me there. I feel exactly the same way. That said, the HY70 numbers are based on testing which gives me confidence in their capacity even if I don't love the picture that you've painted. To play devil's advocate, consider potential weaknesses associated with the HY200's in post-grouted cells:

1) A mason has to break off the face shell, stuff/pour it with grout, and then kinda set a new face shell in there. In my opinion, the face shell is just cosmetic at this point. I've seen some just fall off, seemingly of their own volition.

2) Because of #1, I consider your structurally active system to not include the new, cosmetic face shell. So what you've got is a three sided block, a blob of grout sitting in it, and an anchor that cantilevers 1.25" from the grout.

3) Correctly me if I'm wrong, but I believe that the HY200 values would be based on blocks that were grouted without messing with the face shells. If that's true, then your HY200 values really are not based on testing for this scenario. Scary.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Gopher13]

1) A mason has to break off the face shell, stuff/pour it with grout, and then kinda set a new face shell in there. In my opinion, the face shell is just cosmetic at this point. I've seen some just fall off, seemingly of their own volition.

2) Because of #1, I consider your structurally active system to not include the new, cosmetic face shell. So what you've got is a three sided block, a blob of grout sitting in it, and an anchor that cantilevers 1.25" from the grout.

3) Correctly me if I'm wrong, but I believe that the HY200 values would be based on blocks that were grouted without messing with the face shells. If that's true, then your HY200 values really are not based on testing for this scenario. Scary.

Swear word!!! Now I am hesitant to use both masonry adhesive anchoring systems. Seems to me the best and safest thing to do is to bear the joist on the existing wall and deal with the extra cost of tearing into and repairing the roofing and parapet as a result of it.

 

[KootK]

Meh, I'm sure it'll be fine. it's funny, I've no doubt that through bolting would create the strongest connection but, then, there seems to be no accepted methodology for that whatsoever. And you may not be able to muck with the exterior anyhow.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[haynewp]

When I've given the option, contractors have more often picked reinforcing an existing joist and reinforced the existing connection as needed rather than getting new spliced ones into an existing building with new connections. If the existing joists bear on top of the CMU wall and not on such a bracket then that may be a better way if the bearing checks.

 

[structSU10]

If you are worried about it, you could pocket a steel beam into the wall. That way, you knock out the face shell and full the whole cavity with grout below the new bearing plate. might need to blow out a few blocks across the width to have swing space to get the beam in. This assumes you have a different attachment to the other side, such as a steel beam with a new extended shear tab.