Concrete moment frame with reinforced CMU infill walls - connections to bounding frame

[bones206]

I have a municipal project in Miami that died on the vine a year ago but has now been resurrected. It's a concrete moment frame structure with reinforced CMU infill wall panels, 280' x 135' x 40' high, with a stucco finish. One of the issues I was struggling with early on was how to detail and analyze this structural system. Based on my research, the common practice for building reinforced CMU infill walls in Florida has been to construct the frame bounding first and provide cast-in dowels, form savers or pvc sleeves to allow continuous CMU wall reinforcement and to tie it into the frame members.

I've also learned that the MSJC Appendix B has very specific detailing requirements for infill walls, which I understand are relatively new code requirements. One requirement is that CMU wall connections shall not transfer in plane forces:

B.3.2 In-plane connection requirements for participating infills:

Mechanical connections between the infill and the bounding frame shall be permitted provided that they do not transfer in-plane forces between the infill and the bounding frame.

Code Commentary for B.3.2...

The modeling provisions of Appendix B for participating infills assume that in-plane loads are resisted by the infill by a diagonal compression strut, which does not rely upon mechanical connectors to transfer in-plane load. While mechanical connections, including the use of reinforcement, are permitted, they must be detailed to preclude load transfer between the infill and bounding frame. This is because mechanical connectors between the infill and bounding frame can cause premature damage along the boundaries of the infill under in-plane loading (Dawe and Seah, 1989a). This damage actually reduces the out-of-plane capacity of the infill, as the ability of the infill to have arching action is reduced.

This requirement appears to conflict with what I believe are the traditional construction practices in the area. CMU rebar tying into the bounding frame members will definitely transfer in-plane forces. The commentary even alludes to the use of rebar so long as it is detailed not to transfer in-plane forces. I can't envision what that detail would look like. So my questions are:

1) Is my interpretation of the code correct?
2) Am I correct about the traditional construction practices for this type of system?
3) If so, has the construction industry in Florida begun to adapt their practices to this requirement? Have engineers changed how they detail connections to the bounding frame so that they do not transfer in-plane forces (i.e. no dowels)? Or are people still doing things the way they always have and pretending Appendix B doesn't exist?
4) This is sort of a side issue, but does anyone actually do the equivalent strut frame analysis required by Appendix B in current practice? It seems cumbersome for a large building with dozens of unique infill panels, but I don't see any way around doing it.

 

[GC_Hopi]

1) It appears so... Link
2) I would say so.
3) Again, I would say so but it is a slow adaption. It might pick up if the code official discover this appendix. The move is towards clips, slotted connections, etc Link
4) I would say that my interpretation of code's intent is to push designers to detail the majority of walls as non participating then select specific walls to detail as participating. Something similar to steel design where you have the lateral resisting system in certain bays or something

 

[bones206]

It would be easy to provide embeds in the concrete frame members and welding on retaining plates or clips, but I'm not sure if that would interfere with the stucco finish. Even if I used this detail, the only difference in detailing between participating and non-participating walls would be whether or not I left a gap around the wall or mortared it in, right? What would be the incentive to leave a gap and not take advantage of the shear wall action? It would probably even be cheaper than adding compressible filler material and caulking.

I spoke to a supplier of the dovetail-slot-mounted partition top anchor rods (PTA's) last year and they couldn't confirm that they were code compliant for either hurricane wind loads or for allowing in-plane movement. I've never really trusted anything slotted to reliably allow movement in the real world, and their reluctance to recommend their product for this didn't inspire confidence. I haven't seen any testing to justify using them for this application either.

The other concept I'm struggling with is infill panels with openings. The code says any wall with an opening cannot be treated as participating infill, but the force effects of the wall on the frame must be considered. I'm not sure how to simultaneously model the effects of the infill on the bounding frame while also not taking credit for it's effect on the lateral resistance of the frame as a whole. I guess I have to run the analysis once with the diagonal strut in place, and run it again with the strut removed, and envelope the frame design for both sets of analysis. Either that or leave isolation joints around all sides of any panel with an opening to preclude any interaction.

Unfortunately I don't have a choice in the structural system for this building, since the owner wants to match existing buildings on site. The existing buildings were designed before Appendix B came about and have rebar from the infill walls dowelled into the surrounding frame members on all 4 sides.

 

[GC_Hopi]

There is a minimum gap distance and that appears to be the main differentiation. The incentive is that the code writers are concerned about premature damage at the dowels due to in-plane loading weakening the out-plane capacity.

Interesting all the discussion in Appendix B is geared towards seismic. It makes one wonder if you are designing in the elastic range how applicable is the commentary on mechanical connectors causing premature damage. There is no damage expected if the design is elastic.

 

[MotorCity]

If you want the moment frame to behave like a moment frame, you should not tie the infill wall to the moment frame. Otherwise you may end up with more of a shear wall than a moment frame (depending on their relative stiffness).

 

[bones206]

@GC_Hopi - Yes, it seems geared towards seismic (i.e. high drift ratios) and also seems to have an underlying assumption of an unreinforced infill. I don't think the out-of-plane instability would occur if you have a reinforced infill or low drifts.

@MotorCity - As I understand it, the economy/efficiency of this system has traditionally been from the shared lateral resistance of the hybrid system. Since many infill panels contain openings and the new Appendix B disqualifies those panels from providing lateral resistance, the overall efficiency of the hybrid system seems to be quite diminished. Since it appears to be a radical departure from traditional analysis and detailing practices, I'm interested in the experiences/opinions of others regarding this Appendix B -- especially in Florida.

It's clear from the low volume of responses here that this particular code issue doesn't affect many engineers, which is probably why it hasn't evolved in recent years to clarify the confusing/conflicting requirements.

 

[MotorCity]

bones206 - thanks for the explanation. Out of curiosity, is there a response modification factor (R) that has been established for this lateral system?

 

[GC_Hopi]

Nothing official but TMS 402 commentary says "Until further research is completed, the Committee recommends using the smallest R and Cd value..." So default to your engineering judgement.

 

[bones206]

Seismic isn't required in Florida, so thankfully I don't have to deal with that.

 

[JLNJ]

If you don't intentionally make the infill participating (or non-participating) you might accidentally create a soft story at the non-participating frames. Obviously a seismic no-no.

It seems like an easy thing to postulate about and write into the code, but difficult to detail out and convince contractors it needs to be done. Most cities are full of older buildings with this type of construction.

 

[bones206]

I reached out to the Masonry Association of Florida for their perspective. I'll share if I learn anything.

 

[slickdeals]

Is there a reason you are doing an infill CMU as opposed to doing a load-bearing CMU on the perimeter?

 

[bones206]

It's basically an owner directive to match existing structures that were built on site in the last 10 years. Originally the project was supposed to be a "carbon copy" of an existing building, with minimal budget for engineering. But the layout has been significantly modified, so I have to now redesign the building while remaining faithful to the original structural concept.

If I had my way, I would build it like a Home Depot: tilt up walls with metal decking roof over bar joists, steel columns to support the bridge cranes and roof gravity loads. i think it would cut the building weight and construction time in half, at least. I offered to do a cost comparison and see how much we could save the taxpayers, but was told no thanks...