Diagphragms, Shearwalls and Drag Struts

[strguy11]

There is a debate in my office regarding the use of drag struts when the length of shearwalls are less than the width of the diaphragm

We have a typical apartment style building with a center longitudinal corridor with units on either side. In most cases, the demising walls between the units are the shearwalls and they extend from the exterior wall to the corridor walls. In some cases the shearwalls do not line up across the corridor. The questions/debate is as follows so i am curious what others are doing.

1.When looking at the diaphragm, some have suggested using 2 diaphragms (a leeward and windward). By doing this, the diaphragm chords are the exterior wall and corridor walls. By doing this, the shearwall extends the length of the diaphragm and no drag struts are needed.

2. Look at the building diaphragm as one full-with diaphragm and provide drag struts at each shearwall that is not the full width of the diaphragm. (i.e. at every shearwall since the corridor breaks them up.

3. Look at the building diaphragm as one full-with diaphragm, and check the diaphragm capacity based only on the length of the shearwall. If the capacity is greater, than no dragstrut is required. I.e. provide a dragstrut only when a width of diaphragm greater than the shearwall length is required.


Just out of curiosity, what are most of you doing/seeing? I have been reviewing many example sets of plans from engineers across the country, and it appears that either method #1 or #3 above is being utilized. There are even some cases where the shearwall do not extend to the corridor and no drag strut is called out, leaving me to believe that Option #3 is used.

Thoughts?

 

[DaveAtkins]

At the roof, I have always been more concerned with how the lateral load gets from the diaphragm (at the sloping top chord of the truss) to the shear wall below, which may not be positioned directly below the truss. I typically show extra roof trusses on the drawings to make sure each shear wall gets its own truss, and I specify that the trusses which are over the shear walls are designed to transfer the force from top chord to bottom chord.

But to answer your question, which is also applicable to floor diaphragms, I typically use method 1, since if you have a lot of shear walls, you don't need a large, deep diaphragm.

DaveAtkins

 

[Deker]

My preference is #2. If that's not possible, drag struts need to be provided somewhere within the span of the diaphragm for there to be a complete load path that doesn't tear the diaphragm apart. A single large drag strut can be provided adjacent to the wall, or multiple smaller drag struts can be provided throughout the span of the diaphragm. In either case, a transfer diaphragm needs to be designed to complete the load path. Option #3 doesn't consider a complete load path so I wouldn't consider it. Terry Malone provides multiple ways to analyze this exact condition in section 12.4 of his book which I highly recommend (Link). From that section:

There have been occasions where collectors at transverse shear walls were not installed. The rationale behind this approach is the assumption that by using every transverse wall, the diaphragm forces are reduced to a level where collectors and chords can be ignored. Although to some this may seem reasonable, it is not consistent with code, which requires complete load paths.

 

[MrHershey]

We usually do #3. It seems to be at odds with many of the webinars and other continuing education presentations we've seen, though those are usually looking at "easy" buildings. Main concern with this method would be the diaphragm wanting to essentially separate into the two diaphragms of method 1, as Deker references. Also the vast majority of our flexible diaphragms are metal deck with higher capacities, if we were doing a ton of wood we might feel different. If you've got horizontal members spanning in same direction as wall they may effectively tie the diaphragms together and act as a drag strut from diaphragm to diaphragm even though they're not directly connected to wall. If there aren't members spanning in same direction as wall, more consideration may be needed to make sure diaphragms aren't splitting apart. Whether that's checking connectors at interface between the two diaphragms or whatever.

When they're not quite aligned across a corridor, I've also done a little subdiaphragm similar to what Terry Malone does with disrupted chords in the book referenced above. There's also a shorter paper that is freely available from Terry Malone that covers this, though briefer: Link. Instead of moving a chord force over in plan, you're moving a drag strut force over in plan and then from an analysis standpoint your two walls across the corridor are locked together and thus behave as if they're in the same line. That way you don't have to extend drag strut across the whole building but still maintain a continuous load path.