Reinforced Concrete Wall Designed as Compression Member

[RFreund]

What is the effective length of wall that should be used when analyzing and designing a RC wall as a compression member (when subjected to a concentrated load)? ACI318-08 states in Section 14.8.2.5 that concentrated loads can be distributed down at a 2V:1H. However this is in the alternative design section. Can/should this provision be used when analyzing the section in accordance w/ chapter 10 of the ACI318?

EIT

http://www.HowToEngineer.com

 

[ishvaaag]

Today one can easily model walls through plate elements and then support (with proper continuity) columns at the edge; this would be the natural election for a proper model. However one could investigate a number of cases parametrically to see what width of wall best mimicks, all over the set of hypotheses, the behavior. Typically for column behavior a common rule is not to accept as columns anything with ratios over 4 to 1 in plan.

 

[grantstructure]

If you are having to provide ties for the compression element, you're limited by 10.8.2.

Otherwise, assuming no openings to restrict the redistribution of load, what I do is to spread the load at 2:1 up to the mid-height of the wall. Limit that effective width to 4H plus bearing width where H is the thickness of the wall. (This always controls). You can go up to 1% steel without having to add ties. Equation 14-1 is great unless you're really trying to push it. But never use a k of less than 1.0, even if it's restrained. Ask Dr. Yura why.

Since width drops out of computing r, you're not super sensitive to what you choose here, as long as you check the wall at a realistic load intensity (how much you've spread it out). Finite element is great if you have the tools, but it's not hard to get reasonable designs without it. If you have out-of-plane loads as well be sure you include those and look at p-delta.

 

[RFreund]

Thanks for the replies.
Ishvaaag - Thanks for the input, I was hoping to avoid FEA but I'm sure it would be good.

Grant - Thanks for the input. That was along the lines of what I was thinking. I use 4*t (wall thickness) for masonry is this defined in ACI318 somewhere as well? Good call on the 1% as I was only using steel in compression unless it was tied (another carry over from masonry).

Basically I am trying to go through and produce a very comprehensive retaining wall design spreadsheet. I wanted to get an idea of how to handle heavy point loads (basement wall condition).

Thanks again!


EIT

http://www.HowToEngineer.com

 

[KootK]

@ Grantstructure: Dr. Yura doesn't return my calls any more after the "incident". What is the rationale for using K=1.0, even under restrained conditions? That's what I do. I'd just like to know the argument that you're referring to.

Thanks,

KootK

 

[grantstructure]

Kootk, it has to do with the stiffness of the brace required to really be "non-sway". Pcr/Peuler approaches the k=0.7 line asymptotically, if you assume the case of a fixed base-pinned top condition, but really takes a while to "get there". To really have a k of 0.7, say, you have to have a truly infinitely stiff base, and a very, very stiff system bracing the top of the (pinned) column. It's not generally practical to make the system as stiff as it would have to be for both of those things to be true.

The same effect happens in the alignment chart (on the non-sway side). There, you can account for the actual stiffness/restraint that is supplied at the top and bottom of the member, but (and we're reaching back in my memory here, which hurts a little) when you add the actual displacement that happens (since nothing is truly zero-sway) then the true k value goes up pretty quick. Luckily 1.0 is a nice safe assumption for a wall (pinned top and base; any restraint you do have doesn't actually hurt) and the brace stiffness required is achievable.

As for where 4t shows up in ACI, I don't recall if it does or if that's a rule of thumb I've just always used.