Residential Deck Engineering Seismic

[rollinw]

I am designing a 10x36 covered residential deck attached to the end of a 36' wide house with 4x4 knee braces at the 8' tall posts in an area with high seismicity.

The plan reviewer here requested a seismic load calculation for the 36' long line and I'm curious what R factor is appropriate? From other posts/examples I've seen, it seems most have used the R factor for the overall structure the decks are attached to, and the ACSE 7-05 (code we are using) doesn't really seem to have an appropriate R factor choice for a knee brace situation...

 

[jdgengineer]

I use timber frame under cantilevered column section I believe R = 1.5.

 

[XR250]

What are you guys using for lateral load just due to occupancy?

 

[jdgengineer]

Not to get too far from the OP's question, but XR250 here is a good structure magazine article from a few years ago.

http://www.structuremag.org/?p=1353i

Our seismic forces are usually high enough that we end up designing for ~5 PSF lateral load. This is a bit lower than recommended in the structure magazine article.

I also saw a research paper which tested decks laterally. The conclusions are not great but I think the jist is that for a 1:1 or less aspect ratio that using lag bolts to attach the ledger provides a fair amount of lateral strength. They also recommend using screws for joists hangers. We haven't typically done this but do specify DTT2Z holdowns at 8'-0" oc typically.

http://www.forestprod.org/buy_publi...0savSVJopV0YOeBCg&sig2=aPMl0oHQSKXUUrmkIE9KiQ

 

[KootK]

and the ACSE 7-05 (code we are using) doesn't really seem to have an appropriate R factor choice for a knee brace situation.

Yeah, it's almost as though the timber frame industry forgot to show up to the whole seismic design discussion. Common systems that seem to have no representation:

- knee braced timber frames
- timber braced timber frames
- steel braced timber frames

I imagine that the nature of the connections would usually be such that R is always going to be pretty close to one. Still, it would be nice to see that in print someplace if that's the case.

It's tempting to just assume that all timber frame lateral systems are cantilever column systems just because that's the only place where the phrase "timber frame" shows up. On the other hand, "cantilever column" suggests a fixed base which is pretty rare in timber frame except when columns are embedded fence post style. One can argue that knee braced columns actually cantilever down from the fixed-ish connection at the top. It seems like a bit of a reach though, particularly given the flexibility of the beam forming the top of what is effectively a moment frame.

In Canada, it works like this:

R=3.00; Braced or moment frames in timber with ductile connections - moderate ductility
R=2.25; Braced or moment frames in timber with ductile connections - limited ductility
R=1.00; Other timber stuff not listed

In this context, I'd go R=1.00. To my knowledge, the Canadian codes don't elaborate on which connection systems are limited/moderately/not ductile but it's certainly difficult to see how a knee brace would generate much ductility. And, whatever thee would be, it would probably be limited to just the compression cycle and perhaps prone to ratcheting.

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.

 

[jdgengineer]

KootK brings up some good points. I feel the R = 1.5 for timber frame under cantilevered column felt pretty low and seemed reasonable to me for a residential deck. Plan reviewers have also not had a concern with this approach in the past. I agree, using R= 1.0 would be a more conservative approach since there is pretty limited guidance in the code.

Oddly enough timber braced frames used to be in the UBC before us western states changed to the UBC.

 

[KootK]

@jd:

-know what the UBC R values were? I'm curious.

-I go 1.0 based on the canadian options. I'd be fine with 1.5 in your environment with building officials accepting etc. 1.0 and 1.5 are close enough to the same number here in my book.

-if one can get any reliable ductility out of the connections, it wouldntn't surprise me to find that the overall system is quite ductile. Probably closer to a moment frame. It doesn't take much local joint rotational ductility to translate into a whole lot of global displacement ductility in a moment frame.

- do you have any idea what kind of system the timber cantilever column system is meant to apply to? It just seems like such an improbable choice for the only timber frame option. Pole barns?

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.

 

[rollinw]

The 1991 UCB shows a R value of 4 for Bearg wall systems, 8 for building frame systems with concentrically braced frames... one of the reasons 1 or 1.5 was an eyebrow raiser at first.

Found this relevant FEMA paper looking around:

https://www.fema.gov/media-library-...850c3050c34da135376baa478a1b/P-752_Unit11.pdf

thanks for the input...

 

[XR250]

Not to get too far from the OP's question, but XR250 here is a good structure magazine article from a few years ago.

Thanks.

Slightly OT: In our area most knee braces are 4x4's lagged at each end. Except at the top, they have found a way to get capacity from a lag by installing it between the plys of a double. 2x10 beam. I won't even design a deck with kneebraces anymore. If they really want to use them, I call out, "Bracing per code". Most arch's want to use hidden bracing system anyway which I can design more easily with less push back on the details.

 

[KootK]

Is hidden bracing horizontal bracing under the deck?

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.

 

[XR250]

Correct

 

[msquared48]

Just have to be able to tie in the end members of the deck to the building structure with straps to develop the tension forces generated by the bracing.

I have used this type of system many times.

Mike McCann, PE, SE (WA)

 

[rollinw]

I'm curious about the type of horizontal bracing you would prescribe for a wood deck like I mentioned above XR250? Would these be diagonals in the plane of the diaphragm? Diagonal straps? Sounds like R=4 would be an appropriate basis for seismic load.

As for knee braces and taking a cantilever column approach, ASCE 7-05 section 12.2.5.2 and the 15% of axial design load max can really beef things up...

 

[XR250]

Hers is on I did recently. This one is only designed for the braces to work in tension.

 

[jdgengineer]

If you are using this bracing system I think you could argue using the same R as the main house. Essentially the balcony is just a cantilevered diaphragm. I'd be a bit conservative with the design and make sure to check deflection at the free end of the balcony.

XR250 - how do you anchor the diagonal to the main house? It looked like just a couple of bolts to the rim joist?

 

[XR250]

XR250 - how do you anchor the diagonal to the main house? It looked like just a couple of bolts to the rim joist?

I thru bolt a 6x6 pt block to the house band with (2) bolts. The house band is an LVL. The block is also typically end-nailed to the adjacent joists

 

[rollinw]

Thanks for the post XR250. Is the reason for using a wood brace rather than a metal strap because of exposure? Seems like it would be easier to get the necessary end nailing with a strap in that situation..

 

[XR250]

That should work as well. Using wood is one less material the framers have to order. Also, it can act in compression for additional redundancy.

 

[rollinw]

So the plan checker is asking for diaphragm deflection calcs on the metal strap horizontal strap brace... which I could gather for a cantilever diaphragm you would use delta = S * ( a / Ga ) with Ga being the apparent shear stiffness. Anybody use a different method? I am not sure how to get Ga for this situation.

 

[KootK]

I'd be inclined to look at the brace as a discrete truss and calculate deflections based on that (aggregate effect of PL/AE in each member).

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.