Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Re: Plwd: Rigid Diaphragm Analysis - Opinions Wanted

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Just the type of dialog that I was looking for - thanks. Here are my feelings 
about them - see below in your comments:

In a message dated 7/11/99 2:25:27 PM Pacific Daylight Time, 
jwatson(--nospam--at)inconnect.com writes:

<< In response to 3. - True, but by placing straps at each point of
 intersection you create a situation similar to that of a shearwall.  You
 place a holdown at the end of a shearwall to prevent uplift (with only
 minor displacement).  By placing straps at both sides of the
 intersection of you "U" shape, you will get tension and compression at
 each of you chord locations.  This can give a close model to a fixed
 connection, again think about it like a cantilevered wood shearwall. 
 But this comes back to how you want it to perform.  If you don't place
 the straps, then the building will separate there (needs an expansion
 joint).  If designed to accommodate the separation, then either method
 should be acceptable. I guess it's kind of like a WF baseplate
 connection, you can detail it as either fixed or pinned.

<Dennis>
I think you have essentially proved my point. First, we don't consider a 
shearwall as being fixed to the foundation. The evidence is that we consider, 
nail slippage, and elongation of the tension anchor in the calculation for 
deflection. These are acknowledgments that the connection is not rigid and 
that there will be rotation caused by bending and elongation of the 
restraints. We need not do this if the monolithic material is embedded in 
concrete which none of these represents.
Secondly, the connection of two legs of the diaprhagm are not secured to a 
rigid material such as a foundation. Consider for the moment that the framing 
is parallel to the shearwall on the long leg and perpendicular (bearing on 
the wall) in the short block. You can block and drag, but this will not 
prevent rotation of the joint - especially if the framing does not come 
converge at the same elevation or the plywood panels butt rather than lap the 
line of convergence.
Engineers are very particular about cautioning the community against trying 
to make a rigid connection with wood - Trusses are all pinned and the 
analysis of a continuous chord is still considered pinned by analysis.

I am not convinced that the legs won't deform as the diaphragm shear is 
applied, thus causing the actual location of the center or rotation to be 
only an approximation and not one that is sufficiently accurate to waste the 
time calculating.

<Jake>
 
 In response to 4. - I have seen people increase the shear by
 approximately 10% at shear walls to accommodate the 5% CM offset.  If
 you solve a simple span beam, and place the center of mass off 5%, you
 get about 60/40 ratio for loading.  If you read the SEAOC blue book in
 regards to this, it suggests taking into account the 5% CM displacement
 regardless of diaphragm rigidity. This statement is only found in the
 commentary, and not in the code.  Take that as you will.

<Dennis>

Okay, consider this. The new code converts the base shear to allowable stress 
design and reduces the base shear by 40% (1.4). Historically, I have always 
designed conservativly with an R=6 rather than 8 (therefore a base shear of 
.187Wd rather than the new code which gives 0.14Wd.). It seems that we took a 
simple approach and made it complicated. The argument in favor of the new 
code is valid if and only if you use the results for torsion only and design 
the wall stiffenss based on whichever shear controls over the full area of 
the diaphragm. 
However, this is not what the seminar notes and examples provided by SEA and 
ICBO suggest to the community. It is suggested that the worst case shear be 
evalutated in each line of shear. Therefore, you may have a reaction caused 
by wind and the a seismic reaction in the adjacent block.  The problem is 
that this will provide an improper balance for the walls below and you will 
end up designing one line stiffer than the other, thus forcing rotation to 
occur rather than attempting to control it by balancing all of the wall 
deflections to bewithin a close tolerance.

This may mean that the intention of the code is imperfect and that we can use 
other simplified approaches to make design easier and more productive.

With that said, these features are in the spreadsheet already. They help 
simplfy the tedius math and may be of value to anyone who wants it. However, 
does it make the concept valid?

I've said it before and so far I am not convinced otherwise, that wood (a 
very special material due to it's ductility) was not adequatly researched or 
the methods applied to it's design would be more unique and not the same as 
those of concrete (rigid) monolithically formed materials.

You make some strong points, but I have thought through these and have not 
seen any supportive emperical information to prove them. If the model was 
square or rectangular with a cutout, I may tend to see some validity but 
would still expect deformation (racking) by nature of the lack of rigidity in 
the joints.

Impressive conversation Jake - I think you are ready for your PE:>)

Regards,
Dennis


<Jake>
 
 Jake Watson, E.I.T.
 Salt Lake City, UT >>