Need a book? Engineering books recommendations...

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

Wood - Distribution of Shear

[Subject Prev][Subject Next][Thread Prev][Thread Next]
I just completed the second design based upon the new code. Both were custom 
homes - the first was fairly retangular in shape (although some 
irregularities creating three area's of varying roof elevations) and the 
second a complicated "U" shaped structure.
Both were considered under the '97 UBC code provisions to be a "Monolithic" 
type roof - not allowing for changes in elevation, slope or types of framing 
like shed roof adjacent to gabled roofs. 

In the original design, there was only a minor difference between the 
flexible and torsional (rigid) analysis that resulted in the change of 
nailing on one wall. In almost all other wall conditions, the reserve 
capacity in the plywood sheathing was sufficient to compensate for the 
increase in shear due to torsion.

The "U" shaped structure offered some very noticible differences - with some 
walls differing as much as 1000% (ten times the flexible results). Here are 
some assumptions made in the course of the design:

1. Wind and Seismic reactions were mixed in the flexible analysis. The worst 
condition was calculated for each line of shear and the resisting walls were 
designed accordingly.
2. The total accumulated reactions in each direction (representing wind and 
seismic) were applied laterally in the rigid analysis in each direction. I 
did this in the first design example and found that there was not a 
significant difference in applied loads (approximately 10 to 15% difference). 
I did this to allow for a more consistant comparision of distribution of 
3. The center of gravity of the "U" structure was calculated by AutoCad using 
an addon software and the distance to the assumed origin deducted from the 
origin used to define the resisting walls in the rigid analysis. This insures 
consistancy between the location of applied load and the true center of mass.
4. The final distribution of shear was based upon the worst case results from 
the rigid and flexible analysis.

The final design no longer represents a true model based upon applied loads 
from either wind or seismic, flexible or rigid distribution. Instead, it 
becomes the cumulative worst case of all combinations of loading - what I 
consider to be a severly overly conservative result. 
The reason for rigid analysis is to balance the distribution of shear through 
the diaphragm based upon the stiffness of resisting walls. HOwever, what we 
have done is to change the model and thus change the possible distribution by 
changing the stiffness in each line of shear to represent, not the load 
applied, but the possibility for a combined event which is unlikely to occur.
The question is - who pays for this? Is it necessary to be so overly 
conservative when the performance can not be predicted? Are we creating 
another problem by forcing distribution through the diaphragm to act, not as 
calculated by rigid analysis, but by increasing the stiffness of arbitrary 
lines of shear? How would we work backwards and determine the true 
distribution based upon the artificial stiffness created by combining models?

I am anything but comfortable with the results after doing a few homes. The 
greatest differences occur with "C" or "U" shaped structures. I don't believe 
that this is a realistic design procedure on these types of buildings as we 
ignore the slope of the roof and the assume rigidity where the legs of the 
structure join. 

I intend to bring design examples to the Panel discussion at the SEAOC 
Convention to discuss this issue. I would appreciate any opinions or 
suggestions from the virtual community on what they are noticing in design.

Dennis S. Wish PE