At 03:14 PM 1/21/00 -0500, you wrote:
>Regarding a gable roof, Charles Greenlaw wrote:
>> ... the fact of the roof surfaces being pitched (rather than being
>> essentially parallel to idealized horizontal loading) is not a major
>> complication. The whole matter should reveal its "mysteries" if
>> subjected to some free-body diagram modeling, complete with chasing
>> down of all the resulting boundary reactions with due attention to
>> principles of statics and strength of materials.
Thanks for the reply Charles.
I have not had time to respond to Charles Greenlaw's response to my
question but the last post about this topic reminded me. Until I did an
analysis of this I kept thinking about...for instance a church roof where
the slope of the roof is about 60 degrees from horizontal. The majority of
the horizontal force seems to go into bending the roof deck in the weak
axis. Free body diagrams suggest that the horizontal force is broken down
into a shear in the plane of the diaphragm and a vertical load on the
supporting beams or walls.
I did an analysis of a roof with gable steel trusses 16m long at 1.5m (5')
o.c. over a length of 30.5m. I covered the roof slopes with a FE mesh and
put a horizontal load on the 30.5m long eave at on side. The deck was
modelled as a 22ga flat steel sheet as I did not how to model the deck
otherwise. The tension/comp'n members at the eaves were support beams of
w14x26. The horizontal deflection was 1.25mm. The corresponding tension in
the eave member was 73% of what a simply supported beam would be. The deck
took the other 27%??? The bending of the support beams was 79% of what the
free-body diagram said it was going to be. The shear pattern of the roof
deck was very interesting to see as well. The discontinuity at the ridge
was interesting and was not a problem.
It would seem that the hip roof would fair as well.
We will provide a bent plate over the tops of the ridges under the deck for
deck edge support and shear continuity.
BTW I did the extensive analysis to "make sure" that my hand calcs made
sense and to experiment with Robot97. It did a pretty good job dealing with
the large number of elements in my model. 2700 triangular elements with 22
trusses. If anybody has anything negative to say about this software I
would be interested in hearing about them. There a lot of positives. I have
found some errors with column strength checks with regards to class 4
sections (Canadian code).
Customer support has been excellent so far.
David Handy, P.Eng.
The Thompson Rosemount Group, Cornwall, Ont. Canada
Opinions expressed are personal only.