Need a book? Engineering books recommendations...

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

Re: SCBF question

[Subject Prev][Subject Next][Thread Prev][Thread Next]
In a message dated 12/20/2000 11:11:38 AM Pacific Standard Time, 
Markp(--nospam--at) writes:

<< Section 13.4.a.4 of the 1997 AISC Seismic Provisions Manual states:
 "The top and bottom flanges of the beam at the point of intersection
 of braces shall be designed to support a lateral force that is equal
 to 2 percent of the nominal beam flange strength Fy(bf)(tbf)"
 If the beam cannot be directly laterally supported at this location
 is it acceptable to design the beam using the special loading
 combinations of chapter 16 (97 UBC) along with a lateral force
 equal to Fy(bf)(tbf) in the weak direction at each flange and
 check the member for the combined biaxial bending and compression
 resulting?  A plan checker at DSA suggested using 10% of the axial
 force in the beam resulting from the special loading combinations.
 Mark Pemberton, P.E..

The condition I imagine you have is a brace frame beam along a floor opening 
at the beam midspan so you can't physically place a brace at this location in 
the out-of-plane direction.  

I am not sure if you will have a problem making the beam work strength wise 
for the condition you described using just the special load cases, but the 
overall stiffness of the brace is another issue. 

For strength requirements, both the AISC provisions and the UBC require that 
you check the SCBF beam for an unbalanced force when the compression brace 
buckles and the tension brace pulls down on the beam.  The resulting vertical 
force component of the braces is very large since you are using (FyAgRy - 0.3 
phi Pn) for the brace axial forces and I imagine the beam will weigh in about 
200 lbs/ft to resist this vertical force at the beam midspan in the plane of 
the beam.  If you are checking a W24x192 beam (Fy=50 ksi), in the strong axis 
you have an unbraced length "Lu " of 24.7 feet where maximum allowable design 
stresses are limited to  0.6Fy max and you have a reasonable ry value (ry = 
3.07) for out-of-plane buckling of the beam as a column .   

If you have braces that have a small cross section area, then maybe the beam 
will be smaller, but the tensile force for a HSS 5x5x3/8 tube is AgFyRy = 
(6.58)(46)(1.3) = 393 kips and the vertical component, assumming 45 degree 
brace slope is 0.707 x 393 = 278 kips which is a huge force, which will still 
likely be above 200 kips after you subtract out for the compresson brace.

I would not apply a force in the out-plane direction to the beam flanges (I 
assume it is a typo you have to use Fy(bf)(tbf) for the out-of-plane force 
which would be huge).  When you check the interaction on the beam using the 
ramped up seismic forces by Omega for both the brace to cause bending in the 
beam and axial to cause compression in the beam I would think it would still 
work since you orginally sized the beam for the unbalanced brace buckling 

If it is a chevron brace, isn't half the beam length in tension on one side 
of the brace connection point to the beam midspan and the other half in 
compression, which is another reason for the out-of-plane brace of the beam.  
 If you include an out-of-plane force, I wouldn't use more than 2% of the 
beam flange force, which may cause you to use a larger beam size since now 
you have biaxial bending.   If you are having to check the beam for these 
forces, what does your beam to column connection look like with such huge 
forces resulting from the beam/brace midspan connection.

If you have a two story x brace, then one leg of the two story brace is going 
help control the amount of out-of-plane buckling when the other two story leg 
starts to buckle in compression.

Considering overall brace frame stiffness, if you have two identical brace 
frames, where one is braced at the beam midspan, and the other is not, then 
they will have different stiffnesses when loaded to brace buckling.  I 
imagine the brace frame without the midspan out-of-plane brace will start to 
buckle before the other brace frame does since there is limited torsional 
stiffness at the beam midspan point, thereby shifting the force to the brace 
frame with the midspan brace point that provides torsional stiffness.

If probably would take a non-linear analysis to determine the difference in 
overall brace frame stiffness.  In your computer model, you might want to try 
eliminating the beam on this particular brace (assuming it failed at this 
level) and see where the loads go the other brace frames if they still work 
and see if you can use this as a rational approach.  You might have to upsize 
the columns on this brace frame.

Curious to know more about this, or how you resolve it.

Michael Cochran S.E.