Comments are included in text:
< << John.A.Connor(--nospam--at)nwk02.usace.army.mil writes:
NEHRP directs me that my seismic requirements shall be in accordance with
AISC's Seismic Provisions for Structural Steel Buildings. Based upon the
significance of this building, I have decided to design the steel framing to
meet the requirements of AISC's chapter 13, "Special Concentrically Braced
The braced bays of this building consists of x-braces. The bracing members
are approximately 50' in length. Assuming I use a steel tube shape, chapter
13's requirements for slenderness and also limiting width-thickness ratios
leads me to select a TS 10x6x5/8.>>>
Try and use the TS 10x6x5/8 brace about the weak axis (6" deep) so buckling
will occur in-plane instead of out-of-plane. Now the architect/owner may not
want a wall thickness of at least 10 inches in order to conceal the tube in
the wall. If you place the brace tube about the strong axis (10" deep), then
buckling will occur otu-of -plane.
<<< I used this size as my preliminary member size in my STAAD model. After
running all of the different load cases, the output for this member is about
60 kips for both compression and tension. The "controlling" load case for
this output was from the wind loads.>>>
Would wind still govern if you were using an ordinary CBF with a lower R
value instead of a SCBF. In Kansas, I imagine wind would still govern.
<<<Chapter 13 says that the required strength of my bracing connection shall
the LESSER of the following:
a) "The nominal axial tensile strength of the bracing member, determined as
For this tube: (1.1)*46*16.4=830 kips.>>>
I would recommend for the overstrength Ry, that you use a higher value than
the 1.1 in the AISC Provisions. Tube steel from the mill certs I have seen
has an average yield of 58 ksi instead of 46 ksi nominal, and it is very
common to have values in the 60+ ksi range. To account for an average yield
stress of 58 ksi, Ry becomes approximately 1.3
<<< b) "The maximum force, indicated by analysis, that can be transferred to
brace by the system."
I interpret this as my STAAD output which equals 60 kips.>>
See Mr. Yousefi's comments. Analysis is not always the same as what the
system is capable of transfering to the bracing elements. In most design
offices, this provision is not checked, and you would then design the
connection for the tension capacity of the brace (RyAgFy). The 1997 UBC,
allows you to design the connection for the lessor of the tensile capacity of
the brace or (Omega x Compression force in brace from analysis) for the SCBF.
The AISC specification does allow this provision also only for Ordinary CBF,
therefore, if wind still governs, it may be better to do the brace design
using Ordinary CBF. I believe this provision about (Omega x Compression
force) for SCBF will be disallowed in the the 2000 IBC.
<<<According to chapter 13, the "lesser" would be a connection designed for
Did I interpret the criteria correctly? Did I miss the "fine print"
What I don't understand is, why do I have to have such a "huge" bracing
member, when I only have 60 kips going through the member? In the event of
a seismic event, my connection would fail before the member would "buckle"
or "snap". My seismic event forces are less than my wind forces. It
doesn't seem right requiring a large brace and only have a 3-bolt
During a seismic event, you want the brace to buckle before failing the
connection, therefore the gusset plate connection has to be designed for a
load greater than the "true" buckling capacity of the member. The true
buckling capacity will be dependant upon the actual brace length and yield
strength of steel, assuming Euler buckling is not governing (which has a
factor of safety for allowable axial load of 23/12 (AISC E2-2)). Remember
that your design forces by analysis are not the same as the possible expected
sesimic forces since you reduced the building base shear by a building system
factor (R) for SCBF.
<< Also, when would condition "a" control over condition "b"? For example, if
the analysis says my maximum tension is 500 kips, I would have to select a
member for (0.9)*Fy*Ag. 0.9<1.1 so I would never have this as a controlling
case. Does the code have a mistake in wording, or is it correct?>>>
Review the AISC specifications. If the brace force by analysis is 500 kips,
then the brace must be designed for 500 kips. Since the member you select
will have a capacity greater than 500 kips to satisfy the demand of 500 kips,
the connection needs to be designed for the actual yield capacity of the
member selected. If for architectural reasons you use a brace significantly
larger than required, you might be able to justify using a lessor force for
the connection design since the building system may not be able to transfer
to the brace a force large enough to exceed the capacity of the brace.
For the connection design using ASD, you are allowed a 1.7 increase for the
buckling capacity of the member (23/12 = 1.91 > 1.7). Also a 1.7 increase
for Bolts and Welds. The gusset plate design is critical if you anticipate
buckling out-of-plane of the brace. See the AISC specifications about the 2t
offset requirements for the gusset plate yield line and the fact that the
yield line must intersect the free edge of the gusset plate (not away from
the free edge where it is welded to the beam or column as has been commonly
detailed in recent years).
Some of the connection design may be better completed using LRFD instead of
ASD, since the 1.7 commonly applied to convert ASD to strength design can
result in unconservative answers when compared to the LRFD solution for the
same connection forces.
<<<Any information or comments would be appreciated.
John Connor, EIT
Kansas City, MO
Hope this helps.
Michael Cochran S.E.