I appreciate Mr. Sabelli comments, and would like to add a few additional
comments/questions. Comments/questions inserted within the original message.
<< I The "R" factor for SCBFs corresponds to a brace yielding mechanism,
not to collector failure. Analysis may indicate that not every connection
must resist brace yielding when the yield mechanism forms.>>>
I assume this would be more justified when doing a non-linear analysis. When
doing a linear three dimensional analysis I assume you would want to design
all braces for yielding because of the uncertainty as to the exact brace
yield sequence and redistribution of forces as each brace yields.
<<< II The brace-to-brace connection must also be designed for these forces.
III To limit the connection demand there are a few things that can be done
with the overall frame design to permit using a smaller brace:
A Lower-strength steel should be considered for slender braces. >>
The use of lower strength steel seems like a good concept, but I imagine the
availability would be very limited, especially if you are specifying A36
steel. I imagine that most steels will probably have actual yields in the 50
ksi range, if not higher, due to recycled steel being used in current
domestic steel production.
The use of pipes, I believe, requires that the gusset plate connection be
designed for out-of-plane bending since the pipe will most likely always
buckle out-of-plane unless the gusset plate is reinforced to force the pipe
to buckle in-plane.
<< C Flat-wall width not overall width, should be used when following the
I would agree with you on that. But, for those of us designing with the 1997
UBC, I am afraid we are stuck using the out-to-out dimensions unless we can
convince the building official otherwise.
<< D Full flexural and axial continuity through the brace-to-brace
connection can be provided; this will reduce the effective length to below
the half length. Be aware that for out-of-plane buckling this may induce
significant torsion in the tension brace.>>
Excellent point about the torsion.
<< E If possible, the ends of the brace should be fixed. If this can't be
done, measure the brace length from the line of restraint at the hinge
zone, not from the column centerline.>>
What types of connection would you recommend to obtain complete fixity at the
end of the brace. Are you suggesting to use a connection other than gusset
plates. This would be helpful in eliminating the need for the 2t offset
requirements on the out-of-plane buckling of the gusset plate. I assume
there could be a possible problems with local wall buckling of tubes and
pipes at the point of fixity.
For the design of the brace element, I would still use the face of column to
face of column dimension for sizing of the brace. I am not comfortable using
the clear distance between gusset plates to size the unbraced length of the
brace. To design the gusset plates, I would use the clear length between
gusset plates. This is only really applicable when (Omega x earthquake
forces) govern (are used for) gusset plate design (UBC: SCBF, OCBF; AISC:
OCBF). In cases where the connection is designed based upon tension
yielding (AISC: SCBF), then the actual length of the brace is not necessary.
<<< John, if any of this is unclear or if you would like references for the
above, please feel free to contact me directly. Good luck,