From: Rafael Sabelli <sabelli(--nospam--at)dasse.com>
Date: Tue, 13 Apr 1999 10:15:40 -0700
I would like to thank Mr. Cochran for the light he has shed on the subject
and for the attention he has given to my comments. In response to the
issues that he raised (following my original outline):
I A nonlinear analysis can demonstrate that certain braces will never
reach tensile yielding; this is occurs most often with special
configurations. For example, we often design alternating stories with "V"
and inverted-"V" braces, making a two-story "X", which is designed to
produce a two-story mechanism. If the building has an odd number of
stories, the leftover inverted "V" at the top story is designed to remain
elastic under a conservative range of force distributions. Another example
is the "Zipper" configuration, proposed by Khatib, Mahin, and Pister, and
described in the 1997 AISC Seismic Commentary. Of course, if that behavior
is not designed into the system, analysis will probably not demonstrate it.
II No issues raised.
III A Lower-strength steel may be available for pipes and tubes.
B For X-braced frames using pipes or square tubes and a single
gusset connection (with a hinge zone), buckling will occur in the direction
that the hinge zone allows. That is, the gusset plates may be oriented in
the plane of the frame, allowing out-of-plane rotation, or perpendicular to
the plane of the frame, permitting in-plane rotation (and preventing the
torsional effect mentioned in III D) ("Design of Cross-Braced Frames for
Predictable Buckling Behavior," Sabelli and Hohbach, ASCE Jour. of Struct.
Eng., Feb. 1999; I wasn't going to plug it but the issue was raised).
C Yes, the UBC is more conservative in this regard. The building
official may consider the fact that local buckling is most severe for
braces of low slenderness.
D No issues raised.
E I agree. Except at the base, where braces can be buried in
concrete, I have not been able to design a fixed-end brace connection for
large braces that was economical (compared to larger braces, connections,
The hinge-zone to hinge-zone length (for single diagonals) corresponds to
the actual buckling behavior of the brace and is therefore justified, I
believe. For crossed braces determining the effective length is more
complicated and depends on the center connection detail; for full flexural
and axial continuity at the center, one half the hinge-zone to hinge-zone
distance can be used for brace design. This is the lower bound; the upper
bound depends on the rotational restraint provided by the brace in tension,
and may be significantly higher. The compression capacity of the brace may
require that the gussets be stiffened to prevent their buckling ("Seismic
Behavior and Design of Gusset Plates," Astaneh-Asl, Steel Tips).