Seismic Vertical Bracing
Seismic Vertical Bracing
Seismic Vertical Bracing
William A. Thornton
Cives Engineering Corporation, Roswell, GA, USA
bthornton@cives.com
Larry S. Muir
Cives Engineering Corporation, Roswell, GA, USA
lmuir@cives.com
ABSTRACT
The effect of frame distortions have been routinely neglected in the design of bracing
connections. It is coming to be realized that, because of the large story drifts that
occur during seismic events, that this practice may not be adequate to provide a
structure that can survive an earthquake without collapse. This paper is abstracted
from a forthcoming AISC design guide on vertical bracing connections.
INTRODUCTION
It is coming to be realized for high seismic applications where story drifts of 2-2%
must be accommodated, frame distortion cannot be ignored. These story drifts of 22 % are on the order of ten times the drifts that are expected for wind and low
seismic (R3) design. They occur in part because the actual maximum considered
earthquake (MCE) forces are reduced to about 1/9 of the forces the MCE could
produce. This is done by first using 2/3 of the MCE forces and then dividing them by an
R factor on the order of 6, so the MCE load reduction factor is 6x3/2=9.
The rationale for this reduction factor is twofold: (1) the forces are of short duration
and are reversing, so the response to them does not necessarily achieve the
maximum values, and (2) to allow economical designs to be achieved. The price paid
for this MCE force reduction is the high drift, and the requirement for ductile response
that allows large distortions without fracture and resulting building collapse. If one
used an R of 1, or even 2/3, the drift under even the MCE forces would be no greater
(and probably less because of the duration factor) than traditional wind design. Some
designers of hospitals (Walters et al, 2004) and nuclear power plants do just this.
The current AISC Seismic Provisions (2005) have no requirement to consider frame
distortions and the resulting distortional forces.
DISTORTIONAL FORCES
These forces exist because a braced frame, although considered a pinned structure, is
in reality a braced rigid frame. They would be reduced to essentially zero by the use of
an actual pin as shown in Fig. 1, or they can be controlled by the use of a designed
hinge in the beam as shown in Fig. 2. If no pin or hinge is used, the maximum
distortional forces can be derived from the maximum distortional moment,
In this formula, the column is considered continuous above and below the location
being considered. Fig. 3 shows a statically admissible distortional forces distribution.
These forces are to be added algebraically to those resulting from the Uniform Force
Method (AISC 2005) of bracing connection analysis.
Note that, when the brace force is tension, the distortional forces FD are compression.
These forces tend to pinch the gusset and can cause the gusset to buckle even when
the brace is in tension. This gusset pinching has been observed in physical tests
(Lopez et al, 2004).
AN EXAMPLE
Figure 4 shows a connection designed to satisfy the current Seismic Provisions (AISC,
2005). This design, which does not consider distortional forces, is given in the Design
Guide (AISC, 2008). The statically admissible interface forces for the connection of
Fig. 4 are given in Fig. 5. These forces would be correct if a beam hinge such as
shown in Figs. 1 or 2, were used. However, with no hinge as shown in Fig. 4, the
maximum possible (demand) distortional moment is
= min{1.1(826 ),2(1.1)2260}
= 909kip ft
From the geometry of Figs. 3 and 4,
2
MD
909 14.5
=
FD =
= 609kips
+ eb 14.5 + 8.5 18
HD =
x609 = 474kips
This value, which is compression when the brace force is tension, can be compared to
the 176 kip horizontal force of Fig. 5 between the gusset and the column, which is
tension when the brace force is tension. It can be seen that it is not reasonable to
neglect the distortional forces.
Note that the large distortional forces may not be able to be achieved because of
column and beam web yielding and crippling, and gusset pinching (buckling when the
brace is in tension). The Design Guide (AISC, 2008) proposes using the plate
buckling theory given in the Manual pages 9-8 and 9-9 (AISC, 2005) to control gusset
pinching. The Manual formulations can be written as
Fcr = QFy
Q = 1.0 for 0.7 (yielding)
Q = 1.34 0.486 for 0.7 < 1.41 (inelastic buckling)
Q=
1.30
b
Fy
t
=
1120
5 475 +
2
a
b
where
a = length of free edgedistance between points A and B of Fig. 4.
b = the perpendicular distance from the free edge to the gusset junction point at
the beam and column, point C of Fig. 4.
a
b
= 2.09,
= 28.3
b
t
28.3 50
= 1.48
1120
5 475 +
2.092
Q=
1.30
= 0.594
1.482
fa =
609
= 38.3ksi
0.75 x 21.2
Since 38.3ksi > 26.5ksi , the gusset will buckle in the pinching mode when the brace is
in tension. This buckling will prevent the distortional moment M b = 909k ft from
being achieved, but this out-of-plane buckling is undesirable because it could cause
low cycle fatigue cracks to form in the gusset and its connections.
CONTROL OF DISTORTIONAL FORCES WITH A BEAM HINGE
The idea is shown in Figs. 1 and 2, and has been tested in the context of buckling
restrained braced frames (Fahnestock et al, 2006). A completely designed example
with a beam hinge is shown in Fig. 6. The loads and geometry are the same as the
example of Fig. 4. The Design Guide (AISC, 2008) gives complete calculations for this
example. Because of the beam hinge, the distortional force FD is reduced to 204 kips.
The design shown in Fig. 6 satisfies all the usual limit states, plus gusset pinching, with
the original 3 in. gusset plate.
4
SUMMARY
A forthcoming AISC Design Guide (AISC, 2008) on Vertical Bracing Connections
treats many types of bracing connections and loadings. This paper, which is
abstracted from the Design Guide, presents a rational state of the art treatment of the
distortional forces induced by large seismic drifts.
REFERENCES
AISC (2008), Design Guide for Vertical Bracing Connections, W.A. Thornton and L.S.
Muir, to appear.
AISC (2005), Manual of Steel Construction, 13th Edition, AISC, Chicago, IL.
AISC (2005), Seismic Provisions for Structural Steel Buildings, AISC, Chicago, IL.
Fahnestock, Larry A., Ricles, James M., and Sause, Richard, (2006) Design,
Analysis, and Testing of an EarthquakeResistance Buckling-Restrained Braced
Frame, SEAOC 75th Annual Proceedings.
Lopez, Walterio A., Gwie, David S., Lauck, Thomas W., and Saunders, Mark (2004),
Structural Design and Experimental Verification of a BucklingRestrained Braced
Frame System, AISC Engineering Journal, Fourth Quarter, p. 177-186.
Walters, Mason T., Maxwell, Benjamin H., and Berkowitz, Russell H. (2004), Design
for Improved Performance of BucklingRestrained Braced Frames, SEAOC Annual
Convention Proceedings.
"
12
t
slo
W14x283
3
4"
TAB
(Gr50)
of
th
g
1"
l en
0
1'- 2
869k
HSS8x8x12
7 @ 3" = 1'-9"
3'-
30
3
8
3
8
2414
2414
5
16
7"
8
12
8"
9"
PL4x1(A36)
(BS)
3
4"
30
1
2
1
2
3514
3514
5
16
5
16
W16x100
Shear = 70k
114" TAB
(Gr50)
WP
6"
212"
9
16
9
16
12
12
19
19
WEB
ONLY
BOLTS: 118"A490-X
HOLES STD 1163 "
GUSSET(Gr50)
"
12
lot
fs
o
gth 1 "
len
0
1'- 2
W14x283
869k
HSS8x8x12
30
7
16
7
16
29
29
5
16
30
9
16
9
16
12
3 3
4"x4"
SNIP
3
16
5
16
5
1" 16
2
19
19
912"
@
DOUBLERS
3514
3514
3
4"
WP
2"
5
16
5
16
1314
1314
3@3"
1
2" 22"
WEB
ONLY
GUSSET(Gr50)
(2) SPL PL 143"x1314"x32"
(Gr50)(ClassA Faying Surfaces)
3@3"
1'-112"
7"
3'-
PL4x1(A36)
(BS)
212" 2"
2"
3@3"
2"
W16x100
Shear = 70k
DOUBLER
1
1
1
4"x132"x152" (Gr50)
(B/S of WEB)
(B/S of SPL)
BOLTS: 181"A490-X
HOLES STD 1163 "