Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
welded headed
stud connections
Richard A. Spencer
Assistant Professor
Department of Civil Engineering
University of British Columbia
Vancouver, Canada
Donald S. Neille
Graduate Research Fellow
Department of Civil Engineering
University of British Columbia
Vancouver, Canada
70
shear forces between precast units, and
in some cases also resist pullout forces Synopsis
due to out-of-plane inertia forces on
the units themselves. Simulated earthquake
The accelerations in a severe earth- loading was applied to a
quake may be much greater than those type of welded headed stud
assumed in the equivalent static lat- connection used in precast
eral load analysis. It is generally ac-
cepted that while a building should be box-type buildings.
able to resist moderate earthquakes
Load-deflection hysteresis
with only non-structural damage, yield-
ing and associated structural damage is loops were obtained, and the
acceptable in a severe earthquake. The actual strength of each
structure should be able to undergo connection under cyclic
large deformations without collapse. loading was compared with
The connections between units in the static ultimate design
precast box-type buildings are seldom st ength.
capable of transferring loads which
would cause the precast units to yield. Although the design strength
Therefore, the connections themselves
was exceeded by up to
should be able to yield and sustain large
deflections when loaded to their design 30 percent in the initial
ultimate load and beyond in an earth- cycles, the strength began
quake. to fall as the number of load
The tests described here were made cycles increased, and was
on a commonly used welded headed generally well below the
stud connection. There have been sev-
design strength at failure.
eral static or monotonic tests on con-
nections of this type2' 3' 4 and design Deflections at failure were
data are available for calculating their from seven to twenty-four
static shear and pullout capacity.5'7
times the theoretical elastic
However, there is little or no informa-
tion on their behavior under cyclic or deflection at the design
alternating loads such as those occur- ultimate strength, but
ring during an earthquake. because the connections
lose both strength and
stiffness before failure, their
Welded Headed
Stud Connection ductility does not justify any
reduction in the lateral
The type of connection considered here forces used for earthquake-
is shown in Fig. 1. Short lengths of resistant design.
angle, each with two headed studs
welded to it by the fusion process, are If the connections are
embedded in the edges of each pre- properly designed and
cast unit.
detailed they should perform
A piece of bar or rod is site welded
between two adjacent angles to make satisfactorily in buildings
a connection. The number of connec- designed as box systems.
tions required along each edge is gov-
erred by the design shear and pullout of the connection angle, producing a
forces. moment which is resisted by a com-
pressive force C and a pullout force P,
assumed to act through the studs as
Method of Testing shown. These forces are equal to 10
and 15 percent of F for 12 and 10-in.
The connections were loaded in the
long connection angles, respectively.
test rig shown in Fig. 2. A 2 x 4-ft
This system of forces is not identical
panel with an embedded connection
to the forces resulting from the story
angle was clamped in a vertical posi-
shears and panel inertia forces acting
tion. The cyclic shear force F was ap-
in a box-type building during a severe
plied to this connection angle as shown
earthquake. Further, when units are
in Fig. 3. connected together in a building, each
A 5 x 12 x 5/s-in, steel plate was
connection angle is restrained from ro-
welded to the angle with a Y4-in. spat-
tation as long as adjacent panel edges
er. A displacement controlled hydraulic
remain parallel. However, it is felt that
jack with its axis vertical applied the this test method is sufficiently realistic
cyclic force to the steel plate. The to provide useful data.
loading system did not prevent either
pullout of the studs or rotation of the
connection about a horizontal axis. Connections Tested
The system of forces acting on a
connection is shown in Fig. 3. The Details of the six connection angles
cyclic force F acts 7/8 in. from the face tested are shown in Fig. 4. Each angle
72
has a $/s-in. diameter, 6-in, headed p 4/3 jJ 4/8
stud welded 2-in, from each end. The
studs had a specified ultimate tensile
P uy] ± L'-
v
<1
Design strength
The static ultimate design strength of
each connection was analyzed using
the procedure recommended by the
PCI Connections Manual and PCI De-
sign Handbook.5,7
The ultimate concrete shear strength
for each stud, V'u,, based on the shear-
friction concept, was found to be 14.1
kips. The ultimate pullout strength,
calculated using the area of the
45-deg shear cone, reduced because the
full cone intersects with the surfaces
of the panel, was 29.9 kips for the
Series A connections and 17.5 kips for
the Series B. The ultimate shear ca-
pacity of a stud exclusive of the con-
crete, V', was 13.8 kips, and the ulti-
mate tension capacity exclusive of con-
crete, P'u8, was 16.6 kips.
The design strength of each connec-
tion was then calculated assuming that
the applied shear force F is divided
equally between the two studs, and
that the pullout capacity of the stud-
carrying tension is not affected by the
adjacent stud, which carries compres-
sion. Combined shear, V,,, , and tension,
on a single stud was assumed to
govern, and applied using the PCI in- Fig. 2. Test rig for cyclic loading.
teraction equations. The loading yoke applies cyclic
For concrete capacity: loads directly to the connection.
ro F
.0 0:^..^ b.®CS
vP
C SPACER I i
TRANSVERS E AXIS
OF LOADING YOKE
P
FACEPLATE I^
1k 5^8
^^8 5 N
74
DETAILS OF STUDS &
CONNECTION
PANEL REINFORCEMENT
Al
A2.A3
aao
x3z3l8 L x1 "LONG
B1 °Q ^^^
3 x3 x /Lx12 LONG
oo'o & Q
B2
3 x2^x3%8Lx10`LONG ^
-0aaao
B3 9. n.:
0
0
J
DEFLECTION (in.)
Fig. 5. Load-deflection curve (Connection Al) for
monotonic loading to failure.
76
Fig. 7. Load-deflection loops (Connection A3). Note the relatively large
deflections reached before failure.
to 34 show that the loading was just tained for cycles with constant jack
above the limit for stable behavior, un- displacement allow approximate sta-
til Cycle 35. Approximate stability limit bility Iimit curves to '' be established,
curves are shown in Fig. 6 for this similar to those for Connection A2.
connection. Using the hysteresis loops for the
Degradation of stiffness and increas- cycles when the jack displacement was
ing loop width would be expected if increased (e.g., Cycle 14, Fig. 7), ap-
the load-deflection curve crossed these proximate yield strength envelopes can
curves. Loop behavior would be stable also be constructed. These show a con-
if load reversal occurred before the tinuing loss of strength with an in-
stability limit was crossed. creasing number of cycles of load, and
are quite different from the monotonic
Cyclic loading up to yield load-deflection curve. The maximum
•Connections A3 and B2 were loaded load in each direction in the cycle
up to yield after the initial cycles in preceding failure is shown in Table 1.
the elastic range, and then subjected Connection B1 was also loaded up
to steadily increasing deflections (see to yield, but there was a much bigger
Figs. 7 and 9). After each increase, increase in deflection in each succes-
the displacement amplitude of the jack sive cycle than for Connections A3 or
was held constant for no more than B2 (see Fig. 8). The strength of this
-three or four cycles (e.g., Cycles 10, connection fell for loading in one di-
11, 12, and 13, Fig. 7). The loops ob- rection, but not the other, and it failed
* These values are for Cycle 4. Failure occurred in Cycle 7 after two cycles
at reduced load.
at 32 kips, a load substantially in ex- tests are now being carried out so that
cess of the ultimate design strength more precise predictions can be made.
(although the strength in the other di-
rection had fallen below the design Ductility
strength). The maximum deflection in each di-
From Figs. 7, 8, and 9 it seems clear rection in the cycle prior to failure,
that increasing the number of cycles X,,,^, is given in Table 2. It is clear
above the stability limit will cause that these connections have a consider-
an increasing reduction in the yield able capacity to deflect beyond their
strength, and that the yield strength deflection at yield, and in that sense
envelope for each connection will tend they might be called "ductile." Their
to approach the stability limit as a use would allow a box-type building
lower limit. However, the data from to sustain large, permanent deforma-
these tests are limited, and further tions without collapsing.
Table 2. Summary of deflection data and mode of failure for various types
of connections (also refer to Fig. 4).
Theoretical Maximum deflection (X µQ Mode of Cycles
max ) failure at failure
Connection deflection in cycle before failure (Ratio of
Xmax to XX)
Xt Up Down
(in.) (in.) (in.)
78
7
30
EFLECTION (in.)
-10
-3 0
i ii
1
1AFAPiLU
t
0.3 0.2
^/^
iUifiiiiii
■ !Ilf.hIIII
LII11(I1
.. - .
—YIELD ENVELOPE ''' '
80
There must also have been crushing
and failure of the concrete around the
studs as the test progressed. Further-
more, it is likely that some of this con-
crete simply fell out during each cycle,
allowing the studs to deform, and ex-
plaining in part both the steadily
falling stiffness and the hardening
spring behavior observed in every
cyclic test.
As the applied deflection was in-
creased, the tension cracks in the panel
increased in length and width, and the
connection began to rotate slightly as
first one stud and then the other began
to pull out. The connections in a build-
ing would probably be restrained
against this rotation, but they might
be subject to pullout forces as a result
of inertia loads on the panels.
Connections Al and B1 failed by
concrete failure when a large block of
concrete surrounding the studs broke
away (see Fig. 11). The other connec-
tions (except Connection B3) failed Fig. 11. Connection B1 failed when
when one of the studs broke close to a large block of concrete surround-
the fusion weld (see Fig. 12). ing the studs broke away.
82
V' ultimate concrete shear capac-
Notation ity of a stud
V' ultimate shear capacity of a
C = compressive force in a stud stud exclusive of concrete
subject to test load strength
F = test load applied to connec- V = ultimate shear strength of a
tions stud subject to combined ten-
P = pullout force on a stud sub- sion and shear
ject to test load XZ = theoretical deflection of con-
Pu = ultimate pullout strength of a nection at design ultimate
stud subject to combined ten- strength assuming constant
sion and shear stiffness equal to initial stiff-
P' = ultimate pullout strength of a ness
stud controlled by the concrete Xm = maximum deflection of con-
P' ultimate tension capacity of a nection in cycle preceding
stud exclusive of concrete failure
strength [LI, = ratio of Xma, to Xa