Académique Documents
Professionnel Documents
Culture Documents
f
c
f
c
b
j
the smaller of
b
b
b
c
+ ( ) 2
b
b
mh
c
2
--------- +
b
c
=
where the summation term is applied on each side of the joint
where the column edge extends beyond the beam edge, and
the slope m is 0.3 when the eccentricity between beam and
column centerlines is greater than b
c
/8; otherwise, m is 0.5.
3. To control joint deterioration during earthquake loading
and displacement demands, a specified amount of transverse
reinforcement should be provided within the joint. For a tied
column with rectangular hoops and crossties, the total cross-
sectional area in each direction of transverse reinforcement
within the joint should be at least equal to
(4)
but not less than 0.09sb
c
f
c
/f
yt
.
4. The minimum development length of longitudinal beam
bars with standard hooks anchored in the joint is determined by
(5)
while the minimum development length of headed bars is
0.75l
dh,352
.
Critical sections for the development of longitudinal beam
bars should be taken at the outside edge of the confined core
for Type 2 connections as defined in ACI 352R-02,
10
Section 4.5.1. Either bar heads or hooks should be located
within 50 mm (2 in.) from the back of the confined core,
where heads or hooks can be anchored within the diagonal
strut. In practice, the beam bars should be extended into the
joint as far as possible.
The provisions written in ACI 318-08
1
for the development of
hooked or headed bars are slightly different from ACI 352R-02.
10
Critical sections for the development of longitudinal beam
bars could be taken at beam-column faces, and the minimum
A
sh
0.3sb
c
f
c
f
yt
-----
A
g
A
ch
-------- 1
\ .
| |
=
l
dh 352 ,
f
y
d
b
5 f
c
------------- N-mm
f
y
d
b
60 f
c
---------------- lb-in.
\ .
|
| |
=
Fig. 4Details of mechanical devices for test specimens
(plan view).
332 ACI Structural Journal/May-June 2009
development length for a hooked beam bar anchored in a
joint can be determined following ACI 318-08
1
, Section 21.7.5.1
(6)
Headed bars are not mentioned in Chapter 21 of ACI 318-08
1
for special moment frames, but new provisions for the
development of headed deformed bars in tension can be
found in ACI 318-08,
1
Section 12.6. Based on available
monotonic loading tests,
5-7
ACI Committee 318 wrote that
the minimum development length for a headed bar,
measured from the critical section to the bearing face of the
head, shall be determined by
(7)
for headed deformed bars satisfying: (a) specified f
y
of 420 MPa
(60 ksi) or less; (b) bar size of 36 mm (No. 11) or smaller; (c)
normalweight concrete; (d) relative head area ratio A
brg
/A
b
of 4 or more; (e) minimum clear cover of 2d
b
for each bar;
and (f) minimum clear spacing of 4d
b
between bars.
The length l
dt,318
is longer (more conservative) than
0.75l
dh,352
, even though the effect of load reversals is not
addressed in ACI 318-08,
1
Section 12.6. In all cases, the
aforementioned development lengths shall not be less
than the larger of 8d
b
or 150 mm (6 in.).
Specimen design and reinforcing details
This experimental program was designed using a concrete
compressive strength of 30 MPa (4.35 ksi) and a specified
yield strength of 420 MPa (60.9 ksi) for reinforcing steel.
l
dh 318 ,
f
y
d
b
5.4 f
c
----------------- N-mm
f
y
d
b
65 f
c
---------------- lb-in.
\ .
|
| |
=
l
dt 318 ,
f
y
d
b
5.2 f
c
----------------- N-mm
f
y
d
b
62.5 f
c
-------------------- lb-in.
\ .
|
| |
=
Two sizes of reinforcement meeting ASTM A706 requirements
were used for longitudinal and transverse reinforcement.
Specimens W0 and W150 used traditional deformed bars
with standard 90-degree hooks within the joint, whereas the
other four specimens used screw-deformed bars with
mechanical devices for the anchorage of beam bars. Transverse
reinforcement used consistent properties for all specimens.
Table 1 shows the material properties measured at the date
of testing and the connection design parameters with respect
to the recommendations of ACI 352R-02.
10
Flexural hinging
in the beam was anticipated because the flexural strength
ratios M
r
were much greater than the specified value of 1.2.
Thus, the demand of the joint shear force V
u
was dominated
by the probable moment capacity of the beam, which had a
cross section of 300 x 450 mm (12 x 18 in.) and had four D22
(No. 7) longitudinal bars (steel ratio of 1.29%) at both the top
and bottom (Fig. 3). The relative head area ratio A
brg
/A
b
was
equal to 5.14 for the D22 (No. 7) longitudinal bars terminated
within the joints.
The values of V
u
in Table 1 were computed based on
1.25f
y
(measured) and an internal lever arm of 0.875d in the
beam. Due to the eccentric beams, the nominal joint shear
capacities of eccentric beam-column joints were smaller
than those of concentric joints. Therefore, the joint shear
demand-to-capacity ratios of eccentric joints were greater
than those of concentric joints. Notably, all joint shear
demand-to-capacity ratios were below 1.0, as listed in Table 1.
The joint capacity should be sufficient for the development
of the adjacent beam hinging.
To meet the minimum requirement of transverse reinforcement
in a joint and adjacent column end, D10 (No. 3) hoops with
seismic crossties at a spacing of 100 mm (4 in.) were
provided throughout the column (Fig. 3). As shown in Table 1,
the amount of provided transverse reinforcement was close
to the requirement given in Eq. (4) along both the N-S or
E-W directions. Thus, the joint is considered to be code-
compliant to sustain earthquake loading and inelastic
displacement reversals.
In this experimental program, both headed and hooked
bars were extended into the joints as far as possible, with a
total embedment length of 330 mm (13 in.), or 14.9d
b
was
used for all specimens, leaving a 70 mm (2.8 in.) back cover
behind the outside edge of heads or hooks. For hooked bars
in the test specimens, this embedded length of 14.9d
b
was
less than the development length of 16.7d
b
determined using
Eq. (5) plus cover thickness, but slightly greater than the
development length of 14.2d
b
determined using Eq. (6). For
the other specimens, the embedded length of a headed bar
measured from the column face to the bearing face of the
head was equal to 14.5d
b
, which was close to the development
length of 14.7d
b
per ACI 318-08
1
(Eq. (7)), but greater than
the 12.8d
b
length per ACI 352R-02.
10
All longitudinal beam bars were closely spaced at a center-
to-center spacing of 2.2d
b
, but the outer diameter of the
mechanical device was 2.48d
b
. As shown in Fig. 4, staggering
heads were used to prevent the overlapping of anchor heads.
Notably, the side cover dimension from the north face to the
edge of extreme beam bars was only 3d
b
for eccentric beam-
column joint specimens. The headed bars, however, were not
restrained by additional transverse reinforcement for side-face
blowout failure.
Table 1Material properties and connection
design parameters
Connection
specimen ID
Parameters
Conventional Mechanical anchorage
W0 W150 W0-M1 W150-M1 W0-M2 W150-M2
Concrete strength
f
c
, MPa (ksi)
29.5
(4.28)
30.7
(4.45)
35.8
(5.19)
30.7
(4.45)
35.8
(5.19)
Bar yield strength f
y
,
MPa (ksi)
455 (66) 473 (68.6)
Bar tensile strength
f
u
, MPa (ksi)
682 (99) 667 (96.7)
Flexural strength
ratio M
r
*
3.11 3.11 3.19 3.19 3.19 3.19
Shear demand-
capacity ratio V
u
/V
n
Measured yield strength of transverse reinforcement was 471 MPa (68.3 ksi).
Note: All parameters are based on measured material properties at testing date.
ACI Structural Journal/May-June 2009 333
Test setup and loading procedure
Following an earlier experimental program
16
of eccentric
beam-column joints, the same testing framework was used
for specimens with mechanical anchorage. Each T-shaped
one-way joint specimen was rotated 90 degrees to ease test
setup, as shown in Fig. 5. The column was simply supported
and anchored on the strong floor with four one-dimensional
rollers. This arrangement was used to simulate contraflection
points of zero moment as the column was subjected to lateral
loads, and also to restrain the column for twisting about the
column axis. Prior to applying lateral force, a manually
controlled axial load of 0.1A
g
f
c
was applied along the
column axis during testing. A displacement-controlled actuator
applied a lateral force at the beam centerline using a
predefined procedure consisting of three fully reversed cycles
at gradually increasing drift ratios (0.25, 0.50, 0.75, 1.0, 1.5,
2, 3, 4, 5, 6, 7, and 8%). The predetermined target displacement
at beam tip, , was computed by target drift ratio
(8)
Note that the positive drift ratio corresponds to displacing the
beam tip to the right in Fig. 5, causing a positive bending
moment at the column face.
Using the testing method recommended by ACI 374.1-05,
15
the initial drift ratio should be within an elastic range. Subsequent
drift ratios should be between 1.25 and 1.50 times the previous
drift ratio, and testing should continue to a 3.5% drift ratio. The
loading procedure in this experimental program is relatively
severe with respect to ACI 374.1-05.
15
The presented tests
continued up to an 8% drift ratio for observations of failure
modes. The performance of test specimens should be evaluated
prior to 4% drift, however, because a drift ratio of 4 or 5% is too
large for well-designed moment frames. Figure 5 also shows a
typical instrumentation installed on a test specimen. One pair of
displacement transducers was used to measure the shear
deformation on the north face of the joint, while another two
pairs of displacement transducers were used to measure the
beam plastic hinge (PH) rotation and rotation from beam bar
slip and yield penetration. Furthermore, electrical resistance
strain gauges attached to reinforcing bars at key positions and
were used to monitor the strain histories of reinforcing bars.
EVALUATION OF CYCLIC LOADING RESPONSE
Cyclic loading response and damage progression of
Specimens W0 and W150 has been presented by Lee and
Ko.
16
The performance of four specimens with mechanical
anchorage are evaluated with respect to that of Specimens
W0 or W150, which are referred to as the benchmark
specimens in this paper.
Overall load-displacement response
All six specimens developed the anticipated beam
hinging and behaved in a ductile manner consistent with the
seismic design objectives. Figure 6 shows the measured
lateral load-displacement loops and final damage patterns
for test specimens. The lateral load P was normalized to the
nominal lateral resistance P
n
determined using measured
material properties and a strain compatibility analysis for
the flexural strength of the beam section at the column face.
As shown in Fig. 6, the lateral displacement was also
converted to drift ratios by Eq. (8). Prior to the 4% drift
L
b
0.5h
c
+
------------------------- =
level, the test specimens had similar stiffness and strength
because the anticipated beam hinging dominated the load-
displacement response. After reaching maximum load at the
4% drift level or more, the load-displacement loops showed
pinching due to the beam bar slip from the joint and joint
shear distress. The response of Specimens W0-M2 and
W150-M2 shows relatively more ductility and less pinching
due to the contribution of double mechanical devices on each
beam bar.
As observed by Lee and Ko
16
for Specimens W0 and W150,
the damage progression commenced with a beam flexural
yielding followed by joint cracking, the spalling of the column
concrete cover, crushing of the joint diagonal strut, and eventual
expansion and distortion of the joint region. It can be observed
in Fig. 6 that the significant joint damage of Specimens W0 and
W150 after 5% drift corresponds to significant degradation in
load capacity from the first to third cycle of the 5% drift ratio.
The damage progressions of Specimens W0-M1 and W150-M1,
respectively, were very similar to those of W0 and W150,
except for the pushout of concrete on the east face of the
column. The failure mechanisms of Specimens W0-M1 and
W150-M1 were a complete loss of bond along the headed bars
into the joint and subsequent crushing of the joint diagonal
struts during the 5% drift cycles. As shown in Fig. 6, the load
capacity of Specimens W0-M1 and W150-M1 degraded from
the first to the third cycle of the 5% drift ratio and subsequent
loops became very pinched. Notably, Specimens W0-M1 and
W150-M1 had smaller strength degradation in the post-peak
range than Specimens W0 and W150, respectively. The use of
a single mechanical device in place of the 90-degree hook in the
joint resulted in at least an equivalent or somewhat better
performance under large displacement reversals.
The use of double mechanical devices changed the failure
mechanism and further improved the cyclic response of test
specimens. As shown in Fig. 6, the pushout of concrete on
the east face of the column did not appear at the end of
testing for Specimens W0-M2 and W150-M2. It is evident
that double mechanical devices are very effective for beam
bars subjected to cyclic loading. The most deeply embedded
mechanical device is effective for developing bars in
tension, but might not be as effective in compression because
the concrete cover behind the head is weak. The intermediate
mechanical device was found to effectively transfer the
compression of the beam bar into the joint diagonal strut.
Thus, the pushout of concrete on the east face of the column
Fig. 5Test setup and loading procedure.
ACI Structural Journal/May-June 2009 334
can be prevented. As a result, Specimen W0-M2 could fully
develop a beam PH with minimal joint damage, whereas
Specimen W150-M2 had more visible joint damage on the
east and north face of the column due to the joint eccentricity.
Nevertheless, the resistance of both specimens was maintained
up to 7 or 8% drift, which is more than enough for seismic
design of buildings.
Experimental joint shear stresses
Because all test specimens underwent beam hinging, the
measured maximum resistance P
max
exceeded their nominal
resistances P
n
, as shown in Fig. 6 and Table 2. The experimental
maximum joint shear forces V
j,max
were computed based on
P
max
by assuming an internal lever arm of 0.875d in the
beam section. When dividing V
j,max
by the shear area with
different definitions of effective joint width, as listed in
Table 2, the imposed joint shear stresses were all kept below
the limit of 1.0 MPa (12 psi) set by ACI 352R-02
10
for joints that are not confined on opposite faces.
It can be expected that the greater the imposed joint shear
stress obtained by using effective joint area, the poorer the
seismic performance of such a joint. Data in Fig. 6 show that
the cyclic loading responses of eccentric specimens were
very similar to those of concentric specimens within 4 or 5%
f
c
f
c
Table 3Test results for comparison with acceptance criteria in ACI 374.1-05
Connection
specimen ID
Parameters
Conventional Mechanical anchorage
W0 W150 W0-M1 W150-M1 W0-M2 W150-M2
Initial stiffness for first cycle of 0.25% drift,
kN/mm (kips/in.)
Positive 9.16 (52.3) 9.01 (51.4) 9.19 (52.5) 9.03 (51.5) 9.47 (54.1) 9.71 (55.4)
Negative 8.97 (51.2) 8.18 (46.7) 8.94 (51.0) 9.53 (54.4) 8.97 (51.2) 8.60 (49.1)
Percentage degradation in load capacity from
first to third cycle
of 4% drift
Positive 4% 11% 3% 4% 2% 1%
Negative 5% 20% 5% 5% 3% 2%
Relative energy dissipation ratio
*
0.31 0.22 0.33 0.29 0.32 0.35
Ratio of secant stiffness
Secant stiffness around zero drift was obtained for positive and negative loading direction between 0.4% and +0.4% drift ratios in third compete cycle of 4% drift ratio.
336 ACI Structural Journal/May-June 2009
performed as good as, or even better than, those joints with
90-degree hooks. Figure 7 implies that the use of a single
mechanical device in place of the 90-degree hook is good
enough for a drift ratio of 4%. The use of double mechanical
devices further enhanced the energy dissipation capacities
up to an unrealistic drift level of 7%. Nevertheless, this
innovative detailing of double mechanical devices can be a
possible option to improve the cyclic loading behavior of
headed bars. Further tests for improving bond performance
of beam bars passing through interior joints with double
mechanical devices are under investigation by Lee et al.
17
Fabrication of double mechanical devices is easy for screw-
deformed bars. Doubling the costs of mechanical devices
may be the only drawback.
Drift contributions
The external instrumentation of this experimental program
measured the major drift components, illustrated in Fig. 8.
The three drift components were: 1) rotation from beam bar
slip and yield penetration at the beam end; 2) rotation in the
anticipated beam PH zone within an effective depth of the
beam; and 3) shear deformation measured on the north face
of the joint. Other drift components such as elastic flexural
and shear deformation in the beam and columns were minor
and not relevant in this experimental program.
As shown in Fig. 9, rotation from beam bar slip and yield
penetration contributed most to the total drift angle for three
eccentric connections. At 4% drift, the contribution of beam
bar slip and yield penetration was approximately 40 to 45%
of the total drift angle, whereas that of beam PH rotation was
approximately 30% of the total drift angle. Beyond this drift
level for Specimens W150 and W150-M1, the percent of
beam bar slip and yield penetration increased rapidly and
that of beam PH rotation decreased simultaneously. These
readings are consistent with the damage observations on
both Specimens W150 and W150-M1. As the joint degraded
due to the crushing of joint concrete initiated at 5% drift, the
beam bars lost bond transfers for tension and compression. This
resulted in large slip deformations within the joint and extensive
pushout cracking behind the joint. Consequently, a greater
portion of the rotation was concentrated between the column
face and beam end. The joint shear deformations shown in Fig. 9
were measured on the north face (eccentric side) of the joint.
Beyond 5% drift for Specimen W150 and 6% for Specimens
W150-M1 and W150-M2, the expansion and spalling of joint
concrete made the readings inaccurate.
For Specimen W150-M2 at 5 or 6% drift, the contribution of
beam bar slip and yield penetration was below 50%, and that of
beam PH rotation remained at 30% of the total drift angle. It is
evident that the joint maintained its integrity to support the
complete formation of an adjacent beam PH up to 6% drift. It
seems that the intermediate mechanical device was effective not
only for transferring compression force of the beam bar, but also
for reducing the slip and yield penetration of the beam bar in
tension. The beam bar slip, yield penetration, and PH components
of Specimens W0, W0-M1, and W0-M2 were, respectively,
similar to those shown in Fig. 9. The shear deformation
component measured on the joint face of Specimens W0,
W0-M1, and W0-M2, however, was minimal due to the
distance between the beam and column faces.
EVALUATION OF ANCHORAGE CAPACITY
Anchorage behavior
Strains along one bottom beam bar adjacent to the north
face (eccentric side) were measured and converted to tensile
stress at peak drift ratios by assuming an elastoplastic stress-
strain relationship. Figure 10 compares the bar tensile
stresses f
s1
measured at the column face and f
s2
measured at
Fig. 7Average dissipated energy of three cycles at each
drift level.
Fig. 8Major drift components to total drift angle for test
specimens.
Fig. 9Contributions to total drift after beam yielding for
eccentric specimens.
ACI Structural Journal/May-June 2009 337
0.5h
c
along the beam bar into the column. The measured
strain readings show that all beam bars were adequately
anchored to reach the yield strength at a drift ratio of 1.5% or
more. Before the bar yielded at the column face, the strain
readings showed that 50% of the bar tensile stress was transferred
by the hook or head bearing component plus some bond of
the straight portion in the back 0.5h
c
, while the remainder
was transferred by the bond component of the straight
portion in the front 0.5h
c
. After the bar yielded at the column
face, the bond component in the front 0.5h
c
deteriorated and
vanished at the 4% drift level except for Specimens W0-M2
and W150-M2. This shows that Specimens W0 and W150 as
well as W0-M1 and W150-M1 underwent significant bond
deterioration along the straight portion of the beam bar
because of the adjacent beam PH. Beyond the 4% drift level,
the bond along the straight portion of the beam bar was
almost lost and the entire bar tensile force was transferred by
end anchorage.
The strain histories of Specimens W150-M1 and W150-M2
are compared in Fig. 11. The strain gauge was attached at 0.5h
c
along the beam bar into the column. The strain history of the beam
bar with a single mechanical device in Specimen W150-M1 was
very similar to that of the hooked bar in Specimen W150,
as reported by Lee and Ko.
16
The strain readings for
Specimen W150-M1 remained elastic in the 3% drift cycles
but went into a yielding plateau in the first cycle of 4% drift.
Due to yield penetration and bond deterioration, the shear-
resisting mechanisms changed and the joint shear capacity
decreased as the drift ratio increased. The degradation of the
joint shear capacity can be delayed if good bond is maintained
along the bonded length of beam bars. As shown in Fig. 11,
the strain readings for the beam bar with double mechanical
devices in Specimen W150-M2 remained elastic until the
gauge was damaged in the first cycle of 5% drift. It is evident
that the use of an intermediate mechanical device enhanced
the bond condition under cyclic loading and delayed the
yield penetration along the beam bars into the joint. This
enhancement resulted in Specimens W0-M2 and W150-M2,
sustaining the joint shear forces up to the 8% drift ratio with
limited joint damage (Fig. 6).
On the other hand, the side cover of the north face beam
bars was only 3d
b
for Specimens W150-M1 and W150-M2.
The anticipated side-face blowout failure, however, was not
observed during testing. The provided joint hoops and crossties
were apparently effective in restraining the headed bars.
Section 4.5.3.3 of ACI 352R-02
10
refers to prior knee joints
tested by Wallace et al.
9
and conservatively recommends
providing some U-shaped reinforcement for restraining
headed bars adjacent to a free face. Notably, there is a top
layer of beam bars adjacent to a free face in a knee joint. For
eccentric beam-column joints with a continuous column,
only one of the top or bottom beam bars is adjacent to the
eccentric joint face. The side-face blowout failure of one of
the heads does not result in the loss of anchorage of all
headed bars. Therefore, the additional restraining reinforcement
could be unnecessary.
Comparison of test results with model combining
bond and head bearing
Thompson et al.
8
reviewed numerous test results of
headed bars and proposed a refined model for determining
the anchorage capacity of headed bars by a combination of
head bearing and bond. The following equations based on
the research
4-8
at the University of Texas at Austin can be
taken to represent the state of the art of headed bar
anchorage. First, the anchorage capacity provided by head
bearing can be estimated by
(9)
f
s head ,
f
y
-------------- n
5%
2
c
b
d
b
-----
\ .
| |
A
brg
A
b
----------
f
c
f
y
----- =
Fig. 10Bar tensile stresses measured at beam-column
interface and half column depth into test joints.
Fig. 11Bar strain histories measured at half column depth into
joint: (a) Specimen W150-M1; and (b) Specimen W150-M2.
ACI Structural Journal/May-June 2009 338
where n
5%
is equal to 0.7 for the 5% fractile value for design
purposes. Without the n
5%
term, Eq. (9) estimates the mean
capacity of head bearing.
For the mechanical devices used in Fig. 4, the relative head
area ratio A
brg
/A
b
is equal to 5.14 and the minimum cover
dimension c
b
is equal to 1/2 the center-to-center spacing of
bars being developed. The radial disturbance factor is a
modification factor for stress disturbance caused by the
proximity of a bar to a corner. Thompson et al.
8
recommended
that the factor equals 0.6 + 0.4(c
2
/c
b
) but is not greater
than 2.0, where c
2
is the minimum cover dimension
measured in the direction orthogonal to c
b
. For connections
with a continuous column, as shown in Fig. 3, the c
2
value is
much greater than c
b
. Thus, the factor can be taken as 2.0.
Substituting forgoing parameters except the n
5%
term
into Eq. (9), the predicted head bearing capacity of a
headed bar was 0.65f
y
and 0.76f
y
for Specimens W0-M1
and W150-M1, respectively.
Secondly, Thompson et al.
8
proposed that the bar stress
contributed from bond be calculated using the ratio of available
bonded length l
a
to development length l
d
for straight
deformed bars in tension, assuming that the distribution of
bar stress is linear over l
d
. Thompson et al.
8
observed a
certain degree of bond deterioration when the head bearing
component achieved peak capacity in their tests of headed
bars. Thus, a reduction factor for considering the reduced
bond was also recommended.
(10)
The development length l
d
= 46d
b
was determined using
ACI 318-08,
1
Section 12.1 for the bars used in the test
specimens. The additional bonded length for the headed bars
in the back 0.5h
c
is equal to 74 mm (2.9 in.) or 3.3d
b
(Fig. 4).
Substituting l
d
= 46d
b
and l
a
= 3.3d
b
into Eq. (10), the bar
stress contributed from this bonded length is only 0.02f
y
.
Finally, the summation of the head bearing capacity and
bond capacity from the bonded length in the back 0.5h
c
ranged between 0.67f
y
and 0.78f
y
for Specimens W0-M1 and
W150-M1, respectively, but the measured stress f
s2
achieved
1.0f
y
in all test specimens (Fig. 10). It is concluded that Eq. (9)
underestimates the head bearing capacity for closely-spaced
headed bars used in this experimental program. This is
attributable to the fact that the smaller c
b
= 1.1d
b
, as half of
the center-to-center spacing between bars shown in Fig. 4, is
a conservative assumption recommended by Thompson et al.
8
due to no available test data for anchorage capacity of
multiple headed bars.
In fact, the lower limit of c
b
used in previous tests
5-7
on
headed bars to derive the head bearing model was 2.5d
b
. This
lower limit is also incorporated in ACI 318-08,
1
Section 12.6.1
as a restriction for the use of the development length equation:
(a) minimum clear cover for a headed bar should not be less
than 2d
b
; and (b) clear spacing between headed bars should
not be less than 4d
b
. The test specimens shown in Fig. 4 did
not meet the latter restriction but did demonstrate adequate
anchorage capacities by the staggering heads. The heads of
longitudinal beam bars were anchored within the column
core, which was confined by lateral reinforcement and the
column axial load. Both helped the anchorage of headed bars
in exterior beam-column joints. These parameters, however,
have not been considered in available models or codes. More
tests on the use of staggered headed bars and multiple
layers of headed bars are recommended to refine the model
for their development.
CONCLUSIONS
Based on the results of this experimental program, the
following conclusions are drawn:
1. The use of a single mechanical anchorage device in place
of the 90-degree hook terminating in the joint resulted in an
equivalent or better performance under large inelastic
displacement reversals. Comparing the test results with the
acceptance criteria given in ACI 374.1-05
15
suggested that the
use of mechanical anchorages in place of hooked anchorages
is appropriate, or even better;
2. Attaching double mechanical anchorage devices on
each beam bar within the joint can enhance the anchorage
and further improve the cyclic behavior of beam-column
connections. The experimental data showed that double
mechanical devices could prevent pushout on the back of the
joint and reduce the yield penetration and bond deterioration
along the beam bar into the joint. Therefore, the degradation
of potential joint shear capacity could be delayed beyond the
limiting drift ratio of 4%;
3. For the presented eccentric beam-column joint specimens,
the evaluation of the experimental joint shear stresses
showed that the definition of b
j
= (b
b
+ b
c
)/2 for computing
the shear area of the joint were not conservative, whereas the
ACI 352R-02
10
effective joint width was conservative and
could be further widened;
4. The side-face blowout failure was not observed during
tests of eccentric beam-column joints with mechanical
anchorage, even though there was only 3d
b
side cover for the
extreme headed bar without any additional restraining
stirrups; and
5. The available head bearing model derived from monotonic
loading tests of headed bars underestimates the anchorage
capacity of closely-spaced headed bars used in this cyclic
loading test. It is concluded that the minimum clear spacing
of 4d
b
between headed bars could be further reduced for
longitudinal beam bars terminated within a continuous
column. Due to a limited number of test results, further
investigation of single or multiple layers of closely-spaced
headed bars is recommended.
ACKNOWLEDGMENTS
The authors are grateful to the National Science Council in Taiwan for
funding this project (No. NSC 94-2211-E-224 -014). The Fu-Sheng Co. Ltd.
in Taiwan is also acknowledged for donating screw-deformed bars and
mechanical anchorage devices.
NOTATION
A
b
= nominal bar area, mm
2
(in.
2
)
A
brg
= net bearing area of headed bar, gross head area minus bar area,
mm
2
(in.
2
)
A
ch
= cross-sectional area of column core, measured to outside edges
of hoops, mm
2
(in.
2
)
A
g
= gross area of column section, mm
2
(in.
2
)
A
s
= area of longitudinal tension reinforcement in beam section,
mm
2
(in.
2
)
A
sh
= total cross-sectional area of transverse reinforcement (including
crossties) within spacing s and perpendicular to column core
dimension b
c
, mm
2
(in.
2
)
b
b
= beam width, mm (in.)
b
c
= column width transverse to direction of shear, mm (in.)
b
c
= column core dimension, measured to outside edges of hoops,
mm (in.)
b
j
= effective joint width transverse to direction of joint shear,
mm (in.)
f
s bond ,
f
y
---------------
l
a
l
d
---- where 1 0.7
A
brg
A
b
5
--------------------
\ .
| |
0.3 = =
ACI Structural Journal/May-June 2009 339
c
2
= minimum cover dimension, measured in direction orthogonal to
c
b
, mm (in.)
c
b
= relative concrete cover dimension, smaller of: (a) distance from
bar center to nearest free surface; and (b) one-half of center-to-
center spacing between bars, mm (in.)
d = effective depth of beam section, mm (in.)
d
b
= nominal bar diameter, mm (in.)
f
c
= concrete compressive strength, MPa (psi)
f
s1
= measured bar tensile stress at beam-column interface, MPa (psi)
f
s2
= measured bar tensile stress at 0.5h
c
along bar into joint, MPa (psi)
f
s,bond
= bar stress provided by bond, MPa (psi)
f
s,head
= bar stress provided by head bearing, MPa (psi)
f
u
= ultimate strength of reinforcement, MPa (psi)
f
y
= yield strength of longitudinal reinforcement, MPa (psi)
f
yt
= yield strength of transverse reinforcement, MPa (psi)
h
c
= column depth parallel to direction of joint shear, mm (in.)
L
b
= vertical distance from column face to actuator centerline, 1950
mm (76.8 in.)
L
c
= horizontal distance between roller supports at column ends,
2700 mm (106.3 in.)
l
a
= additional anchorage length, mm (in.)
l
d
= development length for straight deformed bar in tension, mm (in.)
l
dh
= minimum development length for hooked bar, mm (in.)
l
dt
= minimum development length for headed bar, mm (in.)
M
nb
= nominal flexural strength of beam framing into joint, kN-m (ft-kips)
M
nc
= nominal flexural strength of column framing into joint, kN-m
(ft-kips)
M
r
= column-to-beam flexural strength ratio
m = slope to define effective width of joint
n
5%
= 5% fractile coefficient
P
max
= maximum lateral resistance of test specimen, kN (kips)
P
n
= M
nb
/L
b
, nominal lateral resistance of test specimen, kN (kips)
V
col
= column shear calculated based on probable beam moment at
column face, kN (kips)
V
j,max
= experimental maximum joint shear force, calculated based on
P
max
, kN (kips)
V
n
= nominal shear strength of joint, kN (kips)
V
u
= design shear force acting on horizontal plane within joint, kN (kips)
= bond reduction factor for headed deformed bars
= beam tip displacement, mm (in.)
= target drift ratio, or angular rotation of beam chord with respect
to column chord
= radial disturbance factor, 0.6 + 0.4(c
2
/c
b
) 2.0
REFERENCES
1. ACI Committee 318, Building Code Requirements for Structural
Concrete (ACI 318-08) and Commentary, American Concrete Institute,
Farmington Hills, MI, 2008, 465 pp.
2. Ghali, A., and Youakim, S. A., Headed Studs in Concrete: State of the
Art, ACI Structural Journal, V. 102, No. 5, Sept.-Oct. 2005, pp. 657-667.
3. Wright, J. L., and McCabe, S. L., The Development Length and
Anchorage Behavior of Headed Reinforcing Bars, SM Report No. 44,
Structural Engineering and Materials, University of Kansas, Center for
Research, Lawrence, KS, Sept. 1997, 147 pp.
4. DeVries, R. A.; Jirsa, J. O.; and Bashandy, T., Anchorage Capacity in
Concrete of Headed Reinforcement with Shallow Embedments, ACI Structural
Journal, V. 96, No. 5, Sept.-Oct. 1999, pp. 522-530.
5. Thompson, M. K.; Ziehl, M. J.; Jirsa, J. O.; and Breen, J. E., CCT Nodes
Anchored by Headed BarsPart 1: Behavior of Nodes, ACI Structural
Journal, V. 102, No. 6, Nov.-Dec. 2005, pp. 808-815.
6. Thompson, M. K.; Jirsa, J. O.; and Breen, J. E., CCT Nodes Anchored by
Headed BarsPart 2: Capacity of Nodes, ACI Structural Journal, V. 103,
No. 1, Jan.-Feb. 2006, pp. 65-73.
7. Thompson, M. K.; Ledesma, A.; Jirsa, J. O.; and Breen, J. E., Lap Splices
Anchored by Headed Bars, ACI Structural Journal, V. 103, No. 2, Mar.-Apr.
2006, pp. 271-279.
8. Thompson, M. K.; Jirsa, J. O.; and Breen, J. E., Behavior and Capacity of
Headed Reinforcement, ACI Structural Journal, V. 103, No. 4, July-Aug. 2006,
pp. 522-530.
9. Wallace, J. W.; McConnell, S. W.; Gupta, P.; and Cote, P. A., Use of
Headed Reinforcement in Beam-Column Joints Subjected to Earthquake
Loads, ACI Structural Journal, V. 95, No. 5, Sept.-Oct. 1998, pp. 590-606.
10. Joint ACI-ASCE Committee 352, Recommendations for Design of
Beam-Column Connections in Monolithic Reinforced Concrete Structures (ACI
352R-02), American Concrete Institute, Farmington Hills, MI, 2002, 37 pp.
11. Chun, S. C.; Lee, S. H.; Kang, T. H.-K.; Oh, B.; and Wallace, J. W.,
Mechanical Anchorage in Exterior Beam-Column Joints Subjected to Cyclic
Loading, ACI Structural Journal, V. 104, No. 1, Jan.-Feb. 2007, pp. 102-112.
12. LaFave, J. M.; Bonacci, J. F.; Burak, B.; and Shin M., Eccentric
Beam-Column Connections, Concrete International, V. 27, No. 9,
Sept. 2005, pp. 58-62.
13. Shin, M., and LaFave, J. M., Seismic Performance of Reinforced
Concrete Eccentric Beam-Column Connections with Floor Slabs, ACI
Structural Journal, V. 101, No. 3, May-June 2004, pp. 403-412.
14. Canbolat, B. B., and Wight, J. K., Experimental Investigation on Seismic
Behavior of Eccentric Reinforced Concrete Beam-Column-Slab Connections,
ACI Structural Journal, V. 105, No. 2, Mar.-Apr. 2008, pp. 154-162.
15. ACI Committee 374, Acceptance Criteria for Moment Frames Based on
Structural Testing and Commentary (ACI 374.1-05), American
Concrete Institute, Farmington Hills, MI, 2005, 9 pp.
16. Lee, H. J., and Ko, J. W., Eccentric Reinforced Concrete Beam-Column
Connections Subjected to Cyclic Loading in Principal Directions, ACI
Structural Journal, V. 104, No. 4, July-Aug. 2007, pp. 459-467.
17. Lee, H. J.; Kang, J. Y.; and Lin, Y. J., Improving Seismic Behavior of RC
Interior Beam-Column Connections with Headed Reinforcement, Proceedings
of the 9th Korea-Japan-Taiwan Joint Seminar on Earthquake Engineering for
Building Structures (SEEBUS 2007), Hsinchu, Taiwan, 2007, pp. 103-112.