Carl C. Koester

slicedhillockΠολεοδομικά Έργα

25 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

79 εμφανίσεις

1


INFLUENCE OF ACTIVE CONFINEMENT ON

STRAIGHT
-
BAR ANCHORAGES

IN
1

COLUMN
S

2


3

Carl C. Koester and Christopher Higgins

4


5

Biography:
Carl C. Koester
, rec
eived his B.S. and

M.S. degree
s

in Civil Engineering from
6

Oregon State University, and currently works for
CH2M
-
Hill, Corvallis, OR
.

7


8

Christopher Higgins, Ph.D., P.E., M. ACI

is an Associate Professor in the School of Civil and
9

Construction Engineering at Oregon State University.

His research focuses on evaluation and
10

rehabilitation of aging and deteriorated concre
te bridges.

11


12

Abstract:

13

The c
apacity of reinforced concrete deck girder bridge bent cap

straight
-
bar

anchorages
with
14

1950’s vintage details
were
experimentally
investigated. Eleven subassemblage column
15

specimens were examined with different anchorage lengt
hs, bar groupings, transverse steel

16

amounts
, and externally applied
column
axial force
s
.
Test results showed an increase in
17

anchorage capacity due to the presence of transverse steel and externally applied axial force.

18

Experimental results were compared wi
th contemporary design specifications which do not
19

currently account for the beneficial effects of column axial
compression
acting tra
nsverse to the
20

splitting plane.

A modification factor was developed to account for the presence of axial
21

compression

in de
termining the available anchorage capacity.

Results showed that the
22

2


modifi
cation

provided a reasonable and conservative prediction of the available capacity of
1

straight flexural bar anchorages terminating in
exterior
column
s

subjected to
gravity loads
.

2


3

K
EYWORDS

4

Anchorage, Development length, Bond Strength, Flexural Anchorage

5


6

INTRODUCTION

7

RCDG bridges were widely used through the early 1960's and comprise a large proportion of the
8

US bridge inventory.
Many of these
bridges were designed with less steel
than would be required
9

by modern specifications, contain poorly detailed reinforcing bars,
are reaching the end of their
10

originally intended design life
,

and are showing signs of
d
istress
.

Resources are not available for
11

wholesale replacements and therefor
e many of th
ese bridges remain in service.

Agencies
12

responsible for operation of these bridges continue to inspect and evaluate their fitness for
13

purpose
, often relying on
current
design tools for evaluation of existing
vintage
details
.


14


15

RCDG bridges cons
tructed prior to adoption of ASTM A305
-
50
1
, which standardized reinforcing
16

bars, are generally well detailed: containing hooks and bends for anchorage of flexural
17

reinforcing bars. ASTM A305
-
50 however, relaxed the anchorage and bond requirements and
18

thus,

RCDG's designed during the 1950's and early 1960's commonly contain poor flexural
19

detailing

such as cut
-
offs

in flexural tension zones and straight
-
bar
anchorage
s

in supporting
20

columns
.
An example of such detailing practice is shown in Fig. 1.
The particu
lar focus of this
21

research is
the strength and behavior of
such
existing
bent cap
beam flexural
anchorages

in the
22

monolithic supporting columns

to support shear strength assessment of these deep beams
.

This
23

3


paper reports experimental results from sub
-
assem
blage tests of 1950's vintage

detailed

1

reinforced
concrete
bent cap column anchorages
,

compares the
m

with capacity prediction models

2

and
proposes

a modification factor that can account for the observ
ed beneficial effect of column

3

axial compression
.

4


5

RESEA
RCH SIGNIFIGANCE

6

Large numbers of
older

reinforced concrete structures
are in the national inventory and

for
7

evaluation and condition rating,
the capacity of
girder and beam anchorages terminating in
8

columns
must be determined. Using current
ly available

m
ethods, the strength of
some members
9

can be limited by the straight
-
bar flexural anchorage in the supporting column.
This is
10

particularly true for structures constructed at the advent of deformed reinforcing steel in the
11

1950's
. Large
-
size vintage design d
etails of an
RCDG
b
ridge

column
-
girder joint with
straight
-
12

bar

anchorages were tested, evaluated
,

and compared with modern
design specifications
.

Test
13

specimens
contained
large
-
size bars in different multiple bar groupings and under different axial
14

load le
vel
s
. Combining
the
test
results

with

data from previous research
,

a
modification factor
15

was developed to account for
active
transverse pressure in
determining the available strength of
16

straight
-
bar

anchorages.

Th
e proposed modific
ation factor

allows for
h
igher
anchorage
demand
s

17

than permitted

using

current
design
specifications
.

The ability to better characterize the
18

anchorage capacity
enables engineers
to best reveal the available resistance of structural
19

members
with such details
to preclude unnecessary
and costly replacement, posting, or
20

strengthening, for which limited funds are available.


21


22

LITERATURE REVIEW

23

4


Previous

research
on
bond strength
has considered many
different specimen configurations
and

1

diverse loading scenarios

as detailed in
ACI
-
408R
-
03
2
.

Th
e focus of this investigation i
s on
2

anchorage performance
with applied pressure
transverse
to the splitting plane
, such as that
3

represented by the structural member in Fig. 1
.

Previous studies on
bond

combined
with
4

transverse confining pressure
were c
onducted by
Untrauer and Henry
3
,

Doerr
4
,
Gamb
a
rova and
5

Karakoc
5
,

Robins and Standish
6
,7
,

Eligehausen,
et al
.
8
,

Navaratnarjah and Speare
9
,

Malavr
10
,

6

Nagatomo and Kaku
1
1
,

Baldwin and Clark
1
2
,
Gambarova and Rosati
1
3
,

Batayneh
1
4
,
Walker
et
7

al
.
1
5
,16
. The resear
ch demonstrated that
lateral pressure

applied

transverse to the splitting plane
8

increase
d

the measured bond strength
. Some of these findings showed that
confining pressure
9

was not utilized until 20% of the maximum pullout load
5
,
lateral pressure above 0.3f
'
c

did not
10

increas
e the ultimate pullout load
8
,
and that
bond stress increased

with lateral pressure in
11

proportion to √f'
c

and
appeared limited to lateral pressures of

0.25f'
c

9
.


12


13

The p
revious research
focused on
relatively small size
anchorage
bars and s
pecimen geometries
14

that are not
necessarily
reflective of the conditions
and proportions
commonly found in actual
15

bridge structures
.

T
he applicability of these results to the present case is
uncertain
. No
16

generalized method to account for
column

pressure

a
pplied transverse to the splitting plane

is
17

accepted for North American practice.

The current research was undertaken to quantify
18

anchorage performance for groups of large size bars in realistic sized and proportioned vintage
19

columns with axial load effect
s

and to develop predictive methodologies for
characterizing
20

anchorage
strength

to enable more effective rating of existing members
.

21


22

EXPERIMENTAL PROGRAM

23

5


Test

specimens w
ere

constructed to represent 1950's vintage flexural bar anchorages terminating
1

in

re
inforced concrete column sections.
The size, spacing
,

and details of the column anchorage
s

2

were

based on review of
a

database of
over 100 vintage 1950’s
RCDG
bridges
1
7
.

S
quare
column
s

3

with
large size #11
(#36)
bars anchored with straight
-
bar terminations
w
ere

a

commonly
4

identified
anchorage
configuration
. The

selected laboratory
column
specimen

size
s
w
ere

24 in.
5

(610 mm)

square with an overall
length of
72 in.
(
1829 mm)

to facilitate testing without
6

providing confinement to the anchorage zone

as shown in Fi
g. 2
.

The stress state in the tested
7

column section is conservative when compared with
complete

bent cap members, due to the
8

presence of bending in the column subassemblage and shear that are not observed in full
-
size
9

bent caps.

10


11

A total of
elev
en

specimen
s
were

constructed
and
the principal
test variables included
number of
12

anchorage bars,

anchorage length,
amount of
transverse column reinforcement
,

and column axial
13

force.
These were broken into three test series: Series 1 were single bars

without axial lo
ad

but at
14

different anchorage lengths
; Series 2 were 2 bar groups
without

column axial load; and Series
4

15

which were 4 bar groups
with

and without
column axial load.
Table 1
shows

the test matrix used
16

in this study.
The naming convention used for the speci
mens was based on the number of
17

anchorage
bars, transverse reinforcing

type
, axial
compression
load
applied to
the column, and
18

anchorage embedment length.

The number of bars

being tested

was
1
, 2, or

4

bars in the
19

anchorage
.

The
different
anchorage confi
gu
rations are shown in Fig. 3
.

The transverse
20

reinforcement was described either as
M for

medium


or
L for

none
,


with


m
edium


being

#4
21

(
#
13) nominal Grade 40 (275 MPa) ties

spaced at 8 in.

(203 mm)

on
-
center

and “
none


had no
22

ties
in the anchorage zone

(
Fig.
2
)
.

Three axial loads were used: 0, 200 kips (890 kN), and 500
23

6


kips (2224 kN), and each of these was abbreviated as 0, 2, or 5. The embedment lengths of the
1

single
anchorage bars

were 8, 12 or 21 in. (203, 305 or

533 mm)

as measured from the top
2

surfa
ce of the column face
.

Groups of

anchorage bar
s

were all
tested with
21 in. (533 mm)
3

embedment length
s
.

As a reference, t
he required development length

for the #11
(#36)
anchorage
4

bars in nominal 3000 psi
(
21
MPa)
compressive strength concrete
is about

8
3

in
.
(2
104 mm
)

per
5

ACI 318
-
05

section 12.2.2
18

corresponding to an average bond stress of
274 psi (
1.9

MPa)
.

Thus,
6

these anchorage

lengths

are quite short with respect to
modern design

specification
s
, but are
7

typical of vintage in
-
service
beam
-
column
anchor
ages (Fig.

1).

8


9

Material Properties

10

Concrete was provided by a local ready
-
mix supplier for all specimens.

The concrete mix design
11

was based on 1950’s AASHO “Class A” concrete used in previous research at OSU
1
9
.
The
12

average mix proportions were

1:2.9:10.8
:8.8 (water: cement: coarse aggregate: fine aggregate).

13

A slight amount of air
-
entraining admixture was added to obtain desired workability and material
14

properties.
Concrete strength was specified at 3000

psi (21

MPa)
. Compressive strengths
on

the
15

day
-
of
-
t
est

are
reported
in Table
1
.

The aggregate composition for the mix was reported by the
16

supplier to be: 97% passing the 3/4 in. sieve (19 mm), 82% passing 5/8 in. (16 mm), 57%
17

passing 1/2 in. (12.5 mm), 33% passing 3/8 in. (9.5 mm), 21% passing 5/16 in. (8
mm), 9.3%
18

passing 1/4 in. (6.3 mm), 3.0% passing #4 (4.75 mm), 0.6% passing

#8 (2.36 mm) and 0.3%
19

passing

the #200 (0.075 mm) sieve.

The sand composition of the mix was also reported as:
20

99.7% passing the 1/4 in. sieve (6.3 mm), 96.8% passing #8 (2.36 mm),

59.4% passing #16 (1.18
21

mm), 44.9% passing #30 (0.600 mm), 17.9% passing #50 (0.300 mm), 3.7% passing #100 (0.150
22

7


mm) and 1.7% passing

the #200 (0.075 mm) sieve. The coarse aggregate was from Willamette
1

River bed deposits and was smooth rounded basaltic r
ock.

2


3

Reinforcing for the columns
contained
both transverse
ties
and
longitudinal reinforcing bars.

All
4

transverse reinforcing was
ASTM 615/615M
-
05a
20

nominal Gr. 40 (27
5

MP
a
)
.

Longitudinal
5

c
olumn bars for the specimen
s were ASTM 615/615M
-
05a
Gr. 60

(
420
MPa
)

#11
(#
36) bars.

6

Mechanical properties for all rebar used in the specimens were tested
accor
ding to ASTM E8
-
7

04a
21

with the larger reinforcing bars being
machined to

the 505 specimen size,
and results are
8

shown in Table 2
.


9


10

The
anchorage
reinforcing us
ed predominantly in the
actual bridge
members of interest for this

11

study are ASTM A305
-
50 I
ntermediate Grade bars (nominal Gr.40 (27
5

MPa)).

These bars have
12

a lower yield value than the rebar available for this research because modern #11 (
#
36) deformed
13

ba
rs are commonly only available in Grade 60 (420 MPa).
The anchorage bars used for
14

specimens were
modern
ASTM A615/615M
-
05a
Gr.

60

(420 MPa
)

#11
(#
36) bars taken from a
15

single heat of stee
l
.

Due to use of

modern bars

instead of vintage bars, a c
omparison of

ASTM
16

A305
-
50
and ASTM 615A
/615M
-
05a

specifications
was performed
. Review of the geometric
17

and deformation requ
irements
2
2

demonstrated that the
se parameters
are identical for
round
bar
s

18

meeting ASTM A305
-
50

and
the
modern 615A
/615M
-
05a

specification.

The d
eformation
19

spacing

for bars used in this study

was measured as 0.968 in. (24.06 mm), the deformation height
20

as 0.0812 in. (2.06 mm), and the deformation angle as 60
o
.

The deformation face angle was
21

approximately 30
o
, and the measured diameter was 1.4115 in
. (35.8
5

mm).

To obtain these
22

measurements, a sample was machined along the longitudinal axis of the rebar.

The bar section
23

8


was placed on a high resolution scanner with a reference grid, and a computer aided drafting
1

program was used to determine the base
radius.

Using an average of five deformations, obtained
2

from the sectioned bar, the base radius was determined to be 0.1817 in. (4.62 mm).

M
ill
3

markings
on the bar
s

were not
placed
i
n the concrete anchorage zone.

The chemical composition
4

of the bars was re
ported by the material supplier to be 0.44 C, 1.20 Mn, 0.19 P, 0
.033 S, 0.19 Si,
5

0.25 Cu, 0.09 Ni, 0.16 Cr, 0.005 V, 0.028 Mo, 0.005 Nb, and 0.66
Ce
.


6


7

Specimen Construction

and Test Setup

8

Specimens were constructed in the Structural Engineering Research L
aboratory at Oregon State
9

University.

The overall column configuration

and reinforcing layouts

are

shown in Fig.

2
.
To
10

allow access to the ends of the bars for installation of displacement sensors to measure anchorage
11

slip during testing, a

3 in. (76 mm) l
ong PVC pipe
packed with sand and sealed with caulk
was
12

used to debond the
tails of the
anchorage bars from the concrete
.
C
lear c
over
of
1.5 in. (38 mm)

13

was maintained on all sides
of the specimens
.


14


15

After curing,
s
pecimens were anchored
to the strong flo
or

with h
igh
-
strength threaded rods
16

placed 4 ft. (305 mm) on
-
center
for testing.

The overall test setup is illust
rated in Fig
s
.
3
-
5
.
The
17

anchorage bars were loaded with hydraulic cylinders that
reacted on

a stiffened
W12x152

18

loading beam
. The reactions wer
e measured with load cells under the cylinders.

Load was
19

transferred into the anchorage bars by means of mechanical bar couplers designed to butt
-
join
20

two #11 (
#
36) Grade 60 (420 MPa) bars.

Plate washers were placed between the bars and
21

loading beam to ena
ble uniform bearing on the loading beam
.

S
eries 1 tests
were
performed
22

using a single 200 kip (890 kN) hydraulic cylinder and hollow
-
core load cell
. Series 2
and 4
tests
23

9


were performed with two hydraulic cylinders
as illustrated in Fig

5
.

S
ome of the S
erie
s
4

tests
1

were also performed
with
axial force
applied
on the column section

using a
self
-
reacting

system
2

as seen in Fig.
3 and 4
. A hydraulic actuator was used to produce a compression force within the
3

column section and reacted against a stiffened W12x12
0. Four high
-
strength (120 ksi, 827 MPa)
4

bars connected two W12x120
steel
sections at the ends of the column section (
Fig.
4
).

Axial
5

force was applied using a manual hand pump. The hydraulic cylinder, load cell, bearing plates,
6

and
loading beams

were
align
ed

through the geometric center of the column

section
.


7


8

Instruments were applied to the specimens to measure displacements, loads, and strains during
9

testing.

A
nchorage bar movement relative to the concrete was measured as well as overall
10

anchorage bar di
splacements.

Strain gauges were placed after casting on the anchorage bars at
11

the loaded end.

Strain gauges were
also
surface bonded to the column flexural reinforcing bars
12

prior to casting for
Series 4
.

Data from sensors were acquired and stored using com
mercially
13

available PC
-
based data acquisition hardware and software.

Data were colleted at a rate of 4 Hz
14

and archived for subsequent data analysis.

15


16

Loading Sequence

17

Pullout load
on the anchorages
was applied
psuedostatically
in

incrementally larger loa
d steps

18

with a manual hydraulic pump
.

The loading rate was approximately
2.5
kip/sec (
11.1
kN/sec). At
19

each load step the

applied pullout
force was increased to the target level and then unloaded. The
20

load steps were increased

until specimen failure.

An ex
ample of the loading history is shown in
21

Fig.
6

and represents a live load

induced

loading
-
unloading response that is cyclic, but without
22

reversals.
Due to the unloading

phase
, this loading se
quence is more demanding of the

23

10


anchorages th
an
a
conventional m
onotonic test
.
The manual
hydraulic
control
and stiff reaction
1

system
permitted measurement of the descending branch of the anchorage response after
2

reaching ultimate.


3


4

A
xial compression was appli
ed to
some
specimens in Series
4

and t
wo
levels of axial lo
ad were
5

used:

designated as "low" and "high
.
"

The magnitude of axial load for the

low


value was
6

determined from typical service level dead load on a RCDG bridge from the 1950's.

S
tructural
7

details from
a typical
vintage
RCDG b
ridge in the inventory
2
3

wer
e used to determine the
service
8

level
axial force in the columns from the weight of components and wearing surface
.

A loading
9

uncertainty factor of 5% was used to account for additional weight from other components

and
10

overlay

that may have been added to t
he bridge, but not specified on the original plans.

The
11

service
-
level axial loads on the intermediate supporting columns were calculated at

12

approximately

200 kips (8
90

kN), and this magnitude was applied to two
test
specimens.

Th
e

13

induced column compressio
n stress

corresponds to roughly 0.1f

c

for nominal 3300 psi (
22.8
14

MPa) concrete, common for the era.

To model

large

live loads

in addition
to
that of the
15

components and wearing surface,

a "high" applied axial load was selected as 500 kips (2224 kN)
,
16

which
corresponds to 0.26f

c
.

The allowable compression stress for
short
tied
columns in the
17

195
7

AASHO
S
pecification
2
4

section
1.7.8
f

was
approximately
0.
22
f

c

assuming 1% reinforcing
18

ratio, an allowable reinforcing compression stress of 16 ksi

(110 MPa)
, and s
pecified concrete
19

compression
strength

of 3300 psi

(22.8 MPa)
.

The desired axial force level was applied prior to
20

applying the pullout
force

on the anchorage
and maintained during testing.

21


22

EXPERIMENTAL RESULTS AND DISCUSSION

23

11


All specimens were tested to

failure. Failure modes varied for the different specimens and
1

included anchorage yielding, anchorage pullout, and column shear failure.

Specimen
responses

2

were evaluated for maximum applied
pullout

load, slip of the anchorage bar,
average bond
3

strength
,
a
nd failure mode

and
results
are reported in Table
s

3a and 3b
.

In addition, crack pattern
4

and crack angle comparisons were made.
To compare the effects of the various parameters, while
5

reducing the variation that was present in f'
c
, the ultimate rebar stres
s
was normalized

to a
6

reference strength

as
(
c day-of-test
3000/f'
)

12
. Also, i
t has been shown that the bond strength is
7

best characterized by
4
c
f'

25,2
, however for low strengt
h concrete (less than 8000 psi (
55 MPa
)
)
8

c
f'

is considered reasonable
2

and was used in the present study.


9


10

Average bond stresses were determined by taking the applied pullout load at failure and dividing
11

by the
total
surface area of the bar(s) embedded in the column.

Average bond
strength was the
12

highes
t for Specimen 1M0.12.
All specimens containing multiple bar groups had lower average
13

bond strength, even as they exhibited higher

overall anchorage capacities.
Axial compression
in
14

the column and the addition of transverse steel in
the anchorage zone
improved the average
bond
15

strength of the specimens
, although column axial load had a much larger effect.


16


17

T
he normalized
average rebar stress

at ultimate
was used to evaluate the
different
anchorages

as
18

seen in T
able 3b
. T
his was cal
cu
lated by taking the maximum ap
plied force and dividing it by
19

the total area of steel in the anchorage group

and normalizing it with respect to the 3000 psi
(20.7
20

MPa) reference
concrete strength
.

The 12 in. (
305
mm) embedment of Specimen 1M0.12 was
21

suffici
ent to enable
yield
for
a
n Intermediate Grade

(Gr. 40 ksi (
275 MPa
)
) bar
, while the very
22

short embedment of Specimen 1M.08
was sufficient to produce only about half the yield stress of
23

12


an

Intermediate Grade steel
.
The embedment of
Specimen 1M0.21
was
suffi
cient to producing
1

yielding of the Gr. 60 (
420

MPa) bar.

For the Series 2 tests, the
average
bond

strengths
w
ere

2

sufficient to
produce
yield
ing

in
a similar group o
f #11 (#36) Gr. 40 (275 MPa) bars
.

For the
3

Series 4 tests, the embedments were only able to
achieve the yield stress of Intermediate Grade
4

bars when axial load was present.


5


6

The

pullout

load applied to the anchorage
group at a slip of 0.005 in (0.127

mm) and load at
7

initial
slip by offset are shown in T
able 3a
.

Initial

s
lip
was
identified
from i
nspection of the load
-
8

slip
curves

where
the
tangent
stiffness
first deviates from
the initial secant stiffness
.

Slip
9

represents movement of the reinforcing bar relative to the rigid body motion of the concrete
2
6
.
10

F
or specimens 1M0.21 and 1M0.08
,

slip w
as o
btained from the loaded end

using a
reference slip
11

of

0.01 in (0.254 mm)
2
7
.

For
all
the
other specimens, the slip was determined from the unloaded
12

end of the anchorage bars
using
a

reference slip of 0.005 in. (0.127 mm)
2
7
. S
lip was measured
for

13

the unloade
d ends of the anchorage bars
relative to the concrete surface
using displacement
14

sensors placed on the bars through the blockouts described in the previous section.

15


16

Pullout failure of the anchorages was observed for the majority of specimens except for
17

s
pecimens
1M0.21
,
4M0.21,

and

4M5.21.

Specimens in which pullout failures were observed
18

had
wedge
-
shaped
blocks of cracked concrete displaced along with the reinforcing bars at
19

failure.

These
wedge
-
shaped concrete pieces varied in size and shape
depending

o
n the presence
20

of transverse reinforcement,
column
axial
force
, and bar configuration.

Examples of the s
urface
21

c
rack patterns
on the
side
face
of the
column
s

for
Series 4

specimens are shown in Fig.
7a and b
.
22

For specimens with ties

in the anchorage zone
,
the ties restrained the cracked
concrete
23

13


surrounding the anchorage bars
thereby
limiting the

overall depth of the
wedge

to the
cover
1

dimension

of the column
longitudinal
bars.

For specimens without ties, the wedges were deep
er
,
2

penetrating
in some cases
10

in. (254 mm) into the column
.

The concrete below the wedge and
3

around the bar deformations showed visible signs of crushing as the bar was pulled through the
4

concrete.

Transverse reinforcement increased the capacity of the specimens

on average 15%

by
5

prev
enting early separation of the concrete wedges

after initial cracking
.


6


7

T
he presence of axial compression in the column was seen to increase the capacity of the
8

specimens

60% on average.

The increase in anchorage strength with axial load was not
9

proportio
nal to the applied axial load
.
Additionally, increasing axial load
tended to reduce the
10

slip at corresponding
pullout

load values as seen in Fig.
8
.
It was
observed that the
load at which
11

initial cracking
was first
observed
increased with
higher column
axi
al
force.

Additionally,

the
12

extent
of

cracking
decreased as column

axial
force
increased

and
the
anchorage zone
crack
13

angles
changed depending on the axial load levels
(Fig.
7a and b
)
. Crack angle
s

decreased with
14

an increase in

the externally applied

colum
n
axial force. For specimens with
no ties
,

the
15

anchorage zone

crack angle
s

were
37°, 28°, and 15°, with respect to the horizontal

for axial loads
16

of
0, 200, and 500 kips

(0,
8
90, and 2224 kN)
,

respectively
.

A
similar trend was observed for
17

specimens with
t
ies
,
where

crack angles

were
40°, 26°, and 14°
, for the axia
l loads of 0, 200, and
18

500 kips

(0,
8
90, and 2224 kN)
,

respectively
.


19


20

Combining axial load with transverse ties produced a mixed failure mode for Specimen 4M5.21,
21

in which the two inner bars yiel
ded and the two outer bars pulled out, as
shown

in Fig.
8
. As seen
22

in
this figure
,

the
inner and
outer bars
we
re observed to
exhibit
similar slips
at lower loads
, but as
23

14


the maximum load is achieved, the outer bars slip increases much more than the inner b
ars
.

This
1

can be attributed to the interaction of the anchorage with the column reinforcing and the free
2

edges of the column face.

The
free
edge is 4.2

d
b

away from the center of the outside bars
.

O
nce
3

splitting and diagonal cracks form, the outermost
bars

have

reduced
restraint to slip
compared to
4

the interior bars.


5


6

Specimen 4M0.21 p
roduced a column shear failure
. In
comparing
the column shear
strength

7

versus
the demand from the anchorage
for each spe
cimen it was observed that
the column shear
8

demand inc
reased for larger bar groups,
column
axial
load,

and em
bedment length
.

External
9

axial force applied to the column
also
increased the
shear
capacity of the column sections
10

sufficiently to prevent one
-
way shear failure of the column

for later Series 4 tests
,

even with
11

higher demands
produced

by

these
anchorage groups.


12


13

ANALYSIS AND EVALUATION

14

The experimental data were compared to
modern

development length equations for straight bar
s

15

from ACI

318
-
05

as well as comparison with earlier ACI 318
-
56
2
8

allowable
stress design
16

provisions.

ACI 318
-
05 section 12.2.3
specifies

development length as

a function of concrete
17

strength, bar size, bar spacing, yield stress of the steel, transverse reinforcing and different design
18

parameter coefficients.

ACI 318
-
56 specified
different

allowable bond stress
es

for A305 bars
19

that depended primarily on whether they were top or bottom bars. For bottom bars, the allowable
20

bond stress was
0.1f

c
.

21


22

15


An

allowable stress
condition

check
was performed at the observed first slip
by compari
ng the

1

measured average
bond stresses with the 1956 specified allowable stress

as seen in T
able 3a
. For
2

the
Series 1 and 2 specimens, first slip was observed at
normalized
average bond stresses above
3

the
1956
specified allowable by an average of 1.9. For t
est Series 4, specimens without axial load
4

(4M021 and 4L021) exhibited first slip at
normalized
average bond stresses below the
1956
5

specified allowable by an average of 0.7. The presence of axial load significantly increased the
6

normalized
average bond st
ress at first slip above the
1956
allowable bond stress by an average
7

of 2.1 as compared to the 4 bar specimens without axial load.

As
seen in Table 3a
, on average,
8

the presence of axial load in the column
only slightly
increase
d

the
measured
bond stress a
t first
9

slip

as
compared with the Series 1 and 2 results. Thus, at service level conditions, the beneficial
10

effects of axial load in the column are not as prominent and may not
produce

substantially
11

reduced
initial
slip

or well

constrain
crack
s

at these
lo
w
ranges of response. However, for
12

strength evaluation of existing structures, of which many are already cracked, the anchorage
13

capacity may be of greater importance

and is discussed subsequently
.

14


15

Experimental results for
normalized
average bond
stre
ngth

and maximum average
re
bar
stress
16

were used to compare with the previously described
strength
design
methodologies for predicting
17

the capacity of the flexural bar anchorages
in column sections.

ACI 318
-
05 do
es

not provide a
18

means of explicitly evaluating an
chorages in the presence of confining axial loads, and
19

modifications are proposed to account for this beneficial effect.

ACI

318
-
05
use
s

ultimate
20

strength design limit states and define
s

a development length required to produce yielding of the
21

bars. This d
oes not provide a means of directly evaluating ultimate bond strength. However,
22

comparisons between the experimental results an
d the available analysis method

were made by
23

16


taking a ratio of the experimentally provided anchorage length to the specified
deve
lopment

1

length computed from the
design
methods. Multiplying this ratio by the anchorage yield stress
2

determines the fraction of the rebar stress predicted by the provided embedment.

As
seen in
3

previous work
2,12
, increasing the embedment length does not pr
oportionately increase the bond
4

strength, but the relationship is reasonably close to a linear
stress

transfer and thus may be
5

expected to provide reasonable results.


6


7

The c
apacity of the four bar specimens without axial load
was reasonably predicted
when

using
8

ACI

318
-
05
Eqn. 12
-
1 and the
detailed
(c
b
+K
tr
)/d
b

term
.
The contribution of
K
tr

was taken
from

9

the column longitudinal steel acting transverse to the splitting plane

and for all specimens this
10

term reached the limiting value of 2.5
. However, the sam
e approach underpredicted the strength
11

of the specimens with axial load. When the K
tr

term is reduced to zero for analysis
12

simplification, ACI 318
-
05
Eqn. 12
-
1
was quite conservative for all cases, with average
13

exp
erimental to predicted anchorage strength

of 2.3. The

four bar specimens with axial
load
were
14

predicted with
even
higher levels of
conservation

having

average experimental to predicted
15

strengths of 3.4
.
Therefore to
ameliorate

some of the
excessive
conservatism found when
16

evaluating vintage
straig
ht
-
bar

anchorages

in columns with axial load
, a modification

factor

wa
s
17

developed to account for the beneficial presence of
pressure
transverse
to the splitting plane
18

provided by the column axial load
.

19


20

Proposed Modification to

Anchorage Development Length

Due to

Column
Axial Confinement

21

The above
specification design

method

provided overly conservative estimates of anchorage
22

capacity for the specimens, including, in particular,
th
ose

specimens with

service
-
level

axial load
23

17


magnitude
s
. In an effort to estab
lish a
more
reasonable prediction of
the
flexural

anchorage
1

capacity
in column
-
bent connection
s

with applied axial
compression

under gravity loads,

a
2

modification
to the anchorage
strength is proposed that uses
the

reference strength from
ACI
3

318
-
05
. The m
odification takes into account
active confinement
from the

column
axial
4

compression
stress
acting

across the potential
anchorage
splitting plane
.

5


6

Data from
the
present research
,

along with Elighausen
et al
.
7
, Untrauer and Henry
3
,

Robins and
7

Standish
6
,

and

Batayneh
14

were

combined to develop a modification factor for
anchorage
8

strength prediction

that accounts for the column confining pressure.

Confining pressure and
other
9

modeling
input
data
were
determined from reported figures and tables.
Additional deta
ils
10

regarding the specimens, data, and calculations made from thes
e earlier studies is reported by
11

Higgins
et al.
17
.

The specimens in which yielding occurred were eliminated from the analysis,
12

along with specimens in which the lateral pressure applied was
greater than 0.3f'
c,

which was on
13

average
observed to be the upper

limit
of the
benefit seen from adding transverse pressure
6,11,15
.

14

For each individual specimen, the ACI 318
-
05
reference
development lengths were calculated

15

per ACI 318
Eqn. 12
-
1
.

The term
K
tr

was taken as zero for the development of the modifying
16

term since it was seen that the benefit of axial pressure was substantially larger than that of the
17

column reinf
orcing
.

The term
(
c
b
+0)/
d
b

was still limited to a maximum value of 2.5.

The ACI
18

318
-
0
5
computed
reference
development length

was

divided by the

experimentally

provided
19

embedment length

and this

was
then
multiplied by the yield stress to produce a predicted
re
bar
20

stress. The predicted
re
bar stress was then compared to the experimentally rep
orted stress
.
The
21

ratio of predicted to
measured

re
bar stress along with applied column pressure (acting transverse
22

to the splitting plane) w
as

evaluated
.

A

best fit
curve
was determined
by performing a nonlinear
23

18


regression of the data
.

Assuming a normal d
istribution
for

the data, confidence intervals were
1

established. The best fit curve and the confidence intervals are shown in Fig.
9
.
As seen here, the
2

best fit curve is approximately linear with respect to the column axial compression stress.
3

Considering
a

95%
probability
lower bound level and approximating it as linear, a
best fit
4

modification factor was established as:

5


0
.
1
,
25
.
2
800
8
.
0




p


[
1
]

6

w
here p is taken as the service
-
level column compression stress (psi)

on the gross
column
cross
7

section

acti
ng
transverse to

the anchorage splitting plane. An upper bound of 2.25 was
selected

8

(approximately p<
0.3f'
c

for lower strength concrete) due to limitations in the data at very high
9

compression stresses and

prior research

results.

A

lower bound of 1.0 provi
des that
axial
10

compression stresses below 160 psi
(1.1 MPa)
are not relied upon to provide
increased bond

11

stresses
.

The proposed values are shown in T
able 3b

for the given specimens and can be seen to
12

produce

conservative
results
compared with the observed

values for the strength limit state.

13


14


The
modification factor
,

κ
,

can be seen
as increasing the availabl
e bond strength

and thereby
15

decreasing the
development length
and is implemented
as
:


16



L
d

ACI Modified

=
y
s
b
c
b
b
f
ψ
3 1
d
40 f'
κ
c
d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

[
2
]

17

where f
y

is t
he anchorage bar yield stress (psi), f

c

is the concrete compressive strength (psi),

s

is

18

a rebar size factor (0.8 for #6 and smaller and 1.0 for bars #7 and larger), and
d
b

is the bar
19

diameter

(in)
, c
b

is the smaller of t
he distance from the center of th
e

bar to the nearest concrete
20

19


surface or 0.5 times the center
-
to
-
center spacing of bars being developed (in).
Because the
1

research involved only uncoated, bottom bars in normal weight concrete, these

other

factors are
2

not included here.
As seen in this for
mulation,

the modification factor calibration set the K
tr

term
3

to zero.
T
hus it should not be reintroduced when using the reduction factor,

, as the beneficial
4

confinement provided by the column axial compression acting transverse to the potential
5

splitting plane supersedes that provided passively by the reinforcemen
t.

It would be

conservative
6

to use the simplified development length expressi
ons in ACI 318
-
05 section 12.2.2 with the
7

proposed modification factor to predict anchorage strength.
It should be noted that
the proposed
8

modification factor

κ
,

developed
for use
as Eq. 2,

allows a
significant

increase in anchorage
9

strength
than currently

permitted and
is limited to applications
for

existing structures

under
10

gravity loading containing materials, details, and proportions like those considered here rather
11

than for new design
.

12


13

Comparison with
CEB
-
FIP
Recommendations

14

In evaluating the propose
d method, a comparison was made with the CEB
-
FIP
15

Recommendations

(
CEB
-
FIP
)
29

which
includes

a reduction factor to account for the presence of
16

compressive
transverse
stress across

the splitting plane of the developing bars.
CEB
-
FIP
17

Recommendations

Section 2
.4.1.5 describes the reduction factor as follows
:

18


(1-0.04p)
1
1.5


[
3
]

19

where p is the pressure transverse (MPa) to the plane of reinforcement.

This term is applied as an
20

increase to the design bond strength

and

can be taken as 3/2 at an end an
chorage.

While it may
21

appear that Eq. 1 is less conservative than Eq. 3
, the
CEB
-
FIP

procedure
s

produce

shorter
22

20


straight
-
bar

anchorage length
s than those in

ACI 318
-
05. Comparing the simplified development
1

length equations in ACI 318
-
05 for #7 (
#22
) and la
rger bars with clear spacing of at least a bar
2

diameter and at least minimum stirrups against that in

section
2.4.1 of CEB
-
FIP
3

Recommendations considering 36 mm diameter

bars
, over a range of concrete compression
4

strengths from 3000 to 5000 psi (20 to 35 M
Pa), the CEB
-
FIP Recommendation anchorage
5

lengths are on average 1.44 time shorter than those by ACI. As a result, at the limiting values of
6

Eq. 1 and Eq. 3, the proposed approach is only about 3% higher than the CEB
-
FIP
7

Recommendations and produces about
the same anchorage lengths.
At lower column axial
8

stresses, the proposed method
with Eq.
1

would result in
higher

predicted anchorage
lengths than
9

those

from
CEB
-
FIP
.


10


11

CONCLUSIONS

12

Tests were conducted on specimens
representing
1950's vintage
detailed
ben
t

cap beam flexural
13

reinforcing bar
s

terminating in columns
to determine the influence of different bar groupings,
14

column
axial load
s
, and transverse steel parameters on
anchorage
strength
and behavior
.

Eleven
15

anchorage tests were performed.
R
esults were r
eported and compared with current and
past
16

design specifications.

An active confinement term was developed
from the present and
archival

17

experimental
results
to better predict anchorage capacity
using
currently
available design
18

methods
. Based on the report
ed research

findings
, the following conclusions are made:

19



The subassemblage specimens provided a reasonable approximation of the details for
20

1950's vintage flexural anchorages terminating in columns.

The stress state in the tested
21

column section is conser
vative when compared with
complete

bent cap members, due to
22

21


the presence of bending in the column subassemblage and shear that are not observed in
1

full
-
size bent caps.


2



Bar location affected the slip and capacity of bars, with those located near the free e
dge
3

of the specimen exhibiting reduced capacity.

The
se

outer bars also tended to have larger
4

slip values and less

stiffness than those located near

the
center

of the
column
section.


5



The presence of multiple bars decreased the average
bond strength
. The
re
bar stress was
6

lower for the four
-
bar anchorage groups compared with the two
-
bar anchorage groups.

7



Demand on the column section was sufficient to produce one
-
way shear failure for
a

four
8

bar anchorage group. Externally applied axial force increased the
she
ar
demand on the
9

column due to the higher anchorage capacity

but also increased

t
he shear capacity of the
10

column
s

at a higher rate
such that shear failures were not observed for the axially loaded
11

specimens
.

12



Column sections with transverse reinforcement ex
hibited greater capacity than
13

comparable specimens without transverse reinforcement.

The ties were seen to increase
14

capacity on average by 1
5
%.

Ties better
restrained
the concrete
after cracking
, limiting
15

the depth of penetration of the
observed
concrete
w
edge
at
pullout.

16




The
service level dead load
axial force
magnitude
applied to the column increased the
17

anchorage capacity of the specimens by 6
0
% on average.
Further increasing the axial
18

force
to the 1957
AASHO
specified maximum allowable
column
stress
di
d not
19

significantly increase the anchorage capacity.


20



The
application

of
column
axial
compression
force reduced the observed crack angles

21

during pullout testing
, when compared to
otherwise similar

specimens without applied
22

22


axial
compression
. Specimens with

the highest applied axial load exhibited the shallowest
1

crack angle of 15° with respect to horizontal.

2



The benefit of column axial
compression
stress acting transverse to the anchorage
3

splitting plane is neglected by current US
design specifications
.

The
s
e

des
ign
4

specifications underestimated

the
available anchorage strength

more substantially
for
5

s
pecimens with
pressure applied transverse to the splitting plane from
axial compression

6

in the column section
.


7



A development length
modification factor

(κ) was

introduced

to include the beneficial
8

effects of
column
axial
compression stress
acting transverse to the splitting plane. The
9

addition of active confinement effect
s

enabled better prediction of
anchorage

capacity for
10

the
specimens

and may
enable better
es
timate
s

of
member capacities

for vintage RCDG
11

structures

using A305 bars
.

12



U
pper
and lower
limit
s

on κ,
were

established to
restrict the benefit of high column
13

compression stress due to limitations seen in previous research and
neglect the
14

contribution of low column
axial
compression stresses
. The
anchorage length
15

modification factor

κ
is

proposed for

use on

anchorage conditions similar to those tested
.

16

F
urther

research
may be

needed to provide
additional
statistical measures of uncertaint
y

17

over a broader
range

of member proportions, axial force
magnitudes
, and anchorage
18

details.


19


20

23


ACKNOWLEDGMENTS

1

This

research was funded by the Oregon Department of Transportation

and the Federal Highway
2

Administration.
The

authors
would like to thank Mr. Steven M. Soltesz
,
of the Oregon
3

Department of Transportation Research Unit for his assistance in coordinating this
research
4

effort.

T
he
opinions,
findings and conclusions are those of the authors and may not represent
5

those acknowledged
.


6

7

24


REFERENCES


1

1. ASTM A305
-
50T, “Minimum Requirements for the Deformations of Deformed Steel Bars for
2

Concrete Reinforcement,” ASTM I
nternational, 1950, pp. 218
-
220.

3

2. ACI Committee 408R
-
03,

Bond and Development of Straight Reinforcing Bars in Tension,


4

ACI Manual of Concrete Practice Part 5, 2004, pp.1
-
49.

5

3
.

Untrauer, R. E.; and Henry, R. L., "Influence of Normal Pressure on Bond St
rength,"
Journal
6

o
f the

American Concrete Institute Proceedings
, V. 62, No. 5, May 1965, pp. 577
-
585.

7

4
. Doerr, K., "Bond Behavior of Ribbed Reinforcement under Transversal

Pressure,"
Nonlinear
8

behavior of reinforced concrete structures; contributions to
IASS symposium,

International
9

Association for Shell Structures, V. 1, July 3
-
7, 1978, pp. 13
-
24.

10

5
. Gambarova, G.,

and Karakoc, C., "Shear
-
Confinement Interaction at the Bar
-
to
-
Concrete
11

Interface,"
Bond in Concrete
, Ed. Bartos, P.,
Applied Science Publishe
r
,
1982, pp. 82
-
96.

12

6
. Robins, P. J.,

and Standish, I. G., "Effect of Lateral Pressure on Bond of Reinforcing Bars in
13

Concrete,"
Bond in Concrete
, Ed. Bartos, P.,
Applied Science Publisher
, 1982, pp. 262
-
272.

14

7. Robins, P. J.,

and Standish, I. G., "The
I
nf
luence of
L
ateral
P
ressure
U
pon
A
nchorage
B
ond,"
15

Bond in Concrete,
Magazine of Concrete Research
, V.36, No. 129, Dec. 1984, pp. 262
-
272.

16

8
. Eligehausen, R., Popov, E.,

and Bertero, V. V.,
"Local Bond Stress
-
Slip
Relationships of
17

Deformed Bars under General
ized Excitation,"
UCB/EERC Report 83
-
82
, Earthquake
18

Engineering Research Center, University of California at Berkeley, Berkeley, Calif.
, 1983.

19

9
. Navaratnarajah, V., and Speare, P. R. S., "An
E
xperimental
S
tudy of the
E
ffects of
L
ateral
20

P
ressure on the
T
ra
nsfer
B
ond of
R
einforcing
B
ars with
V
ariable
C
over,"
Proceedings of the
21

Institution of Civil Engineers
, V.81, No. 2, Dec. 1986, pp. 697
-
715.

22

25


10
. Malvar, J., "Bond of Reinforcement Under Controlled Confinement,"
ACI Materials Journal
,
1

V. 89, No. 6, Nov.
-
Dec
. 1992, pp. 593
-
601.

2

11
. Nagatomo, K., and Kaku, T.,

Confinement Stress Dependant Bond Behaviors under lateral
3

compressive and tensile stress


Bond in Concrete,
International Conference on Bond in
4

Concrete
, Riga Technical Univers
ity, Oct. 17
-
19, 1992, pp.

69
-
78.

5

12
. Baldwin, M.I.,

and Clark, L.A., "The Assessment of Reinforcing Bars with Inadequate
6

Anchorage,"
Magazine of Concrete Research
, V. 47, No. 171, June 1995, pp. 95
-
102.

7

13
. Gambarova, P. G.,

and Rosati, G., "Bond and
S
plitting in
R
einforced
C
oncre
te:
T
est
R
esults
8

on
B
ar
P
ull
-
O
ut,"
Materials and Structures
, V. 29, June 1996, pp.267
-
276.

9

14
.

Batayneh, M. K.
, “The effects of lateral compression on bond between deformed reinforcing
10

bars and concrete,” Oxford Brookes University, Thesis Dissertation,

No
vember 1993.

11

15
. Walker, P. R.
,
Batayneh, M. K.
,

and Regean, P. E., "Bond
S
trength
T
ests on
D
eformed
12

R
einforcement in
N
ormal
W
eight
C
oncrete,"
Materials and Structures
, V. 30, Aug.
-
Sept. 1997,
13

pp. 424
-
429.

14

16
. Walker, P. R.
;

and Batayneh, M. K.
;

and Regean
, P. E., "Measured and
D
esign
Bo
nd
15

S
trengths

of
D
eformed
B
ars,
I
ncluding the
E
ffect of
L
ater
C
ompression,"
Materials and
16

Structures
,
V. 51, No. 1, Feb.

1999
,

pp. 13
-
26.

17

17
. Hi
ggins, C.,
Senturk, A.E, and Koester, C.C.
, "
Evaluation of Bent Caps in Reinforce
d
18

Concrete Deck Girder Bridges
," Oregon Department of

Transportation, Salem, OR, 2008
.

19


18
. ACI Committee 318, "Building Code Requirements for Structural Concrete," ACI 318
-
05
20

and Commentary ACI 318R
-
05, Michigan: American Concrete Institute, 2005.

21

1
9
.
Higgins, C., Miller, T., Rosowsky, D.V.,

Yim, S.
C.
,

Potisuk, T.
, Daniels, T.
K.
,

N
icholas,
22

B.
S., Robelo,

M. J.
,

Lee, A
.
-
Y
., and Forrest, R. W., "SPR 350 Assessment Methodology for
23

26


Diagonally Cracked Reinforced Concrete Deck Girder," Oregon Department of Tr
a
nsportation,
1

Salem, OR, 2004.

2

20
. ASTM A615/A615M
-
05a, “Standard Specification for Deformed and Plain Carbon
-
Steel
3

Bars for Concrete Reinforcement,”
ASTM International
, 2005, pp. 329
-
333.

4

21
. ASTM E8
-
04, "Standard Test Methods for Tension Testing of Metall
ic Materials,"
ASTM

5

International
, 2005, pp. 64
-
87.

6

22
. Howell, D., and Higgins, C.
,

"Bond and Development of Deformed Square Reinforcing
7

Bars,"
ACI Structural Journal
, V. 104, No. 3, May
-
June 2007, pp. 331
-
341.

8

23
. Oregon State Highway Department Bridge D
ivision, "McKenzie River Bridges," Drawing
9

number 13726, Sheet 3/14, Sept 1957, Accessed Oregon State University Bridge Database 2006.

10

2
4
. American Association of State Highway Officials

(AASHO)
, "Standard Specifications
f
or
11

Highway Bridges," 1957, pp. 36
-
37.

12

2
5.Darwin, D.,

Lutz, A. L.
,

and Zuo, J., "Recommended Provisions and Commentary on
13

Development and Lap Splice Lengths for Deformed Reinforcing Bars in Tension,"
ACI
14

Structural Journal
, Vol. 102, No. 5, Nov.
-
Dec. 2005, pp. 892
-
900.

15

2
6
. Lutz, L. A., and

Gergely, P., "Mechanics of Bond and Slip of Deformed Bars in Concrete,"
16

ACI Journal Proceedings
, V.64, No.11, Nov. 1967, pp.711
-
721.

17

2
7
. Clark, A. P., “Bond of Concrete Reinforcing Bars
,

Journal of the American Concrete
18

Institute Proceedings
, V. 21, No.
3 Nov. 1949, pp. 161
-
184.

19

2
8
. ACI Committee 318, "Building Code Requirements for
Reinforced
Concrete," ACI 318
-
56
,
20

Michigan: A
merican Concrete Institute, 1956
.

21

29
.

CEB
-
FIP Commission 3, "Practical Design Working Group on Recommendations for
22

Practical Desig
n of Structural Concrete," SETO, London, 1999.

23


24