Structure and Infrastructure Engineering

Maintenance, Management, Life-Cycle Design and Performance

ISSN: 1573-2479 (Print) 1744-8980 (Online) Journal homepage: www.tandfonline.com/journals/nsie20

Design and construction of UHPC-filled duct

connections for precast bridge columns in high
seismic zones

Mostafa Tazarv & M. Saiid Saiidi

To cite this article: Mostafa Tazarv & M. Saiid Saiidi (2017) Design and construction of UHPC-
filled duct connections for precast bridge columns in high seismic zones, Structure and
Infrastructure Engineering, 13:6, 743-753, DOI: 10.1080/15732479.2016.1188969

To link to this article: https://doi.org/10.1080/15732479.2016.1188969

Published online: 03 Jun 2016.

Submit your article to this journal

Article views: 829

View related articles

View Crossmark data

Citing articles: 20 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=nsie20
Structure and Infrastructure Engineering, 2017
VOL. 13, NO. 6, 743–753
http://dx.doi.org/10.1080/15732479.2016.1188969

Design and construction of UHPC-filled duct connections for precast bridge columns
in high seismic zones
Mostafa Tazarva and M. Saiid Saiidib
a
Department of Civil and Environmental Engineering, South Dakota State University, Brookings, SD, USA; bDepartment of Civil and Environmental
Engineering, University of Nevada Reno, Reno, NV, USA

ABSTRACT ARTICLE HISTORY

One of the column connection types that has been recently emphasised for accelerated bridge construction Received 7 September 2015
(ABC) in high seismic zones is grout-filled duct connections. Experimental results have shown that the Revised 17 March 2016
Accepted 4 April 2016
seismic behaviour of reinforced concrete columns connected to precast cap beams using this connection Published online
type is emulative of conventional columns. However, an alternative ABC connection is required to connect 4 June 2016
precast columns to shallow cap beams and footings due to insufficient anchorage length in those elements.
Ultra-high-performance concrete (UHPC) was proposed in this study to be used instead of conventional KEYWORDS
grout in the connection to reduce the anchorage length. First, fourteen large-scale pullout tests were Ultra-high-performance
concrete; bond; duct;
performed to investigate the bond strength of UHPC-filled duct connections. The test results showed that development length,
the incorporation of UHPC instead of normal grout as duct filler can reduce the bar embedment length by accelerated bridge
50% or more. Second, design equations were developed based on the pullout test data to compute the bar construction; precast bridge
development length in these connections. Subsequently, two half-scale precast bridge columns utilising columns
this connection type at the column base were tested to verify the design equations. Full column plastic
moment was developed in both specimens under cyclic loading with no UHPC-filled duct connection
failure such as bar pullout even under 12% drift ratio cycles. Finally, a guideline was proposed based on
the pullout and column test data for the design and construction of the UHPC-filled duct connections to
facilitate field deployment.

1. Introduction
& Kreger, 2001). Epoxy-coated Ø36-mm (No. 11) bars were
To reduce traffic impact and onsite construction time and to embedded 8.5db, 12.8db, and 17db in 102-mm (4-in.) diameter
improve total project delivery time and material quality, accel- corrugated steel ducts filled with standard grout. The bars with
erated bridge construction (ABC) has been gaining substantial 8.5db embedment length exhibited pullout failure due to split-
momentum in the United States. Most of the bridge components ting cracks in the concrete around the duct and pullout of the
used in ABC are designed to be precast to reduce onsite activities bar-grout from the duct. Normalised bond strength is defined as
that impact traffic. The quality of materials can also be enhanced the ratio of the bond strength to the square root of compressive
when components are prefabricated in a controlled environment. strength of concrete. The mean normalised bond strength for
In addition, safety of highway workers and travelling public will straight bars was 1.41 MPa0.5 (17 psi0.5) with a standard deviation
be improved by offsite precasting (Culmo, 2011). of 0.25 MPa0.5 (3 psi0.5). Ducts increased the level of confinement
Grout-filled duct (GD) connections, which was originally in the duct system resulting in 2.5 times higher normalised bond
developed to connect cast-in-place columns (CIPs) to bent caps, strength compared to conventional concrete calculated based on
are one of the suitable types of ABC connections for reinforced the ACI318-63 (1963).
concrete (RC) bridge columns in moderate and high seismic Thirty-two pullout specimens similar to those in the afore-
zones (Kapur et al., 2013; Marsh et al., 2011). In this connec- mentioned study were tested by Brenes, Wood, and Kreger
tion type, the extended column longitudinal reinforcing bars are (2006). Three types of 102-mm (4-in.) diameter ducts were
anchored in corrugated ducts placed in a cap beam filled with used: corrugated galvanised strip metal, corrugated high-
conventional grout (Figure 1). density polyethylene, and corrugated polypropylene. Bar pullout
A few experimental studies have focused on either bond was observed in all 32 specimens that had 8db, 12db and 16db
behaviour or seismic performance of this type of connection. embedment lengths. Grout failed then duct seam opened a few
Eight pullout tests were carried out to determine the bond inches below the surface in tests with galvanised steel ducts. The
strength of the GD system (Matsumoto, Waggoner, Sumen, test results showed that initial stiffness of bond-slip curves and

CONTACT Mostafa Tazarv [email protected]

© 2016 Informa UK Limited, trading as Taylor & Francis Group
744 M. Tazarv and M. S. Saiidi

Extended Column properties compared to conventional concrete such as four or

Reinforcing Steel five times higher compressive strength and 0.8% tensile strain
Cap Beam
Grouted Duct capacity, and has been utilised in several bridges in Canada and
B USA mainly to connect precast decks (Graybeal, 2010).
B
Precast Bond behaviour of UHPC-filled duct connections was investi-
Column gated in this paper through 14 large-scale pullout tests. Details of
A A the test specimens, matrix, set-up, and findings of the experiment
Column Section A-A Column Section B-B were presented. Design equations were developed to determine
the development length of the bars in this connection type. To
Figure 1. Grout-filled duct connections. verify the design equation, two precast columns incorporating
UHPC-filled duct connections at the column base were con-
bond strength of galvanised steel ducts were greater than those structed and tested under slow cyclic loading. Based on the find-
of the plastic ducts with the same embedment length. ings of the pullout and column tests, a guideline was proposed
Restrepo, Tobolski, and Matsumoto (2011) performed a series for the design and construction of UHPC-filled duct connections
of 42%-scale column to cap beam connection tests incorporat- to facilitate field deployment.
ing different ABC techniques suitable for seismic zones. In one
specimen a GD connection was utilised. Closely spaced 45-mm
2. Experimental programme
(1.75-in) diameter corrugated ducts were used in the bent cap
filled with high-strength, non-shrink, cementitious grout to To determine bond strength of deformed bars in UHPC-filled
anchor the column longitudinal reinforcement. A displacement duct connections, 14 specimens were tested at the Large Scale
ductility capacity of eight was achieved in the GD model that Structural Laboratory at the University of Nevada Reno (UNR).
was 80% of that of the cast-in-place model. Effect of several parameters on bond behaviour was investigated:
Seventeen pullout tests of large-diameter bars embedded in embedment length, bar size, duct diameter, number of ducts,
grout-filled pipes and ducts were carried out by Steuck, Pang, and bundling of bars. Other parameters, which are important in
Eberhard, and Stanton (2008) and Steuck, Eberhard, and Stanton conventional bond studies such as confinement effect or concrete
(2009). The embedment length ranged from 2 to 14 bar diameter. cover, were not investigated in this study since bars in UHPC-
The bar size ranged from Ø25 mm (No. 8) to Ø57 mm (No. 18). filled duct connections are typically anchored in well-confined
Twenty-centimetre (eight-inch)-diameter corrugated steel pipes footings or cap beams with usually large covers. The test matrix
were used in 15 of specimens, whereas 102-mm (4-in.)-diameter is presented in Table 1. Straight deformed Ø25-mm (No. 8) and
corrugated steel ducts were used in the two remaining speci- Ø36-mm (No. 11) bars were used with embedment lengths of
mens. Note that the pipes were approximately three times thicker 3db, 5db, 8db and 12db. Ducts with nominal size of 75 mm (3 in.),
and two times larger in diameter than the ducts. Bar pullout 100 mm (4 in.) and 125 mm (5 in.) were used.
with some grout attached was observed in specimens with short Table 1 also presents the nominal and actual inner diameters
embedment length. Either grout or bar fracture was observed in (dd) of the ducts used in this study. Corrugated galvanised strip
specimens with longer embedment lengths. metal ducts conforming to ASTM A653/A653M-11 (2011) with
The test results showed that adding fibres to the grout have 26-gauge (0.018-in. or 0.46-mm) wall thickness were used. In
minor effect on the bond behaviour. It was reported that the three of the specimens, two Ø25-mm (No. 8) bars were bundled.
embedment length of 6db in the grout-filled pipe connection The effect of using double ducts spaced at a clear distance of one
is sufficient to yield large-diameter bars and 14db embedment duct diameter, each with a single bar, was investigated in two of
length is sufficient to fully anchor the bars to fracture. The nor- the specimens. The different embedment lengths and the duct
malised bond strength of bar in this connection was approx- material were selected from the findings of the previous studies
imately four times the conventional concrete bond strength. on grouted duct systems such as Brenes et al. (2006).
Pang, Steuck, Cohagen, Stanton, and Eberhard (2008) utilised Several modes of failure are possible in GD connections: (1)
a grout-filled pipe/duct connection in a 40%-scale column-cap bar fracture, (2) bar pullout caused by bond failure, (3) grout
beam connection to investigate the seismic performance. The test mass pullout from the duct, (4) duct pullout by either failure of
results showed that the precast connection behaved the same as bond between concrete surrounding the duct and the duct or
a typical cast-in-place connection. conical failure of concrete, and (5) bar or duct pullout by splitting
Cap beams are usually sufficiently deep to accommodate col- failure of the either surrounding grout or surrounding concrete,
umn longitudinal bars in GDs. One challenge when using GDs respectively (Figure 2). It is obvious that there are two critical
in RC bridge column-to-footing connections is that a typical bond surfaces in this type of connections. One is the bar–grout
footing thickness may not always be sufficient to develop the interface and the other is the grout-duct–concrete interface. The
unhooked reinforcement to develop column base moments. In former was used in all the previous bond studies to obtain ‘bar
the present study, ultra-high-performance concrete (UHPC) was bond strength’ (ub,max) determined as the ratio of the peak ten-
used as duct filler instead of standard grout. UHPC is a type sile force to the surface area of the bar. This approach does not
of cementitious material, a mix of very fine aggregates such as address directly the duct–concrete interface failure. Therefore,
Portland cement, fine sand, silica fume, and ground quartz, very a new term was introduced in this study as ‘duct bond strength’
low water-to-cement ratio, superplasticiser, and steel fibres. The (ud,max), which is the ratio of the peak tensile force to the surface
volumetric ratio of steel fibres is 2%. UHPC exhibits superior area of the duct.
 Structure and Infrastructure Engineering 745

The test specimens were designed in two groups to determine

Note: 1 in. = 25.4 mm; 1 kip = 4.45 kN; 1 psi = 0.00689 MPa; A No. 8 bar has 25-mm diameter; A No. 11 bar has 36-mm diameter; ‘Diam.’ or ‘D’ is an acronym for diameter; ‘SP’ is an acronym for specimen; ‘H’ is the height of the specimen;
the duct bond strength in group I (10 specimens) and the bar

Strength,
Bar Bond

ub,max, psi

5967*

3568*

5920*
4162
N/A
N/A

N/A
N/A

N/A

N/A
N/A
N/A
N/A

N/A
bond strength in group II (4 specimens). Figure 3 depicts a sche-
matic view of the specimens. Both groups had similar geometries
but a large portion of the bars were debonded using 9.5-mm
(3/8-in.)-thick pipe insulation foams in group II. With this detail
Duct Bond
Strength,
ud,max, psi

the entire length of the duct was mobilised in transferring the

720*
467*

712*

544*

700*
435*

500*

595*
N/A

N/A

N/A

N/A
671

662
loads to concrete, and duct pullout was avoided.
All specimens were lightly reinforced longitudinally with
6-Ø10 mm (No. 3) bars and transversely with Ø10-mm (No. 3)
Mode of Failure

spiral at 51-mm (2-in.) pitch. Clear cover was 25 mm (1 in.) in all

Duct Pullout

Duct Pullout
Bar Fracture
Bar Fracture
Bar Fracture

Bar Fracture

Bar Fracture
Bar Fracture
Bar Fracture
Bar Fracture

Bar Fracture
Bar Fracture
Bar Fracture
specimens. Gravity tremie-tube method was used to fill the ducts
Bar Pullout

to avoid air trapping. In this method, a PVC pipe was inserted

into the duct then UHPC was continuously poured from the top
passing through a funnel. At the same time, the pipe was slowly
pulled out, thus allowing the material to fill the duct by its own
Force, kips
56.4
55.0
56.2
67.5
107.7
78.4
54.7
56.0
109.8
102.4
93.9
106.4
110.8
110.9

weight. After filling the ducts with UHPC, anchoring bars were
inserted into the ducts then covered for curing.
There were several test configurations depending on the
bar sizes, the number of ducts, and whether or not the bars
Embed. Length
Embed. Length
Embed. Length

Embed. Length

Embed. Length
Embed. Length
Embed. Length
Test Variable

Bundled Bars
Bundled Bars
Bundled Bars

were bundled. To accommodate these different configura-

Multi Ducts
Multi Ducts
Duct Diam.

Duct Diam.

tions, a self-reacting pullout test set-up was designed and

built at UNR with a pull force capacity of 1335 kN (300 kips).
Figure 4 displays photographs of the test set-up. A 25-mm
(25-in.)-thick steel ring was placed on the surface of the spec-
SP. Size, D. H

24 by 16
24 by 20
24 by 20
24 by 16
24 by 20
24 by 20
24 by 16
24 by 20
24 by 16
30 by 20
30 by 16
30 by 20
30 by 28
30 by 16
in. × in.

imen to uniformly distribute the load on the specimen edge.

Two parallel steel I-beams with a depth of 762 mm (30 in.)
were placed on the top of the steel bearing ring. Another 610-
mm (24-in.)-deep I-beam was used on the top of the bottom
No. of Ducts

I-beams in the perpendicular direction to spread the load. An

889.6-kN (200-kip) hollow-core jack (ram) was installed on
1
1
1
1
1
1
1
1
2
2
1
1
1
1

top to apply the load. Anchored bar/s with threads at the top
end (Figure 3) was bolted to the bottom plate of the connec-
tion. Four high-strength bolts connected the bottom and top
Diam., dd, in.
(Exact) Duct
Nominal

plates. A large high-strength threaded rod attached the top

3 (3.12)
3 (3.12)
3 (3.12)

3 (3.12)
3 (3.12)
3 (3.12)

5 (5.26)
4 (4)
4 (4)
4 (4)
4 (4)

4 (4)
4 (4)
4 (4)

plate to the ram. The ram was controlled manually by an oil

pump with a slow pumping rate resulting in a displacement
rate of 0.076 mm/s (0.003 in/s) on average.
No. of Bars

Strain gauges were installed on opposite faces of bars and

ducts at a depth of 6db from the surface of the group I specimens.
1
1
1
2
2
2
1
1
2
2
1
1
1
1

Since embedment length was very short in group II specimens,

strain gauges were only installed on ducts. However, two strain
Investigating effect of a pair of bars (bundled bars) in a single duct
Embed. Length

gauges were installed on all bars at the height of 6db from speci-
men surface to measure strain of the free bars. Slippage of ducts
12db

12db

12db

12db
8db

3db
8db

3db
8db
5db
8db

8db

3db
8db

was measured by four position transducers, of which two were

installed on the concrete and two on UHPC (Figure 4). Only
Investigating effect of multiple ducts in a specimen.

the average measured displacements are reported in subsequent

sections. More information regarding test set-up and instrumen-
and ‘Embed.’ is an acronym for embedment.
Group No.

tation can be found in Tazarv and Saiidi (2014).

II

II

II

II
I
I

I
I

I
I
I
I

Grade 420 (60) ASTM A615/A615M-09b (2009) bars were

Table 1. Pullout test matrix and results.

used. The measured yield and ultimate strengths of anchored

bars were 517 MPa (75 ksi) and 766.7 MPa (111.2 ksi), respec-
Lower bound bond strength.
Bar Size

2 No. 8
2 No. 8
2 No. 8

2 No. 8
2 No. 8
No. 11
No. 11
No. 11
No. 11

tively. The test day compressive strength of normal concrete and

No. 8
No. 8
No. 8

No. 8
No. 8

UHPC is presented in Table 2. The test day compressive strength

of UHPC was 2.6 to 3.9 times that of conventional concrete.
A study showed that 76-mm (3-in.) cylinder specimen is an
Specimen No.

acceptable alternative to the standard 102-mm (4-in.) cylinder

specimen for compressive testing of UHPC (Graybeal & Davis,
10(b)

2008). Therefore, plastic cylinders with 76-mm (3-in.) diameter

9(b)
4(a)
5(a)
(a)

11
12
13
14

(b)
(a)
1
2
3

6
7
8

*
746 M. Tazarv and M. S. Saiidi

Bar
Duct

Concrete Plan View

Pullout Specimen Bar Fracture Bar Pullout, Bar Pullout, Duct Pullout Duct Pullout Duct/Bar Pullout
Surrounding Grout Mass Surrounding Concrete Conical Splitting Failure
Grout Failure Failure Concrete Failure Failure

Figure 2. Modes of failure in grout-filled duct connections.

Threaded Threaded (3.5 in.) (Figure 5). No damage was observed in UHPC. Only
Area Area in one specimen in group II with 3db embedment length, SP6,
Normal Mild Normal Mild bundled bars pulled out with no visible concrete and UHPC
Concrete Bar Concrete Bar damage. To understand the failure, the specimen was vertically
cut in half (Figure 6). It was observed that both plastic defor-
mation of ribs and crushing of surrounding UHPC next to the
Debonding ribs are the causes of the bar pullout. Note that only four ribs
by Foam were engaged with UHPC in this specimen. In other specimens,
the bar ruptured at the threads after large forces that generally
exceeded the bar yield force were reached (Table 1). The damage
UHPC-Filled Duct UHPC-Filled Duct in the majority of these specimens was limited to some surface
cracks in the concrete.
Figure 3. Schematic view of group I (left) and group II (right) pullout test specimens.
3.2. Bar and duct slippage
were used for the UHPC sampling. The top and bottom of the
Force versus duct slip for all specimens is shown in Figure 7.
cylinders were cut and ground before the compressive testing.
The duct slip was measured at the concrete surface using four
displacement transducers, but only the average of the test data
3. Experimental results and discussion is shown. It can be seen that the bar fracture occurred at small
duct slippage, and resulted in an approximately linear force–duct
3.1. Observed damage
slip behaviour. However, in the specimens in which the duct
In two of the specimens in group I, SP4 and SP11, the ducts pulled out, a complete duct bond–slip relationship was measured
pulled out with conical failure of concrete at a depth of 89 mm and was shown in thick solid lines. Bar slippage was measured

Figure 4. Pullout test setup (left), close-up of connection (middle), and instrumentation (right).
 Structure and Infrastructure Engineering 747

Table 2. Test-day strength of concrete and UHCP in pullout tests.

Specimen No. Test Group No. Concrete, MPa (psi) UHPC, MPa (psi)
1 I 46.3 (6720) 167.5 (24296)
2 I 46.7 (6778) 178.8 (25928)
3 II 51.5 (7471) 146.9 (21307)
4 I 46.6 (6755) 171.4 (24863)
5 I 46.7 (6770) 176.3 (25573)
6 II 51.8 (7506) 147.2 (21356)
7 I 46.7 (6778) 178.8 (25928)
8 II 51.8 (7506) 147.2 (21356)
9 I 46.8 (6786) 181.2 (26284)
10 I 47.4 (6881) 184.6 (26776)
11 I 47.4 (6881) 184.6 (26776)
12 I 48.5 (7039) 168.6 (24457)
13 II 50.9 (7387) 134.1 (19444)
14 I 48.5 (7039) 168.6 (24457)
Note: At least three samples were used for testing. Only average of test data was reported.

Figure 5. Damage of specimen SP4 (left) and specimen SP11 (right).

Duct
UHPC

Bonded
Area

Concrete

Figure 6. Damage of specimen SP6 in a vertical cut-in-half view.

Duct Slip (mm) indirectly by subtracting the bottom plate displacement, bar
0 2 4 6 8 10 12 14
elongation, and duct slippage when the bar above the concrete
120
SP1
SP4
SP2
SP5
SP3
SP6 500 surface remained elastic. Only in SP6, this assumption was valid
100 SP7
SP9-Left
SP8
SP10-Right
SP9-Right
SP10-Left for the entire test while the threaded portion of the bars yielded
SP11 SP12 SP13 400
80
SP14a SP14b in the other group II specimens. The force–slip relationship for
Force (kips)

Force (kN)

300 the bars of SP6 is shown in Figure 8. It can be seen that even with
SP11
60 very short embedment length, 3db, and using bundled bars, a
SP3
40
200 force of 66% of the yield force of the bars was developed in SP6
before the bundled bars pulled out.
20 100
SP4

0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 3.3. Bar and duct bond strengths
Duct Slip (in)
The bar and duct bond strengths are presented in Table 1.
Figure 7. Force vs. duct-slip relationship for all pullout tests. Because in the majority of the tests, bar fractured at the threads
748 M. Tazarv and M. S. Saiidi

Bar Slip (mm) 3.4. Effect of test parameters

0 10 20 30 40 50
400 In group I where duct pullout was desired, the duct bond stress
80 SP6 (ud) was normalised √ to the square root of test day compressive
300 strength of concrete ( fc′ ). In group II in which bar pullout was
desired, the bar bond stress (ub) was normalised √ to the square
Force (kips)

60

Force (kN)
200
root of test day compressive strength of UHPC ( fUHPC ′
). The
40 effect of bar bundling on the duct bond behaviour was investi-
100
gated by comparing the normalised duct bond–slip curves of SP1
20
and SP4 (Figure 9). It can be inferred that the effect of bundling
0 0
is minor on the duct bond behaviour. The effect of bar size on the
0 0.5 1 1.5 2 bond behaviour is shown in Figure 10, which indicates the effect
Bar Slip (in) is negligible. Multiple duct effect could not be determined since
the bars fractured at the threads at small slippage (Figure 11).
Figure 8. Force vs. bar-slip relationship for specimen SP6. The duct slippage was divided by the duct diameter to deter-
mine the effect of duct size (Figure 12). It can be seen that the
duct size had minor effect on the initial duct bond behaviour.
10
Single #8 Bar SP 1 0.8 However, it should be noted that duct diameter is in the denomi-
9 SP 4 nator of the duct bond strength, thus it has a significant effect on
Normalised Duct Bond Stress,

0.7
8
0.6
the bond strength. For example, even though pull force capacity
7
of SP11 was only 85% of that of SP14, the normalised duct bond
ud/ f'c (MPa 0.5)
ud/ f'c (psi 0.5)

6 0.5
strength of SP11 was 13% higher than SP14 due to 24% smaller
5 0.4 duct diameter. In other words, the larger the duct diameter the
4
0.3 lower the duct bond strength. Similar trend exists in conven-
3 Bundled#8 Bars
0.2
tional bar bond strength where larger bars tend to have lower
2 bond strength (Soroushian & Choi, 1989).
1 0.1

0 0
0 0.01 0.02 0.03 4. Design equation for UHPC-filled duct connections
Normalised Duct Slip (Slipduct/Lemb)
4.1. Methodology
Figure 9. Effect of bar bundling on bond-slip relationship of pullout tests. Design bar development length equation for the UHPC-filled
duct connections should include the duct and bar bond strengths.
That is the normalised duct and bar bond strengths were utilised
10
#8 Bar, 8db SP 1
0.8 in design equation derivation. The average normalised measured
9
duct bond strength of group I specimens, excluding SP2, SP7
Normalised Duct Bond Stress,

SP 11 0.7
8
0.6
and SP10, was 0.65 MPa0.5 (7.88 psi0.5) with a standard devia-
7
tion of 0.085 MPa0.5 (1.02 psi0.5) (Figure 13). The design duct
ud/ f'c (MPa 0.5)
ud/ f'c (psi 0.5)

6 0.5
#11 Bar, 8db
bond strength was defined as the average normalised duct bond
5 0.4 strength less one standard deviation leading to 0.57 MPa0.5 (6.86
4
0.3 psi0.5), which is a lower bound bond strength to increase the level
3
0.2
of the safety in design. The coefficient was rounded down to
2 0.56 MPa0.5 (6.75 psi0.5). The average and the design duct bond
0.1
1 strengths are shown in Figure 13 with solid back lines. The design
0 0 embedment length is calculated using the design bond strength.
0 0.01 0.02 0.03 0.04 0.05
There was only three acceptable test data regarding the bar
Normalised Duct Slip (Slipduct/Lemb)
bond strength that could be used from this study. To enrich the
Figure 10. Effect of bar size on bond-slip relationship of pullout tests.
database, three pullout tests of bars anchored in UHPC cylinders
[diameter of 400 mm (15.74 in.) and height of 200 mm (7.87 in.)]
carried out by Graybeal (2010) were also used (Table 3). It should
be noted that no duct was used in Graybeal’s study. The aver-
instead of bar/duct pullout, the values that are listed are the age normalised bar bond strength of six tests was 2.89 MPa0.5
lower bound bond strength marked with an asterisk in the (34.9 psi0.5) with a standard deviation of 0.46 MPa0.5 (5.56 psi0.5)
table. Some of specimens with the longer embedment length, (Figure 14). Similar to what was done for the design duct bond
SP2, SP7, SP8 and SP10, are excluded from further analyses strength, the design bar bond strength was recommended to
since similar specimens with shorter embedment length had the be the average less one standard deviation. The coefficient was
same mode of failure of bar fracture at the thread. The average rounded to 2.49 MPa0.5 (30 psi0.5). The average and the design
bar bond strength of group II specimens (excluding SP8) was bar bond strengths are shown in Figure 14 with solid black lines.
36.9 MPa (5350 psi), which is approximately eight times that Since the normalised bar bond strength in Graybeal’s tests
of conventional concrete. (specimens without any transverse steel or duct confinement)
 Structure and Infrastructure Engineering 749

10
0.8 was only 12% lower than that of the UHPC-filled duct system
9 tests on average, it can be inferred that the bar bond strength in
Normalised Duct Bond Stress, 0.7
8 UHPC is inherently high (approximately eight times of conven-
7 0.6 tional concrete bond strength), and the effect of confinement

ud/ f'c (MPa 0.5)

ud/ f'c (psi 0.5)

6 0.5 by spiral and duct was not the reason of achieving relatively
5 0.4 high bar bond strength in the UHPC-filled duct connections.
4 Nevertheless, external confinements may improve UHPC bond
0.3
3 strength (12% increase in bond strength observed in the pres-
2 SP 1
0.2 ent study may be attributed to confinement effects) but further
1
SP9-Right Duct 0.1 study is needed. It was reported by Kook, Shin, Kwahk, and Yoon
SP9-Left Duct
0 0 (2010), Lee and Lee (2015), Marchand et al. (2015), and Lagier,
0 0.002 0.004 0.006 0.008 Massicotte, and Charron (2015) that the bar bond strength of
Normalised Duct Slip (Slipduct /Lemb) UHPC (without ducts) is 5 to 10 times that of conventional con-
crete, which supports the findings of the present study.
Figure 11. Effect of multiple ducts on bond-slip relationship of pullout tests. Based on the design duct and bar bond strengths, the bar
development length in UHPC-filled duct connections (Lemd) to
10
fracture the bar is the greater of Equations (1) and (2):
0.8
9 SP 11
US Customary Units SI Units
SP14
Normalised Duct Bond Stress,

0.7
8
0.6 db2 .fs db2 .fs
7 Ld,duct = Ld,duct = (1)
ud/ f'c (MPa 0.5)

√ √
ud/ f'c (psi 0.5)

6 0.5 27dd . fc� 2.24dd . fc�

4 in. Duct, 8db
5 0.4
4 db .fs db .fs
5 in. Duct, 8db 0.3 Ld,bar = Ld,bar =
3
√
�
√
�
(2)
0.2
120 fUHPC 9.96 fUHPC
2
1 0.1
where Lemd is the development length of a straight deformed
0 0
bar in UHPC-filled duct connections. It is expected that the duct
0 0.05 0.1 0.15
bond strength controls the embedment length in typical cases
Normalised Duct Slip (Slipduct /dd)
(Equation (1)) due to the lower strength compared to the bar
Figure 12. Effect of duct size on bond-slip relationship of pullout tests. bond strength. In other words, the critical bond surface in the
UHPC-filled duct connections is the duct-concrete interface. The
superior mechanical properties of UHPC as well as very high
bond strength of the bar in UHPC made the bar-duct component
to behave as a large diameter reinforcing element that is anchored
in concrete. Note that Equation (2) can be used to determine the
embedment length of straight deformed bars in UHPC blocks
such as UHPC-footings.

4.2. Comparison with codes and previous studies

The embedment length can be substantially reduced when
UHPC-filled duct connections are used. Figure 15 illustrates
bar stress versus bar development length for different design
equations. Bar was assumed to be Ø25 mm (No. 8) with an ulti-
Figure 13. Normalised and design duct bond strengths.
mate strength of 758 MPa (110 ksi). The compressive strength

Table 3. Bond strength of bars in FHWA pullout tests (Graybeal, 2010) [With Author Permission].

Force, Strength of UHPC, Duct Bond Strength, ud,max, Bar Bond Strength, ub,max,
U.S. Bar Size Embed. Length(a) Mode of
Sample ID (mm) mm (in.) kN (kips) MPa (psi) MPa (psi) MPa (psi) Failure
T-13-B No. 4 (Ø13) 75 89.9 103.4 No ducts was used 29.6* Bar Fracture
(2.95) (20.2) (15000) (4291.5)
T-16-B No. 5 (Ø16) 100 144.5 103.4 No ducts was used 28.5* Bar Fracture
(3.93) (32.5) (15000) (4134.7)
T-19-B No. 6 (Ø19) 125 205.5 103.4 No ducts was used 27.0* Bar Fracture
(4.92) (46.2) (15000) (3922.2)
(a)
Bars were anchored in 400-mm × 200-mm (15.74-in. × 7.87-in.) UHPC Cylinders. No ducts and no transverse reinforcements were used. ‘Embed.’ is an acronym for
‘Embedment’.
*
Lower bound bond strength.
750 M. Tazarv and M. S. Saiidi

of concrete and UHPC were assumed to be 34.5 MPa (5000 psi)

and 137.9 MPa (20,000 psi), respectively. All the ACI318-11
(2011) development length factors were assumed to be one. For
Equation (1), three graphs are shown that are for different duct-
to-bar diameter ratios of 2.83, 3.12 and 4.0.
To rupture the bar in a 101-mm (4-in.)-diameter duct, the bar
needs to be embedded 381 mm (15 in. or 15db) into the UHPC-
s

filled duct connection whereas the bar development lengths

in the GD (Brenes et al., 2006) and conventional connections
(ACI318-11 or AASHTO, 2012) are 889 mm (35 in. or 35db)
and 1194 mm (47 in. or 47db), respectively. Therefore, more than
50% reduction of embedment length can be achieved in this
Figure 14. Normalised and design bar bond strengths.
connection compared to other anchorage systems. Note that the
other studies are independent of the duct diameter. The embed-
ment length in grout-filled pipe system (Steuck et al., 2008) using
48.3-MPa (7000 psi) grout needs to be 14db to rupture the bar. It
should be noted that the design equation developed for grout-
filled pipe system was based on 17 tests, 15 of which had 203-mm
(8-in.)-diameter pipes. As shown in the present study, the duct
size has a significant effect on the bond strength. That is, the
embedment length in a UHPC-filled duct connection with a 203-
mm (8-in.) duct is 7db, which is 50% less than that of the grout-
filled pipe system. Overall, substantial reduction of embedment
length can be achieved by using UHPC-filled duct connections
to be incorporated either in shallow cap beams or footings.

5. Application in bridge column connections

Figure 15. Bar stress vs. normalised development length.
Two half-scale precast bridge columns were constructed and
tested at UNR to verify the proposed design equations. One
D column, entitled as “PNC”, was built with conventional materi-
Transverse als in the plastic hinge zone and the other, referred to as “HCS”,
Reinforcement was constructed with different materials such as shape memory
Debonded

Precast
Longitudinal alloy and engineered cementitious composite (ECC) in the plas-
Length

Column
Reinforcement tic hinge zone. UHPC-filled duct connections shown in Figure 16
4db were utilised to connect these columns to the footings. Table 4
Footing presents the measured test-day material strength and Table 5
Extended 4db
Column describes the column properties. A CIP model with similar geom-
Corrugated Reinforcing Lemd etry and properties tested by Haber, Saiidi, and Sanders (2014)
Galvanised Bars
Duct w/ UHPC
was selected as a reference column. The straight bar embedment
length in the PNC and HCS footing ducts was respectively 50
and 0.5% larger than the required embedment length (based on
dd the greater of Equations (1) and (2) using the test-day measured
strength of materials). These embedment lengths were only 85%
Figure 16. UHPC-filled duct connections for bridge columns. of the straight portion of the 90-degree hooked bars anchored in
the footing of CIP. All columns were tested under similar slow

Table 4. Measured strength of materials for column model tests, MPa (psi).
Material Element CIP PNC HCS
Conventional Concrete(a) Footing 37.3 (5415) 37.8 (5485) 47.4 (6880)
Conventional Concrete(a) Column 30.6 (4445) 22.7 (3290) 46.7 (6775)
Self-Consolidating Concrete(a) Core and Head N/A 65.6 (9510) 34.9 (5065)
UHPC(b) UHPC-Filled Duct N/A 158.4 (22970) 145.0 (21030)
ECC(c) Plastic Hinge N/A N/A 44.0 (6390)
Longitudinal Steel Bar Yield Strength 474.4 (68800) 453.4 (65800) 495.1 (71800)
Longitudinal Steel Bar Ultimate Strength 767.4 (111300) 632.0 (91700) 684.8 (99300)
a
152-mm (6-in.) by 305-mm (12-in.) cylinder.
b
76-mm (3-in.) by 152-mm (6-in.) cylinder.
c
102-mm (4-in.) by 203-mm (8-in.) cylinder.
 Structure and Infrastructure Engineering 751

Table 5. Column model and UHPC-filled duct connection details.

Property CIP PNC HCS
Diameter, m (in.) 0.61 (24) Same as CIP Same as CIP
Height, m (in.) 2.74 (108) Same as CIP Same as CIP
Longitudinal bars 11 Ø25-mm (No. 8) steel Same as CIP 10 Ø32-mm (No. 10) SMA
10 Ø36-mm (No. 11) steel
Transverse bars Ø10 mm (No. 3) at 51 mm (2 in.) Same as CIP Same as CIP
Exact inner duct diameter, mm (in.) N.A. 79 (3.12) 102 (4)
Required embedment length in N.A. 401 (15.8) 607 (23.9)
UHPC-filled ducts using measured
material strength, mm (in.)
Provided embedment length(a) and 711 (28), 90-degree hook 610 (24), straight bar 610 (24), straight bar
anchorage type, mm (in.)
UHPC-filled duct connection perfor- N.A. No damage after 10% drift ratio cycles No damage after 12% drift ratio cycles
mance
a
Excluding debonded length of bars in ducts.

80 350 test (Figure 17). The UHPC-filled duct connections incorpo-

Lateral Strength (kips)

70 rated in PNC and HCS respectively withstood 10 and 12% drift

Lateral Strength (kN)
300
60 Max AASHTO 250 ratios (the ratio of column lateral displacement to the column
= 5.0

50 Column = 7.36
Displacement 200 height) with no bar pullout from the ducts, with no duct pull-
40
30
Ductility Demand 150 out from the surrounding footing concrete, and with no conical
PNC Column
20 100 failure of the footing concrete (Figure 18). The mode of failure
HCS Column
10 CIP Column
50 was column longitudinal bar fracture and was the same as that
0 0 in CIP. The AASHTO Seismic Guide Specifications (AASHTO,
0 1 2 3 4 5 6 7 8 9 10 11
2014) limits the displacement ductility demand of columns to
Drift (%) 5.0, which is equivalent to a drift ratio of 6.75% for CIP. With the
drift ratio capacity of 10% or more in PNC and HCS, it can be
Figure 17. HCS, PNC and CIP column average force-drift envelopes (Tazarv and
Saiidi, 2015a) [With permission from ASCE]. seen that the UHPC duct connection provided at least 48% more
capacity than the maximum drift demand. Furthermore, the col-
umn test data confirmed that the proposed design equation for
reversed cyclic displacement-control loading to failure. Complete UHPC-filled duct connections is reliable since these connections
information regarding the test procedure, objectives, detailed showed excellent seismic performance under extreme inelastic
test results, and analytical studies for the PNC and HCS columns deformations.
can be found in Tazarv and Saiidi (2015a) and Tazarv and Saiidi
(2015b), respectively.
6. Recommended design guidelines
Even though the straight bar embedment length in the footing
of PNC and HCS was 15% lower than the CIP bar embedment Design and construction recommendations are presented herein
length, full plastic moment was developed in both precast col- to facilitate the field application of precast columns incorporating
umns and the plastic moment was maintained during the entire UHPC-filled duct connections:

Figure 18. Plastic hinge damage of (a) PNC and (b) HCS after testing.
752 M. Tazarv and M. S. Saiidi

R1 – UHPC-filled duct columns and adjoining members such are viable precast column connections for ABC in high
as footings and cap beams should be designed in accordance with seismic regions.
a legally adopted bridge code.
R2 – Three-day and 28-day compressive strength of UHPC
Notations
should not be less than 70 MPa (10,000 psi) and 140 MPa
(20,000 psi), respectively. D  the diameter (or the largest side dimension) of spec-
R3 – Only corrugated galvanised strip metal ducts conform- imen, (in., mm)
ing to ASTM A653 should be used. H the height of specimen, (in., or mm)
R4 – Duct diameter (dd) should not be less than three column
Ld,bar  the bar development length based on bar bond
longitudinal bar diameter (3db).
strength, (in., mm)
R5 – Duct wall thickness should not be less than 0.46 mm
(0.018 in.). Ld,duct  the bar development length based on duct bond
R6 – Development length (Lemd) of column bars anchored in strength, (in., mm)
adjoining members should be the greater of Equations (1) and Lemb  the development length of straight deformed bar in
(2). However, this length shall not be greater than the adjoin- UHPC-filled duct, (in., mm)
ing member depth less the required cover concrete for those db the bar diameter, (in., mm)
members. dd the inner diameter of duct, (in., mm)
R7 – A length of four column longitudinal bar diameter (4db) f ’c the compressive strength of concrete, (psi, MPa)
above and below the column–footing interface shall be debonded f ’UHPC the compressive strength of UHPC, (psi, MPa)
with an appropriate debonding method. fy the bar yield strength, (psi, MPa)
fu. the bar ultimate strength, (psi, MPa)
fs  the bar stress, (psi, MPa); 1.5fy or fu whichever is
7. Summary and Conclusions greater
To be able to deploy grouted duct connections in shallow cap ub the bar bond stress, (psi, MPa)
beams and footings, UHPC was proposed to be used instead of ud the duct bond stress, (psi, MPa)
conventional grout. Fourteen pullout tests were carried out to
determine the bond strength of UHPC-filled duct connections,
and two precast columns utilising UHPC-filled duct connections
at the column base were tested. A summary of the findings of the Acknowledgements
experimental study is as follows: Special thanks are due Dr. Saad El-Azazy and Dr. Charles Sikorsky, the
Caltrans Research Program Managers for their support and advice. The
(1)  The bond strength of bars embedded in UHPC is
interest and comments of Mr. Mike Keever, Dr. Mark Mahan, and Mr. Ron
approximately eight times that of bars anchored in Bromenschenkel are also much appreciated. The authors are indebted to
conventional concrete. Dr. Patrick Laplace and Mr. Mark Lattin of UNR for their assistance in
(2)  Using UHPC-filled duct connections results in at execution of the tests. The authors would like thank Mr. Vic Perry, for-
least 50% reduction of the embedment length com- mer vice-president and general manager of Lafarge North America Inc.
for donating the UHPC material. The assistance of Mr. Peter Seibert and
pared to conventional construction.
Mr. Kyle Nachuk of Lafarge North America Inc. for mixing and casting of
(3)  The critical surface in UHPC-filled duct connections UHPC is also highly appreciated. Dr. Ali Mehrsoroush and Dr. Zachary
is the duct–concrete interface when the duct and bar Haber are thanked for their assistance in this study. The authors would
have the same embedment length. like thank Dr. Benjamin Graybeal of Federal Highway Administration for
(4)  The bar size had minor effect on the duct bond behav- sharing test data.
iour. Furthermore, bundling of bars had negligible
effect on the duct bond strength. The duct size had Disclosure statement
minor effect on the initial bond behaviour. However, No potential conflict of interest was reported by the authors.
the duct bond strength was reduced when the duct
size increased.
(5)  Both precast columns withstood large inelastic defor- Funding
mations without any UHPC-filled duct connection The study presented in this paper was funded by the California Department
failure. Actually, full plastic moment was developed of Transportation (Caltrans) through contract No. 65-A0372.
and was maintained during the entire test even under
drift ratios of 10% or more.
References
Overall, the proposed design equation to determine the AASHTO. (2012). AASHTO LRFD bridge design specifications. Washington,
bar development length in UHPC-filled duct connections DC: American Association of State Highway and Transportation
is satisfactory since: (1) it captures the important param- Officials. 505 pp.
eters affecting the bond strength, and (2) it was verified AASHTO. (2014). AASHTO guide specifications for LRFD seismic bridge
design. Washington, DC: American Association of State Highway and
through two half-scale bridge column tests. Design and Transportation Officials.
construction guidelines were proposed to facilitate the ACI318-63. (1963). Building code requirements for reinforced concrete.
field application of UHPC-filled duct connections, which Detroit, MI: American Concrete Institute. 144 pp.
 Structure and Infrastructure Engineering 753

ACI318-11. (2011). Building code requirements for reinforced Concrete. (2015). Bond behaviour of reinforcing bars in UHPFRC-experimental
Detroit, MI: American Concrete Institute, 503 pp. investigation. Materials and Structures, 49, 17 pp. doi:http://dx.doi.
ASTM A615/A615M-09b. (2009). Standard specification for deformed and org/10.1617/s11527-015-0628-0
plain carbon-steel bars for concrete reinforcement. West Conshohocken, Marsh, M. L., Wernli, M., Garrett, B. E., Stanton, J. F., Eberhard, M. O.,
PA: ASTM International, 6 pp. & Weinert, M. D. (2011). Application of accelerated bridge construction
ASTM A653/A653M-11. (2011). Standard specification for steel sheet, zinc- connections in moderate-to-high seismic regions (Report No. 698).
coated (galvanized) or zinc-iron alloy-coated (galvannealed) by the hot- Washington, DC: National Cooperative Highway Research Program
dip process. West Conshohocken, PA: ASTM International, 14 pp. (NCHRP), 324 pp.
Brenes, F. J., Wood, S. L., & Kreger, M. E. (2006). Anchorage requirements Matsumoto, E. E., Waggoner, M. C., Sumen, G., & Kreger, M. E. (2001).
for grouted vertical-duct connectors in precast bent cap systems (FHWA/ Development of a precast bent cap system (FHWA Report No. FHWA/
TX-06/0-4176-1). Center for Transportation Research, The University TX-0-1748-2). Center for Transportation Research, The University of
of Texas at Austin, 251 pp. Texas at Austin, 404 pp.
Culmo, M. P. (2011). Accelerated bridge construction-experience in design, Pang, J. B. K., Steuck, K. P.,Cohagen, L., Stanton, J. F. and Eberhard, M. O.
fabrication and erection of prefabricated bridge elements and systems (2008). Rapidly constructible large-bar precast bridge-bent seismic
(Federal Highway Administration Report No. FHWA-HIF-12-013). connection (Report No. WA-RD 684.2). Seattle, Washington: University
McLean, VA: Federal Highway Administration, 346 pp. of Washington, 255 pp.
Graybeal, B. (2010). Behavior of field-cast ultra-high performance concrete Restrepo, J. I., Tobolski, M. J., & Matsumoto, E. E. (2011). Development
bridge deck connections under cyclic and static structural loading of a precast bent cap system for seismic regions (NCHRP Report 681).
(Federal Highway Administration Report No. FHWA-HRT-11-023). Washington, DC. 166 pp.
McLean, VA: Federal Highway Administration, 116 pp. Soroushian, P., & Choi, K. B. (1989). Local bond of deformed bars with
Graybeal, B., & Davis, M. (2008). Cylinder or cube: Strength testing of 80 different diameters in confined concrete. ACI Structural Journal, 86,
to 200 MPa (11.6 to 29 ksi) ultra-high-performance fiber-reinforced 217–222.
concrete. ACI Materials Journal, 105, 603–609. Steuck, K. P., Eberhard, M. O., & Stanton, J. F. (2009). Anchorage of large-
Haber, Z. B., Saiidi, M. S., & Sanders, D. H. (2014). Seismic performance diameter reinforcing bars in ducts. ACI Structural Journal, 106, 506–
of precast columns with mechanically spliced column-footing 513.
connections. ACI Structural Journal, 111, 339–650. Steuck, K. P., Pang, J. B. K., Eberhard, M. O., & Stanton, J. F. (2008).
Kapur, J., Bardow, A., Dekelbab, W., Keever, M., Saiidi, M. S., Sletten, J., Anchorage of large-diameter reinforcing bars grouted into ducts (Report
Tobias, D., & Yen, W. P. (2013). Best practices regarding performance of No. WA-RD 684.1). Seattle, Washington: University of Washington, 149
abc connections in bridges subjected to multihazard and extreme events pp.
(National Cooperative Highway Research Program, NCHRP Project Tazarv, M., & Saiidi, M. S. (2014). Next generation of bridge columns
20–68A, Scan 11–02). Washington, DC, 96 pp. for accelerated bridge construction in high seismic zones (Report No.
Kook, K., Shin, H., Kwahk, I., & Yoon, Y. (2010). Bond characteristics of CCEER-14-06). Reno, Nevada: Center for Civil Engineering Earthquake
ultra high performance concrete. Journal of the Korea Concrete Institute, Research, Department of Civil and Environmental Engineering,
22, 753–760. University of Nevada, 400 pp.
Lagier, F., Massicotte, B., & Charron, J.P. (2015). Bond strength of tension Tazarv, M., & Saiidi, M. S. (2015a). UHPC-filled duct connections for
lap splice specimens in UHPFRC. Construction and Building Materials, accelerated bridge construction of RC columns in high seismic zones.
93, 84–94. Engineering Structures, 99, 413–422. doi:http://dx.doi.org/10.1016/
Lee, J. K., & Lee, S. H. (2015). Flexural behavior of ultra-high-performance j.engstruct.2015.05.018
fiber-reinforced concrete moment connection for precast concrete Tazarv, M., & Saiidi, M. S. (2015b). Low-damage precast columns for
decks. ACI Structural Journal, 112, 451–462. accelerated bridge construction in high seismic zones. ASCE Journal
Marchand, P., Baby, F., Khadour, A., Battesti, T., Rivillon, P., Quiertant, M., of Bridge Engineering, 21, 13 pp. doi:http://dx.doi.org/10.1061/(ASCE)
Nguyen, H. H., Genereux, G., Deveaud, J. P., Simon, A., & Toutlemonde, F. BE.1943-5592.0000806