Sie sind auf Seite 1von 11

See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/256486356

Determining Slipping Stress of Prestressing


Strands in Confined Sections
ARTICLE in ACI STRUCTURAL JOURNAL NOVEMBER 2012
Impact Factor: 0.96

CITATION

DOWNLOADS

VIEWS

72

147

2 AUTHORS:
Mohamed K. ElBatanouny

Paul Ziehl

Wiss Janney Elstner Associates Inc.

University of South Carolina

31 PUBLICATIONS 52 CITATIONS

82 PUBLICATIONS 213 CITATIONS

SEE PROFILE

SEE PROFILE

Available from: Mohamed K. ElBatanouny


Retrieved on: 17 August 2015

ACI STRUCTURAL JOURNAL

TECHNICAL PAPER

Title no. 109-S66

Determining Slipping Stress of Prestressing Strands in


Confined Sections
by Mohamed K. ElBatanouny and Paul H. Ziehl
Development length and slipping stress of prestressing strands
subjected to confining stress is not well-quantified and the appropriateness of the ACI 318-11 equation under such conditions can
be questioned. In 1992, a test was performed on nineteen 14 in.
(356 mm) square prestressed concrete piles with a clamping force
applied during testing under lateral load. The findings indicated
that the ACI 318-11 equation for development length of prestressing
strands may not be suitable when used for sections subjected to
confining stress. In this study, a modified equation that accounts for
the effect of concrete confinement is discussed and compared to the
published 1992 results and the ACI 318-11 equation. The moment
strength of the sections is also compared using moment-curvature
analysis by comparing three different slipping values: 1) those
obtained from experimental results; 2) the ACI 318-11 equation;
and 3) the modified equation.
Keywords: confining stress; development length; moment capacity; slipping.

INTRODUCTION
The use of precast, prestressed concrete piles in bridge
construction is common in the United States; however,
the performance of such units under seismic loading is
not entirely clear. The behavior of the connection between
prestressed piles and cast-in-place (CIP) reinforced concrete
caps is particularly not well-understood. Current South
Carolina Department of Transportation (SCDOT) connection details1,2 require the plain embedment of the pile into
the bent cap one pile diameter with a construction tolerance
of 6 in. (152 mm). Plain embedment requires no special
detailing to the pile end or the embedment region and no
special treatment of the pile surface, such as roughening
or grooving. The ductility and moment capacity of such
connections is of interest because this short embedment
length is often much less than the length required for development of the full tensile strength of the prestressing strands
within the embedded region.
Generally, the development length of prestressing strands
is calculated from ACI 318-11, Eq. (12-4).3 In the case of
piles embedded in CIP caps, the embedment length is usually
far less than the development length. Therefore, the strands
are predicted to slip at a level of stress less than their nominal
capacity. This stress is referred to as the slipping stress.
The ACI 318-11 equation was developed for the case of
superstructure elements not subjected to confining stress.
Therefore, the application of this equation to substructure
elements having significant confining stress may not be
appropriate. A pile embedded in a CIP cap is subjected to
the shrinkage of the confining concrete in the cap, which
creates confining stress (also known as clamping force)
on the pile, which serves to enhance the bond between
the prestressing strand and the surrounding concrete. This
leads to a decrease in the development length and an associated increase in the slipping stress of the prestressing
ACI Structural Journal/November-December 2012

strand.4,5 This effect became very apparent during the testing


of a series of precast concrete piles embedded in CIP bent
caps at the University of South Carolina Structures Laboratory.5 Because the embedment length of the piles was much
less than the development length of prestressing strands, the
strands were expected to slip prior to achieving the nominal
capacity. Significant differences were found between the
experimental results and those predicted by ACI 318-11,
Eq. (12-4).5
Shahawy and Issa4 discussed the findings of a significant
experimental investigation related to the effect of confinement from CIP caps to prestressed concrete piles with
emphasis on the resulting behavior under lateral load. The
results showed that the development length of prestressing
strands was enhanced due to confining stress. They concluded
that using the ACI 318-11 equation without consideration of
confinement will lead to very conservative values.
This study makes use of the experimental results reported
by Shahawy and Issa4 to investigate the appropriateness of
a potential modification to the ACI 318-11 equation.5 The
theoretical slipping stress calculated from the modified
equation and ACI 318-11, Eq. (12-4), are compared.
The calculation of the moment strength of piles in seismic
regions is a critical issue. A moment-curvature analysis6 was
performed to calculate the moment strength of the sections
using the modified equation and the ACI 318-11 equation.
These were then compared to calculated moments using
measured slipping stress values from the Shahawy and
Issa4 study. Furthermore, finite element models were created
to investigate the value and distribution of the confining
stress due to shrinkage for use in the modified equation.
RESEARCH SIGNIFICANCE
Several important investigations have addressed the suitability of ACI 318-11, Eq. (12-4), for development length7,8;
however, only a few have considered the effect of confinement.4,5 The ACI 318-11 equation for calculating the development length of prestressing strands was derived for
unconfined sections and does not account for the effect of
confinement. To account for the effect of confinement, a
potential modification to the ACI 318-11 equation is developed and described. The results from the modified equation
are compared to published experimental results that directly
addressed the effect of confinement for the development
length of prestressing strands.
ACI Structural Journal, V. 109, No. 6, November-December 2012.
MS No. S-2010-340.R1 received November 18, 2011, and reviewed under Institute
publication policies. Copyright 2012, American Concrete Institute. All rights
reserved, including the making of copies unless permission is obtained from the
copyright proprietors. Pertinent discussion including authors closure, if any, will be
published in the September-October 2013 ACI Structural Journal if the discussion is
received by May 1, 2013.

767

long and were cut from 80 ft (24.4 m) long prestressed


concrete piles. Number 5 (No. 16) gauge steel was used
as spiral reinforcement that varied in pitch depending on
location. The end sections of the full-length piles were
provided with more spiral reinforcement than the interior
sections (middle sections), as shown in Fig. 1. Four embedment lengths of 36, 42, 48, and 60 in. (914, 1067, 1219,
and 1524 mm) were used in this study. Cores with diameters
of 6 in. (152 mm) were taken from the specimens to determine the concrete compressive strength. A summary of the
experimental program is provided in Table 1.

Mohamed K. ElBatanouny is a Graduate Research Assistant in the Department of


Civil & Environmental Engineering at the University of South Carolina, Columbia,
SC. He received his BS from Helwan University, Cairo, Egypt, in 2008, and his MS
from the University of South Carolina in 2010.
ACI member Paul H. Ziehl is an Associate Professor in the Department of Civil &
Environmental Engineering at the University of South Carolina. He received his PhD
from the University of Texas at Austin, Austin, TX. He is a member of ACI Committees
335, Composite and Hybrid Structures, and 437, Strength Evaluation of Existing
Concrete Structures.

SUMMARY OF EXPERIMENTAL RESULTS4


Test specimens
Nineteen 14 in. (356 mm) square prestressed concrete pile
specimens with 8.5 in. (13 mm) diameter prestressing strands
were tested in this study. The specimens were 12 ft (3.66 m)

Test procedure and analytical study


The experimental investigation was designed to simulate the behavior of a CIP cap. A steel test frame was used
to restrain the pile cap against translation and rotation. An
initial test was conducted to assess the value of confining
stress exerted from the shrinkage of the cap. A CIP cap was
cast with a pile embedded in its center for the initial test. The
CIP cap had a cross section of 42 x 54 in. (1.1 x 1.4 m) with
a depth of 48 in. (1219 mm). The pile was instrumented with
vibrating wire strain gauges spaced at 12 in. (305 mm) along
the embedment length (48 in. [1219 mm]). A schematic of
this confining stress test is shown in Fig. 2. After 28 days, the
principal strains were measured with a resulting maximum
value of 245 me. Using a conservative value for the Youngs
modulus of 3.6 106 psi (24,800 MPa), the maximum
confining stress was calculated to be 880 psi (6.1 MPa).
The authors4 of this study assumed that this value
would approach zero at the ends of the embedment length
following a parabolic distribution of a maximum measured
value of 245 me at a depth of 30 in. (762 mm); therefore,

Fig. 1Details of test specimens.4


Table 1Details of test program4
Section

fc, ksi

Measured steel stress


at failure, ksi

fse*, ksi

Embedment length, in.

Transfer length, in.

Available flexural
bond length, in.

A-1E

End

7.10

256

162

36

26.9

9.10

A-2E

End

5.84

263

178

36

29.6

6.40

A-3I

Interior

6.59

254

161

36

26.9

9.10

A-4I

Interior

5.60

153

164

36

27.3

8.70

B-1E

End

6.70

262

173

42

28.8

13.2

B-2E

End

6.45

261

172

42

28.6

13.4

B-3E

End

5.98

257

169

42

28.1

13.9

B-4E

End

7.80

260

168

42

28.0

14.0

B-5E

End

6.48

263

174

42

29.0

13.0

B-6I

Interior

6.48

259

169

42

28.2

13.8

C-1E

End

6.96

260

168

48

28.0

20.0

C-2E

End

6.50

258

166

48

27.6

20.4

C-3I

Interior

7.76

262

170

48

28.3

19.7

C-4I

Interior

6.50

258

165

48

27.5

20.5

C-5I

Interior

6.50

260

170

48

28.4

19.6

C-6E

End

6.50

258

167

48

27.8

20.2

D-1E

End

7.20

262

169

60

28.2

31.8

D-2I

Interior

6.50

261

172

60

28.6

31.4

D-3E

End

6.50

260

170

60

28.3

31.7

Specimen number

Effective prestressing stress back calculated from Shahawy and Issa data.
Embedment length minus transfer length.
Notes: 1 in. = 25.4 mm; 1 ksi = 6.895 MPa.

768

ACI Structural Journal/November-December 2012

an average confining stress was computed to be 525 psi


(3.6 MPa). Using this average confining stress value as an
upper limit, a clamping force of 200 kips (888 kN) was
applied to the upper and lower faces of the embedment
length of the pile specimens to represent the confining stress.
This clamping force was applied using post-tensioning
thread bars, as shown in Fig. 3. The lateral faces of the pile
embedment length were not subjected to confinement. The
confining stress varied with embedment length, resulting in
397, 340, 298, and 238 psi (2.74, 2.34, 2.05, and 1.64 MPa)
for embedment lengths of 36, 42, 48, and 60 in. (914, 1067,
1219, and 1524 mm), respectively. The highest applied
confining stress value was taken to be 75% of the average
confining stress measured in the initial test.
A hydraulic jack placed at 6 ft (1.84 m) from the face of
the supporting frame was used to apply lateral load on the
piles in increments of 3 kips (13.3 kN) up to a load of 18 kips
(80.1 kN); thereafter, the load increments were much smaller
until failure was achieved. At each load step, cracks were
marked and displacements and strains were recorded. Details
of the test setup are shown in Fig. 3.
The piles were next analyzed using a nonlinear material model. The time-dependent effects due to load history,
temperature history, creep, shrinkage, and relaxation of
steel were considered in the computer program.4 The
program4 was used to calculate the structural response
through the elastic and inelastic range up to ultimate load.
At each load step, nonlinear equilibrium equations using the
displacement formulation of the finite element method were
derived for the geometry and material properties.4
Findings of Shahawy and Issa4
The effect of transverse reinforcement was examined and
transverse reinforcement was found to have a slight effect
in terms of the moment capacity of the piles. The ultimate moment of the piles cut from the end sections was
slightly higher than those of the piles cut from the middle
sections by approximately 6%, as shown in Table 2. The
experimental slipping stress of the prestressing strands was
determined by measuring the strain along the length of the
strand at various levels of load until failure. The measured
slipping stress determined by this method is presented in
Table 2. For development length, the embedment length
of the piles was compared to the theoretical development
length required to obtain the same slipping stress using
three different equations:
1. The ACI 318-11 equation3 for development length
of prestressing strands, which is shown in Eq. (1a). This
equation divides the development length into two parts:
1) transfer length; and (2) flexural bond length. In this equation, Ld is the development length (in.), which is equal to
the embedment depth of the pile; fse is the effective stress of
prestressing strand (psi); fps is the nominal flexural strength
of the prestressing strand (psi); and db is the nominal diameter of the prestressing strand (in.). The slipping stress can
be calculated by rearranging the ACI 318-11 equation, as
shown in Eq. (1b). The nominal flexural strength of the
prestressing strand, fps, is renamed as the slipping stress, fss
(psi), for clarity.

Ld =

f ps - fse
fse
db +
db
3000
1000

ACI Structural Journal/November-December 2012

(1a)

Fig. 2Confining stress test setup.4

Fig. 3Lateral loading test setup with applied clamping force.4


fss = 1000

Ld 2
+ f f ps
db 3 se

(1b)

2. A modification to Eq. (1a), as proposed by Shahawy and


Issa.4 The proposed modification incorporates an average
bond stress term in the second part of the equation, as shown
in Eq. (2a). The Shahawy and Issa4 equation is rearranged
as Eq. (2b), which can be used to calculate slipping stress.
The calculated average bond stress (psi) is uave, which can be
calculated using Eq. (3). In this equation, P is the resisting
steel strength based on the strand slipping stress at failure
(lb); T is the resisting concrete strength, which is assumed
to be zero at ultimate due to cracking (lb); le is the available
embedment length (in.); and db is the nominal strand diameter (in.).

Ld =

f ps - fse
fse
db +
db
3000
4uave

fss = 4uave

Ld 4uave
+ 1fse f ps
db 3000

uave =

(P - T )
ple db

(2a)

(2b)

(3)
769

Table 2Shahawy and Issa4 test results and calculated slipping stresses
Slipping stress, ksi

Specimen number

Embedment length, in.

Theoretical ultimate
moment, kip-in.

Measured ultimate
moment, kip-in.

Measured steel stress


at failure, ksi

Eq. (1b)

Eq. (2b)

Eq. (4b)

A-1E

36

1560

1840

256

180

256

167

A-2E

36

1460

1800

263

190

263

177

A-3I

36

1530

1550

254

180

254

166

A-4I

36

1440

1550

253

181

253

167

B-1E

42

1530

1620

262

199

262

182

B-2E

42

1520

1870

261

199

261

183

B-3E

42

1480

1760

257

197

257

180

B-4E

42

1600

1560

260

196

260

179

B-5E

42

1520

1840

263

200

263

185

B-6I

42

1520

1600

259

197

259

181

C-1E

48

1550

1510

260

208

260

190

C-2E

48

1520

1690

258

206

258

188

C-3I

48

1600

1760

262

209

262

190

C-4I

48

1520

1660

258

206

258

188

C-5I

48

1520

1690

260

210

260

192

C-6E

48

1520

1700

258

207

258

189

D-1E

60

1570

1730

262

233

261

210

D-2I

60

1520

1730

261

235

261

212

D-3E

60

1520

1620

260

233

260

210

Notes: 1 in. = 25.4 mm; 1 kip-in. = 0.11298 kN-m; 1 ksi = 6.895 MPa.

3. An equation proposed by Zia and Mostafa,8 as shown


in Eq. (4a). Their approach used the same parameters
used in calculating the development length of prestressing
strands, with the exception of two terms: fsi, which is the
stress in prestressing steel at transfer (ksi), and fc, which
is the compressive stress of concrete at the time of initial
prestressing (ksi). The effective stress of the prestressing
strand, fse, and the nominal flexural strength of the
prestressing strand, fps, should be used (ksi).
Ld =

1.5 fsi
db - 4.6 + 1.25 f ps - fse db
fc

(4a)

Ld
f
3.68
- 1.2 si +
+ fse f ps
db
fc
db

(4b)

fss = 0.8

Comparisons between these three approaches are


summarized in Table 2. The comparisons indicate that
the ACI 318-11 equation is conservative when confining
stress is applied to the concrete section. The Zia and
Mostafa8 proposed equation (Eq. (4b)) was more conservative than the ACI 318-11 equation. The Shahawy and
Issa4 proposed equation (Eq. (2b)) has a good match with the
experimental data. However, this equation uses the slipping
stress of the strand as an input. More detailed discussion is
provided in the following sections.
770

HISTORY OF ACI 318-11 EQUATION


The expression for the development length of prestressing
strands found in ACI 318-11 was proposed by Mattock9 and
members of ACI Committee 423.10 The expression divides
the development length into two parts: transfer length and
flexural bond length. To develop the expression for transfer
length,11 results of a study by Hanson and Kaar12 and Kaar et
al.13 were used. They stated a value for average transfer bond
stress ut = 400 psi (2.76 MPa). For flexural bond length,
another approach was used based on the definition of general
bond slip introduced by Janney.14 Janney14 stated that when
the peak of the high bond stress wave reaches the transfer
length, general bond slip occurs and leads to a reduction in
the frictional resistance resulting from the Hoyer effect.15
Hanson and Kaar12 agreed with Janneys14 explanation, but
they did not state a value for the average flexural bond stress.
Due to the difficulty of codifying this concept, Mattock9 and
the members of ACI Committee 42310 used the data of
Hanson and Kaars12 beam tests to formulate an approach
based on an average flexural bond stress. They constructed
a straight-line relationship by subtracting the estimated
transfer length from the embedment length of the strand. The
increase in strand stress due to flexure was determined to be
the difference between the strand stress at the load causing
slip and the effective stress due to prestressing. The use of
a constant slope for the flexural bond length implies a value
of average flexural bond stress ufb = 140 psi (0.96 MPa).5 It
is worth noting that the assumption of average flexural bond
stress was made to simplify the approach and makes it easier
to codify. The expressions for transfer length and flexural
bond length are shown in Eq. (5) and (6), respectively.
ACI Structural Journal/November-December 2012

Aps fse

fse
f
Lt =
=
db = se db
7.36 ut
3000
So ut

L fb =

(f

ps

- fse

)d = (f

7.36 u fb

ps

- fse

1000

)d

(5)

(6)

where Lt is transfer length (in.); Lfb is flexural bond length


(in.); So is the strand perimeter (So = 4/3 * p * db [in.]);
and Aps is the strand cross-sectional area (Aps = 0.725 * p *
db2/4 [in.2]).
EFFECT OF CONFINEMENT ON EFFECTIVE
TRANSFER LENGTH
Transfer length in the absence of confinement is a function
of diameter, effective prestress, and average transfer bond
stress. Mechanisms contributing to the value of average
transfer bond stress can be categorized into three groups:
adhesion, friction, and mechanical interlock.11 Adhesion
is destroyed by the relative slip between the strand and the
surrounding concrete and the contribution of mechanical
interlock in the average transfer bond stress can be neglected
due to unwinding.16
Frictional bond stress is developed as a result of the radial
compressive stresses, which are attributed to the Hoyer
effect,15 where longitudinal contraction results in radial
expansion of the tendon. This Poissons expansion induces
compression perpendicular to the steel-concrete interface. In
the absence of confinement, the value of the average transfer
bond stress is assumed to be ut = 400 psi (2.76 MPa). When
confinement occurs, it is convenient to represent the effect
as an increase in the apparent bond stress. It should be noted
that confinement does not change the transfer length itself.
Rather, confinement decreases the potential for slipping of
the strands within the transfer zone.
To account for this behavior, a new term referred to as the
effective transfer length (Lte) is proposed. This term takes
into account both the unconfined average transfer bond and
the increase in bond stress due to confinement. The value of
the resulting apparent bond stress is determined by adding
the average transfer bond stress (400 psi [2.76 MPa]) to the
average bond stress that is due to confinement. The average
bond stress due to confinement is calculated by multiplying
the confining stress by the coefficient of friction between the
steel and concrete (m). The resulting average bond stress is
shown in Eq. (7).
utc = 400 + m s cav

(7)

where utc is the average confined bond stress within the


transfer zone (psi); scav is the average confining stress
applied to the prestressed concrete section; and m is the coefficient of friction between the steel and concrete (generally
taken as m = 0.417).
EFFECT OF CONFINEMENT ON
FLEXURAL BOND LENGTH
For the confined flexural bond stress ufbc (psi), the same
approach was used, assuming that the confining stress would
only affect the friction stress. Due to the reduction in strand
diameter resulting from the increase in strand stress in the
ACI Structural Journal/November-December 2012

average flexural bond stress zone, the Hoyer effect15 is


reduced and ufb is implied in the ACI 318-11 equation to be
equal to 140 psi (0.96 MPa).5 The reduction of the Hoyer
effect15 leads to a decrease in the frictional forces resulting
from the confining stress. A ratio between the average transfer
bond stress and the average flexural bond stress was used
to decrease the effect of the confining stress, where ut /ufb =
2.86. Therefore, Eq. (8a) is introduced to assess the average
flexural bond stress, including the effect of confining stress.
Another reason to use this factor is the fact that microcracks
will form in the pile/bent-cap system at higher levels of load
(average flexural bond stress only appears after cracking14),
causing the confining stress from the shrinkage of the bent
cap to decrease. In the test program considered, however, the
confining stress is not expected to decrease, as it was applied
to the specimens permanently via a clamping force.4 Therefore, for the test program considered, the reduction factor of
2.86 was neglected, as presented in Eq. (8b).
m s cav
2.86

(8a)

u fbc = 140 + m s cav

(8b)

u fbc = 140 +

POTENTIAL MODIFICATION TO ACI 318-11


EQUATION (ACCOUNTING FOR CONFINEMENT)
Replacing the average flexural bond stress term in the
expression of flexural bond stress given in Eq. (6) by the
confined flexural bond stress will modify the equation to
account for confining stress. In the cases where confining
stress is present, the values of both confined transfer bond
stress (Eq. (7)) and confined flexural bond stress (Eq. (8a))
are greater than those of the average transfer bond stress and
average flexural bond stress, respectively, thereby decreasing
the development length (Eq. (9a)) and increasing the slipping
stress of prestressing strands (Eq. (9b)). It is noted that the
effect of confinement does not change the transfer length.
Rather, it reduces the potential for slipping of the strands
due to an increase in the apparent bond stress. The first part
of the equation represents the effective transfer length, while
the second part represents the flexural bond length, where
Ldc is the confined development length (in.). Equations (7)
and (8a) are used to define the values for confined transfer
bond stress and confined flexural bond stress, respectively.
Ldc =

f ps - fse
fse
db +
db
7.36 utc
7.36 u fbc

fss = 7.36 u fbc

Ld utc - u fbc
+
fse f ps
db
utc

(9a)

(9b)

MOMENT-CURVATURE ANALYSIS
A detailed moment-curvature analysis was conducted
using a numerical program.6 Using the compressive strength
data in Table 1, each of the 19 piles was modeled according
to its material properties. Two concrete material models were
771

Table 3Relation between embedment length


and confinement
Embedment
length, in.

Confining
stress, psi

scav,
psi

utc
(Eq. (7)),
psi

ufbc
(Eq. (8a)),
psi

ufbc
(Eq. (8b)),
psi

36

397

199

479

168

220

42

340

170

468

164

208

48

297

149

460

161

200

60

238

119

448

157

188

Notes: 1 in. = 25.4 mm; 1 psi = 0.006895 MPa.

used in the analysis: unconfined concrete for the cover and


confined concrete for the core of the section where lateral
reinforcement surrounds the concrete. The confined concrete
stress-strain curve differs from the unconfined concrete
stress-strain curve according to the percentage of transverse
reinforcement in the section. This is accounted for within
the numerical program through use of the Mander18 confined
concrete model.
Slipping stress of prestressing strands
The slipping stress of a prestressing strand is affected
by the embedment length in the pile-to-cap connection.
Considering the ACI 318-11 equation (Eq. (12-4)), as shown
in Table 2, it is predicted that the strands will not develop
their full tensile capacity (fpu = 270 ksi [1860 MPa]) due to
insufficient development length. Three slipping stress values
were investigated in the modeling of each pile:
1. The experimentally determined slipping stress obtained
from the experiments performed by Shahawy and Issa4;
2. The slipping stress as calculated from the ACI 318-11
equation (Eq. (1b) of this paper), knowing the value of
development length and effective prestressing stress calculated from the results of Shahawy and Issa4; and
3. The slipping stress as calculated from Eq. (9b), which
accounts for the effect of confinement. Within this equation,
the confined transfer bond stress and confined flexural bond
stress are calculated using Eq. (7) and (8b), respectively,
where the average confining stress is acquired by averaging
the applied confining stresses and distributing it over the four
faces of the pile. As the value of confining stress changes
according to the different development lengths used in this
study, the values of the confined transfer and flexural bond
stresses will change accordingly, as shown in Table 3.
The three slipping stress values calculated for each pile
were incorporated in the numerical models, and momentcurvature plots were generated to examine the differences
in moment capacity for each specimen. The analysis in each
model is terminated when the strands reach the maximum
allowable strain.
CONFINING STRESS
The confining stress described in this paper was introduced using a known force and was therefore calculated and
used in Eq. (9b); however, for design purposes, this confining
stress will be exerted from the shrinkage of the CIP bent cap
cast around the embedded pile. Therefore, an equation for
the estimation of the confining stress for purposes of design
has been developed and is described in the following.5
The equation uses Lams equations for the calculation of
stresses in thick-walled cylinders. For this equation, ideal772

ized circular geometries are modeled for both the pile and
bent cap for simplicity. In Eq. (10), sc is the confining stress
(psi); do and Do are the least dimensions (in.) of the pile
and bent cap, respectively; and Ep, vp and EBC, vBC are the
Youngs modulus (psi) and Poissons ratio of the pile and
bent cap, respectively.
Do e sh sc =

do
E BC

do s c
1 - vp
Ep

D 2 + d0 2
d
* o2
+ vBC + 0 1 - v p
2
Do - d0
Ep

(10)

where esh is the shrinkage strain at a given time, calculated in accordance with ACI 209R-92.19 In this paper,
the shrinkage strain can be calculated as esh = t/(35 + t)
(esh)u, where t is the time in days and (esh)u is the ultimate
shrinkage strain (780 me). The value of ultimate shrinkage
given in ACI 209R-9219 is only applicable for cases where
the reinforcement of the bent cap is minimal. For other
cases, the effect of reinforcement on the shrinkage strain
should be considered.
The confining stress from the Shahawy and Issa4 initial
test was calculated using Eq. (10), with dimensions and
material properties taken from the Shahawy and Issa4 experiment. A finite element model was created to assess the accuracy of the equation in predicting the confining stress at a
given strain.20 The results are described in the Results and
Discussion section.
Effect of creep
An analytical creep model was used to assess the
effect of creep on the confining stress. A restrained creep
model21,22 was used to model creep in the pile, as the piles
have spiral reinforcement, which will affect the creep, while
an unrestrained creep model23 was used to model creep in
the cap, as it did not have any reinforcement. A daily based
creep analysis was performed and, as expected, the effect of
creep decreased the value of the confining stress. To incorporate creep in the confining stress equation, an approximation was introduced to simplify the approach by introducing
creep as a reduction factor (Rcr). The value of Rcr is defined
in the following. This simplification was done due to the
complexity of dealing with two time-dependent variables:
shrinkage and creep. In Eq. (11), the confining stress calculated from Eq. (10) is reduced by the creep effect and is
referred to as the average confining stress scav. The average
confining stress scav should be calculated and used in Eq. (7)
and Eq. (8a) or (8b).
s cav = s c (1 - Rcr )

(11)

RESULTS AND DISCUSSION


Results related to confining stress, slipping stress, moment
capacity, and development length are discussed in this section.
Confining stress
The confining stress due to shrinkage of the bent cap
in the Shahawy and Issa4 initial test was calculated using
Eq. (10), which does not account for creep, and Eq. (11),
where Rcr was calculated using a daily-based shrinkage/
ACI Structural Journal/November-December 2012

Table 4Results of slipping stresses


Specimen number

Ld, in.

Experimental fss, ksi

fss (ACI 318-11)


(Eq. (1b)), ksi

Ratio*, %

fss (Eq. (9b)), ksi


(ufbc [Eq. (8a)])

Ratio, %

fss (Eq. (9b)), ksi


(ufbc [Eq. (8b)])

Ratio, %

A-1E

36

256

180

70

194

76

204

80

A-2E

36

263

190

73

204

78

213

81

A-3I

36

254

180

71

194

76

204

80

A-4I

36

253

181

71

195

77

205

81

B-1E

42

262

199

76

214

82

225

86

B-2E

42

261

199

76

213

82

224

86

B-3E

42

257

197

77

211

82

222

87

B-4E

42

260

196

75

210

81

222

85

B-5E

42

263

200

76

215

82

225

86

B-6I

42

259

197

76

211

82

223

86

C-1E

48

260

208

80

223

86

236

91

C-2E

48

258

206

80

221

86

235

91

C-3I

48

262

209

80

224

86

237

91

C-4I

48

258

206

80

221

86

234

91

C-5I

48

260

210

81

225

86

237

91

C-6E

48

258

207

80

222

86

235

91

D-1E

60

262

233

89

249

95

264

101

D-2I

60

261

235

90

250

96

265

102

D-3E

60

260

233

90

249

96

264

102

Ratio between slipping stress from ACI 318-11 equation (Eq. (1)) and experimental slipping stress.

Ratio between slipping stress from modified ACI 318-11 equation (Eq. (9b)) using ufbc from Eq. (8a) and experimental slipping stress.

Ratio between slipping stress from modified ACI 318-11 equation (Eq. (9b)) using ufbc from Eq. (8b) and experimental slipping stress.
Notes: 1 in. = 25.4 mm; 1 ksi = 6.895 MPa.

creep analysis to be 20%. The results from Eq. (11) were


then compared to a finite element model, which uses the
actual rectangular geometry of the pile and bent cap from the
Shahawy and Issa4 test. Using Eq. (10), the confining stress
was significantly overpredicted with a value of 1230 psi
(8.5 MPa). The average confining stress (accounting for
creep) calculated using Eq. (11) was more reasonably estimated with a value of 980 psi (6.8 MPa). This compares to
the maximum measured confining strain by Shahawy and
Issa4 of 880 psi (6.1 MPa), which is 90% of the calculated
average confining stress by Eq. (11). This is considered to
be a reasonable match when consideration is given to the
uncertainty involved with the material properties used in
Eq. (10) and (11). From the finite element model using the
actual rectangular geometry for both the pile and the bent
cap, the average confining stress was found to be 907 psi
(6.3 MPa). This value compares favorably and is within 3%
of the measured confining stress value.
The aforementioned confining stresses represent the
maximum value of the confining stress acting over the
embedment length of the pile. Shahawy and Issa4 reported an
average value of the confining stress to be 525 psi (3.6 MPa)
by assuming a parabolic distribution of confining stress along
the embedment length. For design purposes, a proposed
simplification is to assume that the value of the confining
stress varies linearly over the embedment length, where the
minimum value is zero (at the soffit) and the maximum value
is calculated using Eq. (11). Therefore, a recommended
approach is to use one-half of the confining stress value
calculated using Eq. (11) due to the assumed linear distribution. The confining stress for purposes of design would
ACI Structural Journal/November-December 2012

therefore be 490 psi (3.4 MPa), which compares favorably


with the average confining stress value reported by Shahawy
and Issa4 (within 7%).
Slipping stress
When sufficient development length is provided, slipping
of the strands should not occur and the strands will reach
their nominal tensile capacity fpu. In the study by Shahawy
and Issa,4 the available development length (embedment
length) was less than the theoretical value required by the
ACI 318-11 equation. This condition therefore predicts slipping prior to reaching the nominal tensile capacity of the
strands. The measured steel stress at failure as reported in
Shahawy and Issa4 (fps; Table 1, Column 4) is compared to
the theoretical values calculated with the ACI 318-11 equation (Eq. (1b)) and the modified equation (Eq. (9b)), which
accounts for confining stress. The results are shown in
Table 4. The ratios (in percentage) between slipping stress
calculated from the ACI 318-11 equation, experimental slipping stress, and the modified equation are listed. Overall,
a better match is achieved with the modified equation
(Eq. (9b)). The slipping stress from Eq. (9b) was calculated
twice using both Eq. (8a) and (8b). There is a better match
with the experimental results when Eq. (8b) is used to calculate the confined flexural bond stress because this equation is
more representative for the study discussed.
Moment capacity
The moment capacity of the piles is dependent on the slipping stress and fps of the prestressing strands. Using the three
slipping stress values discussed previously, the values of
773

Table 5Ultimate moments for different slipping stresses using moment-curvature analysis and required
development length to achieve experimental slipping stress
Calculated moment, kip-in.

Development length

Specimen
number

Ld, in.

Theoretical ultimate
moment, kip-in.

Using experimental
slipping stress

Using ACI 318-11


(Eq. (1b)) slipping stress

Using Eq. (9b)


slipping stress

ACI 318-11
(Eq. (1a)), in.

Eq. (9a), in.

A-1E

36

1560

1540

1250

1360

74.2

52.2

A-2E

36

1500

1490

1280

1380

72.2

51.5

A-3I

36

1530

1520

1240

1350

73.2

51.5

A-4I

36

1480

1470

1220

1330

72.2

51.0

B-1E

42

1540

1530

1350

1450

73.2

54.1

B-2E

42

1530

1520

1350

1440

73.2

54.0

B-3E

42

1510

1490

1320

1410

72.2

53.3

B-4E

42

1580

1560

1360

1480

74.2

54.6

B-5E

42

1530

1520

1350

1450

73.2

54.1

B-6I

42

1530

1510

1340

1440

73.2

53.9

C-1E

48

1550

1540

1400

1490

74.2

56.3

C-2E

48

1530

1520

1380

1470

73.9

56.0

C-3I

48

1580

1560

1440

1510

74.2

56.4

C-4I

48

1530

1510

1380

1460

73.8

55.9

C-5I

48

1530

1520

1400

1470

73.2

55.7

C-6E

48

1530

1510

1390

1470

73.2

55.6

D-1E

60

1560

1550

1500

1560

74.2

59.0

D-2I

60

1530

1520

1480

1520

73.2

58.4

D-3E

60

1530

1520

1480

1520

73.2

58.2

Notes: 1 in. = 25.4 mm; 1 kip-in. = 0.11298 kN-m.

Fig. 4Moment versus curvature for Specimen D-2I (60 in.


[1524 mm] embedment).

Fig. 5Average ultimate moment versus embedment depth.

ultimate moment are calculated and summarized in Table 5.


As the embedment length increases, the average moment
capacity for each embedment length increases in all cases.
However, the moment capacity as calculated using the modified equation (Eq. (9b)) compares favorably with the ones
calculated using the experimental slipping stress.
Figure 4 shows a moment-versus-curvature plot for
Specimen D-2I (60 in. [1524 mm] embedment). The ratios
between the moment capacities calculated using the proposed
equation (Eq. (9b)) and the one using experimental slipping
stress for embedment lengths of 36, 42, 48, and 60 in. (914,
1067, 1219, and 1524 mm) are 91%, 95%, 97%, and 100%,

respectively. If the ACI 318-11 equation (Eq. (1b)) is used


for calculating moment capacity and compared to the values
calculated using experimental slipping stress, the ratios are
82%, 88%, 92%, and 97%, respectively. Figure 5 shows
the average calculated moment capacity versus embedment
length for the different models.

774

Development length
The development length required to reach the measured
experimental slipping stress was calculated using the
ACI 318-11 equation (Eq. (1a)) and the modified equation
(Eq. (9a)), as shown in Table 5. Using the embedment length
ACI Structural Journal/November-December 2012

as a benchmark, the results obtained from the modified equation (Eq. (9a)) provide a better match than those obtained
with the ACI 318-11 equation (Eq. (1a)).
Design recommendation and limitations
The confining stress is a function of shrinkage; therefore,
it is predicted that the value of confining stress will continue
to increase with time. At higher levels of confining stress,
microcracks may form to relieve the high stress, which leads
to a drop in the magnitude of the confining stress. Therefore,
an upper limit of 750 psi (5.2 MPa) is proposed to take into
account the effect of microcracking at high levels of confinement. This value is partially based on an ongoing laboratory
investigation, where piles are plainly embedded in CIP bent
caps and tested under reverse lateral cyclic loading to check
the moment capacity and ductility of the connection.24,25 This
upper-limit value is assumed to be the maximum confining
stress acting on the embedded end of the pile. A simplified
equation (Eq. (12)) is proposed by substituting this upperlimit value in Eq. (9a).5
Ldc =

f ps - fse
fse
db +
db
5000
1800

(12)

The results from the actual pile-to-CIP-bent-cap connections show that Eq. (12) has a better comparison with the
experimental results than the ACI 318-11 equation5; however,
the use of Eq. (12) with the data described in this paper is not
appropriate, as the confining stress was artificially simulated
with steel plates for the Shahawy and Issa4 study.
The ACI 318-11 equation is more conservative than the
modified equation. Therefore, it is not recommended that
the modified equation approach be used in practice in the
absence of further investigation and verification. The results
presented in this study are limited to the use of 0.5 in.
(13 mm) low-relaxation seven-wire prestressing strands.
The appropriateness of using Eq. (12) with a different strand
diameter requires further investigation.
SUMMARY AND CONCLUSIONS
The appropriateness of ACI 318-11, Eq. (12-4), for the
calculation of development length for prestressing strands
in confined sections was studied. A modified equation
was developed and introduced in this paper to account for
confinement. The experimental results of Shahawy and
Issa4 were used to develop a moment-curvature analysis.
The results were compared to calculated results from the
ACI 318-11 equation and the modified equation. The conclusions of this study can be drawn as follows:
1. Confining stress affects the bond between prestressing
strands and concrete by increasing the effective average
bond stress within the transfer zone and the average flexural
bond stress. This enhances (increases) the stress required to
cause slipping.
2. Equation (9a) was developed for calculating development length in cases where confining stress takes place. One
such case occurs when precast piles are embedded in CIP
bent caps.
3. A better fit to the published experimental data was
obtained for confined sections with Eq. (9b) than with the
ACI 318-11 equation (Eq. (1b)). The results of both equations
are conservative when compared to the experimental results.
ACI Structural Journal/November-December 2012

4. The difference between the modified equation and the


ACI 318-11 equation becomes more significant as shorter
embedment lengths are used in the pile/bent-cap system.
5. The embedment length of prestressed piles in CIP bent
caps has a notable effect on the slipping stress of prestressing
strands and the moment capacity of the section.
6. The modified equation (Eq. (9a) and (9b)) provided
a reasonable fit to the experimental data described in this
paper. Further consideration is recommended prior to implementation of these equations for purposes of design. Among
other items, it is recommended that confining stresses be
monitored in field applications.
ACKNOWLEDGMENTS

The authors wish to express their gratitude and sincere appreciation to


the South Carolina Department of Transportation (SCDOT) and the Federal
Highway Administration (FHWA) for financial support. The opinions, findings, and conclusions expressed in this paper are those of the authors and
not necessarily those of SCDOT or FHWA.

REFERENCES

1. SCDOT Bridge Design Manual, South Carolina Department of Transportation, Columbia, SC, 2006, pp. 19.1-19.22.
2. SCDOT Seismic Design Specifications for Highway Bridges, Version
2.0, South Carolina Department of Transportation, Columbia, SC, 2008,
153 pp.
3. ACI Committee 318, Building Code Requirements for Structural
Concrete (ACI 318-11) and Commentary, American Concrete Institute,
Farmington Hills, MI, 2011, 503 pp.
4. Shahawy, M. A., and Issa, M., Effect of Pile Embedment on the
Development Length of Prestressing Strands, PCI Journal, V. 37, No. 6,
1992, pp. 44-59.
5. ElBatanouny, M. K.; Ziehl, P. H.; Larosche, A.; Mays, T.; and Caicedo,
J. M., Bent-Cap Confining Stress Effect on the Slip of Prestressing Strands,
ACI Structural Journal, V. 109, No. 4, July-Aug. 2012, pp. 487-496.
6. XTRACT Program, v3.0.8, TRC/Imbsen Software Systems, Rancho
Cordova, CA, 2007.
7. Deatherage, J. H.; Burdette, E. G.; and Key Chew, C., Development
Length and Lateral Spacing Requirements of Prestressing Strands for
Prestressed Concrete Bridge Girders, PCI Journal, V. 39, No. 1, Jan.-Feb.
1994, pp. 70-83.
8. Zia, P., and Mostafa, T., Development Length of Prestressing Strands,
PCI Journal, V. 22, No. 5, Sept.-Oct. 1977, pp. 54-65.
9. Mattock, A. H., Proposed Redraft of Section 2611Bond of the
Proposed Revision of Building Code Requirements for Reinforced Concrete
(ACI 318-56), ACI Committee 323 Correspondence, 1962.
10. Bennett, W. B. Jr., Preliminary Draft of the Proposed Revision of
Section 211Bond and Anchorage of the Tentative Recommendations for
Prestressed Concrete, Letter to the Members of ACI Committee 423, 1963.
11. Tabatabai, H., and Dickson, T. J., The History of the Prestressing
Strand Development Length Equation, PCI Journal, V. 38, No. 6, Nov.Dec. 1993, pp. 64-75.
12. Hanson, N. W., and Kaar, P. H., Flexural Bond Tests of Pretensioned
Prestressed Beams, ACI JOURNAL, Proceedings V. 55, No. 1, Jan. 1959,
pp. 783-802.
13. Kaar, P. H.; La Fraugh, R. W.; and Mass, M. A., Influence of
Concrete Strength on Strand Transfer Length, PCI Journal, V. 8, No. 5,
Oct. 1963, pp. 47-67.
14. Janney, J. R., Nature of Bond in Pretensioned Prestressed Concrete,
ACI JOURNAL, Proceedings V. 50, No. 5, May 1954, pp. 717-736.
15. Hoyer, E., and Friedrich, E., Beitrag zur frage der haftspannung
in eisenbetonbauteilen (Contribution to the Question of Bond Stress in
Reinforced Concrete Elements), Beton und Eisen 38, 1939, pp. 107-110.
16. Stocker, M. F., and Sozen, M. A., Investigation of Prestressed
Concrete for Highway Bridges, Part V: Bond Characteristics of Prestressing
Strand, Bulletin 503, University of Illinois Engineering Experiment
Station, Urbana, IL, 1970, 119 pp.
17. PCI Design Handbook: Precast and Prestressed Concrete, sixth
edition, Precast/Prestressed Concrete Institute, Chicago, IL, 2004, 736 pp.
18. Mander, J. B.; Priestley, M. J. N.; and Park, R., Theoretical StressStrain Model for Confined Concrete, Journal of Structural Engineering,
ASCE, V. 114, 1988, pp. 1804-1826.
19. ACI Committee 209, Prediction of Creep, Shrinkage, and Temperature Effects in Concrete Structures (ACI 209R-92) (Reapproved 1997),
American Concrete Institute, Farmington Hills, MI, 1992, 48 pp.

775

20. Dassault Systmes, ABAQUS Analysis Users Manual V. 6.9,


ABAQUS Inc., Dassault Systmes, Providence, RI, 2009.
21. Baant, Z. P., Prediction of Concrete Creep Effects Using AgeAdjusted Effective Modulus Method, ACI JOURNAL, Proceedings V. 69,
No. 4, Apr. 1972, pp. 212-219.
22. Dilger, W. H., Creep Analysis of Prestressed Concrete Structures
Using Creep-Transformed Section Properties, PCI Journal, V. 27, No. 1,
1982, pp. 99-118.

776

23. Comit Euro-International du Bton, CEB-fib Model Code 1990,


Thomas Telford Services, Ltd., London, UK, 1993, 437 pp.
24. ElBatanouny, M. K., Determining Slipping Stress of Prestressing
Strands in Piles Embedded in CIP Caps, MS thesis, University of South
Carolina, Columbia, SC, Dec. 2010, 90 pp.
25. Larosche, A., The Connection Behavior of Precast Prestressed Piles
to Cast-in-Place Bent Caps, MS thesis, University of South Carolina,
Columbia, SC, May 2011, 130 pp.

ACI Structural Journal/November-December 2012

Das könnte Ihnen auch gefallen