You are on page 1of 10

Engineering Structures 41 (2012) 403–412

Contents lists available at SciVerse ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Bond of 13 mm prestressing steel strands in pretensioned concrete members


J.R. Martí-Vargas ⇑, P. Serna, J. Navarro-Gregori, L. Pallarés
Institute of Concrete Science and Technology (ICITECH), Universitat Politècnica de València, 4G, Camino de Vera s/n, 46022 Valencia, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents an experimental research work to determine both the transmission and the anchor-
Received 5 May 2011 age lengths of seven-wire prestressing steel strands in different concrete mixes. A testing technique
Revised 22 February 2012 based on a bond behavior analysis by measuring the force supported by the prestressing strand on a ser-
Accepted 30 March 2012
ies of specimens with different embedment lengths has been used. Relationships between the average
Available online 7 May 2012
bond stress for both the transmission length and anchorage length as a function of the concrete compres-
sive strength have been found. Equations to compute transmission and anchorage lengths of 13 mm pre-
Keywords:
stressing strands have been obtained. The experimental results have been compared with the theoretical
Bond
Concrete
prediction from proposed equations in the literature and with experimental results from other authors.
Strand Ó 2012 Elsevier Ltd. All rights reserved.
Prestress
Transmission length
Anchorage length
Transfer length
Development length

1. Introduction ment, have been conducted over the years. Bond strength, as well
as transmission and anchorage lengths, depend on several factors
In pretensioned concrete members, the prestressing force in the [1,3]: concrete strength at the time of the prestress transfer, initial
reinforcement is transferred to the concrete by the bond in the end reinforcement stress, concrete cover, prestress transfer process
region of the member. Furthermore, when a pretensioned concrete condition, reinforcement geometry, reinforcement surface condi-
member is loaded, the activation of bond stress increases the pre- tion, concrete strength at the time of loading, etc. However, no con-
stressing reinforcement force. Therefore, in pretensioned concrete sensus has been reached on the main parameters to be considered
members it is essential a correct design and an accurate prediction in the equations to calculate both transmission and anchorage
of the lengths affected in the end region of the member by means lengths. An example of this is ACI Code 318-11 [2], provisions for
of the required bond stress. transmission (transfer) and anchorage (development) lengths
In the end region of a pretensioned concrete member and after which are not a function of concrete strength. On the other hand,
the prestress transfer operation, the stress in the prestressing rein- Eurocode 2 [4] and Model Code 2010 [5] provisions for transmis-
forcement varies from zero at the free end to a maximum value sion and anchorage lengths include concrete properties.
(effective stress) along the distance, defined as the transmission Nowadays, it is assumed that bond performance is essential for
length in agreement with the terminology presented in [1] (trans- an adequate response of pretensioned prestressed concrete appli-
fer length according to [2]). When a pretensioned concrete mem- cations. ACI 318-11 [2] indicates that the quality assurance proce-
ber is loaded, a complementary bond length beyond the dures for bonded applications should be used to confirm that the
transmission length is required to develop the corresponding rein- bond properties of reinforcement are adequate. However, there
forcement stress from the effective prestress. The embedment are no minimum requirements for the bond performance of pre-
length from the free end required to reach a design stress is known stressing strands in [2], or in standards like in [6,7].
as the anchorage length [1] (development length according to [2]). In spite of the large number of experimental research works
The anchorage length is obtained as the sum of the transmission carried out, there is no consensus on a standard test method for
length and the complementary bond length. bond quality [1]. Recently, an experimental methodology has been
Several theoretical and experimental works about bond and developed, the ECADA test method, which is based on both the
transmission, and on anchorage lengths of prestressing reinforce- measurement and analysis of the force supported by the reinforce-
ment in specimen series with different embedment lengths [8],
⇑ Corresponding author. Tel.: +34 96 387700775612; fax: +34 96 3877569. and its feasibility has been verified [9].
E-mail addresses: jrmarti@cst.upv.es (J.R. Martí-Vargas), pserna@cst.upv.es (P. The purpose of this research study is to develop an analytical
Serna), juanagre@cst.upv.es (J. Navarro-Gregori), luipalru@cst.upv.es (L. Pallarés). bond model to predict the transmission and anchorage lengths of

0141-0296/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.engstruct.2012.03.056
404 J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412

13 mm prestressing steel strands for bond characterization. To this in prestressing reinforcement after prestress transfer, rp2 the stress
end, an experimental program to determine transmission and in prestressing reinforcement at loading, Ap the cross-sectional area
anchorage lengths, as well as the average bond stress along both of prestressing reinforcement, LA the anchorage length.
the transmission length and the complementary bond length in Fig. 1 shows the idealized increase of the prestressing reinforce-
12 concretes of different compositions and properties, by means ment stress with the embedment length from the free end accord-
of the ECADA test method, has been set up. The experimental re- ing to the bilinear model presented.
sults have been compared with other theoretical and experimental According to Fig. 1 (LA = LT + LC), several equations based on
studies found in the literature. experimental results have been proposed by several codes and
authors to predict the transmission and the anchorage lengths.
2. Background Table 1 shows some of these equations for seven-wire prestressing
strands. For each reference, the equations for transmission length
According to [10], the uniform bond stress distribution hypoth- (Equations (a)) and for complementary bond length (Equations
esis is an unattainable limit since a portion zone that behaves in an (b)) are indicated, resulting in the corresponding equation for
elastic way will always exist in both the transmission length and anchorage length as Equation (c) = Equation (a) + Equation (b).
the complementary bond length. An analytical bond model for Complementary bond length is deduced as LC = LA  LT for the case
anchorage length that considers both the plastic and elastic zones of Ref. [15]. Once the notation of the different equations has been
along the transmission and complementary bond lengths was pro- adapted from their original form to SI Units, then:
posed in [11]. These elastic zones are located one after the other at
the end of the transmission length, and also at the beginning of the / nominal diameter of prestressing strand
complementary bond length. However, a plastic response along the rpt initial prestress in prestressing strand prior to release
almost entire transmission length [12,13], and a very small elastic rpi effective stress in prestressing strand just after prestress
zone in the case of complementary bond length [14], have been transfer
reported. rpa maximum stress in strand at loading (for design stress
Currently, the uniform bond stress distribution hypothesis is [5], at nominal strength [2])
generally accepted by several codes [2,4,5] and authors [15–28], rpcs effective stress in prestressing strand after all prestress
which assumes linear variations of the prestressing reinforcement losses
stress for both the transmission and complementary bond lengths, fci concrete compressive strength at time of release
resulting in a bilinear model. fcl concrete compressive strength at loading
In order to obtain the equilibrium of a prestressing reinforce- fc concrete compressive strength at 28 days
ment, the transfer bond force over the transmission length and
the anchorage bond force along the complementary bond length
must equal the force in the prestressing reinforcement according It should be noted that Table 1 includes several equations for
to Eqs. (1) and (2), respectively; consequently, the anchorage transmission length: rp1 = rpi [4,5,16–20,23,24] and others
length can be obtained from Eq. (3): rp1 = rpcs [2,11]. Some cases correspond to the variations proposed
for the ACI Code provisions, which first appeared in 1963 [29] and
U T Rp LT ¼ rp1 Ap ð1Þ were derived from Eq. (3) taking into account UT = 2.76 MPa and
UC = 0.94 MPa [30]. This equation has remained up to date in [2]
U C Rp LC ¼ ðrp2  rp1 ÞAp ð2Þ and is applied for all types of concrete in spite of a considerable
number of proposed changes that includes concrete strength
rp1 Ap ðrp2  rp1 ÞAp [11,16,21,24]. In addition, several authors [16,17,19–21,24]
LA ¼ LT þ LC ¼ þ ð3Þ
U T Rp U C Rp consider that the use of term rpi, rather than rpcs, to compute
where UT is the average bond stress along the transmission length, transmission length is more rational, and in [17,19,20,22] the
UC the average bond stress along the complementary bond length, UT = 2.76 MPa is retained resulting in greater transmission lengths.
Rp the perimeter of prestressing reinforcement, LT the transmission For design purposes, it is generally considered that the trans-
length, LC the complementary bond length, rp1 the effective stress mission length (with rp1 = rpi or rp1 = rpcs) established at the time
of the prestress transfer does not significantly change with time.
The anchorage length prediction takes into account the term
rp1 = rpcs in the complementary bond length in all the cases pre-
sented in Table 1, except in [25] (rp1 = rpt). This exception is con-
sidered in [25] to obtain the best coefficient of correlation (R2) in
several simple regression models based on measured complemen-
tary bond lengths (R2 = 0.47 when rp1 = rpcs and R2 = 0.69 when
rp1 = rpt; for transmission length, R2 = 0.40 is obtained).
Figs. 2–4 present the transmission length, complementary bond
length, and anchorage length of a seven-wire 13 mm prestressing
steel strand, respectively. These lengths have been predicted from
the equations in Table 1 for concrete compressive strength at the
time of prestress transfer fci, which is equal to 30 MPa and
50 MPa in these comparisons. The following relationships have
been adopted: rpt = 0.75fpu (fpu = 1860 MPa, specified tensile
strength of prestressing strand), rpi = 0.93rpt, rpcs = 0.8rpt,
rpa = 0.9fpu and fcl = 1.5fci. For the ap2 factor included in the
MC2010 [5] to consider the action effect to be verified in design
(ap2 = 1 for calculation of anchorage length when moment and
shear capacity is considered; ap2 = 0.5 for verification of transverse
Fig. 1. Idealized strand stress profile along the anchorage length. stress in anchorage zone), a value of ap2 = 0.75 has been adopted by
J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412 405

Table 1
Proposed equations for transmission length and anchorage length from the literature (in MPa and mm).

Source [reference] Eq. Equations (a) for Equations (b) for Remarks
number transmission length (LT) complementary bond length
(LC)
ACI 318-63 [29] ACI 318-11 [2] (4) r /
pcs
LT ¼ 20:7 LC ¼ 0:145ðrpa  rpcs Þ/
Martin and Scott [15] (5) LT ¼ 80/ LC = LA  LT /
LA ¼ 2:69 ðrpa  1595 Þ
/1=6
Zia and Mostafa [16] (gradual release) (6) LT ¼ a
rpi /
b LC ¼ 0:181ðrpa  rpcs Þ/ For gradual release: a = 1.3; b = 58 For sudden release:
fci
a = 1.5; b = 117
pffiffiffiffi
Cousins et al. [11] (7) rpcspAp
ffiffiffiffi þ 0:5 U 0t fci LC ¼
ðrpa rpcs ÞAp
pffiffiffi For uncoated strands: U 0t ¼ 0:556; B = 0.0815 MPa/mm;
LT ¼ B p/U0c fc
p/U 0t fci U 0c ¼ 0:110
Shahawy et al. [17] (LT) (8) rpi / LC ¼ 0:145
r  rpcs Þ/
ð pa k = 2 for piles k = 1 for slabs and slender members k = 0.5
LT ¼ 20:7 k
Shahawy [18] (LC) when LA/h 6 3 (h = overall thickness of member)
Deatherage et al. [19] (9) LT ¼
rpi / LC ¼ 0:218ðrpa  rpcs Þ/
20:7
Buckner [20] (10) LT ¼
rpi / LC ¼ k  0:145ðrpa  rpcs Þ/ 16k62
20:7
qffiffiffiffiffiffiffi qffiffiffiffi
Mitchell et al. [21] (11) LT ¼
rpi / 20:7
LC ¼ 0:145ðrpa  rpcs Þ/ 30
20:7 fci fc
ðrpcs =0:8Þ/
 rpa  rpa
Tadros and Baishya [22] (12) LT ¼ LC ¼ 186  8 ðrpa  rpcs Þ/ 186 8P1
20:7
Lane [23] (13) 4rpt / 6:4ðrpa rpcs Þ/
LT ¼ fc
 127 LC ¼ fc
þ 381
Mahmoud et al. [24] (14) rpi / ðrpa rpcs Þ/ at = 2.4 for steel strands af no reported rpcs = effective
LT ¼ 0:67
at fci
LC ¼ af fcl0:67
prestress at loading
Kose and Burkett [25] (15) LT ¼ 0:05
rpt p
ð1/Þ
ffiffiffi
2
LC ¼ 203:2 þ 0:22
ðrpa rpt Þð1/Þ
pffiffiffi
2

fc fc
EC2-2004 [4] (16) r ðr r Þ a1 (type of release); a2 (area factor,); gp1 and gp2 account
LT ¼ a1 a2 / g gpif LC ¼ a2 / gpa g fpcs
p1 1 ctdi p2 1 ctd
for the tendon type; g1 account the bond conditions; fctdi
and fctd (concrete tensile strength)
MC 2010 [5] (17) A r A ðr r Þ ap1(type of release); ap2 (action effect to be verified); ap3
LT ¼ ap1 ap2 ap3 p/p g g pi f LC ¼ p/p g pag fpcs
p1 p2 ctdi p1 p2 ctd
(bond situation); Ap/p/ = 7//36; gp1 (tendon type); gp2
(tendon position); fctdi and fctd (concrete tensile strength)

EC2-2004/MC2010 [4,5] Eq. (16) (17) (sudden) 1,5 LT,30 / LT,50


ratio
EC2-2004/MC2010 [4,5] Eq. (16) (17) (gradual) 1,5

Kose and Burkett 2005 [25] Eq. (15) 1,3


Mahmoud et al.1992 [24] Eq. (14) 1,4
Lane 1998 [23] Eq. (13) 1,8
Tadros and Baishya 1996 [22] Eq. (12) 1,0

Mitchell et al.1993 [21] Eq. (11) 1,3

Buckner 1995 [20] Eq. (10) 1,0


Deatherage et al. 1994 [19] Eq. (9) 1,0

Shahawy et al. 1992 [17] Eq. (8) 1,0


Cousins et al. 1990 [11] Eq. (7) 1,2

Zia and Mostafa 1977 [16] Eq. (6) (sudden) 1,9

Zia and Mostafa 1977 [16] Eq. (6) (gradual) 1,8

Martin and Scott 1976 [15] Eq. (5) 1,0

ACI 318-11 [2] Eq. (4) 1,0

0 500 1000 1500 2000

Transmission length for fci = 30 MPa (LT,30) Transmission length (mm)


Transmission length for fci = 50 MPa (LT,50)

Fig. 2. Transmission lengths as predicted by different pre-existing equations.

the authors from the established values ap2 = 1 and ap2 = 0.5 for the which are related to concrete strength. The ratios of the lengths ob-
upper bound and lower bound values of transmission length, tained for both concrete compressive strengths are shown in the
respectively. With ap2 = 0.75, the provisions for the transmission figures.
length from [4,5] coincide. Regarding to the experimental results of transmission and
Figs. 2–4 show the wide ranges of predicted values by means of anchorage lengths obtained in the literature, the values of the
different equations of transmission length, complementary bond transmission lengths for 13 mm prestressing steel strands are
length and anchorage length, respectively. In addition, it may be around 600–700 mm, with minimum values of 330–350 mm
seen that these lengths are always decreased when concrete [21,24] and maximum of 1800 mm [11]. Anchorage length values
strength increases for the lengths predicted from those equations are often above 2000 mm [11,19], although some are also around
406 J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412

EC2-2004/MC2010 [4,5] Eq. (16) (17) (sudden) 1,5


EC2-2004/MC2010 [4,5] Eq. (16) (17) (gradual) 1,5

Kose and Burkett 2005 [25] Eq. (15) 1,2 LC,30 / LC,50
ratio
Lane 1998 [23] Eq. (13) 1,4

Tadros and Baishya 1996 [22] Eq. (12) 1,0


Mitchell et al.1993 [21] Eq. (11) 1,3
Buckner 1995 [20] Eq. (10) (λ=2) 1,0
Buckner 1995 [20] Eq. (10) (λ=1) 1,0
Deatherage et al. 1994 [19] Eq. (9) 1,0

Shahawy 1993 [18] Eq. (8) (k=0,5) 1,0

Shahawy 1993 [18] Eq. (8) (k=2) 1,0


Cousins et al. 1990 [11] Eq. (7) 1,3

Zia and Mostafa 1977 [16] Eq. (6) 1,0

Martin and Scott 1976 [15] Eq. (5) 1,0

ACI 318-11 [2] Eq. (4) 1,0

0 500 1000 1500 2000 2500

Complementary bond length for fci 30 MPa (LC,30) Complementary bond length (mm)
Complementary bond length for fci 50 MPa (LC,50)

Fig. 3. Complementary bond lengths as predicted by different pre-existing equations.

EC2-2004/MC2010 [4,5] Eq. (16) (17) (sudden) 1,5 LA,30/LA,50


ratio
EC2-2004/MC2010 [4,5] Eq. (16) (17) (gradual) 1,5
Kose and Burkett 2005 [25] Eq. (15) 1,3

Lane 1998 [23] Eq. (13) 1,6


Tadros and Baishya 1996 [22] Eq. (12) 1,0
Mitchell et al.1993 [21] Eq. (11) 1,3
Buckner 1995 [20] Eq. (10) (λ=2) 1,0
Buckner 1995 [20] Eq. (10) (λ=1) 1,0
Deatherage et al. 1994 [19] Eq. (9) 1,0

Shahawy (et al.1992)-93 [17,18] Eq. (8) (k=0,5) 1,0


Shahawy (et al.1992)-93 [17,18] Eq. (8) (k=2) 1,0
Cousins et al. 1990 [11] Eq. (7) 1,3
Zia and Mostafa 1977 [16] Eq. (6) (sudden) 1,2

Zia and Mostafa 1977 [16] Eq. (6) (gradual) 1,2


Martin and Scott 1976 [15] Eq. (5) 1,0
ACI 318-11 [2] Eq. (4) 1,0

0 500 1000 1500 2000 2500 3000 3500

Anchorage length for fci 30 MPa (LA,30) Anchorage length (mm)


Anchorage length for fci 50 MPa (LA,50)

Fig. 4. Anchorage lengths as predicted by different pre-existing equations.

700 mm [21]. Moreover, the UT/UC ratios to characterize the differ- concrete specimens. Specimens are made in pretensioning frames
ent bond situations have been determined theoretically and exper- with an adjustable strand anchorage as shown in Fig. 5. At the
imentally [20,27,31]. opposite end, an Anchorage-Measurement-Access (AMA) system
is placed to simulate the sectional stiffness of the specimens. The
test equipment is completed with a hydraulic jack that can be
3. Test procedure and instrumentation placed at the pretensioning frames ends.
The force in the strand is controlled and registered while the
The ECADA test method consists in sequentially analysing the test is being carried out by means of a hollow force transducer
transmission and anchorage process of the strand in pretensioned placed in the AMA system. Relative displacements between the
J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412 407

strand and concrete are also continuously measured and registered sand 0/4 and a polycarboxylic ether superplasticiser. All the con-
by means of a displacement transducer at the free end of the spec- crete mixes were designed with a constant gravel/sand ratio of
imen. A pressure transducer completes the instrumentation and is 1.14.
used to control the hydraulic jack. No internal measuring devices The prestressing strand was a low-relaxation seven-wire steel
are used in the specimens tested in order to not distort the bond strand, 13 mm in diameter, at a prestress level of 75% of the guar-
phenomenon. anteed ultimate strength (1860 MPa), specified as UNE 36094:97 Y
Once the equipment test is set up as shown in Fig. 5, with the 1860 S7 13.0 [6]. The main characteristics were taken from the
hydraulic jack connected to the frame at the free end, the different manufacturer: diameter, 12.9 mm, section, 99.69 mm2, nominal
test procedure phases are as follows. strength, 192.60 kN, yield stress at 0.2%, 177.50 kN, and modulus
of elasticity, 196.70 GPa.
3.1. Preparation stage All the specimens were subjected to the same consolidating and
curing conditions. The prestress transfer was gradually performed
 Lining up the strand in the frame. at 24 h after casting to avoid dynamic shock effects [33,34]. A 2-h
 Tensioning the strand: The hydraulic jack pulls the anchorage stabilization period from the prestress transfer was established.
plate and separates it from the adjustable strand anchorage. The pull-out operation was carried out after this stabilization per-
 Anchorage of the strand: The adjustable strand anchorage is set iod (consequently, fcl = fci) to reach a reference force (PR) of 158 kN
up to contact the anchorage plate thus bearing the force intro- in the prestressing strand, corresponding to the strand’s nominal
duced into the strand. The hydraulic jack is then unloaded. yield strength at 0.1% [6].
 Specimen cast: The concrete is mixed, placed into the form pre-
pared in the frame, and consolidated. 5. Determining transmission and anchorage lengths
 The specimen is cured to achieve the desired concrete proper-
ties and is then demoulded before testing. With the ECADA method, both the transmission and anchorage
lengths are determined by measuring and analysing the force sup-
3.2. Testing stage ported by the strand in a series of pretensioned concrete specimens
with different embedment lengths [8,9]. By way of example, Fig. 6
 Prestress transfer: The hydraulic jack is loaded to recover the shows the results of transferred prestressing and pull-out forces
force in the strand supported by the adjustable strand anchor- versus the embedment length for a concrete mix design.
age which is relieved. The strand prestress transfer takes place The transferred prestressing force values after stabilization per-
at a controlled speed through the unloading of the hydraulic iod PT are ordered according to specimen embedment lengths
jack. The prestressing force is transferred to the concrete and (Fig. 6). The obtained curves present a bilinear tendency, with an
the concrete specimen is supported at the stressed end of the ascendent initial branch and a sensibly horizontal branch corre-
frame. sponding to the effective prestressing force PE (PE = rpiAp). Trans-
 Stabilization period: The force in the strand after release (PT) is mission length LT corresponds to the shorter specimen
measured. embedment length with PT = PE; that is, it corresponds to the short-
 Pull-out operation: The hydraulic jack is positioned to increase er specimen embedment length that marks the beginning of the
the force in the strand by separating the anchorage plate of horizontal branch.
the AMA system from the frame. The maximum force achieved The pull-out force values PA are ordered according to the spec-
during the pull-out operation before the strand slip at the free imen embedment lengths (Fig. 6). The obtained curves present an
end (PA) is measured. ascendent trend. Anchorage length LA corresponds to the shorter
specimen embedment length of the test specimens in which the
4. Experimental program reference force PR in the strand is reached in the pull-out operation
without a strand slip at the free end of the specimen; that is, it cor-
An experimental program to determine the transmission and responds to the first specimen of the series with PA P PR. The com-
anchorage lengths of a 13 mm prestressing steel strand in different plementary bond length is obtained as LC = LA  LT.
concrete mixes has been carried out. The resolution in determining the transmission and anchorage
Specimens cross-sections were 100  100 mm2 with a concen- lengths will depend on the sequence of the specimen lengths
trical single prestressing strand. Tests were carried out on 12 dif- tested. For the specimen embedment length equal to the measured
ferent concrete mixes with water/cement ratios (w/c) ranging transmission length, the force reached during the pull-out opera-
from 0.3 to 0.5, cement content (C) from 350 to 500 kg/m3 and a tion before the strand slip (P A ) is slightly greater than the effective
compressive strength at the time of testing fci ranging from 24 to prestressing force PE. This fact indicates that the transmission
55 MPa. The concrete components were: cement CEM I 52.5 R length obtained for the adopted embedment length sequence is
[32], crushed limestone aggregate 7/12, washed rolled limestone somewhat longer than the real transmission length.

Pretensioning frame
Embedment length AMA
Adjustable system
anchorage

Hydraulic jack Displacement Concrete Sleeve Hollow force Anchorage


Strand
transducer specimen transducer

Fig. 5. Test equipment.


408 J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412

180 PE in Eq. (18) was chosen instead of PA to directly consider the re-
Anchorage length sults obtained after the stabilization period of the prestress transfer
PR
160
from the embedment length sequence tested at a resolution of
140 50 mm. In this way, slighter average bond stresses than the real
ones were determined for the transmission zone. The P A value will
Strand force (kN)

120 coincide with the PE value if the transmission length is a multiple


100 point of the embedment length resolution.
The effect of concrete strength on the average bond may be
80 illustrated by redefining UT and UC as [11,21,24]:
Transmission length
60
U T ¼ U 0T ðfci Þa ð20Þ
40 Pull-out force at end slip
20 Force after transfer and stabilization U C ¼ U 0C ðfci Þa ð21Þ
Force before release
In order to determine U 0T and U 0C and the appropriate a expo-
0
0 100 200 300 400 500 600 700 800 900 1000 nent, several regression analyses of the test results have been car-
Embedment length (mm) ried out by substituting UT and UC in Eqs. (18) and (19),
respectively, for Eqs. (20) and (21). Figs. 7 and 8 show the obtained
Fig. 6. Test results for concrete C400/0.35. adjustments. Therefore, the proposed equations for both transmis-
sion and complementary bond lengths derived from the test re-
sults of this study are:
6. Results and discussion

PE
6.1. Test results LT ¼ ð22Þ
Rp 0:4fci2=3
For each specimen, the prestress transfer and the pull-out of the
strand operations performed with the ECADA test method have PA  P A
LC ¼ ð23Þ
been carried out sequentially. For each concrete mix, transmission Rp 0:25fci2=3
and anchorage lengths are determined from a series of specimens
with different embedment lengths. From the adjustments, the obtained UT/UC ratio is 1.6 (0.4/0.25);
Table 2 summarizes the main results for all the concrete mix consequently, anchorage length can be obtained from the follow-
designs. The effective prestressing force PE is the average value of ing equation:
the force in the prestressing strand in those specimens with an
embedment length equal to or longer than the transmission length 2:5
obtained by the ECADA test method for each concrete mix design LA ¼ ½PE þ 1:6ðP A  PA Þ ð24Þ
Rp fci2=3
after the stabilization period.
Figs. 9–11 show the comparisons of the predicted values from
6.2. Proposed bond model Eqs. (22), (23), (24), respectively, versus the measured transmis-
sion lengths, complementary bond lengths and anchorage lengths.
The average bond stresses values along both transmission and The quality of the adjustments is comparable to that obtained in
complementary bond lengths from the measured data in this [24]. Therefore, the 50 mm resolution applied to determine these
study, according to Eqs. (1) and (2), are obtained from the follow- lengths from the sequences of specimen lengths is reliable.
ing equations: In this experimental study for the bond characterization of
13 mm prestressing steel strand, the testing loading time coincides
PE
UT ¼ ð18Þ with the time of the prestress transfer (fcl = fci). For fcl > fci, the UC
Rp LT values can be expected to be above the obtained UC values. As a re-
sult, when fcl > fci, Eq. (23) is conservative for the complementary
P A  P A
UC ¼ ð19Þ bond length prediction, while Eq. (24) proves conservative for
Rp LC the anchorage length prediction.

Table 2
Test results from the experimental program.

Designation Cement (kg/m3) w/c ratio fci (MPa) PE (kN) P A (kN) PA (kN) LT (mm) LA (mm) LC (mm)
C350/0.50 350 0.50 26.1 132.49 144.47 158.49 550 650 100
C350/0.45 0.45 37.3 131.49 143.63 161.97 550 700 150
C350/0.40 0.40 46.7 131.49 137.72 159.47 550 700 150
C400/0.50 400 0.50 24.2 127.01 140.35 166.69 650 850 200
C400/0.45 0.45 28.3 133.88 135.02 157.93 550 700 150
C400/0.40 0.40 41.4 129.90 145.20 164.86 550 700 150
C400/0.35 0.35 45.3 131.09 140.36 159.02 500 600 100
C450/0.40 450 0.40 36.3 129.30 137.42 159.20 550 700 150
C450/0.35 0.35 46.6 130.10 131.90 161.35 500 650 150
C500/0.40 500 0.40 30.8 128.90 133.08 158.02 600 800 200
C500/0.35 0.35 46.6 132.19 138.46 160.82 450 600 150
C500/0.30 0.30 54.8 129.50 133.80 169.24 400 600 200
Average 38.7 130.61 138.45 161.42 533 687 154
J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412 409

7 800
Average measured bond

Predicted transmission length (mm)


700
y = 0,40x
stress Ut (MPa)

5
600
4 R2 = 0,56
3 500

2
400
1

0 300
0 2 4 6 8 10 12 14 16
fci 2/3 (in MPa) 200

Fig. 7. Bond stress regression analysis based on transmission length test results. 100

0
7 0 100 200 300 400 500 600 700 800
Average measured bond

6 Measured transmission length (mm)


stress Uc (MPa)

5 Fig. 9. Predicted versus measured transmission lengths.

4 y = 0,25x
300

Predicted complementary bond length (mm)


3

2 250

200 R2 = 0,63
0
0 2 4 6 8 10 12 14 16

fci 2/3 (in MPa)


150
Fig. 8. Bond stress regression analysis based on complementary bond length test
results.
100

According to the notation used in Table 1, Eqs. (22) and (24) can
be rewritten as Eqs. (25) and (26), respectively: 50

2:5Ap rpi
LT ¼ ð25Þ 0
Rp fci2=3 0 50 100 150 200 250 300
Measured complementary bond length (mm)
2:5Ap
LA ¼ ½rpi þ 1:6ðrpa  rpa Þ ð26Þ Fig. 10. Predicted versus measured complementary bond lengths.
Rp fci2=3
1000
Term rpa in Eq. (26), obtained as P A =Ap , coincides with rpi when P A
900
= PE in the ECADA test methodology. For a general case, rpa should
Predicted anchorege length (mm)

be replaced with rpi in Eq. (26). 800


In order to obtain equations for design, additional experimental
works on transmission and anchorage lengths with fcl > fci should 700 R2 = 0,53
be conducted, and term rpa in Eq. (26) should be replaced with
rpcs. Moreover, the 95% confidence intervals for the transmission 600
and anchorage lengths should be established.
500

400
6.3. Comparison with others research works and code provisions
300
The experimental results obtained in this study have been com-
pared with the theoretical predictions obtained from the equations 200
included in Table 1 and the proposed equations by considering the
experimental conditions of this study in all cases. For this purpose, 100
the following relationships have been adopted: rpt = 0.75fpu = 1395
0
MPa, rpi = rpcs = 1310 MPa (obtained on average as PE/Ap), rpa = 0 100 200 300 400 500 600 700 800 900 1000
0.9fpu = 1674 MPa (implies PA = 166.88 kN, average LA = 725 mm Measured anchorage length (mm)
and average LC = 192 mm by extrapolation with the experimentally
obtained UC values) and fcl = fci = 38.7 MPa. Fig. 11. Predicted versus measured anchorage lengths.
410 J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412

Proposed equations Eq. (22) (23) 1,6

EC2-2004/MC2010 [4,5] Eq. (16) (17) (gradual) 2,7 Ut / Uc


ratio
Kose and Burkett 2005 [25] Eq. (15) 3,5

Lane 1998 [23] Eq. (13) 2,6

Tadros and Baishya 1996 [22] Eq. (12) 3,0


Mitchell et al.1993 [21] Eq. (11) 3,7

Buckner 1995 [20] Eq. (10) (λ


(l=2)
=2) 6,0

Buckner 1995 [20] Eq. (10) (λ(l=1)


=1) 3,0

Deatherage et al. 1994 [19] Eq. (9) 4,5

Shahawy (et al.1992)-93 [17,18] Eq. (8) (k=0,5) 6,0

Shahawy (et al.1992)-93 [17,18] Eq. (8) (k=2) 1,5

Cousins et al. 1990 [11] Eq. (7) 4,9

Zia and Mostaf a 1977 [16] Eq. (6) (gradual) 6,0

Martin and Scott 1976 [15] Eq. (5) 7,0

ACI 318-11 [2] Eq. (4) 3,0

0 1 2 3 4 5 6
experimental values for Ut Ut for fci 38,7 MPa (Ut)
Bond stress (MPa)
experimental values for Uc Uc for fcl 38,7 MPa (Uc)

Fig. 12. Prediction of bond stress based on test results according to proposed and pre-existing equations.

Proposed equations Eq. (25) (26) 0


0,69 LT / LA
ratio
EC2-2004/MC2010 [4,5] Eq. (16) (17) (gradual) 0,57

Kose and Burkett 2005 [25] Eq. (15) 0,51


Lane 1998 [23] Eq. (13) 0,60
Tadros and Baishya 1996 [22] Eq. (12) 0,60

Mitchell et al.1993 [21] Eq. (11) 0,49

Buckner 1995 [20] Eq. (10) (λ=2)


(l=2) 0,37

λ=1)
Buckner 1995 [20] Eq. (10) ((l=1) 0,55
Deatherage et al. 1994 [19] Eq. (9) 0,44
Shahawy (et al.1992)-93 [17,18] Eq. (8) (k=0,5) 0,37
Shahawy (et al.1992)-93 [17,18] Eq. (8) (k=2) 0,71

Cousins et al. 1990 [11] Eq. (7) 0,42


Zia and Mostaf a 1977 [16] Eq. (6) (gradual) 0,37
Martin and Scott 1976 [15] Eq. (5) 0,34

ACI 318-11 [2] Eq. (4) 0,55

0 500 1000 1500 2000 2500 3000 3500


measured transmission lengths Transmission length for fci 38,7 MPa (LT)
Member length (mm)
measured anchorage lengths Anchorage length for fci 38,7 MPa (LA)

Fig. 13. Prediction of transmission and anchorage lengths test results according to proposed and pre-existing equations.

Fig. 12 shows the comparison for the average bond stresses out in the set of predictions as it comes closer to the obtained
along both the transmission and anchorage lengths, while Fig. 13 results.
shows the comparison for transmission length and anchorage The predicted UT/UC ratios are 1.5–7, as observed in Fig. 12, with
lengths. These figures also include the average values obtained an average value of 4. A ratio of 4.0 was derived to correlate trans-
for UT (4.6 MPa), UC (2.8 MPa), LT (533 mm) and LA (725 mm). mission to pull-out bond stress–slip relationships [27]. Other the-
Fig. 12 depicts the wide ranges of predicted values. For UT, oretical studies [20] indicate values of 1–8 for the UT/UC ratio, with
Eqs. (6a) and (11a) provide a good prediction of the experimental an average value of 2.4. Moreover, experimental results with UT/
results of this study. For UC, results are greater than the predicted UC = 1.4 (in beams) are presented in [21], and are offered in [31]
values. Only the prediction made by Eq. (8b) with k = 2 stands with UT/UC = 2 (in cylindrical concrete specimens). The UT/UC ratio
J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412 411

obtained in this work is 1.6, similar to the aforementioned exper- anchorage lengths values that vary considerably and differ from
imental results and the prediction by Eq. (8) with k = 2. each other.
Fig. 13 shows that Eqs. (6a) and (11a) offer a good prediction of  The predicted transmission length generally overestimates the
the average measured LT in this study. Generally, the measured measured transmission length, with predictions that provide
transmission length is over evaluated by the remaining equations, transmission length values more than twice the measured
with predictions that provide transmission length values more than transmission lengths.
twice the measured transmission lengths. Similar experimental re-  From the experimental results of this study a high LT/LA ratio has
sults for transmission length are presented in [15,16,18,23,24]. been obtained.
For anchorage lengths, and in agreement with the greater UC in
relation to that predicted, the test results are distinguished by
short lengths (see Fig. 13), resulting in a poor prediction of the
Acknowledgments
experimental anchorage lengths from the equations found in the
literature. Similar experimental results are found in [11] for coated
The content of this article forms part of the research work that
strands, and in [20] for uncoated strands with fci = 48 MPa and
the Institute of Concrete Science and Technology (ICITECH) is pres-
fcl = 65 MPa.
ently conducting in conjunction with PREVALESA and ISOCRON.
The predicted LT/LA ratios are 0.34–0.71, as observed in Fig. 13,
This study has been funded by the Spanish Ministry of Education
with an average value of 0.5. The LT/LA ratio obtained from the equa-
and Science and ERDF (Project BIA2006-05521). The authors wish
tions proposed in this study is 0.69 (LT/LA = 528/ 763 – the experi-
to thank the above companies as well as the concrete structures
mental ratio is LT/LA = 533/725 = 0.73 –), indicating that the
laboratory technicians at the Universitat Politècnica de València
complementary bond lengths obtained are relatively short. In addi-
for their cooperation. Finally, the authors also wish to pay their re-
tion to the proposed equations, the predicted ratio of 0.71 by Eq. (8)
spects to C.A. Arbeláez.
with k = 2 is the best prediction of the experimental LT/LA ratio.

References
7. Conclusions
[1] FIB. Bond of reinforcement in concrete. State of the art report. Fib Bulletin No.
An experimental program to determine transmission and 10. Lausanne: International Federation for Structural Concrete; 2000.
anchorage lengths and the average bond stress along both the [2] ACI Committee 318. Building code requirements for reinforced concrete (ACI
318-11). Farmington Hills, MI: American Concrete Institute; 2011.
transmission length and the complementary bond length of [3] CEB. Anchorage zones of prestressed concrete members. Bulletin d’Information
13 mm prestressing steel strand has been conducted by means of No.181. Lausanne: Comité Euro-International du Béton; 1987.
the ECADA test method. The influence of concrete strength on [4] CEN. Eurocode 2: design of concrete structures – Part 1–1: general rules and
rules for buildings. European standard EN 1992-1-1:2004:E. Brussels: Comité
transmission and anchorage lengths has been analyzed.
Européen de Normalisation; 2004.
The main conclusions drawn from this experimental study are: [5] FIB. Model Code 2010. First complete draft. Fib Bulletin No. 55, vol. 1.
Lausanne: International Federation for Structural Concrete; 2010.
 An increase of the concrete compressive strength at the testing [6] AENOR. UNE 36094:1997 Alambres y cordones de acero para armaduras de
hormigón pretensado. Madrid: Asociación Española de Normalización y
time results in an increase of the bond stress along both the Certificación; 1997.
transmission and complementary bond lengths. However, this [7] ASTM. A416/A416M-10 standard specification for steel strand, uncoated
fact is not considered in the current ACI 318 Code provisions. seven-wire for prestressed concrete. West Conshohocken, PA: American
Society for Testing and Materials; 2010.
 An average bond stress along the transmission length as a func- [8] Martí-Vargas JR, Serna-Ros P, Fernández-Prada MA, Miguel-Sosa PF, Arbeláez
tion of the concrete compressive strength at the time of the pre- CA. Test method for determination of the transmission and anchorage lengths
stress transfer has been obtained as UT = 0.4 fci2=3 . in prestressed reinforcement. Mag Concr Res 2006;58(1):21–9.
[9] Martí-Vargas JR, Arbeláez CA, Serna-Ros P, Castro-Bugallo C. Reliability of
 An average bond stress along the complementary bond length transfer length estimation from strand end slip. ACI Struct J
as a function of the concrete compressive strength at loading 2007;104(4):487–94.
has been obtained as UC = 0.25 fcl2=3 . [10] Guyon Y. Béton précontrainte. Étude théorique et expérimentale. Paris: Ed.
Eyrolles; 1953.
 The obtained UT/UC ratio is 1.6, which is in agreement with [11] Cousins ThE, Johnston DW, Zia P. Transfer and development length of epoxy-
other experimental results reported by other authors. coated and uncoated prestressing strand. PCI J 1990;35(4):92–103.
 With these relationships for UT and UC, the estimation of the [12] Janney J. Nature of bond in pretensioned prestressed concrete. ACI J
1954;25(9):717–37.
transmission length, the complementary bond length, and the
[13] Barnes RW, Grove JW, Burns NH. Experimental assessment of factors affecting
anchorage length is a good adjustment to the lengths measured transfer length. ACI Struct J 2003;100(6):740–8.
for the 50 mm resolution in the specimen lengths sequences, [14] Brearley LM, Johnston DW. Pull-out bond tests of epoxy-coated prestressing
strand. J Struct Eng – ASCE 1990;116(8):2236–52.
and is reliable.
[15] Martin L, Scott N. Development of prestressing strand in pretensioned
 The following equation to predict the transmission length of members. ACI J 1976;73:453–6.
13 mm prestressing steel strand is proposed: [16] Zia P, Mostafa T. Development length of prestressing strands. PCI J
1977;22(5):54–65.
2:5Ap rpi [17] Shahawy M, Moussa I, Batchelor B. Strand transfer lengths in full scale AASHTO
LT ¼ prestressed concrete girders. PCI J 1992;37(3):84–96.
Rp fci2=3 [18] Shahawy M. An investigation of shear strength of pretensioned concrete
AASHTO Type II Girders. Florida: Structures Research Center, FDOT; 1993.
 The following equation to predict the anchorage length of [19] Deatherage JH, Burdette E, Chew ChK. Development length and lateral spacing
13 mm prestressing steel strand when the testing loading time requirements of prestressing strand for prestressed concrete bridge girders.
coincides with the prestress transfer time is proposed: PCI J 1994;39(1):70–83.
[20] Buckner CD. A review of strand development length for pretensioned concrete
2:5Ap members. PCI J 1995;40(2):84–105.
LA ¼ ½rpi þ 1:6ðrpa  rpi Þ [21] Mitchell D, Cook WD, Khan AA, Tham Th. Influence of high strength concrete
Rp fci2=3 on transfer and development length of pretensioning strand. PCI J
1993;38(3):52–66.
 The test results obtained in this study have been compared with [22] Tadros MK, Baishya MC. Discussion of A review of strand development length
the theoretical predictions obtained from the different equations for pretensioned concrete members. PCI J 1996;41(2):112–27.
[23] Lane SN. A new development length equation for pretensioned strands in
proposed by several authors and codes to determine transmis- bridge beams and piles. Research FHWA-RD-98-116. Mclean, VA: Federal
sion and anchorage lengths. Predictions give transmission and Highway Administration; 1998.
412 J.R. Martí-Vargas et al. / Engineering Structures 41 (2012) 403–412

[24] Mahmoud ZI, Rizkalla SH, Zaghloul ER. Transfer and development lengths of [29] ACI Committee 318. Building code requirements for reinforced concrete (ACI
carbon fiber reinforcement polymers prestressing reinforcing. ACI Struct J 318-63). Detroit, MI: American Concrete Institute; 1963.
1999;96(4):594–602. [30] Tabatabai H, Dickson Th. The history of the prestressing strand development
[25] Kose MM, Burkett WR. Formulation of new development length equation for length equation. PCI J 1993;38(5):64–75.
0.6 in. prestressing strand. PCI J 2005;50(5):96–105. [31] Abrishami HH, Mitchell D. Bond characteristics of pretensioned strand. ACI
[26] Hegger J, Bülte S, Kommer B. Structural behaviour of prestressed beams made Mater J 1993;90(3):228–35.
with self-consolidating concrete. PCI J 2007;52(4):34–42. [32] CEN. Cement. Part 1: compositions, specifications and conformity criteria for
[27] Pozolo A, Andrawes B. Analytical prediction of transfer length in prestressed common cements. European standard EN 197-1:2000. Brussels: Comité
self-consolidating concrete girders using pull-out test results. Constr Build Européen de Normalisation; 2000.
Mater 2011;25:1026–36. [33] Belhadj A, Bahai H. Friction-slip: an efficient energy dissipating mechanism for
[28] Martí-Vargas JR, Serna P, Navarro-Gregori J, Bonet JL. Effects of concrete suddenly released prestressing bars. Eng Struct 2001;23:934–44.
composition on transmission length of prestressing strands. Constr Build [34] Moon DY, Zi G, Kim JH, Lee SJ, Kim G. On strain change of prestressing strand
Mater 2012;27:350–6. during detensioning procedures. Eng Struct 2010;32:2570–8.