STUDY AND DESIGN OF ANCHORAGE CORNER BLISTERS OF

POST-TENSIONING TENDONS

Carla Marchão1, Válter Lúcio1 and Hans Rudolf Ganz2

1
CERIS/ICIST, Civil Engineering Department, Faculty of Sciences and Technology, Universidade
NOVA de Lisboa, Caparica, Portugal
2
Ganz Consulting, Switzerland

ABSTRACT
The anchorage corner blisters of internal continuity post-tensioning tendons of bridges built with the
cantilever method are singular zones where the simplified formulas for the design of the reinforcement
required to resist transverse tensile forces due to the application of point loads, proposed by the current
standards, are not sufficient as they do not cover all the effects that require reinforcement. Besides being
eccentric and intermediate anchorages, as the anchorages are located in blisters, there is the need to transfer
the prestressing force to the bottom slab and web of the girder.
In addition to the difficulties in design, during construction the high density of steel reinforcement in
anchorage blisters is often responsible for the existence of zones with low concrete compacity, which may
lead to concrete crushing failures under the anchor plates. To reduce the density of steel reinforcement, in
the research project which is the basis of this paper a high-performance fibre reinforced self-compacting
mixture (HPFRC) was used
An experimental program was carried out to study the transmission of prestressing force to the slab and web
of the box girder, assess the strut-and-tie models used in design and to study the feasibility of using a high-
performance fiber reinforced self-compacting mixture (HPFRC) only in the blister.
It can be concluded that the use of HPFRC in anchorage blisters is a very interesting solution regarding the
saving in materials and the reduction of the steel reinforcement density near the local anchorage zone.

Keywords: post-tensioning, anchorage zones, high performance concrete, box girder bridge, steel fibres.

1. Introduction
Previous studies (Haroon et al.,2006; Leung et al., 2009; Gomes, F. and Nunes, S., 2010) have shown the
efficiency of high performance concretes with regard to the reduction of confining steel reinforcement in
anchorage zones. So, with the aim of decrease the amount of reinforcement needed in anchorage blisters
and also reduce its concrete cross-section a high performance fiber reinforced self-compacting mixture was
developed. It is noteworthy that, because of its high strength, the volume of HPFRC needed is about 1/3 to
1/2 of that of conventional concrete for comparable structural elements (Ganz, 2008).
With the aim of studying the behavior of anchorage blisters regarding the transmission of stresses to the
web and the bottom slab of the box girder, as well as the feasibility of using high performance concrete only
in the blister, three half scale models of the inferior corner of a box girder existing bridge were studied: a
reference specimen of ordinary reinforced concrete (ORC) and two high performance fiber reinforced
concrete (HPFRC) blister specimens.
The specimens were self-balanced with anchorages at both ends of the tendons and included a corner blister
for two anchorages with a portion of the slab and the contiguous web. In the HPFRC blister specimens, the
section of the blister was reduced.
The design of the reinforcement of the specimens was based in the tensile forces obtained from strut-and-
tie models. The strut-and-tie models were developed taking into consideration the stress fields obtained
from a linear-elastic analysis.
2. Experimental Program

2.1 Specimens geometry

The Lezíria Bridge over the Tagus River, near Lisbon, was used as a prototype for the specimens considered
in this study. A half-scale model of the inferior corner of the box girder was considered. A self-balanced
model with anchorages at both ends of the tendons was chosen, as shown in Figure 1. The model includes
a corner blister for two anchors, with a portion of the slab and the contiguous web.

Fig. 1. View of the reference specimen

The length of the specimens was defined taking into account the balance between the need to disperse the
compressive forces between anchorages and the optimization of dimensions in order to obtain the smallest
specimen (see Figure 2 for the case of the reference specimen).
The stressing anchorage used was a VSL type GC for seven strands of 0.6”, with a 180×180mm2 bearing
plate. The typical cross-section of the specimens is presented in Figure 3a and was defined based on the
dimensions of the prototype. The width and height of the cross-section were chosen to ensure an adequate
distribution of the prestressing force, taking into account the position of the prestressing tendons.

Fig. 2. Side view

For the reference specimen, both minimum center distance and edge distance of anchorages, recommended
by VSL International (2012) for the type of anchorage used, were considered. For the HPFRC blister model
the adopted center distance and edge distances of anchorages were assessed through load transfer tests
(Marchão, 2014). In both specimens, the necessary distances from anchorages to bottom slab and web were
defined in order to provide sufficient clearance for the stressing jack. The geometry of the blisters resulted
in those shown in Figure 3.
 (a) Reference specimen (b) HPFRC blister specimens

Fig. 3. Cross-section C-C

The tendons layout, whose profile is schematically shown in Figure 1, was set to minimize the post-
tensioning friction losses and the size of the anchorage blister.

2.2 Strut-and-tie models for design and detailing

The strut-and-tie models were developed taking into account the stress trajectories obtained from the linear-
elastic analysis and the need to keep the equilibrium with the applied loads in the ultimate limit state. In the
choice of the ties adopted position and direction has been taken into account the fact that the ties should
coincide with the corresponding reinforcement. An example of the strut-and-tie models adopted in the
longitudinal sections is shown in Figure 4.

Fig. 4. Example of strut-and tie model – longitudinal section along a tendon

As shown in Figure 4, in the longitudinal section, due to the dispersion of the concentrated prestressing
force over the cross-section, it was assumed that the reinforcement needed should be parallel to the front
face of the blister (perpendicular to the applied force). The resultant tie force (tie 1) was located in the
theoretical section where the compression stress field, due to the dispersion of the concentrated force,
intersects the upper face of the blister.
In the slab between anchorages, the compression was considered to be uniformly distributed across the slab
thickness, so that, two equal struts located at one quarter and three quarters of the regular slab thickness
were adopted. According to Rogowsky and Marti (1996), a tie back force of 25% of the prestressing force
was assumed, uniformly distributed across the slab thickness.
The transverse ties 2 were designed to resist to half of the radial forces due to tendon deviation. The other
half was resisted by the concrete struts that converge in each top node. The orientation of the left strut of
each node was determined by the establishment of the equilibrium at the node. The corbel and local bending
effects (ties 3 and 4, respectively) were also considered.
The design of the specimen’s reinforcement was made based on the tensile forces obtained with the strut-
and-tie models previously presented. By the fact that the force applied during the test was limited to the
failure load of the prestressing tendons and assuming that concrete failure was not relevant, the steel
reinforcement used in the blister zone was intentionally undersized with the purpose of trying to reach the
failure of the model during the test. So, the reinforcement was designed considering a post-tensioning force
of half of the specified characteristic value of 0,1% proof force (Fp0,1k) for seven strands type 0,6" (with a
nominal area of 150 mm2) and a reinforcing steel tensile strength of 600 MPa, since the reinforcing steel
grade that was planned to use would have a characteristic yield stress of fyk = 500 MPa.
The reinforcement detailing is presented in Figure 5. Comparing both specimens, it can be observed that, in
addition to saving materials, the great advantage of adopting HPFRC in the anchorage blister was the
decrease of the steel reinforcement density near the local anchorage zone, with the obvious improvements
in concrete quality.

(a) Reference specimen

(b) HPFRC blister specimens

Fig. 5. Reinforcement in the anchorage blister zone

2.3 Specimens execution

The specimens were produced in a precast company. The concrete strength class requested was a C30/37
and a reinforcing steel with a characteristic yield stress of fyk = 500 MPa and ductility class C was used.
In the HPFRC blister specimens only the blister was cast with HPFRC. The remaining elements (bottom
slab, web and end block) were cast with ordinary concrete. In the case of the HPFRC precast blister
specimen, the blister was cast before the slab and web. To allow the precast blister to settle in the
reinforcement of the bottom slab, simplifying its positioning during the casting of the remaining model, its
dimensions plus 15 mm was considered corresponding to the concrete cover of the web and bottom
slab, as shown in Figure 6.

Fig. 6. Side view of the HPFRC precast blister specimen

The HPFRC blister was cast upside down because there was the need to leave the reinforcement of the
anchorage blister in standby to make the subsequent connection to the bottom slab and web of the model. To
increase bond at the interface between the two concretes, the surfaces in contact with the bottom slab and web
were roughened with a pneumatic chipping hammer.
The characteristics of the materials at the date of the tests are presented in Tables 1 and 2. In of the HPFRC
precast blister specimen Table 1, the concrete mean compressive strength (fcm,cube) is referred to 150 mm width
cubes.

Table 1. Concrete characteristics

Specimen Reference HPFRC blister HPFRC precast blister

Type of concrete OC OC HPFRC OC HPFRC
fcm,cube (MPa) 34.6 46.1 117.1 49.2 126.7

Table 2. Steel reinforcement characteristics

Bar diameter (mm) 6 8 10 12 20

fy (MPa) 564 535 542 557 546

2.4 Test procedure and measurements

The bottom surface of the specimens was not levelled, so the tests were performed with the specimens
rotated in relation to the actual position, as shown in Figure 7.

Fig. 7. Specimen before testing

The tendons were tensioned at the end block and in order to quantify the prestressing force value in the
anchorage blister, two TML KCM-2MNA load cells with 2000 kN capacity were used.
A total of 66 and 8 strain gauges, for the reference specimen and for the HPFRC blister specimens,
respectively, were affixed to diametrically opposite sides of selected reinforcing bars. The crack openings at
the concrete surface were measured with relative displacement transducers TML type PI on the locations that
corresponded to the zones with the highest tensile stresses. The displacement of the anchorage blister in
relation to the bottom slab and web was measured using TML CDP displacement transducers.
The adoption of larger diameter ducts or anchorage bearing plates for a larger number of strands would
misrepresent the behavior of the specimen, so the elements adopted were the ones usually used for seven
strand tendons. Obviously, this decision limited the test, since the force that could be applied was dependent
on the failure force of the prestressing strands. Thus, in a first phase of the reference specimen test, two
tendons with 7 strands of 0.6” type prEN 10138-3(2006) Y1860S7 with a nominal cross-section of 150 mm2
each were used. Given that for the maximum tensioning force of the tendons the model was far from failure,
it was decided to perform a second phase, using 9 strands of the same type in each duct. It was also decided
to use 9 strands in each tendon for the subsequent tests.

2.5 Experimental results

2.5.1 Reference specimen

The prestressing force was applied in increments of 154.7kN up to 1701.7 kN in the first phase of the test,
and in increments of 133.6 kN up to 1870.4 kN in the second phase. The evolution of the blister
displacements with the average measured load is presented in Figure 8. As can be seen, the measured
displacements are smaller as closer to the web the transducer is. A residual displacement at the end of the
first phase of the test can be observed.

Fig. 8. Evolution of the blister displacements with the average load

Figure 9 presents the evolution of the relative displacements in the concrete surface with the average load
for relative displacement transducers 3 and 5, which recorded crack openings greater than 0.3 mm that can
be seen in Figure 10.

(a) Relative displacement transducer 3 (b) Relative displacement transducer 5

Fig. 9. Concrete surface relative displacements vs average load

 Fig. 10. Location of relative displacement transducers 3 and 5
The evolution of the strains of the reinforcing steel with the average measured load is presented in Figure 11,
for strain gauges 16 to 22, which recorded the larger strain values.

Fig. 11. Strain of the reinforcing steel vs average measured load

Although the strain values are not directly related to the tensile force on the ties of the strut and tie models
used in the design, but with the steel reinforcement stresses, it can be seen that:
 Strain gauges 16 to 22 have large strains because they are located in the reinforcement that corresponds
to tie 1 of the strut-and-tie longitudinal models (see Figure 4), which is the one with largest tensile
force. In this case, the ratio between the adopted and the required reinforcement is 1.0 and 1.3 for
tendons A and B, respectively.
 Strain gauges 16, 18, 19 and 21 present similar strains because they are located in the legs of the stirrups
between the two tendons. Among these, the strain gauges 19 and 21 have the largest strain, since they
are located in the reinforcement that corresponds to tie 1 of tendon A (see Figure 4), which is the one
with largest tensile force and smallest ratio between the adopted and the required reinforcement.

2.5.2 HPFRC blister specimen

The tendons were prestressed to 83.7% of the tensile strength, value that corresponds to the failure of the
first wires. The prestressing force was applied in increments of 125.1 kN up to approximately 2101.3 kN.
The evolution of the blister displacements with the average measured load is presented in Figure 12.

Fig. 12. Evolution of the blister displacements with the average measured load
The largest measured crack widths correspond to the relative displacement transducers 3, 4, 6 and 7 and are
below 0.15 mm. The variation of the concrete surface relative displacements with the average measured
load, for these transducers (whose location is shown in Figure 14), is presented in Figure 13.

(a) Relative displacement transducer 3 (b) Relative displacement transducer 4

(c) Relative displacement transducer 6 (d) Relative displacement transducer 7

Fig. 13. Concrete surface relative displacements vs average load

 4 6

Fig. 14. Location of relative displacement transducers 3, 4, 6 and 7

2.5.3 HPFRC precast blister specimen

The tendons were tensioned to 89.7% of the tensile strength, corresponding to the failure of the first strand.
The prestressing force was applied in increments of 125.1 kN up to approximately 2251.4 kN. The evolution
of the blister displacements with the average measured load is presented in Figure 15. Apparently, transducer
1 did not work properly up to the load of 875 kN.

Fig. 15. Evolution of the blister displacements with the average measured load
The measured crack widths were similar to the ones obtained for the previous specimen, except for the
crack between the anchorage blister and the bottom slab which has reached a width of approximately
0.3 mm, as can be seen in Figure 16.

Fig. 16. Concrete surface relative displacements vs average load

Regarding the safety and serviceability requirements for the blister, the most unfavorable situation
corresponds to the application of the post-tensioning force (since the prestressing force is lower after
tensioning due to the wedge draw-in of the anchorage device and time dependent losses), so long term-tests
were not performed. In future works is recommended to study the influence of the HPFRC creep in the
behaviour of the blister.

3. Conclusions
Comparing the test results of the HPFRC blister specimens, the two models showed a very similar behavior
during the test satisfying all requirements for serviceability and strength. No significant difference
regarding cracking was observed, except for the crack formed in the connection between the blister and the
bottom slab. Measurements of the relative displacement transducers in that zone showed that the relative
displacement become higher for the precast blister model from a load value of approximately 1250 kN.
However, the crack width did not exceed 0.3 mm for the maximum load achieved in the test.
Regarding the steel reinforcement, although the strain values are not directly related to the tensile force on
the ties of the strut-and-tie models used for design but with the steel reinforcement stress, in most cases the
largest strains were measured in the strain gauges located in the reinforcement that corresponded to ties with
largest tensile force. The measured values were all below the steel yielding strain.
The use of HPFRC in anchorage corner blisters is a very interesting solution regarding the saving of
materials and the reduction of the steel reinforcement density near the local anchorage zone, with the
obvious advantages in concrete quality.
The adoption of solutions with HPFRC precast blisters avoids producing and casting with two different
concretes on site. This solution can be valuable regarding quality control in the execution of HPFRC blisters.

4. Acknowledgements
Collaboration and materials supplying by VSL, Concremat, Secil, Comital, MC-Bauchemie, Sika, and
KrampeHarex is gratefully acknowledged.

References
EN 1992-1-1 (2004), Eurocode 2 – Design of Concrete Structures – Part 1.1: General rules and rules for
buildings. EN 1992-1-1:2004. European Committee for Standardization, Brussels.
Ganz, H. R. (2008), Prestressed structural concrete: New developments and applications. Encontro
Nacional Betão Estrutural 2008, Guimarães, Portugal, pp. 3-20.
Gomes, F., Nunes, S. (2010), Betão auto-compactável reforçado com fibras. Aplicação em zonas de ancoragem
de cabos de pré-esforço. Encontro Nacional Betão Estrutural 2010, Lisboa, Portugal, pp. 3-20.
Haroon, S., Yazdani, N., and Tawfiq, K. (2006), Posttensioned anchorage zone enhancement with fiber-
reinforced concrete. Journal of Bridge Engineering, Vol. 11, No. 5, pp. 566–572.
Leung, C. K. Y., Zhang, X., and Xia, Q. (2009), The application of pseudo-ductile cementitious composites
in the anchorage zone of post-tensioned concrete members. Materials and Structures, Vol. 42,
pp. 1221–1231.
Marchão, C., Lúcio, V., Ganz, H. R.(2014), Application of a high performance fiber reinforced self-
compacting concrete in post-tensioning anchorage zones. Fib Congress 2014, Mumbai.
Rogowsky, D., Marti, P. (1996), Detailing for Post-Tensioning. VSL Report Series 3. VSL International
Ltd., Berne, Switzerland.
VSL International (2012), European Technical Approval Nº ETA-06/0006 – Version of 4th May 2012.
SETRA. European Organisation for Technical Approvals, Paris.