Concrete Repair, Rehabilitation and Retrofitting II Alexander et al (eds)

2009 Taylor & Francis Group, London, ISBN 978-0-415-46850-3

FRP tendon anchorage in post-tensioned concrete structures

J.W. Schmidt & B. Tljsten
Technical University of Denmark, Kgs. Lyngby, Denmark

A. Bennitz
Lule University of Technology, Lule, Sweden

J.W. Schmidt & H. Pedersen

COWI A/S, Lyngby, Denmark

ABSTRACT: Strengthening of building structures by the use of various external post-tensioning steel tendon
systems, is known to be a very efficient method. However, FRP as material in external post-tensioning projects
has been investigated during the last decade. The advantages for this material are the high effective Youngs
modulus and the high stress capacity in the linear elastic range of the material. The use of external tendons
increases the requirements on the anchorage systems. This is in particular important when using un-bonded
tendon systems, where the anchorage and deviators are the only force transfer points. The demand for high
capacity anchorage tendons is fulfilled for steel tendons, but no competitive mechanical anchor has yet been
developed for FRP tendon. A new small, reliable and more user friendly anchor has to be developed, before
FRP tendons can be utilized with all of its capacity. Thus, several attempts of developing a mechanical FRP
anchor have been made worldwide with promising results. Some of these attempts are presented in this paper
together with an insight into a present research collaboration project at the Technical University of Denmark,
Lule University of Technology, Sweden, and COWI A/S, Denmark.

INTRODUCTION

The implementation of external post-tensioning systems

is a very efficient way to upgrade existing concrete
structures. Traditional post tensioning steel systems
have been used in this area for decades. Often the
same anchorage systems as in ordinary bridge constructions are used. It seems obvious to look for an
application where Carbon Fibre Reinforced Polymer (CFRP) rods are implemented due to their high
corrosion resistance, high stress caspacity and low
weight. As for steel tendons the real benefit from the
materials high stress capacity can only be achieved
when the CFRP rod is prestressed and installed in
this state. By this the condition of the concrete structure is improved in the serviceability limit state (e.g.
crack widths are decreased) and the load capacity
of the structure is also increased to a desired level
(due to the compression introduced into the concrete
structure).
Standard anchor systems for steel tendons have
been in use for several decades. The modelling of
such devices is simple, since steel is an isotropic
material making it possible to determine the shear

stresses obtained by introducing a normal stress

perpendicular to the tendon.
CFRP is, however, an anisotropic material with
low compression strength in the transverse direction
of the fibres. It is difficult, due to this fact, to establish a model, which determines the size of the shear
stress which occurs when the rod is subjected to a
normal stress perpendicular to the rod. A sufficient
shear resistance in the CFRP rod can therefore not be
obtained in the same way as in traditional pre-stressing anchor systems using steel tendons. It is foreseen
that development of a reliable standard anchor for
CFRP rods shall to a large extent be based on experiments and tests. The tests shall comprise both the
instant pre-stressing state as well as the long-term
behaviourdecrease of prestressing force with time.
Much research work has been, and still is, carried out
in this field. However, a better understanding of the
given CFRP anchorage systems is needed in order to
develop newer systems which are commercially competitive in the civil engineering industry.
The requirements for FRP anchorage system has
been discussed by researchers Reda Taha (1997 and
2003) and Sayed-Ahmed (2002) and different standards

1181

ASTM (2006), ACI (2004) and associations such as

IABSE-AIPC-IVBH(2003), and FIB (1993), However, there is still a great demand for commercial FRP
anchorage device systems, which can compete on the
same level as the ones used for post-tensioned steel. An
important factor is the amount of force each anchor has
to transfer. The force in each anchor is less, in bonded
applications, as pre- or post-tensioning with continuous posterior bonding to the concrete through epoxy or
grout. Once a jack stress of 5065% is reached, anchors
in the mentioned cases, are relieved and the remaining
percents are left for the continuous bond to handle as
the structure is exposed to live loads, ACI (2004). This
limit stress is about 25% lower than the jacking stresses
allowed in high strength steel due to larger standard
deviations for ultimate strength and the lack of yield
behaviour. The anchors have to be able to handle 100%
of ultimate tensile strength of the tendon if no bonding
is appliedthus giving rupture of the tendon outside
the anchorage zone. Altogether, challenges are the largest for un-bonded post-tensioned field applications,
and perhaps more precisely for external post-tensioning
applications where the aesthetic factor is more important. Focus should, with that in mind, be at developing a
small, simple, strong, durable and aesthetic anchor suitable for industrialised prestressing of FRPs.

AVAILABLE ANCHOR DEVICES

There are mainly two techniques which are used for

anchorage of the CFRP tendon: Bonding and friction.
Bonded anchors keep the tendon in place through
plain adhesion between the adherent and adhesives,
while the friction anchor is a mechanical anchor,
which uses compression force and, through this, friction to grab the tendon.

2.1

Figure 1. Bonded anchors: Example of contoured sleeveand straight sleeve bonded anchor.

Bonded anchors

Epoxy or similar bonding agents are widely used

in FRP applications because of their versatility and
strength. It is also a natural part in many pre-stressing applications involving FRP tendons. In literature
the most commonly used anchor for unbonded posttensioned tendons is the bonded sleeve anchorage.
This type of anchor is composed of an outer sleeve
filled with the bonding agent and the inserted tendon
(Figure 1). Also other bonded anchor methods exist,
such as the double-lap anchor presented by Maravegias and Triantafillou (1996) and the gradient method
presented by Czaderski and Motavalli (2007). However, the latest of these is a new research area and
effort is still put into the development of this technique. Literature concerning bonded anchor systems
for FRP tendons is still limited. The reason for this

Figure 2. Mechanical anchor: Example of a clamping

anchor with an inside sleeve.

could be the consequence of results presented by

research groups during the last decade. Nanni et al.
(1996) tested the commercial anchor systems available at that time for ultimate and short term sustained
loading performance. Their results were overall positive and the ultimate strength stated by the FRP manufacturer was reached in all tests irrespective of whether
epoxy or cementitious grout was used as adhesive. A
165 mm sleeve was used for the epoxy samples and a
500 mm sleeve for the grouted sleeve. Similar results
are also reported by other researchers who have
reached enough capacity in the bonded anchors, but
the bonding zone is, however rather large. Lees et al.
(1995), for instance, used 150 mm anchors to anchor
a 15 kN FRP force, which results in a poor strength
to length ratio compared to available anchors for steel
tendons. However, the size of the bonded anchor types
makes them fit for ground anchorage purposes, but
unfit for the implementation in external post-stressed
applications for concrete structures. The method has
therefore been abandoned by the concrete industry but fully adopted by the geologists, Zhang et al.
(2001). Stronger bond and almost instantaneous

1182

hardening has to be achieved, if bonded anchors are to

be appreciated for post-tensioning applications.
2.2

Mechanical anchors

The split wedge anchor system is the most widely

used system in civil engineering application, and it
is the most commonly used anchor type in steel prestressing. However, the use of wedge anchors implies
compression of the FRP tendon, which causes problems due to the anisotropic properties. The strength
properties in the transverse direction of the fibres are
very poor, and mechanical friction anchors use large
compression forces to anchorage the tendon. This
introduces very high principal stresses in front of the
anchor device, where both tension and compression
forces are represented. FRP tendon anchorages are
therefore, at present time, only reliable if they have
a length exceeding 70 mm and a considerable presetting force is applied.
It would be possible to use them in larger systems
and compete with given steel pre- or post-stressing
systems, if this length can be decreased with about
1/3, due to the fact that steel wedges require an anchorage length of 2550 mm. An example of a wedge
anchor system is shown in Figure 3. The number of
wedges and the material properties might be varied,
in order to grab carefully around the FRP tendon.
Enka (1986) used plastic wedges for pre-tensioning
mechanical gripping anchors and reported these to
work satisfactory. Kakihara et al. (1991) used a twopiece split-wedge anchor for pre-stressed concrete
using AFRP, which gave an efficiency of 97%. The
split-wedge anchor has the advantage of being easy to
assemble and use on construction site and is therefore
favoured in pre-stressing applications. Shear failure
of the tendon due to excessive shear stresses in the
anchor zone has been evaluated by Sayed-Ahmed et
al. (1998), who also presented research of FRP tendon anchorage, which dealt with the optimizing of
wedge anchor dimension through change of angle in

Figure 4. Geometrical configuration of anchor with longitudinal curved wedge, Al-Mayah et al. (2007).

Figure 5. FEM analysis. Pre-setting stresses of anchor with

longitudinal curved wedges (Al-Mayah et al. 2007).

the surface between wedge and barrel and the implementation of a copper sleeve between wedge and
CFRP rod Sayed-Ahmed et al. (1998). The reason
for the change in angle is to move the high compressive forces into the anchor-back end and through this
prevent high principal stresses in front of the anchor,
where both tension- and compressive forces are represented. The copper sleeve prevents crushing of the
fibres due to direct contact between the wedges and
CFRP rod. Al-Mayah (2007) and his team worked in
general extensive on the development of the wedge
anchor and gained great progress in developing a
wedge anchor system. Al-Mayah et al. (20012007),
Figure 4 and Figure 5. Al-Mayah presented in 2008 a
paper Al-Mayah et al. (2008) regarding the study of
interfacial mechanics of CFRPmetal couples under
different contact pressures using a clamping anchor
(a clamping anchor example is shown in Figure 2).
Aluminium- and copper sleeves were used in direct
contact with the rod. The tests showed that the shear
stresses increased if smooth machined CFRP rods
were used and also when soft sleeves were used
together with a rod of high ultimate capacity.

Figure 3.

Example of 3 wedge anchor with inside sleeve.

MATHEMATICAL MODELING
OF ANCHORS

The first model of analyzing an anchor using FEM

modelling was made by Mitchell et al. (1974). It was
shown that a bond release between the resin plug
and the metallic socket reduces the shear stresses
significantly. Noisternig et al. (1996) made 3D FEM

1183

model analysis, where they studied the mechanicaland bonded anchor systems for CFRPs. In the
research they optimized the compressive stress distribution in order to find an optimal soft transition
between the anchor and the rod. They increased the
angle of the outer conical tube in order to reduce the
critical stresses. These results are, however, not applicable for evaluation of wedge anchors. Sayed-Ahmed
et al. (1998) developed a system where the barrel
system anchor with 4 wedges had direct contact with
the FRP tendon. The barrel and wedges had an angle
difference of 0.1 degrees in order to increase the compression force in the back of the anchor and through
this decrease the principal stresses at the entrance of
the anchor. The anchor had no sleeve and was both
resin and steel free. Campbell et al. (2000) verified
the results of Sayed-Ahmed et al. (1998) through
orthotropic FEM analysis. However, Sayed-Ahmed
et al. (1998), Sayed-Ahmed (2002) and Campbell
et al. (2000) investigated the anchor system trough a
linear elastic analysis. They reported an extension of
the hoop stresses inside the barrel in the cases, where
the friction force between the barrel and the wedges
was low.
In 2001 AI-Mayah et al. (2001) made an axisymmetric linear quadratic FE model for the analysis
of anchors with a metallic sleeve between the tendon
and the wedge. The study showed that pre-setting
force had a significant influence on the prevention
of tendon slip compared to the friction between the
anchor components.
Al-Mayah et al. (2007) developed an anchor system which was longitudinally curved (instead of an
angle change) in order to control the evenness of the
compression force in the anchor. The phenomenon
is also seen in Figure 5 where the highest stresses
are obtained in the back when presetting the anchor.
Al-Mayah and his team worked on the development
of the wedge anchor where they tested according
to PTI (2006) recommendation and changed different parameters in the wedge anchorage system.
Al-Mayah made in 2005 a FEM study concerning
the barrel thickness effects on the anchorage system and also what effects sandblasting would have
on the interior of the copper and aluminium sleeve.
The highest interfacial shear stresses were recorded
when sandblasted soft annealed sleeves were used
Al-Mayah et al. (2006).
4

have already been obtained. The project involves

measuring of shear stresses in the rod through optical fibres, development of a new metal anchor, testing
of different anchor lengths and verification through
FEM-analysis and analytical theory.
4.1

Test setup

The test setup is seen in Figure 6 where the 8 mm rod

tendon is anchored by a clamp anchor (Figure 7) in
one end (dead end) and a wedge anchor in the other
end. The presetting procedure is done though a special designed duct in which the CFRP tendon can be

Figure 6 Test setup. Pre-setting procedure (left) and test

setup (right).

PRESENT RESEARCH PROJECT

The research project is a co-operation between Technical University of Denmark (DTU), Lule University
of Technology (LTU) and COWI A/S and focuses on
the development of a new anchor type for FRP systems. The project is still ongoing but promising results

Figure 7.

1184

Clamping anchor (dead end anchor).

placed. The wedges in the anchor are placed against a

support and the whole system is compressed with the
decided pre-setting force. The test setup consists of
the two mechanical anchors, the rod and the tension
system with applied measurement equipment.
The optical fibres are mounted in a groove at the
rod surface and covered with epoxy for protection.
The clamping anchor in the dead end has shown
very good reliability in the performed tests, which is
important for the test procedure of the new invented
anchor.
Pilot tests have been done on the clamping anchor
in order to verify the reliability of the dead end before
testing the wedge anchor. A clamping anchor was
placed in both ends of the rods and tested in the test
setup. 100% of the tests showed a fracture between
the anchors which occurred at 130kN. 6 preliminary
tests were made with the new developed wedge
anchor, which showed a fracture at 110112 kN at a
presetting force of 100 kN. These rods were however
taken from another delivery batch, which could mean
a change of capacity in the 8 mm rod. The wedges
in the wedge anchor consist of aluminium with provided yield strength of 150MPa, and the barrel is
produced in steel with provided yield strength of
235 MPa. Further tests have to be done on the actual
aluminium- and steel material for verification and
implementation in the future FEM-model analysis
and analytical approaches. The CFRP HR 2500 rod
type HS from STO which is used for future tests is
approved according to EN 2561 and the properties are
shown in Table 1.
The measurements at present time have been made
to ensure fracture between the anchors and reliability
of the clamping- and wedge anchor. The measurement
devices are seen in Figure 8, and the obtained results
will be clarified by the use of optical fibres.
The barrel strain gauge measurements should
result in values which we can use for the prediction
of the hoop stresses and contact pressure which leads
in the direction of predicting the friction between the
barrel and wedges through analytical evaluations.
The optical fibres are mounted in order to show the
shear stresses in the rod. It is the goal, through these
measurements, to verify the shear stresses at the rod
surface. The deformation measurer (LVDT) measures
Table 1.

Material properties of the HR 2500 CFRP rod.

Property

HR 2500 CFRP Rod

Mean tensile E-modulus [GPa]

Min tensile E-modulus [GPa]
Mean tensile strength [MPa]
Min tensile strength [MPa]
Min elongation at break [%]

165
158
2 500
2 200
1.3

Figure 8. Placing of measuring devices.

the mutual movement of the individual anchor pieces

and the rod. An external measurement point will be
measuring the movement of the supports.
5

CONCLUSION

No sufficient mechanical anchor system has been

made yet, despite worldwide efforts to find a competitive FRP tendon anchorage. Several wedge shapes
and different materials have been used in order to
solve the problem of too high principal stresses at
the anchor front. It is seen that no FRP anchor is yet
implementable in external post-tensioning applications, which means that further improvements are
necessary in order to make FRP pre-stressing systems commercially compete able. No such FRP systems are available on the market yet, which results
in different evaluations of each pre-stressing system.
This makes the systems very expensive and the reliability could be questioned, due to the fact that no
system experience has been gained. It is very important to verify the theoretical evaluations through test
results. However, it can be hard to measure friction
values between the rod, wedges and barrel. It is therefore problematic to obtain proper correspondence in

1185

the analytical methods. Optical fibres could be one

of the tools in the search of new test methods. It will
be possible to measure shear stresses in the rod and
verify these through theory.
REFERENCES
ACI 2004. ACI 440.4r-04: Pre-stressing concrete structures
with FRP tendons, Technical report, ACI.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2001a. Experimental and analytical investigation of a stainless steel
anchorage for CFRP prestressing tendons, PCI Journal
46(2), 88100.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2001b. Mechanical behavior of CFRP rod anchors under tensile loading,
Journal of Composites for Construction 5(2), 128135.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2005a. Variable
thickness barrel anchor for CFRP prestressing rods. Ned
Burns symposium on historic innovations in prestressed
concrete, Journal of Composites for Construction ACI
SP-231, 237252.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2005b. Effect
of sandblasting on interfacial contact behavior of carbon-fiber-reinforced polymer-metal couples, Journal of
Composites for Construction pp. 289295.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2006a. Development and assessment of a new CFRP rod-anchor system
for prestressed concrete, Applied Composite Materials
13(5), 32134.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2006b. Fem and
mathematical models of the interfacial contact behaviour
of CFRP-metal couples, Composite Structures 73 pp.
3340.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2006c. Effect of
sleeve material on interfacial contact behavior of CFRPmetal couples, Journal of Materials in Civil Engineering
Nov/Dec 2006, 825830.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2007. Novel
anchor system for CFRP rod: Finite-element and mathematical models, Journal of Composites for Construction 11(5), 469.
Al-Mayah, A., Soudki, K. and Plumtree, A. 2008. Effect of
rod profile and strength on the contact behavior of CFRPmetal couples, Composite Structures 82(1), 1927.
ASTM 2006. Astm -d7205/d 7205 m -06: Standard test
method for tensile properties of fiber reinforced polymer
matrix composite bars, Technical report, ASTM.
Enka, B. 1986. The twaron resin prestressing element for
concrete, Technical report, Arapree.
Erki, M. and Rizkalla, S. 1993. Anchorages for FRP,
Concrete International: Design and Construction 15(6),
5459.

FIP 1993. Recommendations for the acceptance of posttensioning systems, Technical report.
IABSE, AIPC, IVBH 2003. Use of fibre reinforced polymers
in bridge construction, Technical report.
Kakihara, R., Kamiyoshi, M.K.S. and Noritake, K. 1991. A
new aramid rod for the reinforcement of prestressed concrete structures, Advanced Composite Materials in Civil
Engineering Structures, Proceedings of the Specialty
Conference pp. 134142.
Lees, J.M., Gruffydd-Jones, B. and Burgoyne, C.J., 1995,
Expansive Cement Couplers: A Means of Pretensioning Fiber-Reinforced Plastic Tendons, Construction and
Building Materials, V. 9, No.6, pp. 413423.
Maravegias, S. and Triantafillou, T. 1996. Numerical study
of anchors for composite prestressing straps, Composite
Structures 35(3), 323330.
Mitchell, R.A., Woolley, R.M. and Halsey, N. 1974. Highstrength end fittings for FRP rod and rope, Journal of
the Engineering Mechanics Division, ASCE V. 100,
No. EM4, 687703.
Nanni, A., Bakis, C.E., ONeil, E.F. and Dixon, T.O. 1996.
Performance of FRP tendon-anchor systems for prestressed concrete structure s, PCI Journal 41(1), 34.
Noisternig, J.E. and Jungwirth, D. 1996. Design and analysis of anchoring systems for a carbon fiber composite
cable, Advanced Composite Materials in Bridges and
Structures pp. 935942.
PTI 2006, 1997, 1985. Post-tensioning manual,
Post-Tensioning Institute 6:th, 5:th and 4:th edition.
Reda Taha, M.M., Sayed-Ahmed, E.Y. and Shrive, N.G.
1997. Towards a new non-metallic anchorage system for
post-tensioned applications with carbon fibre reinforced
plastic tendons, International SAMPE Symposium and
Exhibition (Proceedings) 42(1), 288297.
Reda Taha, M.M. and Shrive, N.G. 2003. New concrete
anchors for carbon fiber-reinforced polymer posttensioning tendonspart 1: State-of-the-art review/
design, ACI Structural Journal -American Concrete
Institute 100(1), 86.
Sayed-Ahmed, E.Y. 2002. Single-and multi-strand steel
anchorage systems for CFRP tendons/stays, Proceedings, Annual Conference -Canadian Society for Civil
Engineering 2002, 13811390.
Sayed-Ahmed, E.Y. and Shrive, N.G. 1998. New steel
anchorage system for post-tensioning applications using
carbon fibre reinforced plastic tendons, Canadian Journal of Civil Engineering 25(1), 113127.
Schmidt, R.J., Dolan, C.W. and Holte, L.E. 1994. Anchorage of non-metallic prestressing tendons, Structures
Congress XII pp. 14151420.
Zhang, B., Benmokrane, B., Chennouf, A. & Mukhopadhyaya, P. 2001.Tensile behaviour of FRP tendons for
prestressed ground anchors, Journal of Composites for
Construction, 8593.

1186