Construction and Building Materials 112 (2016) 28–38

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Iron-based shape memory alloys for prestressed near-surface mounted

strengthening of reinforced concrete beams
Moslem Shahverdi ⇑, Christoph Czaderski, Masoud Motavalli
Empa, Swiss Federal Laboratories for Materials Science and Technology, Überlandstrasse 129, 8600 Dübendorf, Switzerland

h i g h l i g h t s

Manufacturing of long ribbed iron-based shape memory alloy (Fe-SMA) strips.

Activation of long ribbed Fe-SMA strips by resistive heating.

Demonstration of application of near surface mounted Fe-SMA strips for strengthening.

Prestressing of RC beams with long ribbed Fe-SMA strips.

Experimental examination of six large scale RC beams.

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

Article history: Low-cost Fe-Mn-Si-based shape memory alloys (SMAs) have attracted much attention in the research
Received 13 June 2015 community and in practice over the past two decades as a cost-effective alternative to the expensive
Received in revised form 5 February 2016 Ni-Ti-based SMAs. The shape memory effect refers to the phenomenon in which SMAs, if they are
Accepted 23 February 2016
deformed, return to their former shape upon heating. Near-surface mounted (NSM) strengthening tech-
Available online 1 March 2016
niques can be used to strengthen concrete beams. The advantages associated with NSM strengthening are
its ability to significantly reduce the probability of harm that results from corrosion, fire, acts of vandal-
Keywords:
ism, mechanical damage, and aging effects. Iron-based SMA (Fe-SMA) strips can be used as NSM rein-
Iron-based shape memory alloy
Activation
forcements. The NSM Fe-SMAs can more easily be prestressed than NSM fiber reinforced polymers,
Near-surface mounted because the prestressing of SMAs does not require any mechanical jacks and anchorheads.
Strengthening In this study, an experimental program was established to investigate the flexural behavior of rein-
Prestressing forced concrete (RC) beams that were strengthened and prestressed with Fe-SMA strips. The specific
focus was on the demonstration of the feasibility of this strengthening technique. A total of six RC beams
were experimentally examined under deflection control in a four-point bending loading rig. The experi-
ments consisted of one beam strengthened by Fe-SMA strips but not prestressed, three beams strength-
ened by prestressed Fe-SMA strips, and one beam strengthened by one CFRP strip. Additionally, one beam
with no strengthening, which served as the reference beam, was examined. The results showed that the
cracking loads and mid-span deflections of prestressed beams compared to other beams were, respec-
tively, higher and lower. Significant strengthening effects were achieved in NSM Fe-SMA-strengthened
beams compared to the reference beam. These direct effects showed that the application of near-
surface mounted Fe-SMA strips worked well as prestressing elements in concrete beams.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction on a concrete surface [1]. In addition to the application on the sur-

Many existing concrete bridges must be strengthened due to
aging or adapting to increase their load capacity. A popular
strengthening technique currently is the application of fiber rein-
forced polymer (FRP) strips or fabrics by means of epoxy adhesives
⇑ Corresponding author.
E-mail address: moslem.shahverdi@empa.ch (M. Shahverdi).
face, another technique exists, where the FRPs are inserted and
glued into grooves in the concrete cover, which is called the
near-surface mounted (NSM) strengthening technique [2]. This
strengthening technique requires cutting grooves in the concrete
cover, and no surface preparation work is needed afterwards. The
advantages that are associated with NSM strengthening compared
to the externally bonded technique are its ability to significantly
reduce the probability of harm that results from corrosion, fire, acts

http://dx.doi.org/10.1016/j.conbuildmat.2016.02.174
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38
of vandalism, mechanical damage, and aging effects. Furthermore,
better bonding behavior due to a confining effect in the grooves
can be expected.
In some of the cases, strengthening FRPs are prestressed. The
advantage of prestressing is that existing deformations and crack
widths can be reduced, and furthermore, the cracking and yielding
loads are higher, and the FRP material is better utilized [3]. For the
prestressing of externally bonded FRPs, different systems can be
purchased from the market; however, prestressing systems for
NSM FRPs are not available yet. The problem with prestressing
NSM FRP strips/bars is that gripping the FRPs in the groove and
prestressing them is very difficult. An overview of existing NSM
prestressing systems used in laboratories is given in [4]. The appli-
cation of iron-based shape memory alloy (Fe-SMA) strips instead of
FRP strips will overcome this problem. Preliminary experiments to
investigate the feasibility of the application of prestressed SMA
strips for the strengthening of concrete structures have been pre-
sented by Czaderski et al. [5].
The available SMAs on the market are Nickel-Titanium (NiTi,
Nitinol). However, this material is far too expensive for the con-
struction industry [6]. Low-cost Fe-SMAs can be an interesting
alternative. The Fe-SMA material was discovered by Sato et al.
[7] in 1982. An extensive explanation of how this material works
and some key properties of these materials, such as the recovery
stresses and the corrosion resistance, have been presented by Cla-
dera et al. [8]. A novel iron-based shape memory alloy Fe–17Mn–
5Si–10Cr–4Ni–1(V,C) (ma.-%) has been developed at Empa in
Switzerland [9–13].
Shape memory alloys (SMAs) have several unique properties.
The two most important properties are the shape memory effect
(SME) and superelasticity [6]. The SME refers to the phenomenon
that if SMAs are deformed, they return back to their former shape
upon heating. Superelasticity refers to the phenomenon that SMA
can undergo a large amount of inelastic deformation and recover
its shape after unloading automatically without heating. Due to
the presence of the superelasticity property, SMAs are used in civil
applications as a passive vibration damping and energy dissipation
material [6]. Most of the research work in the field of SMAs for con-
struction is focuses on these topics, for example [6,14–19]. Fur-
thermore, recently 16 new seismic dampers were installed in JP
Tower Nagoya in Japan using an Fe-SMA [20]. Moreover, in Japan,
the material is used for the prestressing of crane rail joints using
fishplates made of Fe-SMA [21,22] and for pipe joining [22].
The shape memory effect can be used to prestress an SMA strip/
bar. For example, in principle, if a straight SMA bar is deformed by
pulling and then heated above a certain temperature, it will return
to its original length. However, if the deformation recovery is
restrained (e.g., by embedding the SMA in concrete), a mechanical
stress occurs in the material when it is heated and cooled after-
wards. This stress is called ‘‘recovery stress”, and it can be used
for introducing prestressing forces in concrete structures to
improve its serviceability. One of the advantages of such a pre-
stressing technique compared to conventional prestressing is that
there are no frictional losses due to the development of uniform
tension force along the total length of the embedded SMA tendon.
Therefore, such a prestressing technique would be even suitable
for curved concrete members or if the tendon profile is strongly
curved. Furthermore, in comparison to the conventional prestress-
ing techniques, SMA tendons can be used for prestressing extre-
mely thin concrete members without any need for anchorheads,
oil hydraulic cylinders, ducts and grout injection.
Several investigations on the prestressing of small concrete or
mortar prisms by using SMAs can be found in the literature, for
example [6,23–30]. Additionally, SMAs are used to actively confine
concrete cylinders [31,32], large-scale concrete columns [33] and
non-circular concrete elements [34]. Tran et al. [32] performed
29
some crush tests of concrete cylinders confined by nickel–titanium
SMA wires. They studied active confinement by SMA wires previ-
ously prestrained in martensitic state and then subjected to the
memory effect by heating. For the comparison, they also studied
passive confinement by using the same SMA, but in austenite state.
The results obtained in their study show that stiffness, strength,
and ductility of the concrete were improved in both confinement
systems. However, all of these investigations are still ongoing,
and more research work is necessary.
The current study is a continuation of the study described in [5].
The same Fe-SMA strips described in [5] were used as prestressed
NSM for the strengthening of RC beams. Six reinforced concrete
beams with spans of 2 m were examined in the framework of the
current study. The SME of the Fe-SMA strips was used to
strengthen the RC beams. The obtained results showed the feasibil-
ity of the application of prestressed Fe-SMA strips for the flexural
strengthening of concrete beams.
2. Experiments
Experimental work was conducted at Empa, the Swiss Federal Laboratories for
Materials Science and Technology, on six two-meter span beams. The experimental
program was designed to demonstrate the application of near-surface mounted
prestressed Fe-SMA strips for the flexural strengthening of concrete beams. Beams
were constructed with the dimensions shown in Fig. 1. All of the beams were loaded
in a four-point bending test scheme with a span of 2.0 m. The experiments included
a reference beam with no strengthening, three beams strengthened with two acti-
vated Fe-SMA strips, one beam strengthened with two non-activated Fe-SMA strips,
and one beam strengthened with one CFRP strip for comparison. To have similar
strengthening effects for all of the strengthened beams, only one CFRP strip was
used as a strengthening element. An overview of the test program appears in
Table 1. The material properties of the Fe-SMA strips, concrete, steel reinforcement,
grout, CFRP strips, and adhesive used in the current study, the measurement set-up
and the experimental procedure are presented in this section.
2.1. Materials
2.1.1. Iron-based shape memory alloys (Fe-SMA)
The novel iron-based shape memory alloy Fe–17Mn–5Si–10Cr–4Ni–1(V,C)
(ma.-%) developed at Empa in Switzerland [9–12] was used in this study. A large
quantity of the new alloy was produced and strips were manufactured. The detailed
production procedure of the Fe-SMA strips is described in [5]. Tensile failure exper-
iments, prestraining and heating/cooling experiments on the Fe-SMA strips are also
presented in [5].
The Fe-SMAs used in this study were initially in the form of long ribbed strips.
They were delivered to Empa in two batches. To ensure a good bond between the
Fe-SMA strips and the concrete, ribs (Fig. 2) were applied on the strips by cold form-
ing. The ribs were applied at an angle of approximately 40° on one side and 130° on
the other side of the strip to ensure a regular strain pattern along the strip [5]. The
presented lap-shear tests on the Fe-SMA strips proved the feasible bond behavior of
the ribbed Fe-SMA strips [5].
The nominal thickness, initial width and initial length of the Fe-SMA strips were
1.7 mm, approximately 25 mm and more than 3 m, respectively. However, these Fe-
SMA strips were cut into strips with a length of 2.6 m and a width of 20 mm (Fig. 2).
The strips were ground on the edges to remove the edge cracks. The remaining
short pieces of the strips were used for material characterization.
2.1.1.1. Characterization tests of the Fe-SMA strips. The short Fe-SMA specimens were
prestrained in a 20 kN Zwick tensile testing machine. Then, they were heated by
means of a climate chamber that was added to the Zwick testing machine to deter-
mine the recovery stress (prestress). A special clip-on extensometer with a gauge
length of 100 mm was used for the heating experiments. The temperature expan-
sion of the extensometer was continuously compensated during the experiments
by respective control in the testing machine.
2.1.2. Concrete, steel reinforcement, grout, and CFRP strips
A concrete mix of Type I Portland cement (350 kg/m3) and a coarse aggregate
with a maximum size of 16 mm and a water cement ratio of 0.50 by weight was
used to cast the beams. Additional concrete samples of 150  150  150 mm were
casted for each beam and were tested at the age of 28 days and on the day of per-
forming the experiments. The average compressive strength, splitting tensile
strength, and elastic modulus of the concrete after 28 days were, respectively,
53.4 MPa, 3.4 MPa, and 35.4 GPa. The used internal reinforcement of the examined
beams is presented in Fig. 1. The diameter of the four flexural reinforcement bars
was 8 mm (Ø8). The internal stirrups also had a diameter of 8 mm and a spacing
30 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38

Fig. 1. Drawing of Beams No. 3, 4, and 6, T1 to T7 are thermocouples, all dimensions are in mm.

Table 1
Overview of the experimentally examined beams. of 150 mm. The Elastic modulus and yielding strength of the internal flexural rein-
forcement were 200 GPa and 508 MPa, respectively. The CFRP strip (S&P Laminates
Beam No. Strengthening with
CFK 150/2000) was received from S&P Clever Reinforcement in Seewen, Switzer-
1 Reference beam, no strengthening land. The measured elastic modulus of the CFRP strip was approximately
2 Strengthened by two SMA strips but not activated 150 GPa. The CFRP strip was glued into the groove with an epoxy adhesive. The
3 Strengthened by two SMA strips and activated Fe-SMA strips were glued into the grooves with a cement-based mortar. Because
4 Strengthened by two SMA strips and activated the Fe-SMA strips should be activated by heating up to approximately 160 °C in a
5 Strengthened by one CFRP strip short time, it would create problems in an epoxy adhesive, having a glass transition
6 Strengthened by two SMA strips and activated temperature that was well below 160 °C. The cement-based mortar was a flowable
and expanding grout (SikaGrout-311) from Sika in Switzerland, and it was pur-
chased from the market. According to the technical datasheet from the company,

Fig. 2. Left: a photo of the long Fe-SMA strips, which were embedded in Beams 3 and 4 (photographed just before embedding), right: a zoomed photo of Fe-SMA strip ends
where sealing foam and ribs are shown.
 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38 31

the maximum grain diameter size, compression strength after 28 days and elastic
modulus of this mortar are 1.2 mm, 80–90 MPa and approximately 37.2 GPa,
respectively.

2.2. Experimental procedure

The beam experimental procedure was based on the following main steps:

Prestraining of the strips

Grouting of the Fe-SMA into the grooves

Activation, i.e., the prestressing of Fe-SMA strips embedded in the concrete
beams

Loading of beams up to failure

2.2.1. Prestraining
The long Fe-SMA strips test specimens, which were embedded in the grooves of
the RC beams in a later stage (length of 2.6 m), were prestrained to a 2% elongation
at room temperature by a manually operated oil hydraulic jack and then relaxed to
Fig. 4. The left bottom side of Beam No. 4 which is equipped for activation, i.e. the
a stress-free state with the recovery of the elastic strain. The Fe-SMA strips were
resistive heating of the Fe-SMA strips. The photo shows the contacting of the power
clamped at the ends with steel clamps and then fixed in a vertical test set-up on
supply by copper cables and copper clamps on the Fe-SMA strips.
the strong floor.

2.2.2. Grouting
The grooves with dimensions 25  6 mm at the bottom of the beams were
made on a saw cut table machine. The SMAs were inserted into the grooves and
the grooves were filled by a cement-based grout that was prepared according to
the supplier’s recommendations (Fig. 3). Finally, the excess grout was removed,
and the surface was leveled.
Before embedding the Fe-SMA strips in the concrete, each Fe-SMA strip was
equipped with three Type K thermocouples, Figs. 1 and 2. The thermocouples
mounted on each strip measured the temperature at the surface of the strips inside
the concrete beams. The numbering of the thermocouples can be found in Fig. 1.

2.2.3. Activation
The activations were performed after the grout for the Fe-SMA strips was fully
cured. The activations were performed by resistive heating. A programmable elec-
trical power supply was used to provide the resistive heating. This power supply
was controlled by a LabView program. The program acquired the signals from the
thermocouples and controlled the current supply for resistive heating. A relatively
high current density of approximately 9 A/mm2 was applied to minimize the heat-
ing times and reduce the heat flow into the concrete. When the temperature that
was measured by selected thermocouples T1 to T3 or T4 to T6, which measured Fig. 5. The right side of the beam with the copper cables which connects the two
the strip temperatures in the concrete, reached the target temperature of 160 °C, Fe-SMA strips in order to close the current circuit.
the power supply was switched off. Copper clamps (Figs. 4 and 5) were used to
secure the contacts between the cables from the power supply and the Fe-SMA
strips. The Fe-SMA strips in Beam No. 3 were activated one by one. The two Fe-
SMA strips in Beams No. 4 and 6 were connected by a copper cable (Fig. 5) and could
therefore be heated together (i.e. in series). During the activation and the subse-
quent cooling phase, the mid-span displacements were measured by using two
LVDT’s. The strain on the top side and bottom side of the beam were also measured
by using a strain gauges and a 1 m long LVDT on the bottom side of the concrete
beams and the temperatures were measured by means of thermocouples (Fig. 1).
In addition, an infrared camera was placed under the activated beams to measure
the surface temperature of the bottom side of the concrete beam during the activa-
tion and cooling to room temperature.

Fig. 6. Photo of loading setup.

Fig. 3. Embedding the Fe-SMA strips into grooves in the concrete cover with
cement based grout.
2.2.4. Loading
The specimens were simply supported at the ends and were loaded using two
concentrated loads with 400-mm spacing (Fig. 6). The load was applied simultane-
ously using two 50-kN hydraulic jacks. The tests were performed under cylinder
displacement control of the left cylinder with a loading rate of 0.02 mm/s. Fig. 1
shows the set-up for the loading phase, and Fig. 6 presents a photo of the loading
set-up. Loading was applied in a quasi-static manner until a deflection of 1/500
of the span length (i.e., 4 mm) was reached at the mid-span (the serviceability limit
state). Afterwards, except for Beam No. 6, the load was released to a mid-span
deflection of 2 mm, and then, the load was applied again. The loading–unloading
was repeated five times for each beam. After the fifth loading cycle, the load was
32 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38

applied monotonically until the load capacity was reached. A linear voltage dis-
placement transducer (LVDT) was placed at the load point locations during the
experiments, to control the speed of the hydraulic jack. Deflections were measured
at the beam centerline using LVDTs, as shown in Fig. 1. Full-field displacements
were measured at the front side of each beam on a surface of approximately
400 mm by 150 mm in the middle of the beam with a 3D digital image-
correlation (DIC) measurement system, see Fig. 1. An irregular pattern of black
speckles on a white painted background was applied on the beam.

3. Results and discussions

3.1. Prestraining

Stress–strain diagrams of representative experiments for the

long and short Fe-SMA test specimens are given in Fig. 7. The
unloading path after prestraining reached a strain of approxi-
mately 1.5%. Similar stress–strain curves of the two different
experiments can be observed. The Fe-SMA strips are in the austen- Fig. 8. Recovery stress of short Fe-SMA strips tested in the 20 kN Zwick tensile
ite phase before the prestraining. The nonlinear deformation testing machine with a climate chamber.
behavior during loading is due to a phase transformation from
austenite to martensite, and in the higher strain regions, there
Table 2
are possibly also plastic deformations. During unloading, the strain
Measured recovery stresses in short Fe-SMA strips examined in a 20 kN Zwick tensile
also deviates from the linear-elastic behavior. The deviation from testing machine with climate chamber.
the linear-elastic unloading is called pseudoelastic strain [11],
Specimen Used in beam rini (MPa) rend (MPa) Dr (MPa)
which means a phase transformation from martensite back to
austenite. R1 Beam No. 3 41.7 243.0 201.3
R2 Beam No. 3 41.7 230.0 188.3
R5 Beam No. 4 42.2 220.0 177.8
3.2. Recovery stresses R6 Beam No. 2 40.8 241.0 200.2
R11 Beam No. 6 45.8 246.6 200.8
Determination of the recovery stresses on the short Fe-SMA R12 Beam No. 6 51.2 245.7 194.5

specimens resulted in the stress–temperature diagrams depicted

in Fig. 8. This figure presents the heating–cooling route to deter-
mine the recovery stresses. While the testing machine held the
strain constant and the temperature was increased to 160 °C at a
rate of 2 °C min1, there were distinctive decreases in the stresses
until a temperature of approximately 40 °C was reached. To avoid
compression in the test specimens during heating due to the ther-
mal expansion effect, each specimen was loaded to a preload value
(rini) that corresponded to approximately 30–40 MPa prior to the
heating–cooling route (Table 2). The decrease is attributed to the
thermal expansion effect on the examined specimens during the
increase in the temperature. At temperatures that are higher than
40 °C, the stresses increased. These stress increases are attributed
to phase transformations (from martensite to austenite), which
causes the shape memory effect. The stresses reached approxi-
mately 150 MPa at the maximum heating temperature of 160 °C.
After a waiting time of approximately one hour at 160 °C, the tem-
perature was decreased again at a rate of approximately 2 °C min1
and went down to 23 °C. By cooling down the Fe-SMA strips, the
stresses increase further due to the thermal contraction effect dur-
ing cooling. The inclinations of the curves are approximately paral-
lel to the starting part of the curves, which indicates similar
coefficients of thermal expansion and elastic modulus for the two
areas. At the end of the curves, the inclinations of the curves
decrease. This decrease is attributed to a partial transformation
from austenite to martensite. Finally, the obtained recovery stres-
ses at RT were reached. The applied tensile stresses prior to the
heating–cooling route (rini) and the tensile stresses at the end of
the waiting time (rend) are presented in Table 2 for all of the exam-
ined specimens. The difference between rend and rini, which is indi-
cated by Dr, is also presented in this table.
In this study, the tensile stresses in the Fe-SMA strips obtained
at the end of the heating–cooling routes, rend, were taken as the
recovery stress. The recovery stresses varied between 220 and
250 MPa.

3.3. Activation of the Fe-SMA strips embedded in beams

As described above, resistive heating was used for the activation

Fig. 7. Representative prestraining of the short and long strips. The short Fe-SMA
test specimens were prestrained in 20 kN Zwick tensile testing machine and the
long Fe-SMA test specimen were prestrained by a manually operated oil hydraulic
jack.
of the embedded Fe-SMA strips. Recorded temperatures for Beam
No. 3 over time are illustrated in Fig. 9. In the same figure, the dis-
placement at mid-span along the time axis is given. Two heating
steps of approximately 15–16 s up to around 160 °C were applied.
It is visible that the beam deformed at the beginning downward
due to temperature expansion, which is similar to in the character-
ization experiments discussed above. However, after a very short
time, the activation of the Fe-SMA strips worked and their pre-
stressing reacted in the other direction; the beam deformed
upward and compensated the thermal expansion. Fig. 10 displays
the displacement at mid-span over a longer time period as in
Fig. 9. It illustrates that a mid-span displacement of approximately
 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38 33

Fig. 9. Activation of Beam No. 3: the first 30 min of the displacement at mid-span Fig. 11. Activation of Beam No. 4: the first 10 min of the displacement at mid-span
and the corresponding temperatures in the Fe-SMA strips. diagram and the corresponding temperatures in the Fe-SMA strips.

Fig. 10. Long term displacement behavior of Beam No. 3 after the activation. With Fig. 12. Long term displacement behavior of Beam No. 4 after the activation. With
cooling the prestress in the Fe-SMA develops and the beam lifts up. More uplift cooling the prestress in the Fe-SMA develops and the beam lifts up. More uplift
develops because of concrete creep. develops because of concrete creep.

0.15 mm was reached after some hours when the beam was fully
cooled down. Then, the concrete started to creep, and an additional
mid-span displacement of approximately 0.02 mm occurred after
3 days.
Figs. 11–14 represent the same curves as explained above for
Beams No. 4 and 6. Generally, the behavior was similar to Beam
No. 3, although both of the Fe-SMA strips were activated in one
heating step of approximately 36–37 s. The mid-span uplift after
the complete cooling down of the beams was again approximately
0.15 mm. In the figures, the used currents and voltages are also
given. Current densities of 8.82 A/mm2 for Beam No. 3 and
7.65 A/mm2 were used for Beam No. 4 and 6. The voltage depends
on the length of the Fe-SMA strips and was therefore clearly higher
in the case in which both strips were heated in series (37–38 V)
compared to the case in which only one strip was heated (23 V).
The measured temperature on the strip surface basically should
have been the same, however, for Beam No. 6 there was relatively
high scatter. The reason for such scatter is probably due to varia-
tion in the fixation of the thermocouples on the SMA strips. The fix-
ation is quite challenging because the thermocouple cannot be in
direct contact with the SMA as there is a current running through Fig. 13. Activation of Beam No. 6: the first 10 min of the displacement at mid-span
the SMA during the activation (electrically isolation is needed), on diagram and the corresponding temperatures in the Fe-SMA strips.
34 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38
Fig. 14. Long term displacement behavior of Beam No. 6 after the activation. With
cooling the prestress in the Fe-SMA develops and the beam lifts up. More uplift
develops because of concrete creep.
the other hand it must be as near as possible to the SMA surface
because the temperature gradient in the mortar and concrete
beside the strip is strong.
Fig. 15 shows a photo that was taken by the infrared video cam-
era from the left-side bottom of the beam during the activation. It
is visible at the shown state, that the temperature in the embedded
Fe-SMA strip is significantly higher than the temperature of the
concrete.
Because it was not possible to directly measure the prestress
load in the Fe-SMA strips, the force in the strips had to be esti-
mated based on the strain measurements and/or the mid-span
deflection of the beams. For these estimations, losses due to con-
crete creep, concrete shrinkage, and Fe-SMA strip relaxations were
assumed to be small and negligible. Hence, the prestress force was
taken to be constant along the beam length between the supports.
Right after the cooling of the beam to room temperature and fully
obtaining recovery stress in the Fe-SMA strips, i.e., approximately
two hours after the resistive heating, the mid-span deflection of
the beam can be calculated by
ML2 ASMA rrec eL2
dmid ¼ ¼ of the load–deflection diagrams, and the diagrams followed the
8EI 8EI
where ASMA is the cross-section area of the embedded Fe-SMA strips,
L is the beam span, rrec is the recovery stress of the Fe-SMA strips, e
is the eccentricity of the strips from the natural axis of the
Fig. 15. Photos taken by infrared camera from left-side bottom of Beam No. 4: (a) normal photo and (b) thermal photo at the maximum captured temperature, 37 s after
resistive heating start.
uncracked concrete beams, and EI is the bending stiffness of the
uncracked concrete beams. EI was determined from the first loading
phase by fitting a line to the P-dmid diagrams. By rearranging the
above equation, and replacing the parameters for Beam No. 3 the
rrec for Beam No. 3 can be estimated by
8EI 8  2769037
rrec ¼ dmid ¼  0:15 ¼ 190 MPa
ASMA eL2 2  20  1:7  65  20002
The estimated rrec for Beam No. 4 and No. 6 are presented in
Table 3. Calculations showed that the recovery stresses obtained
in the beams were smaller than those obtained in the climate
chamber presented in Table 2. Thus far, the reason for this differ-
ence is not known yet, however, one hypothesis is that the activa-
tion process of the Fe-SMA strip in the concrete beams by resistive
heating is slightly different from the activation in a climate cham-
ber. Another hypothesis is that the recovery stress in the Fe-SMA
strip might be slightly reduced because of the shortening of the
concrete beams under loading. Additionally, the temperature dis-
tribution in the strip in the concrete is not uniform because it is
a climate chamber. Furthermore, the initially applied stress in
the climate chamber rini might have an influence (in the beams rini
in the Fe-SMA strips is zero). More investigations will be per-
formed in the group at Empa to verify the differences between
these two approaches for the activation.
3.4. Loading behavior of the beams
3.4.1. Load-mid-span behavior
The load-mid-span deflection diagrams of all of the examined
beams are depicted in Fig. 16 (loading up to failure and loading
up to the serviceability limit state). In all of the examined beams,
except for Beam 5, three changes in the slope are evident. The first
is due to concrete cracking; see Fig. 16(b). The second change in the
slope of the load-mid-span displacement occurred when the inter-
nal steel reinforcement began to yield and the third change
occurred when the Fe-SMAs acted in their low stiffness region.
The load–deflection diagrams were linear prior to the concrete
cracking, see Fig. 16(b). After the concrete cracking, the slope of
the load–deflection decreased clearly. The cycling loading at the
serviceability limit state did not have any clear effect on the slope
same slope as prior to the cycling loading. Comparing the load–de-
flection diagrams from Beam No. 3 and 4 with the one from Beam
No. 6, without the loading–unloading, we also verified that the cyc-
lic loading at the serviceability did not have any significant effect
 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38
Table 3
Mid-span deflection of beams due to prestressing Fe-SMA strips and back calculated recovery stresses.
Beam No. Reinforcement
3 2 Steel ribbed bars Ø8
+2 SMAs (activated)
4 2 Steel ribbed bars Ø8
+2 SMAs (activated)
6 2 Steel ribbed bars Ø8
+2 SMAs (activated)
Fig. 16. Load-mid-span displacement diagrams of the examined beams (a) up to
failure, and (b) at the serviceability limit state.
on the load-mid-span deflection diagram. However, the slopes of
the loading–unloading diagrams, i.e., the cyclic loading for the
strengthened beams were significantly higher than for the un-
strengthened reference beam, see Fig. 17(a), which shows the
effect of the additional reinforcement.
In the prestressed beams, smaller deflections were exhibited.
For example, at a load of 8 kN, see Fig. 17(b) and Table 4, the
mid-span deflection for Beam No. 2 was 5.6 mm while for Beam
No. 3,4, and 6 it was on average 4.2 mm, (4.6 + 3.9 + 4.1)/3 = 4.2.
At this load level, prestressing caused a decrease of 25% in the
mid-span deflection.
3.4.2. Cracking load and crack patterns
The loads at which the first concrete crack in the bottom side of
the beams along the constant moment region occurred, called
Pcrack, for all of the beams are presented in Table 4. The load at
the concrete cracking was estimated from the DIC measurement
system and verified by the slope changes in the load versus mid-
span diagrams. The Pcrack of the strengthened beams was much
35
Displacement shortly after Recovery stress shortly after
activation (mm) activation (MPa)
0.150 190
0.152 193
Average of (0.133 and 0.171)
0.168 213
Average of (0.161 and 0.175)
Fig. 17. Comparison of load mid-span displacement behaviors under serviceability
limit state of examined beams, (a) effect of cyclic loading of representative beams,
and (b) effect of activation.
higher than for the reference beam. Furthermore, by comparing
the Pcrack from Beam No. 2 with Beam No. 3, 4, and 6, the effect
of prestressing is determined. The Pcrack of the prestressed-Fe-
SMA strengthened beams on average, (4.5 + 4.7 + 4.2) kN = 4.47 kN,
was approximately 80% greater than for the non-prestressed-Fe-
SMA strengthened beams, i.e., (4.47/2.5 = 1.78).
During the testing of all of the beams, many cracks occurred. In
all of the beams, as a crack formed, it grew almost instantaneously
to a height of more than 100 mm above the bottom side of the
beam. The rapid crack growth was attributed to the low internal
reinforcements. The crack patterns for each beam type were differ-
ent. However, the cracks were fairly uniformly distributed. Fork-
like patterns were noted in the cracks in the constant moment
region, whereas the cracks that occurred in the shear span tended
to not fork but instead to propagate toward the load points, see
Fig. 18. The cracking patterns, as shown in Fig. 18 included vertical
flexural cracks around the mid-span and diagonal shear-flexure
cracks in the shear spans. Reference Beam No. 1 had the largest
crack spacing.
36 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38
Table 4
Failure modes, maximum load, mid-span deflection and concrete cracking load of beams.
Beam Failure mode Maximum load,
No. Pmax
(kN)
1 Concrete crushing after steel yielding 9.8
2 Concrete crushing after steel and SMA yielding 16.8
3 Concrete crushing after steel and SMA yielding 16.9
4 One SMA tensile failure after steel yielding, followed by 16.8
concrete crushing
5 Concrete crushing after CFRP tensile failure 22.9
6 One SMA tensile failure after steel yielding, followed by 16.4
concrete crushing
Fig. 18. Side view of all beams after failure test, observed cracking patterns for each beam.
3.4.3. Loading up to failure and failure modes
Failure of the reference beam, Beam No. 1, occurred due to
crushing of the concrete in compression after yielding of the longi-
tudinal steel reinforcement. High localized strains were generated
in the concrete at the top side in the mid-span location, and final
failure was due to the concrete crushing. The maximum compres-
sive strain measured on the top at mid-span was approximately
0.004. The failure of the strengthened beam with one CFRP strip,
Beam No. 5, was associated with a sudden tensile failure of the
CFRP strip, as shown in Fig. 16, followed rapidly by concrete crush-
ing at the top. In case of the strengthened beam with Fe-SMA, the
failure was concrete crushing after the yielding of the longitudinal
normal steel reinforcement and the so-called yielding of the Fe-
SMA strips. However, the failure modes of Beam 4 and 6 were con-
crete crushing after a tensile failure in one of the Fe-SMA strips.
The maximum achieved loads for the examined beams (Pmax)
are depicted in Table 4. The maximum loads of the strengthened
beams were much higher than for the reference beam. The maxi-
mum load represents the applied load, in excess of the beam
dead-weight, that resulted in the failure of the beam. The load
capacities of the Fe-SMA strengthened beams were up to 80%
greater than for the reference beam. Beam No. 5 with the CFRP
strip (the highest strength) had the highest failure load. However,
Beam No. 2 exhibited the largest deflection at failure compared
with the other strengthened beams. A comparison of the examined
beams in terms of the mid-span deflections at the maximum
reached load is presented in Table 4. The results presented in
Figs. 16, 17(b), and Table 4 show that the application of prestressed
Fe-SMAs caused a decrease in the beam deflection and conse-
quently an increase in the load capacity at the serviceability limit
state.
The ultimate load capacities for prestressed and unprestressed
beams were the same, which is the usual behavior in conventional
Mid-span deflection at Cracking load, Mid-span deflection at
Pmax Pcrack P = 8 kN
(mm) (kN) (mm)
48.7 2.0 17.0
70.6 2.5 5.6
56.9 4.5 4.6
51.1 4.7 3.9
55.9 2.4 7.1
52.0 4.2 4.1
prestressed concrete beams also. Experimental results and the
observed modes of failure and the load–deflection behavior of
the examined beams are shown in Tables 3 and 4.
In Beams No. 4 and No. 6, drops in the load occurred due to
one Fe-SMA strip rupture. However, a large amount of deflection,
more than 50 mm equal to 1/40 of the span length, had taken
place prior to the Fe-SMA rupture. The strains in the Fe-SMA
strips were estimated based on the fulfilled displacement record
on the side view of the concrete beams, using the DIC measure-
ment system. A typical result of the DIC during the loading after
the concrete cracking is depicted in Fig. 19(a). This figure shows
the major strains and the occurrence of four main cracks in the
mid-span of Beam No. 3 up to a load level of 11.0 kN. A line that
is located at approximately 850–1150 mm from the left support of
the beam was defined between two points at the central height of
the Fe-SMA strips, 10 mm from the bottom side of the concrete
beam. The length change of this line divided by the initial length
was assumed to be equal to the overall strain of the Fe-SMA strips
along the constant moment region of the beam. This estimation
was performed for all four Fe-SMA strengthened beams from
the start of the loading up to the failure load. The estimated
strains of the Fe-SMA strips at the mid-span region are illustrated
in Fig. 19(b). In Beam No. 2 and 3, no tensile Fe-SMA failure
occurred; however, in Beam No. 4 and 6, tensile failure of one
of the Fe-SMA strips occurred. The reached strains in Beam No.
4 and 6 were approximately 2.2% and 3.8%. These strains are in
addition to the residual strains of approximately 1.5% in the strips
after the prestraining. Therefore, the maximum obtained strains
at tensile failure were 3.7% and 5.3%. These values are relatively
low compared to the determined tensile failure strains, which
are in the range of 8–14% [5]. The reason for this behavior is
not fully clear yet. Possibly, micro-cracks could decrease the ten-
sile failure strain of the Fe-SMA strips.
 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38
Fig. 19. Digital image correlation (DIC) measurements: (a) major strain at loading
of around 11 kN at the middle area of beam No. 3 and (b) estimated Fe-SMA strip
strains along the constant moment region of all beams.
4. Conclusions
In the present study, the application of an iron-based shape
memory alloy (Fe-SMA) developed at Empa for near-surface
mounted (NSM) strengthening was successfully demonstrated.
The following main conclusions can be drawn:
1. The general feasibility of ribbed Fe-SMA strips for reinforcing
and prestressing concrete beams was shown. The Fe-SMA strips
could be activated, i.e., prestressed, by resistive heating, and the
prestressing force could be transferred to the concrete section.
2. The cracking load of the beams that were strengthened with
prestressed Fe-SMA strips was on average 80% higher than the
cracking load of a beam strengthened with the same Fe-SMA
without prestressing. The displacement at a load of 8 kN was
75% less for the prestressed beams compared to the non-
prestressed beam with the same amount of Fe-SMA
strengthening.
3. To conduct the described experiments, several new ideas and
approaches have been successfully invented and employed,
including the manufacture of ribbed Fe-SMA strips with reason-
able dimensions for structural strengthening, the activation of
long Fe-SMA strips embedded in concrete beams by the
approach of resistive heating, the application of Fe-SMAs as pre-
stressed NSM strengthening, and strain measurements on the
embedded Fe-SMA strips by DIC.
37
4. Prestressing offers many advantages, including reduced crack
widths, reduced deflections, reduced stress in the internal steel,
and possibly increased fatigue resistance. Similar to conven-
tional prestressing, the failure loads do not increase.
5. Fe-SMA strips can be more easily prestressed than CFRP strips.
Because prestressing the SMAs does not require mechanical
jacks and anchorheads, the additional openings on the concrete
surface beside the grooves that are necessary to clamp the NSM
strips are significantly smaller. Fe-SMAs can be prestressed
even if they are embedded without the need for a duct in the
concrete.
6. The bond behaviors of the ribbed Fe-SMA glued in grooves with
cement-based mortar were good and were judged to be suffi-
cient for strengthening applications. The determined recovery
stresses (i.e., prestresses after activation) were approximately
200 MPa.
The results of the performed experiments showed that the pre-
sented procedure of strengthening and prestressing of RC beams by
Fe-SMAs was accomplished successfully. Further studies to
demonstrate the application of Fe-SMA in real structures are con-
tinuing at Empa.
Acknowledgements
This study was financially supported by the Swiss Commission
for Technology and Innovation (CTI project 14496.1 PFIW-IW),
which is greatly appreciated. Furthermore, thanks go to the Insti-
tute of Ferrous Metallurgy at the Montanuniversität Leoben in Aus-
tria, the Institute of Metal Forming at the TU Bergakademie
Freiberg in Germany and the company G. Rau GmbH & Co. KG,
Pforzheim in Germany, which produced the iron-based shape
memory strips. Thanks go also to the company S&P Clever Rein-
forcement, Seewen, Switzerland, which provided the CFRP strips
and epoxy-based adhesive. In addition the authors want to thank
the technicians from the Bauhalle at EMPA, Switzerland. Finally,
the authors would like to thank the company re-Fer AG in Wol-
lerau, Switzerland, for their support.
References
[1] U. Meier, Strengthening of structures using carbon fibre epoxy composites,
Constr. Build. Mater. 9 (6) (1995) 341–351.
[2] L. De Lorenzis, J.G. Teng, Near-surface mounted FRP reinforcement: an
emerging technique for strengthening structures, Compos. Eng. 38 (2) (2007)
119–143.
[3] M. Motavalli, C. Czaderski, K. Pfyl-Lang, Prestressed CFRP for strengthening of
reinforced concrete structures – recent developments at Empa Switzerland, J.
Compos. Constr., ASCE 15 (2) (2011) 194–205.
[4] R. El-Hacha, K. Soudki, Prestressed near-surface mounted fibre reinforced
polymer reinforcement for concrete structures – a review, Can. J. Civ. Eng. 40
(11) (2013) 1127–1139.
[5] C. Czaderski, M. Shahverdi, R. Brönnimann, C. Leinenbach, M. Motavalli,
Feasibility of iron-based shape memory alloy strips for prestressed
strengthening of concrete structures, Constr. Build. Mater. 56 (2014) 94–105.
[6] L. Janke, C. Czaderski, M. Motavalli, J. Ruth, Applications of shape memory
alloys in civil engineering structures – overview, limits and new ideas, Mater.
Struct. 38 (279) (2005) 578–592.
[7] A. Sato, E. Chishima, K. Soma, T. Mori, Shape memory effect in c¢e
transformation in Fe-30Mn-1Si alloy single crystals, Acta Metall. 30 (6)
(1982) 1177–1183.
[8] A. Cladera, B. Weber, C. Leinenbach, C. Czaderski, M. Shahverdi, M. Motavalli,
Iron-based shape memory alloys for civil engineering structures: an overview,
Constr. Build. Mater. 63 (2014) 281–293.
[9] Z. Dong, U.E. Klotz, C. Leinenbach, A. Bergamini, C. Czaderski, M. Motavalli, A
novel Fe-Mn-Si shape memory alloy with improved shape recovery properties
by VC precipitation, Adv. Eng. Mater. 11 (1–2) (2009) 40–44.
[10] C. Leinenbach, H. Kramer, C. Bernhard, D. Eifler, Thermo-mechanical properties
of an Fe-Mn-Si-Cr-Ni-VC shape memory alloy with low transformation
temperature, Adv. Eng. Mater. 14 (1–2) (2012) 62–67.
[11] W.J. Lee, B. Weber, G. Feltrin, C. Czaderski, M. Motavalli, C. Leinenbach, Phase
transformation behavior under uniaxial deformation of an Fe–Mn–Si–Cr–Ni–
VC shape memory alloy, Mater. Sci. Eng., A 581 (2013) 1–7.
38 M. Shahverdi et al. / Construction and Building Materials 112 (2016) 28–38
[12] W.J. Lee, B. Weber, G. Feltrin, C. Czaderski, M. Motavalli, C. Leinenbach, Stress
recovery behaviour of an Fe–Mn–Si–Cr–Ni–VC shape memory alloy used for
prestressing, Smart Mater. Struct. 22 (12) (2013) 9.
[13] W.J. Lee, B. Weber, C. Leinenbach, Recovery stress formation in a restrained Fe-
Mn-Si-based shape memory alloy used for prestressing or mechanical joining,
Constr. Build. Mater. 95 (2015) 600–610.
[14] O.E. Ozbulut, S. Hurlebaus, R. Desroches, Seismic response control using shape
memory alloys: a review, J. Intell. Mater. Syst. Struct. 22 (14) (2011) 1531–
1549.
[15] A. Zafar, B. Andrawes, Experimental flexural behavior of SMA-FRP reinforced
concrete beam, Front. Struct. Civ. Eng. 7 (4) (2013) 341–355.
[16] A. Cladera, E. Oller, C. Ribas, Pilot experiences in the application of shape
memory alloys in structural concrete, J. Mater. Civ. Eng. 26 (11) (2014).
[17] M.S. Saiidi, M. Sadrossadat-Zadeh, C. Ayoub, A. Itani, Pilot study of behavior of
concrete beams reinforced with shape memory alloys, J. Mater. Civ. Eng. 19 (6)
(2007) 454–461.
[18] M.S. Saiidi, M. O’Brien, M. Sadrossadat-Zadeh, Cyclic response of concrete
bridge columns using superelastic nitinol and bendable concrete, ACI Struct. J.
106 (1) (2009) 69–77.
[19] H. Tran, X. Balandraud, J.F. Destrebecq, Curvature effect on the mechanical
behaviour of a martensitic shape-memory-alloy wire for applications in civil
engineering, Smart Mater. Struct. 24 (2) (2015).
[20] T. Sawaguchi, I. Nikulin, K. Ogawa, K. Sekido, S. Takamori, T. Maruyama, Y.
Chiba, A. Kushibe, Y. Inoue, K. Tsuzaki, Designing Fe-Mn-Si alloys with
improved low-cycle fatigue lives, Scripta Mater. 99 (2015) 49–52.
[21] T. Maruyama, T. Kurita, S. Kozaki, K. Andou, S. Farjami, H. Kubo, Innovation in
producing crane rail fishplate using Fe-Mn-Si-Cr based shape memory alloy,
Mater. Sci. Technol. 24 (8) (2008) 908–912.
[22] T. Maruyama, H. Kubo, 12 - Ferrous (Fe-based) shape memory alloys (SMAs):
properties, processing and applications, in Shape Memory and Superelastic
Alloys, in: K. Yamauchi et al. (Eds.), Shape Memory and Superelastic Alloys,
Woodhead Publishing, 2011, pp. 141–159.
[23] A.K. Maji, I. Negret, Smart prestressing with shape-memory alloy, J. Eng. Mech.
124 (10) (1998) 1121–1128.
[24] P. Soroushian, K. Ostowari, A. Nossoni, H. Chowdhury, Repair and
strengthening of concrete structures through application of corrective
posttensioning forces with shape memory alloys, in Transportation Research
Record, 2001. pp. 20–26.
[25] Y. Watanabe, E. Miyazaki, H. Okada, Enhanced mechanical properties of Fe-
Mn-Si-Cr shape memory fiber/plaster smart composite, Mater. Trans. 43 (5)
(2002) 974–983.
[26] T. Sawaguchi, T. Kikuchi, K. Ogawa, S. Kajiwara, Y. Ikeo, M. Kojima, T. Ogawa,
Development of prestressed concrete using Fe-Mn-Si-based shape memory
alloys containing NbC, Mater. Trans. 47 (3) (2006) 580–583.
[27] S. El-Tawil, J. Ortega-Rosales, Prestressing concrete using shape memory alloy
tendons, ACI Struct. J. 101 (6) (2004) 846–851.
[28] M. Sherif, O. Ozbulut, A. La, R.F. Hamilton, Self-post-tensioning for concrete
beams using shape memory alloys, in: ASME 2014 Conference on Smart
Materials, Adaptive Structures and Intelligent Systems, SMASIS 2014, 2014.
[29] C. Czaderski, B. Hahnebach, M. Motavalli, RC beam with variable stiffness and
strength, Constr. Build. Mater. 20 (9) (2006) 824–833.
[30] K. Moser, A. Bergamini, R. Christen, C. Czaderski, Feasibility of concrete
prestressed by shape memory alloy short fibers, Mater. Struct. 38 (279) (2005)
593–600.
[31] M. Shin, B. Andrawes, Experimental investigation of actively confined concrete
using shape memory alloys, Eng. Struct. 32 (3) (2010) 656–664.
[32] H. Tran, X. Balandraud, J.F. Destrebecq, Improvement of the mechanical
performances of concrete cylinders confined actively or passively by means of
SMA wires, Arch. Civ. Mech. Eng. 15 (1) (2015) 292–299.
[33] M. Shin, B. Andrawes, Lateral cyclic behavior of reinforced concrete columns
retrofitted with shape memory spirals and FRP wraps, J. Struct. Eng. 137 (11)
(2011) 1282–1290.
[34] Q. Chen, M. Shin, B. Andrawes, Experimental study of non-circular concrete
elements actively confined with shape memory alloy wires, Constr. Build.
Mater. 61 (2014) 303–311.