Construction and Building Materials 274 (2021) 121793

Contents lists available at ScienceDirect

Construction and Building Materials

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

Iron based shape memory alloys as shear reinforcement for bridge

girders
Christoph Czaderski a,⇑, Moslem Shahverdi a, Julien Michels b
a
Empa, Structural Engineering, Dübendorf, Switzerland
b
re-fer AG, Seewen, Switzerland

h i g h l i g h t s

 Iron-based shape memory alloy in the form of U-shaped ribbed bars.

 Used in combination with sprayed cement based mortar.
 For shear strengthening of reinforced concrete (RC) beams.
 Prestressed shear strengthening.
 Reduction of beam deflections, stresses in internal steel stirrups, crack widths.

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

Article history: Considerable age of a very high number of bridges conjointly with a steadily increasing amount of traffic
Received 17 August 2020 and changes in design philosophy (e.g. earthquake engineering) have made maintenance needed and ret-
Received in revised form 5 November 2020 rofitting become more and more important over the years. Retrofitting can become necessary both for
Accepted 20 November 2020
flexure or shear enhancements. Existing steel solutions for shear strengthening are very laborious and
Available online 6 December 2020
complex, and the durability of the steel construction is questionable. As an alternative to steel solutions,
carbon fiber reinforced polymer (CFRP) sheets or strips are used for shear strengthening of reinforced
Keywords:
concrete beams. But, prestressing of CFRP sheets or strips is hardly applicable. However, a prestressing
Shear strengthening
Reinforced concrete
of a shear strengthening has the advantages that the width of existing shear cracks can be reduced
Shape memory alloys and the stresses in the internal steel stirrups are reduced.
Prestressed strengthening Therefore, in this study, a new iron-based shape memory alloy (‘memory-steel’) in the form of U-
Digital image correlation shaped (stirrups) ribbed bars with a nominal diameter of 12 mm were used in combination with sprayed
Shear cracks mortar for shear strengthening of reinforced concrete (RC) structures. The memory-steel bars were acti-
vated with electric resistive heating. The activation resulted in a prestress of about 300 N/mm2 in the
memory-steel reinforcement and consequently in vertical compressive stresses in the web of the RC
beams. Large-scale experiments on T-beams with a height of 0.75 m and a total length of 5.2 m were per-
formed to show the practicability and efficiency of the memory-steel shear strengthening. Promising
results have shown that the new strengthening system works well in practice. The shear capacity could
be increased significantly. Furthermore, at the serviceability limit state, the prestressed memory-steel
stirrups reduced the overall beam deflections, the stresses in the internal steel stirrups, the number of
cracks, and the crack widths.
Ó 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

1. Introduction integrity in combination with their construction and maintenance

costs is one of the key challenges civil engineers are confronted
Built infrastructure, such as buildings, bridges, highways, dams, with. Since infrastructure (beside buildings) is in its large majority
tunnels, water supply systems, and many others guarantee the public-owned, investments often directly derive from tax money
proper functioning of our daily living standards. Their structural and hence are of high public interest. For several decades already,
reinforced concrete is a very well-known and efficient building
⇑ Corresponding author at: Empa, Empa, Structural Engineering Research Labo- material and represents nowadays by far the most spread con-
ratory, Überlandstrasse 129, 8600 Dübendorf, Switzerland. struction material all over the world. Considerable age of a very
E-mail address: [email protected] (C. Czaderski).

https://doi.org/10.1016/j.conbuildmat.2020.121793
0950-0618/Ó 2020 The Author(s). Published by Elsevier Ltd.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

high number of for instance bridges conjointly with a steadily 15]. The hydraulic jacks and the mechanical anchorage system of
increasing amount of traffic and changes in design philosophy conventional prestressing techniques would become obsolete.
(e.g. earthquake engineering) have made maintenance needed Therefore, prestressing is easier and cheaper when using SMA,
and retrofitting become more and more important over the years. because ducts, anchor heads, and oil hydraulic jacks are not neces-
Maintenance costs represent nowadays a very important part of all sary anymore. The material in the shape of a ribbed bar is initially
investments related to a specific construction type. In the United prestrained and can subsequently be placed into a concrete struc-
States, for instance, maintenance costs for bridges are above 10 bil- ture or externally installed and being surrounded afterwards with
lion USD per year according to a study by the American Society of an additional cementitious layer (e.g. shotcrete or sprayed mortar),
Civil Engineers [1]. [16]. Once the surrounding material has sufficiently cured, heating
Prestressing a reinforced concrete structure is an even more is applied to the steel bars by resistive heating. Since the bars can-
efficient technique in terms of durability and serviceability, as not move back to their initial state as they are blocked by the con-
cracks and deformations can be reduced or at least prevented from crete, they will be prestressed and consequently, the concrete will
further growing. This technique is a common construction method undergo compression stresses. (Iron-based) shape memory alloys
all over the world: a large number of bridges are fully or partially further have the advantage that in case of a future demolition,
designed this way, especially to overcome large spans and to guar- the material can be fully recycled and hence offers a strong contri-
antee the mentioned satisfactory behavior under service loads. bution to the circular economy. In Europe and North America,
Prestressed tendons are mainly used in the longitudinal direction stainless steel production is almost solely based on recycling
(flexural reinforcement). Very rare is the usage of prestressing for (melting and re-casting) scrap material. Iron-based shape memory
shear strengthening because it is very complicated from a practical alloys can, due to their metallurgical composition (see description
point of view. However, according to for example the Swisscode below for the alloy composition of the Fe-SMA used in this study),
SIA 166 ‘‘Externally bonded reinforcement” [2], the provision is be treated in a similar manner as stainless steel and hence be re-
given that, if a beam with shear cracks at serviceability state shall introduced into the casting process.
be strengthened for shear, then, the strengthening shall envelop Lastly, fire protection is also far less invasive compared to for
the total beam height and must be prestressed. The advantages if instance FRP products, which due to the low glass transition tem-
the shear strengthening is prestressed are [3]: (i) the shear cracks perature of the resin require substantial measures to protect them
can be reduced or even closed, (ii) the existing internal steel stir- from high temperatures. The behavior of iron based shape memory
rups are partly unloaded what increases their fatigue resistance alloys under elevated temperatures is comparable to steel [17],
and (iii) the strengthening takes immediately additional loads. which leads to simpler and less cost-intensive protection systems
However, prestressing of the shear strengthening is hardly possible compared to FRP’s.
today with the conventional systems. Fig. 1 shows a (prestressed) The most commercially available shape memory alloy on the
shear strengthening example on a Swiss bridge. It can be seen that market is Nickel/Titanium (NiTi). However, the commercial appli-
such solutions are very laborious and complex, and the durability cation of NiTi is limited due to the high costs of raw materials
of the steel construction is questionable. As an alternative to the and processing. The effectiveness of NiTi for prestressing is often
steel solutions, carbon fiber reinforced polymer (CFRP) sheets or restricted by their relatively narrow thermal hysteresis, which is
strips are used for shear strengthening of reinforced concrete inadequate for the large and stable recovery stresses required for
beams [4–9]. However, prestressing of the CFRP sheets or strips civil engineering applications. In a study presented in [18], a con-
is hardly possible. The only CFRP system, which can prestressed, crete beam was reinforced with NiTi wires. It was found that if
is the unbonded carbon fiber–reinforced polymer (CFRP) strap sys- the NiTi wires are heated, they are prestressed and have a larger
tem from the company Carbo-Link, a spin-off from Empa in stiffness, however, if they are not heated, they lose the prestress
Switzerland, which was originally developed at Empa [3,5,10–12]. and have a reduced stiffness. As an alternative, iron-based shape
However, the application of shape memory alloy stirrups can memory alloys such as Fe-Mn-Si based alloys are more suitable
provide the possibility to implement prestressed shear strengthen- for this type of application due to their wide transformation hys-
ing. Shape memory alloys have the characteristic that they, after teresis, lower cost, and high elastic stiffness compared to conven-
having been pre-deformed, move partly back to their initial shape tional Nitinol alloys [15,19]. Sato et al. discovered the shape
when subjected to high temperatures. This characteristic leads to a memory effect in these alloys in 1982 [20]. A Fe-Mn-Si based alloy
completely new philosophy in terms of a prestressing technique especially feasible for construction was invented at Empa [21]. The
for concrete constructions based on shape memory effects [13– composition of the alloy is Fe-17Mn-5Si-10Cr-4Ni-1(V,C) (mass%).

Fig. 1. Example of a conventional shear strengthening on a Bridge near Solothurn in Switzerland.

2
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Related products were then developed in collaboration with re-fer and several measurements during the loading up to failure were
AG founded in 2012 to commercialize these reinforcements [22]. performed to quantify the effect of the prestressed stirrups. A fur-
The Fe-Mn-Si based alloy is generally known as ‘memory-steel’. ther aim of the investigation was to show the practicability of the
memory-steel is produced in the form of ribbed bars and strips new strengthening technique for on-site applications.
[14,23–25].
The research in the field of SMA for shear strengthening started
already in 2001 [26]. Strengthened successfully a RC beam in the 2.1. Strengthening procedure
laboratory and also a real bridge in the US by using iron based
shape memory alloy rods. Unfortunately, according to the knowl- The strengthening procedure on the RC beams included the fol-
edge of the authors, this research was not continued. Only again lowing working steps:
in 2016 [27], published a study on pseudoelastic Ni-Ti (commonly
known as Nitinol or NiTi) continuous spiral shear reinforcements 1. Drilling of holes in the flange (Fig. 5)
for rectangular RC beams to obtaining more ductile shear failures. 2. Concrete hammering in the strengthening area (Fig. 2)
The beams reinforced with the Ni-Ti rectangular spirals showed 3. Sandblasting of the hammered zones to remove all loosely con-
high ductility in shear, with very high deflections and large crack crete parts
widths at failure. In the same institute, proof-of-concept 4. Installation of the U-shaped ribbed memory-steel stirrups
researches on concrete beams with the dimensions (Fig. 3)
80x150x900 mm were performed. Firstly [28], presented an inves- 5. Application of formwork for the repair mortar (Fig. 4)
tigation on shear strengthening of the rectangular beams by using 6. Application of a bonding primer (Sika MonoTop-910N) to the
3 mm diameter Ni-Ti-Nb wires externally installed around the concrete surface
beams, forming a pseudo-spiral, to actively confine, or transver- 7. Application of the repair mortar (Sika MonoTop-412N) by
sally prestress, the RC members. Secondly [29], used 12.5 and spraying (Fig. 4)
25 mm wide and 0.5 mm thick iron based shape memory alloys 8. Filling of the holes in the flange with cementitious grout
(Fe-SMA) strips. The strips were wrapped around the rectangular (SikaGrout-314)
cross-section, connected with a buckle, and then activated (pre- 9. Activation of the memory-steel stirrups by resistive heating
stressed) what means that they were heated with a heat gun.
The shear strengthened beams showed a significant higher failure The thickness of the sprayed mortar layer was approximately
load and for the beams strengthened with activated strips, the 30 mm. Therefore, the U-shaped ribbed memory-steel stirrups
appearance of shear cracks was delayed, there were fewer shear had a cover of 10–20 mm in the test beams. However, similar as
cracks, and the deflections of beams were significantly lower, in usual reinforced concrete, the concrete protects the reinforce-
which revealed the benefits of this active shear strengthening tech- ment for corrosion and larger covers are recommended for real
nology. Lastly [30], recently published a study on the shear applications.
strengthening of real-scale RC T-beams by means of U-shaped
Fe-SMA strips. The T-beams had a length of 5.5 m and a height
2.2. Beam geometry and materials
of 0.55 m. All the beams failed in shear and the retrofitted beams
showed a 30% increased shear strength. The active shear strength-
The geometries of the beams were chosen so that they are rep-
ening delayed the appearance of cracks and reduced their crack
resentative of onsite application; the height, the web width, and
width at service loads. However, in the study, problems with the
the span were 0.75 m, 0.16 m, and 4.3 m, respectively (Fig. 5
mechanical fasteners to fix the Fe-SMA strips to the web occurred
and Fig. 7). A further aim of the experiments was to reach shear
so that the proper performance of the Fe-SMA strips were pre-
failure so that the behavior of the new shear strengthening method
vented and it was concluded for future research another type of
under shear loading and shear failure could be investigated. There-
mechanical anchorage or a fully wrapping method shall be used.
fore, the moment/shear ratio a/d (a = distance between load and
[31] used also 25 mm wide and 0.5 mm thick iron-based shape
support, d = static height) was chosen as 1800/665 = 2.7, what is
memory alloys (Fe-SMA) strips to shear strengthen 2.3 m long and
a good ratio for shear failure (Fig. 7) according to the ‘‘Kani valley”,
0.35 m high T-beams. Although there were some issues with the
see e.g. [32]. Additionally, a high strength flexural reinforcement (4
anchorage of the strips at the flanges and more development work
bars Ø=30 mm, yield strength of 843 MPa, tensile strength of
would be necessary, generally, they found a significant increase in
945 MPa) was used to avoid premature flexural failure. The inter-
both strength and ductility of the strengthened beams compared to
nal steel stirrups with a diameter of 8 mm had a yield strength of
the control specimens.
504 MPa and tensile strength of 560 MPa. Three stirrups were
In this research, a large-scale experimental campaign on RC
placed in each shear span. The locations of the stirrups are dis-
beams, which were strengthened for shear by a new innovative
played in Fig. 7.
method is presented. The new method uses U-shaped ribbed
memory-steel bars, which are embedded in an additional sprayed
mortar layer and are prestressed after the mortar has cured.

2. Experiments

Large-scale reinforced concrete (RC) beams with a shear defi-

cient reinforcement were constructed, strengthened for shear by
using U-shaped ribbed memory-steel bars embedded in an addi-
tional sprayed mortar layer, and lastly loaded to failure. The aim
was to show the efficiency of the new memory-steel shear
strengthening method. The main focus was on the effect of the pre-
stress of the stirrups. To study the effect of the prestress, reference
beams that were not prestressed were examined as comparison Fig. 2. Concrete hammering on one of the test beams.

3
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 3. Installed memory-steel stirrups on one of the test beams. The red bolts are
polymer dowels for fixation of the memory-steel stirrups on the concrete surface.
(For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.) Fig. 6. Stress–strain diagram of a ribbed memory-steel bar, similar as used in this
study.

with a mean strain at failure Agt of 20%. A typical stress–strain dia-

gram is displayed in Fig. 6. The first part of it corresponds to Action
(1) as discussed below, see Fig. 8.
On the surface of the hammered concrete, the bonding primer
Sika MonoTop-910 N was used. According to the technical data
sheet of the supplier, it is a cementitious, polymer-modified one-
component coating material containing silica fume and has a com-
pressive strength after 28 days of approximately 35 N/mm2. The
repair mortar Sika MonoTop-412 N is, according to the technical
data sheet of the supplier, a 1-component, fiber reinforced, low
shrinkage repair mortar, and has a compressive and flexural tensile
strength after 28 days of approximately 55 and 8 N/mm2, respec-
Fig. 4. Application of spray mortar to one of the test beams. Timber formwork on
tively. For the filling of the holes in the flange, SikaGrout-314
the sides and the bottom.
was used. According to the technical data sheet of the supplier, it
is a 1-component, cementitious, expanding, low shrinkage grout
and has a compressive strength after 28 days of approximately
90 N/mm2.

2.3. Experimental program

The experimental program is presented in Table 1 and Fig. 7.

Beam 1 was the unstrengthened reference beam. Beams 2 and 3
had the same strengthening scheme (i.e. three memory-steel stir-
rups), however, the memory-steel stirrups were only activated
(prestressed) in Beam 2, so that the beneficial effect of prestress
could be studied. Beams 5 and 6 were strengthened with five
memory-steel stirrups, whereas again only the memory-steel stir-
rups in Beam 5 were activated.
Furthermore, the reference beam without strengthening (Beam
1) was repaired after the failure test and tested again (Beam 4),
[33].

2.4. Prestraining, activation, and loading of memory-steel stirrups

The main actions for the shear strengthening by using memory-

steel are illustrated in Fig. 8a): Action (1): The memory-steel rein-
forcements were pre-strained to a strain level of about 4%, after-
Fig. 5. Cross-section of the test beams including the strengthening method. ward fully released and then formed in U-shape. In the next step,
they were embedded in the sprayed mortar as described above
(Fig. 4). After the mortar had cured, the beams were transported
The concrete was produced and delivered from a local concrete to the laboratory. Then, the memory-steel U-shaped stirrups were
plant. It had a maximum aggregate size of 16 mm and the obtained activated by resistive heating, Action (2) in Fig. 8a) and b). For that
strengths are given in Table 1. reason, the electric power supply from the company re-fer AG,
The ribbed memory-steel bars with a nominal diameter of which is specially designed for activation of memory-steel, was
12 mm (effective approximately 11 mm) were delivered by the connected by cooper clamps (Fig. 9 and Fig. 10). By using a current
company re-fer. They were prestrained to 4% and bent afterwards of 400 A (current density of 3.5 A/mm2), one stirrup after the other
in a U-shape. Tensile mean strength Rm of the bars was 810 N/mm2 was heated to approximately 160 °C, see the measured tempera-
4
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 7. Geometry and reinforcement of the beams.

Table 1
Overview of the experimental program. Concrete strength of the test beams at the day of failure tests (fcm,cube: compression test on three cubes 150 mm3, fctm: splitting tensile test
on three cubes 150 mm3).

Internal steel stirrups in External memory-steel stirrups memory-steel fc,cube fct

a shear span in a shear span prestressed MPa MPa
Beam 1 3 – – 48.1 3.90
Beam 2 3 3 Yes 46.5 3.45
Beam 3 3 3 No 43.5 3.45
Beam 4 (the damaged 3 Yes – –
repaired Beam 1)
Beam 5 3 5 Yes 58.7 4.85
Beam 6 3 5 No 42.1 3.25

5
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 8. a) Schematic of three main actions of application of memory-steel to strengthen and prestress a RC structure. b) Schematic illustration of the activation, Action (2), and
the recovery stress development.

memory-steel bars, a prestress of approximately 300 MPa can be

expected for heating temperatures measured in Fig. 11. Therefore,
if a web width of 220 mm is assumed (including the sprayed mor-
tar), Beam 2 and 5 with three and five memory-steel stirrups had
before the start of the loading experiment vertical compression
stress in the concrete web of approximately 0.6 MPa and
0.9 MPa, respectively.
Lastly, Action (3), when the concrete structure was loaded, the
memory-steel U-shaped stirrups performed as shown in Fig. 8a.

2.5. Experimental set-up

The beams were investigated in a four-point bending loading

scheme (Fig. 7 and Fig. 12). The loading was applied by using
two 1000 kN hydraulic jacks. The beams were supported by two
Fig. 9. One side of Beam 2 with the sprayed mortar area, the thermocouples for the
rolling supports, one of them horizontally fixed.
temperature measurements during activation, and the U-shaped memory-steel
stirrups which are visible on the top side.

2.6. Measurements

Beside conventional mid-span displacement with LVDT’s, strain

gauge measurements (designation SG in Fig. 13) and load cells, also
digital image correlation (DIC) measurements were performed
(Fig. 12) with two measurement fields (gray area in Fig. 13) on both
shear spans to monitor the crack pattern and crack width develop-
ments. Strain gauges were not attached to the memory-steel stir-
rups, because the memory-steel stirrups were heated to 160–
200 °C, what could damage the strain gauges.

2.7. Repair of a damaged beam

As described in the overview of the experimental program in

Table 1, besides the investigation of new beams, also a beam,
which was loaded to failure was repaired and then tested again.
Fig. 10. Detail of the cooper clamps for the connection of the electric power supply To be more specific, the reference beam without strengthening
(resistive heating) with the U-shaped memory-steel stirrups on Beam 2. (Beam 1) was repaired after the failure test. Firstly, the large shear
cracks in Beam 1 were closed at the sides with a mortar (Sika
FastFix-121) and subsequently injected with a two-part injection
tures given in Fig. 11. The temperatures were measured with ther-
grout (Sika InjectoCem-190), Fig. 14. Afterward, the same strength-
mocouples (Fig. 9), which were attached to the stirrups before the
ening steps as described above (Steps 1 to 9) were performed
casting. It is visible in Fig. 11, that the maximum temperatures
(Fig. 15). Then, the beam was loaded to failure. The experiment
were reached after 1.5 min, whereas the cooling down to room
on the repaired beam was designated as Beam 4. The internal steel
temperature needed approximately 20 min. For the ribbed
stirrups, which ruptured during the failure test on Beam 1 were not

6
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 11. Measured temperatures of the memory-steel stirrups in Beam 2 during their activation. The numbering of the stirrups is given in Fig. 7. For each stirrup, two
measurements were executed.

failure of the flange under the loading point (Fig. 16b and c), which
was not observed in the reference test beam (Beam 1), see Fig. 16a.
Additional to the shear cracking behavior in the web, visible in
Fig. 16, cracks in the flange on the top side of the beams could be
observed. Fig. 18 displays a photo of the cracks on the left (side of
shear failure) top side of Beam 5 after the shear failure. The trans-
verse cracks (No. 1 in Fig. 18) in cross-direction of the beam are
bending cracks in the longitudinal direction, originating from the
small shear stiffness in the longitudinal direction. However, the
longitudinal cracks (No. 2 in Fig. 18 and Fig. 19) indicate a bending
in cross-direction, which might originate from the shear failure of
the beam. Furthermore, in some cases, a slight slip of the stirrups
(Fig. 20) on the top side of the flange was observed after the failure
tests, what might also originate from the shear failure of the beam.

Fig. 12. Set-up for the large-scale experiments (1: tripods and 2: cameras for the
both DIC systems, 3: beam supports, 4: lamps for DIC measurements, 5: oil 3.2. Force-mid-span displacement
hydraulic loading cylinders, 6: wooden plates for light shielding for DIC measure-
ments, 7: test beam, 8: area of sprayed mortar). The force-mid-span displacement behaviors of all five beams
are given in Fig. 21. The strengthening with the memory-steel stir-
rups and shotcrete mortar (increase of the web width from 160 to
repaired. The results of the Beam 1 and 4 were presented already in 220 mm) increased the failure forces by 86% (Table 2). Addition-
[33]. ally, the mid-span displacements at a serviceability force of
200 kN are also compared in Table 2 and Fig. 21b. Shear strength-
ening decreased the mid-span considerably down to a minimum of
3. Results 61% for the case of Beam 5.
By comparing the non-prestressed and prestressed memory-
3.1. Failure modes steel strengthening in Table 2 and Fig. 21, the following conclu-
sions can be derived: (1) the failure force was not influenced by
All beams failed in shear (Fig. 16). In all beams, the internal stir- a prestress of the U-shaped memory-steel stirrups, however, (2)
rups yielded and ruptured before the shear failure of the beams. the beam displacements at serviceability was influenced by a pre-
Rupture of the internal steel stirrups was hearable and some were stress. In the case of three and five stirrups, the mid-span displace-
visible via the large cracks (Fig. 17). Yielding of the stirrups ments were reduced if the stirrups were prestressed from 5.5 mm
resulted in the formation of shear cracks with large widths to 4.8 mm and from 4.8 mm to 4.4 mm (approximately 10% reduc-
(Fig. 16). However, the memory-steel stirrups did not rupture tion), respectively. Smaller displacement is due to the higher shear
(Fig. 17). At shear failure, the Beams 2 to 6 showed a compression stiffness of the cross-section if the stirrups are prestressed.

7
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 13. Overview of the measurements performed during the loading experiments up to failure. Displacements, designation SG displays the strain gauge measurements on
the internal steel reinforcements and on the concrete surface, load cells for the load, the gray area designates the fields of the DIC measurements.

Fig. 14. Beam 1 after failure tests. Injected shear cracks. The repaired beam was designated as Beam 4.

repaired and strengthened (Beam 4). It is visible, that the initial

stiffness of Beam 4 is slightly lower as this is due to the flexural
stiffness and the flexural cracks were already produced during
the first loading. However, afterward, the shear stiffness is consid-
erably higher what indicates that the shear strengthening is effec-
tive and works well. The failure mode of Beam 4 was the same as
the Beams 2 to 6, namely a compression failure of the flange under
the loading point Fig. 23). The failure load was even higher as Beam
1 which also shows that the repair and strengthening were
successful.

3.3. Force-strain in flexural reinforcement and concrete at mid-span

The curves of the force-tensile strain in the flexural reinforce-

Fig. 15. Beam 4 after injection of shear cracks, concrete hammering, sandblasting, ment and the force-compressive strain in the concrete, both at
drilling of the holes in the flange, installation of the U-shaped ribbed memory-steel
beam mid-span (Fig. 13), are given in Fig. 24. It is visible, that
stirrups. The strengthening steps 5 to 9 are still missing in the photo.
the tensile reinforcement stayed in elastic state and the concrete
was clearly below 3000 lm/m, at which concrete crushing is usu-
Similar results were reported in [30], in which also beam failure ally expected. These measurements confirm that the design of the
tests with activated and non-activated memory-steel shear test beams was correct, predicting that shear failure occurs prior to
strengthening were presented. The failure loads of the beams with bending failure.
activated and non-activated memory-steel shear strengthening
were similar, however, the appearance of shear cracks in the beams 3.4. Force-strain in internal steel stirrups
with activated shear strengthening were delayed.
The force-mid-span displacement behavior of the beam, which As displayed in Fig. 13, three strain gauges were applied to the
was loaded to failure two times is given in Fig. 22. As described internal steel stirrups. However, some of them were damaged dur-
before, firstly, it was tested as the reference unstrengthened beam ing the strengthening work so that for Beam 2 only one and for
(Beam 1), and secondly, it was loaded to failure after it was Beam 6 only two strain measurement results were available. The

8
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 16. Failure modes of a) Beam 1 (reference beam), b) Beam 2, c) Beam 3, d) Beam 5 and e) Beam 6.

9
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 17. Tensile failures of the internal steel stirrups after the failure test: a) in Beam 5, b) in Beam 6 (1: internal steel stirrups, 2: external applied memory-steel stirrups in
sprayed mortar layer).

Fig. 20. Beam 5: arrows are indicating the slip of the memory-steel stirrups after
the shear failure.
Fig. 18. Beam 5: 1) transverse cracks and 2) longitudinal cracks on the left beam
top side (side of shear failure).
measured strains in the internal steel stirrups are presented in
Fig. 25. Firstly, the beneficial effect of the shear strengthening is
visible, because the strains in the internal stirrups are lower in
the strengthened beams compared to the reference beam. Sec-
ondly, the prestressing of the memory-steel stirrups reduced the
strain in the internal steel stirrups even more, which can also be
observed in Fig. 25a for the Beams 2 and 3 and Fig. 25b for the
Beams 5 and 6.

3.5. Crack pattern and crack widths

As mentioned before, on both shear spans, the crack pattern

and crack width developments were monitored by using a DIC
system. The crack pattern development of Beam 2 and 3 is dis-
played in Fig. 26 for various load steps at the serviceability state
in the force range of 50 to 150 kN. The influence of the pre-
stressing of the memory-steel stirrups is mainly visible from
the different inclinations of the cracks. The cracks in Beam 3
(non-prestressed) on the left side of the figure show more
inclined cracks.
Fig. 19. Beam 5: only one longitudinal crack on the right beam top side (the side As Fig. 26 only shows the qualitative crack pattern develop-
with no shear failure). ment, the software ACDM was used to evaluate the continuous

10
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 21. a) Force-displacement diagrams of Beams 1 to 6 (without 4), b) zoom of a).

Table 2
Results of the large-scale experiments.

Strengthening configuration Maximum load F [kN] Mid-span displacement

[mm] at F = 200 kN
Beam 1 313.4 100% 7.2 100%
Beam 2 3 stirrups prestressed 535.5 171% 4.8 66%
Beam 3 3 stirrups 532.7 170% 5.5 77%
Beam 4 (the repaired damaged 409.1 131% 5.4 75%
Beam 1)
Beam 5 5 stirrups prestressed 583.6 186% 4.4 61%
Beam 6 5 stirrups 583.0 186% 4.8 67%

crack widths. The software was developed at the Institute of Struc-

tural Engineering at the ETH Zürich in Switzerland [34]. In Fig. 27
and Fig. 28, the continuous crack widths at the load stage 150 kN of
Beam 2, Beam 3, and Beam 5, Beam 6 are given, respectively. Addi-
tionally, in the figures, the mean values of all cracks, the mean
absolute value of the crack slips, and the number of cracks are
presented.
The influence of prestressing on the crack widths is visible in
Fig. 27; crack widths are smaller and the cracks are less inclined
in Beam 2, which had prestressed memory-steel stirrups. Also,
the number of cracks is smaller. The same can be observed in
Fig. 28 with Beam 5, which was prestressed. For example, it is vis-
ible in Fig. 27, that Beam 2 with prestressed memory-steel stirrups
Fig. 22. Force-displacement diagram of Beams 1 and 4. had on the left shear span a mean crack width of 0.059 mm,
whereas Beam 3 with non-prestressed memory-steel stirrups had
a mean crack width of 0.092 mm, what is more than 50% larger.
With Fig. 27 and Fig. 28, the influence of using three or five stir-
rups can be observed. The beams with five stirrups had less
inclined cracks than the beams with three stirrups, however, the
crack widths were similar.

3.6. Design

From this research work, it can be concluded, that the ultimate

limit state of shear design can performed by neglecting the pre-
stressing force in the shear reinforcement and using the usual
equations as given in the codes and guidelines. However, the ser-
viceability limit state of shear design is strongly influenced by
the prestressing force in the shear reinforcement. More work is
Fig. 23. Failure mode of Beam 4. necessary to work out a design concept to consider this effect.

11
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 24. Force-tensile strain in flexural reinforcement and compressive strain in the concrete of Beam 1 compared with a) Beam 2 and 3 (three stirrups) and b) Beam 5 and 6
(five stirrups) at beam mid-span (Fig. 13).

Fig. 25. Force-tensile strain in internal steel stirrups (shear reinforcement) in Beam 1 compared with a) Beam 2 and 3 (three stirrups) and b) Beam 5 and 6 (five stirrups). (*:
the location of the internal stirrups with strain gauges were not on the beam side with the shear failure).

12
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Force Beam 3 Beam 2

kN 3 memory-steel stirrups 3 memory-steel stirrups
non-prestressed prestressed

55

92

126

146

Fig. 26. Results of the DIC measurements, which were performed on the concrete surfaces on the beams 2 and 3 during the failure experiments as shown in Fig. 13. The
figures display the maximum principal strains.

13
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

Fig. 27. Crack patterns and widths of the beams at a load of 150 kN: a) and b) Beam 2 (three prestressed stirrups), c) and d) Beam 3 (three non-prestressed stirrups).

a) b) Beam 5, right shear span, load =150 kN:

Beam 5, left shear span, load =150 kN: Mean value of crack widths: 0.045 mm
Mean value of crack widths: 0.057 mm Mean absolute value of crack slips: 0.025 mm
Mean absolute value of crack slips: 0.033 mm Number of cracks: 15
Number of cracks: 7

c) d)
Beam 6, left shear span, load =150 kN:
Beam 6, right shear span, load =150 kN:
Mean value of crack widths: 0.060 mm
Mean value of crack widths: 0.076 mm
Mean absolute value of crack slips: 0.023 mm
Mean absolute value of crack slips: 0.033 mm
Number of cracks: 13
Number of cracks: 13

Fig. 28. Crack patterns and widths of the beams at a load of 150 kN: a) and b) Beam 5 (five prestressed stirrups), c) and d) Beam 6 (five non-prestressed stirrups).

14
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

4. Conclusion [3] Masoud Motavalli, Christoph Czaderski, Kerstin Pfyl-Lang, Prestressed CFRP for
strengthening of reinforced concrete structures: recent developments at Empa,
Switzerland, J. Compos. Constr. 15 (2) (2011) 194–205.
This study on iron-based shape memory alloys for shear [4] Christoph Czaderski, Masoud Motavalli, Fatigue behaviour of CFRP L-shaped
strengthening of concrete girders has revealed several aspects plates for shear strengthening of RC T-beams, Compos. B Eng. 35 (4) (2004)
279–290.
summarized as follows:
[5] J.M. Lees, A.U. Winistörfer, U. Meier, External prestressed carbon fiber-
reinforced polymer straps for shear enhancement of concrete, J. Compos.
& The presented technique on U-shaped ribbed bars embedded in Constr. 6 (4) (2002) 249–256.
[6] Robert M. Foster, Monika Brindley, Janet M. Lees, Tim J. Ibell, Chris T. Morley,
the shotcrete or sprayed mortar layer is a technically feasible
Antony P. Darby, Mark C. Evernden, Experimental investigation of reinforced
method for on-site application without any major difficult concrete T-beams strengthened in shear with externally bonded CFRP sheets, J.
tasks. All the performed steps starting with preparing the Compos. Constr. 21 (2) (2017) 04016086, https://doi.org/10.1061/(ASCE)
memory-steel stirrups over the surface preparation of the gird- CC.1943-5614.0000743.
[7] S.J.E. Dias, J.A.O. Barros, NSM shear strengthening technique with CFRP
ers to finally embedding and activating the shape memory laminates applied in high T cross section RC beams, Compos. B Eng. 114
effect were demonstrated. (2017) 256–267.
& The strengthening efficiency could be demonstrated by several [8] Ahmed Godat, Farah Hammad, Omar Chaallal, State-of-the-art review of
anchored FRP shear-strengthened RC beams: a study of influencing factors,
static large-scale experiments. By adding additional stirrups, Compos. Struct. 254 (2020) 112767, https://doi.org/10.1016/
load-bearing capacity could be substantially increased. The fail- j.compstruct.2020.112767.
ure occurred in shear by rupture of the internal steel stirrups [9] E. Moradi, H. Naderpour, A. Kheyroddin, An experimental approach for shear
strengthening of RC beams using a proposed technique by embedded through-
and a subsequent crushing in the upper compression zone. No section FRP sheets, Compos. Struct. 238 (2020) 111988, https://doi.org/
interface failure between the existing (and subsequently rough- 10.1016/j.compstruct.2020.111988.
ened) concrete surface and the new sprayed mortar layer was [10] A. Winistörfer, Development of non-laminated advanced composites straps for
civil engineering applications (Ph.D. thesis), 1999, University of Warwick, U.K.
observed. The memory-steel stirrups did not rupture.
p. 170.
& Prestressing the memory-steel stirrups allowed to increase the [11] J.M. Lees, A.U. Winistörfer, Nonlaminated FRP strap elements for reinforced
load at which shear cracks appear. Furthermore, at identical concrete, timber, and masonry applications, J. Compos. Constr. 15 (2) (2011)
146–155.
load level, the width of shear cracks and beam mid-span dis-
[12] Feifei Jin, Janet M. Lees, Experimental behavior of CFRP strap–strengthened RC
placements is smaller in case of prestressing than if initially beams subjected to sustained loads, J. Compos. Constr. 23 (3) (2019)
unstressed memory-steel stirrups are used. 04019012, https://doi.org/10.1061/(ASCE)CC.1943-5614.0000936.
& Prestressing the memory-steel stirrups further reduced the [13] Julien Michels, Moslem Shahverdi, Christoph Czaderski, Flexural strengthening
of structural concrete with iron-based shape memory alloy strips, Struct.
stress level in the internal steel stirrups. This is particularly ben- Concr. 19 (3) (2018) 876–891.
eficial for applications in infrastructure constructions under [14] J. Michels, M. Shahverdi, C. Czaderski, R. El-Hacha, Mechanical performance of
cyclic loading and related fatigue problems. iron-based shape-memory alloy ribbed bars for concrete prestressing, ACI
& Mater. J. 115 (6) (2018) 877–886.
Lastly, the investigation showed also that it was possible to [15] L. Janke, C. Czaderski, M. Motavalli, J. Ruth, Applications of shape memory
repair a heavily damaged beam by using the strengthening alloys in civil engineering structures – overview, limits and new ideas, Mater.
method presented in the paper. Struct. 38 (279) (2005) 578–592.
[16] Moslem Shahverdi, Christoph Czaderski, Philipp Annen, Masoud Motavalli,
Strengthening of RC beams by iron-based shape memory alloy bars embedded
in a shotcrete layer, Eng. Struct. 117 (2016) 263–273.
CRediT authorship contribution statement
[17] Elyas Ghafoori, Martin Neuenschwander, Moslem Shahverdi, Christoph
Czaderski, Mario Fontana, Elevated temperature behavior of an iron-based
Christoph Czaderski: Conceptualization, Methodology, Investi- shape memory alloy used for prestressed strengthening of civil structures,
gation, Validation, Writing - original draft, Supervision, Funding Constr. Build. Mater. 211 (2019) 437–452.
[18] Christoph Czaderski, Bernd Hahnebach, Masoud Motavalli, RC beam with
acquisition, Project administration. Moslem Shahverdi: Conceptu- variable stiffness and strength, Constr. Build. Mater. 20 (9) (2006) 824–833.
alization, Methodology, Investigation, Writing - review & editing, [19] A. Cladera, B. Weber, C. Leinenbach, C. Czaderski, M. Shahverdi, M. Motavalli,
Validation. Julien Michels: Conceptualization, Methodology, Inves- Iron-based shape memory alloys for civil engineering structures: an overview,
Constr. Build. Mater. 63 (2014) 281–293.
tigation, Resources, Writing - review & editing, Validation. [20] 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.
Declaration of Competing Interest [21] Zhizhong Dong, Ulrich E. Klotz, Christian Leinenbach, Andrea Bergamini,
Christoph Czaderski, Masoud Motavalli, A novel Fe-Mn-Si shape memory alloy
with improved shape recovery properties by VC precipitation, Adv. Eng. Mater.
The authors declare that they have no known competing finan-
11 (1-2) (2009) 40–44.
cial interests or personal relationships that could have appeared [22] https://www.re-fer.eu/.
to influence the work reported in this paper. [23] Moslem Shahverdi, Julien Michels, Christoph Czaderski, Masoud Motavalli,
Iron-based shape memory alloy strips for strengthening RC members: material
behavior and characterization, Constr. Build. Mater. 173 (2018) 586–599.
Acknowledgements [24] Moslem Shahverdi, Christoph Czaderski, Masoud Motavalli, Iron-based shape
memory alloys for prestressed near-surface mounted strengthening of
reinforced concrete beams, Constr. Build. Mater. 112 (2016) 28–38.
The authors would like to express their thankfulness to the [25] C. Czaderski, M. Shahverdi, R. Brönnimann, C. Leinenbach, M. Motavalli,
Commission for Technology and Innovation of Switzerland, which Feasibility of iron-based shape memory alloy strips for prestressed
strengthening of concrete structures, Constr. Build. Mater. 56 (2014) 94–105.
supported this research financially (CTI Project No. 18528.1
[26] Parviz Soroushian, Ken Ostowari, Ali Nossoni, Habibur Chowdhury, Repair and
PFIW-IW). Further thanks go to the industry partner re-fer from strengthening of concrete structures through application of corrective
Seewen in Switzerland, who delivered all the memory-steel and posttensioning forces with shape memory alloys, Transp. Res. Rec. 1770 (1)
also supported this research financially. Lastly, the financial and (2001) 20–26.
[27] Benito Mas, Antoni Cladera, Carlos Ribas, Experimental study on concrete
technical support of the Empa Structural Engineering Research beams reinforced with pseudoelastic Ni-Ti continuous rectangular spiral
Laboratory is greatly appreciated. reinforcement failing in shear, Eng. Struct. 127 (2016) 759–768.
[28] Joan M. Rius, Antoni Cladera, Carlos Ribas, Benito Mas, Shear strengthening of
reinforced concrete beams using shape memory alloys, Constr. Build. Mater.
References 200 (2019) 420–435.
[29] Luis A. Montoya-Coronado, Joaquín G. Ruiz-Pinilla, Carlos Ribas, Antoni
Cladera, Experimental study on shear strengthening of shear critical RC
[1] ASCE. https://www.infrastructurereportcard.org/cat-item/bridges/. 2017.
beams using iron-based shape memory alloy strips, Eng. Struct. 200 (2019)
[2] SIA166, Klebebewehrungen (Externally bonded reinforcement), 2004, Zurich,
109680, https://doi.org/10.1016/j.engstruct.2019.109680.
Switzerland: Schweizerischer Ingenieur- und Architektenverein, SIA. pages 44.

15
C. Czaderski, M. Shahverdi and J. Michels Construction and Building Materials 274 (2021) 121793

[30] Antoni Cladera, Luis A. Montoya-Coronado, Joaquín G. Ruiz-Pinilla, Carlos reinforcement: RIBAS GONZÁLEZ AND FERNÁNDEZ RUIZ, Struct Concrete 18
Ribas, Shear strengthening of slender reinforced concrete T-shaped beams (5) (2017) 720–732.
using iron-based shape memory alloy strips, Eng. Struct. 221 (2020) 111018, [33] M. Shahverdi, C. Czaderski, J. Michels, ‘memory steel’ for Shear Reinforcement
https://doi.org/10.1016/j.engstruct.2020.111018. of Concrete Structures, SMAR 2019, 2019.
[31] L. Zerbe, M. Reda, M. Dawood, A. Belarbi, A. Senouci, B. Gencturk, M. Al-Ansari, [34] Nicola Gehri, Jaime Mata-Falcón, Walter Kaufmann, Automated crack
J. Michels, Behavior of Retrofitted Concrete Members Using Iron-Based Shape detection and measurement based on digital image correlation, Constr.
Memory Alloys, in: SMAR 2017, Zurich, Switzerland, 13-15 September 2017, Build. Mater. 256 (2020) 119383, https://doi.org/10.1016/
2017. j.conbuildmat.2020.119383.
[32] Carlos Rodrigo Ribas González, Miguel Fernández Ruiz, Influence of flanges on
the shear-carrying capacity of reinforced concrete beams without web

16