ArticlePDF Available

High-Speed and High-Efficiency Shape Memory Alloy Actuation


Abstract and Figures

When standard voltage levels commonly adopted in industry are used to activate thermal shape memory alloy (SMA) wire actuators, they often result in very high electrical currents which may eventually damage or destroy the actuators. To improve performance of SMA wire actuators operating in industrial environments, in this paper we investigate a novel, fast and energy-efficient actuation strategy based on short pulses in the millisecond range. The use of higher voltages leads to a highly dynamic activation process, in contrast to commonly used quasi-static activation based on low voltage. A test setup is designed to examine the effects of the control parameters (i.e., supply voltage, activation pulse duration, SMA wire pre-tension and wire diameter) on the measured displacement and force output of the SMA wire. It is shown that actuation times in the range of 20 ms and strokes of more than 10 % of the SMA wire length can be reached. Additionally, energy savings of up to 80 % with respect to conventional quasi-static actuation are achieved. Possible applications for this activation method are release mechanisms, switches or safety applications.
Content may be subject to copyright.
High-Speed and High-Efficiency Shape Memory Alloy
Paul Motzki, Tom Gorges, Mirco Kappel, Marvin Schmidt, Gianluca Rizzello, Stefan Seelecke
Center for Mechatronics and Automation Technologies (ZeMA) gGmbH, Saarbruecken, Germany
Department of Systems Engineering, Department of Materials Science and Engineering, Saarland
University, Saarbruecken, Germany
When standard voltage levels commonly adopted in industry are used to activate thermal shape
memory alloy (SMA) wire actuators, they often result in very high electrical currents which may
eventually damage or destroy the actuators. To improve performance of SMA wire actuators operating
in industrial environments, in this paper we investigate a novel, fast and energy-efficient actuation
strategy based on short pulses in the millisecond range. The use of higher voltages leads to a highly
dynamic activation process, in contrast to commonly used quasi-static activation based on low voltage.
A test setup is designed to examine the effects of the control parameters (i.e., supply voltage, activation
pulse duration, SMA wire pre-tension and wire diameter) on the measured displacement and force output
of the SMA wire. It is shown that actuation times in the range of 20 ms and strokes of more than 10 %
of the SMA wire length can be reached. Additionally, energy savings of up to 80 % with respect to
conventional quasi-static actuation are achieved. Possible applications for this activation method are
release mechanisms, switches or safety applications.
Keywords: Shape Memory Alloy; SMA; Actuator; High-speed; High-voltage; Energy efficiency
1. Introduction
Shape memory alloys (SMAs) are known for their high
energy density which allows for the construction of
compact and light-weight actuator systems [1][7].
NiTi-based wires represent a common SMA component
used to develop actuator systems. When heated to their
specific transformation temperature, these wires can
contract due to a phase transformation from martensite
to austenite. To generate a stroke, an SMA wire has to
be stretched in its martensitic state prior to the
activation, e.g., with a pre-tensioned spring. Typical
transformation temperatures for commercially available
SMA wires range from 70 °C to 90 °C [8], [9]. In most
applications, an SMA wire is controlled by supplying an
electrical current, which results into a Joule heating that
increases the wire temperature. SMA wire
manufacturers list specific activation currents for each
wire diameter in their data sheets [8], [9]. These current
values lead to voltages in the order of 1-10 V and may
be continuously applied without overheating the wire.
The use of those control values, however, commonly
results into a slow activation of the SMA wire (order of
seconds), which potentially limits the applicability of
the technology [10][13]. It is remarked how the
electrical resistance, and consequently the desired
supplied voltage and power, depend on the geometry of
the SMA actuator, i.e., wire diameter and length, as well
as phase composition and temperature [14], [15]. All of
those quantities change during actuation, therefore the
supplied current is non-constant for a constant applied
voltage. This implies a need for application-specific
power electronics to ensure correct current values for
SMA activation, which may severely complicate the
overall system design. In addition, many researchers
investigated different driving strategies such as pulse-
width modulated controllers, or PI and PID controllers
for position control via resistance feedback [16][24].
This paper presents a study of a different activation
approach for SMA wire actuators. The total electrical
energy that is needed to heat up the wire to
transformation temperature can be controlled not only
by voltage level but also by the length of the activation
pulse. So instead of using power electronics to
transform the given supply voltage to a desired level,
the duration of the activation pulse is adjusted. That
way, the activation of the SMA wire takes place at a
high electrical power for a short time interval. First
studies on SMA pulse activation were conducted by
Vollach, Shilo and coworkers [25][28]. In their
investigation, kilovolt pulses of the duration of
microseconds were used to activate SMA material,
while acquiring at the same time frequency response,
displacement and SMA force. However, the presented
studies were conducted mainly with a material science
perspective, since the focus of the work was the kinetic
of the Nickel-Titatium phase transformation. In
addition, the adopted voltage of the order of kV is
difficult to be reproduced in real-life applications, e.g.,
industrial ones. It is pointed out how the use of such a
control strategy is still relatively unexplored for
application-oriented SMA devices which operate in
industrial environments, and thus subject to constraints
in terms of maximum voltage levels.
The goal of the study presented in this paper is the
exploration of potential application scenarios for pulse-
controlled SMA, compatibly with voltage levels
commonly used in industrial environments. Typical
supply voltages used in real-life applications range from
12 V to 48 V in the automotive sector, 24 V for
industrial bus systems, and up to 110-125 V in
aerospace industry. These values are significantly
higher than the ones commonly used to drive SMA
wires (on the order of 1-10 V), and therefore they lead
to higher currents than recommended by datasheets. At
the same time, these values are significantly lower than
the ones investigated in previous studies on pulse
driving (on the order of kV). In this paper, it is shown
that it is possible to use these given supply voltages to
generate pulse activation in the millisecond range,
achieving fast and energy-efficient SMA actuation with
a compact and lightweight power electronics solutions.
In conventional quasi-static operation, one typically
focuses on stroke as a direct consequence of the low-
voltage activation. On the other hand, the high-voltage
activation of SMA wires, and the related phase
transformation in the material, result in an instantaneous
generation of force which leads to a highly dynamic
stroke as a consequence of accelerations and velocities.
In addition to simplified power electronics and faster
actuation time, a further advantage of pulse actuation is
energy-efficiency. In current state-of-the-art actuators,
the SMA wire activation takes place under non-
adiabatic conditions, i.e., a large amount of the heating
energy is lost due to heat exchange with the
environment during the activation process. With high
electrical power and a fast heating pulse, it is possible
to reach the transformation temperature and start
actuation before heat is lost to the environment, thus
leading to an adiabatic activation. A first preliminary
study on this topic has been presented in conference
paper [29]. This paper extends the previous work by
Detailed description of the experimental setup and
the test procedure
Comparison between conventional and high-speed
Sequential high-speed optical camera recordings to
illustrate “ballistic” behavior
Systematic experiments to investigate tendencies in
SMA behavior with focus on typical industrial
standard voltage supply:
o Variation of supply voltage
o Variation of counterforce
o Variation of SMA wire diameter
o Variation of activation pulse length
Theoretical energy analysis
Experimental investigation of energy efficiency in
relation to the supply voltage for different SMA
wire diameters
The remainder of the paper is structured as follows.
Section 2 focuses on the description of the experimental
setup, as well as the test procedure. Section 3 presents
the experimental results, first introducing the effects of
high-speed activation. Afterwards, the systematic
measurement results are introduced and discussed, and
an energetic analysis is performed. Section 4 provides
concluding remarks and future directions of the ongoing
2. Experimental Setup
The block diagram of the experimental setup is shown
in Fig. 1 (upper part). A power supply is used to heat the
SMA wire with an electric pulse. A laser displacement
sensor, model Keyence LK-G87, measures the SMA
wire stroke. The time derivatives of the displacement
signal, i.e., velocity and acceleration, are calculated and
plotted in post-processing. SMA force is recorded by a
load cell of the type Futek LSB200. The voltage on the
SMA wire is measured via 4-wires sensing, while
electrical current measurement is realized via a current
clamp. Additionally, the whole experiment is recorded
with an optical high speed camera system Olympus i-
SPEED TR for a better interpretation of the
measurement results. Signal processing and data
acquisition during the experiment are realized via an NI
CompactRIO system. The mechanical structure of the
setup, including the positioning of the key components,
is shown in Fig. 1 (lower part, [29]). The SMA wire is
guided vertically and clamped at top and bottom. The
specifically designed clamps combine mechanical and
electrical connection, by connecting the SMA wire
directly to a small PCB part. The upper clamp is
attached to the load cell which is connected to the
aluminum frame structure. The bottom clamp is directly
connected to the air bearing shaft. The total weight of
the clamp including shaft is 22.5 g. A compression
spring can be placed between the bottom clamp and a
micro-adjustment stage, allowing to vary the SMA wire
pre-tension. In applications, the use of springs as a bias
mechanism for the SMA wire is a common design
solution [30]. Additionally, the spring is used to adjust
bias forces to bigger SMA wire diameters. The air
bearing guides the vertical motion with minimal
friction. The laser displacement sensor detects the
motion of the air bearing shaft. The SMA wire length in
its austenitic state in this setup is 225 mm.
Fig. 1: Block diagram (upper part) and mechanical structure (lower
part, [29]) of the experimental setup.
2.1 Test Procedure
The SMA wire used in these experiments is a SAES
Getters SmartFlex actuator wire with a transition
temperature As = 90 °C [8]. The SMA wire length is
measured in its austenitic state, in order to guarantee
reproducibility. At low temperature (martensitic) state,
the SMA wire shows a two-way effect. This means that,
after cooling down, the SMA wire shows a non-defined
elongation, even without external loads [2], [7]. This
elongation is not uniquely defined, and may vary from
sample to sample. To ensure that the SMA wire length
is always measured in the same crystallographic state, a
separate measuring setup is performed. In this setup, the
SMA wire is heated in a nearly load-free configuration
to reach complete phase transformation to austenite.
The desired length is measured and marked in this high
temperature state. The wire is then cut at the markers
and clamped in the experimental setup. The micro-
adjustment stage, in combination with the compression
spring and the load cell, allows the generation of a
desired pre-stress and pre-strain in the SMA wire. The
experiment is then started via LabView.
SMA wires show a training effect after early activation
cycles. This effect is visible in a change of its remnant
strain after cooling down and returning to its initial
state. A new SMA wire in the test setup needs a few
training cycles until the remnant strain stays constant.
Therefore, the force signals and the displacement
signals are compared to the initial values after each
experiment. The measurements are started, once the
force and displacement stays at constant values after
few activations. During the measurements, mechanical
stress in the SMA wire can reach critical levels which
may lead to plastic deformations and eventually damage
the wire. This aspect can be monitored by the force and
displacement values after each experiment. In case
critical stress is overcame, the SMA wire has to be
3. Experimental Results
Typical activation of a conventional SMA wire is
displayed in Fig. 2. An SMA wire with a diameter of
76 µm is activated according to the manufacturer’s data
sheet suggestion, i.e., a supply voltage of 8 V. The
resulting electrical current is of approximately 150 mA
for 1 s. The initial force is set to 0.28 N, which results
to a mechanical stress of 60 MPa in the SMA wire. The
bias spring has an elastic constant of 0.056 N/mm. The
graph shows the trigger signal (brown), electrical
current (black, [A]), force (purple, [N]), displacement
(green, [mm]), velocity (yellow, [10-1 m/s]) and
acceleration (cyan, [m/s2]) signals. After 1 s, a
maximum displacement of 8.33 mm is reached, which
corresponds to a stroke of 3.7 %. The maximum
mechanical stress in the SMA wire reaches 165 MPa.
This measurement will serve as a reference for the
energy analysis in the following experiments at higher
voltages, which will use identical parameters for initial
force and spring rate.
In all of the experiments performed with quasi-static
actuation mode, the displacement is measured as a
direct output. The force and displacement signals are
always in phase, implying that inertial effects are
negligible. From a mechanical point of view, the system
is described by the equilibrium of forces, including the
SMA force FSMA, the spring force Fspr and the weight
force Fweight:
   . (1)
Note that the inertia force is neglected in (1), due to the
quasi-static condition. After the supply voltage is turned
off, the SMA force and the displacement return to their
initial starting values. The displacement signal shows
the effects of the thermal cooling of the SMA wire.
The energy balance for this conventional activation is
described by:
 
    
 
. (2)
Fig. 2: Conventional SMA activation as reference measurement
showing a synchronized evolution of force and displacement in a
quasi-static condition.
The electrical energy input is calculated by integrating
the electrical power (P=U∙i, voltage U, current i) in the
time interval [t0, t1], where t0 is the initial time and t1 is
the time at which the transformation temperature Ttrans
is reached and the phase transformation is completed.
The total electric energy equals the sum of three
contributions, reported on the right-hand side of (2). The
first term corresponds to the energy required to heat the
SMA wire from room temperature Tenv to
transformation temperature, given by the product of
(Ttrans Tenv), specific heat capacity c and SMA wire
mass mSMA. The second term h∙mSMA describes the latent
heat of the phase transformation. Finally, the third term
accounts for the heat loss due to thermal exchange
between the SMA wire and the environment, where α is
the convective cooling coefficient, and ASMA is the
surface area of the SMA wire. Note that both i and TSMA
are assumed to be non-constant during the time interval
[t0, t1].
In contrast to the conventional activation, Fig. 3 shows
experimental results of SMA activation at higher
voltages. In these experiments, the same SMA wire with
a diameter of 76 µm is activated at 24 V
(Fig. 3, upper part), 48 V (Fig. 3, center part) and 125 V
(Fig. 3, lower part).
Fig. 3: High-speed SMA activation at 24 V (upper part), 48 V (center
part) and 125 V (lower part).
The first qualitative difference observed is the
oscillation of the spring-mass-system before it returns
to its initial starting values. The force and displacement
signals are not synchronous anymore. In this dynamic
system, the displacement is the result of the acceleration
caused by the SMA force. Mechanically, the system
dynamics have to be included by adding the inertia force
Fi to (1):
    (3)
For a discussion of the quantitative changes, expanded
views of the three measurements in Fig. 3 are shown in
Fig. 4. A comparison of the initial current values at the
three different supply voltages shows that there is a
variation of the measured resistance. This results from
the manual clamping of the SMA wires at both ends.
The clamping forces, and thus the contact and transition
resistance, can vary after mounting a new SMA wire.
Therefore, all the measurement series concerning
energy consumption are performed with the same SMA
wire. The current signal in the measurement at 24 V
(Fig. 4, upper part) is similar to the conventional current
signal (shown in Fig. 2). First, the current slightly
decreases due to the rising temperature and thus rising
electrical resistance in the SMA wire. Once the SMA
wire starts contracting, the global geometric changes, in
particular the shorter SMA wire length and the larger
diameter, lead to a decrease of the electrical resistance
and the electrical current starts to increase. The
measurements at 48 V (Fig. 4, center part) and 125 V
(Fig. 4, lower part) also show the decrease in the current
upon activation. The higher supply voltages even result
in a faster temperature increase and thus in a more
significant decrease in the current signals. Because of
the shorter pulse length in the measurements at 48 V and
125 V, the decrease of the electrical resistance due to
the contraction of the SMA wire cannot be observed in
the corresponding current signals. The contraction of
the SMA wire is only minimal when the electrical
current is turned off. The displacement signals reach
higher maximum values in comparison to the
conventional activation, but the force signals are always
at 0 N when maximum displacement is reached. At
these points during activation, the SMA wire is slack so
that the load cell is unable to detect any force. This
means that maximum displacement is not reached
through conventional phase transformation, but due to
the extreme acceleration of the bottom mass. This effect
will be referred to as “ballistic” activation. For these
three measurements, velocities and accelerations are up
to two orders of magnitudes higher with respect to the
conventional actuation. The force signals at 48 V and
125 V show an instantaneous response to the current
signal, and reach their maximum values with the first
peak. The behavior of the force at 24 V is slightly
different, as the maximum force is reached with the
second peak. The displacement at the second force peak
is almost 5 mm, which is much higher in comparison to
the two measurements at higher voltages. This
displacement is an indicator for the actual SMA wire
contraction, i.e., the measurement at 24 V shows the
transition from quasi-static to highly dynamic actuation.
In this particular case, dynamic effects are barely
visible, but do not dominate the response.
Fig. 4: Zoomed-in measurements of SMA high-speed activations at
24 V (upper part), 48 V (center part) and 125 V (lower part).
The input energy for all three measurements in Fig. 3 is
about 0.4 J, which is much lower than the input energy
of 1.2 J for the conventional activation. In conventional
activation, energy is lost to the environment during
actuation, which is described by the last term on the
right-hand side in (2). Very fast actuation times lead to
very small time intervals [t0, t1] and the heat exchange
term becomes negligible. The energy balance (2) for
this adiabatic activation of SMA wires is reduced to:
    . (4)
A detailed discussion of the energy-efficiency follows
in section 3.2, which will show that all the following
experiments with supply voltages of 48 V or higher can
be considered to be occurring under adiabatic
Fig. 5 shows a measurement with synchronized high-
speed camera recordings to visualize the ballistic effect.
The rapid heating pulse and the resulting phase
transformation in the SMA wire accelerate the attached
mass to a velocity that causes the SMA wire to become
slack. The moving mass can result in displacement
values that cannot be reached by conventional phase
transformation. In this experiment, a 76 µm SMA wire
is activated by a 2 ms pulse at 125 V. The maximum
displacement of 21 mm implies a stroke of 9.3 %. The
force in the SMA wire is present immediately and
reaches its maximum of 4 N when the current is turned
off at 2 ms. This force is equivalent to a mechanical
stress of 880 MPa. After that, the force decreases as the
SMA wire starts contracting and the mass is accelerated.
The acceleration results in high velocities of the moving
mass and the SMA wire becomes slack, as shown in the
optical camera screenshots. At these points, the force
signal is at 0 N. The excited spring-mass system
continues then to oscillate until it reaches its initial state.
The high frequency oscillation in the force signal is due
to the natural frequency of the load cell. Maximum
displacement is reached 38 ms after the electrical
current is turned off. Interestingly, the acceleration
signal also reaches its maximum with a delay of about
15 ms after activation, and after the maximum force
peak. That means that the SMA wire is accelerated for
a certain time interval even though there is no external
energy input. This effect can be explained by an
increase of the austenite transformation temperature
under increasing mechanical stress, which leads to the
SMA wire acting like an energy reservoir. In fact,
transformation temperature typically increases with
mechanical stress by a factor of 0.11-0.14 K/MPa [5].
In the presented measurement (Fig. 5) the mechanical
stress reaches 880 MPa immediately after activation.
The austenite start temperature of 90 °C is defined at a
mechanical stress of 200 MPa [8]. Thus, at 4 N the
austenite start temperature increases up to 185 °C.
Fig. 5: Synchronized high-speed recording of a “ballistic” activation of an SMA wire.
The SMA wire heats up to high temperatures before the
force starts to decrease as the SMA wire starts
contracting. As the force decreases, the transformation
temperature also drops again. The SMA wire’s
temperature is now higher than the transformation
temperature, which leads to more energy in the hot wire
being used for the phase transformation, even though
there is no external energy input. After the force signal
reaches 0 N, the acceleration starts decreasing, because
the mechanical stress in the SMA wire and thus the
transformation temperature cannot decrease any further.
During the experiment, the measured force is not in a
linear relationship with the acceleration of the moving
mass according to    . The load cell measures
the resulting forces at the top clamp of the SMA wire,
while the bottom clamp represents the accelerated mass.
In addition to the initial force resulting from the mass of
the bottom clamp, the pre-loaded spring and the phase
transformation, the load cell also records so-called force
vibrations as soon as the system is in motion [26]. The
last screenshot in the sequence of high-speed recordings
(Fig. 5) is taken at 0.1545 s and indicates a slack wire.
However, the load cell records a force of about 1.5 N at
this time, which implies the presence of an additional
force. Vollach and Shilo describe these force vibrations
as a genuine mechanical response of the SMA wire due
to waves that propagate along the SMA wire.
3.1 Systematic Measurements
In the experimental analysis, SMA wire diameters of
50 µm, 76 µm and 100 µm are evaluated at three
different supply voltages and working against two
different loads. The goal is to study the effects of these
parameters on some performance indices which are
relevant for actuator applications. The maximum force,
the maximum displacement and the time delay to reach
the maximum displacement are extracted from the
measurements at 24 V, 48 V and 125 V supply voltage
and graphically presented to demonstrate experimental
trends. At each supply voltage, five different activation
pulse lengths (pulse widths) are applied, which results
in a stepwise increasing energy input. The stiffness of
the compression spring working against the SMA wire
equals 0.056 N/mm. Each measurement is performed
with initial loads of 0.28 N (weight plus preloaded
spring) and 0.67 N (same weight, increased spring
preload). For a clear presentation of the measured data,
the experimental results of the 76 µm wire at 48 V
working against the lower load of 0.28 N is used as a
reference. The variations of the initial load, the SMA
wire diameter and the supply voltage are then displayed
in separate graphs, always in comparison to this
reference measurement.
In Fig. 6, the experimental results of the reference
measurement with the 76 µm wire are displayed. At the
supply voltage of 48 V, activation pulses of 2 ms, 4 ms,
6 ms, 8 ms and 10 ms are used. With increasing
activation pulse time, the generated SMA force
increases (Fig. 6, upper part), because the total energy
input is higher and more SMA material is transformed
from martensite to austenite. Because of the higher
forces, the weight is accelerated more. This leads to
higher total displacements of the moving mass
(Fig. 6, center part). Note that the maximum
displacement obtained in most of these measurements
does not correspond to the actual SMA wire contraction,
but results from the high acceleration of the mass. The
ballistic effect displayed in Fig. 5 occurs every time,
when the force signal drops to 0 N. The time to reach
the maximum stroke also increases, because the total
displacement is much higher (Fig. 6, lower part) but the
average speed of the moving mass is much higher for
longer activation pulses.
The effect of different initial loads is displayed in Fig. 7.
At 48 V supply voltage, the 76 µm wire’s preload is
increased from 0.28 N to 0.67 N by increasing the
spring compression. The higher preload in the SMA
wires leads to slightly higher maximum forces, but the
difference becomes negligible (Fig. 7, upper part).
Fig. 6: Measurement of the 76 µm SMA wire at 48 V, recording
maximum force (upper part), maximum stroke (center part) and time
delay to reach maximum stroke (lower part) for five different
activation pulse lengths. The initial load is 0.28 N.
In case of activation with a 10 ms pulse, the difference
in maximum force is only 0.15 N. The high force peak
right after activation results from the phase
transformation. The SMA wire wants to contract
immediately, but is unable to because of the load’s
inertia. A higher initial load results in a higher
transformation temperature, that means more energy is
used to heat up the wire and less energy is used for the
phase transformation. Less phase transformation means
that less force is added to the initial load in the first peak
of the force signal. The higher initial load leads to lower
maximum strokes (Fig. 7, center part). Because of the
higher initial spring load but constant mass, the negative
acceleration is higher, which leads to less total
displacement and shorter times to reach the maximum
stroke (Fig. 7, lower part).
Fig. 7: Comparison of maximum force (upper part), maximum stroke
(center part) and time delay to reach maximum stroke (lower part)
for two different loads (0.28 N, 0.67 N). The experiments are run at
48 V supply voltage using a 76 µm SMA wire.
The next experimental results show the effect of the
SMA wire diameter (Fig. 8). A bigger SMA wire
diameter results in a lower electrical resistance, which
leads to a higher power during activation at a constant
voltage. Also the total energy input at defined activation
pulse lengths is higher for bigger diameters. This results
in more energy being used for phase transformation and
thus higher SMA forces generated (Fig. 8, upper part).
For small energy inputs at short pulses (2 ms) the
difference in force is not as large because the major
portion of the energy is used to heat up the wire. The
heating of the different wire diameters at constant
voltage is very similar for these short pulses because the
electrical current density is the same (and no energy is
lost to the environment). The higher forces in the bigger
wire diameters lead to higher strokes (Fig. 8, center
part), because of the higher energy input and stronger
acceleration of the weight. The stronger acceleration
also leads to higher speeds of the mass and thus shorter
times to reach the maximum stroke (Fig. 8, lower part).
Fig. 8: Comparison of maximum force (upper part), maximum stroke
(center part) and time delay to reach maximum stroke (lower part)
for three different SMA wire diameters (50 µm, 76 µm, 100 µm). The
experiments are run at 48 V supply voltage and an initial load of
0.28 N.
When activated with a 2 ms pulse, all three SMA wire
reach the same maximum displacement, but the 100 µm
wire is 10 ms faster than the 50 µm wire. When
activated with a 10 ms pulse, all three wires reach their
maximum displacement after about 45 ms, but the
100 µm wire shows a stroke of 23.9 mm whereas the
50 µm wire only reaches 9.2 mm.
Additionally, a comparison of the time delays to reach
the reference displacement of 8.33 mm is presented in
Fig. 9. This is of particular interest for possible
application scenarios. For this comparison, the relevant
experiments at 48 V for all three SMA wire diameters
are selected. The reference stroke of 8.33 mm is reached
by all three wires with 10 ms pulses and by the two
bigger diameters at 8 ms pulses. The 100 µm SMA wire
reaches the reference stroke in less than 23 ms.
Fig. 9: Time delays to reach the reference stroke of 8.33 mm for
different SMA wire diameters activated with an 8 ms and a 10 ms
pulse at 48 v supply voltage.
The last parameter study considers three different
supply voltages (Fig. 10). At different voltage levels,
the activation pulse lengths have to be adjusted so the
energy input stays similar. Therefore, this study
investigates maximum force (Fig. 10, upper part),
maximum stroke (Fig. 10, center part) and time to reach
the maximum stroke (Fig. 10, lower part) in relation to
the input energy. Again, the 76 µm wire is used in these
experiments with an initial load of 0.28 N. Higher
voltages result in higher forces for the same input
energy. Higher voltage equals higher electrical power.
As previously explained, the force increases as long the
load’s inertia prevents the SMA wire from contracting.
The faster the total energy is injected into the SMA wire,
the more phase transformation takes place before the
weight starts moving. The maximum stroke is almost
identical at 48 V and 125 V. As shown in Fig. 4 and
Fig. 5, the maximum displacement is reached by the
“ballistic” acceleration of the mass. The highest point
can be related to a potential energy. If no energy is lost
during the experiment, the same input energy results in
the same output. At 24 V, the SMA wire is not able to
generate the same stroke, which means energy is lost
during activation. The activation pulses at 24 V are
much longer, which results in cooling of the wire during
the activation. The generated stroke of about 12.5 mm
for an energy input of 0.4-0.5 J is still the result of the
ballistic activation, but the maximum stroke does not
increase further. The activation pulse at 0.4 J is 40 ms
long. This time interval is long enough for the system to
reach a thermodynamic equilibrium, so that even longer
activation pulses do not result in higher forces and
higher output strokes. Again, the high acceleration of
the bigger wire diameter leads to faster activation times.
As before, also the time delays to reach the reference
stroke are compared for the different supply voltage
levels (Fig. 11). For an energy input of 0.4 J, the stroke
of 8.33 mm is reached at all three voltages. At 125 V the
reference displacement is reached in under 22 ms.
Fig. 10: Comparison of maximum force (upper part), maximum
stroke (center part) and time delay to reach maximum stroke (lower
part) for three different voltage supply levels (24 V, 48 V, 125 V).
The experiments are run with a 76 µm SMA wire and an initial load
of 0.28 N.
Fig. 11: Time delays to reach the reference stroke of 8.33 mm for
different supply voltages. The 76 µm SMA wire is always activated
with a total energy of 0.4 J.
Even at the moderate voltage of 24 V, Fig. 10 shows
only a small decrease in energy-efficiency and
performance for the activation compared to the higher
voltages. The following section takes a closer look at
the energy balance during high-speed SMA activation.
3.2 Energy-Efficiency
The minimum energy for a phase transformation is
reached under adiabatic conditions. This means that no
energy is lost through heat flow to the environment
during the process. This minimum energy can be
calculated with (4). The energy to heat a SMA wire
(76 µm) from room temperature to transformation
temperature (ΔT = 65 K) is calculated with the specific
heat capacity c = 500 J/(kg·K) and the mass
mSMA = 6.58 m. The term  in (4) describes the
latent heat of the phase transformation with h ≈ 20 J/g
[9]. The theoretical minimum energy for a full phase
transformation is calculated as Wmin = 0.345 J.
In the first exemplary experiment, which serves as the
reference, a 76 µm SMA wire is activated in the
conventional way using the suggested current of
150 mA in the datasheet [8]. The SMA wire is activated
with an activation pulse width of 1 s. The displacement
reaches the value of 8.33 mm, which corresponds to a
3.7 % stroke. The same wire is then activated at
increasing voltages. At each set voltage, the activation
pulse time (pulse width) is slowly increased until a
displacement of 8.33 mm is reached. In all experiments,
the initial force in the SMA wire is 0.28 N and a spring
rate of 0.056 N/mm is used. The results of these
experiments are shown in Tab. 1. The activation delay
describes the time interval from the start of activation
until the bottom clamp of the SMA wire reaches
maximum displacement of 8.33 mm.
The experimental results and observed tendencies are
presented in Fig. 12. The necessary pulse width to reach
the reference displacement decreases dramatically
between 8 V and 16 V, and stays almost constant after
24 V (a). The energy consumption shows a very similar
behavior (c), as it is directly linked to the pulse
width (d). Interestingly, also the activation delay
mimics this behavior (b). The reason for this is the
experimental procedure focusing on energy
consumption analysis. If no energy is lost because the
experiment runs under adiabatic conditions, the same
total energy amount is necessary to generate a constant
displacement output, which relates to a constant
potential energy. At higher electrical powers, the initial
force and thus the acceleration of the mass increases but
at the same time the pulse width, which describes the
acceleration interval, is decreased. This leads to a
constant average velocity and activation delay. If
shorter activation delays at the same displacement are
desired, the pulse width can be increased at high
electrical power. In this scenario, the displacement
could be limited by a hard stop in the construction.
Fig. 12: Graphical illustration of experimental results regarding activation pulse (a) activation speed (b) and energy consumption (c) in relation
to the supply voltage and energy consumption in relation to the activation pulse (d) for a 76 µm diameter SMA wire.
The energy consumption for pulse widths between 1 ms
and 30 ms is very similar. With a 77 ms activation pulse
at 16 V, the energy consumption starts to increase more
rapidly. The results of these experiments suggest that
activation pulse widths up to almost 30 ms result in
nearly adiabatic conditions. To verify these
observations from a theoretical point of view, an
estimate time interval for adiabatic conditions can be
calculated. The temperature of a body according to
Newton’s law of cooling is determined by
. (5)
The time constant in this thermal process is described
with the product of thermal resistance and heat capacity:
    (6)
With a heat capacity of       and
a thermal resistance of      
the time constant for the cooling process can be
calculated. The heat transfer coefficient for the
convective cooling process of the SMA wire at room
temperature can vary between 10-100 W/(K·m2).
Usually values close to 10 W/(K·m2) are chosen to
represent conditions with minimal air flow. The
parameter A describes the heat transfer surface area of
the SMA wire. With the time constant and (5) the time
interval for a near adiabatic phase transformation can
now be determined. Considering a 1 % cooling of the
wire, this time interval is calculated as
  . Using the parameters in this
experiment, time intervals between 6.2-61.6 ms are
theoretically possible. The observed value of 30 ms
relates to a heat transfer coefficient of 20 W/(K·m2),
which is highly reasonable for the considered setup. For
this SMA wire diameter, the experimental results for the
energy consumption during a high voltage SMA
actuation are even slightly lower (22.2 %) than the
calculated theoretical minimum (28.8 %). This
minimum was calculated assuming a complete phase
transformation, but the displacement in the experiments
relates to a stroke of only 3.7 %. For complete phase
transformation strokes of 4-5 % can be expected.
Additionally, the displacement in the experiments at
voltage values equals or larger than 16 V was reached
due to the ballistic effect, which means that even less
phase transformation has taken place. On the other
hand, the calculation of this energy minimum does not
take into account the increasing austenite start
temperature for the phase transformation under
mechanical stress. As stated above, the transformation
temperatures generally changes with a factor of 0.11-
0.14 K/MPa. Since the mechanical stress is not constant
during the actuation, the transformation temperature
will change constantly during the activation process. In
summary, the actual energy balance in these
experiments cannot be trivially calculated, but a
magnitude of possible exemplary energy savings can be
determined. In addition to the absolute energy value,
also the experimentally identified time interval for
adiabatic conditions matches the theoretical
The same measurement sequence is performed for the
SMA wire diameters of 50 µm and 100 µm. The total
energy consumption at the different supply voltages is
illustrated in Fig. 13. At low voltages, the thin wires
need more energy to reach the same stroke. Thin wires
have a higher surface-to-volume ratio and thus lose
more energy to the environment during non-adiabatic
activation. Between 12 V and 16 V the SMA wire start
to reach the adiabatic region and the energy
consumption settles at near constant values. At these
voltages, the thin wires need the least amount of energy.
The two components of energy have been introduced in
(4). First, the SMA has to be heated to the
transformation temperature ( ). After that,
energy is consumed by the latent heat of the phase
transformation ( In small diameter SMA wires is
less material to be heated up before the actual phase
transformation can begin.
The different energy levels for the SMA wire diameters
lead to almost identical pulse widths at set voltages
(Fig. 14).
Displacement [mm]
Pulse width [ms]
Activation delay [ms]
Energy [J]
Ratio [%]
Tab. 1: Comparison of activation speed and energy consumption at different activation pulse widths for a 76 µm SMA wire.
Fig. 13: Total energy consumption for three different SMA wire
Fig. 14: Comparison of the activation pulse widths in relation to the
supply voltage (upper part) with zoomed-in graph (center part) and
in relation to the energy consumption (lower part) for three different
SMA wire diameters.
The expanded diagram (Fig. 14, center part) shows that
the pulse widths are always slightly higher for thinner
SMA wires. In relation to the total energy consumption,
the pulse width of the thin wires rapidly increases once
the activation starts leaving adiabatic regions
(Fig. 14, lower part). In that case, the thicker SMA
wires quickly become more energy-efficient because of
their smaller surface-to-volume ratio.
Similarly to the pulse width, the activation speed is
almost identical for the different SMA wires, provided
that they are activated under adiabatic conditions (Fig.
15). The zoomed-in graph (Fig. 15, lower part) indicates
that the thicker SMA wires are slightly faster than the
smaller wire diameters, which is related to higher
acceleration forces in thicker SMA wires due to the
higher electrical power for activation.
Fig. 15: Comparison of the activation speed in relation to the
supply voltage (upper part) with zoomed-in graph (lower part).
4. Conclusions
This paper has presented an experimental investigation
on an alternative SMA wire actuation strategy based on
high voltage pulses. Instead of relying on complex and
cost-intensive power electronics for current control, the
novel control strategy can be performed with simplified
hardware capable of applying a short voltage impulse
for a short period of time. Additional advantages of the
proposed strategy are high actuation speeds and lower
energy consumption. An experimental setup for
investigating the SMA wire behavior has been
described, and systematic measurement results have
been discussed. The load and the pre-tension of the
SMA wire affect the overall actuator dynamics. In
addition, it is shown how high accelerations can lead to
ballistic effects, which are responsible for a large
actuation stroke. Once activated, the actuator system
starts oscillating, according to a mass-spring-damper
dynamics. The presented exemplary experiment
showed activation times in the range of 40 ms for
strokes of close to 10 %. With shorter activation pulses,
energy consumption during activation is minimized,
because no heat is lost to the environment. This
adiabatic heating is already reached at moderate
voltages around 24 V for the 76 µm wire. An analytical
method of estimating the minimum energy for the SMA
actuation and the maximum time interval for adiabatic
conditions has also been demonstrated and discussed.
Energy savings up to 80 % have been achieved.
Possible applications for this SMA activation could be
in the field of release mechanisms, safety applications
or switches, where only a single actuation or low
frequency activations are needed. As an example, an
SMA actuator could be used to release a spring loaded
system, in which dynamic effects like oscillations don’t
have any negative influence on the actuator
In a next step, the experimental setup will be improved
with an additional force measurement at the moving end
of the SMA wire for a better force-acceleration
correlation during the actuation. Also, the mentioned
effect of the varying contact resistance will be addressed
by an improved clamp design. Further research will
focus on the investigation of SMA actuation with high
AC voltage, specifically looking at mains voltages like
130 V in the United States and 230 V in Europe.
The authors would like to acknowledge Thomas Würtz
for the technical support in the building of the test setup,
specifically the electronic components and SAES
Getters for providing the SMA material.
[1] H. Janocha, Adaptronics and Smart Structures, 2.
Berlin Heidelberg: Springer Verlag, 2007.
[2] H. Janocha, Unkonventionelle Aktoren - Eine
Einführung, 2. München: Oldenburg Verlag, 2013.
[3] J. M. Hollerbach, I. W. Hunter, and J. A. Ballantyne,
“Comparative analysis of actuator technologies for
robotics,” Robot. Rev. 2, pp. 299342, 1992.
[4] H. Funakubo, Shape Memory Alloys. Amsterdam:
Gordon and Breach Science Publ., 1987.
[5] S. Langbein and A. Czechowicz,
Konstruktionspraxis Formgedächtnistechnik, vol. 1.
Wiesbaden: Springer Fachmedien, 2013.
[6] P. L. Reece, Progress in smart materials and
structures. New York: Nova Science Publishers,
Inc., 2007.
[7] J. Mohd Jani, M. Leary, A. Subic, and M. A.
Gibson, “A review of shape memory alloy research,
applications and opportunities,” Materials and
Design, vol. 56. pp. 10781113, 2014.
[8] SAES Getters, “SmartFlex Brochure,” 2017.
[Online]. Available:
tFlex Brochure_2.pdf. [Accessed: 01-Mar-2018].
[9] Dynalloy Inc., “Technical Characteristics of
Flexinol Actuator Wires,” 2017. [Online].
[Accessed: 01-Mar-2018].
[10] R. A. Russell and R. B. Gorbet, “Improving the
response of SMA actuators,” Robot. Autom. 1995.
Proceedings., 1995 IEEE Int. Conf., 1995.
[11] C. S. Loh, H. Yokoi, and T. Arai, “Natural Heat-
Sinking Control Method for High-Speed Actuation
of the SMA,” Int. J. Adv. Robot. Syst., vol. 3, no. 4,
p. 42, Dec. 2006.
[12] J. Qiu, J. Tani, D. Osanai, Y. Urushiyama, and D.
Lewinnek, “High-speed response of SMA
actuators,” Int. J. Appl. Electromagn. Mech., 2000.
[13] R. Featherstone and Y. H. Teh, “Improving the
speed of shape memory alloy actuators by faster
electrical heating,” Springer Tracts Adv. Robot.,
[14] S. J. Furst and S. Seelecke, “Modeling and
experimental characterization of the stress, strain,
and resistance of shape memory alloy actuator wires
with controlled power input,” J. Intell. Mater. Syst.
Struct., vol. 23, no. 11, pp. 12331247, May 2012.
[15] S. J. Furst, J. H. Crews, and S. Seelecke, “Numerical
and experimental analysis of inhomogeneities in
SMA wires induced by thermal boundary
conditions,” Contin. Mech. Thermodyn., 2012.
[16] V. Shankar, G. N. Dayananda, P. Senthil Kumar, M.
Subba Rao, and R. Balasubramaniam,
“Development of electronic actuation system for
shape-memory-alloy-based aerospace structures,” in
SPIE, Smart Materials, Structures, and Systems,
2003, vol. 5062, pp. 914921.
[17] N. Ma and G. Song, “Control of shape memory
alloy actuator using pulse width modulation,” Smart
Mater. Struct., vol. 12, no. 5, p. 712, 2003.
[18] Z. Guo, Y. Pan, L. B. Wee, and H. Yu, “Design and
control of a novel compliant differential shape
memory alloy actuator,” Sensors Actuators A Phys.,
vol. 225, pp. 7180, Apr. 2015.
[19] A. Ianagui and E. A. Tannuri, “A sliding mode
torque and position controller for an antagonistic
SMA actuator,” Mechatronics, vol. 30, pp. 126139,
Sep. 2015.
[20] N. T. Tai and K. K. Ahn, “A hysteresis functional
link artificial neural network for identification and
model predictive control of SMA actuator,” J.
Process Control, vol. 22, no. 4, pp. 766777, Apr.
[21] Y. M. Alsayed, A. A. Abouelsoud, and A. M. R.
Fath El Bab, “Hybrid sliding mode fuzzy logic-
based PI controller design and implementation of
shape memory alloy actuator,” in 2016 8th
International Congress on Ultra Modern
Telecommunications and Control Systems and
Workshops (ICUMT), 2016, pp. 1015.
[22] Z. Shi, T. Wang, D. Liu, C. Ma, and X. Yuan, “A
fuzzy PID-controlled SMA actuator for a two-DOF
joint,” Chinese J. Aeronaut., vol. 27, no. 2, pp. 453
460, Apr. 2014.
[23] T. Ross Lambert, A. Gurley, and D. Beale, “SMA
actuator material model with self-sensing and
sliding-mode control; experiment and multibody
dynamics model,” Smart Mater. Struct., vol. 26, no.
3, p. 35004, 2017.
[24] A. Gurley, T. Ross Lambert, D. Beale, and R.
Broughton, “Dual measurement self-sensing
technique of NiTi actuators for use in robust
control,” Smart Mater. Struct., vol. 26, no. 10, p.
105050, 2017.
[25] S. Vollach and D. Shilo, “The mechanical response
of shape memory alloys under a rapid heating
pulse,” Exp. Mech., vol. 50, no. 6, pp. 803811,
[26] S. Vollach, D. Shilo, and H. Shlagman, “Mechanical
Response of Shape Memory Alloys Under a Rapid
Heating Pulse - Part II,” Exp. Mech., vol. 56, no. 8,
pp. 14651475, 2016.
[27] S. Vollach, R. Caciularu, and D. Shilo, “Equilibrium
stress during the response of shape memory alloys to
an abrupt heat pulse,” Scr. Mater., vol. 141, pp. 50
53, Dec. 2017.
[28] Y. Malka and D. Shilo, “A fast and powerful release
mechanism based on pulse heating of shape memory
wires,” Smart Mater. Struct., vol. 26, no. 9, p.
95061, Sep. 2017.
[29] P. Motzki, T. Gorges, T. Würtz, and S. Seelecke,
“Experimental Investigation of High-Speed/High-
Voltage SMA Actuation,” in ASME 2017
Conference on Smart Materials, Adaptive Structures
and Intelligent Systems - SMASIS17, 2017, p.
[30] S. J. Furst, “Design, Fabrication, and Control
Methods for Exploiting the Multifunctional Sensing
and Actuation Capabilities of Shape Memory Alloy
Wires,” Saarland University, 2012.
... This result is confirmed by the study [32]-a 41% reduction in energy is reported when driving an intelligent soft composite actuator. The authors Crystals 2021, 11, 1120 3 of 20 of [33] developed a PWM control with increased voltage (12 V ÷ 125 V) and reduced pulse length (226 ms ÷ 0.94 ms) for an SMA wire activation, resulting in 2 to 5 times lower power consumption and 4 to 25 times shorter time to reach a prescribed deformation. The advantages of PWM modulation and the easy fabrication of control devices through the modern hardware base have led to its use for experimental technical devices [34] and incorporation in industrially manufactured products [35]. ...
... This result is confirmed by the study [32]-a 41% reduction in energy is reported when driving an intelligent soft composite actuator. The authors of [33] developed a PWM control with increased voltage (12 V ÷ 125 V) and reduced pulse length (226 ms ÷ 0.94 ms) for an SMA wire activation, resulting in 2 to 5 times lower power consumption and 4 to 25 times shorter time to reach a prescribed deformation. The advantages of PWM modulation and the easy fabrication of control devices through the modern hardware base have led to its use for experimental technical devices [34] and incorporation in industrially manufactured products [35]. ...
Full-text available
The paper presents a theoretical and experimental investigation of a thermo-mechanical model of an actuator composed of a shape memory alloy wire arranged in series with a bias spring. The developed mathematical model considers the dynamics of the actuator in the thermal and mechanical domains. The modelling accuracy is increased through the developed algorithm for modelling the minor and sub minor hystereses, thus removing the disadvantages of the classical model. The algorithm improves the accuracy, especially when using pulse-width modulation control, for which minor and sub minor hystereses are likely to occur. Experimental studies show that the system is very sensitive, and there are physical factors whose presence cannot be considered in the mathematical model. The experimental research has shown that setting constant values of the duty cycle is impossible to obtain a stable value of displacement and force. The comparison between the developed mathematical model results and the experimental results shows that the differences are acceptable. The improved modelling serves as a basis for designing such actuators and creating an improved automatic feedback control system to maintain a given displacement (force) or trajectory tracking.
... Based on their performance capabilities, SMA actuator applications can be classified into fast responding systems [12,13] and those that move very high loads [15]. This paper outlines a system which combines high forces with high-speed actuation. ...
Full-text available
Conference Paper
Today, Nickel-Titanium (NiTi)-based shape memory alloy (SMA) wires are already used in a wide range of commercial actuator applications. The combination of their unique properties, such as their high energy density as well as their biocompatibility provides a broad range of applications. The former in particular allows for designing light-weight actuator systems with high forces using small installation spaces. In systems like emergency brakes or switch disconnectors, which require high forces as well as high actuation speed, the high-power capability of NiTi actuators is exploited. The presented work discusses the development of a giant power catapult demonstrator, that combines the high-speed and high-force capability of SMA wires. To illustrate the vast force, speed, and power potential of SMA wires, a bowling ball is launched from its resting position into the air using SMA wire bundles. For demonstration purposes a target altitude for the bowling ball of 500 mm is chosen, requiring a theoretical actuation force of F ≅ 920 N. To be able to launch the bowling ball to the desired height, an instantaneous energy release with an overall peak-power of P ≅ 0.5 MW is required.
... Based on their performance capabilities, SMA actuator applications can be classified into fast responding systems [24] [25] [26] [27] [28] and those that move very high loads [29] [30]. The content of this paper deals with combining both topics in a descriptive way, resulting in a mechatronic system for technology demonstration purposes. ...
Full-text available
Nickel-Titanium (NiTi) based shape memory alloy (SMA) wires are already often used in industrial actuator applications. Their high energy density allows to build light-weight actuator systems with high forces using small installation spaces. Combined with the biocompatibility of NiTi, a huge field of applications can be covered by SMA actuated systems. In systems like emergency brakes or switch disconnectors, which require high forces as well as high actuation speed, the high-power capability of NiTi actuators is exploited. The presented work details the development and characterization of a giant power catapult demonstrator, that combines the high-speed and high-force capability of SMA wires. To illustrate the vast force, speed, and power potential of SMA wires, a bowling ball is launched from its resting position vertically into the air using SMA wires. For demonstration purposes a target altitude for the bowling ball of 500 mm is chosen. With the height and the overall accelerated mass given, an actuation force F ≅ 920 N is needed. The instantaneous energy release from the designed power source results in the targeted flight height and an overall peak power of P ≅ 0.5 MW.
... By taking advantage of self-sensing or including an external position sensor, a feedback control for the activation current can be included. By reducing the activation times with a high electrical heating power and getting close to the adiabatic state, the energy efficiency can easily be optimized [13]. Also the decoupling of the antagonistic SMA wires can reduce switching cycle times [14]. ...
Full-text available
Conference Paper
In this paper, the development process of a shape memory alloy (SMA) driven pinch valve is proposed. The features of SMA actuator wires like small installation space, small weight, and the high energy density, allow for designing compact systems with high force outputs. A functional prototype is presented, allowing for closing forces of up to 250 N and 3 mm of stroke, which enables pinching hoses up to a diameter of 5 mm. It is driven by four mechanically parallel 500 µm diameter wires to open the valve and two 300 µm diameter to close the valve and cut off the medium flow inside of the hose. This SMA driven pinch valve features a spring-loaded toggle mechanism. It leads to stable and energy free position holding in both, opened and closed state of the valve, compared to commonly used valves, which allow for only one energy free position. As the system can be built without using any ferromagnetic materials the valve is predestined to be used in Magnetic Resonance Imaging (MRI). Due to the higher energy efficiency and weight reduction in comparison to commonly used solenoid pinch valves, they can be a substitute in a variety of applications.
... spring coils) are the most common form of the existing SMA actuators. In addition to being simple to fabricate or assemble [11,15,16], the fundamental reason is that the wires can be easily activated by electrical heating [17]. However, the one-dimensional configuration is difficult to be used directly as functional structures, but instead, as discrete actuators or in combination with other structures [18,19]. ...
Full-text available
Leveraging advances in shape memory alloys (SMAs) and flexible thermoelectric devices (f-TEDs), this paper presents a structural and functional integrity composite sheet to address the inefficient and bulky activation of 2-D SMAs. A series of experimental tests were performed to reveal the generation, change, and transfer mechanisms of different types of electrically-induced heat in the f-TED, as well as the temperature-induced SMA phase transformation behaviors. The results show that the composite sheet exhibited good bidirectional thermal management capacity. The austenite deformation can be completed within 10s at an operating current of 2 A. The cooling recovery, in particular, performs much better than in conventional modes (the temperature declines exponentially with time). Finally, through two functional prototypes, a light switch and a flexible gripper, the application potential of the proposed composite was further experimentally demonstrated.
... In case of excessive loads, the fracture site could be protected from damage by an abrupt increase in implant stiffness. Composite materials and metals, such as shape memory alloys and dielectric elastomers have opened up a whole new world of opportunities to build active implants [21][22][23][24] . The ability of shape memory alloys such as nitinol to shorten could be employed to create an implant capable of varying its length, just as in nitinol cardiovascular implants and implants used for arthrodesis in foot surgery [25][26][27][28] . ...
Full-text available
Nonunion is a complication of long bone fractures that leads to disability, morbidity and high costs. Early detection is difficult and treatment through external stimulation and revision surgery is often a lengthy process. Therefore, alternative diagnostic and therapeutic options are currently being explored, including the use of external and internal sensors. Apart from monitoring fracture stiffness and displacement directly at the fracture site, it would be desirable if an implant could also vary its stiffness and apply an intervention to promote healing, if needed. This could be achieved either by a predetermined protocol, by remote control, or even by processing data and triggering the intervention itself (self-regulated ‘intelligent’ or ‘smart’ implant). So-called active or smart materials like shape memory alloys (SMA) have opened up opportunities to build active implants. For example, implants could stimulate fracture healing by active shortening and lengthening via SMA actuator wires; by emitting pulses, waves, or electromagnetic fields. However, it remains undefined which modes of application, forces, frequencies, force directions, time durations and periods, or other stimuli such implants should ideally deliver for the best result. The present paper reviews the literature on active implants and interventions for nonunion, discusses possible mechanisms of active implants and points out where further research and development are needed to build an active implant that applies the most ideal intervention. Statement of significance : Early detection of delays during fracture healing and timely intervention are difficult due to limitations of the current diagnostic strategies. New diagnostic options are under evaluation, including the use of external and internal sensors. In addition, it would be desirable if an implant could actively facilitate healing (‘Intelligent’ or ‘smart’ implant). Implants could stimulate fracture healing via active shortening and lengthening; by emitting pulses, waves, or electromagnetic fields. No such implants exist to date, but new composite materials and alloys have opened up opportunities to build such active implants, and several groups across the globe are currently working on their development. The present paper is the first review on this topic to date.
... 3,8 However, limiting aspects are the need for an active cooling solution to achieve a fast switching speed and the lack of a powerless halting state. 2 To circumvent these shortcomings TCPAs have been incorporated in low energy structures. 9 Nevertheless, in the aspect of a versatile biomimetic actuator, the limitations persist. DE actuators exhibit many advantageous properties, for instance, a high strain and strain rate, self-sensing capabilities, and halting in the activated state without further power consumption. ...
... When looking for alternative actuator technologies with the intention of replacing state-of-art drives in applications, SMA technology is oftentimes overlooked because they are known for their limited actuation frequency and feasible temperature range. While this holds true for the shape memory material itself, these drawbacks can be attacked and overcome by smart system design and control strategies [18], [19]. The need for fast switching frequencies as well as functionality in higher ...
Full-text available
This work presents the development of an innovative shape memory alloy (SMA) actuator principle, which allows high‐speed switching cycles through the decoupling of antagonistically arranged SMA wires. Being optimized for the use at high ambient temperatures up to 65 °C, a possible application area is the active venting of injection molds where it can be used to expel air, which is trapped during the injection mold process. The patented actuator principle is based on a decoupled agonist‐antagonist SMA‐spring‐system and allows a high‐speed closing movement by a compact and lightweight design. Another innovation compared to conventional antagonistic SMA actuator systems is the integrated fail‐safe mechanism, which guarantees a defined closed state in case of power failure. Subsequently, in the motivation the need for active venting valves for injection molding is first described. Secondly, the novel actuator principle is introduced, and the development of an electronics concept is discussed. Finally, the design process, assembly, and validation of two iterations of the actuator prototype are presented. The final prototype validation measurements showcase high performance by valve strokes of 1 mm within 100 ms at ambient temperature of 65 °C. This article is protected by copyright. All rights reserved.
... While research shows that Joule heating with high electrical currents can increase the heating rate, SMA-based structures still faces problems of overheating and damage to the actuator (Jani et al. 2014;Featherstone and Teh 2006;Song et al. 2016). In addition, low heat conduction efficiency is the most significant factor limiting bandwidth during the cooling process (Jani et al. 2014;Lee et al. 2018;Motzki et al. 2018). With the classical experiment-simulation design approaches restricting the shape of SMA-based smart structure to rods, plates, ribbons, springs and wires (Rodrigue et al. 2017), the lack of geometric design freedom limits the performance of SMA-based actuators in terms of response time. ...
Full-text available
We present a novel multiphysics and multimaterial computational design framework for shape-memory alloy-based smart structures. The proposed framework uses topology optimization to optimally distribute multiple material candidates within the design domain, and leverages a nonlinear phenomenological constitutive model for shape-memory alloys (SMAs), along with a coupled transient heat conduction model. In most practical scenarios, SMAs are activated by a nonuniform temperature field or a nonuniform stress field. This framework accurately captures the coupling between the phase transformation process and the evolution of the local temperature field. Thus, the resulting design framework is able to optimally tailor the two-way shape-memory effect and the superelasticity response of SMAs more precisely than previous algorithms that have relied on the assumption of a uniform temperature distribution. We present several case studies, including the design of a self-actuated bending beam and a gripper mechanism. The results show that the proposed framework can successfully produce SMA-based designs that exhibit targeted displacement trajectories and output forces. In addition, we present an example in which we enforce material-specific thermal constraints in a multimaterial design to enhance its thermal performance. In conclusion, the proposed framework provides a systematic computational approach to consider the nonlinear thermomechanical response of SMAs, thereby providing enhanced programmability of the SMA-based structure.
Full-text available
Herein, a new torsional device based on shape memory alloy elements and operating as torque limiter is presented. It consists of a set of pseudoelastic NiTi tapes working in series configuration. The tapes are assembled radially with respect to the rotational direction and work sequentially in beam-like bending mode. The trend of the resulting torque may span from pulsating to continuous by changing the number of tapes. Furthermore, the use of NiTi material enables a temperature-dependent response. Due to the particular mounting, the flag-shaped pseudoelastic mechanical behavior typical of NiTi is partially hidden; however, the use of pseudoelastic elements provides a stable cyclic response and enables the possibility to tune the output torque.
Full-text available
Shape memory alloy (SMA) actuator is a potential advanced component for servo-systems of aerospace vehicles and aircraft. This paper presents a joint with two degrees of freedom (DOF) and a mobility range close to ±60° when driven by SMA triple wires. The fuzzy proportional-integral-derivative (PID)-controlled actuator drive was designed using antagonistic SMA triple wires, and the resistance feedback signal made a closed loop. Experiments showed that, with the driving responding frequency increasing, the overstress became harder to be avoided at the position under the maximum friction force. Furthermore, the hysteresis gap between the heating and cooling paths of the strain-to-resistance curve expanded under this condition. A fuzzy logic control was considered as a solution, and the curves of the wires were then modeled by fitting polynomials so that the measured resistance was used directly to determine the control signal. Accurate control was demonstrated through the step response, and the experimental results showed that under the fuzzy PID-control program, the mean absolute error (MAE) of the rotation angle was about 3.147°. In addition, the investigation of the external interference to the system proved the controllable maximum output.
Conference Paper
The thermal shape memory effect describes the ability of a deformed material to return to its original shape when heated. This effect is found in shape memory alloys (SMAs) such as nickel-titanium (NiTi). SMA actuator wire is known for its high energy density and allows for the construction of compact systems. An additional advantage is the so-called “self-sensing” effect, which can be used for sensor tasks within an actuatorsensor-system. In most applications, a current is used to heat the SMA wires through joule heating. Usually a current between zero and four ampere is recommended by the SMA wire manufacturers depending on the wire diameter. Therefore, supply voltage is adjusted to the SMA wire’s electrical resistance to reach the recommended current. The focus of this work is to use supply voltages of magnitudes higher than the recommended supply voltages on SMA actuator wires. This actuation method has the advantage of being able to use industry standard voltage supplies for SMA actuators. Additionally, depending on the application, faster actuation and higher strokes can be achieved. The high voltage results in a high current in the SMA wire. To prevent the wire from being destroyed by the high current, short pulses in the micro- and millisecond range are used. As part of the presented work, a test setup has been constructed to examine the effects of the crucial parameters such as supply voltage amplitude, pulse duration, wire diameter and wire pre-tension. The monitored parameters in this setup are the wire displacement, wire current and force generated by the SMA wire. All sensors in this setup and their timing is validated through several experiments. Additionally, a highspeed optical camera system is used to record qualitative videos of the SMA wire’s behavior under there extreme conditions. This optical feedback is necessary to fully understand and interpret the measured force and displacement signals.
Using a shape memory alloy actuator as both an actuator and a sensor provides huge benefits in cost reduction and miniaturization of robotic devices. Despite much effort, reliable and robust self-sensing (using the actuator as a position sensor) had not been achieved for general temperature, loading, hysteresis path, and fatigue conditions. Prior research has sought to model the intricacies of the electrical resistivity changes within the NiTi material. However, for the models to be solvable, nearly every previous technique only models the actuator within very specific boundary conditions. Here, we measure both the voltage across the entire NiTi wire and of a fixed-length segment of it; these dual measurements allow direct calculation of the actuator length without a material model. We review previous self-sensing literature, illustrate the mechanism design that makes the new technique possible, and use the dual measurement technique to determine the length of a single straight wire actuator under controlled conditions. This robust measurement can be used for feedback control in unknown ambient and loading conditions.
Previous studies of the responses of shape memory alloys (SMA) under a rapid heat pulse revealed the existence of a critical plateau stress that determines the performances of high rate SMA actuators. In this letter, we investigate the effects of temperature and initial stress on the plateau stress. Calculations based on the integration of the Clausius-Clapeyron equation while considering the inhomogeneity of the transformation temperature were in good agreement with the measured data. Our results indicate that the plateau stress represents equilibrium conditions and that actuation performances can be increased significantly by increasing the initial stress.
This article presents a novel actuator and a new concept for a release mechanism that are especially useful in applications that require fast motion of large masses over long distances. The actuator is based on ultra-fast pulse heating of NiTi wires, which provide a unique combination of large work per volume, short response time and enhanced energy efficiency. The release mechanism utilizes the fast and powerful actuator to form conditions in which the latch (safety pin) moves faster than the deployed device. As a result, the contact between these two masses is disconnected and the resulting friction forces are decreased to approximately zero. The actuator and release mechanism address the two major drawbacks of conventional shape memory alloy (SMA) actuators: slow actuation time and low energy efficiency. Using a dedicated setup, the experimental results validate the disconnection between the masses and map the effects of several variables on the performance of the actuator and release mechanism. In particular, we map the energetic efficiency and find the optimal operating conditions for a successful release using a minimal amount of input energy. At the optimal conditions, the actuator response time and the consumed input energy are smaller by an order of magnitude with respect to performances of previous SMA-based release mechanisms with comparable requirements.
Shape memory alloys (SMA) can be used to create actuators that are simple, high strength, and inexpensive. These benefits come at the cost of low electrical efficiency, moderate lifetime, and complex mechanical behavior that makes them difficult to design into new applications and products. To improve the integration of SMA actuators - in particular thin SMA wires heated by passing electric current through them - into modern mechanical applications, we have created tools for modeling SMA mechanical and thermal behavior in dynamic systems and under feedback controls. Thermo-electro-mechanical constitutive models are implemented in a multibody dynamics software where they are easily applied to an actuator emplaced in a multibody dynamic system. Mechanical behavior is modeled with 1D constitutive equations. The material state determines the electrical resistivity of the material which drives ohmic heating, while thermal cooling is based on a heat transfer analysis of thin cylinders. These models contain states which are very difficult to measure experimentally (such as crystal phase fraction) and thus provide insight into the material behavior and design that experimental results cannot offer. This thermomechanical model is used in conjunction with sliding mode control - historically difficult to simulate in numerically integrated models - to develop a working ball-on-a-beam setup in which the ball position is controlled via current passed through an SMA wire and with application of an original self-sensing method. The constitutive model is developed in the multibody dynamics software MSC ADAMS and validated through the simulation of the same system.
We report on a new experimental system for exploring the thermo-mechanical response of shape memory wires under a rapid heating pulse. The new setup enables the force and displacement generated by the phase transforming wire to be measured with a μs time resolution. In addition, part of the tests incorporated high-speed infrared and visible light photography. The experimental system enables exploring several unique dynamic effects. In particular, stress levels of 1.6 GPa with negligible plastic deformation and elastic strain rates of 103 s−1 are observed. Force measurements reveal a dead time of ~20 μs between the end of the electric pulse and the onset of the stress rise in the SMA wire. The high strain rate, generated by the phase transformation, induces string-like vibrations of the SMA wire that result in strong stress vibrations. The developed experimental system opens the way for studying the kinetics of the phase transformation.