See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245367052

Design of Shape Memory Alloy Actuators for Morphing Laminar Wing With
Flexible Extrados

Article  in  Journal of Mechanical Design · September 2009

DOI: 10.1115/1.3160310

CITATIONS READS
65 222

5 authors, including:

Thomas Georges V. Brailovski

Pratt & Whitney Canada École de Technologie Supérieure
12 PUBLICATIONS   191 CITATIONS    244 PUBLICATIONS   2,866 CITATIONS   

SEE PROFILE SEE PROFILE

Daniel Coutu Patrick Terriault

Pratt & Whitney Canada École de Technologie Supérieure
17 PUBLICATIONS   323 CITATIONS    96 PUBLICATIONS   1,289 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Porous metallic structures for biomedical applications View project

Additive manufacturing View project

All content following this page was uploaded by V. Brailovski on 16 July 2018.

The user has requested enhancement of the downloaded file.

 Thomas Georges Design of Shape Memory Alloy
e-mail: thomas.georges.1@ens.etsmtl.ca

Vladimir Brailovski
Professor
e-mail: vbrailovski@mec.etsmtl.ca
Emeric Morellon
e-mail: emeric.morellon@gadz.org
Daniel Coutu
e-mail: daniel.coutu.4@ens.etsmtl.ca
Patrick Terriault
Professor
e-mail: patrick.terriault@etsmtl.ca
Department of Mechanical Engineering,
École de Technologie Supérieure (ETS),
1100 Notre-Dame Street West,
Montréal, QC H3C 1K3, Canada
Actuators for Morphing Laminar
Wing With Flexible Extrados
An active structure of a morphing wing designed for subsonic cruise flight conditions is
composed of three principal subsystems: (1) flexible extrados, (2) rigid intrados, and (3)
an actuator group located inside the wing box. The four-ply laminated composite flexible
extrados is powered by two individually controlled shape memory alloy (SMA) actuators.
Fulfilling the requirements imposed by the morphing wing application to the force-
displacement characteristics of the actuators, a novel design methodology to determine
the geometry of the SMA active elements and their adequate assembly conditions is
presented. This methodology uses the results of the constrained recovery testing of the
selected SMA. Using a prototype of the morphing laminar wing powered by SMA actua-
tors, the design approach proposed in this study is experimentally validated.
关DOI: 10.1115/1.3160310兴
Keywords: morphing wing, flexible extrados, SMA actuators, Ti–Ni shape memory
alloys, design

1 Introduction wing functions as follows: according to the changing flow condi-

Common requirements for actuators designed for robotic, auto-
motive, aeronautic, and spatial applications 关1,2兴 are small size
and weight, low cost, and integration compatibility. Compared
with traditional 共electrical or pneumatic兲 or advanced 共piezoelec-
tric, magnetostrictive, or electrostrictive兲 actuators, innovative ac-
tuators based on the shape memory alloy 共SMA兲 technology pos-
sess the highest power density 关3,4兴. Other advantages offered by
the SMA devices are their compliance with harsh environments,
the simplicity of their actuation mechanisms, their silent and
smooth motion, and their autosensing ability. SMA actuators are
thus interesting candidates for aeronautic applications, particularly
for airfoil profile control systems 关4–7兴.
This work is focused on the design of the actuator system for a
morphing wing with flexible extrados. The design arrangement of
the SMA-based actuator system will be presented and analyzed in
terms of governing equations. The stress-strain design diagram
will be built from a series of constrained recovery experiments
with the selected SMA. Taking the functional requirements for the
actuation system of the morphing wing and the recovery stress/
strain performances of the SMA into account, the design of the
actuators will be completed by calculating the cross section and
length of the SMA active elements and by dimensioning other
components of the power transmission system. Finally, the pro-
posed design configuration will be validated experimentally.
2 The Active Structure
The active structure of the morphing wing consists of a flexible
extrados connected to actuators via a transmission system, and it
is subjected to variable flow conditions 共aerodynamic forces兲. As
presented in Fig. 1, modification of the airfoil profile occurs when
actuators located inside the wing box apply individually con-
trolled displacements to the flexible extrados. The overall stiffness
and integrity of the experimental wing is provided by the rigid
intrados. The front and rear edges of the flexible extrados are
connected to the rigid intrados in such a way that profile tangency
continuity and accommodation to shape modifications are assured
and aerodynamic forces are sustained. The morphing laminar
Manuscript received December 1, 2008; final manuscript received May 27, 2009;
published online August 18, 2009. Review conducted by Nancy Johnson.
tions, a controller commands the actuators to modify the wing
shape to assure its conformity with a series of optimized profiles
stored in the control system database 共open-loop control兲 关8兴. The
2D WTEA supercritical airfoil 共Institute for Aerospace Research,
National Research Council of Canada兲 is selected as the reference
profile for the morphing laminar wing with a 500 mm chord and a
990.6 mm span. This WTEA airfoil has a maximum thickness/
chord ratio of 16% and is similar to a family of natural laminar
flow-capable, supercritical airfoils 关9兴. Measuring from the lead-
ing edge, the flexible part of the wing extrados is contained be-
tween 3% 共15 mm兲 and 70% 共350 mm兲 of the chord 共Fig. 1兲. The
control system database covers a flow condition range encompass-
ing 7 Mach numbers 共M = 0.2– 0.35 incremented by 0.025兲 and 7
angles of attack 共␣ = −1 – 2 deg incremented by 0.5 deg兲 for a total
of 49 optimized wing profiles, and thereby defining the target
morphing envelope for the active structure 关10,11兴.
Based on the results of a multicriteria optimization of the active
structure 关10兴, the flexible extrados is made from laminate com-
posite consisting of four carbon/Kevlar plies and epoxy matrix
and moved by two spanwise oriented actuation lines located at
25.3%c and 47.6%c, as shown in Fig. 1.
3 Design of the Actuator Group
3.1 Functional Requirements of the Actuators. The active
system of the morphing wing with two actuators is schematized in
Fig. 2. For the selected four-ply-two-actuator configuration, the
forces and displacements required from each actuator to reach the
target morphing profiles were directly obtained from the finite
element 共FE兲 structural analysis for each of the flow cases studied
共Figs. 3共a兲 and 3共b兲兲 关10兴. In Fig. 3共c兲, the working envelopes for
Actuators 1 and 2 are superposed and one actuation case corre-
sponding to the angle of attack ␣ = 0.5 deg and Mach number
M = 0.275 is highlighted as an example. The optimal profile, for
this case, is reached when Actuators 1 and 2 impose vertical dis-
placements of 6.57 mm and 3.34 mm to the flexible extrados. For
Actuator 1, a force of 1032 N acting against the aerodynamic
suction is necessary to keep the skin in its nominal position 共null
displacement兲. As the aerodynamic load 共suction兲 pulls the skin
upwards, the force in the actuation system can be reduced to 533
N by applying 499 N in the same direction as that of the aerody-

Journal of Mechanical Design Copyright © 2009 by ASME SEPTEMBER 2009, Vol. 131 / 091006-1

Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 1 Conceptual design of the morphing laminar wing

namic load to allow 6.57 mm displacement of the flexible extra-
dos. When Actuator 1 moves, Actuator 2 should generate an ad-
ditional 100 N 共from 900 N to 1000 N兲 to allow its displacement
of 3.34 mm.
The maximum and minimum actuation forces and displace-
ments obtained from Fig. 3 for Actuators 1 and 2 are displayed in
Table 1. It can be observed that the forces required from both
actuators have a positive sign, thus indicating that they are in
Fig. 2 Actuators and flexible extrados
“pull” 共or retaining兲 mode because the aerodynamic suction forces
prevail on the forces needed to deform the flexible extrados.
3.2 Conceptual Design of the Actuator Group. The actuator
group, illustrated schematically in Fig. 4 consists of four main
subsystems: the SMA active element, the transmission system 共in-
cluding the slider and crank mechanism兲, the flexible extrados,
and the compression bias spring 共for simplicity, only one actuator
is shown here兲. The SMA element is connected to the flexible
extrados through the transmission system. The bias spring is
placed in parallel with the SMA element. The use of the transmis-
sion system allows adjustment of the force-stroke actuator char-
acteristics to the force-displacement characteristics of the flexible
extrados 关11兴.
The compression bias spring pushes the slider to the left, which
corresponds to the nominal extrados position 共Fig. 4共a兲兲, while the
aerodynamic forces pull the flexible extrados upward to move the
slider to the right, which corresponds to the deployed extrados
position 共Fig. 4共b兲兲. The SMA active element can reach two ex-
treme states: 共a兲 cold 共“off”兲 and 共b兲 hot 共“on”兲. The cold state
corresponds to the nominal 共low兲 extrados position 共Fig. 4共a兲兲,
and the hot state to the morphed 共deployed兲 extrados position
共Fig. 4共b兲兲.

Fig. 3 Force-displacement characteristics of the four-ply-two-actuator active structure for 49 benchmark flow
cases: „a… Actuator 1, „b… Actuator 2, and „c… Actuators 1 and 2 †10‡

091006-2 / Vol. 131, SEPTEMBER 2009 Transactions of the ASME

Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 1 Force-displacement envelopes for Actuators 1 and 2

Actuator 1 Actuator 2
Parameters Min Max Min Max

Vertical extrados displacement, ␦ext 共mm兲 0 8.16 0 4.1

Aerodynamic force per 1 m of span, Faero 共N兲 510 1800 410 1420
Actuation force per 1 m of span, Fact 共N兲 15 1800 420 1760
Extrados stiffness coefficient, kext 共N/mm兲 43.8 110.6 ⫺28.6 613.4

3.2.1 Nominal Extrados Position (SMA is Off). In this con-
figuration, the SMA element is in its cold state 共off兲 and the com-
pression bias spring acts against the aerodynamic suction forces
while stretching the SMA active element. The extrados is in its
nominal position and does not apply any force. This situation is
illustrated schematically in Fig. 4共a兲.
Given the presence of the transmission system with transmis-
sion ratio R, the aerodynamic forces and the forces related to the
extrados reaction, as presented by the active element and the bias
spring, will be decreased by a factor of R, while the stroke pro-
vided by the active element will be increased by the same factor
R. Using the free-body diagram 共FBD兲 of Fig. 4共a兲, the reaction of
c
the cold-state SMA element FSMA can be determined:
Faero
c
FSMA = Fspr − 共1兲
R c
where Fspr is the bias spring force and Faero is the aerodynamic
force when the extrados is in its nominal position.
3.2.2 Deployed Extrados (SMA is On). In this configuration,
the SMA active element is heated to move the extrados from its
nominal to the deployed position and, therefore, it works against
both the bias spring and the flexible extrados but is assisted by the
aerodynamic forces 共Fig. 4共b兲兲. Using the FBD of Fig. 4共b兲, the
h
force to be provided by the SMA active element FSMA can be
calculated by the following equation:
共Fext − Faero兲
h
FSMA = 共Fspr + kspr␦兲 + 共2兲
R
共Fspr + kspr␦兲 is the bias spring reaction when the slider moves to
the right, which corresponds to the deployed extrados position,
kext is the stiffness coefficient of the bias spring, ␦ is the actuator
stroke, Faero and Fext are the aerodynamic force and the force
needed to move the extrados from its nominal to the deployed
position, respectively, and Fext = kext␦ext so that kext is the stiffness
coefficient of the flexible extrados and ␦ext = ␦ / R, which is the
flexible extrados displacement at the actuation point.
The actuator stroke from the nominal to the deployed extrados
position can be calculated as follows:
␦ = LSMA
c h
− LSMA
h
where LSMA and LSMA are the lengths of the SMA element in its
cold and hot states, calculated as follows:
c
LSMA = LSMA + ␦0 + ␦ and h
LSMA = LSMA + ␦0 共3兲
where LSMA is the length of the stress-free SMA element before its
installation in the system. During installation, this element is
stretched to apply an elongation, which is the sum of the hot state
elongation ␦0 and the actuator stroke ␦.
To summarize, the behavior of the active structure is the result
of the interaction between the active SMA element, the bias

Fig. 4 Schematic representation and simplified mechanical models of the actuator in „a… nominal
extrados position and „b… deployed extrados position

Journal of Mechanical Design SEPTEMBER 2009, Vol. 131 / 091006-3

Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 5 Experimental testing bench for SMA characterization
spring, and the flexible extrados submitted to variable aerody-
namic forces. If the flexural extrados stiffness, the pressure distri-
bution and the extrados displacements from the nominal to the
target position are known, the design of the active structure can
then be completed by selecting an appropriate SMA element and
bias spring.
3.3 SMA Active Element—Material Characterization and
Design Diagram. The material selected to manufacture the SMA
active elements is a 쏗1 mm Ti–50.26 at. % Ni wire supplied by
Special Metals Corp. 共NY兲 subjected to 30% cold work by cold
rolling 共CR兲 and postdeformation annealing 共PDA兲 at 300° C 共1
h兲. After cold rolling, the cross section of the SMA wire reduces to
0.7 mm2. To characterize this material, the experimental setup
shown in Fig. 5 was developed. SMA specimens 共1兲 共initial length
of 70 mm and cross section of 0.7 mm2兲 are thermally cycled
between ambient temperature 共25° C, approximately 5 ° C lower
than the martensite finish temperature兲 and 150° C 共50° C higher
than the austenite finish temperature to complete the martensite-
to-austenite transformation兲 by resistive heating. A power supply
共2兲 共Sorensen DML 40-15, Ametek Programmable Power Inc.,
CA兲 is used to inject electrical current into the specimen, and the
heating is controlled using the temperature signal as a feedback.
The temperature is measured using a K-thermocouple 共3兲 placed
directly on the surface of the specimen and electrically isolated
from it by a Wakefield thermal compound. The force generated by
the actuator is measured by the load cell 共4兲 共LC 101-500, Omega
Eng. Inc., CT兲 fixed on the linear guide 共5兲; the displacement by
the linear variable displacement transducer 共6兲 共Trans Tek series
240, Intertechnology Inc., ON, Canada兲; and the electrical resis-
tance by a Keithley 2010 multimeter 共7兲 共Keithley Instruments
Inc., OH兲. The data acquisition and the control are performed
using a PC programed for this application with LABVIEW software
共8兲 共LABVIEW 8.2, National Instrument Corp., TX兲.
Testing bench was designed in order to define a functional en-
velope for the SMA active element, thus constructing a design
diagram for the SMA active element to be used in the morphing
091006-4 / Vol. 131, SEPTEMBER 2009
Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
wing actuation system 关12兴. To fulfill this task, four testing modes
are used 共Fig. 6兲. Each testing was repeated 200 times in order to
stabilize the SMA behavior.
3.3.1 Constant-Stress Mode. A constant stress of 220 MPa ap-
plied to the SMA active element in its martenstic state determines
its initial testing strain of 3%. When thermomechanical cycling is
repeated, an evolution of the material behavior is observed. After
200 cycles, the SMA behavior stabilizes and its martensitic 共␧m兲,
austenitic 共␧a兲, and recovery 共␧rec = ␧m − ␧a兲 strains can be ex-
tracted 关12兴. The martensitic strain of 5.5% is then selected as a
starting point for the subsequent testing modes.
3.3.2 Fixed-Support Mode. The stiffness of the active element
support is close to infinity. In these conditions, the SMA can only
generate stresses, the displacement is null. The recovery stress
increment generated upon heating up to 150° C can be defined as
␴rec = ␴a − ␴m, where ␴m are martensitic and ␴a are austenitic
stresses.
3.3.3 Elastic Bias Mode. Representing an intermediate load-
ing case, the diagram shown here is associated with the stress-
strain behavior of an active SMA element connected to an elastic
bias spring.
3.3.4 Combined Constant-Stress–Fixed-Support Mode. The
previously described actuation system with limited nominal and
deployed extrados positions can also be tested using the testing
bench of Fig. 5. To this end, the specimen is heated first under
constant stress 共1-2, Fig. 6兲; but when a certain amount of strain
induced by this force is recovered, the specimen’s extremities are
blocked, while heating continues with an additional increase in
stress 共2-3, Fig. 6兲. Upon cooling, the sequence is reversed. First
the stress decreases without a change in the specimen length 共3-2,
Fig. 6兲 and then, if cooling continues, an additional elongation
occurs 共2-1, Fig. 6兲.
The results obtained from each of these testing modes after the
200 cycle stabilization allow the working space of a given active
Transactions of the ASME
 Fig. 6 SMA design diagram „stress-strain space…

element to be delimited as shown in Fig. 6. The “cold assembly should be noted that whereas the force to be provided by the
point” determines the SMA element’s equilibrium position in its actuator in the cold state should be determined using the maxi-
martensitic state and the “hot curve” determines the SMA el- mum value of the aerodynamic forces, the force to be provided in
ement’s equilibrium position in its austenitic state, for variable its hot state is determined using its minimum value. This approach
bias conditions. leads to a conservative design of the actuator because the SMA
To guarantee the functional stability of the SMA active el- element is designed under the most stringent loading conditions.
ement’s performances during repetitive recovery strain and stress
generations, the following limitations are imposed in the present c Faero
work: the strain varies between 3% and 5% and the stress varies FSMA = Fspr −
R
between 250 MPa and 500 MPa, as shown by the hatched area in
Fig. 7. The application of these constraints prevents loosening of
the SMA wires when the actuator cools down and their overload- 共Fext − Faero兲
h
FSMA = 共Fspr + kspr␦ext
max
兲+
ing and, therefore, premature degradation when it is heated up. R
3.4 Design Procedure. In our design case, Actuator 1 should
meet higher functional requirements than Actuator 2, and it is thus
c
LSMA − LSMA = ␦0 + ␦
a critical element of the application. The mandatory information
for the Actuator 1 design combines the force-stroke requirements ␦ = R␦ext
max
共4兲
from Table 1 as well as the allowable strains and stresses from
Fig. 7, as regrouped in Table 2. Since the SMA active element is a wire, its geometry is defined by
The complete actuator is described by a system in Eq. 共4兲. It the wire cross section 共S兲 and stress-free length 共LSMA兲, and the

Fig. 7 SMA design diagram in strain-stress space with delimited actuation

area

Journal of Mechanical Design SEPTEMBER 2009, Vol. 131 / 091006-5

Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 2 Design information for Actuator 1

Parameters Value

Maximum vertical extrados displacement, ␦ext

max
共nm兲 8.16
min
Minimum and maximum aerodynamic forces, Faero and
max
Faero 共N兲 510 and 1800
max
Maximum extrados stiffness, kext 共N/mm兲 110.6
Minimum and maximum SMA strain, ␧min and ␧max 共%兲 3 and 5
Minimum and maximum SMA stress, ␴min and ␴max
共MPa兲 250 and 500

maxima and minima recovery forces and displacements generated

by this element can be expressed as follows:
Fig. 8 Manufacturing setup of the flexible extrados †2‡
c
FSMA = ␴minS
h
FSMA = ␴maxS c
LSMA = LSMA + ␧maxLSMA = 1225 ⴱ 共1 + 0.05兲 = 1286.25 mm
h
LSMA − LSMA = ␧minLSMA or to 5% wire elongation.
The maximum actuation force corresponds to the actuator’s on
c
LSMA − LSMA = ␧maxLSMA 共deployed extrados兲 position and to the shortest length of the
SMA element
c
LSMA h
− LSMA = R␦ext
max
共5兲 h
LSMA c
= LSMA − 3␦ext
max
= 1286.25 − 3 ⴱ 8.16 = 1261.18 mm
By combining systems 共4兲 and 共5兲, one can write or to 3% wire elongation 共see the actuation trajectory in the stress-
Faero strain space of Fig. 7兲.
␴minS = Fspr −
R
4 Experimental Validation of the SMA Actuator De-
共kext ␦ext − Faero
max max min
兲 sign
␴maxS = 共Fspr + kspr␦ext
max
兲+ 共6兲
R
4.1 Morphing Wing Prototype. To fulfill the function of the
flexible extrados, the laminated composite structure used in the
共␧max − ␧min兲LSMA = R␦ext
max
present study is manufactured from the Huntsman 8602 reinfusion
Two equations in respect to the cross section and length of the epoxy resin and two types of plies, as shown in Fig. 8, a low-
SMA element can therefore be obtained as follows: modulus unidirectional carbon and a 2 ⫻ 2 twill woven hybrid
carbon/Kevlar. During vacuum-assisted resin transfer molding, the
共Faero
max min
− Faero 兲 + ␦ext 共kext + R2kspr兲
max max
S= laminate is formed to the reference shape of the WTEA upper
R共␴max − ␴min兲 profile using a machined aluminum mold. To minimize the lami-
共7兲 nate weight, all of the plies are oriented so that their higher elastic
R␦ext
max
moduli are aligned with the wing chord, leading to a typical
LSMA = 共0H / 0C / ¯兲 laminate stacking sequence. To ensure shape modifi-
␧max − ␧min
cation uniformity in the span direction without affecting flexural
The use of a compression gas spring 共GS兲 as a bias element is behavior in the chord plane, steel stiffeners are fixed to the inner
considered in this work because this device applies quasiconstant surface of the laminate, precisely where the actuators are con-
forces, thus decreasing the power requirements for SMA active nected to the structure.
elements and approaching the “d” testing mode of Fig. 6. If the To validate the proposed design methodology, the wing proto-
selected transmission ratio R equals 3 and the stiffness coefficient type with two actuators was tested under laboratory conditions
of the bias spring is negligible 共kspr ⬇ 0 N / mm兲, we can now 共Faero = 0兲 using the testing bench of Fig. 9. The SMA wires 共1兲,
calculate, using Eq. 共7兲 and the data from Table 2, that an SMA forming an active element, are connected to the rigid base-plate
wire with a cross section of 3.5 mm2 and a length of 1225 mm 共2兲 and to the mobile slider 共3兲. During actuation, the SMA wires
will be sufficient to accomplish the task and the forces to be pull the slider 共3兲 and the crank 共4兲, which is attached to the rail,
provided by the bias spring and the SMA element will be as fol- transfers the movement to the flexible extrados 共5兲. A crank
lows: mechanism 共6兲 connects the slider 共3兲 to the rigid base-plate 共2兲.
c
FSMA = ␴minS = 875 N 共8兲 The displacement of the slider 共3兲 is measured by the linear po-
tentiometer 共7兲 共LP804 series, Omega Eng. Inc., CT兲. The gas
max
Faero spring 共8兲 that assures the return of the morphing wing to its
Fspr = c
+ FSMA = 1475 N 共9兲 reference position is connected to the base-plate 共2兲 and the slider
3 共3兲 to counteract the SMA active element 共1兲. The prestrain of the
SMA wires is assured by the tensioner, composed of the threaded
共kext ␦ext − Faero
max max min
兲 rod connected to the gas spring 共8兲 and the rigid base-plate 共2兲.
h
FSMA = 共Fspr + 3kspr␦ext
max
兲+
3 This testing bench has two functioning modes: 共a兲 the manual
共10兲 mode to characterize the flexible skin and the transmission system
h
FSMA = 1600 N and 共b兲 the automatic mode to test the SMA-powered actuators.
The minimum actuation force thus calculated corresponds to the 4.1.1 Characterization of the Flexible Skin and the Transmis-
actuator’s off 共nominal extrados兲 position and to the maximum sion System (Manual Mode). For this testing, the SMA actuators
length of the SMA element and the gas spring are not used and are replaced by a mechanical

091006-6 / Vol. 131, SEPTEMBER 2009 Transactions of the ASME

Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 9 Schema of the flexible extrados testing bench

actuation system composed of the threaded rod 共9兲 connected to
the slider 共3兲 at one extremity, and to the rigid base plate 共2兲 at the
other 共Fig. 10兲. When the rod 共9兲 is rotated, the slider 共3兲 moves,
thus deforming the flexible extrados. The force applied is mea-
sured by the load cell 共10兲 共LC 201-100, Omega Eng. Inc., CT兲.
The force-displacement characteristics of the active structure
corresponding to both actuation lines are presented in Fig. 11. It
can be observed that the behavior of the active structure is hys-
teretic and can be interpreted as linear with two stiffness coeffi-
cients for the loading and for the unloading 共Table 3兲. In this case,
Actuator 1 is manually activated using the rod 共9兲 共Actuator 1 is
active兲; while Actuator 2 is dragged by the flexible extrados
movement 共Actuator 2 is passive兲. To allow independent control
of each actuator, it should be possible to keep Actuator 2 motion-
less when Actuator 1 moves and vice versa. This indicates that
when more than one actuator is used in the system, blocking
forces 共Fbloc兲 corresponding to each actuator should be deter-
mined and considered in the selection of the bias spring for each
actuator 共see Table 3兲.

Fig. 10 Schema of the extrados testing bench for characterization of the

flexible skin and the transmission system

Fig. 11 Force-displacement characteristics of the active structure „Actuator 1 is active

and Actuator 2 is passive…

Journal of Mechanical Design SEPTEMBER 2009, Vol. 131 / 091006-7

Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 3 FE calculated and measured data for Actuator 1
Parameters
Maximum vertical extrados displacement, ␦ext
max
共mm兲
Extrados stiffness coefficient, kext 共N/mm兲
Blocking forces, Fbloc 共N兲, for Actuators 1 and 2
The difference between the experimental and the modeled ac-
tive structure characteristics can be explained as follows: 共a兲 the
displacement of 8.16 mm was intentionally limited to 8 mm be-
cause of the design and manufacturing constraints and 共b兲 the
manufacturing tolerances and friction explain the hysteretic be-
havior and the maximal measured value of 7.7 mm. The latter
value represents physical limitation of the system and cannot be
exceeded because any additional displacement is prevented by the
stop slider 共see a drastic increase in reaction forces when the 7.7
mm displacement is reached, Fig. 11兲.
4.1.2 SMA Actuator Testing. The current arrangement inside
the wing box implies a paired number of wires and allows a maxi-
mum number of six Ti–50.26 at. % Ni wires per actuator group.
The SMA wires are disposed along the span direction, and a pul-
ley system is used to double their working length. The SMA pre-
strain can be regulated with the adjustment screw installed be-
tween the rigid base plate and the T-support. The total SMA wires
cross section for one actuation group will therefore be S = 6 ⴱ 0.7
= 4.2 mm2 and the length as LSMA = 1800 mm. These values give
safety factors of 1.5 for the stroke and 1.2 for the force available
for actuation as compared with the calculated data 共see Sec. 3.4兲.
The spring force can be calculated as follows 共see Eq. 共9兲 with
Faero = 0 N and Table 3兲:
max
Fbloc + Faero
c
Fspr = FSMA + = 1100 N 共11兲
3 ment screw 共Fig. 9兲.
The selected gas spring 共Industrial Gas Spring Inc., PA兲 has a
1000 N-rated load capacity and 2 N/mm stiffness for loading and
unloading curves 共see the experimental force-displacement plots
Fig. 12 Force-displacement characteristics of the gas spring „1000 N, In-
dustrial Gas Spring Inc., PA…
091006-8 / Vol. 131, SEPTEMBER 2009
Downloaded From: http://mechanicaldesign.asmedigitalcollection.asme.org/ on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Value
FE calculated Measured
8.16 7.70
100 共loading兲
110.6
82 共unloading兲
153.6 共Actuator 1兲 100 共Actuator 1兲
174.4 共Actuator 2兲 180 共Actuator 2兲
of Fig. 12兲. During setup, the GS was loaded to 120 mm 共1270 N兲
and then slightly unloaded to the requested 1100 N, which corre-
sponds to the nominal extrados position 共1兲. During actuation,
SMA wires move the GS from its initial position of 120 mm to its
final position of 120+ 7.7ⴱ 3 = 143.1 mm corresponding to the
force of 1285 N 共1 → 2 → 3兲. To regain the nominal extrados po-
sition, the GS moves backward 共3 → 4 → 1兲 and then the cycle can
be repeated.
SMA actuators have been thermomechanically cycled before
their integration in the prototype in order to stabilize their me-
chanical properties. After 200 training cycles under a constant
stress of 230 MPa, corresponding to 3% of the initial strain, the
SMA in the cold state 共Fig. 6兲 had a martensitic strain of 5.5%
c
corresponding to LSMA = 1900 mm. The wires were then installed
on the rigid base plate 共Fig. 9兲 and the slider was blocked in the
nominal extrados position corresponding to 5% of the SMA strain.
By heating of the SMA active elements, the exceeding 0.5% cold
strain was recovered, thus establishing the necessary installation
conditions.
The force in the SMA wires 共1兲 was first removed by untwisting
the adjustment screw, and then the flexible extrados 共5兲 was fixed
to the cranks 共4兲 with the slider 共3兲 in the nominal extrados posi-
tion. The gas spring 共7兲 was then fixed to the rigid base plate 共2兲
and a prestrain of 120 mm was imposed with the tensioner. Finally
the prestrain on the SMA wires 共1兲 was restored using the adjust-
To control the SMA active elements, a power supply 共SPS 100-
33, AMREL Power Product Inc., CA兲 was used to inject electrical
current into the specimen, and the current was controlled using the
Transactions of the ASME

Fig. 13 SMA design diagram for Actuator 1 „force-
displacement space…
displacement signal as a feedback. To avoid overheating, speci-
men’s temperature was measured using a K-thermocouple placed
directly on the surface of the specimen and electrically isolated
from it by a Wakefield thermal compound 共the maximum tempera-
ture was set at 130° C兲. The data acquisition and the control were
performed using a PC programed for this application with LAB-
VIEW software.
For Actuator 1, the experimentally measured 共solid line兲 and the
calculated 共dotted line兲 force-displacement trajectories are super-
posed in Fig. 13, the latter being obtained from the stress-strain
design diagram of Fig. 7 using the known geometry of the SMA
active element 共S = 4.2 mm2 and LSMA = 1800 mm兲. As can be
observed, after 50 s of heating at 9 A and 15 V 共135 W兲, the SMA
temperature was 112° C and a maximum force of 1550 N was
necessary to generate 7.7 mm of flexible extrados displacement
共with the transmission ratio R = 3, this corresponds to 23.1 mm of
the SMA active element stroke兲. Upon cooling, the active struc-
ture takes 115 s to come back to the initial profile. The time
response for a complete actuation cycle is 165 s, thus correspond-
ing to a 10−2 Hz frequency, which can be increased using forced
cooling or antagonist actuator configuration 共2 SMA connected in
series兲. Note that in this case, the SMA active element was not
used at its maximum potential in terms of stress and strain gen-
eration. Increasing heating temperature of the SMA or its initial
tension will definitely increase force generation. However, both
manipulations could lead to premature degradation of the material
during cycling and thus shortens the active life of the system.
5 Conclusion
The objective of this work was to develop a design methodol-
ogy for SMA actuators applied to an experimental morphing wing.
The conceptual design of such a 500 mm chord, 990 mm span,
2D-aspect laminar morphing wing prototype combines three main
subsystems: 共1兲 flexible extrados made of laminar composite ma-
terial, 共2兲 rigid intrados, and 共3兲 an actuator system located inside
the wing box designated to perform extrados shape modifications.
The adaptive structure designed here consists of a four-ply flexible
extrados powered by two individually-controlled SMA actuators.
The forces that these actuators must provide in order to fit target
wing profiles were calculated for a given flexible structure. Each
actuator includes linear SMA active elements with one extremity
connected to the flexible extrados via a cam transmission system,
and the other extremity connected to the rigid structure of the
Journal of Mechanical Design
DownloadedViewFrom:
publicationhttp://mechanicaldesign.asmedigitalcollection.asme.org/
stats on 08/22/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
wing via a gas spring. To meet the functional requirements of the
application, the geometry 共length and cross section兲 of the SMA
active elements and the bias spring characteristics were calculated
and experimentally validated in the absence of aerodynamic pres-
sure.
For the wind tunnel test, the cold and hot state forces recalcu-
c
lated with the aerodynamic load for Actuator 1 will be FSMA
h
= 1050 N and FSMA = 1835 N. For this testing, the gas spring
should have Fspr = 1700 N load rating capacity. Given that the
maximum force of 1835 N that the designed SMA active element
共S = 4.2 mm2 and LSMA = 1800 mm兲 must provide will be less
than the 2100 N corresponding to the maximum allowable stresses
of 500 MPa, the experimental wing should safely morph under all
the aerodynamic conditions considered in this study. This last as-
sertion will be verified during the forthcoming wind tunnel test-
ing, as well as the stability of the SMA active element behavior
during repetitive actuation.
Acknowledgment
The authors acknowledge the contributions of C. Fischer, S.
Bérubé, and J.-S. Ratelle to the design and manufacture of the
experimental morphing wing prototype used in this study. The
authors would like to thank the Consortium for Research and In-
novation in Aerospace in Quebec, the Natural Sciences and Engi-
neering Research Council of Canada, Bombardier Aerospace, and
Thales Canada for their financial support. The authors acknowl-
edge G.-H. Simon from Thales Group as the project initiator.
References
关1兴 Butero, F., 2008, “Shape Memory Actuators for Automotive Applications,”
Advanced Materials and Processes, ASM International, 166共3兲, pp. 37–40.
关2兴 Furuya, Y., and Shimada, H., 1991, “Shape Memory Actuators for Robotic
Applications,” Mater. Des., 12共1兲, pp. 21–28.
关3兴 Mavroidis, C., Pfeiffer, C., and Mosley, M., 1999, “Conventional Actuators,
Shape Memory Alloys and Electrorheological Fluids,” Automation, Miniature
Robotics and Sensors for Non-Destructive Testing and Evaluation, Y. Bar-
Cohen, ed., American Society for Non-Destructive Testing Inc. 共ASNT兲, pp.
189–214.
关4兴 Seelecke, S., and Müller, I., 2004, “Shape Memory Alloy Actuators in Smart
Structures: Modeling and Simulation,” Appl. Mech. Rev., 57共1兲, pp. 23–46.
关5兴 Elwaleed, A. K., Mohamed, N. A., Nor, M. J. M., and Mustafa, M. M., 2007,
“A New Concept of a Linear Smart Actuator,” Sens. Actuators, A, 135共1兲, pp.
244–249.
关6兴 Campanile, L. F., Seack, O., and Sachau, D., 2000, “Belt-Rib Concept for
Variable-Camber Airfoils: Recent Developments,” Proceedings of the Confer-
ence on Smart Structures and Integrated Systems, Newport Beach, Society of
Photo-Optical Instrumentation Engineers 共SPIE兲, 3985, pp. 110–120.
关7兴 Campanile, L. F., and Keimer, R., 2003, “A Shape-Memory Actuator for Sur-
face Geometry Control,” Proceedings of the International Conference on Mar-
tensitic Transformation (ICOMAT’02), Espoo, Finland, EDP Sciences, 112共2兲,
pp. 1189–1192.
关8兴 Coutu, D., Brailovski, V., Terriault, P., and Fischer, C., 2007, “Experimental
Validation of the 3D Numerical Model for an Adaptive Laminar Wing With
Flexible Extrados,” Proceedings of the 18th International Conference of Adap-
tive Structures and Technologies (ICAST ’2007), Ottawa, Canada.
关9兴 Eggleston, B., Poole, R. J. D., Jones, D. J., and Khalid, M., 1987, “Thick
Supercritical Airfoils With Low Drag and Natural Laminar Flow,” J. Aircr.,
24共6兲, pp. 405–411.
关10兴 Coutu, D., Brailovski, V., and Terriault, P., 2009, “Promising Benefit of an
Active-Extrados Morphing Laminar Wing,” J. Aircr., 46共2兲, pp. 730–731.
关11兴 Brailovski, V., Terriault, P., Coutu, D., Georges, T., Morellon, E., Fischer, C.,
and Bérubé, S., 2008, “Morphing Laminar Wing With Flexible Extrados Pow-
ered by Shape Memory Alloy Actuators,” Proceedings of the ASME Confer-
ence on Smart Materials, Adaptive Structures, and Intelligent Systems
(SMASIS 2008), Ellicott City, MD, Paper No. 377.
关12兴 Georges, T., Brailovski, V., Coutu, D., and Terriault, P., 2007, “Design Dia-
gram for Linear SMA Actuators Integrated in a Morphing Wing Structure,”
Proceedings of the International Conference on Shape Memory and Superelas-
tic Technologies (SMST ’2007), Tsukuba, Japan, ASM International, pp. 455–
462.
SEPTEMBER 2009, Vol. 131 / 091006-9