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

net/publication/319293829

Shape Memory Effect and Performance of a Nitinol Engine

Article · January 2017

CITATIONS READS

0 3,220

1 author:

Karthikeyan Palani
Jain University
16 PUBLICATIONS   43 CITATIONS   

SEE PROFILE

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

Computational Fluid Dynamics - Two-phase Flows View project

Heat Transfer in Nanofluids View project

All content following this page was uploaded by Karthikeyan Palani on 26 August 2017.

The user has requested enhancement of the downloaded file.

 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

SHAPE MEMORY EFFECT AND PERFORMANCE OF A NITINOL ENGINE

Vyavahare P.V
School of Mechanical and Building Sciences, VIT University Chennai, India

Karthikeyan C.P
School of Mechanical and Building Sciences, VIT University Chennai, India

ABSTRACT

This paper mainly focuses on the parameters that affect power output of Nitinol heat engine and
phase transformation characteristics of Nitinol wire. In order to conduct this study, a two-pulley
engine is fabricated and tested for power generation. Thermal parameters influencing the phase
transformation of the Ni-Ti wire are determined. Results show that temperature of bath has a
significant impact on power output. A mathematical model is developed for nitinol engines using the
shape memory effect. Efficiency estimated using the model developed is compared with the Carnot
efficiency. The comparative results of the efficiency were found to be inversely proportional to each
other.

Keywords: Shape memory alloy, transformation temperature, phase transformation, austenite,

martensite.

1. Introduction

Despite several deformations, shape memory alloys (SMA) can still remember their predetermined
shape. SMA can be austenite, martensite or a mixture of them depending on the temperature.
Normally nitinol transforms between the high temperature phase called B2 phase (also called as
austenite - denoted by P) and the low temperature phase called B19 phase (also called as martensite,
denoted by M). However due to thermal and mechanical effects such as thermal cycling, chemical
composition, heat treatment, deformation may appear as an intermediate phase known as
rhombohedral or R-phase (denoted by R) between the transformation of austenite to martensite
which result in a two-stage transformation. Numerous elements can influence the transformation
characteristics in NiTi-based shape memory alloy: variation in nickel content, after solution treatment
aging, combined thermo-mechanical treatment, thermal cycling, doping with ternary alloying
elements and processing techniques [1]. There are many factors which affect the power output of
nitinol engine such as dimensions of the sink, diameter of nitinol wire, centre distance between two
pulleys, number of turns of wire, bath temperature. Another factor of interest in the testing of the
SMAs is heating/cooling rate which is impacted by heat transfer for a prolonged period, as observed
in most of SMA’s. Until now there are many publications which concentrate on the thermal and
mechanical effects on the transformation behaviour of NiTi SMA [2]. However, the study of different
parameters and transient study of SMA is still limited till date.

2. History of SMA

www.australianresearchjournals.com.au 109
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

While conducting experiments with an alloy of gold (Au) and cadmium (Cd), an interesting
phenomenon was discovered by Arne Olandder in 1932. When cooled, the Au-Cd alloy could be
plastically deformed and then be heated to return to predetermined shape. This phenomenon is
called the Shape Memory Effect (SME) and the alloys that show this behavior are called Shape
Memory Alloys (SMA). At the Brussels World’s fair, researchers Chang and Read demonstrated the
Shape Memory Effect in 1958. By lifting a weight using an Au-Cd SMA, they showed that the SME
could be used to perform mechanical work. After this, further research has been carried to
determine other materials demonstrating the same phenomenon. In 1962, a group of U. S. Naval
Ordnance Laboratory analysts, lead by William Beuhler, unearthed a huge revelation in the field of
SME and SMA. An alloy of nickel and titanium was being tested for heat and corrosion resistance and
they came to know that the alloy of nickel and titanium also exhibits SME phenomenon. The Ni-Ti
SMA was found to have some advantages over already discovered alloys; they were significantly less
expensive, easier to work with, and less dangerous (from health standpoint). These factors create
interest and motivate others to research in the Shape Memory Effect and applications of shape
memory effect [3].

2.1. Principle of Operation

Shape Memory alloys are the alloys of metallic materials which, when subjected to suitable thermal
procedure, return to their original position. Au-Cd, Cu-Sn, Cu-Zn-(X), Cu- Al-Ni, In-Ti, Ni-Al, Ni-Ti, Fe-
Pt, Mn-Cu, and Fe-Mn-Si are some examples of these alloys. Between the two different phases,
called austenite and martensite, a temperature and stress based shift occurs in the material’s
crystalline structure and thus the shape memory effect (SME) is observed. Soft phase is called
martensite phase while hard phase is called austenite phase. Following is a simple example to explain
SME in action.

Consider the SMA in austenite phase. If it is cooled below its phase transition temperature, the
crystalline structure will change to hard phase, i.e. the martensite phase. If bar is plastically
deformed by any means, say bending, and heated above transition temperature then the bar will
return to its original position due to phase transformation from martensite to austenite. Let us make
this phenomenon simpler by a simplified two-dimensional representation of the material’s crystalline
arrangement, as shown in Figure 1. Every single box represents a grain of material with border
representing grain boundary. Due to grains, a heavily twinned structure is formed; this means that
the grains are oriented across grain boundaries in a symmetric order. The twinned structure permits
the interior cross section of individual grains to change while keeping the same interface with
contiguous grains. Therefore, Shape Memory Alloys are fit for encountering large macroscopic
deformations while maintaining considerable order within their microscopic structure. For instance,
if a bit of SMA begins as austenite [Figure 1a], the interior nuclear grid of every grain is cubic, making
grains with pretty much right points.

If phase transformation temperature is reached by cooling action, the crystalline structure changes
to martensite [Figure 1b] and the grains breakdown to the structures like jewels. Note that the grains
are oriented in different directions for different layers. Presently, if adequate anxiety is connected,
the martensitic structure shown in Figure 1b will begin to yield and "de-twin" as the grains re-situate,
such that they are all balanced in the same direction [Figure 1c]. This conduct can be better
comprehended by looking at a regular anxiety strain bend for the martensite stage [Figure 2].
www.australianresearchjournals.com.au 110
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

Figure 1. Material Crystalline Arrangement during the Shape Memory Effect

At first, for little hassles, the structure shown in Figure 1b acts flexibly from area 0 to 1. At position 1,
the material yields and de-twinning happens somewhere around 1 and 2. At 2, the martensitic
structure is totally de-twinned as shown in Figure 1c. In the blink of an eye, a second flexible area
happens from 2 to 3. At 3, an enduring plastic distortion starts that is not recoverable by the SME.

Figure 2. Stress -Strain Relationship of a Shape Memory Alloy

2.2. Nickel-titanium (Ni-Ti) shape memory alloy

Out of all the Shape Memory Alloys that have been developed so far, Nickel-Titanium (Ni-Ti) has
proven to be the most beneficial and flexible in engineering applications. Some of the following
characteristics of Ni-Ti make it better than the other SMA’s: better ductility, recoverable motion is
fast, excellent corrosion resistance (comparable to series 300 stainless steels), high biocompatibility,
stable transformation temperature, and the ability to be electrically heated for shape recovery.

Ni-Ti SMA is the paired, equiatomic intermetallic compound of nickel and titanium. It is made up of
approximately 50 atomic% Ni and 50 atomic% Ti. The valuable qualities of the intermetallic
compound are moderate solvency range for overabundance Ni or Ti as well as for most other metallic
elements, and ductility comparable to most ordinary alloys. The solvency permits Ni-Ti to be alloyed
with different components which enables improvements in mechanical properties and stage change
temperature (where stage change temperature is taken to mean AF). Adding additional Ni to the
binary compound (up to 1% extra) strongly depresses the phase transformation temperature and
increases the yield strength of the austenite. Iron and chromium can likewise be added to bring
www.australianresearchjournals.com.au 111
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

down the transition temperature. By shifting these and different components, the change
temperature can be fluctuated from –200 to 110°C (- 325 to 230°F). Copper can be used as an
additive to decrease the hysteresis and lower the deformation stress (de-twinning stress) of the
martensite. Table 1 shows the key physical properties of equiatomic Ni-Ti SMA.

Table 1 Properties and phase transfromation of Nitinol

Properties Austenite Martensite

Melting temperature, ̊ C 1300
Density, g/cm3 6..54
Resistivity, Ω-cm Approx. 100 Approx. 70
Thermal conductivity,
18 8.5
W cm/o C
Similar to 300 series stainless steel or titanium
Corrosion resistance
alloys
Young’s Modulus, GPa Approx. 83 Approx. 28
Yield strength, Mpa 195 to 690 70 to 140
Ultimate tensile
895
strength, MPa
Transformation
-200 to 110
temperature, ̊C
Latent heat of
167
transformation, ̊C
Shape memory strain 8.5% maximum

2.3. Manufacturing and shaping of Nitinol

Manufacturing Nitinol SMA and shaping it for a specific purpose is very tough task. Inert atmosphere
has to be maintained while melting Ti as it is a very reactive element. Plasma-arc melting, electron-
beam melting, and vacuum induction melting are some of the common methods used during this
process. Standard hot-forming and cold-working processes are used in order to initially shape the Ni-
Ti ingots. During cold working, alloy must annealed frequently as it gets work harden very quickly. By
minimizing the stress needed to de-twin the martensite and increasing the strength in the austenite
phase, we can improve SMA's performance. For this, we can use processes like work hardening and
heat treatment processes. Machining Ni-Ti through cutting methods is difficult, as is welding,
soldering, and brazing. For creating specific shapes grinding, shearing, and punching are better
methods. The “memory configuration” that is giving shape memory to SMA is done by holding the
part in desired shape, and then heat-treating at 500 to 800°C (950 to 1450°F). Prefabricated SMA
elements like wire, rod, ribbon, strip, sheet, and tubing are provided by companies like Dynalloy, Inc.
and Shape Memory Applications, Inc. This companies also provides custom element to user
specification.

3. Methodology

3.1 Modeling

Model consists of simply two pulleys. A nitinol wire of 0.5 mm diameter is run between two pulleys.
www.australianresearchjournals.com.au 112
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

Two pulleys were made up of glass. Diameter of the bigger pulley is 70mm and diameter of the
smaller pulley is 40mm. Center distance between the two pulleys can be varied by fixing adjustable
bolt at different positions. Solidworks model of two pulley system helps to visualize the model and
the challenges in fabrication.

3.2 Experimental setup

The engine [Figure 3a], generates power by using nitinol wire loop. The loop made of nitinol runs in
between the two pulleys. Hot water on hot side and cool ambient air on cool side is used by this
device in order to perform action. The smaller wheel of engine is partially dipped in a hot liquid.
Memory given to the nitinol loop wire in engine is 'straight' shape. Whenever a part of loop gets in
contact with a hot liquid, it gets heated above its transition temperature and tries to straighten out.
Figure 3a explains the phenomenon. From Figure 3b, it is clear that at position 1, nitinol wire is
straight and cool. As the wire travels from position 1 to position 2, it gets bent around the small glass
pulley and comes in contact with hot liquid. As the wire moves from position 2 to position 3, it comes
in contact with hot water and thus gets heated above its transition temperature due to which it tries
to straighten out. As the wire tries to straighten out, it takes form the form shown by dotted line.
While doing so, the wire exerts a tugging force of magnitude F along the loop. As the wire moves
from position 3 to 4, it comes in contact with ambient air due to which its temperature decreases
and, thus, austenite phase starts converting to martensite phase. As the wire moves from position 1
to position 4, it travels over the bigger pulley and a sufficient long time is available for the wire to
cool below its transition temperature; the wire get ready for another cycle. In simple words, one side
of loop stiffens due to high temperature while the other side i.e. air side of loop, cools and relaxes. A
wheel pulley rotates due to mechanical force. Sometimes it becomes necessary to jump start the
engine by giving rotation to the bigger pulley manually. Interestingly, the engine hasn’t a set
rotational direction. Whichever way it is started it will continue to rotate. The engine can also be
powered by using solar energy. A magnifying lens focusing sunlight on the smaller glass wheel also
supplies sufficient temperature gradient to power the engine.

(a) (b)

Figure 3. Working principle of nitinol engine

www.australianresearchjournals.com.au 113
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

3.3 Experiments performed

Aim of this experiment is to find out variation of speed of nitinol engine with respect to decrease in
temperature of bath. Water is takeen as a heating fluid as it is easily available and it has many
advantages over other fluids as heating source. Water is heated to 90 degrees Celsius with help of
conventional heating process and then the heating source is removed. Water is then poured in a flask
and nitinol engine is placed over the flask. The temperature on cool side is room temperate which
was found to be 32 degrees Celsius during the experiment.
As explained in working principle in the above section, the engine starts rotating. Speed of engine is
measured with the help of laser tachometer. Figure 4 shows the laser tachometer used for
measuring speed of nitinol engine in RPM. As the time passes, temperature of water bath goes on
decreasing due to heat loss to surrounding. Variation in speed of the engine is measured with the
help of laser tachometer.

Figure 4. Laser tachometer

4. Results and Discussion

4.1. Transient analysis: -
Table 2 shows the variation of speed of engine with respect to decrease in temperate of water bath.
Figure 5 shows the output of these results in graphical form. From Table 2 and the aforementioned
graph we can conclude that,

• Engine does not start unless temperature of bath reaches 45 degrees Celsius
• From 45 to 75-degree Celsius speed goes on increasing with the bath temperature
• After 75-degree Celsius, speed remains constant irrespective of increase in temperature of bath

Table 2 (a) and (b) Variation of speed with temperature of bath

(a)
Temperature
S. No RPM
(degree Celsius)
1 90 180
2 87 180
3 84 180
4 81 180
5 78 180
6 75 157
7 72 129
www.australianresearchjournals.com.au 114
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

8 69 110
9 66 100

(b)
Temperature
S. No RPM
(degree Celsius)
10 63 93
11 60 86
12 57 82
13 54 80
14 51 75
15 48 75
16 45 72
17 42 0
18 40 0

Figure 5. Graph of temperature Vs speed

5. Mathematical model for finding efficiency

NITINOL engine is fed by the heat energy of LP water by means of heat transfer at the boundary of
Wire – LP water. Let us consider this process similar to heat transfer of metal surface at the boundary
with the moving water. Then, the intensity of the flow of heat transfer will be equal to: [21]

W= (σ0 + σ1 ) SW ΔT Watt (1)

Where, V is the velocity of heated section of NITINOL engine with the length of lw relative to water,
ΔT=Twarm – Tcool is the temperature gradient (thermal head) – the difference between the
temperatures of LP water, Twarm (hot) and Tcool (cold) water, and σ0 ,σ1 are the coefficients of heat
transfer equal to [21]

σ0=350 W/m2k , σ1=2100 (W/m2k).(s/m)1/2 (2)

The contact surface Sw of wire with LP water for engine is determined by,

SW=n2πr0w=1*2π*0.5*10-3*0.37= 1.16238*10-3 mm2 (3)

Where r0 = radius of nitinol wire which is 0.5 mm in our case

www.australianresearchjournals.com.au 115
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

And lw = length of nitinol wire running between two pulleys =370 mm in our case

Apparently, W is the sum of two summands:

W= WQ+WA (4)

Where, WQ is a part of W consumed continuously for heating of wire at inverse martensitic

transition M→A, and WA is a part of W at the expense of which engine develops the power.

Q is the quantity of heat transferred to wire for specific time at operating conditions, necessary for
M→A transition

Q=M.CP.ΔT Watt (5)

The value CP is the average value of the product of specific heat capacity of nitinol C, from the
experiment, it is known that in the case of nitinol [21]

Where ,

CP=5.2*106 J/m3k (6)

And M= mass flow rate of wire through fluid (m3/s)

M=πr20 V (7)

By definition, the thermal efficiency is the relation between the work and the quantity of heat
received by engine at heating, i.e.

η=WA/Q (8)

Here WA is considered instead of W because only WA is a part of work developed by engine while
other part i.e. WQ is absorbed by engine

Hence,
(9)
WA= W/2

Now, let us find thermal efficiency of engine according to Carnot, for comparing Carnot efficiency of
engine with actual efficiency consider the readings obtained while performing experiment of
temperature vs. speed.

At 90 degrees’ Celsius temperature of water, speed given by NITINOL engine was N=180 RPM

According to Carnot, efficiency is determined as

ηK =1- (Tcool / TWarm) (10)

In this case, TCOOL = Temperature of cool air (i.e. room temperature) = 305 K TWARM= Temperature of
hot water = 363 K.

Putting this values in Eq(10), we get

(11)

www.australianresearchjournals.com.au 116
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

ηk = 0.1597= 15.97%

Now for finding actual efficiency of engine we will make use of Eq (8),

Putting values of Eq(2) and Eq(3) in Eq(1) and substituting ΔT=363-305=58 k, we get

W=(350+2100*)*1.16238*10-3*58 (12)

0Now V= (ΠDN/60) (13)

Where D= diameter of rotating wheel = 70 mm

V= 0.6597 m/sec (14)

Putting Eq (14) in Eq (12), we get,

W= 138.58 W
(15)

From Eq (9)

WA = = 69.29 W (16)

Now from Eq (5),(6) AND (7),

Q=M.CP.ΔT

Q=π*(0.5*10-3)2*0.6597*5.2*106*58

Q=156.2753 W (17)

Putting Eq (16) and Eq (17) in Eq (8), we get

η= = 0.4433= 44.33% (18)

We can compare the Carnot and actual efficiency of nitinol engine now by using Eq (11) and Eq (18);
actual efficiency of nitinol engine is greater than Carnot efficiency. This happens because nitinol
engine does not require any work output for its operation like Carnot cycle requires. Thus, it is not
limited to Carnot efficiency. However, Carnot efficiency of each case is obtained just for comparison
with other engine working on Carnot cycle.

For different cases, comparison between actual and Carnot efficiency is tabulated in Table 3.

The tabulated results are represented by graph in Figure 5.

www.australianresearchjournals.com.au 117
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

Table 3. Efficiency calculation

Sr. Tcool Twarm ΔT Vrel W WA Q

ηk η
no (K) (K) (K) (m/s) (w) (w) (w)
1 305 363 58 0.6597 138.59 69.29 156.27 0.1597 0.4434
2 305 360 55 0.6597 131.42 65.71 148.19 0.1527 0.4434
3 305 357 52 0.6597 124.25 62.12 140.10 0.1456 0.4434
4 305 354 49 0.6597 117.08 58.54 132.02 0.1384 0.4434
5 305 351 46 0.6597 109.91 54.95 123.94 0.1310 0.4434
6 305 348 43 0.5754 97.11 48.55 101.05 0.1235 0.4805
7 305 354 49 0.4728 102.18 51.09 94.61 0.1384 0.5399
8 305 342 37 0.4030 72.39 36.19 60.91 0.1081 0.5942
9 305 339 34 0.3665 64.08 32.04 50.90 0.1002 0.6294
10 305 336 31 0.3408 56.79 28.39 43.15 0.0922 0.6570
11 305 333 28 0.3151 49.76 24.88 36.03 0.0840 0.6904
12 305 330 25 0.3000 43.63 21.81 30.69 0.0757 0.7107
13 305 327 22 0.2924 37.99 18.99 26.27 0.0672 0.7228
14 305 324 19 0.2748 32.04 16.02 21.33 0.0586 0.7511
15 305 321 16 0.2748 26.98 13.49 17.96 0.0498 0.7511
16 305 318 13 0.2638 21.59 10.79 14.01 0.0408 0.7704

From Table 3, we conclude that efficiency of engine goes on increasing as the temperature gradient
decreases. This happens because the decrease in W is very less as compared to the decrease in Q.
Due to this ratio, efficiency increases.

Figure 5. Graph of temperature difference Vs efficiency

6. Summary
Thermal parameters influencing the speed of the Ni-Ti engine are determined. Results show that the
temperature of bath plays a major role on power output. From mathematical results we can
conclude that the actual efficiency of nitinol engine is inversely proportional to the temperature
gradient.
7. References
1. Zheng, Y., Cui, L., Schrooten, J., 2004. Temperature memory effect of a nickel–titanium shape
memory alloy. Appl. Phys. Lett. 84-1, 31.

www.australianresearchjournals.com.au 118
 AUSTRALIAN JOURNAL OF ENGINEERING AND TECHNOLOGY RESEARCH (AJETR)
Vol. 2, Issue 1 - 2017

2. Kurita, T., Matsumoto, H., Abe, H., 2004. Transformation behavior in rolled NiTi. J. Alloys Compd.
381, 158–161.
3. Huang, W.M., 2005. Transformation front in shape memory alloys. Mater. Sci. Eng. A 392, 121.
4. Huang, X., Liu, Y., 2001. Effect of annealing on the transformation behavior and super-elasticity of
NiTi shape memory alloy. Scr. Mater. 45, 153–160.
5. Kurita, T., Matsumoto, H., Sakamoto, K., Abe, H., 2005. Transformation behavior of shock-
compressed Ni48Ti52. J. Alloys Compd. 400, 92.
6. Liu, Y., McCormick, P.G., 1990. Factors influencing the development of two-way shape memory in
NiTi. Acta Metall. Mater. 38-7, 1321–1326.
7. Liu, Y., Humbeeck, J.V., Stalmans, R., Delaey, L., 1997. Some aspects of the properties of NiTi
shape memory alloy. J. Alloys Compd. 247, 115–121.
8. Liu, Y., Yang, H., Voigt, A., 2003. Thermal analysis of the effect of aging on the transformation
behavior of Ti 50.9 at.% Ni. Mater. Sci. Eng. A 360, 350.
9. Chrobak, D., Str ´o˙ z, D., Morawiec, H., 2003. Effect of early stages of precipitation and recovery
on the multi-step transformation in deformed and annealed near-equiatomic TiNi alloy. Scr.
Mater. 48-5, 571
10.Otsuka, K., Wayman, C.M., 1998. Shape Memory Materials. Cambridge University Press.
11.Somsen, C., Z¨ahres, H., K¨ astner, J., Wassermann, E.F., Kakeshita, T., Saburi, T., 1999. Influence
of thermal annealing on the martensitic transitions in Ni–Ti shape memory alloys. Mater. Sci. Eng.
A 273–275, 310–314.
12.Lahoz, R., Puertolas, J.A., 2004. Training and two-way shape memory in NiTi alloys: influence on
thermal parameters. J. Alloys Compd. 381, 130–136.
13.Frick, C.P., Ortega, A.M., Tyber, J., Maksound, A.El.M., Maier, H.J., Liu, Y., Gall, K., 2005. Thermal
processing of polycrystalline NiTi shape memory alloys. Mater. Sci. Eng. A 405, 34–49.
14.Gorbet, R., and Wang, D., 1998, “A Dissipativity Approach to Stability of a Shape Memory Alloy
Position Control System”, IEEE Transactions on Control Systems Technology, Vol. 6, No.

www.australianresearchjournals.com.au 119

View publication stats