Renewable Energy 139 (2019) 600e610

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Numerical studies of co-axial travelling-wave thermoacoustic cooler

powered by standing-wave thermoacoustic engine
Patcharin Saechan a, Artur J. Jaworski b, *
a
Department of Mechanical & Aerospace Engineering, Faculty of Engineering, King Mongkut’s University of Technology North Bangkok, Bangkok, 10800,
Thailand
b
School of Computing and Engineering, University of Huddersfield, Huddersfield, HD1 3DH, United Kingdom

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

Article history: Design and optimisation of a co-axial travelling-wave thermoacoustic cooler (TWTC) powered by a
Received 12 September 2018 standing-wave thermoacoustic engine (SWTE) is conducted based on the linear thermoacoustic theory.
Received in revised form The target application is storage of medical supplies in rural communities of developing countries with
4 February 2019
no access to electricity grid, where waste heat from cooking can be used as energy input. The design is
Accepted 11 February 2019
Available online 13 February 2019
guided by construction simplicity and low cost. Thus the SWTE/TWTC coupling is enclosed in a constant-
diameter resonator which accommodates a compact co-axial configuration of the TWTC. Following
earlier experiments, working fluid is air at 10 bar, and operating frequency is 50 Hz. Parameters such as
Keywords:
Coupled thermoacoustic engine-cooler
the ratio of the cross-sectional area of the cooler to feedback inertance and stack and regenerator ge-
Travelling-wave ometries, have been optimised for maximum system efficiency. Here, at the target input power of 600 W
Standing-wave a drive ratio of 7.75% can be obtained, leading to the cooling load of 133 W at 250 K. This corresponds to
Design and optimisation the SWTE thermal-to-acoustic efficiency of 17% and TWTC COP of 1.9, and the combined efficiency of
Sustainable system 22.2%. The simulations are to inform the build of an enhanced prototype of the proposed device. A
Developing countries detailed discussion and guidelines for practical builds are provided.
© 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

1. Introduction and background
Thermoacoustic (TA) technologies deal with the conversion
between thermal and acoustic energies. In TA engines an acoustic
oscillation is powered by thermal energy; in TA refrigerators
(coolers) the “heat pumping” from a lower temperature “source” to
a higher temperature “sink” is driven by the acoustic power [1].
They are considered environmentally friendly since inert gases, or
simply air, used as working media do not cause ozone layer
depletion. In addition, with no moving parts, they are highly reli-
able and exhibit a consistent performance [2]. TA devices can utilise
any source of thermal input, including industrial waste heat [3,4] or
renewable energy sources such as solar energy [5], sustainable
biomass/biogas combustion or geothermal energy. Moreover, a
combined system of a TA engine (TAE) driving a TA cooler (TAC)
may be considered “sustainable” if the heat source for the engine is
one of those identified above.
Current work arises from recent studies [6] aiming to use waste
* Corresponding author.
E-mail addresses: a.jaworski@hud.ac.uk, a.jaworski@yahoo.co.uk (A.J. Jaworski).
heat from biomass-fuelled cooking to provide refrigeration capa-
bilities to store medical supplies in rural communities of devel-
oping countries with no access to electricity. Fig. 1(a) illustrates the
concept whereby TAE, excited by a temperature difference between
cookstove flue gases and the ambient, generates acoustic power
(W_ ) used in turn by TAC working between the temperatures
AC;cool
of the ambient and the “cool box”.
The concept, tested in laboratory conditions, used electrical
heaters to mimic both the heat input and cooling load. Fig. 1(b) and
(c) show the schematic and photograph of the experimental
apparatus (overall length 3.7 m). A standing-wave thermoacoustic
engine (SWTE) powers a travelling-wave thermoacoustic cooler
(TWTC) of a co-axial design (cf. [7], both sharing a constant-
diameter resonator as pressure containment e in this early
demonstrator an easily available 6-inch pipework, with internal
diameter 161.5 mm, was used). The rig produced the minimum
cooling temperature of
19.7  C (at no cooling load), the maximum
cooling load of 120 W at the cooling temperature of þ8  C (using
2.5 kW heat input from resistive heaters), the maximum COP of 2.4,
and maximum overall efficiency (cooling load to thermal input) of
5.5%.

https://doi.org/10.1016/j.renene.2019.02.059
0960-1481/© 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610 601

(a)

(b)

(c)

Fig. 1. (a) Functional diagram of the coupled TAE and TAC for application in developing countries. (b) Schematic diagram of SWTE coupled to co-axial TWTC; symbols denote: BS e
bounce space, HHX e hot heat exchanger, STK e stack, AHXSWTE e ambient heat exchanger (engine), TBT e thermal buffer tube, CHX - cold heat exchanger, REG e regenerator,
AHXTWTC e ambient heat exchanger (cooler); (c) Photograph of the laboratory test rig with instrumentation.

However, the optimisation of the complete SWTE/TWTC system
has not been carried out. The build of the engine “alone” as a first
step was based on a pre-selected resonator diameter of 6 inch.
Subsequently, the cooler was optimised and built to be coupled to
the existing 6-inch engine. Therefore, this paper presents a new
investigation in order to ascertain how the device can be optimised
in the global (holistic) sense, where engine and cooler parameters
could be varied at the same time, while releasing the constraint of
the fixed resonator diameter. This could lead to selecting a different
containment vessel diameter (especially smaller) to save both the
weight (added portability) and cost of materials, while increasing
the overall thermodynamic performance. Subsequently, new
design parameters could be found for an improved demonstrator to
be driven in future by heat input from sustainable biomass com-
bustion or other renewable energy sources.
Previous work has also provided the design trade-offs for the
device application in rural communities: the working medium is
chosen as air at 10 bar (achievable from truck tyre pumps), while
the operating frequency is around 50 Hz to relax the fabrication
constraints due to small thermal penetration depths. Hence, these
two values are not re-visited in the current optimisation proced-
ures, but will be treated in this paper as assumed constants.
2. Literature review
Hofler [8] demonstrated a successful prototype of TAC driven by
a loudspeaker; subsequently many TACs were developed and
studied, e.g. Garrett et al. [2]; Tijani [9] and Ueda et al., [10]. Since
the early coolers showed relatively low performance, many pa-
rameters/components of TACs had to be investigated to improve
the technology. Studies included the system configuration [11],
stack geometry [12], working fluid [13], resonator [14] and elec-
trical driver [15]. The best cooling performance so far appears to be
30% of a coefficient of performance (COP) relative to Carnot for a
cooling power of 210 W at 233 K [16].
Methods of coupling of TAEs to TACs or pulse-tube coolers have
been investigated previously using both simulations and experi-
ments [17,18]. Such devices can achieve cryogenic temperature
range [19,20], down to about 2 K [21]. Specialised applications
include gas liquefaction [22,23], mixture separation [24] or ice
cream storage [25]. Recently, Sharify & Hasegawa [26] presented a
TAC driven by a multistage travelling-wave engine. A new design of
thermoacoustically-driven thermoacoustic refrigerator (TADTAR)
with an adjustable resonator length has been investigated experi-
mentally by Alcock et al., [27].
Various design strategies were developed for designing TACs e
these can focus on individual components (e.g. stack), cooler sub-
systems (e.g. design of cooler excluding the acoustic driver) or
entire systems including engine/linear alternator together with the
cooler system. Wetzel & Herman [28] investigated the design and
optimisation procedures of a TAC by employing the short stack
boundary layer approximation. Tijani [9] also presented the design
strategy of the standing-wave TACs driven by a loudspeaker. The
optimisation of different parts of TAC has also been discussed.
Ishikawa & Hobson [29] studied the optimum heat exchanger
surface area of a standing-wave TAE by applying minimum entropy
602 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610

generation criteria. Zoontjens et al. [30] presented optimisation of 2 0   13

an inertance segment connecting the standing-wave type TAC to 1 Tm b fk
~f n
TAE and validated their simulations with results of numerical finite H_ 2 ¼ Re4p1 U
~ @1

1  A5
element and experimental approaches. Ueda et al. [31] simulated
2 ð1 þ εs Þð1 þ sÞ 1
~f n
the performance of a travelling-wave TAC having a looped tube by 0   1
rm cp jU1 j2 dTm fk
~f n 1 þ εfs fn
varying the regenerator parameters. Bassem et al. [32] demon-
þ Im @~f n þ k
A
strated the design process of a looped-tube travelling-wave TAC 2Auð1
sÞj1
fn j2 dx ð1 þ εs Þð1 þ sÞ
driven by a linear motor, including optimisation of the regenerator
radius and position. Ghorbanian & Karimi [33] presented the dTm

ðAk þ Asolid ksolid Þ
design and optimisation with a simplified model of a compact dx
travelling-wave TAC driven by a travelling-wave TAE. The position, (4)
length and hydraulic radius of the TAC regenerator were optimised
for the maximum COP. where u is angular frequency, a, rm , Tm , cp , g, k, b and s are speed of
However, the design process of a coupled TAE/TAC system still sound, mean density, mean temperature, isobaric specific heat ca-
involves significant technical challenges due to its complexity. pacity, ratio of specific heats, thermal conductivity, gas expansion
There is no widely accepted “optimal” geometry as this depends on coefficient and Prandtl number of working fluid, respectively. A is
the particular design targets and operating conditions e at the flow area of the channel. i is imaginary unit. Re[], Im[] and tilde
same time the overall performance is very sensitive to the choice of symbol (~) denote the real part, imaginary part and conjugation of a
design parameters from a large variable space. However, the vol- complex quantity, respectively. fn , fk and εs are viscous function,
ume of research available on the holistic design and optimisation of thermal function and correction factor for finite solid heat capacity,
a TAE/TAC coupling is limited. In addition to papers reviewed respectively. These depend on the geometry of individual compo-
herein, Babaei and Siddiqui [34] investigated the geometrical nents and, for instance, for a parallel plate configuration are pre-
optimisation of a standing-wave TAE driving a standing-wave TAC sented as follows (Ward et al. [37]):
which is clearly a different configuration from the present, i.e. h i
ð1þiÞy0
containing a co-axial travelling-wave cooler. Current paper ex- tanh dn
plores the inter-dependencies between the overall efficiency and fv ¼ ð1þiÞy0
(5)
design parameters of the standing-wave engine and co-axial trav- dn
elling-wave cooler in a holistic way. This provides a better under-
h i
standing of the complex interaction between these parameters and ð1þiÞy0
tanh dk
their relative importance, and will enable other researchers to focus fk ¼ ð1þiÞy0
(6)
their investigations on the parameters that are most critical to the dk
overall performance of the selected device planform.
 
krm cp 1=2 tanh½ð1 þ iÞy0 =dk 
3. Design concept and DeltaEC model εs ¼ (7)
krs cs tanh½ð1 þ iÞl=ds 
The design and optimisation in the present work are imple-
where rs , cs and ds are density, heat capacity and penetration depth
mented based on the linear thermoacoustic theory, which assumes
of solid. y0 is half of plate-to-plate spacing in a parallel plate stack
that the device geometry and flow of energy are one-dimensional
and l is the length of the stack. The thermal penetration depth (dk )
and along a path extending from one termination to another
and viscous penetration depth (dn ) in equations (5) e (7) can be
along the central axis of gas oscillation. The fundamental physics
calculated as follows:
concerned with the linear thermoacoustics theory can be described
by the continuity, energy and momentum equations which are the sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
functions of temperature, angular frequency, pressure and volume 2k
dk ¼ (8)
flow rate amplitudes, system geometry and gas properties (see for rm cp u
instance Ref. [35]; or [36]). The changes in pressure amplitude (p1 ),
velocity amplitude (U1 ), mean temperature (Tm ) and total power sffiffiffiffiffiffiffiffiffi
2m
flow (H_ ) along a thermoacoustic segment can be determined as
2 dn ¼ (9)
presented below:
rm u

dp1 iurm where m is the dynamic viscosity of working fluid.

¼
U (1)
dx Að1
fn Þ 1 The acoustic power produced or dissipated in any segment per
unit length can be estimated from the expression given by Swift

 [35] as:
dU1 i uA ðg
1Þfk bðfk
fn Þ dTm
¼
1 þ p1 þ U
dx rm a2 1 þ εs ð1
fn Þð1
sÞð1 þ εs Þ dx 1 _

dW 1 ~ dp1 þ p dU1
(2)
AC
¼ Re U 1 ~ 1 (10)
dx 2 dx dx

  In reaching the optimum design configuration, the efficiency of
Tm bðfk
~f n Þ
H_ 2
12 Re p1 U
~ 1
SWTE (hSWTE ) is considered the first crucial parameter for evalu-
dTm 1 ð1þεs Þð1þsÞð1
~f n Þ
¼ 0  1 ating the performance of the system. It is given by:
dx ðfk
~f n Þ 1þεsfkfn
rm cp jU1 j2 ~
@f n þ A
ðAk þ Asolid ksolid Þ
2 Im ð1þεs Þð1þsÞ _
2Auð1
sÞj1
fn j W
hSWTE ¼ _ AC (11)
QH
(3)
_
where W AC is the acoustic power generated by the SWTE measured
 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610
at the end of the ambient heat exchanger and Q_ H is the input heat
power supplied by the hot heat exchanger. Generally, W _ can be
AC
approximated by using the following equation which excludes
viscous losses [1]:
_ 1 Tm b u 2
2 Working gas
W AC ¼ Pdk Dx p ðG
1Þ (12)
4 rm cp 1
where P is the total effective circumference around the stack,
perpendicular to the heat flow direction, Dx is the length of the
stack, and G is the normalised temperature gradient. The latter is
the ratio of temperature gradient along the stack to the critical
temperature gradient and is given as follows:
VTm
G¼ (13)
VTc
Tm bup1
VTc ¼ (14)
rm cp U1
Performance of the cooler is indicated by its COP. It is defined in
this work as the ratio of the heat load of the TWTC (Q_ C ) to the
_ ) supplied by the SWTE:
acoustic power (W AC
Q_ C
COP ¼ (15)
W_
AC
Due to the cooler being driven by the SWTE, the acoustic power
generated from the SWTE varies as the geometry changes. Since
input acoustic power is variable, it is difficult to consider COP alone
as a measure of system performance. Instead the “overall system
efficiency” (referred to as “cooling efficiency” here) is seemingly the
best description of the overall performance of the proposed device
as it combines the performance of both SWTE and TWTC. It can be
written as:
Q_
hcool ¼ hSWTE COP ¼ _ C (16)
QH
The above parameters are employed in this work to justify the
design choices in the pursuit of the optimal design.
The first element of the combined SWTE/TWTC system is the
engine itself. As already discussed, in this work compressed air at
10 bar is the preferred working fluid, as noble gases may not be
easily available in remote rural communities. Its thermoacoustic
performance at 50 Hz is thought to be acceptable, and its use leads
to a cost-effective device. Briefly, main components of SWTE are: a
stack, two heat exchangers and a resonator, as shown in Fig. 1(b);
their main features are listed in Table 1. The stack for the engine is
made of parallel stainless steel plates because of the simplicity of
fabrication. The central part of the TWTC is the regenerator which
consists of stacked-up layers of stainless steel mesh screens with a
small hydraulic radius compared to the thermal penetration depth
(rh ≪dk ). Therefore, a very good thermal contact between the gas
and solid in the regenerator can be achieved; the temperature of
the gas can be assumed to be the same as the temperature of the
adjacent solid. All heat exchangers in both SWTE and TWTC are of
parallel-plate configuration with the working gas oscillating be-
tween the plates. The details of possible practical implementations
will be discussed later in section 6.
In the design of the TWTC, the cooling temperature of 250 K at
the cold heat exchanger is the target. The TWTC employs an
acoustic network based on the same principle as the torus-shaped
design [7]. It consists of a feedback inertance, compliance and
acoustic resistance as also shown schematically in Fig. 1(b). The
603
Table 1
Selected design parameters of SWTE and TWTC.
Global parameters Values
Frequency 50 Hz
Mean pressure 10 bar
Air
Standing-wave engine (SWTE) Design features
Hot heat exchanger (HHX) Parallel-plate configuration
Stack Stainless steel layered parallel-plates
Ambient heat exchanger (AHX) Parallel-plate configuration
Resonator Stainless steel pipe
Travelling-wave cooler (TWTC) Design features
Cold heat exchanger (CHX) Parallel-plate configuration
Regenerator Stainless steel mesh screen
Ambient heat exchanger (AHX) Parallel-plate configuration
Compliance Stainless steel pipe
regenerator and heat exchangers generate an acoustic flow resis-
tance. The annular space between the resonator and the regener-
ator holder forms the feedback inertance. The cylindrical volume of
gas confined between the closed end of the resonator and the TAC’s
AHX plays the role of acoustic compliance. The acoustic network is
designed to create the travelling-wave phasing necessary to oper-
ate the thermoacoustic Stirling cycle and helps to create high
acoustic impedance in the regenerator. The acoustic power pro-
duced by SWTE is delivered to TWTC in which the direction of heat
flow is opposite to the acoustic wave propagation. This means that
the acoustic waves move from the hot side to the cold side through
the regenerator, whereas heat is pumped in the opposite direction
as reported by Ceperley [38] and Huang et al. [39]. In these cir-
cumstances, the acoustic energy is consumed to pump heat from
the low temperature to the high temperature end. The components
of TWTC are also listed in Table 1.
The design of the coupled SWTE/TWTC and its performance
analysis were aided by the simulation programme DeltaEC [37].
This computer code is widely used to design and predict the effi-
ciency of the thermoacoustic engines and refrigerators. The one-
dimensional wave equations (1) e (4) based on low-amplitude
approximation are solved simultaneously to determine the
change of the variables of state of the working fluid (for example
oscillatory pressure, oscillatory volume flow rate, acoustic power
flow, total power flow etc.) along all segments of a thermoacoustic
device. With a geometry given by the user, a solution found in each
segment is required to match at the junctions the solutions for the
neighbouring segments. The use of DeltaEC has several advantages
including relative model simplicity and short calculation time e
especially compared to fully three-dimensional models. The dis-
advantages include convergence problems in cases where initial
conditions are relatively far from the final solution and where
appropriate guesses and targets for iterative calculations are
defined incorrectly e often many of such issues need to be over-
come through sufficient user experience. Similarly, it is well rec-
ognised in the community that DeltaEC does not offer reliable
models for thermal inputs and thermal losses; certain design as-
sumptions need to be made and these often lead to overestimated
efficiencies. For instance, comparisons between simulations and
experiments, based on the same cooling temperature, conducted
for the previous rig [6] led in the worst-case scenarios to the overall
efficiency reduction from 6.4% (simulation) to 4.35% (experiment).
Use of DeltaEC and similar packages (e.g. SAGE) as sole thermoa-
coustics research tools has been reported by Li et al. [22]; Al-Kayem
& Yu [40] and Hu et al. [41]; amongst many others.
In simulation, the proposed system in Fig. 1(b) can be
604 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610
Fig. 2. Block diagram of the segments in DeltaEC model. Abbreviations have following
meanings: BS e bounce space, HHX e hot heat exchanger, STK e stack, AHX e ambient
heat exchanger (denoted in Fig. 1(b) as AHXSWTE and AHXTWTC, for engine and cooler,
respectively), RES - resonator, TBT e thermal buffer tube, CHX e cold heat exchanger,
REG e regenerator.
characterised as block diagram as illustrated in Fig. 2. The calcula-
tion begins at the bounce space (BS) segment and then ends at hard
end segment which is the usual part of a closed thermoacoustic
system.
When defining the boundaries for the simulation, the volu-
metric velocity must become zero at the main resonator ends due
to rigid walls:
U1 ¼ 0 at x¼0 and x ¼ L;
L being the main resonator length. According to the momentum
and acoustic power equations, these boundary conditions result in:
vp1 _ the surroundings are assumed to take place only through the heat
¼0 and W AC ¼ 0 at x¼0 and x¼L
vx
Due to the flow split at the beginning of the TWTC, the flow
criteria are then set up to ensure that complex pressure amplitude
and temperature at the union point of the main component trunk
and the inertance branch are the same. These conditions are:
p1;Trunk ¼ p1;Branch and TTrunk ¼ TBranch
For the volumetric velocity at the “union”, the condition is of
course flow continuity:
U1;Resonator ¼ U1;Trunk þ U1;Branch
4. Optimisation procedure
In the optimisation process, the boundary conditions are set up,
and initially the geometries of each component in Fig. 2 are
assumed. The solutions are then guessed in order to meet the tar-
gets by solving the differential momentum, continuity and total
power equations simultaneously as described above. Fourth-order
Runge-Kutta integration method is employed in the DeltaEC
design code. If the solutions diverge, the guessed values are then
adjusted and re-calculated until the convergence is achieved. The
shooting method [37] is used in this work to find the guess values
which are consistent with target results. In the first converged
solutions, the efficiency of the system obtained is typically quite
low, as may be expected, because the geometries of the thermoa-
coustic device have not been optimised. Therefore, the optimisation
process is subsequently executed to achieve the maximum effi-
ciency of the entire system based on the multivariable search
method. The optimisation procedure is performed by varying the
values of the parameters in each component of the SWTE, including
the cross sectional area, the plate spacing and the length of the
stack, the plate spacing and the length of the heat exchangers. The
parameters involved in the optimisation process of the TWTC are
the length and the hydraulic radius of the regenerator, cooler
diameter, the lengths of thermal buffer tube and the heat ex-
changers, and the compliance volume. The optimisation process
stops when the optimal configuration of both SWTE and TWTC is
achieved giving the maximum overall efficiency of cooling. Table 2
shows the investigated ranges of parameters during the SWTE and
TWTC parametric optimisation. The obtained values are shown in
the penultimate column. For comparison, the last column shows
optimised values for the demonstrator device described by Saechan
& Jaworski [6]; where the resonator diameter (161.5 mm) was fixed.
However, some parameters are kept constant in this work due to
the limitations of the manufacturing and cost effectiveness of the
design. These parameters are input power at SWTE, plate thickness
of the stack and the solid temperature of both ambient heat ex-
changers. The 600 W of input power at SWTE is employed in the
model which is comparable to the heat supplied by the anticipated
biomass combustion process. Moreover, the plate thickness of
0.2 mm in the stack is the thinnest in the market at a reasonable
price. Although a thinner stack plate thickness may yield a higher
efficiency, one must take into account reasonable costs of the
thermoacoustic device as one of the main objectives of the current
design. The solid temperature of ambient heat exchangers in both
(17) SWTE and TWTC are assumed 300 K which corresponds to the
cooling fluid temperature. To avoid the effect of the entrance losses,
the porosity of the heat exchangers is assumed to be equal to that
for the stack [9]. In addition, ideal thermal insulation around the
thermoacoustic device is assumed, thus the heat interactions with
(18)
exchangers.
As has been reported by Swift [1]; the optimal length of the heat
exchanger can be determined from the peak-to-peak displacement
amplitude of the gas at the heat exchanger location, corresponding
to the distance within which the working fluid transfers heat be-
tween the stack and heat exchanger. This guideline is taken into
(19) account in the optimised simulation. In summary, the assumptions
adopted in the present study can be indicated in Table 3.
5. Results
(20)
Fig. 3 illustrates the nature of solutions obtained from DeltaEC
software. Here, as examples, the pressure and volumetric velocity
amplitude distributions along the system are shown for the final
optimised case. Overall, the system exhibits a half-wavelength
behaviour; the velocity nodes are at the ends and the pressure
node is located in the centre. There is a pressure drop along the
stack of SWTE caused by the flow resistance of the stack. At the
“union” (junction) point near the front of the TWTC (where iner-
tance branch separates from TBT “trunk”), the oscillatory volu-
metric velocity separates into two components. In order to balance
the pressure drop across the TWTC, the larger component is pre-
sent at the inlet of the inertance tube due to a low resistance while
the smaller component is present at the inlet to the TBT as the
regenerator present in this branch constitutes a higher flow
obstruction. This effect can be seen on the pressure amplitude
distribution as a pressure drop at the regenerator.
Further discussion will focus on the influence of the individual
parameters affecting the cooling performance of the TWTC driven
by the SWTE. As has been investigated by many researchers, the
geometries of the stack in SWTE and regenerator in TWTC are a
major consideration in a thermoacoustic system that impacts its
effectiveness. However, the supplementary constituents such as the
heat exchangers, the resonator, the thermal buffer tube, feedback/
inertance tube and the compliance cavity are also critical for sus-
taining an efficient flow of acoustic power. Thus, these parameters
are investigated within the optimisation process of the cooler’s
performance. The optimisation analysis is split into sections con-
cerned with SWTE and TWTC subsystems for ease of explanations.
 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610 605

Table 2
Investigated parameters of SWTE and TWTC.

Component Parameter (unit) Range investigated Optimised value (current) Values from Saechan & Jaworski [6]

Standing-Wave Engine (SWTE)

Hot heat exchanger (HHX) Length (mm) 10e50 23.03 18
Plate spacing (mm) 1e10 1.46 1.7
Stack Length (mm) 50e500 370 226
Plate spacing (mm) 0.3e1.2 0.52 0.52
Ambient heat exchanger (AHX) Length (mm) 20e100 62.41 55
Plate spacing (mm) 1e10 1.15 4.5
Diameter of the SWTE Length (mm) 80e170 93 161.5 (fixed)
Travelling-Wave Cooler (TWTC)
Cold heat exchanger (CHX) Length (mm) 5e15 11.4 16.2
Plate spacing (mm) 0.5e5 0.5 0.6
Regenerator Length (mm) 10e70 26 30
Hydraulic radius (mm) 10e60 40.71 38.36
Ambient heat exchanger (AHX) Length (mm) 5e15 12.4 9.2
Plate spacing (mm) 0.5e5 0.5 0.5
Thermal buffer tube Length (mm) 10e120 68 335
Compliance volume Length (mm) 10e250 168 199
Diameter of the TWTC Length (mm) 50e150 79.6 125

Numerical values in bold are the final figures obtained from current optimization.

Table 3
The assumptions implemented in the present model.

Parameters Values

Input power at SWTE 600 W

Plate thickness of the stack 0.2 mm
Cooling temperature of CHX of TWTC 250 K
Solid temperature of AHXs 300 K
Porosity of the heat exchangers equal to the porosity of the stack/regenerator
Length of the heat exchangers peak-to-peak displacement amplitude of the gas

Fig. 4. Effects of SWTE geometric parameters on the overall efficiency of cooling.

sensitivity of the cooling efficiency of the entire system to each

Fig. 3. Pressure (a) and volumetric velocity (b) amplitude distribution along the
system.
5.1. Optimisation of SWTE
The variation of the cooling efficiency against the dimensionless
geometry which is normalised by its optimised dimension is pre-
sented in Fig. 4. The optimal geometry ratio equals unity at the
maximum cooling performance. Thus, this represents the
parameter of SWTE. As can be seen, the cross-sectional area of the
SWTE (or its diameter) and plate spacing of the stack have the most
significant effect on the overall efficiency of cooling. The diameter
of the SWTE which is the outer border of the feedback inertance
tube has a significant effect on acoustic power flow in TWTC. This
will be discussed later. As is well known, the porosity of the stack is
the key factor in the acoustic power generation, and this porosity is
also represented in terms of stack plate spacing and will be dis-
cussed later on. In contrast, the cooling efficiency is almost insen-
sitive to the plate spacing variation of heat exchangers. The details
606 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610

of the effects of each parameter on the efficiency of the cooler will the calculated volumetric velocity amplitudes at the AHX and HHX
be discussed below. which are 4.4331  10
2 m3/s and 1.6358  10
2 m3/s, respectively.
As has been reported in the available literature, the most The effect of the length and porosity of stack on the cooling
important part of the thermoacoustic engine is the stack, here efficiency is examined as presented in Fig. 5. Here, the overall ef-
composed of multiple plates aligned parallel to the resonator tube. ficiency of cooling is presented as a contour plot with respect to the
The most critical parameter of the stack is its hydraulic radius (rh ) plate spacing in the stack and the stack length. As has been dis-
which is the ratio of the cross sectional area and the perimeter of cussed, low porosity of stack (small spacing) results in increased
the channel or the ratio of the gas volume to the gas-solid contact heat transfer, but yields unnecessarily high friction loss. On the
area. In parallel plate stack design, the hydraulic radius is equiva- other hand, acoustic power generated depends on stack length,
lent to half the plate spacing (y0 ) [12]. The plate spacing establishes while the friction loss is also proportional to this length. The
the degree of thermal interaction between the gas flowing between combination of these effects leads to the optimal porosity and
the plates and the stack material which is the kind of interaction length of stack in the current design of 0.72 and 0.37 m, respec-
that needs to be optimised to obtain the maximum performance. If tively. Of course, the optimal porosity of 0.72 corresponds to
the gaps are too narrow, the viscous effects will cause the gas to optimal plate spacing of 0.52 mm (cf. Fig. 5).
lose too much energy to friction. If the gaps are too large, the gas
cannot effectively transfer heat to and from the solid material in the 5.2. Optimisation of TWTC
stack. From the current optimisation procedure, the optimal plate
spacing is about 0.52 mm in the design discussed. Due to the par- Similarly, as in the SWTE analysis, the significant parameters of
allel plate geometry, the porosity of the stack can also be estimated cooler affecting the cooling performance are investigated in terms
directly. With the given plate spacing, the stack void fraction cor- of the sensitivity as illustrated in Fig. 6. Again, the cooling efficiency
responds to 0.72. is extremely sensitive to the cross sectional area and the hydraulic
The next parameter to be considered is the length of the stack. radius of regenerator, while the change of thermal buffer tube
Judging from equation (12), the effect of the length of the stack Dx length has a small effect on cooling performance. From this, it can
on the cooling efficiency would be a direct proportionality as the be clearly inferred that the areas of SWTE and TWTC, and the po-
increased stack length enhances the acoustic power generation. rosities of stack and regenerator are the vital parts in the simula-
However, when the stack is too long, there will be a significant tions of the coaxial thermoacoustic cooler driven by the
dissipation, which also reduces the overall efficiency of cooling. thermoacoustic engine and that a careful design must be under-
Moreover, there is another limitation to the growth of the stack taken due to the sensitivities found. The influence of individual
length. It is the normalised temperature gradient (G) as presented factors on the cooling efficiency will be explained in what is to
in equation (13). In the design of an engine (prime mover), the follow.
value of G must be higher than unity for the acoustic power to be The increase of regenerator length enhances the acoustic power
generated within the stack. consumption and reduces the temperature gradient across the
The cross sectional area of the engine is the total area including regenerator. The overall efficiency of cooling decreases rapidly as
the area of gas and solid. The generated acoustic power from the the regenerator length is less than 15 mm. This might be due to the
stack is also directly proportional to the cross sectional area as fixed temperature gradient across the regenerator as the design
implied in the term of effective circumference around stack, P, in target. Thus, as the regenerator length decreases, the acoustic
equation (12). Due to the assumed constant heat input into the consumption also decreases resulting in smaller cooling capacity.
engine, the increase of the cross sectional area is limited to produce
the maximum acoustic power. In the current design, the suitable
cross sectional area for the engine is 0.0068 m2.
Heat exchangers are the necessary components in the ther-
moacoustic device to supply and extract heat at the ends of the
stack. Here, the critical parameters of the heat exchangers will be
discussed including the length, porosity and the plate spacing of the
heat exchangers. Firstly, the optimum length is determined from
the movement of gas at the heat exchanger location from the peak
to peak displacement amplitude (equal to 2jU1 j=u). It approximates
the axial distance of a gas parcel moving in one cycle of oscillation
which is proportional to the volumetric velocity amplitude at the
given frequency. Tijani [9] recommends that the porosity of the
heat exchangers should be equivalent to the porosity of the stack to
avoid the entrance problems of the gas when leaving the stack and
entering the heat exchanger or vice versa. Therefore, here the
porosity of 0.72 has been applied to match the porosity of the stack.
Therefore, only the plate spacing of heat exchangers is varied to
evaluate the maximum heat transfer performance in this simula-
tion. The effect of the plate spacing of heat exchangers on acoustic
power is similar to that of plate spacing in the stack. Fig. 4 presents
the variation of efficiency as a function of the plate spacing of the
HHX and AHX. The maximum efficiency values are obtained at the
plate spacing of the AHX and HHX of 1.15 mm and 1.46 mm,
respectively. In addition, the optimum lengths of the AHX and HHX
determined from the peak to peak displacement amplitudes are
62.41 mm and 23.03 mm, respectively. These values correspond to Fig. 5. Contour plot of the effect of the stack length and the stack plate spacing on the
overall efficiency of cooling.
 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610
Fig. 6. Effects of TWTC geometrical parameters on the overall cooling efficiency.
The effect of hydraulic radius of the regenerator on the cooling
performance is represented in terms of the dimensionless Lautrec
number (NL ) [42] which is the ratio between hydraulic radius (rh )
and thermal penetration depth (dk ); NL ¼ rh =dk . This number in-
dicates the level of the thermal contact between working fluid and
solid. If NL is low, the perfect thermal contact between working
fluid and solid wall will be realised. However, then the pressure
drop across the regenerator increases considerably, which results in
a low volume flow rate amplitude.
According to results in Fig. 7, the maximum cooling efficiency
obtained is approximately 22.2% which is found at NL ¼ 0.37 and
the regenerator length of 26 mm. To maintain the high efficiency, a
longer regenerator length should be designed with a larger hy-
draulic radius mesh screen. Interestingly, at the left-top corner of
the contour plot (Lreg < 0.02 and NL > 0.25), the cooler performance
drops significantly in response to even very small changes in these
Fig. 7. Contour plot of the effect of hydraulic radius and regenerator length on the
overall cooling efficiency.
607
variables. Therefore, this “pitfall zone” should be avoided in the
design and the regenerator length range of 22e35 mm with the
mesh screen of hydraulic radius 40e50 mm is recommended in the
current design.
The effect of the cooler cross sectional area and thermal buffer
tube length on the cooling efficiency is examined as presented in
Fig. 8. Here, cooling efficiency is presented as a contour plot with
respect to the ratio of the regenerator holder cross sectional area to
the total cross sectional area and the length of the buffer tube. The
acoustic power consumption of the regenerator is directly pro-
portional to the cross sectional area [35]. However, the cooling
capacity of the TWTC has to be a compromise between the area of
the cooler and inertance tube (i.e. the annular space) to obtain the
maximum cooling capacity because the feedback inertance space is
also needed to return the acoustic power to the cooler and adjust
the phase difference between the oscillating pressure and volu-
metric velocity to operate the reversed Stirling cycle in the TWTC.
The thermal buffer tube is located between the cold heat
exchanger and the flow straightener of the TWTC and used as a
thermal cushion. The temperature of air then rises gradually from
the cooling temperature to mean temperature along the thermal
buffer tube. According to the simulation results in Fig. 8, the
maximum cooling efficiency is obtained at the thermal buffer tube
length of 68 mm and the cross sectional area ratio of 0.76. This
fraction implies that the area of the feedback inertance tube is 24%
of the total cross sectional area. Moreover, it can be seen that the
variation of thermal buffer tube length has a smaller effect on the
cooling efficiency than the cross sectional area ratio. Again, in the
top region of the contour plot, the cooling efficiency reduces rapidly
due to the smaller free flow area of the feedback tube leading to
enormous viscous losses. In the current coaxial thermoacoustic
cooler design, the recommended range for the thermal buffer tube
length is 60e80 mm and 74e77% of total area is suggested for the
regenerator area.
The compliance cavity is provided by the volume of gas enclosed
at the end of the TWTC acting as the space for the feedback acoustic
power returning to the cooler unit. The increased volume of the
compliance leads to an increase of volume flow rate amplitude
Fig. 8. Contour plot of the effect of thermal buffer tube length and cross sectional area
ratio on the overall efficiency of cooling.
608 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610
resulting in higher cooling capacity. However, a larger volume of
the compliance will give a negative effect in acoustic power dissi-
pation and reduce the acoustic power flow to the cooler. According
to the simulation results in Fig. 6, the optimal compliance volume is
around 1150 cm3.
In the design of thermoacoustic devices, there are many
geometrical solutions to achieve the workable devices as shown by
the results above, but only one solution is the actual optimum to
match the given criteria. In this work, the design of thermoacous-
tically driven thermoacoustic cooler to accomplish the maximum
cooling performance at temperature of 250 K is carried out sys-
tematically leading to an optimal configuration from the point of
view of the overall cooling efficiency. The simulation results are
summarised in Table 4. The optimisation process leads to the final
model with optimised parameters presented earlier in Table 2.
It is worth comparing these results to the analogous numerical
simulations carried out by Saechan & Jaworski [6] for the device
based on a fixed 6-inch pipe size. The maximum cooling load ob-
tained from the optimal configuration was approximately 185 W at
the design solid temperature of 250 K and at the cold heat
exchanger and heat input of 2.25 kW (8.2% of the overall efficiency).
In the current numerical work, for the new resonator diameter of
93 mm, 133 W is predicted from 600 W heat input (22.2% overall
efficiency) which means an almost three-fold improvement in ef-
ficiency, whilst reducing the size and weight of the device. On the
other hand, re-running the code for the device based on 6-inch
resonator but for the heat input of only 600 W gives the cooling
load of 30.2 W at the solid temperature of 250 K.
Furthermore, the direction of the acoustic power flow within
the modelled device is presented in more detail to better under-
stand the performance of the system. The variation of acoustic
power and total power flowing through the system at the condition
of maximum cooling efficiency can be illustrated by a Sankey dia-
gram shown in Fig. 9. The value of the total power flow is propor-
tional to the magnitude of the arrows. This quantitative energy flow
analysis allows to gain insight into the effect of each component on
the performance of the system. Starting at the SWTE, 600 W of
thermal power is supplied to the HHX while 497.98 W of thermal
Table 4
Simulation results of the design of TWTC driven by SWTE.
Parameters Results
The maximum acoustic power conversion efficiency of the SWTE 17%
Acoustic pressure amplitude 0.775 bar
Drive ratio 7.75%
The cooling capacity at cooling temperature of 250 K 133 W
COP 1.9
The overall efficiency of cooling 22.2%
Fig. 9. Sankey diagram of the acoustic power flow inside the SWTE/TWTC system.
power is exhausted to the cooling water at the AHX in order to
convert this energy into the acoustic power of 107.83 W at the stack.
The net acoustic power of 102.02 W is drawn from the AHX of
SWTE. The amount of power delivered to the cooler junction drops
to 68.67 W due to the viscous losses in the long resonator, and this
power is added to 172.78 W of the “recycled” power which flows
out from the regenerator core. Thus, the acoustic power of 241.45 W
passes the inertance tube to drive the cooler. 3.87 W of this power is
lost in the inertance tube and compliance cavity. Acoustic power of
59.27 W is required by the regenerator to remove 133.05 W of heat
load at the cold temperature (250 K) to ambient heat exchanger
(300 K). Finally, the acoustic power that leaves the regenerator is
fed back and added to the acoustic power stream that comes from
the SWTE.
6. Conclusions and future work
This paper presents a detailed numerical study of the co-axial
TWTC powered by the SWTE in order to find the optimal design
of a device that could use biomass burning for generating a cooling
capability for storing vital medical supplies in remote and rural
areas of developing countries with no access to electricity grid. The
general concept of such engine/cooler coupling, which uses a
constant-diameter resonator for simplicity of construction, has
been practically demonstrated [6], but no systematic study of all the
important geometrical parameters has been ever performed in a
holistic approach. Current paper addresses this important gap by
multi-variable modelling based on the linear thermoacoustic the-
ory through the use of DeltaEC.
As discussed in the previous sections, a large selection of
geometrical parameters was considered; however, the “overall
cooling efficiency” is clearly most sensitive to the diameters of
SWTE and TWTC and/or their ratio. It is interesting to note that
although the re-designed system has a smaller diameter (93 mm, cf.
Table 2) compared to the previous prototype (6 inch pipe), the
predicted acoustic power is far higher. Of course, it is unreasonable
to expect a bespoke resonator design of 93 mm internal diameter.
However, there are standard pipe sizes that are close enough to the
optimised value; for instance, a standard pipe size of 3.5 inch gives
the internal pipe diameter of 90.1 mm or 95.5 mm for schedules 40
or 10, respectively.
Using air at 10 bar as working medium and 50 Hz as operating
frequency, the cooling temperature of 250 K at TWTC and input
power of 600 W at the hot heat exchanger of SWTE have been set as
the targets of present design. According to the optimisation results,
the drive ratio of 7.75% can be produced by the SWTE with the
thermal-to-acoustic energy conversion efficiency of 17.0%. The
cooling load of 133 W can be extracted from the optimised system
which is equivalent to COP of 1.9. The overall efficiency of cooling
can reach the maximum of 22.2%.
The findings from the present study provide an important
illustration of principles behind the design process of the coupled
engine-cooler thermoacoustic system. Clearly, when the parame-
ters of both the engine and cooler are optimised in a holistic
manner, the complexity of calculations increases significantly
because of an increasingly multi-dimensional parameter space.
However, the design rewards are also higher because much higher
predicted efficiencies are possible in a geometrically more compact
layout.
Future work on the concept outlined here must focus on two
aspects: firstly, solving a range of manufacturing/engineering is-
sues and subsequent build and test of the enhanced prototype;
secondly, exploring the viability of such future product in terms of
cost and scale. The first aspect can be illustrated by the issues
related to the heat exchanger design. These have been modelled in
 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610
the simulations as parallel-plate configurations for simplicity, but
the functionality of the actual prototype will inevitably require
some design variations and compromises. On the TWTC side, it is
envisaged that the parallel-plate heat exchangers can be easily
replaced by fin-and-tube design closely related to parallel-plate
type. The ambient heat exchanger of this type has already been
successfully applied by Saechan & Jaworski [6] with water flowing
inside the tubes. For the cold heat exchanger, it is envisaged to
utilise heat pipes (cf. [43,44]) as tubes in the fin-and-tube heat
exchanger design. These would protrude outside the resonator into
a suitable “cool box” placed outside the resonator.
On the SWTE side, the ambient heat exchanger can also be
conveniently handled through the application of an appropriate
fin-and-tube design with running water in the tubes. However, the
biggest challenge seems the HHX, in particular from the point of
view of employing combustion gases as the heat source. As shown
in Table 2, the optimisation process leads to hot heat exchanger fins
of circa 23 mm length and spacing of 1.46 mm (which leads to plate
thickness 0.57 mm using the same porosity value as in the stack).
This would lead to difficulties of manufacturing fin plates with very
small spacing. Fortunately, the optimisation process also suggests
(Fig. 4) that the sensitivity of the overall cooling efficiency to the
spacing in the HHX is negligible, and so for example multiplying the
plate thickness and plate spacing by a factor of 3 would lead to
1.71 mm thick plates with 4.38 mm spacing giving reasonable
manufacturability. This would lead to having around 15 fins on the
resonator diameter of 93 mm. It is envisaged at this stage that the
heat input could be delivered using fins welded externally onto the
resonator (the flue gasses flowing through this arrangement)
through a heat conduction process across the resonator wall and
into the internally welded fins of the hot heat exchanger inside the
resonator.
To cover the second aspect (that came up in the refereeing
process), it is worth elaborating on the scale and cost of the target
practical/real system. It should be remembered here that previous
work already outlined the minimum cooling load needed for safe
storage of medical supplies. This was defined as 30 W over a duty
cycle of 2 h in 24-h period (cf. Patcharin & Jaworski [6]), and was
already exceeded by the previous demonstrator rig (120 W at the
expense of over 2 kW of thermal input). However, the previous rig
was problematic as a starting point to develop a “product” mainly
because of size and resulting mass e approximately 60 kg.
Reducing the size, while increasing overall efficiency has been
precisely the purpose of current optimisation study: the size ob-
tained would lead to a mass of approximately 30 kg; at the same
time, the performance optimisation leads to a similar cooling load
as before (133 W) at an expense of 600 W of thermal input. In this
particular application, one of the highest costs is that of the steel
pressure enclosure (typically around 1.25 USD/kg) and the overall
mass reduction by a factor of two is important.
Of course, the pressure enclosure is not the only component.
One must also think about fabrication and component costs for the
remainder of the appliance. Here, a clear advantage is the use of
local labour in rural communities of developing countries who have
vested interest in developing cooler systems proposed. Overall, it is
projected that an appliance as described in this paper would not
exceed 100 US dollars in total cost. It should also be mentioned that
such costing exercises were already conducted under the umbrella
of the “parent-project” SCORE (Stove for COoking, Refrigeration and
Electricity supply, funded by EPSRC UK). For comparison, a device
based on thermoacoustic effect generating small amounts of elec-
tricity (ca. 20 W) for the rural communities has already been
described by Abdoulla-Latiwish et al., [45]. Here the target cost was
around 50 US dollars.
609
Acknowledgments
Artur J. Jaworski would like to thank EPSRC (UK) for funding
under grants EP/E044379 and EP/R023328; and would also like to
thank Royal Society (UK) for the funding under Industry Fellowship
(ref. IF110094). Partners in the SCORE consortium: University of
Nottingham, City, University of London, Queen Mary University of
London and Practical Action are acknowledged for useful collabo-
rations. Dr Mao and Dr Yu are acknowledged for useful discussions
with Dr Saechan in relation to some of the technical content.
References
[1] G.W. Swift, Thermoacoustic engines, J. Acoust. Soc. Am. 84 (1988) 1146e1180.
[2] S.L. Garrett, J.A. Adeff, T.J. Hofler, Thermoacoustic refrigerator for space ap-
plications, J. Thermophys. Heat Tran. 7 (4) (1993) 595e599.
[3] H. Babaei, K. Siddiqui, W.A. Chishty, Sustainable thermoacoustic refrigeration
system for gas turbine power plants, in: 17th Symposium of Industrial
Application of Gas Turbines (IGAT), Banff, Canada, 2007.
[4] D.L. Gardner, C.Q. Howard, in: Waste-heat-driven Thermoacoustic Engine and
Refrigerator Proceedings of ACOUSTICS, Adelaide, Australia, 2009.
[5] J.A. Adeff, T.J. Hofler, Design and construction of a solar powered thermoa-
coustically driven thermoacoustic refrigerator, J. Acoust. Soc. Am. 107 (2000)
37e42.
[6] P. Saechan, A.J. Jaworski, Thermoacoustic cooler to meet medical storage
needs of rural communities in developing countries e high pressure system,
Therm. Sci. Eng. Prog. 8 (2018) 31e46 (December).
[7] M.E.H. Tijani, S. Spoelstra, Study of a coaxial thermoacoustic-Stirling cooler,
Cryogenics 48 (1) (2008) 77e82.
[8] T.J. Hofler, Thermoacoustic Refrigeration Design and Performance, Ph.D.
thesis, University of California, 1986.
[9] M.E.H. Tijani, Loudspeaker-driven Thermo-Acoustic Refrigeration, Ph.D. thesis,
Eindhoven University of Technology, 2001.
[10] Y. Ueda, T. Biwa, T. Yasaki, U. Mizutani, Construction of thermoacoustic stir-
ling cooler, Physica B 329e333 (2003) 1600e1601.
[11] T. Yazaki, T. Biwa, A. Tominaga, A pistonless Stirling cooler, Appl. Phys. Lett. 80
(1) (2002) 157e159.
[12] M.E.H. Tijani, J.C.H. Zeegers, A.T.A.M. de Waele, The optimal stack spacing for
thermoacoustic refrigeration, J. Acoust. Soc. Am. 112 (1) (2002) 128e133.
[13] M.E.H. Tijani, J.C.H. Zeegers, A.T.A.M. de Waele, Prandtl number and ther-
moacoustic refrigerators, J. Acoust. Soc. Am. 112 (1) (2002) 134e143.
[14] X. Xie, Q. Li, Z. Li, Q. Li, Investigation on the different types of resonators in the
thermoacoustic Stirling prime mover, Ultrasonics 44 (2006) e1503ee1505.
[15] I. Paek, L. Mongeau, J.E. Braun, S.Y. Kang, Phase of acoustic impedance and
performance of standing wave thermoacoustic coolers, J. Mech. Sci. Technol.
23 (5) (2009) 1476e1484.
[16] M. Pierens, J.P. Thermeau, T. Le Polles, P. Duthil, Experimental characterization
of a thermoacoustic travelling-wave refrigerator, in: International Conference
on Fluid Mechanics, Heat Transfer and Thermodynamics (ICFMHTT 2011),
2011, pp. 13e15. July 2011, Amsterdam, Netherlands.
[17] W. Dai, E. Luo, Y. Zhang, H. Ling, Detailed study of a traveling wave ther-
moacoustic refrigerator driven by a traveling wave thermoacoustic engine,
J. Acoust. Soc. Am. 19 (5) (2006) 2686e2692.
[18] B. Yu, E.C. Luo, S.F. Li, W. Dai, Z.H. Wu, Experimental study of a
thermoacoustically-driven traveling wave thermoacoustic refrigerator, Cryo-
genics 51 (2011) 49e54.
[19] W. Dai, E. Luo, J. Hu, H. Ling, A heat-driven thermoacoustic cooler capable of
reaching liquid nitrogen temperature, Appl. Phys. Lett. 86 (2005), 224103.
[20] L.M. Qiu, D.M. Sun, W.L. Yan, P. Chen, Z.H. Gan, X.J. Zhang, G.B. Chen, Inves-
tigation on a thermoacoustically driven pulse tube cooler working at 80 K,
Cryogenics 45 (2005) 380e385.
[21] C. Wang, G. Thummes, C. Heiden, A two-stage pulse tube cooler operating
below 4 K, Cryogenics 37 (3) (1997) 159e164.
[22] L. Li, Z. Wu, J. Hu, G. Yu, E. Luo, W. Dai, A novel heat-driven thermoacoustic
natural gas liquefaction system. Part I: coupling between refrigerator and
linear motor, Energy 117 (2) (2016) 523e529.
[23] G.W. Swift, J.J. Wollan, Themoacoustics for liquefaction of natural gas, Gas TIPS
8 (4) (2002) 21e26.
[24] P.S. Spoor, G.W. Swift, Thermoacoustic separation of a He-Ar mixture, Phys.
Rev. Lett. 85 (8) (2000) 1646e1649.
[25] M.E. Poese, R.W.M. Smith, S.L. Garrett, R. v. Gerwen, P. Gosselin, Thermoa-
coustic Refrigeration for Ice Cream Sales, International Institute of Refrigera-
tion 6th Gustav Lorentzen Conference - Natural Working Fluids, Glasgow, UK,
2004.
[26] E.M. Sharify, S. Hasegawa, Traveling-wave thermoacoustic refrigerator driven
by a multistage traveling-wave thermoacoustic engine, Appl. Therm. Eng. 113
(2017) 791e795.
[27] A.C. Alcock, L.K. Tartibua, T.C. Jen, Experimental investigation of an adjustable
thermoacoustically-driven thermoacoustic refrigerator, Int. J. Refrig. 94 (2018)
71e86.
[28] M. Wetzel, C. Herman, Design optimization of thermoacoustic refrigerators,
610 P. Saechan, A.J. Jaworski / Renewable Energy 139 (2019) 600e610
Int. J. Refrig. 20 (1) (1997) 3e21.
[29] H. Ishikawa, P.A. Hobson, Optimization of heat exchanger design in a ther-
moacoustic engine using a second law analysis, Int. Commun. Heat Mass
Transf. 23 (3) (1996) 325e334.
[30] L. Zoontjens, C.Q. Howard, A.C. Zander, B.S. Cazzolato, Modelling and opti-
misation of acoustic inertance segments for thermoacoustic devices, in: Pro-
ceedings of Acoustics 2006, Christchurch, New Zealand, 2006.
[31] Y. Ueda, M. B. M, T. Kohei, A. Atsushi, Optimization of the regenerator of a
traveling-wave thermoacoustic refrigerator, J. Appl. Phys. 107 (3) (2010),
034901.
[32] M.M. Bassem, Y. Ueda, A. Akisawa, Design and construction of a traveling
wave thermoacoustic refrigerator, Int. J. Refrig. 34 (4) (2011) 1125e1131.
[33] K. Ghorbanian, M. Karimi, Design and optimization of a heat driven ther-
moacoustic refrigerator, Appl. Therm. Eng. 62 (2) (2014) 653e661.
[34] H. Babaei, K. Siddiqui, Design and optimization of thermoacoustic device,
Energy Convers. Manag. 49 (2008) 3585e3598.
[35] G.W. Swift, Thermoacoustics: a Unifying Perspective for Some Engines and
Refrigerators, Acoustical Society of America through the American Institute of
Physics, 2002.
[36] A. Piccolo, Design issues and performance analysis of a two-stage standing
wave thermoacoustic electricity generator, Sustain. Energy Technol. Assess.
26 (2018) 17e27.
[37] B. Ward, J. Clark, G.W. Swift, Design Environment for Low-Amplitude Ther-
moAcoustic Energy Conversion e DeltaEC Version 6.4b2.7, Los Alamos
National Laboratory, New Mexico, USA, 2017.
[38] P.H. Ceperley, A pistonless Stirling engine-the travelling wave heat engine,
J. Acoust. Soc. Am. 66 (1979) 1508e1513.
[39] Y. Huang, E. Luo, W. Dai, Z. Yu, A traveling wave thermoacoustic refrigerator
within room temperature range, Cryocooler 13 (5) (2004) 189e194.
[40] A. Al-Kayiem, Z. Yu, Numerical investigation of a looped-tube travelling-wave
thermoacoustic engine with a bypass pipe, Energy 112 (2017) 111e120.
[41] J.Y. Hu, E.C. Luo, L.M. Zhang, Y.Y. Chen, Z.H. Wu, B. Gao, Analysis of a displacer-
coupled multi-stage thermoacoustic-Stirling engine, Energy 145 (2018)
507e514.
[42] S.L. Garrett, Resource letter: TA-1: thermoacoustic engines and refrigerators,
Am. J. Phys. 72 (1) (2004) 11e17.
[43] B. Delpech, M. Milani, L. Montorsi, D. Boscardin, A. Chauhan, S. Almahmoud,
B. Axcell, H. Jouhara, Energy efficiency enhancement and waste heat recovery
in industrial processes by means of the heat pipe technology: case of the
ceramic industry, Energy 158 (2018) 656e665.
[44] H. Jouhara, S. Almahmoud, A. Chauhana, B. Delpech, G. Bianchi, S.A. Tassou,
R. Llera, F. Lago, J.J. Arribas, Experimental and theoretical investigation of a flat
heat pipe heat exchanger for waste heat recovery in the steel industry, Energy
141 (2017) 1928e1939.
[45] K.O.A. Abdoulla-Latiwish, X. Mao, A.J. Jaworski, Thermoacoustic micro-
electricity generator for rural dwellings in developing countries driven by
waste heat from cooking activities, Energy 134 (2017) 1107e1120.