Professional Documents
Culture Documents
PII: S0140-7007(13)00196-5
DOI: 10.1016/j.ijrefrig.2013.07.018
Reference: JIJR 2571
Please cite this article as: Ouadha, A., Baghdad, M., Addad, Y., Effects of variable thermophysical
properties on flow and energy separation in a vortex tube, International Journal of Refrigeration (2013),
doi: 10.1016/j.ijrefrig.2013.07.018.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Energy separation and flow within a vortex tube were studied numerically. > A 3D computational domain was
generated considering the quarter of the geometry. > The flow predictions are based upon the RSM model. > Air
is selected as the working fluid with temperature-dependent thermophysical properties.> The mean pressure and
temperature of the cold and hot exits are used in the exergy analysis.
PT
RI
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
PT
Technologie Mohamed BOUDIAF d’Oran, Oran El-Mnouar, 31000 Oran, Algérie
b
Khalifa University of Science, Technology and Research, P.O.Box 127788, Abu Dhabi, UAE
RI
SC
Abstract
U
A numerical study has been conducted to investigate the effects of variable fluid properties on the prediction of a
AN
basic tube vortex design. Beforehand, a literature review is presented to highlight some of the recent advances in the
enhancements of the device design and its efficiency. The three-dimensional computations with constant and variable
M
properties revealed that the constant thermophysical assumption might not have a dramatic effect if the aim is to
predict global values only, but extra caution should be taken for an in-depth flow assessment. The exergy analysis
D
conducted suggests that the highest exergy efficiency, for the current device design, ranges from 38% to 46%
depending on the inlet pressure value. Based on the current numerical analysis; rather large exergy losses are due to
TE
irreversibility occurring at either; the lowest or the highest cold mass fraction boundary conditions.
Nomenclature
C
A Area (m2)
AC
Ex Exergy (W)
*
Corresponding author. Tel.: +213 6 61204325. Fax : +213 41290466
E-mail address: ah_ouadha@yahoo.fr (A. Ouadha)
1
ACCEPTED MANUSCRIPT
p Pressure (Pa)
PT
r Radial distance from axis
T Temperature (K)
RI
ui Absolute fluid velocity component in direction xi
xi Cartesian coordinate (i = 1, 2, 3)
SC
∆Tch Temperature difference between cold and hot end
Greek symbols
γ Adiabatic exponent
U
ξ Cold gas fraction
η Efficiency
AN
λ Thermal conductivity (W m-1 K-1)
M
µ Dynamic viscosity (kg m−1 s−1)
Subscripts
o Environmental state
c Cold gas
EP
ex Exergetic
h Hot gas
C
k Kinetic
o Outlet
p Potential
ph Physical
s Static
t Total
2
ACCEPTED MANUSCRIPT
1. Introduction
Vortex tubes are an attractive means for producing cold and hot streams from a compressed gas. Due to their
simplicity, ease of manufacturing and their potential low cost, vortex tubes could become viable candidates for partly
replacing the conventional cooling systems. In recent years, tremendous advances have been made in increasing the
efficiency of vortex tubes. Many research efforts have been focused on the optimal geometry that maximizes the
PT
COP. In particular, the number of inlet nozzles and their relative locations has been the subject of a number of
theoretical, experimental and numerical studies. For instance, some of the early proposals aiming at increasing the
RI
device performance by: increasing the number of inlet nozzles, changing the tube diameter, varying the cold and hot
outlet diameters, and/or changing the working fluid are the ones due to Deissler and Perlmutter, 1960, Linderstrbm-
SC
Lang, 1964, Yu and Tankel, 1974, Marshall, 1977, Takahama et al., 1979, Collins and Lovelace, 1979, Stephan et al.,
U
Since the beginning of the last decade, research is focussed more towards producing detailed data on the flow in
AN
different vortex tube configurations by both; experimental measurements and numerical predictions. For example,
Fröhlingsdorf and Unger (1999) simulated numerically the compressible flow and energy separation phenomena
using the CFD code CFX. They extended an axisymmetric model by integrating relevant terms for the shear-stress-
M
induced mechanical work. Ahlborn and Gordon (2000) showed that the thermal and fluid dynamics of the vortex tube
bear the signature of a classic cooling cycle. They developed simple analytical formulas for the temperature and
D
pressure profiles within the tube and compared successfully the principal model predictions to experimental
TE
measurements. Saidi and Valipour (2003) performed an experimental investigation in order to provide information
data on the classification of the parameters affecting vortex tube operation. They divided these parameters into two
EP
different types; geometrical and thermophysical ones. The Results showed that these parameters have a non
negligible influence on the cold temperature difference, thus, the efficiency of the vortex tube. Behera et al. (2005)
conducted numerically a detailed parameters analysis of a vortex tube. The velocity components and the flow patterns
C
have been evaluated using the CFD code Star-CD. Optimal design parameters of the vortex tube, such as number of
AC
nozzles, nozzle profiles, cold-end diameter, length to diameter ratio and cold and hot gas fractions, have been also
determined. Gao et al. (2005) manufactured a simple vortex tube in order to investigate pressure, temperature, and
velocity distributions using nitrogen as the working fluid. Pressure and velocity were measured using a special Pitot
tube while the temperature field was obtained using thermocouples. They reported result for different entrance
conditions and their study was further strengthened by including a thermodynamic analysis. Aljuwayhel et al. (2005)
investigated numerically the energy separation mechanism and flow phenomena within a counter-flow vortex tube. A
3
ACCEPTED MANUSCRIPT
two-dimensional axi-symmetric computational domain was used for their study. Then, the computational predictions
were compared to experimental data obtained from a laboratory vortex tube operating with room temperature
compressed air. The work also included a parametric study to investigate the effects of varying the diameter and
length of the vortex tube. Promvonge and Eiamsa-ard (2005) analysed the effects of the number of inlet nozzles, the
cold-end diameter, and the tube insulations on the temperature reduction and the isentropic efficiency of the vortex
PT
tube. Skye et al. (2006) presented a comparison between the performance predicted by numerical analysis and
experimental measurements using a commercially available vortex tube. In their study, they considered a two-
RI
dimensional steady axisymmetric computational domain. They presented results obtained with both; the standard and
renormalization group (RNG) k-ε turbulence models. Wu et al. (2007) used three innovative technologies in order to
SC
improve the energy separation and the efficiency of vortex tubes. A new nozzle with equal gradient of Mach number
and a new intake flow passage of nozzles with equal flow velocity have been designed and developed to reduce the
U
flow loss. This newly invented diffuser by the authors has been installed to reduce friction loss of air flow energy at
the hot tube-end, which has been shown to greatly improve the vortex tube performance. Eiamsa-ard and Promvonge
AN
(2007) presented a numerical analysis of flow field and temperature separation in a uni-flow vortex tube type. In
particular, they studied the effects of the turbulence modelling (k–ε model and ASM), effects of numerical schemes
M
(hybrid, upwind and second-order upwind) and grid density; on the calculation of energy separation in the vortex
tube. They argued that the use of the ASM improves slightly the accuracy of the predictions in comparison with those
D
obtained with the k–ε model. Farouk and Farouk (2007) used the large eddy simulation (LES) approach to predict the
TE
flow and temperature fields in a vortex tube. The temporal evolutions of the axial, radial and azimuthal components
of the velocity along with the temperature, pressure and density fields in the RHVT have been simulated. The
predictions from the numerical model were compared with published experimental results and k–ε model predictions.
EP
The authors found that the temperature separations predicted by the LES model was closer to the experimental
results. The LES results however under predicted the total temperature separation at the cold exit. Arjomandi and Xue
C
(2007) investigated the effect of the hot nozzle size on the performance of the Ranque–Hilsch vortex tube. Series of
AC
plugs have been used in the experiment in order to find the relationship between the diameter of hot end plug and the
performance of the vortex tube. Behera et al. (2008) generated a three-dimensional computational grid of a vortex
tube and conducted a numerical study using the CFD code Star-CD to analyse the flow parameters and energy
separation mechanism inside the tube. Computations have been conducted for different fluid properties and flow
parameters to understand the energy transfer mechanisms. Xue and Arjomandi (2008) studied the effect of the angle
of rotating flow on the performance and efficiency of a vortex tube. They showed that a smaller vortex angle resulted
in a larger temperature difference thus, a better performance for the heating efficiency of the vortex tube. However,
4
ACCEPTED MANUSCRIPT
the findings revealed that this improvement was achieved only with relatively small values of input pressure.
Rattanongphisat et al. (2008) presented a three dimensional numerical predictions using the standard k–ε model to
simulate the physical behaviour of the flow variables such as temperature and pressure inside the vortex tube. A good
agreement was obtained for the outlet temperatures predictions and the experimental data. Also Akhesmeh et al.
(2008) used the standard k-ε turbulence model with a two-dimensional axisymmetric domain to study the flow fields
PT
and the associated temperature separation within a vortex tube. Simulations have been carried out for various cold
outlet mass-flow rates and showed a reasonable agreement with experimental data. Dincer et al. (2008) studied the
RI
effect of length to diameter ratio and nozzles number on the performance of a counter-flow vortex tube using
artificial neural networks on the basis of experimental data. Farouk et al. (2009) used large eddy simulation (LES) to
SC
predict the species and temperature separation within a counter flow Ranque–Hilsch vortex tube using two-phase
nitrogen–helium mixture as a working fluid. The results presented include; temporal evolutions of all three velocity
components, the temperature distribution, the pressure, mass density, and species concentration fields within the
U
vortex domain. Over the entire cold mass fraction range very small amount of gas separation was predicted. The gas
AN
separation was observed to be in the range of 0.8 × 10−4-1.0 × 10−4 which is negligible for any practical consideration.
Dincer et al. (2009) determined experimentally the optimal position, diameter and angle of a mobile plug, located at
M
the hot outlet side in the vortex tube. Experiments have been conducted with a supply pressure varying from 200 to
380 kPa using 2, 4 and 6 nozzles. They concluded that the maximum difference in the temperatures of hot and cold
D
streams was obtained for the plug diameter of 5 mm, tip angles of 30° and 60°, 4 nozzles and by keeping the plug
TE
location at the far extreme end. Similarly, Nimbalkar and Muller (2009) carried out a series of experiments focusing
on various diameters of the cold end side for different inlet pressures and cold fractions. Eiamsa-ard et al. (2010)
studied the effects of cooling of the hot tube on the temperature separation and cooling efficiency in a counter-flow
EP
vortex tube. They found that the mean cold air temperature reduction and cooling efficiency of the vortex tube with
the hot tube cooling are respectively, 5.5 to 8.8% and 4.7 to 9% higher than those of the vortex tube without cooling.
C
AC
Despite the advantage of the exergy analysis to; identify the magnitude, the type and the cause of losses in a system,
this concept seems to have attracted less attention in previous tube vortex studies. For instance, Saidi and Yazdi
(1999) used a thermodynamic model to investigate the energy separation in a vortex tube and a Gouy-Stodola type
relation has been used to evaluate the total irreversibility. The authors developed then an approach based on the
exergy analysis method to optimise dimensions and operating conditions of the vortex tube. In agreement with the
experimental studies mentioned above, they highlighted that the temperature differences increase with increasing
inlet pressure (i.e. exergy loss decreases while increasing the inlet pressure). Increasing vortex tube length would also
5
ACCEPTED MANUSCRIPT
results in an increase of temperature differences and a significant decrease of the exergy loss. The exergy destruction
has been minimized at cold gas fraction, ξ , approaching 0.7 which means the efficient working point of this vortex
tube is at ξ = 0.7. Additionally, using materials with more smooth surfaces (to reduce friction forces) and lower
thermal conductivities (to increase wall thermal resistance) have better second law efficiency. Interestingly, it has
been reported that the second law efficiency is improved (exergy destruction decreases) with increasing the nozzle
PT
diameter. Kirmaci (2009) conducted an exergy analysis in attempt to give better understanding to the effects of the
orifice nozzles number and the inlet pressure on the heating and cooling performance of the counter-flow vortex tube.
RI
The paper reported the measurement and the exergy analysis for a cold mass flow fraction equal to 0.5. Also, Dincer
et al. (2010) carried out an experimental investigation and an exergy analysis to determine the performance of a
SC
vortex tube with various nozzle cross-section areas under inlet pressures of 260 and 300 kPa. The total inlet exergy,
total outlet exergy, total lost exergy and exergy efficiency have been reported. It has been found that the exergy
efficiency of that particular vortex tube design varies between 1 and 39%.
U
AN
It is clear from the above literature survey that the vortex tube device has received a considerable interest among the
research society during the last years and a substantial amount of work both; numerical and experimental, has been
M
published on the subject. However, the experimental work is most of the time limited to integral values which is
understandable due to the small dimensions of the device and the relatively high operating pressures. On the other
D
hand, recent developments in the CFD field, makes it now possible to examine in more detail the mechanism of
TE
thermal separation in vortex tubes. Nevertheless, in the above mentioned work, there seems to be only few attempts
to take into account the 3D nature of the flow combined with advanced RSM turbulence models. Furthermore, most
previous studies have considered constant thermophysical properties of the working fluid within the vortex tube.
EP
However, the flow within the vortex tube involves significant temperature and properties variations. Temperature-
dependent thermophysical properties of the working fluid may have significant effects on flow and heat transfer in a
C
vortex tube. It is important to show the influence of these variations on the flow and heat transfer behaviours.
AC
Hence, following the above mentioned studies, this paper presents an exergy analysis by quantifying the exergy
losses within a basic vortex tube design and its exergy efficiency as a function of the cold fraction and the inlet air
pressure. Beforehand, the predictions with a second moment closure (RSM) of Launder et al. (1975) and variable heat
capacity, viscosity, and thermal conductivity on a steady 3-dimentional turbulent flow in a vortex tube for various
inlet pressures are compared to available experimental data. Vis-à-vis the choice of the turbulence model, the RSM
model has been opted for due to its superiority in mimicking the flow behavior in comparison the two-equation type
6
ACCEPTED MANUSCRIPT
2. Mathematical model
A schematic representation of the vortex tube configuration highlighting the device dimensions and the system axis is
shown in Fig. 1. This configuration has been the subject of an experimental study conducted by Dincer et al. (2009),
followed by a numerical study presented by Baghdad et al. (2011). Air is introduced to the tube tangentially through
PT
four identical nozzles of an inlet section Ain = 2x2 mm2. The cold end diameter is dc = 5 mm and hot end is defined by
a cone-shaped valve having an angle of 30°, situated at the extreme end of the tube. According to Dincer et al. (2009),
RI
this configuration allows for the optimal temperature difference between the hot and cold streams.
SC
2.2. Governing equations
The working fluid (air) is assumed compressible and Newtonian, resulting in the continuity equation (1), Favre
U
averaged Navier–Stokes equations (2) and the energy equation (3) as follows:
∂ρ ui
=0
∂xi AN (1)
∂ ∂p ∂τ ij ∂
∂x j
( ρ ui u j ) = − + +
∂xi ∂x j ∂x j
(
− ρ ui′u ′j ) (2)
M
∂ ∂u τ ∂q
∂xi
( ρ Eu j ) = i ij − j
∂x j ∂x j
(3)
D
where ρ , ui , p and E denote density, velocity components in x, y and z directions, the static pressure, and the
TE
volumetric total energy. τ ij is the Favre-averaged Reynolds stress tensor expressed as:
∂ui ∂u j 2 ∂ui
EP
τ ij = µ + − µ δ ij (4)
∂x
j ∂xi 3 ∂xi
∂T
q j = −λ (5)
∂x j
1
ρ E = ρ e + ρ ui ui (6)
2
7
ACCEPTED MANUSCRIPT
In compressible flows, the conservation equations are associated to a law relating pressure, density and energy. For
p
= r ⋅T (7)
ρ
The specific heat capacity, thermal conductivity and dynamic viscosity are assumed to vary with temperature
PT
according to a simple power law having the general following form:
RI
The constants ai (see Table 1) are determined using a regression method based on available literature data (Borgnakke
SC
2.3. Turbulence modelling
A selection of Hi and Low-Reynolds turbulence models are available in the commercial CFD code Fluent. In the
U
present study the original Reynolds Stress Model, RSM, of Launder et al. (1975) has been selected. The latter has
AN
been found (Baghdad et al., 2011) to mimic better the experimental data compared to the classical two-equation
models, namely, standard k-ε, the k-ω and the k-ω SST models.
M
In this model, the Favre-averaged Reynolds stress tensor is related to the local velocity gradients by an eddy
viscosity, µ t , which is estimated assuming the Boussinesq approximation. Effectively, the turbulent viscosity, µ t , is
D
computed from scalar quantities determined from the transport equations specific to the turbulence model. Details
TE
related to the turbulence model are omitted from the present paper seeking briefness but these can be easily found in
RSM is a Hi-Reynolds model and requires the near-wall treatment. The wall modelling in this case is based on the so-
C
called enhanced wall treatments also known as the two-layer model. This model is activated as follows:
AC
- In the fully turbulent region ( Re y > Re*y ; Re*y = 200 , with Re y = ρ y k µ ), the RSM is used.
- In the viscosity-affected near-wall region, the one-equation model of Wolfstein (1969) is employed.
At the inlet and outlets (cold and hot), pressure values are specified. The mass flow rate at the hot outlet is prescribed
as fraction of the inlet, while the one at the cold outlet is obtained as an output from the code satisfying continuity.
This type of boundary condition should mimic exactly the experimental set up in which; as the valve position is
8
ACCEPTED MANUSCRIPT
changed, the proportions of hot and cold air change, but the total flow remains the same. Thus the amount of air
exiting the cold end can be varied from 0 to 100% of the inlet mass flow rate. The amount of this air is known as the
cold fraction:
m& c
ζ = (9)
m& i
PT
A turbulence intensity of 5% was specified at the inlet (as no information about turbulence levels were provided in
the experimental paper). Finally, in all the runs presented herein, the tube walls are considered to be adiabatic. The
RI
values of different input conditions are summarized as follows:
SC
x = y =0, z =-1 0 ξ i = mcold/min = 0.1÷0.9
U
3. Numerical procedure
the quarter of the device and using periodic conditions which should account for any secondary motion that might
D
exist. In addition, the present grid is generated using equidistant hexahedral cells to minimize all the errors associated
TE
with cells extrusion, distortion and so on. Fig. 2 shows two different views of the mesh generated.
In any numerical study, the mesh quality and resolution have a strong influence on the solution accuracy. For this
EP
purpose, sensitivity tests are carried out to guarantee a grid independent solution. The latter is assessed by the
analysis of four different multi-blocks mesh sizes of 80,000, 140,000, 200,000 and 380,000 cell volumes for the case
of an inlet pressure of 380 kPa and a cold mass fraction fixed to 0.3.
C
First, the four meshes are compared quantitatively by considering numerical uncertainties in predicting the mean
AC
temperature difference between hot and cold exits (Th ‒Tc). Table 2 shows that the mean temperature difference
values for meshes 3 and 4 are almost identical. It is noted that only -0.6 % variation in the mean temperature
difference occurs when the grid passes from 200,000 to 380,000 cells. In addition, the axial and radial velocities
plotted for the four grids considered (see Fig. 3 and 4), clearly prove a rather monotonic convergence of the velocity
profiles achieving a grid independent results with the meshes containing more than 200,000 control volume cells.
9
ACCEPTED MANUSCRIPT
Based on these above mentioned uncertainties and velocity profiles comparison; it is hence, estimated safe to use
mesh 3 for the remaining of the study as mesh 4, was observed to dramatically increase the computational costs of the
All computations are run in steady mode. The density based, implicit solver is used to solve the governing equations.
PT
Second order upwind discretization schemes are used for the convective terms in the momentum, energy, and
turbulence and the pressure-velocity coupling is ensured using the SIMPLE algorithm. The code default relaxation
RI
factors are used which might suggest that all the runs are stable and further emphasize that the grid density is
SC
4. Results and discussions
U
The results in this section are presented as follows; first, the turbulence model predictions for both constant and
AN
variable thermophysical properties are compared with experimental data, then the flow within the vortex tube is
analyzed by considering the velocity and temperature fields from all these runs. In addition, model predictions in
terms of profiles at different locations of the domain and temperature and turbulent energy contours at the mid-plane
M
of the tube are analyzed. Finally, the exergy analysis conducted involving the determination of the amount of exergy
lost within the vortex tube and its exergy efficiency is presented.
D
TE
Dincer et al. (2009) have presented their results as measurements of the mean temperature difference between the hot
and cold streams as function of the cold fraction and the pressure inlet. Accordingly and in absence of more measured
EP
parameters, the computed results are compared to experimental data in terms of the mean temperature difference
defined as:
C
∆T = Th − Tc (10)
AC
To establish the validity of the numerical model, comparisons of the computed mean temperature differences between
the hot and cold ends with those measured by Dincer et al. (2009) for a similar vortex tube are carried out. Fig. 5
shows a comparison of the mean temperature difference between the hot and cold streams temperatures as predicted
by the numerical model for four different pressure inlets ranging from 200 to 380 kPa. As shown in the figure, the
model is capable of returning reasonable and acceptable predictions within less than 5 % for the low pressure inlet
case but worsen slightly in the case of the highest inlet pressure. It is shown also that variable thermophysical
10
ACCEPTED MANUSCRIPT
properties (VTP) enhances the prediction of mean temperature difference compared to experimental measurements.
The remaining differences might well be attributed to the fact that the current numerical results are obtained for an
idealized case with smooth wall and no heat loss to the surroundings (adiabatic walls are used herein).
In order to give a detailed insight in the mechanisms that occur in the vortex tube, the latter is analysed for the
PT
specific case of a pressure inlet equal to 380 and a cold fraction equal to 0.3, which, according to the experimental
works, returns the lowest possible cold temperature. First, the air introduced in the vortex tube via the inlets
RI
positioned at a given angle of the main pipe, creates vortices in the internal flow. Fig. 6 shows that the model is
capable of predicting this flow behaviour which is expected as this is imposed mainly by the angular position of the
SC
inlet sections.
In addition, the figure illustrates the existence of two vortices, both; rotating in the same direction and with the same
U
angular velocity. The first is a free vortex that characterizes the outer regions and tends to escape through the hot exit,
AN
whereas the second one is a forced vortex representing the reverse flow that appears in the core region and moves
back toward the cold exit crossing all the hot tube and the vortex chamber centre. It can also be noticed that the
reverse flow starts moving toward the cold exit at a specific location along the main tube. Interestingly, it can be
M
observed in the figure that there is a non-negligible discrepancy between fixed and variable thermophysical properties
Figs. 7 and 8 show the two main velocity components; tangential, and axial as function of the dimensionless radial
TE
coordinate r/R at specific streamwise locations z/L=-0.04, 0.04, 0.5, and 0.93. A noticeable difference is observed
between fixed and variable thermophysical properties predictions for both; the axial and tangential velocity
EP
components in the hot tube, especially at the locations; z/L=0.04 and 0.5. In agreement with the above observations,
using constant variables would result in larger values of the tangential velocity component, while under-predicting
the axial component values at the tube centre, hence affecting the starting point strength of the inner vortex. However,
C
this difference is somewhat negligible near the cold and hot exits (z/L=-0.04 and 0.93).
AC
Figs. 9 and 10 show the distribution of static and total temperatures in the mid-plane of the vortex tube. A low-energy
zone is located around the tube axis near the inlet. The maximum air energy accumulates in the hot end annular
region. The flow of the warm fluid towards the hot exit is accompanied by an energy gain which increases
temperature. Although the total temperature difference between the outer air zone and that of the tube core is higher,
11
ACCEPTED MANUSCRIPT
the static temperature difference is not as important which could be attributed to a relatively high kinetic energy in
this region. The total temperature of a compressible fluid is calculated using the following expression:
γ −1 2
Tt = Ts 1 + M (11)
2
The static temperature approaches the total temperature in regions close to the hot exit because the velocity is reduced
PT
and the kinetic energy is converted to thermal energy by viscous dissipation and shear effects. The temperature drop
of this warm air coincides with the flow reversal point, giving its energy to outer counter-flow during its movement
RI
back toward the cold exit. Observed in the figures, rather significant differences, in terms of temperature distribution
and minimum/maximum values, are obtained between constant thermophysical properties and variable ones.
SC
Illustrated in Figure 11, the turbulent energy distribution is also considerably affected by the constant variables
assumption. It might be argued then, that these differences are at the origin of the disagreement between the global
U
temperature differences, ∆T, illustrated in Fig. 5.
the process due to irreversibilities. Exergy analysis is based on both the first and second laws of thermodynamics. It
D
aims mainly to detect and quantify the losses that occur in a real process in order to show where efforts should be
TE
focused to improve the overall efficiency of the system. The advantages of the exergy analysis are detailed in Kotas
(1995). The numerical simulation presented above, is thus, complemented by an exergy analysis.
In absence of chemical reactions, the total exergy of a system consists of three main components, namely, physical,
EP
& = Ex
Ex & + Ex& + Ex
& (13)
C
ph k p
where,
AC
& is the physical exergy due to the deviation from the environment state, calculated using:
Ex ph
& = m& ( h − h − T ( s − s ) )
Ex (14)
ph 0 0 0
& is the kinetic exergy induced by the system velocity, and is evaluated using:
Exk
& = 1 m& v 2
Ex (15)
k
2
12
ACCEPTED MANUSCRIPT
& is the potential exergy due to the system elevation, computed using:
and Ex p
& = m& g z
Ex (16)
p
where h and s are enthalpy and entropy of the system at given temperature and pressure (T, p), g is the gravitational
acceleration, z is the net height, defined as the difference between the height of the hot outlet and the height of the
PT
inlet of the vortex tube, h0 and s0 are enthalpy and entropy at the environmental temperature and pressure (T0, p0) and
RI
The exergy balance of a vortex tube can be represented in the following form using exergy values of streams entering
and leaving it, as shown in Figure 12. The exergy balance is stated around a control volume delimited by specific
SC
boundaries. The exergy flow to the control volume is always greater than that from the control volume. The
U
& − Ex
Ex & − ∆Ex
& = ∆Ex
& (17)
i o t
AN
The difference between entering and leaving exergy streams is called total exergy losses ∆Ex
& . The Total exergy loss
t
represents all kinds of internal irreversibilities in the system. They are related to the entropy generated in the system
M
by the following formula:
∆Ex
& = m& ⋅ T ⋅ ∆s (18)
D
t 0
The advantage of exergy analysis is that permits the expression of exergetic efficiency. This later can be defined as
TE
(Cornelissen, 1997):
EP
&
Ex
ηex = & o (19)
Ex i
C
Assuming air to be a perfect gas, specific air enthalpy and entropy are related to the specific heat capacity by the
AC
following relations:
h = h00 + c p (T ) dT
∫
c (T ) (20)
s = s00 + ∫ p dT − r ln ( p )
T
where h00 and s00 are the integration constants determined using an arbitrary reference state and r is the gas constant.
13
ACCEPTED MANUSCRIPT
Using the specific heat capacity given in equation (8) and after integration, relations that are used to calculate specific
a1 2 a2 3 a3 4
h = h00 + a0T + 2 T + 3 T + 4 T
(21)
s = s + a ln (T ) + a T + a2 T 2 + a3 T 3 − r ln( p)
00 0 1
2 3
PT
where constants ai are given in Table 1.
RI
i
SC
following relations:
&
Exi = m& i ( hi − h0 − T0 ( si − s0 ) ) + 2 m& i vi
1 2
U
c = m& c ( hc − h0 − T0 ( sc − s0 ) ) + m& c v c
Ex
& 1 2
2 (22)
&
Exh = m& h ( hh − h0 − T0 ( sh − s0 ) ) + m& h v h + m& h g zh
& = ξ Ex & + (1 − ξ ) Ex
1
2
2 AN
Exo c
&
h
M
& i is the inlet mass flow rate, hi is the inlet specific enthalpy, si is the inlet specific entropy, vi is the inlet
where m
D
velocity, m& c is the mass flow rate at the cold end, hc is the outlet specific enthalpy of cold end, sc is the outlet specific
& h is the mass flow rate at the cold end, hh is the outlet specific
entropy of cold end, vc is the velocity at the cold end, m
TE
enthalpy of the hot end, sh is the outlet specific entropy of the hot end, vh is the velocity at the hot end of the vortex
tube, h0 is the environment specific enthalpy, s0 is the environment specific entropy, T0 is the environment
EP
The total exergy loss of the vortex tube, difference between the inlet exergy and the total outlet exergy, is calculated
C
as follows:
AC
∆Ex
& = Ex
& − Ex
i
&
o
(23)
The exergy efficiency of the vortex tube can be evaluated using Eq. 19.
Fig. 13 highlights the amount of exergy produced at the inlet and outlet of the vortex tube as function of the inlet
pressure and the cold fraction. The inlet exergy is the sum of exergy of the four inlets, while the outlet exergy is the
sum of exergy of the cold and hot ends of the vortex tube. The increase in the inlet pressure results in an increase of
the inlet exergy but is found to be independent of the cold fraction variation. On the other hand, the outlet exergy
14
ACCEPTED MANUSCRIPT
presents a different behaviour against cold fraction values. For a given pressure, the outlet exergy follows a parabola
shape with maximum values at cold fraction ranging from 0.4 to 0.6. In addition, similarly to the inlet exergy, the
Finally the performance of the actual device design is analysed by means of total exergy losses and exergy efficiency
presented in Fig. 4 (a) and (b) respectively. At low inlet pressures, the total exergy losses seems to be less dependent
PT
on the cold fraction, ξ, while its effects start to become more apparent at larger values (pi > 320 kPa). All the curves
are characterized by having minimum values located around the centre, between ξ = 0.4 and 0.6, while reaching the
RI
maximum at both extremes. Here again, the inlet pressure values are observed to be affecting the exergy losses
magnitude. This might suggest that for the actual device design, irreversibility losses become more significant if a
SC
higher flow rate is forced toward one particular exit. In attempt to illustrate the best operating conditions for the
actual design, a dotted line is added in the figure which follows the minimum exergy losses for all inlet pressures.
Unlike the coefficient of performance (the traditional criteria measuring performance of refrigeration systems) the
U
exergy efficiency is firmly based on both the first and second laws of thermodynamics. Figure 4 (b) illustrates the
AN
exergy efficiency curves for all four inlet pressure values. As one would expect from Figure 4 (a) discussion above,
minimum exergy losses correspond to higher efficiency and vice-versa. In addition, it worth noticing here that the
M
exergy efficiency curves follow almost the same trend as the temperature differences presented in Fig. 5. Especially
the maximum points are located at the same cold fraction values. Also the curves reveal that for an inlet pressure of
D
200 kPa, the maximum exergy efficiency is only 38 %. However, the exergy efficiency reaches much higher values if
TE
the inlet air pressure is increased. For instance, the maximum exergy efficiency is as high as 46 % with an inlet
5. Conclusions
A numerical study has been carried out to investigate the energy separation mechanism, flow phenomena, and the
C
exergy efficiency, within a vortex tube using RSM turbulence model and considering variable thermophysical
AC
properties. Satisfactory predictions of mean temperature difference between the hot and cold exits are obtained in
comparison with the available experimental data (Dincer et al., 2009). The present results show that although the
variable thermophysical properties might seem to have only little to insignificant influence on the numerical
predictions (if comparison is limited to global values only); the variable properties are actually found herein to be
The exergy analysis indicates that the exergy efficiency of the actual vortex tube design ranges from 15 to 47 %. As
the exergy efficiency curves are observed to have almost the same trend regardless of the inlet pressure value, future
15
ACCEPTED MANUSCRIPT
attempts to improve in the current design should rather focus more on geometrical enhancements such as convergent
inlet channel to increase the flow velocity at the inlet, new design to increase the vortices rotation and/or round
References
PT
Ahlborn, B.K., Gordon, J.M., 2000. The vortex tube as a classic thermodynamic refrigeration cycle. J. Appl. Phys.
88, 3645−3653.
RI
Ahlborn, B.K., Camirey, J., Kellerz, J.U., 1996. Low−pressure vortex tubes. J. Phys. D: Appl. Phys. 29, 1469−1472.
Akhesmeh, S., Pourmahmoud, N., Sedgi, H., 2008. Numerical Study of the Temperature Separation in the
SC
Ranque−Hilsch vortex tube. American J. Eng. Appl. Sci. 1, 181−187.
Aljuwayhel, N.F., Nellis, G.F., Klein, S.A., 2005. Parametric and internal study of the vortex tube using a CFD
U
model. Int. J. Refrig. 28, 442−450.
Arjomandi, M., Xue, Y., 2007. An investigation of the effect of the hot end Plugs on the efficiency of the
AN
ranque−hilsch Vortex tube, J. Eng. Sci. Technology. 2, 211−217.
Baghdad, M., Ouadha, A., Imine, O., Addad, Y., 2011. Numerical study of energy separation in a vortex tube with
M
different rans models. Int. J. Thermal Sci. 50, 2377-2385.
Behera, U., Paul, P.J., Dinesh, K., Jacob, S., 2008. Numerical investigations on flow behaviour and energy separation
D
Behera, U., Paul, P.J., Kasthurirengan, S., Karunanithi, R., Ram, S.N., Dinesh, K., Jacob, S., 2005. CFD analysis and
experimental investigations towards optimizing the parameters of Ranque−Hilsch vortex tube. Int. J. Heat Mass
Borgnakke, C., Sonntag, R.E., 1997. Thermodynamic and Transport Properties, Wiley & Sons Inc., New York.
Collins, R.L., Lovelace, R.B., 1979. Experimental study of two−phase propane expanded through the Ranque−Hilsch
C
Cornelissen, R.L., 1997. Thermodynamics and sustainable development−the use of exergy analysis and the reduction
Deissler, R.G., Perlmutter, M., 1960. Analysis of the flow and energy separation in a turbulent vortex. Int. J. Heat
Dincer, K., Avci, A., Baskaya, S., Berber, A., 2010. Experimental investigation and exergy analysis of the
performance of a counter flow Ranque−Hilsch vortex tube with regard to nozzle cross−section areas. Int. J. Refrig.
33, 954−962.
16
ACCEPTED MANUSCRIPT
Dincer, K., Baskaya, S., Uysal, B.Z., Ucgul, I., 2009. Experimental investigation of the performance of a
Ranque−Hilsch vortex tube with regard to a plug located at the hot outlet. Int. J. Refrig. 32, 87−94.
Dincer, K., Baskaya, S., Uysal, B.Z., 2008. Experimental investigation of the effects of length to diameter ratio and
nozzle number on the performance of counter flow Ranque−Hilsch vortex tubes. Heat Mass Transfer. 44, 367−373.
Eiamsa−ard, S., Wongcharee, K., Promvonge, P., 2010. Experimental investigation on energy separation in a
PT
counter−flow Ranque−Hilsch vortex tube: Effect of cooling a hot tube. Int. Comm. Heat Mass Transfer. 37, 156−162.
Eiamsa−ard, S., Promvonge, P., 2007. Numerical investigation of the thermal separation in a Ranque−Hilsch vortex
RI
tube. Int. J. Heat Mass Transfer. 50, 821−832.
Farouk, T., Farouk, B., Gutsol, A., 2009. Simulation of gas species and temperature separation in the counter−flow
SC
Ranque−Hilsch vortex tube using the large eddy simulation technique. Int. J. Heat Mass Transfer. 52, 3320−3333.
Farouk, T., Farouk, B., 2007. Large eddy simulations of the flow field and temperature separation in the
U
Fröhlingsdorf, W., Unger, H., 1999. Numerical investigations of the compressible flow and the energy separation in
AN
the Ranque−Hilsch vortex tube. Int. J. Heat Mass Transfer. 42, 304−311.
Gao, C.M., Bosschaart, K.J., Zeegers, J.C.H., de Waele, A.T.A.M., 2005. Experimental study on a simple
M
Ranque−Hilsch vortex tube. Cryogenics. 45, 173−183.
Hilsch, R., 1947. The use of the expansion of gases in a centrifugal field as cooling process. The Review Sci.
D
Kirmaci, V., 2009. Exergy analysis and performance of a counter flow Ranque−Hilsch vortex tube having various
nozzle numbers at different inlet pressures of oxygen and air. Int. J. Refrig. 32, 1626−1633.
Kotas, T.J., 1985. The Exergy Method of Thermal Plant Analysis. Anchor Brendon Ltd.: Tiptree, Essex.
EP
Launder, B.E., Reece, G.J., Rodi, W., 1975. Progress in the Development of a Reynolds−Stress Turbulence Closure.
Linderstrbm−Lang, C.U., 1964. Gas separation in the Ranque−Hilsch vortex tube. Int. J. Heat Mass Transfer. 7,
AC
1195−1206.
Marshall, J., 1977. Effect of operating conditions, physical size and fluid characteristics on the gas separation
performance of a Linderstrom−Lang vortex tube. Int. J. Heat Mass Transfer. 20, 227−231.
17
ACCEPTED MANUSCRIPT
Menter, F.R., 1994. Two−Equation Eddy−Viscosity Models for Engineering Applications. AIAA J. 32, 1598−1605.
Nimbalkar, S.U., Muller, M.R., 2009. An experimental investigation of the optimum geometry for the cold end
Promvonge, P., Eiamsa−ard, S., 2005. Investigation on the vortex thermal separation in a vortex tube refrigerator.
PT
Rattanongphisat, W., Riffat, S.B., Gan, G., 2008. Thermal separation flow characteristic in a vortex tube: CFD
RI
Saidi, M.H., Allaf Yazdi, M.R., 1999. Exergy model of a vortex tube system with experimental Results. Energy. 24,
625−632
SC
Saidi, M.H., Valipour, M.S., 2003. Experimental modeling of vortex tube refrigerator. Appl. Thermal Eng. 23,
1971−1980
Singh, P.K., Tathgir, R.G., Gangacharyulu, D., Grewal, G.S., 2004. An experimental performance evaluation of
U
vortex tube. IE (I) Journal−MC. 84, 149−153
AN
Shannak, B.A., 2004. Temperature separation and friction losses in vortex tube, Heat Mass Transfer. 40, 779−785.
Skye, H.M., Nellis, G.F., Klein, S.A., 2006. Comparison of CFD analysis to empirical data in a commercial vortex
M
tube, Int. J. Refrig. 29, 71−80
Stephan, K., Lin, S., Durst, M., Huang, F., Seher, D., 1983. An investigation of energy separation in a vortex tube,
D
Takahama, H., Kawamura, H. , Kato, S., Yokosawa, H., 1979. Performance characteristics of energy separation in a
Wolfstein, M., 1969. The velocity and temperature distribution of one−dimensional flow with turbulence
EP
augmentation and pressure gradient, Int. J. Heat Mass Transfer. 12, 301−318.
Wu, Y.T., Ding, Y., Ji, Y.B., Ma, C.F., Ge, M.C., 2007. Modification and experimental research on vortex tube. Int.
C
Xue, Y., Arjomandi, M., 2008. The effect of vortex angle on the efficiency of the Ranque−Hilsch vortex tube, Exp.
Yu, D.R., Tankel, L.E., 1974. Influence of vortex−tube saturation and length on the process of energetic gas
18
ACCEPTED MANUSCRIPT
Tables
a0 a1 a2 a3
cp (J/kg K) 1015.9683 -0.12802 2.94104E-4 2.93413E-8
µ (kg/m s) -2.61819E-7 8.04006E-8 -7.04884E-11 3.50834E-14
k (W/m K) -8.07779E-4 1.04110E-4 -5.11705E-8 1.56157E-11
PT
RI
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
Table 2. Mesh sensitivity study on the mean temperature difference between hot and cold exits
Mesh Number of cell volumes Th Tc (K) Error (%)
1 80 000 63.88 21.65
2 140 000 54.75 8.58
3 200 000 49.75 -0.60
4 380 000 50.05 -
PT
RI
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
Figures
a. b.
PT
Fig. 1. Physical problem: a. schematic representation and dimensions (in mm units); b. axis system.
RI
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
PT
Fig. 2. Computational mesh of the vortex tube.
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
a. b.
1.0 1.0
Grid 4 Grid 4
Grid 3 Grid 3
0.5 Grid 2 0.5 Grid 2
Grid 1 Grid 1
0.0 0.0
r/R
r/R
PT
-0.5 -0.5
-1.0 -1.0
RI
-500-400-300-200 -100 0 100 200 300 400 500 -500-400-300-200 -100 0 100 200 300 400 500
Axial velocity (m/s) Axial velocity (m/s)
SC
c. d.
1.0 1.0
U
Grid 4 Grid 4
Grid 3 Grid 3
0.5 Grid 2 0.5 Grid 2
0.0
Grid 1
AN 0.0
Grid 1
r/R
r/R
-0.5 -0.5
M
-1.0 -1.0
-500-400-300-200 -100 0 100 200 300 400 500 -500-400 -300-200-100 0 100 200 300 400 500
D
Fig. 3. Influence of grid size on axial velocity profiles: a. z/L = -0.04; b. z/L = 0.04; c. z/L = 0.5; d. z/L = 0.93.
C EP
AC
ACCEPTED MANUSCRIPT
a. b.
1.0 1.0
Grid 4 Grid 4
Grid 3 Grid 3
0.5 Grid 2 0.5 Grid 2
Grid 1 Grid 1
0.0 0.0
r/R
r/R
PT
-0.5 -0.5
-1.0 -1.0
RI
-500 -400 -300 -200 -100 0 100 200 300 400 500 -500 -400 -300 -200 -100 0 100 200 300 400 500
Tangential velocity (m/s) Tangential velocity (m/s)
SC
c. d.
1.0 1.0
Grid 4 Grid 4
U
Grid 3 Grid 3
0.5 Grid 2 0.5 Grid 2
Grid 1 Grid 1
0.0
AN 0.0
r/R
r/R
-0.5 -0.5
M
-1.0 -1.0
-500 -400 -300 -200 -100 0 100 200 300 400 500 -500 -400 -300 -200 -100 0 100 200 300 400 500
D
Fig. 4. Influence of grid size on tangential velocity profiles: a. z/L = -0.04; b. z/L = 0.04; c. z/L = 0.5; d. z/L =
0.93.
C EP
AC
ACCEPTED MANUSCRIPT
a. b.
60 60
Num. VTP Num. VTP
Num. FTP Num. FTP
50 Exp. (Dincer et al., 2009) 50 Exp. (Dincer et al., 2009)
40 40
∆T (K)
∆T (K)
30 30
PT
20 20
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
RI
ξ ξ
c. d.
SC
70 70
Num. VTP
Num. FTP
60 Exp. (Dincer et al., 2009) 60
50 50
U
∆T (K)
∆T (K)
40 40
30
AN 30 Num. VTP
Num. FTP
Exp. (Dincer et al., 2009)
20 20
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
M
ξ ξ
Fig. 5. Performance of the numerical model in predicting experimental measurements: a. p = 200 kPa; b. p = 260
D
a.
b.
PT
RI
SC
Fig. 6. Streamlines colored by flow velocity: a. Fixed thermophysical properties; b. Variable thermophysical
properties.
U
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
a. b.
1.0 1.0
FTP FTP
VTP VTP
0.5 0.5
0.0 0.0
r/R
r/R
PT
-0.5 -0.5
-1.0 -1.0
-400 -300 -200 -100 0 100 200 300 400
RI
-400 -300 -200 -100 0 100 200 300 400
Tangential velocity (m/s) Tangential velocity (m/s)
c. d.
SC
1.0 1.0
U
0.5 0.5
0.0
AN r/R
0.0
r/R
-0.5 -0.5
M
-1.0 -1.0
-400 -300 -200 -100 0 100 200 300 400 -400 -300 -200 -100 0 100 200 300 400
Tangential velocity (m/s)
Tangential velocity (m/s)
D
Fig. 7. Tangential velocity profiles: a. z/L = -0.04; b. z/L = 0.04; c. z/L = 0.5; d. z/L = 0.93.
TE
C EP
AC
ACCEPTED MANUSCRIPT
a. b.
1.0 1.0
FTP FTP
VTP VTP
0.5 0.5
0.0
PT
0.0
r/R
r/R
-0.5 -0.5
RI
-1.0 -1.0
-500 -400 -300 -200 -100 0 100 200 300 400 500 -500 -400 -300 -200 -100 0 100 200 300 400 500
Axial velocity (m/s) Axial velocity (m/s)
SC
c. d.
1.0 1.0
FTP FTP
U
VTP VTP
0.5 0.5
0.0 AN 0.0
r/R
r/R
-0.5 -0.5
M
-1.0 -1.0
-200 -100 0 100 200 -200 -100 0 100 200
Axial velocity (m/s) Axial velocity (m/s)
D
Fig. 8. Axial velocity profiles: a. z/L = -0.04; b. z/L = 0.04; c. z/L = 0.5; d. z/L = 0.93.
TE
C EP
AC
ACCEPTED MANUSCRIPT
a.
340 K
335 K
330 K
325 K
320 K
315 K
310 K
305 K
b. 300 K
295 K
PT
290 K
285 K
280 K
275 K
270 K
265 K
RI
260 K
Fig. 9. Static temperature contours: a. Fixed thermophysical properties; b. Variable thermophysical properties.
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
a.
340 K
335 K
330 K
325 K
320 K
315 K
310 K
305 K
b. 300 K
295 K
PT
290 K
285 K
280 K
275 K
270 K
265 K
RI
260 K
Fig. 10. Total temperature contours: a. Fixed thermophysical properties; b. Variable thermophysical properties.
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
a.
4000 m²/s²
3800 m²/s²
3600 m²/s²
3400 m²/s²
3200 m²/s²
3000 m²/s²
2800 m²/s²
2600 m²/s²
2400 m²/s²
2200 m²/s²
b. 2000 m²/s²
1800 m²/s²
1600 m²/s²
PT
1400 m²/s²
1200 m²/s²
1000 m²/s²
800 m²/s²
600 m²/s²
400 m²/s²
200 m²/s²
RI
Fig. 11. Turbulent kinetic energy: a. Fixed thermophysical properties; b. Variable thermophysical properties.
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
a. b.
80 30
p = 200 kPa p = 200 kPa
70 = 260 kPa = 260 kPa
25
60 = 320 kPa = 320 kPa
= 380 kPa 20 = 380 kPa
50
Exin (W)
Exo (W)
40 15
30
10
PT
20
5
10
0 0
RI
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
ξ ξ
Fig. 13. Exergy versus inlet pressure: a. Inlet exergy; b. Outlet exergy.
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
a. b.
50 80
p = 200 kPa p = 200 kPa
= 260 kPa 70 = 260 kPa
40 = 320 kPa = 320 kPa
60
= 380 kPa = 380 kPa
30 50
∆Extot (W)
ηex (%)
40
20
PT
30
10 20
10
0
RI
0.0 0.2 0.4 0.6 0.8 1.0
0
0.0 0.2 0.4 0.6 0.8 1.0
ξ
ξ
SC
Fig. 14. Combined influence of the cold fraction and the inlet air pressure on: a. Total exergy losses; b. Exergy
efficiency.
U
AN
M
D
TE
C EP
AC