CFD Study of Orifice Pulse Tube Cryo-Cooler

B. S. Gawali, P. A. Mane
Mechanical Engineering Department,
Walchand College of Engineering
Sangli.

A commercial computational fluid dynamics (CFD) software package is used to model the
oscillating flow inside a pulse tube cryocooler. The cryocooler comprises compressor, after
cooler, regenerator, pulse tube with cold and hot heat exchanger, orifice and a reservoir. A
two-dimensional axis symmetric model is considered for simulation. This simulation
demonstrates the time varying temperature at cold heat exchanger along with computation of
the heat fluxes at the cold heat exchanger (CHX). Also the phase difference in pressure and
mass flow rate at CHX is also predicated. The only externally imposed boundary conditions
are a cyclically moving piston wall at one end of tube and constant temperature boundaries at
the external walls of the hot and cold heat exchangers. The present work aims to analyze the
CFD simulation of orifice pulse tube cryocooler and swirling phenomenon in the pulse tube.
And also make an attempt to reduce the swirling phenomenon by using mesh.

Keywords:-Orifice pulse tube cryocooler, CFD, cold heat exchanger (CHX), Regenerator.

INTRODUCTION inertance pulse tube cryocooler (IPTC) the

Pulse tube cryocooler (PTC) is one of the
promising competitors to the conventional
Gifford-McMahon and Stirling cryocoolers,
because it has no moving parts in low
temperature region and thus advantages of
long life, high reliability and low vibration. A
PTC is multifaceted flow loop that consists of
components in which complex solid fluid
thermal interactions takes place. Design of
pulse tube cryocooler systems has advanced
incessantly since invention of basic PTC [1]
leading to improvements in their performance
and lower achievable refrigeration
temperature. The major geometrical variations
introduced so far include the modified OPTC
[2], the multi-pass OPTC and recently the
exact nature of physical phenomenon
underlining operation of PTC is not well
understood a fundamental difficulty in all
PTCs is that their working fluid compression
and expansion processes are not well defined,
and poorly-understood thermal relaxation and
phase-lag phenomena which dominate their
operation. Crucial among these is the phase
angle between pressure and mass flow. This
phase angle is influenced by the wave
resonance phenomena in BPTCs, and is
adjusted by orifices and/or valves in
various OPTC designs.
Theoretical approaches like cyclic, nodal,
phasor analysis have been used for the
analysis of PTCs [3,4,5]. These models,
although simple and convenient to use, do not
adequately consider the flow and heat transfer
details. Computational models have recently
been developed for the simulation of the
regenerator, or the entire PTC. The
commercially available CFD code, such as
Fluent [6] is a powerful tool to numerically Figure 1: Simulated OPTC system
solve the Navier-stokes equations with finite
volume descritization. The main difficulties are The flow in the PTC is considered as two
modeling oscillating flow phenomenon in the dimensional therefore the axis symmetric
system and heat transfer in regenerators and geometry is used to reduce the computational
heat exchangers. With the available features time and number of cells. Quadrilateral and
in commercial codes it is possible to triangular mesh structure is used for meshing
overcome the above mentioned difficulties. the OPTC system with different features of the
The CFD simulation of PTC is not Gambit [12]. Total numbers of cells for the
straightforward and can produce meaningful model with zero thickness are 9282.
results only when reasonably closure relations
representing the fluid solid interactions are Table 1: Dimensions of the Simulated System
applied.
The applications of this simulation code to Radius Length
PTCs have been demonstrated by some Component
(mm) (mm)
researchers. Hozumi et. al. [7] performed axis A(compressor) 10 6.5
symmetric 3D simulations of BPTC and OPTC B(Transfer line) 2 10
with an interest in the effects of gravity and C(WHX1) 4.3 9.21
orientation on system performance. Flake and D(Regenerator) 4.3 52
Razni [8] conducted an axis symmetric E(CHX) 4.3 9.85
analysis of BPTC and an OPTC, using the F(Pulse tube) 3.1 80
fluent commercial CFD package and indicated G(WHX2) 3.1 9.89
the occurrence of recirculation patterns and H(Orifice) 0.21 ----
streaming effects in their simulated pulse I (Reservoir) 3.1 50
tube. Cha et.al. [9] also used CFD software J (Reservoir) 40 100
fluent to simulate the inertance pulse tube
cryocooler (IPTC) and studied the The different components of OPTC are
multidimensional flow and heat transfer subjected to different boundary conditions i.e
effects. adiabatic wall, constant wall temperature etc.
In the present work sinusoidal moving piston Boundary conditions of these components are
face is used to model the compressor. listed in table 2.
Regenerator and heat exchangers are
modeled using porous media condition. This Table 2. Boundary conditions for the
paper further verifies the feasibility of CFD simulation
simulation of OPTC.
Initial Value
Model / Boundary
SIMULATED SYSTEM
Condition
Compressor Wall Adiabatic
A schematic of OPTC is shown in Figure 1.
Transfer-Line Wall Adiabatic
The cryocooler is identical to the experimental
WHX1 Wall 300 K
set up of Gawali et al. [10]. The geometric
Regenerator Wall Adiabatic
dimensions of all the components are listed in
CHX Wall Adiabatic /
Table 1
1watt load
 Pulse Tube Wall Adiabatic For modeling the porous media the resistance
WHX2 Wall 300 K coefficients α and C2 are obtained using
Tube Wall Adiabatic equation 1 and 2 [11].
Reservoir Volume Wall Adiabatic
Initial Condition 300 K D 2
  3

Operating Pressure 14 bar  

p
(1)
1 5 0  (1   ) 2
Working medium (Helium All properties
gas) 3 . 5  (1   )
C 2  (2)
D p   3

CFD MODELLING Continuum based conservation equation is

applied in the system. The mass, momentum
The CFD code FLUENT (6.3.26) is used for and energy equations solved by FLUENT are
simulation. The OPTC system is simulated by as follows:
assuming cylindrical and linear alignment,
axis-symmetric, two-dimensional flow, with f
working media helium as ideal gas. The
1 
 rf vr   f vx   0 (3)
transient analysis is done because the t r r x
process is time dependant.
Compressor is modeled by using dynamic   
  f v   .   f vv   . p  .  (4)
mesh feature of the FLUENT. The frequency t    
of the compressor is 39 Hz. The dynamic
mesh model in fluent is used to model flows   
where the shape of the domain is changing t     
  f E   . v   f E  p    .  k f T   .v 

(5)

with time due to motion of the domain p v2

boundaries. The dynamic mesh model can Where E  h  
also be used for steady-state applications,
 2
where it is beneficial to move the mesh in the All properties represent the properties of
steady state solver. The motion is prescribed the working fluid helium. The above equations
motion by specifying the linear and angular apply to all components, except for CHX,
velocities about the center of gravity of solid WHX1, WHX2 and the regenerator i.e
body with time. In order to model compressor, compressor, pulse tube, reservoir etc.
a user defined function with necessary
modification was developed in C programming The latter four components are modeled as
to simulate the piston cylinder effect. porous media, assuming that within the
volume containing the distributed resistance,
PistonDisplacement  X  X a sin t there exists everywhere local balance
between pressure and resistance forces [11].
For porous media the eqn.6, 7, 8 of mass,
Regenerator, Warm Heat Exchangers (WHX1 momentum, and energy are used.
and WHX2), and Cold Heat Exchanger
(CHX1) were modeled as porous media. ( f ) 1 
Stainless steel wire mesh screens (mesh size  [ r  f  r ]  ( f  x )  0 (6)
250) are usually selected as the regenerator t r r x
packing material, since they offer high heat   
( f  ) ( f . ) 
transfer surface areas, high heat capacity, and t
low thermal conductivity. Copper mesh (mesh (7)
1 
size 150) is used for WHX1, WHX2, and CHX. P [ ]  [ 1 j  c f j j]
2
 
[f Ef  (1)f Es ] [(f Ef  P)]
t (8)

[ Kf  (1)ksT  ( .)

All the simulations are carried out as transient

processes, starting with an initial system
temperature of 300 K. Simulations can be
continued until steady-periodic state is
obtained.
The criterion for steady periodic condition is
that the cycle-average temperature of the cold
end heat exchanger (CHX) would reach a
steady state.
RESULTS AND DISCUSSION
Temperature variation along axial direction is
shown in Figure2. Significant temperature
gradients are predicted in the regenerator and
pulse tube.
Temperature contour along axial direction is
shown in Figure3. The gas piston is also
Figure3. Temperature (K) contour along axial
direction
The variation in cycle average
temperature at CHX with time is shown in
Figure 4. Steady state temperature (cycle
average) of 100.9 K is attained after 120 sec.
The experimental value of cycle average
temperature at CHX reported by Gawali et al
is 98.9K. And the time required to reach
steady state is around 20 min. This variation
in time is due to thermal loss in actual system.
350
300

observed at the centre of the pulse tube and it 250

Temperature ( K)

varies with crank angle. 200

150

350 100

50
300
0
250 0 20 40 60 80 100 120 140
WHX 1 Time ( Sec.)
200
Temperature ( K)

REGENERATOR
150 CHX
PULSE-TUBE
100
Figure 4. Cool down characteristics
WHX2
50

0 The variation in pressure and mass flow rate

0 50 100 150 200 with crank angle at CHX is shown Figure 5.
Length ( mm) The phase difference in pressure and mass
flow rate of 400 is observed. The mass flow
Figure2. Temperature variation in axial rate is lagging the pressure wave.
Direction The phase difference reported by Gawali.et.al
is 400 respectively. These experimental results
validate the CFD predictions.
 Pressure Such phenomenon in the pulse tube is
20 0.025
Mass flow rate partially responsible for the overall
0.02
16 0.015
performance of the cryocooler like unsteady
0.01
cold end temperature. Similar results were

Mass flow rate (kg/sec)

12 observed by Zhang [13] and suggested that
Pressure ( bar)

0.005
0 swirling phenomenon can reduced using wire
8
-0.005 mesh in the pulse tube. Figure 7 and 8 shows
4 -0.01 velocity snapshots in the pulse tube with and
-0.015 without screen.
0 -0.02
0 200 400 600 800
Crank Angle

Figure. 5 Phase shift at CHX

Figure.7. Swirling pattern in Pulse Tube
The Figure 6 shows the temperature variation
without and with (1 watt) load at CHX. When
there is no load the steady state temperature
observed is 100.9 K. When 1 watt load is
applied at CHX the steady state temperature
observed is 115 K.
Figure.8 Flow visualization with screen in
pulse tube
116 1W load 101.6
No load Screen is located at 20 mm from CHX to study
115.8 101.4 its effect on fluid flow in the pulse tube.
115.6 Swirling phenomenon is reduced with
Temperature (K)

Temperature (K)

101.2
115.4
installation of screen. The influence of location
101 and size of screen can be a parameter of
115.2
interest.
100.8
115

114.8 100.6
CONCLUSION
114.6 100.4
0 200 400 600 800
The entire OPTC system operating in steady
Crank angle
periodic state under variety of boundary
conditions is numerically simulated using CFD
Figure 6. Temperature variations at CHX with code. The present CFD simulation confirms
no load and load condition. that the commercial available codes like fluent
The trends predicated with the help of CFD are capable to model the complex fluid flow
tool fluent i.e. Figure 3,4,5,6 are in good phenomenon and predict property variation in
agreement with the experimental results of OPTC system. Also the attempt made to
Gawali et.al. reduce the swirling is partially achieved; still
further analysis is required to reduce the
undesirable effects of swirling in the pulse
FLOW PHENOMENON IN THE PULSE tube.
TUBE

In the present simulation swirling

phenomenon is observed in the pulse tube.
REFERENCES
Nomenclature
1. Gifford W. E. and Longsworth, “Pulse tube Dp Mean diameter of particles [m].
refrigeration progress.” Advances in
Cryogenic Engineering, Vol. 10, pp. 69-79. h Enthalpy [J.kg-1]
2. Mikulin E. I., Tarasov A. A. and k Thermal Conductivity [W.m-1.K-1]
Shkrebyonock, “Low temperature expansion p Pressure [N.m-2]
Pulse Tube.” Advances in Cryogenic T Temperature [K].
Engineering, Vol. 29, pp. 629-637, Plenum t Time [s]
press, New York, (1984).  & c2 Resistance coefficients
3. Zhu S.W and Chan Z.Q, “Isothermal model Greek symbols
of pulse tube refrigerator”, Cryogenics (1994)  Stress tensor [N.m-2]
Vol.34 PP591-595.  Porosity
4. Atrey M. D., Bapat S. L., and
Subscripts:
Narayankhedkar K. G., “Cyclic simulation
of Stirling cryocoolers.” Cryogenics, 29, pp. f Fluid
341-347.  Velocity [m.s-1]
5. Gawali B.S., Atrey M.D. and  Molecular Viscosity [kg.m-1.s-1].
Narayankhedkar K.G. “Performance  Fluid Density [kg.m-3]
Prediction and Experimental Investigation of r Radial coordinate
Orifices Pulse Tube Cryocooler” ICEC 19, s Solid
2003, pp 391-394. x Axial coordinate
6. Jon D Anderson, Jr. “ Computational Fluid r Radial coordinate
Dynamics”, Publication: Mcgraw-Hill, Inc.
7. Hozumi Y, Shiraishi M, Murakami M.
“Simulation of thermodynamics aspects about
pulse tube cryocooler.” In: Proceedings of the
international cryogenics engineering
conference, September 2003. p. 1500–7.
8. Flakes B, Razani A. “Modeling pulse tube
cryocoolers with CFD.” In: Proceedings of the
international cryogenics engineering
conference, September 2003. p. 1493–9.
9. Cha J S, Ghiaasiaan S M, Desai P V,
Harvey J P, Kirkconnell C S. “CFD simulation
of multi-dimensional effects in an inertance
pulse tube refrigerator.” In: Proceedings of
13th international cryocooler conference, vol.
13; 2004.p. 285-91
10. Gawali B. S,”Design and development of a
miniature integral Pulse Tube Nitrogen
Liquefier”, Ph. D Thesis, I.I.T Bombay, 2005
11. Fluent 6.3.26 Users guide, Fluent. Inc.
12. Gambit 2.3.16 Users guide.
13 Zhang X.B. , Qiu L.M., Gan Z.H., He Y.L.,
“CFD study of a simple orifice pulse tube
cooler” cryogenics 47, pp 315-321 (2007).