CFD Modelling of Flow and Heat Transfer within the Parallel Plate Heat Exchanger in Standing Wave Thermoacoustic System

1
CFD Modelling of Flow and Heat Transfer within
the Parallel Plate Heat Exchanger in Standing
Wave Thermoacoustic System
Fatimah A. Z. Mohd Saat and Artur J.
Jaworski Department of Engineering University of
Leicester, University Road, UK, LE1
7RH, a.jaworski_at_leicester.ac.uk, Tel44 (0) 116
223 1033 Fax[phone]
2
Presentation outline

• Background
• Introduction
• Computational Model
• Terminology for data analysis
• Validation
• Results and Discussions
• Effect of Temperature
• Effect of gravity and orientation
• Conclusions and Future Work

2/17
3
Background
TA effect is based on the interaction between
sound wave and solid boundary producing power or
heat pumping effect.

• Gas particle move back/forth standing sound wave
• Compress-expands change heat with solid wall
  (stack)

Stirling cycle
Thermoacoustic cycle
Plate

• Challenges
• commercialization interest - upgrade efficiency
• high efficiency high amplitude i.e. non-linear
  effects, temperature-driven phenomena.

3/17
4
Introduction

• Motivation Understanding the detailed
  fluid-flow heat transfer could be the key to
  improving thermodynamic performance of TA
  systems.
• Fundamental work disregard temperature
• Mao et al. Experiments in Fluids,
• 2008, 45 833-846. PIV,
• vortex shedding and flow pattern
• evolution at higher acoustic
• excitations.
• Jaworski AJ et. al. Exp Thermal and Fluid
• Science,2009, 33 495-502
• CFD-assisted analysis on entrance effect
• Entrance length changes over time of a flow cycle
• Entrance length deeper into the channel
• Practical application temperature effect
• Shi L. et. al., Meas. Sci. Tech., 2010, 21115405
• Experimental method for thermal analysis (PLIF)
• of parallel plate heat exchanger
• flow changes due to temperature

4/17
5
Simplified investigation(Shi L. et.al. ,2010)
Introduction

• Simplify the configuration of stack and heat
  exchangers as shown above
• focus on the generic fluid flow and heat transfer
  processes rather than the intricacies of
  practical thermoacoustic systems
• for the convenience of performing PIV and PLIF
  experiments especially simplified illumination
  and temperature range considerations

5/17
6
Introduction Experiment work

• Shi L. et. al., Meas. Sci. Tech., 2010, 21 115405

PLIF
6/17
Tw K-type thermoacouple
7

• CFD reliable technique for design and analysis
• Piccolo A. (2011) modelled 1D sound wave with 2D
  time-averaged temperature and energy flux. No
  information on time variation of flow and heat
  transfer phenomena. Heat estimated with an error
  30-50 compared to experiment.
• Current study developing reliable TA CFD method
  on the basis of the detailed velocity and
  temperature field distributions for future design
  and optimization.
• Effect of gravity (Lin Y. et. al. 2008) and
  device orientation (Shen C. et. al. 2009) on flow
  and heat transfer across heat exchangers.

7/17
8
Computational model
U
P
?/4 model
Pressure antinode
loudspeaker
4.6m
7.4m
Moving wall
Fine mesh near HX area Course elsewhere- 1D
steady flow
Heat exchangers wall temperature
Pressure antinode
Distance from pressure antinode
Sound speed
8/17
9
Computational model
9/17

• Solver setting
• PISO pressure-velocity coupling
• Spatial discretization 2nd order upwind
• Time discretization 1st order implicit
• (dynamic mesh)
• Time steps-1200 per cycle.
• UDF code - moving wall,
• - HX wall temperature,T(x).

ANSYS-FLUENT 13.0
Unsteady 2D model
N2 compressible gas
Gravity ?(p, T)
Laminar viscous heating
T-dependent properties
Frequency 13.1 Hz
Pm 0.1MPa
Drive ratio PA/Pm ()0.3
70 cyclesReal time 5.3s Computational time ? 10
days
Pend wall (Pa)
Convergence - v, mass 1 x 10 -4 - Energy 1 x
10-7 Steady state oscillatory flow
steady oscillatory state
10
Terminology for data analysis
positive direction
negative direction
g
ve direction
-ve direction

• At appropriate location and phase cycle, the
  effect of temperature, gravity and device
  orientation will be presented as
• Velocity profile
• Temperature profile
• Heat flux

joint
CHX
HHX
1 mm
15 mm
15 mm
10/17
11
Validation
Normalized velocity validated at 1mm from the
joint above CHX very good match

• Max. deviation between CFD and
• EXPERIMENT
• - HHX (18C)
• - CHX (17C)
• Max experimental standard
• deviation (16?C)
• Qualitatively good match to
• experiment (trend, shape)

11/17
12
Effect of Temperature
12/17
CASE 1
Symmetrical distribution
Flow symmetry is distorted by the effect of
temperature
CASE 2
CASE 1 no temperature CASE 2 with temperature
13
Effect of Temperature
joint
dT flow leaning to the left
HHX
CHX
15 mm
15 mm
Likely related to temperature-driven buoyancy
effect
13/17
14
?9, CHX and HHX
dT0 similar Ucentreline at CHX and HHX
dT - Ucentreline smaller at CHX, larger
at HHX (different gas displacement) -
Velocity gradient at HHX shifted up
joint
HHX
CHX
15 mm
15 mm
14/17
15
Effect of gravity and orientation
?9 - flow in positive direction (from HHX
to CHX)
0

0?
g
-90
Orientation 90?(HHX above CHX) - lower
velocity -90? (CHX above HHX) - higher velocity
90
The temperature-driven buoyancy effect either
assist or obstruct the flow depending on the
device orientation, flow direction
15/17
16
Heat transfer characteristics
16/17
g Ty3mm smaller compared to
g0 Orientation 90? (HHX above CHX) -90? (CHX
above HHX)
Space-time-averaged q
Case Error Error
HOT COLD
g 0 -29 -30
g0 0? -42 -27
17
Conclusion and Future Work

• Chosen model and boundary conditions are able to
  simulate the oscillatory flow and T-driven
  natural convection effect.
• Geometry is symmetry, but flow is not due to the
  presence of T- symmetry or periodic model is not
  appropriate.
• Velocity profiles good match
• Temperature profile same trend differences
  possibly due to heat accumulation and the essence
  of 3D feature of experimental rig.
• Effect of temperature flow structure distorted
• Flow and heat transfer depends on the location of
  hot heat exchanger, direction of flow and
  temperature-driven buoyancy effect.
• Further investigation - wider range of fluid
  displacement.
• Variation of T(x) and u(x) for flow between
  plates investigating the vortex structure at
  the end of plates and possibility of entrance
  effects between the plates.

17/17
18
Thank you for your attention
19
CFD
?1
PLIF
CFD
?11
PLIF

• Temperature-driven buoyancy effect
• Hot end successfully replicated but with
  slightly different pattern
• Cold end buoyancy effect not seen possibly due
  to initialisation of
• model fluid temperature
  (initialized at 80?C, to save
• computational time and match
  the heat accumulation in
• experimental data
• Between plates Temperature contour matched
  reasonably well

11/18
20
70 cyclesReal time 5.3s Computational time ? 10
days
Pend wall (Pa)