This action might not be possible to undo. Are you sure you want to continue?
Scholar Commons
Theses and Dissertations
612007
Numerical analysis of heat transfer during jet
impingement on curved surfaces
Cesar F. HernandezOntiveros
University of South Florida
This Thesis is brought to you for free and open access by Scholar Commons. It has been accepted for inclusion in Theses and Dissertations by an
authorized administrator of Scholar Commons. For more information, please contact scholarcommons@usf.edu.
Scholar Commons Citation
HernandezOntiveros, Cesar F., "Numerical analysis of heat transfer during jet impingement on curved surfaces" (2007). Theses and
Dissertations. Paper 2211.
http://scholarcommons.usf.edu/etd/2211
Numerical Analysis of Heat Transfer During Jet Impingement on
Curved Surfaces
by
Cesar F. HernandezOntiveros
A thesis submitted in partial fulfillment
of the requirements for the degree of
Master of Science in Mechanical Engineering
Department of Mechanical Engineering
College of Engineering
University of South Florida
Major Professor: Muhammad Mustafizur Rahman, Ph.D.
Frank Pyrtle III, Ph.D.
Autar Kaw, Ph.D.
Date of Approval:
March 30, 2007
Keywords: steady state, transient analysis, hemispherical plate, cylindrical
plate, heat flux
© Copyright 2007, Cesar F. HernandezOntiveros
Dedication
To
God
My Father and Mother
Without them this could not have been possible
To
My Advisor Professor Muhammad Mustafizur Rahman
Thank you for your immense patience and guidance
Acknowledgments
To
My Friends: Son H. Ho, Jorge C. Lallave, and Phaninder Injeti
for their help and support throughout this very long journey!
i
Table of Contents
List of Figures iii
List of Symbols x
Abstract xiii
Chapter 1 Introduction 1
1.1 Literature Review 2
1.1.1 Round Jet Impingement 2
1.1.2 Slot Jet Impingement 4
1.2 Overview of Literature 8
1.3 Thesis Aim 8
Chapter 2 Mathematical Models and Computation 10
2.1 Hemispherical Model 10
2.1.1 Governing Equations : Steady State Heating 12
2.1.2 Boundary Conditions : Steady State Heating 12
2.1.3 Governing Equations : Transient Heating 14
2.1.4 Boundary Conditions : Transient Heating 15
2.2 Cylindrical Plate Model 16
2.2.1 Governing Equations: Steady State Heating 17
2.2.2 Boundary Conditions : Steady State Heating 17
2.2.3 Governing Equations : Transient Heating 19
2.2.4 Boundary Conditions : Transient Heating 19
2.3 Numerical Computation 21
2.3.1 Steady State Process 21
2.3.2 Transient Process 23
2.4 Mesh Independence and Time Step Study 24
2.4.1 Cylindrical Coordinates 24
2.4.2 Cartesian Coordinates 26
Chapter 3 Hemispherical Model Results 29
3.1 Steady State Heating 29
3.2 Transient Heating 52
ii
Chapter 4 Cylindrical Plate Model Results 66
4.1 Steady State Heating 66
4.2 Transient Heating 85
Chapter 5 Discussion and Conclusions 98
References 100
Bibliography 104
Appendices 105
Appendix A: CFD Code for Axisymetric Model 106
Appendix B: CFD Code for 2D Model 114
iii
List of Figures
Figure 2.1 Three Dimensional Schematic of a Hollow Hemisphere
During an Axisymetric Liquid Jet Impingement
11
Figure 2.2 CrossSectional View of a Hollow Hemisphere During an
Axisymetric Jet Impingement
11
Figure 2.3 Schematic View of a Curved Plate During a 2D Symmetric
Liquid Jet Impingement
16
Figure 2.4 Mesh Plot of a Curved Plate During a 2D Symmetric Liquid
Jet Impingement
21
Figure 2.5 Local Dimensionless Interface Temperature for Different
Number of Elements in z (or r) and Φ Directions for Water
as Fluid and Silicon as Solid (Re=750, b/d
n
=0.5, ß=2.0,
q=250 kW/m
2
)
25
Figure 2.6 SolidFluid Interface Dimensionless Maximum Temperature
Variation for Silicon Hemisphere at Different Time
Increments (Re=750, b/d
n
=0.5, ß=2.0, H
n
= 0.30 cm, q=250
kW/m
2
)
26
Figure 2.7 Local Dimensionless Interface Temperature for Different
Number of Elements in x or φ, and y Directions (Re=750,
ß=2.5, H
n
= 0.30 cm)
27
Figure 2.8 SolidFluid Interface Dimensionless Maximum Temperature
Variation for Silicon Plate at Different Time Increments
(Re=750, b/d
n
=0.5, ß=2.5, H
n
= 0.30 cm, q=250 kW/m
2
)
28
iv
Figure 3.1
Velocity Vector Distribution for Silicon Hemisphere with
Water as the Cooling Fluid (Re = 750, b = 0.60 mm, Q =
5.665x10
7
m
3
/s, β = 2.0, H
n
= 0.30 cm, q = 250 kW/m
2
)
30
Figure 3.2
Pressure Distribution for Silicon Hemisphere with Water as
the Cooling Fluid (Re = 750, b = 0.60 mm, Q = 5.665x10
7
m
3
/s, β = 2.0, H
n
= 0.30 cm, q = 250 kW/m
2
)
31
Figure 3.3 Temperature Distribution for Silicon Hemisphere with Water
as the Cooling Fluid (Re = 750, b = 0.60 mm, Q = 5.665x10

7
m
3
/s, β = 2.0, H
n
= 0.30 cm, q = 250 kW/m
2
)
32
Figure 3.4 Free Surface Height Distribution for Different Reynolds
Numbers and Water as the Cooling Fluid (b=0.60 mm,
β=2.0, H
n
=0.30 cm, q=250 kW/m
2
)
33
Figure 3.5 Dimensionless Interface Temperature Distribution for a
Silicon Hemisphere at Different Reynolds Numbers, and
Water as the Cooling Fluid (β=2.0, H
n
= 0.30 cm, b=0.60
mm, T
o
=373 K)
35
Figure 3.6 Local Nusselt Number Distribution for a Silicon Hemisphere
at Different Reynolds Numbers, and Water as the Cooling
Fluid (β=2.0, H
n
=0.30 cm, b=0.60 mm, T
o
=373 K)
35
Figure 3.7 Dimensionless Interface Temperature Distribution for a
Silicon Hemispherical Plate at Different Reynolds Numbers,
and Water as the Cooling Fluid (β=2.0, H
n
= 0.30 cm, b=0.60
mm, q=250 kW/m
2
)
36
Figure 3.8 Local Nusselt Number Distribution for a Silicon plate at
Different Reynolds Numbers, and Water as the Cooling
Fluid (β =2.0 , H
n
= 0.30 cm, b=0.60 mm, q=250 kW/m
2
)
37
Figure 3.9 Average Nusselt Number and Heat Transfer Coefficient
Variation for Different Reynolds Numbers for Constant Heat
Flux (q=250kW/m
2
) and Isothermal (T
o
= 373 K) Boundary
Conditions (β=2.0, H
n
= 0.30cm)
38
Figure 3.10 Dimensionless Interface Temperature for Silicon Hemisphere
at Different Nozzle to Target Spacing Ratio (β) for Water as
the Cooling Fluid (Re=500, Q=3.776x10
7
m
3
/s, d
n
=1.2mm,
b=0.60 mm, q=250 kW/m
2
)
39
v
Figure 3.11 Local Nusselt Number for Silicon Hemisphere at Different
Nozzle to Target Spacing Ratio (β) for Water as the Cooling
Fluid (Re=500, Q=3.776x10
7
m
3
/s, d
n
=1.2mm, b=0.60mm,
q=250kW/m
2
)
40
Figure 3.12 Dimensionless Interface Temperature for Silicon
Hemisphere at Different Thickness (b) for Water as the
Cooling Fluid (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm,
b=0.60 mm, q=250 kW/m
2
)
41
Figure 3.13 Local Nusselt Number for Silicon Hemisphere at Different
Thickness (b) for Water as the Cooling Fluid (Re=750,
Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60mm, q=250kW/m
2
)
42
Figure 3.14
Dimensionless Interface Temperature for Silicon Plate at
Different Thickness (b) for Water as the Cooling Fluid
(Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60 mm,
T
o
=373 K)
43
Figure 3.15 Local Nusselt Number for Silicon Hemisphere at Different
Thickness (b) for Water as the Cooling Fluid (Re=750,
Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60mm, T
o
=373 K)
44
Figure 3.16 Dimensionless Interface Temperature for Silicon
Hemispherical Plate for Different Cooling Fluids (Re=750,
Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60 mm, T
o
=373 K)
45
Figure 3.17 Local Nusselt Number for Silicon Hemisphere for Different
Cooling Fluids (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm,
b=0.60mm, T
o
=373 K)
45
Figure 3.18 Local Nusselt Number and Dimensionless Interface
Temperature for Different Cooling Fluids (Re = 1500,
Q=1.133x10
6
m
3
/s, β=2.0, H
n
=0.30 cm, b=0.60 mm, q=250
kW/m
2
)
46
Figure 3.19 Local Nusselt Number and Dimensionless Interface
Temperature for Different Hemisphere Materials with Water
as the Cooling Fluid (Re=1000, β=2.0, b/d
n
=0.5, q=250
kW/m
2
)
48
Figure 3.20 Dimensionless Interface Temperature for Silicon
Hemispherical Plate for Different Materials (Re=750,
Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60 mm, T
o
=373 K)
49
vi
Figure 3.21 Local Nusselt Number Distribution for Silicon Plate for
Different Solid Materials (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60mm, T
o
=373 K)
50
Figure 3.22 Stagnation Nusselt Number Compared with Liu et al. [36],
Scholtz and Trass [37], and Nakoryakov et al. [38] with
Actual Numerical Results under Different Reynolds
Numbers
51
Figure 3.23 Dimensionless Local Heat Flux Variation at the SolidFluid
Interface for Silicon Hemisphere at Different Time Instants
(Re=750, b/d
n
=0.5, ß=2.5)
53
Figure 3.24 Dimensionless Interface Temperature Variation for Silicon
Hemisphere at Different Time Instants (Re=750, b/d
n
=0.5,
ß=2.5)
54
Figure 3.25. Local Nusselt Number Distribution for Silicon Hemisphere at
Different Time Instants (Re=750, b/d
n
=0.5, ß=2.5)
55
Figure 3.26 Distribution of Dimensionless Maximum Temperature at the
SolidFluid Interface, within the Solid, and Maximumto
Minimum Temperature Difference with Time for Two
Reynolds Numbers (Silicon Hemisphere, b/d
n
=0.5, ß=2.5)
56
Figure 3.27 Variation of Average Nusselt Number with Time For Silicon
Hemisphere at Two Different Reynolds Numbers (b/d
n
=0.5,
ß=2.5)
57
Figure 3.28 Time Required to Reach SteadyState for Silicon
Hemisphere at Different Reynolds Numbers (b/d
n
=0.5,
ß=2.5)
58
Figure 3.29
Distribution of Dimensionless Maximum Temperature at the
SolidFluid Interface, within the Solid, and Maximumto
Minimum Temperature Difference with Time for Different
Plate Thicknesses (Silicon Hemisphere Re=750, ß=2.5)
59
Figure 3.30 Variation of Average Nusselt Number with Time for
Constantan Hemisphere at Three Different Thicknesses
(Re=750, ß=2.5)
60
vii
Figure 3.31 Distribution of Dimensionless Maximum Temperature at the
SolidFluid interface, within the Solid, and Maximumto
Minimum Temperature Difference with Time for Different
Materials (Re=750, ß=2.5)
61
Figure 3.32 Variation of Average Nusselt Number with Time for Different
Materials (Re=750, b/d
n
,=0.5, ß=2.5)
62
Figure 3.33 Time Needed to Reach SteadyState for Different
Hemispherical Plate Thicknesses and for Different Materials
(Re=750, ß=2.5)
63
Figure 3.34 Isotherms Countour Lines at Varius Time Instants for Silicon
Hemisphere Plate of b/d
n
=0.5 (Re=750, ß=2.5, Top Left
Corner (z=0, r=0))
64
Figure 3.35 Isotherms Countour Lines at Varius Time Instants for Silicon
Hemisphere Plate of b/d
n
=1.5 (Re=750, ß=2.5, Top Left
Corner (z=0, r=0)
65
Figure 4.1 Velocity Vector Distribution for a Silicon Curved Plate with
Water as the Cooling Fluid (Re = 750, b= 0.60 mm, Q =
5.665x10
7
m
3
/s, β = 2.5, H
n
= 0.3 cm, q = 250 kW/m
2
)
67
Figure 4.2 Free Surface Height Distribution for Different Reynolds
Numbers and Water as the Cooling Fluid (b=0.60 mm, β=2.0,
H
n
=0.30 cm, q=250 kW/m
2
)
68
Figure 4.3 Dimensionless Interface Temperature Distribution for a
Silicon Curved Plate at Different Reynolds Numbers and
Water as the Cooling Fluid (β=2.5, H
n
= 0.30 cm, b=0.60 mm,
q=250 kW/m
2
)
69
Figure 4.4 Local Nusselt Number Distribution for a Silicon Curved Plate
at Different Reynolds Numbers and Water as the Cooling
Fluid (β =2.5, H
n
=0.30 cm, b=0.60 mm, q=250 kW/m
2
)
70
Figure 4.5 Average Nusselt Number and Maximum Temperature Within
the Solid for Different Reynolds Numbers for Constant Heat
Flux Conditions (q
= 250 KW/m
2
, β=2.5, H
n
= 0.30cm)
71
viii
Figure 4.6 Dimensionless Interface Temperature for Silicon Curved
Plate for Different Nozzle to Target Spacing Ratio (β) and
Water as the Cooling Fluid (Re=750, d
n
=1.2mm, b=0.60
mm, q=250 kW/m
2
)
72
Figure 4.7 Local Nusselt Number Distribution for Silicon Curved Plate
for Different Nozzle to Target Spacing Ratio (β) and Water
as the Cooling Fluid (Re=750, d
n
=1.2mm, b=0.60mm,
q=250kW/m
2
)
73
Figure 4.8 Dimensionless Interface Temperature for Silicon Curved
Plate for Different Inner Plate Radius of Curvature to Nozzle
Diameter Ratio (R
i
/d
n
) for Water as the Cooling Fluid (β=2.5,
H
n
= 0.30 cm Re=750, d
n
=1.2mm, b=0.60 mm, q=250
kW/m
2
)
74
Figure 4.9 Local Nusselt Number Distribution for Silicon Curved Plate
for Different Inner Plate Radius of Curvature to Nozzle
Diameter Ratio (R
i
/d
n
) for Water as the Cooling Fluid (β =2.5
, H
n
= 0.30 cm Re=750, d
n
=1.2mm, b=0.60 mm, q=250
kW/m
2
)
75
Figure 4.10 Average Nusselt Number and Maximum Temperature in the
Solid Variations (Θ
max
) for Silicon Curved plate for Different
Inner Plate Radius of Curvature to Nozzle Diameter Ratio
(R
i
/d
n
) for Water as the Cooling Fluid (β =2.5 , H
n
= 0.30 cm
Re=750, d
n
=1.2mm, b=0.60 mm, q=250 kW/m
2
)
76
Figure 4.11 Heat Transfer Coefficient and Nusselt Number Distributions
for Different Nozzle Slot Widths (Q=0.0006 m
3
/s, β =2.5 , H
n
=
0.30 cm β=2.5, Re=750, q=250 kW/m
2
77
Figure 4.12 Dimensionless Interface Temperature for Different Plate
Thicknesses (β =2.5, Re=750, d
n
=1.2mm, q=250 kW/m
2
)
78
Figure 4.13 Local Nusselt Number Distribution for Different Plate
Thicknesses (β =2.5, Re=750, H
n
= 0.30 cm, q=250 kW/m
2
)
79
Figure 4.14 Average Nusselt Number and Maximum Temperature in the
Solid Variations with Plate Thickness (β =2.5, Re=750)
80
Figure 4.15 Dimensionless Interface Temperature for Different Cooling
Fluids (β=2.5, Re=750)
81
ix
Figure 4.16
Heat Transfer Coefficient and Nusselt Number Distributions
for Different Cooling Fluids (Re=750, β=2.5)
82
Figure 4.17
Local Nusselt Number Compared with Bartoli and Faggiani
[21] at Different Reynolds Numbers (φ=90°, d
n
=2.0 mm,
b=0.50 mm, q=50 kW/m
2
)
83
Figure 4.18 Average Nusselt Number Compared with Gori and Bossi [29]
and Whitaker [40] for Different Reynolds Numbers (d
n
=2.5
mm, b = 0.2 mm, q=2.35 kW/m
2
)
84
Figure 4.19 Dimensionless Local Heat flux Variation at the SolidFluid
Interface for Silicon Plate at Different Time Instants (Re=750,
b/d
n
=0.5, ß=2.5)
86
Figure 4.20 Dimensionless Interface Temperature Variation for Silicon
Plate at Different Time Instants (Re=750, b/d
n
=0.5, ß=2.5)
88
Figure 4.21 Local Nusselt Number Distribution for Silicon Plate at
Different Time Instants (Re=750, b/d
n
=0.5, ß=2.5)
89
Figure 4.22 Distribution of Dimensionless Maximum Temperature at the
SolidFluid Interface, within the Solid, and Maximumto
Minimum Temperature Difference with Time for Two
Reynolds Numbers (Silicon Plate, b/d
n
=0.5, ß=2.5)
90
Figure 4.23 Variation of Average Nusselt Number with Time For Silicon
Plate at Two Different Reynolds Numbers (b/d
n
=0.5, ß=2.5)
91
Figure 4.24 Time Required to Reach SteadyState for Silicon Plate at
Different Reynolds Numbers (b/d
n
=0.5, ß=2.5)
92
Figure 4.25 Distribution of Dimensionless Maximum Temperature at the
SolidFluid Interface, within the Solid, and Maximumto
Minimum Temperature Difference with Time for Different
Plate Thicknesses (Silicon Plate Re=750, ß=2.5)
93
Figure 4.26 Variation of Average Nusselt Number with Time for
Constantan Plate at Three Different Thicknesses (Re=750,
ß=2.5)
94
x
Figure 4.27 Distribution of Dimensionless Maximum Temperature at the
SolidFluid interface, within the Solid, and Maximumto
Minimum Temperature Difference with Time for Different
Materials (Silicon Plate, Re=750, ß=2.5)
95
Figure 4.28 Variation of Average Nusselt Number with Time for Different
Materials (Re=750, b/d
n
,=0.5, ß=2.5)
96
Figure 4.29 Time Needed to Reach SteadyState for Different Cylindrical
Plate Thicknesses and for Different Materials (Re=750,
ß=2.5)
97
xi
List of Symbols
b Plate thickness, r
o
 r
i,
R
o
 R
i,
[m]
d
n
Diameter of the nozzle [m]
F
o
Fourier number, α
f
t/ d
n
2
g Acceleration due to gravity [m/s
2
]
h Heat transfer coefficient [W/m
2
K], q
int
/(T
int
T
j
)
h
av
Average heat transfer coefficient [W/m
2
K], defined by Eqn.(14)
H
n
Distance of the nozzle from the point of impingement [m]
k Thermal conductivity [W/m K]
Nu Nusselt number, (h⋅d
n
)/k
f
Nu
av
Average Nusselt number for the entire surface, (h
av
⋅d
n
)/k
f
Nu
max
Maximum Nusselt number for the entire surface, (h
max
⋅d
n
)/k
f
n Coordinate normal to the surface
p Pressure [Pa]
Pr Prandtl number, ν
f
/α
f
q Heat flux [W/m
2
]
Q Fluid flow rate [m
3
/s]
r Radial coordinate [m]
r
i
Inner radius of hemisphere [m]
r
o
Outer radius of hemisphere [m]
xii
R Radius of curvature [m]
Re Reynolds number, (V
J
⋅d
n
)/ν
f
R
i
Inner radius of curvature of the plate [m]
R
o
Outer radius of curvature of the plate [m]
s Coordinate along the arc length, R
o
Φ, r
o
Φ [m]
T Temperature [K]
T
o
Constant Temperature at the inner surface of hemisphere (isothermal
case) [K]
t Time [s]
V
J
Jet velocity [m/s]
V
r, z
Velocity component in the r, zdirection [m/s]
V
x, y
Velocity component in the x, ydirection [m/s]
x coordinate along xaxis [m]
y coordinate along yaxis [m]
z Axial coordinate [m]
Greek Symbols:
α Thermal diffusivity [m
2
/s]
β Dimensionless nozzle to target spacing, H
n
/d
n
δ Liquid film thickness [m]
ν Kinematic viscosity [m
2
/s]
θ Angular coordinate [rad]
Θ Dimensionless temperature (constant heat flux boundary condition), 2⋅k
f
⋅(T
int
T
J
)/ (q⋅d
n
)
xiii
Φ Angular coordinate for curved plate [rad]
Φ Azimuthal coordinate for hemisphere [rad]
ρ Density [kg/m
3
]
σ Surface tension [N/m]
Ω Dimensionless temperature (isothermal boundary condition), (T
int
T
J
)/ (T
o

T
J
)
Subscripts:
atm Ambient
av Average
f Fluid
i Initial Condition
int Solidfluid Interface
j Jet or inlet
max Maximum
n Nozzle
s solid
SS Steady State
w Inner surface of hemisphere
xiv
Numerical Analysis of Heat Transfer During Jet Impingement on Curved
Surfaces
Cesar F. HernandezOntiveros
ABSTRACT
The flow structure and convective heat transfer behavior of a free liquid jet
ejecting from a round nozzle impinging vertically on a hemispherical solid plate
and a slot nozzle impinging vertically on a cylindrical curved plate have been
studied using a numerical analysis approach. The simulation model incorporated
the entire fluid region and the solid hemisphere or curved plate.
Solution was done for both isothermal and constant heat flux boundary
conditions at the inner surface of the hemispherical plate and the constant heat
flux boundary condition at the inner surface of the cylindrical shaped plate.
Computations for the round nozzle impinging jet on the hemispherical plate and
cylindrical plate were done for jet Reynolds number (Re
J
) ranging from 500 to
2000, dimensionless nozzle to target spacing ratio (β) from 0.75 to 3, and for
various dimensionless plate thicknesses to diameter nozzle ratio (b/d
n
) from
0.0831.5. Also, computations for the slot nozzle impinging jet on the cylindrical
plate were done for inner plate radius of curvature to nozzle diameter ratio (R
i
/d
n
)
of 4.1616.66, plate thickness to nozzle diameter ratio (b/d
n
) of 0.081.0, and
different nozzle diameters (d
n
), Results are presented for dimensionless solid
fluid interface temperature, dimensionless maximum temperature in the solid,
xv
local and average Nusselt numbers using the following fluids: water (H
2
O),
flouroinert (FC77), and oil (MIL7808) and the following solid materials:
aluminum, copper, Constantan, silver, and silicon.
Materials with higher thermal conductivity maintained a more uniform
temperature distribution at the solidfluid interface. A higher Reynolds number
increased the Nusselt number over the entire solidfluid interface. Local and
average Nusselt number and heat transfer coefficient distributions showed a
strong dependence on the impingement velocity or Reynolds number. As the
velocity increases, the local Nusselt number increases over the entire solidfluid
interface. Decreasing the nozzle to target spacing favors the increasing of the
Nusselt number. Increasing the nozzle diameter decreases the temperature at
the curved plate outer surface and increases the local Nusselt number. Similarly,
local and average Nusselt number was enhanced by decreasing plate thickness.
Numerical simulation results are validated by comparing with experimental
measurements and related correlations.
1
Chapter 1 Introduction
The implementation of jet impingement due to its high convective heat
transfer rate in mechanical and chemical processes has led to many industry
applications. These include cooling and heating of metal plates, cleaning of iced
aircraft wings, cooling of turbine blades, drying of wetted surfaces, cleaning of
industrial equipment, and cooling of electrical equipment. The advantage of
using jets translates into reduced operational cost by improving the cooling
efficiency of the system or equipment and reduces early failure of it. Thus, the
need for heating/cooling processes able to transfer or remove very high heat
fluxes appeals to liquid jet impingement as a choice. Enhancement of such
heating/cooling processes will require information on the effects of jet velocity,
thermophysical properties of fluid and solid, thickness of the solid, and height
from nozzle to impingement surface. Freesurface jets are created when a liquid
discharges into ambient air or other type of gaseous environment [1]. The free
surface, then, begins to form instantly at the nozzle exit and remains throughout
the impingement section and wall region as the fluid moves downstream along
the plate after impingement. The structure of the free surface depends on
surface tension, gravitational and pressure forces.
2
1.1 Literature Review
1.1.1 Round Jet Impingement
An early investigation involving liquid jet impingement using round nozzles
was carried out by Stevens and Webb [2] who considered an axisymmetric,
singlephase liquid jet impinging on a flat uniformly heated surface. This
experimental study investigated the effects of Reynolds number, nozzle to plate
spacing, and jet diameter. Garimella and Rice [3] experimentally measured the
local heat transfer from a small heat source to a normally impinging,
axisymmetric and submerged liquid (FC77) jet from a round nozzle for a range
of Reynolds number of 4,00023,000 and nozzle to heat source spacing ratios of
1 to 14. They concluded that secondary peaks in the local heat transfer
coefficient resulted form increasing nozzle diameters for a given Reynolds
number. Gomi and Webb [4] performed an investigation on heat transfer from a
vertical heated surface to an obliquely impinging circular freesurface jet of
transformer oil for various Reynolds number (2351,746) with jet angles from 45°
to 90°. They concluded that the maximum heat transfer coefficient was found to
decrease with increasing jet inclination. Lee et al. [5] studied heat transfer from a
convex surface with low curvature using liquid crystals to measure the local
surface temperature at Reynolds numbers ranging from 11,00050,000. The
experimental study concluded that the stagnation point Nusselt number
increases with increasing surface curvature. Kornblum and Goldstein [6]
analyzed the flow of circular jets impinging on semicylindrical surfaces (convex
and concave) for relatively small jet to semicylindrical diameter ratios ranging
3
from 0.01970.0394 using a flow visualization technique. Lee et al. [7] employed
an apparatus consisting of various components (heat exchanger, flow meter, and
air blower) to target a convex hemispherical surface for Re=11,00087,000 at
different dimensionless nozzletosurface distance and at constant hemisphere
tonozzle diameter ratio to obtain information on the stagnation and local Nusselt
numbers. Tong [8] numerically studied convective heat transfer of a circular
liquid jet impinging onto a substrate to understand the hydrodynamics and heat
transfer of the impingement process using the volumeoffluid method to track the
free surface of the jet.
Cornaro et al. [9] visualized fluid flow using a smokewire technique on
concave and convex surfaces to examine the effects of curvature, nozzle to
surface distance, and Reynolds number. The convex surface was examined for
jet diameters of 0.0472, 0.0726, and 0.0986 m with Reynolds numbers ranging
from 6,00015,000. In addition, Cornaro et al. [10] showed the effect of
increasing Reynolds number on the local Nusselt number for different types of
curvature. Fleischer et al. [11] used a smokewire flow visualization technique to
investigate the behavior of a round jet impinging on a convex surface. They
showed the initiation and development of ring vortices and their interaction with
the cylindrical surface. The effects of Reynolds number, jet to surface distance,
and relative curvature were studied. Baonga et al. [12] experimentally studied
the hydrodynamic and thermal characteristics of a free round liquid jet impinging
into a heated disk for nozzle to plate spacing of 312 times nozzle diameter and
for Reynolds number of 6009,000.
4
1.1.2 Slot Jet Impingement
A number of past studies have considered liquid jet impingement from a
twodimensional slot nozzle to a heated flat plate. One of the early investigations
on jet impingement on a flat plate was conducted by Inada et al. [13] who studied
laminar flow between a plane surface and a twodimensional water jet with
constant heat flux using the RungeKutta method to obtain solutions for the
boundary layer momentum and energy equations. Similar analytical solution to
the momentum equation using the first order CauchyEuler ordinary differential
equation and the integral equation approach to solve for the energy equation was
employed by Carper [14]. He attempted to obtain a solution for the ratio of
thermal to hydrodynamic boundary layer thickness for a certain range of
combinations of jet Prandtl number and jet Reynolds number for an axisymmetric
liquid jet impinging on a flat surface. Also, Liu and Lienhard [15] analytically and
experimentally investigated convective heat transfer to an impinging liquid jet
from a heated flat surface kept at a uniform heat flux. Solutions to the thermal
boundary layer and film thickness were approximated using the integral method
to obtain information about the local Nusselt number and temperature
distribution. They concluded that the average Nusselt number for a constant
heat flux is larger than for a constant wall temperature. Also, the average
Nusselt number for a plane jet is generally larger than that for an axisymmetric
jet. Wadsworth and Mudawar [16] performed an experiment to investigate single
phase heat transfer from a smooth simulated chip to a twodimensional jet of
dielectric liquid FC72 delivered from a very thin rectangular slot jet into a
5
channel confined between the chip surface and nozzle plate. The experiment
studied the effects of jet width, confinement channel height, and impingement
velocity for Reynolds number ranging from 1,000100,000. Gau and Chung [17]
reported heat transfer for air jet impingement on semicylindrical curved surfaces
at low and high Reynolds numbers (6,00035,000) by varying slot widths to
surface diameter ratios (0.0220.125) using smoke particles as means to
visualize the air flow. The authors observed rotating vortices at the stagnation
line on the convex surface which increased the heat transfer characteristics of
the flow. Stevens and Webb [18] experimentally characterized the flow structure
under an impinging liquid jet hitting a flat surface. However, their research only
focused in measuring the velocity of the free surface of the jet. To achieve it,
they used a novel laserDoppler velocimetry technique to capture the fluctuations
of the free surface for Reynolds numbers ranging from 16,000 to 47,000.
Teuscher et al. [19] studied jet impingement cooling of an inline array of
discrete heat sources (fins of different configurations) by a submerged slot jet of
FC77. Experimental results of convective heat transfer were obtained for
various slot widths (0.2540.508 mm) and Reynolds number of 10500. Ma et al.
[20] measured heat transfer coefficients resulting from the impingement of
transformer oil jet issuing from various slot widths for fluid Prandtl numbers of
200270 and Reynolds number between 55 and 415. Bartoli and Faggiani [21]
experimentally studied heat transfer form a circular cylinder to a slot jet of water
at different Prandtl numbers for 3500≤Re≤20000 for tests performed at different
angles 0°≤ θ≤ 180° in order to obtain the local and average Nusselt number. Slot
6
jet impingement involving heat transfer from circular cylinders using air as the
cooling fluid was experimentally studied by McDaniel and Webb [22] for
Reynolds numbers in the range 6008000 varying cylinder diameter to jet width
spacings of 0.66, 1 and 2, and for jet exit to nozzle width spacing from 1 to 11.
Gori and Bossi [23] investigated heat transfer for the flow over the surface of a
cylinder for Reynolds number (400020,000) and determined that the distance
between the exit of the slot jet and the cylinder surface affects the local Nusselt
number. Kayansayan and Küçüka [24] performed an experimental and
numerical study of jet impingement cooling on a concave channel. The results of
the experimental investigation concluded that for a range of Reynolds numbers
from 200 to 11,000 and slot to surface spacing from 2.2 to 4.2 of the slot width
the heat transfer rates at the surface of the concave channel are improved due to
the curvature of the channel.
Rahman et al. [1] numerically studied the conjugate heat transfer during
impingement of a confined liquid jet. By varying slot width, impingement height,
and plate thickness for Reynolds number ranging from 4451,545. Shi et al. [25]
carried out a numerical study to examine the effects of Prandtl number (0.771),
nozzle to target spacing (210) and Reynolds number ranging from 0.01100 on
heat transfer under a semiconfined laminar slot jet. Olsson et al. [26] simulated
heat transfer from a slot air jet impinging on a cylinder at various Reynolds
numbers ranging from 23,000100,000 and varying jet to cylinder distances and
cylinder curvatures. They stated that the flow characteristics and the heat
7
transfer distribution around cylinders are found to be dependent on the distance
and the opening between the jets.
Chan et al. [27] employed a liquid crystal thermographic system to
experimentally determine the effects of jet Reynolds number, dimensionless slot
nozzle width to impingement surface distance ratio, and slot nozzle width to
circumferential distance on the local heat transfer for an air impinging slot jet on a
semicircular convex surface. Chan et al. [28] measured the mean flow and
turbulence of a turbulent air slot jet impinging on two different semicircular
convex surfaces at Reynolds number 12,000 using a hotwire Xprobe
anemometer. Gori and Bossi [29] experimentally determined the optimal height
in the jet cooling of an electrically heated circular cylinder for various Reynolds
numbers measuring mean and local Nusselt numbers. Rahimi et al. [30]
investigated an underexpanded jet impinging on a heated cylindrical surface
varying the nozzle to surface spacing (310) for various Reynolds numbers.
Yang and Hwang [31] carried out a numerical simulation of flow of a turbulent slot
jet impinging on a semicylindrical convex surface for Reynolds number (6,000
20,000) of the inlet flow and by varying slot jet width to jet to surface distance
ratio.
Chen et al. [32] performed a theoretical analysis to characterize heat
transfer from horizontal surfaces to single phase free surface laminar slot jets
using a heat flux condition for different working fluids and different nozzle sizes.
Zuckerman and Lior [33] employed numerical models to understand the heat
transfer behavior on circular cylinders cooled by radial slot jets. These models
8
attempted to simulate a cylinder exposed to a radial array of slot jets (28) for
Reynolds number ranged from 5,000 to 80,000 and different target diameter to
nozzle hydraulic diameter. They concluded that the highest average Nusselt
number would occur when having a lower number of jets.
1.2 Overview of Literature
The literature mentioned above describes most studies to date on jet
impingement over a curved or flat surface using air or liquid as the working fluid
and most of them have considered the steady state condition only. A few
attempted to obtain local heat transfer distribution of concave, convex, or flat
surfaces taking into account the transient nature of the problem. Therefore, a
through study of liquid jet impingement on a curved surface is needed as this is
encountered in many industrial processes.
1.3 Thesis Aim
The present study attempts to carry out a comprehensive numerical
investigation of steady and transient local conjugate heat transfer for laminar free
surface jet impingement over a hollow hemisphere and a curved cylindrical
shaped plate. Computations using water (H
2
O), flouroinert (FC77), oil (MIL
7808), and ammonia (NH
3
) as working fluids were carried out for different flow
configurations, plate configurations, and different plate materials. The
computations were carried out using the finite element method using FIDAP
version 8.6. software package focused in computational fluid dynamics (CFD) of
9
the FLUENT corporation. The finite element method (FEM) is used for finding
approximate solution of complex partial differential equations (PDE) as well as of
integral equations. The solution method uses a mesh domain (elements grid) to
solve the differential equation completely by applying to each element a system
of simultaneous equations. The system of equations is solved for unknown
values using the techniques of linear algebra or nonlinear numerical schemes,
as appropriate, or rendering the PDE into an equivalent ordinary differential
equation, which is then, solved using finite difference. This approach differs from
the finite element method (FEM) such that it uses finite differences schemes
(forward, backward, and central) to approximate derivatives, being the central
difference the one that provides the more accurate approximation. Also, the
finite difference method is an approximation to the differential equation whereas
the finite element method is an approximation to its solution. One of the most
attractive features of the FEM over the FDM is its ability to handle complex
geometries and boundaries (moving boundaries) with relative ease. That is the
reason why the FEM was chosen as the preferred method to solve for the
mathematical model presented in this study.
10
Chapter 2 Mathematical Models and Computation
2.1 Hemispherical Model
The physical hemispherical model studied here is threedimensional (3D)
by nature as depicted in Figure 2.1. However, the problem can be greatly
simplified into an axisymmetric liquid jet model that impinges on the outer surface
of a hollow hemisphere subjected to an isothermal or constant heat flux boundary
condition at its inner surface. Thus, the free jet discharges from the round nozzle
and impinges perpendicularly at the top center of the hemisphere while the
hemisphere is dissipating heat from within. Figure 2.2 shows a 2D cross
sectional view of the system including the origin and axes used to write the
boundary conditions. The fluid is Newtonian and the flow is incompressible and
axisymmetric.
11
Figure 2.1 Three Dimensional Schematic of a Hollow Hemisphere During an
Axisymmetric Liquid Jet Impingement
Figure 2.2 CrossSectional View of a Hollow Hemisphere During an Axisymetric Jet
Impingement
12
2.1.1 Governing Equations: Steady State Heating
The equations describing the conservation of mass, momentum (r and z
directions respectively), and energy can be written as [34]:
0 V = ⋅ ∇
r
(2.1.1)
( )
(
¸
(
¸
− ∇ +
∂
∂
− = ∇ ⋅
2
r
r
V
r
V
2
ν
r
p
f
ρ
1
r
V V
r
(2.1.2)
( )
z
2
f
z
V ν
z
p
ρ
1
g V V ∇ +
∂
∂
− − = ∇ ⋅
r
(2.1.3)
( )
f
2
f f
T α T V ∇ = ∇ ⋅
r
(2.1.4)
The conservation of energy inside the solid can be characterized by the following
equation:
0 T
s
2
= ∇ (2.1.5)
2.1.2 Boundary Conditions: Steady State Heating
The following boundary conditions were used:
0
r
T
: 0 z b 0, r At
s
=
∂
∂
≤ ≤ − = (2.1.6)
0
r
T
, 0
r
V
, 0 V : H z 0 , 0 At
f Z
r n
=
∂
∂
=
∂
∂
= ≤ ≤ = r
(2.1.7)
0
z
T
: r , r r r At
S
o o
=
∂
∂
= ≤ ≤ z
i
(2.1.8)
atm o o o
p p : r z δ), (r r r At = = + ≤ ≤ (2.1.9)
J f r J Z
n
n
T T , 0 V , V V :
2
d
r 0 , H At = = − = ≤ ≤ = z (2.1.10)
13
At inner surface of hemisphere:
o
T T or q
r
S
T
s
k :
2
π
0 = =
∂
∂
− ≤ Φ ≤ (2.1.11)
At outer surface of hemisphere:
n
f
T
f
k
n
S
T
s
k ,
S
T
f
T 0,
Z
V
r
V :
2
π
0
∂
∂
=
∂
∂
= = = ≤ Φ ≤ (2.1.12)
The boundary condition at the free surface 
¹

\

≤ Φ ≤
2
π
2
d
n
included the
kinematic condition and balance of normal and shear stresses. The kinematic
condition related velocity components to local slope of the free surface. The
normal stress balance took into account the effects of surface tension. In the
absence of any significant resistance from the ambient gas, the shear stress
encountered at the free surface is essentially zero. Similarly, a negligible heat
transfer at the free surface results in zero temperature gradient.
The local and average heat transfer coefficients can be defined as:
) (
J
T T
q
−
=
int
h
(2.1.13)
) (
) ( Φ Φ ⋅ −
−
=
∫
d sin
1
h
2
0
int
int
av
π
J
J
T T h
T T
(2.1.14)
where int T is the average temperature at the solidliquid interface. The local and
average Nusselt numbers are calculated according to the following expressions.
f
n
k
d h
Nu
⋅
=
(2.1.15)
14
f
n av
av
k
d h
Nu
⋅
= (2.1.16)
2.1.3 Governing Equations: Transient Heating
A liquid jet axially discharging from a round nozzle impinges on the outer
surface of a hollow hemisphere subjected to a constant heat flux boundary
condition at its inner surface. At t=0 the power source is turned on and heat
begins to flow only after an initially isothermal fluid flow has been established on
the hemisphere. The fluid is Newtonian and the flow is incompressible and
axisymmetric. Therefore, the equations describing the conservation of mass,
momentum (r and z directions respectively), and energy [34]:
0 V = ⋅ ∇
r
(2.1.17)
( )
(
¸
(
¸
− ∇ +
∂
∂
− = ∇ ⋅
2
r
r
V
r
V
2
r
p
f
ρ
1
r
V V ν
r
(2.1.18)
( )
z
2
f
z
V
z
p
ρ
1
g V V ∇ +
∂
∂
− − = ∇ ⋅ ν
r
(2.1.19)
( )
f
2
f f
T α T V ∇ = ∇ ⋅ +
∂
∂
r
t
T
f
(2.1.20)
The conservation of energy inside the solid can be characterized by the following
heat conduction equation:
t
T
T
s
s
∂
∂
= ∇
2
α (2.1.21)
15
2.1.4 Boundary Conditions: Transient Heating
The following boundary conditions were used:
0
r
T
: 0 z b 0, r At
s
=
∂
∂
≤ ≤ − = (2.1.22)
0
r
T
, 0
r
V
, 0 V : H z 0 , 0 At
f Z
r n
=
∂
∂
=
∂
∂
= ≤ ≤ = r
(2.1.23)
0
z
T
: r , r r r At
S
o o
=
∂
∂
= ≤ ≤ z
i
(2.1.24)
atm o o o
p p : r z δ), (r r r At = = + ≤ ≤ (2.1.25)
J f r J Z
n
n
T T , 0 V , V V :
2
d
r 0 , H At = = − = ≤ ≤ = z (2.1.26)
At inner surface of hemisphere:
0 , q
r
S
T
s
k :
2
π
0 ≥ =
∂
∂
− ≤ Φ ≤ t (2.1.27)
At outer surface of hemisphere:
n
f
T
f
k
n
S
T
s
k ,
S
T
f
T 0,
Z
V
r
V :
2
π
0
∂
∂
=
∂
∂
= = = ≤ Φ ≤ (2.1.28)
The plate was assumed to be at thermal equilibrium with jet fluid before the
transient heating of the plate was turned on. The velocity field at this condition
was determined by solving only the continuity and momentum equations in the
fluid region. Thus,
At t = 0:
T
f
= T
s
=
T
j
,
V
i
= V
r, z
(2.1.29)
16
2.2 Cylindrical Plate Model
The physical model corresponds to a twodimensional symmetric liquid jet
that impinges on the outer surface of a curved hollow cylindrical shaped plate
subjected to a uniform heat flux boundary condition at the inner surface as shown
in Figure 2.3. The fluid is Newtonian and the flow is incompressible and
symmetric about the midplane. Figure 2.3 also shows the origin and axes used
to write the boundary conditions.
Figure 2.3 Schematic View of a Curved Plate During a 2D Symmetric Liquid Jet
Impingement
17
2.2.1 Governing Equations: Steady State Heating
The equations describing the conservation of mass, momentum (x and y
directions respectively), and energy using a 2D coordinate system can be
written as [34]:
0 V = ⋅ ∇
r
(2.2.1)
( )
y
V
2
µ
y
p
f
ρ
1
y
g
y
V V ∇ +
∂
∂
− − = ⋅ ∇ ⋅
r
(2.2.2)
( )
x
V
2
µ
x
p
f
ρ
1
x
V V ∇ +
∂
∂
− = ⋅ ∇ ⋅
r
(2.2.3)
( )
f
T
2
f
α
f
T V ∇ = ⋅ ∇ ⋅
r
(2.2.4)
The conservation of energy inside the solid can be characterized by the following
equation:
0 T
s
2
= ∇ (2.2.5)
2.2.2 Boundary Conditions: Steady State Heating
The following boundary conditions were used:
0
x
T
: 0 y b  0, x At
s
=
∂
∂
≤ ≤ = (2.2.6)
0
x
T
0,
x
V
0, V : H y 0 0, x At
f
y
x n
=
∂
∂
=
∂
∂
= ≤ ≤ = (2.2.7)
0
y
T
: x , R x R At
s
o o
=
∂
∂
− = ≤ ≤ y
i
(2.2.8)
atm o
p p y = = + ≤ ≤ : x ), R ( x R At
o o
δ (2.2.9)
18
J f x J y
n
n
T T , 0 V , V V :
2
d
x 0 , H At = = − = ≤ ≤ = y (2.2.10)
At inner surface of plate:
q
n
s
T
s
k :
max i
R s 0 =
∂
∂
− ≤ ≤ φ (2.2.11)
At outer surface of plate:
n
f
T
f
k
n
s
T
s
k ,
f
T
s
T 0,
y
V
x
V :
max
o
R s 0
∂
∂
=
∂
∂
= = = ≤ ≤ φ (2.2.12)
The boundary condition at the free surface included the kinematic condition and
balance of normal and shear stresses. The kinematic condition related velocity
components to local slope of the free surface. The normal stress balance took
into account the effects of surface tension. In the absence of any significant
resistance from the ambient gas, the shear stress encountered at the free
surface is essentially zero. Similarly, a negligible heat transfer at the free surface
results in zero temperature gradient.
) (
J
T T
q
−
=
int
h (2.2.13)
) (
) ( φ φ
π
φ
d sin
4
h
max
0
int
int
av
∫
⋅ −
−
=
J
J
T T h
T T
(2.2.14)
where int T is the average temperature at the solidliquid interface. The local and
average Nusselt numbers are calculated according to the following expressions
f
n
k
d h
Nu
⋅
=
(2.2.15)
f
n av
av
k
d h
Nu
⋅
= (2.2.16)
19
2.2.3 Governing Equations: Transient Heating
The physical model corresponds to a twodimensional symmetric liquid jet
that impinges on the outer surface of a curved hollow cylindrical shaped plate
subjected to a uniform heat flux boundary condition at the inner surface. At t=0
the power source is turned on and heat begins to flow only after an initially
isothermal fluid flow has been established on the plate. The fluid is Newtonian
and the flow is incompressible and symmetric about the midplane. The
equations describing the conservation of mass, momentum, and energy in the
fluid region can be written as [34]:
0 V = ⋅ ∇
r
(2.2.17)
( )
y
V
2
µ
y
p
f
ρ
1
y
g
y
V V ∇ +
∂
∂
− − = ⋅ ∇ ⋅
r
(2.2.18)
( )
x
V
2
µ
x
p
f
ρ
1
x
V V ∇ +
∂
∂
− = ⋅ ∇ ⋅
r
(2.2.19)
( )
f
T
2
f
α
f
T V ∇ = ⋅ ∇ ⋅ +
∂
∂ r
t
f
T
(2.2.20)
The conservation of energy inside the solid can be characterized by the following
equation:
t
T
s
s
∂
∂
= ∇
s
2
T α
(2.2.21)
2.2.4 Boundary Conditions: Transient Heating
The following boundary conditions were used:
0
x
T
: 0 y b  0, x At
s
=
∂
∂
≤ ≤ = (2.2.22)
20
0
x
T
0,
x
V
0, V : H y 0 0, x At
f
y
x n
=
∂
∂
=
∂
∂
= ≤ ≤ = (2.2.23)
0
y
T
: x , R x R At
s
o o
=
∂
∂
− = ≤ ≤ y
i
(2.2.24)
atm o
p p y = = + ≤ ≤ : x ), R ( x R At
o o
δ (2.2.25)
J f x J y
n
n
T T , 0 V , V V :
2
d
x 0 , H At = = − = ≤ ≤ = y (2.2.26)
At inner surface of plate:
0 t q,
n
s
T
s
k :
max i
R s 0 ≥ =
∂
∂
− ≤ ≤ φ (2.2.27)
At outer surface of plate:
n
f
T
f
k
n
s
T
s
k ,
f
T
s
T 0,
y
V
x
V :
max
o
R s 0
∂
∂
=
∂
∂
= = = ≤ ≤ φ (2.2.28)
The plate was assumed to be at thermal equilibrium with jet fluid before the
transient heating of the plate was turned on. The velocity field at this condition
was determined by solving only the continuity and momentum equations in the
fluid region. Thus,
At t = 0:
T
f
= T
s
=
T
j
,
V
i
= V
x, y
(2.2.29)
21
2.3 Numerical Computation
2.3.1 Steady State Process
The governing equations in conjunction with the boundary conditions
described above were solved using the Galerkin finite element method employed
by FIDAP. Four node quadrilateral elements were used as shown in Figure 2.4.
Figure 2.4 Mesh Plot of a Curved Plate During a 2D Symmetric Liquid Jet Impingement
For every element, the velocity, pressure, and temperature fields were
approximated until convergence was achieved. The method to solve the set of
the resulting nonlinear equations was the NewtonRaphson algorithm due to its
22
coupled nature for two dimensional problems and its ability to solve all
conservation equations in a simultaneous manner. The approach used to solve
the free surface problem described here was to introduce a new degree of
freedom representing the position of the free surface. This degree of freedom
was introduced as a new unknown into the global system of equations. A scaled
dense grid distribution was used to adequately capture large variations near the
solidfluid interface of the meshed domain. Due to nonlinear nature of the
governing transport equations, an iterative procedure was used to converge at
the solution for the velocity and temperature fields. Since the solution of the
momentum equation required only two out of the three boundary conditions at
the free surface, the third condition was used to upgrade the position of the free
surface at the end of each iteration step. The height of the free surface was
adjusted after each iteration by satisfying the kinematic condition that relates the
slope of the free surface to local velocity components at the free surface. The
NewtonRaphson solver employed spines to track the free surface and re
arranged grid distribution with the movement along the free surface. The spines
are straight lines passing through the free surface nodes and connect the
remaining nodes underneath the free surface. The movement of the free surface
affected only the nodes along the spine. The solution was considered converged
when relative change in field values from a particular iteration to the next, and the
sums of the residuals for each variable became less than 10
6
.
23
2.3.2 Transient Process
In order to determine the initial velocity field (V
i
), the equations for the
conservation of mass and momentum were solved. Once the initial free surface
height distribution and the flowfield for the isothermal equilibrium condition were
satisfied, the power of the heat source was turned on and heat began to flow.
The computation domain included both solid and fluid regions, and continuity,
momentum, and energy were solved simultaneously as a conjugate problem.
The height of the free surface was adjusted after each iteration by satisfying the
kinematic condition that relates the slope of the free surface to local velocity
components at the free surface. The solution was considered converged when
relative change in field values from a particular iteration to the next, and the sums
of the residuals for each variable became less than 10
6
. The computation
continued towards the steady state condition; however, because of large
changes at the outset of the transient and very small changes when the solution
approached the steadystate condition, a variable time step was used for the
computation.
24
2.4 Mesh Independence and Time Step Study
2.4.1 Cylindrical Coordinates
Several grids or combinations of number of elements were used to
determine the accuracy of the numerical solution as shown in Figure 2.5. The
numerical solution becomes grid independent when the number of divisions
equal to 40x119 in the axial (z) (radial in thin film after impingement) and arc (Φ)
directions respectively is used. Numerical results for a 40x119 grid gave almost
identical results compared to those using 34x79 and 40x87 grids. A quantitative
difference in grid independence was calculated using the following expression:
e
N
D
C
int
T + = (2.3.1)
C, D, and e are constants to be evaluated. N represents the number of divisions
along a chosen axis. T
int
is the solidfluid interface temperature at a given Φ
location of the hemispherical plate. Thus, equation (2.3.1) has three unknowns
at three sets of interface temperatures taken at three different grid sizes. This
results in a set of nonlinear equations with three variables. The value of e was
obtained based on doing a particular number of iterations. At s or Φ=0.001276
m, the value of e was calculated to be 8.554. Once the value of e had been
obtained, it was substituted into the system of equations and the values of C and
D were calculated to be 318.9112 and 0.72205E19 respectively. In addition, the
percentage error is calculated used the following expression:
x100
C
C
int
T −
(2.3.2)
25
The error at Φ=0.001276 m was found to be 3.987% for 40x119 grids and 4.26%
for 48x87 grids.
Figure 2.5 Local Dimensionless Interface Temperature for Different Number of Elements
in z (or r) and Φ Directions for Water as Fluid and Silicon as Solid (Re=750,
b/d
n
=0.5, ß=2.0, q=250 kW/m
2
)
Figure 2.6 plots the transient variation of the maximum dimensionless
temperature encountered at the hemisphere’s outer surface (solidfluid interface)
for different time increments as a function of time using the Fourier number (F
o
)
as a dimensionless number to represent time. Observe that the simulation is not
very susceptible to the time increments chosen. However, a very small time
increment (0.00001 s) ensures the initial condition to be revealed. For this
numerical study, an appropriate time increment of 0.01 s was selected in order to
accomplish a smooth variation. Notice how the maximum dimensionless
320
325
330
335
340
345
350
355
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
S
o
l
i
d

F
l
u
i
d
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
K
nzxnΦ = 22x57
nzxnΦ = 26x57
nzxnΦ = 30x79
nzxnΦ = 34x79
nzxnΦ = 40x119
nzxnΦ = 40x87
26
temperature increases rapidly as the time increases all the way to the steady
state condition.
Figure 2.6 SolidFluid Interface Dimensionless Maximum Temperature Variation for
Silicon Hemisphere at Different Time Increments (Re=750, b/d
n
=0.5, ß=2.0,
H
n
= 0.30 cm, q=250 kW/m
2
)
2.4.1 Cartesian Coordinates
Several grids or combinations of number of elements were used to
determine the accuracy of the numerical solution as shown in Figure 2.7. The
numerical solution becomes grid independent when the number of divisions
equal to 30x132 in the y and x or φ (in thin film after impingement) directions is
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Fourier Number, Fo
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
a
t
t
h
e
I
n
t
e
r
f
a
c
e
,
Θ Θ Θ Θ
m
a
x
Time Increment = 0.00001 s
Time Increment = 0.01 s
Time Increment = 0.03 s
Time Increment = 0.05 s
Time Increment = 0.08 s
Time Increment = 0.1 s
Time Increment = 0.2 s
27
used. Numerical results for a 30x132 grid gave almost identical results compared
to those using 32x132 and 26x92 grids. The average difference was 0.0466%.
Also, a quantitative difference in grid independence was carried on using
equation (2.3.1). At s or φ=0.001277 m, the value of e was calculated to be
8.554. Once the value of e had been obtained, it was substituted into the system
of equations and the values of C and D were calculated to be 338.2391 and 
0.7070E19 respectively. The error percentage is calculated from equation (2.3.2)
and found to be 1.51% for 32x132 grids and 1.63% for 26x92 grids at
φ=0.001277 m.
Figure 2.7 Local Dimensionless Interface Temperature for Different Number of Elements
in x or φ, and y Directions (Re=750, ß=2.5, H
n
= 0.30 cm)
326
328
330
332
334
336
338
340
342
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
S
o
l
i
d

F
l
u
i
d
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
K
nyxnx or nΦ=10x32
nyxnx or nΦ=16x32
nyxnx or nΦ=20x92
nyxnx or nΦ=26x92
nyxnx or nΦ=30x132
nyxnx or nΦ=32x132
28
Figure 2.8 shows the transient variation of the maximum dimensionless
temperature found at the plate’s outer surface (solidfluid interface) for different
time increments as a function of time using the Fourier number (F
o
)
as a
dimensionless number to represent time. Note that the numerical simulation is
not very susceptible to the time increments chosen. For this study, a time
increment of 0.01 s was chosen in order to obtain a smooth variation for
temperature. Notice how the maximum dimensionless temperature at the
transient increases rapidly all the way to the steadystate condition.
Chapter 3
Figure 2.8 SolidFluid Interface Dimensionless Maximum Temperature Variation for
Silicon Plate at Different Time Increments (Re=750, b/d
n
=0.5, ß=2.5, H
n
=
0.30 cm, q=250 kW/m
2
)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 0.05 0.1 0.15 0.2 0.25 0.3
Fourier Number, Fo
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
a
t
t
h
e
I
n
t
e
r
f
a
c
e
,
Θ Θ Θ Θ
i
n
t
Time Increment = 0.00001 s
Time Increment = 0.01 s
Time Increment = 0.03 s
Time Increment = 0.05 s
Time Increment = 0.08 s
Time Increment = 0.1 s
Time Increment = 0.2 s
29
Chapter 3 Hemispherical Model Results
3.1 Steady State Heating
A typical velocity vector distribution is shown in Figure 3.1. It can be
observed that the velocity remains almost uniform at the potential core region of
the jet. However, the velocity decreases and the fluid jet diameter increases as
the fluid gets closer to the surface during the impingement process. The direction
of motion of the fluid particles shifts along the angle of curvature, afterwards, the
fluid accelerates creating a region of minimum sheet thickness. That is the
beginning of the boundary layer zone. It can be noticed that as the fluid moves
downstream along the convex surface, the boundary layer thickness increases
and the frictional resistance from the wall is eventually transmitted to the entire
film thickness where the fully viscous zone develops.
30
Figure 3.1 Velocity Vector Distribution for Silicon Hemisphere with Water as the Cooling
Fluid (Re = 750, b = 0.60 mm, Q = 5.665x10
7
m
3
/s, β = 2.0, H
n
= 0.30 cm, q
= 250 kW/m
2
)
A pressure distribution plot is shown in Figure 3.2. As it is seen, the
pressure at the impingement regions is higher due to the fluid impacting the
hemisphere outer surface in comparison with the remaining portion of the
hemisphere. A pressure gradient distribution is more uniform, however, along
the arc length of the solid.
31
Figure 3.2 Pressure Distribution for Silicon Hemisphere with Water as the Cooling Fluid
(Re = 750, b = 0.60 mm, Q = 5.665x10
7
m
3
/s, β = 2.0, H
n
= 0.30 cm, q = 250
kW/m
2
)
Figure 3.3 depicts temperature distribution along the solid and fluid region.
Higher temperatures are encountered along the end of the arc length of the solid
fluid interface whereas lower temperatures are encountered at the stagnation
region. After impact has occurred, the particles start to accelerate towards the
remaining portion of the hemisphere, hence, increasing the velocity boundary
layer and temperature boundary layer thickness.
32
Figure 3.3 Temperature Distribution for Silicon Hemisphere with Water as the Cooling
Fluid (Re = 750, b = 0.60 mm, Q = 5.665x10
7
m
3
/s, β = 2.0, H
n
= 0.30 cm, q
= 250 kW/m
2
)
Figure 3.4 shows the free surface height distribution for different Reynolds
numbers when the jet strikes the hemisphere’s surface. It can be seen that the
fluid spreads out moving along the outer surface of the hemisphere. As the
Reynolds number increases, the film diminishes in thickness due to a larger
impingement velocity that translates to a higher fluid velocity in the film. For the
conditions considered in the present investigation, the flow was supercritical and
a hydraulic jump did not occur within the computation domain. These
observations concur with the experimental work of Stevens and Webb [35] for
free jet impingement on a flat surface. It was observed that the velocity remains
33
almost uniform at the potential core region of the jet. However, the velocity
decreases and the fluid jet diameter increases as the fluid get closer to the
surface during the impingement process. The direction of motion of the fluid
particles shifted along the angle of curvature, afterwards, the fluid accelerated
creating a region of minimum sheet thickness. That was the beginning of the
boundary layer zone. It was noticed that as the fluid moves downstream along
the convex surface, the boundary layer thickness increases and the frictional
resistance from the wall is eventually transmitted to the entire film thickness
where the fully viscous zone developed. The three different regions observed in
the present investigation agreed with the experiments of Liu et al. [36] for jet
impingement over a flat surface.
Figure 3.4 Free Surface Height Distribution for Different Reynolds Numbers and
Water as the Cooling Fluid (b=0.60 mm, β=2.0, H
n
=0.30 cm, q=250
kW/m
2
)
0.52
0.46
0.4
0.34
0.28
0.22
0.16
0.1
0.04
0.02
0.08
0.14
0.2
0.26
0.32
0 0.06 0.12 0.18 0.24 0.3 0.36 0.42 0.48 0.54 0.6
Radial Distance (cm)
A
x
i
a
l
D
i
s
t
a
n
c
e
(
c
m
)
Re=500
Re=750
Re=1000
Re=1500
Re=2000
34
Figures 3.5 and 3.6 show dimensionless interface temperature and local
Nusselt number distributions as a function of dimensionless distance (s/d
n
) along
the solidfluid interface at different Reynolds numbers under an isothermal
heating condition. The curves in Figure 3.5 reveal that the dimensionless
interface temperature decreases with jet velocity or (Reynolds number). The
dimensionless interface temperature has the lowest value at s/d
n
= 0.4 and
increases along the arc length (s) reaching the highest value at the end of it. The
local Nusselt number depicted in Figure 3.6 increases rapidly over a small
distance (core region) measured from the stagnation point, reaching a maximum
around s/d
n
= 0.5, and then decreases along the radial distance as the boundary
layer develops further downstream. The location of the maximum Nusselt
number can be associated with the transition of the flow from the vertical
impingement to horizontal displacement where the boundary layer starts to
develop. Figures 3.5 and 3.6 confirm how an increasing Reynolds number
contributes to a more effective cooling by the enhancement of convective heat
transfer coefficient.
35
Figure 3.5 Dimensionless Interface Temperature Distribution for a Silicon Hemisphere at
Different Reynolds Numbers, and Water as the Cooling Fluid (β=2.0, H
n
= 0.30
cm, b=0.60 mm, T
o
=373 K)
Figure 3.6 Local Nusselt Number Distribution for a Silicon Hemisphere at Different
Reynolds Numbers, and Water as the Cooling Fluid (β=2.0, H
n
=0.30 cm,
b=0.60 mm, T
o
=373 K)
0.60
0.70
0.80
0.90
1.00
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Ω
Ω
Ω
Ω
i
n
t
Re=500
Re=750
Re=1000
Re=1500
Re=2000
0
15
30
45
60
75
90
105
120
135
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
Re=500
Re=750
Re=1000
Re=1500
Re=2000
36
Figures 3.7 and 3.8 show dimensionless interface temperature and local
Nusselt number distributions as a function of dimensionless distance (s/d
n
) along
the solidfluid interface at different Reynolds numbers under a constant heat flux
condition. The curves in Figure 3.7 reveal that the dimensionless interface
temperature decreases with jet velocity or (Reynolds number). The
dimensionless interface temperature has the lowest value at the stagnation point
(underneath the center of the axial opening) and increases along the arc length
(s) reaching the highest value at the end of it.
Figure 3.7 Dimensionless Interface Temperature Distribution for a Silicon Hemispherical
Plate at Different Reynolds Numbers, and Water as the Cooling Fluid (β=2.0,
H
n
= 0.30 cm, b=0.60 mm, q=250 kW/m
2
)
0
0.055
0.11
0.165
0.22
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
Re=500
Re=750
Re=1000
Re=1500
Re=2000
37
The local Nusselt number shows a very similar kind of distribution as the
case where the bottom surface of the hemispherical plate is kept at an isothermal
condition. The values are however different. This also shows that an increasing
Reynolds number contributes to a more effective cooling by the enhancement of
convective heat transfer coefficient.
Figure 3.8 Local Nusselt Number Distribution for a Silicon plate at Different Reynolds
Numbers, and Water as the Cooling Fluid (β =2.0 , H
n
= 0.30 cm, b=0.60 mm,
q=250 kW/m
2
)
Figure 3.9 plots the average Nusselt number and average heat transfer
coefficient as a function of Reynolds number. It may be noted that average
Nusselt number increases with Reynolds number. As the flow rate (or Reynolds
number) increases, the magnitude of fluid velocity near the solidfluid interface
that controls the convective heat transfer rate increases. It may be also noted
0
25
50
75
100
125
150
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, S/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
Re=500
Re=750
Re=1000
Re=1500
Re=2000
38
that effects of boundary condition is more at lower Reynolds number and the
curves come closer as the Reynolds number increases.
Figure 3.9 Average Nusselt Number and Heat Transfer Coefficient Variation for
Different Reynolds Numbers for Constant Heat Flux (q=250kW/m
2
) and
Isothermal (T
o
= 373 K) Boundary Conditions (β=2.0, H
n
= 0.30cm)
The solidfluid dimensionless interface temperature and local Nusselt
number distributions for six different nozzletotarget spacing for water as the
coolant and Reynolds number of 500 are shown in Figures 3.10 and 3.11
respectively.
0
5
10
15
20
25
30
400 600 800 1000 1200 1400 1600 1800 2000
Reynolds Number, Re
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
10000
15000
20000
25000
30000
35000
A
v
e
r
a
g
e
C
o
n
v
e
c
t
i
v
e
H
e
a
t
T
r
a
n
s
f
e
r
C
o
e
f
f
i
c
i
e
n
t
,
h
a
v
(
W
/
m
2
K
)
Isothermal Nu number
Constant heat flux Nu number
Isothermal h
Constant heat flux h
39
Figure 3.10 Dimensionless Interface Temperature for Silicon Hemisphere at Different
Nozzle to Target Spacing Ratio (β) for Water as the Cooling Fluid (Re=500,
Q=3.776x10
7
m
3
/s, d
n
=1.2mm, b=0.60 mm, q=250 kW/m
2
)
It may be noticed that the impingement height quite significantly affects
the dimensionless interface temperature as well as the Nusselt number only at
the stagnation region and the early part of the boundary layer region.
0.070
0.085
0.100
0.115
0.130
0.145
0.160
0.175
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
β=0.75
β=1.0
β=1.5
β=2.0
β=2.5
β=3.0
40
Figure 3.11 Local Nusselt Number for Silicon Hemisphere at Different Nozzle to Target
Spacing Ratio (β) for Water as the Cooling Fluid (Re=500, Q=3.776x10
7
m
3
/s, d
n
=1.2mm, b=0.60mm, q=250kW/m
2
)
At larger arc length the values get closer for all impingement heights. It is
quite expected since the impingement height essentially controls the change in
velocity the fluid particles encounter during the free fall from nozzle exit to target
hemisphere’s surface and therefore affects areas controlled by direct
impingement. Figure 3.12 shows dimensionless interface temperature for a
constant q. It can be observed that there is a larger variation of interface
temperature at a smaller thickness. As the thickness increases, the temperature
at the solidfluid interface becomes more uniform due to higher distribution of
heat within the solid by conduction.
0
10
20
30
40
50
60
70
80
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
β=0.75
β=1.0
β=1.5
β=2.0
β=2.5
β=3.0
41
It may be also noticed that the average temperature at the solidfluid
interface decreases when the thickness of the solid hemisphere increases. This
decrease in temperature is due to lower heat load at increased thickness. In the
present investigation, the outer radius of the hemisphere was kept constant ( to
preserve same β ) while the inner radius was varied to get different thicknesses.
Since the heat flux imposed in the inner surface of the hemisphere was also kept
constant, a smaller inner radius resulted in smaller heat input rate at that
boundary. In addition, the resistance of the material to the path of heat flow
increases with thickness.
Figure 3.12 Dimensionless Interface Temperature for Silicon Hemisphere at Different
Thickness (b) for Water as the Cooling Fluid (Re=750, Q=5.665x10
7
m
3
/s,
d
n
=1.2mm, b=0.60 mm, q=250 kW/m
2
)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
b/dn=0.083
b/dn=0.25
b/dn=0.5
b/dn=1.0
b/dn=1.5
42
Figure 3.13 plots the distribution of local Nusselt number along the surface
of the hemisphere for different values of wall thickness. Nusselt number changes
by only a small amount over the thicknesses considered in the present
investigation when silicon is used as the solid material. A larger peak Nusselt
number is obtained at a smaller thickness.
Figure 3.13 Local Nusselt Number for Silicon Hemisphere at Different Thickness (b) for
Water as the Cooling Fluid (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm,
b=0.60mm, q=250kW/m
2
)
A somewhat similar numerical results were obtained for the isothermal
boundary condition of the hemisphere at different thicknesses. Figure 3.14 shows
the dimensionless interface temperature for the isothermal case. It can be seen
that increasing the thickness contributes to a lower interface temperature at the
outer surface of the hemispherical plate. A larger thickness offers a larger
0
10
20
30
40
50
60
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
b=0.01
b=0.03
b=0.06
b=0.12
b=0.18
43
thermal resistance between inner and outer surface of the solid and therefore
results in larger temperature drop. Unlike the constant flux case, the variation of
interface temperature is more for a thicker plate. It may be also noticed that
temperature at the stagnation region is very significantly affected by plate
thickness. The local Nusselt number (Figure 3.15) along the plate’s interface
increased up to s/d
n
=0.9 as the plate became thinner. After that, Nusselt number
became almost independent of thickness variation and continued decaying along
the hemisphere’s outer surface.
Figure 3.14 Dimensionless Interface Temperature for Silicon Hemisphere at Different
Thickness (b) for Water as the Cooling Fluid (Re=750, Q=5.665x10
7
m
3
/s,
d
n
=1.2mm, b=0.60 mm, T
o
=373 K)
0.70
0.80
0.90
1.00
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Ω Ω Ω Ω
i
n
t
b=0.01
b=0.03
b=0.06
b=0.12
b=0.18
44
Figure 3.15 Local Nusselt Number for Silicon Hemisphere at Different Thickness (b) for
Water as the Cooling Fluid (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm,
b=0.60mm, T
o
=373 K)
Figure 3.16 compares the solidfluid interface temperature for the
isothermal boundary condition for the present working fluid (water) with two other
coolants that have been considered in previous thermal management studies,
namely flouroinert (FC77) and oil (MIL7808). It may be noticed that water
(T
J
=310 K) presents the lowest interface temperature when compared with FC77
(T
J
=273 K) and MIL7808 (T
J
=371 K). The highest Nusselt (Figure 3.17) number
is obtained when FC77 is used as the working fluid (T
j
=273.15 K). These results
were obtained for a constant Reynolds number of 750.
0
10
20
30
40
50
60
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
b=0.01
b=0.03
b=0.06
b=0.12
b=0.18
45
Figure 3.16 Dimensionless Interface Temperature for Silicon Hemispherical Plate for
Different Cooling Fluids (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60
mm, T
o
=373 K)
Figure 3.17 Local Nusselt Number for Silicon Hemisphere for Different Cooling Fluids
(Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60mm, T
o
=373 K)
0.88
0.90
0.92
0.94
0.96
0.98
1.00
1.02
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Ω Ω Ω Ω
i
n
t
H20(Pr=5.49)
FC77(Pr=31.80)
MIL7808(Pr=124.44)
0
20
40
60
80
100
120
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
Nu, H20(Pr=5.49)
Nu,FC77(Pr=31.80)
Nu,MIL7808(Pr=124.44)
46
Figure 3.18 compares the hemisphere’s solidfluid interface temperature
and local Nusselt number results of water with flouroinert (FC77) and oil (MIL
7808) for the constant heat flux boundary condition. It may be noticed that water
presents the lowest interface temperature and highest Nusselt number
distribution in comparison with FC77 and MIL7808. The lowest Nusselt number
is obtained when FC77 is used as the working fluid. These results were obtained
for a constant Reynolds number of 1500.
Figure 3.18 Local Nusselt Number and Dimensionless Interface Temperature for
Different Cooling Fluids (Re = 1500, Q=1.133x10
6
m
3
/s, β=2.0, H
n
=0.30 cm,
b=0.60 mm, q=250 kW/m
2
)
Figure 3.19 shows the dimensionless interface temperature and local
Nusselt number distribution plots as a function of a dimensionless radial distance
(s/d
n
) for different solid materials with water as the working fluid. The studied
0
25
50
75
100
125
150
175
200
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
0
0.11
0.22
0.33
0.44
0.55
0.66
0.77
0.88
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
Nu, H20(Pr=5.49)
Nu, FC77(23.66)
Nu, MIL7808(Pr=124.44)
H20 Temp
FC77 Temp
Mil7808 Temp
47
materials were silicon, silver, aluminum, copper, and Constantan, having different
thermophysical properties. Constantan shows the lowest dimensionless
temperature at the impingement zone and the highest at the outlet in comparison
with other solid materials. Copper and silver show a more uniform distribution
and higher temperature values at the impingement zone due to their higher
thermal conductivity. The dimensionless temperature and local Nusselt number
distributions of these two materials are almost identical due to their similar
thermal conductivity values. The crossover of curves for all five materials
occurred due to a constant fluid flow and heat flux rates that provide a thermal
energy balance. Solid materials with lower thermal conductivity show higher
maximum local Nusselt number. The choice of material is also crucial in
determining the magnitudes of these temperatures. A material with larger thermal
conductivity will facilitate a faster rate of heat transfer, and therefore will result in
a lower maximum temperature at the solidfluid interface and within the
hemispherical plate. The temperature difference at the interface is an indication
of the level of temperature nonuniformity at the impingement surface.
48
Figure 3.19 Local Nusselt Number and Dimensionless Interface Temperature for
Different Hemisphere Materials with Water as the Cooling Fluid (Re=1000,
β=2.0, b/d
n
=0.5, q=250 kW/m
2
)
Dimensionless interface temperature and local Nusselt number were also
investigated for various solid plate materials such as silicon, silver, aluminum,
copper, and Constantan using the same working fluid (water) for the isothermal
boundary condition. Figure 3.20 demonstrates that materials with higher thermal
conductivity such as silver and copper maintain a higher and more uniform
temperature distribution along the solidfluid interface transferring heat faster
towards the hemisphere’s outer surface. Constantan has significantly lower
interface temperature due to higher thermal resistance offered by it. Also, the
interface temperature varies over a larger range due to lower rate of conduction
heat transfer within the solid. In Figure 3.21, it can be observed that the local
0
20
40
60
80
100
120
140
160
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
Silver
Copper
Aluminum
Silicon
Constantan
Temp, Silver
Temp, Copper
Temp, Aluminum
Temp, Silicon
Temp, Constantan
49
Nusselt number changes only slightly with solid properties and curves are very
close to each other.
Figure 3.20 Dimensionless Interface Temperature for Silicon Hemispherical Plate for
Different Materials (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60 mm,
T
o
=373 K)
0.40
0.60
0.80
1.00
1.20
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Ω Ω Ω Ω
i
n
t
Silver
Copper
Aluminum
Silicon
Constantan
50
Figure 3.21 Local Nusselt Number Distribution for Silicon Plate for Different Solid
Materials (Re=750, Q=5.665x10
7
m
3
/s, d
n
=1.2mm, b=0.60mm, T
o
=373 K)
Three of the papers used for the validation of this numerical study were
the analytical works carried out by Liu et al. [36], Scholtz and Trass [37], and
Nakoryakov et al. [38] using fluid for Prandtl number greater than unity (Pr>1) as
coolants. The fluids were tested for heat removal under free liquid jet
impingement on a heated flat surface maintained at uniform heat flux. The
graphical representation of actual numerical Nusselt number results at the
stagnation point at different Reynolds number are shown in Figure 21. The
results shown at Figure 21 compared within 7.20% of Scholtz and Trass [37],
within 8.48% of Nakoryakov et al. [38], and within 7.76% of Liu et al. [36]. The
local Nusselt number under Reynolds number at 500, 750, 1000, 1250 and 1500
0
10
20
30
40
50
60
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
Silver
Copper
Aluminum
Silicon
Constantan
51
correlates with a margin of 12.17%, 5.65%, 3.35%, 6.00%, and 11.85%
respectively. Considering the errors inherent in any experimental study as well as
discretization and roundoff errors in the simulation, this comparison is quite
satisfactory. In addition, the surface curvature may have affected the heat
transfer characteristics near the impingement region that was not present in
experimental studies.
Figure 3.22 Stagnation Nusselt Number Compared with Liu et al. [36], Scholtz and
Trass [37], and Nakoryakov et al. [38] with Actual Numerical Results under
Different Reynolds Numbers (d
n
= 1.2mm, b=0.6mm, q=250kW/m
2
)
15
20
25
30
35
40
45
50
55
60
65
250 500 750 1000 1250 1500 1750
Reynolds Number, Re
S
t
a
g
n
a
t
i
o
n
N
u
s
s
e
l
t
n
u
m
b
e
r
,
N
u
Present Values
Liu et al.
Scholtz and Trass
Nakoryakov et al.
52
3.2 Transient Heating
In order to understand the thermal response of solid materials to the
flowing of heat when the power source is turned on (t > 0), it is necessary to
analyze the transient heat transfer process when a local heat flux travels
throughout the entire solid up to its outer surface and to the cooling fluid. The
local heat flux variation along the solidfluid interface for different time intervals is
shown in Figure 3.23. Since an initial isothermal condition was assumed at the
beginning of the process, the interfacial heat flux is zero at t=0 s. As expected,
the heat flux at the solidfluid interfaces increases with time. A much larger heat
flux is seen at the stagnation region since the cold fluid at the jet strikes that
region and keeps the minimum temperature at that location. This behavior
occurs due to the constant renewal of cold fluid to dissipate the heat. The heat
dissipated is utilized to rise the temperature of the solid as well as the fluid and
reduces thermal storage within the solid due to convective heat transfer. Another
maximum heat flux is encountered around s/d
n
=0.5, and then it decreases
downstream. This is due to the transition of the fluid from the vertical
impingement to a thin film flow along the curved surface where the boundary
layer starts to develop. This peak can be associated with the start of the thermal
boundary layer in the thin film structure. Figure 3.23 also presents q
av,int
r
o
2
/q
w
r
i
2
which is the ratio of the energy transmitted to the fluid to the energy input at the
bottom face of the hemisphere. It may be noticed that a large amount of energy
is absorbed by the solid at the early part of the transient and more and more
53
energy is dissipated to the fluid as the transient progresses. The interfacial heat
flux reaches within 1% of the steadystate equilibrium condition at F
o
= 0.103.
Figure 3.23 Dimensionless Local Heat Flux Variation at the SolidFluid Interface for
Silicon Hemisphere at Different Time Instants (Re=750, b/d
n
=0.5, ß=2.5)
Figure 3.24 illustrates the dimensionless interface temperature for different
time instants. As can be observed at the very beginning of the heat transfer
process, the solidfluid interface maintains a uniform temperature compared to
that when approaching the steadystate condition. This pattern is due to the
thermal storage in the fluid necessary to develop the thermal boundary layer
since an isothermal condition was present at the beginning of the problem. As
time goes on, the thickness of the thermal boundary layer increases and
therefore the temperature rises. Figure 3.24 also illustrates the difference of
0
0.5
1
1.5
2
2.5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
i
a
l
H
e
a
t
F
l
u
x
(
q
i
n
t
/
q
)
Fo=0.00154, =0.199
Fo=0.0054, =0.386
Fo=0.010, =0.546
Fo=0.056, =0.957
Fo=0.103, =0.990
Fo=0.231, =0.995
Fo=0.372, =0.995
q
av,int
r
o
2
/q
w
r
i
2
q
av,int
r
o
2
/q
w
r
i
2
q
av,int
r
o
2
/q
w
r
i
2
q
av,int
r
o
2
/q
w
r
i
2
q
av,int
r
o
2
/q
w
r
i
2
q
av,int
r
o
2
/q
w
r
i
2
q
av,int
r
o
2
/q
w
r
i
2
54
maximumtominimum temperature at the interface being the maximum located
at the outer edge of the hemisphere and the minimum at the stagnation point of
the hemisphere due to the constant renewal of cold fluid from the nozzle. Thus,
such temperature difference increases with time as more heat flows throughout
the hemispherical solid and transmitted to the fluid. The range of temperature
encountered at the solidfluid interface increases with time and reaches a
constant value at the steady state.
Figure 3.24 Dimensionless Interface Temperature Variation for Silicon Hemisphere at
Different Time Instants (Re=750, b/d
n
=0.5, ß=2.5)
Fig. 3.25 shows the variation of local Nusselt number along the solidfluid
interface at different time instants. The local Nusselt number is controlled by
local temperature and heat flux at the solidfluid interface. Both of these
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance (s/d
n
)
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
Fo = 0.0015, ΘmaxΘmin(int) = 6.27E04
Fo = 0.0054, ΘmaxΘmin(int) = 7.44E03
Fo = 0.010, ΘmaxΘmin (int) = 1.70E02
Fo = 0.026, ΘmaxΘmin(int) = 3.99E02
Fo = 0.056, ΘmaxΘmin (int) = 5.17E02
Fo = 0.103, ΘmaxΘmin (int) = 5.46E02
Fo = 0.231, ΘmaxΘmin (int) = 5.50E02
Fo = 0.372, ΘmaxΘmin (int) = 5.51E02
55
quantities increase with time. The local Nusselt number shows a higher value at
early stages of the transient process due to smaller temperature difference
between the jet and the outer surface of the hemisphere. This essentially means
that all heat reaching the solidfluid interface via conduction through the solid is
more efficiently convected out as the local fluid temperature is low everywhere at
the interface. The local Nusselt number decreases with time until it reaches the
steadystate equilibrium distribution. Figure 3.25 also provides the integrated
average Nusselt number for the entire hemispherical surface. As expected, the
average Nusselt number is large at the early part of the transient and
monotonically decreases with time ultimately reaching the value for the steady
state condition.
Figure 3.25 Local Nusselt Number Distribution for Silicon Hemisphere at Different Time
Instants (Re=750, b/d
n
=0.5, ß=2.5)
0
10
20
30
40
50
60
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
Fo = 0.0015, Nu = 36.51
Fo = 0.0055, Nu = 24.72
Fo = 0.010, Nu = 20.02
Fo = 0.056, Nu = 18.26
Fo = 0.103, Nu = 18.20
Fo = 0.231, Nu = 18.19
Fo = 0.372, Nu = 18.09
av
av
av
av
av
av
av
56
Figure 3.26 presents results for dimensionless maximum temperature at
the interface, maximum temperature in the solid, and maximumtominimum
difference temperature at the interface for two different Reynolds numbers. The
maximum temperature within the solid was encountered at the outlet plane next
to the heated surface (z=r
o
, r=r
i
). As noticed, the temperature begins to rise with
time as the hemispherical solid begins to store heat starting after the specified
initial condition (T
J
=310 K) showing a rapid response at the earlier part of the
heating process until its thermal storage capacity reduces up to its limit (steady
state). It is important to mention that the time necessary to reach steadystate
depends strongly on the Reynolds number.
Figure 3.26 Distribution of Dimensionless Maximum Temperature at the SolidFluid
Interface, within the Solid, and MaximumtoMinimum Temperature
Difference with Time for Two Reynolds Numbers (Silicon Hemisphere,
b/d
n
=0.5, ß=2.5)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0 0.05 0.1 0.15 0.2 0.25 0.3
Fourier Number, Fo
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m

t
o

M
i
n
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
D
i
f
f
e
r
e
n
c
e
a
t
t
h
e
I
n
t
e
r
f
a
c
e
,
Θ Θ Θ Θ
m
a
x

Θ Θ Θ Θ
m
i
n
Θmax(interface), Re=500
Θmax(interface), Re=1000
Θmax(solid), Re=500
Θmax(solid), Re=1000
ΘmaxΘmin(interface), Re=500
ΘmaxΘmin(interface), Re=1000
57
Figure 3.26 also confirms that a higher Reynolds number increases
convective heat transfer and therefore lowers the hemisphere’s temperature.
The control of maximum temperature is important in many critical thermal
management applications including electronic packaging.
Figure 3.27 plots the average Nusselt number variation along the solid
fluid interface for two different Reynolds numbers over the entire transient start
up of the heat transfer process. As expected, the average Nusselt number
becomes larger as the Reynolds number increases due to higher velocity of the
fluid particles moving along the hemisphere’s outer surface, hence, increasing
the rate of heat transfer.
Figure 3.27 Variation of Average Nusselt Number with Time For Silicon Hemisphere at
Two Different Reynolds Numbers (b/d
n
=0.5, ß=2.5)
0
20
40
60
80
100
120
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Fourier Number, F
o
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
Nu , Re=500
Nu , Re=1000
av
av
58
The time required to reach steadystate for different Reynolds numbers is
presented in Figure 3.28. Fo
ss
was defined as the Fourier number at which the
solidfluid interface temperature everywhere on the hemispherical plate reached
within 0.001% of the steady equilibrium distribution. The time to reach thermal
equilibrium condition decreases as the Reynolds number increases in value.
This is due to more fluid flow rate available to carry away the heat and faster
development of thermal boundary layer that is smaller in thickness.
Figure 3.28 Time Required to Reach SteadyState for Silicon Hemisphere at Different
Reynolds Numbers (b/d
n
=0.5, ß=2.5)
The effects of varying the thickness in the hemispherical plate on
maximum temperature at the interface, maximum temperature within the solid,
and maximumtominimum temperature difference at the interface can be seen
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
400 700 1000 1300 1600
Reynolds Number, Re
F
o
u
r
i
e
r
N
u
m
b
e
r
,
F
o
S
S
59
on Figure 3.29. The plate thickness significantly affects the temperature
distribution. It may be note that as the thickness of the hemispherical plate
increases, the time needed to achieve steadystate conditions increases. This is
due to more storage capacity of heat within the solid. Also, the temperature at
the solidfluid interface remains lower due to higher thermal resistance of the
solid to the path of heat flow.
Figure 3.29 Distribution of Dimensionless Maximum Temperature at the SolidFluid
Interface, within the Solid, and MaximumtoMinimum Temperature
Difference with Time for Different Plate Thicknesses (Silicon Hemisphere
Re=750, ß=2.5)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 0.1 0.2 0.3 0.4 0.5 0.6
Fourier Number, Fo
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
m
a
x
0
0.01
0.02
0.03
0.04
0.05
0.06
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m

t
o

M
i
n
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
D
i
f
f
e
r
e
n
c
e
a
t
t
h
e
I
n
t
e
r
f
a
c
e
,
Θ Θ Θ Θ
m
a
x

Θ Θ Θ Θ
m
i
n
Θmax(interface), b/dn=0.5
Θmax(interface), b/dn=1.0
Θmax(solid), b/dn=0.5
Θmax(solid), b/dn=1.0
ΘmaxΘmin(interface), b/dn=0.5
ΘmaxΘmin(interface), b/dn=1.0
60
Figure 3.30 shows the average Nusselt number variation as a function of
time for three distinct plate thicknesses using Constantan as the solid material.
The average Nusselt number is higher for higher plate thickness.
Figure 3.30 Variation of Average Nusselt Number with Time for Constantan Hemisphere
at Three Different Thicknesses (Re=750, ß=2.5)
The maximum temperature at the solidfluid interface, maximum
temperature within the solid, and maximum to minimum temperature difference
(as a measure for temperature nonuniformity) for different solid materials is
presented in Figure 3.31. As expected, the temperature changes occur faster at
the beginning of the heating process for all materials and the slope gradually
decays when the steadystate approaches. It can be observed that materials
0
25
50
75
100
125
150
175
200
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Fourier Number, F
o
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
b/d =0.083
b/d =0.5
b/d =1.0
n
n
n
61
with a very low thermal conductivity such as Constantan maintains a higher
temperature at the hemisphere’s outer surface and within the solid as the thermal
conductivity controls how effective the heat flows and distributes throughout the
material. The thermal diffusivity of the material also contributes to the transient
behavior of the solid. As noticed, Silicon and Copper reach the steadystate
faster than Constantan due to their higher thermal diffusivity, which controls the
rate of heat being transferred through the solid material.
Figure 3.31 Distribution of Dimensionless Maximum Temperature at the SolidFluid
interface, within the Solid, and MaximumtoMinimum Temperature
Difference with Time for Different Materials (Re=750, ß=2.5)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 0.1 0.2 0.3 0.4 0.5
Fourier Number, Fo
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
m
a
x
0
0.02
0.04
0.06
0.08
0.1
0.12
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m

t
o

M
i
n
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
D
i
f
f
e
r
e
n
c
e
a
t
t
h
e
I
n
t
e
r
f
a
c
e
Θ Θ Θ Θ
m
a
x

Θ Θ Θ Θ
m
i
n
Θmax(interface), Constantan
Θmax(interface), Silicon
Θmax(interface), Copper
Θmax(solid), Constantan
Θmax(solid), Silicon
Θmax(solid), Copper
ΘmaxΘmin(interface), Constantan
ΘmaxΘmin(interface), Silicon
ΘmaxΘmin(interface), Copper
62
Figure 3.32 shows the distribution of average Nusselt Number with time
for the three materials used in this study. Constantan shows a higher average
heat transfer coefficient compared to Silicon or Copper over the entire transient
process. A significant difference is seen at the earlier part of the transient and
the curves come close together as the steadystate approaches.
Figure 3.32 Variation of Average Nusselt Number with Time for Different Materials
(Re=750, b/d
n
,=0.5, ß=2.5)
It will be also important to know how the materials responded in reaching
thermal equilibrium based on their thickness. Figure 3.33 presents Fo
ss
for these
materials for different plate thicknesses. As the thickness increases in value, the
time to reach steadystate also increases. Constantan takes longer in reaching
steadystate due to its lower thermal diffusivity compared to Copper and Silicon.
0
10
20
30
40
50
60
70
80
90
0 0.01 0.02 0.03 0.04 0.05
Fourier Number, F
o
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
Constantan
Copper
Silicon
63
Also, the property of the solid plays more significant role in determining the
duration of the transient heat transfer process when the thickness is increased.
Figure 3.33 Time Needed to Reach SteadyState for Different Hemispherical Plate
Thicknesses and for Different Materials (Re=750, ß=2.5)
The development of isothermal lines within the solid can be observed in
Figure 3.34 at different time instants for b/d
n
=0.5. It is important to notice that at
early stages of the transient heat transfer process, the isothermal lines grow
parallel to the inner (bottom) heated surface of the hemispherical plate. As time
goes on, the isothermal lines start moving upward toward lower temperature
regions until they reach the solidfluid interface. After that, they start to form
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Dimensionless Plate Thickness, b/d
n
F
o
u
r
i
e
r
N
u
m
b
e
r
,
F
o
Constantan
Copper
Silicon
64
concentric lines near the stagnation point and expand further down into the solid
until a steadystate condition is achieved.
Figure 3.34 Isotherms Countour Lines at Varius Time Instants for Silicon Hemisphere
Plate of b/d
n
=0.5 (Re=750, ß=2.5, Top Left Corner (z=0, r=0))
T
max
=310.24 K
T
min
=310.15 K
t=0.00028 s
T
max
=310.93 K
T
min
=310.48 K
t=0.002 s
T
max
=313.45 K
T
min
=312.14 K
t=0.01 s
T
max
=319.39 K
T
min
=318.14 K
t=0.05 s
T
max
=326.05 K
T
min
=324.99 K
t=0.1 s
T
max
=343.60 K
T
min
=340.57 K
t=0.5 s
T
max
=344.91 K
T
min
=342.65 K
t=1.2 s
T
max
=345.20 K
T
min
=343.15 K
t=2.2 s
65
Figure 3.35 shows the same phenomenon for b/d
n
=1.5. The temperatures inside
the solid are much lower as compared to a thinner plate and the time to reach
steadystate equilibrium thermal condition increases.
Figure 3.35 Isotherms Countour Lines at Varius Time Instants for Silicon Hemisphere
Plate of b/d
n
=1.5 (Re=750, ß=2.5, Top Left Corner (z=0, r=0)
T
max
=310.43 K
T
min
=310.15 K
t=0.0002 s t=0.002 s
T
max
=310.94 K
T
min
=310.19 K
T
max
=311.68 K
T
min
=310.58 K
t=0.01 s
T
max
=313.72 K
T
min
=312.30 K
t=0.05 s
T
max
=316.13 K
T
min
=314.97 K
t=0.1 s
T
max
=329.57 K
T
min
=327.05 K
t=1.03 s
T
max
=330.75 K
T
min
=329.64 K
t=2.06 s
T
max
=330.98 K
T
min
=329.87 K
t=2.80 s
66
Chapter 4 Cylindrical Plate Model Results
4.1 Steady State Heating
A velocity vector distribution is shown in Figure 4.1. It can be observed
that the velocity remains almost uniform at the potential core region of the jet.
However, the velocity decreases and the fluid jet diameter increases as the fluid
gets closer to the surface during the impingement process. The direction of
motion of the fluid particles shifts along the angle of curvature, afterwards, the
fluid accelerates creating a region of minimum sheet thickness. That is the
beginning of the boundary layer zone. It can be noticed that as the fluid moves
downstream along the convex surface, the boundary layer thickness increases
and the frictional resistance from the wall is eventually transmitted to the entire
film thickness where the fully viscous zone develops.
67
Figure 4.1 Velocity Vector Distribution for a Silicon Curved Plate with Water as the
Cooling Fluid (Re = 750, b= 0.60 mm, Q = 5.665x10
7
m
3
/s, β = 2.5, H
n
= 0.3
cm, q = 250 kW/m
2
)
Fig. 4.2 shows the free surface height distribution for different Reynolds
numbers when the jet strikes the curved plate’s surface. It can be seen that the
fluid spreads out moving along the outer surface of the curved plate. As the
Reynolds number increases, the free surface increases in thickness due to
higher flow rate. In addition, a higher Reynolds number causes a higher
rebounding velocity of the fluid particles at the impact that also contributes to a
higher film thickness, particularly near the impingement region. For the conditions
considered in the present investigation, the flow was supercritical and no
hydraulic jump was present.
68
.
Figure 4.2 Free Surface Height Distribution for Different Reynolds Numbers and Water
as the Cooling Fluid (b=0.60 mm, β=2.0, H
n
=0.30 cm, q=250 kW/m
2
)
Figures 4.3 and 4.4 show dimensionless interface temperature and local
Nusselt number distributions as a function of dimensionless distance (s/d
n
) along
the solidfluid interface at different Reynolds numbers. The curves in Figure 4.3
reveal that the dimensionless interface temperature decreases with jet velocity or
Reynolds number. The dimensionless interface temperature has the lowest
value near the stagnation point (underneath the jet opening) where the fluid
particles start to move along the disk surface (after the impingement process)
and increases along the arc length reaching the highest value at the end of it.
This also shows that an increasing Reynolds number contributes to a more
effective cooling by the enhancement of convective heat transfer coefficient.
0.42
0.46
0.50
0.54
0.58
0.62
0.66
0.70
0.74
0.78
0.82
0.86
0.00 0.06 0.12 0.18 0.24 0.30 0.36 0.42 0.48
Radial Distance (cm)
A
x
i
a
l
D
i
s
t
a
n
c
e
(
c
m
)
Re=500
Re=750
Re=1000
Re=1200
Re=1500
Re=1800
69
Figure 4.3 Dimensionless Interface Temperature Distribution for a Silicon Curved Plate at
Different Reynolds Numbers and Water as the Cooling Fluid (β=2.5, H
n
= 0.30
cm, b=0.60 mm, q=250 kW/m
2
)
The local Nusselt number shown in Figure 4.4 increases rapidly over a
small distance measured from the stagnation point (core region), reaches a
maximum around s/d
n
= 0.23, and then decreases along s as the boundary layer
develops further downstream. The location of the maximum Nusselt number can
be associated with the transition of the flow from the vertical impingement to
displacement along the disk surface where the boundary layer starts to develop.
Figure 4.4 also confirms that an increasing Reynolds number contributes to a
more effective cooling by the enhancement of convective heat transfer
coefficient. The local Nusselt number increased very significantly all along the
plate surface with the increase of Reynolds number.
0
0.1
0.2
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
Re=500
Re=750
Re=1000
Re=1200
Re=1500
Re=1800
70
Figure 4.4 Local Nusselt Number Distribution for a Silicon Curved Plate at Different
Reynolds Numbers and Water as the Cooling Fluid (β =2.5, H
n
=0.30 cm,
b=0.60 mm, q=250 kW/m
2
)
Figure 4.5 plots the average Nusselt number (Nu
av
) and maximum
temperature in the solid (Θ
max
) as a function of Reynolds number. It may be
noted that average Nusselt number increases with Reynolds number and
maximum temperature within the solid decreases with increasing Reynolds
number. The maximum temperature happens at the outer edge of the plate
adjacent to the inner surface which is heated. As the flow rate (or Reynolds
number) increases, the magnitude of fluid velocity near the solidfluid interface
that controls the convective heat transfer rate increases. That results in lowering
of the maximum temperature. The control of maximum temperature may be
crucial in the thermal management of electronic equipment.
0
10
20
30
40
50
60
0 0.5 1 1.5 2 2.5 3 3.5 4
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
Re=500
Re=750
Re=1000
Re=1200
Re=1500
Re=1800
71
Figure 4.5 Average Nusselt Number and Maximum Temperature Within the Solid for
Different Reynolds Numbers for Constant Heat Flux Conditions (q
= 250
KW/m
2
, β=2.5, H
n
= 0.30cm)
The solidfluid dimensionless interface temperature and local Nusselt
number distributions for six different nozzletotarget spacing for water as the
cooling fluid and Reynolds number of 750 are shown in Figures 4.6 and 4.7
respectively. As the curved plate surface gets closer to the nozzle (smaller
nozzletotarget spacing ratio), the temperature at the solid fluid interface
decreases due to higher jet momentum at impingement that causes higher
velocity of fluid particles adjacent to the plate enhancing the heat transfer.
0
5
10
15
20
25
30
35
40
400 600 800 1000 1200 1400 1600 1800 2000
Reynolds Number, Re
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
i
n
S
o
l
i
d
,
Θ Θ Θ Θ
m
a
x
Nu
Θ
av
max
72
Figure 4.6 Dimensionless Interface Temperature for Silicon Curved Plate for Different
Nozzle to Target Spacing Ratio (β) and Water as the Cooling Fluid (Re=750,
d
n
=1.2mm, b=0.60 mm, q=250 kW/m
2
)
Therefore, in Figure 4.7, a higher Nusselt number is seen all along the arc
length. It can be noticed that the impingement height affects the Nusselt number
more at the stagnation region and the early part of the boundary layer region. At
larger arc length the values are in a small range for all impingement heights. This
is quite expected since the jet momentum more strongly affects the areas
subjected to direct impingement. It can be also noticed that no significant
change is seen at heights higher than β=2, where impingement height does not
play a strong role in determining the convective heat transfer process.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
β=0.75
β=1.0
β=1.5
β=2.0
β=2.5
β=3.0
73
Figure 4.7 Local Nusselt Number Distribution for Silicon Curved Plate for Different
Nozzle to Target Spacing Ratio (β) and Water as the Cooling Fluid (Re=750,
d
n
=1.2mm, b=0.60mm, q=250kW/m
2
)
Therefore, in Figure 4.7, a higher Nusselt number is seen all along the arc
length. It can be noticed that the impingement height affects the Nusselt number
more at the stagnation region and the early part of the boundary layer region. At
larger arc length the values are in a small range for all impingement heights. This
is quite expected since the jet momentum more strongly affects the areas
subjected to direct impingement. It can be also noticed that no significant
change is seen at heights higher than β=2, where impingement height does not
play a strong role in determining the convective heat transfer process.
The effects of varying the inner radius of curvature of the impingement
plate are demonstrated in Figures 4.84.10. Figure 4.8 shows the dimensionless
0
10
20
30
40
50
60
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
β=0.75
β=1.0
β=1.5
β=2.0
β=2.5
β=3.0
74
solidfluid interface temperature of water flowing on the curved plate with different
R
i
/d
n
as a function of dimensionless distance (s/d
n
). The corresponding
distribution for a flat plate (infinite radius of curvature) is also shown for
comparison. In order to keep the same energy input to the plate in these runs,
the value of arc length s was kept constant. Therefore, half cylindrical plate
(shown in Fig. 1) was possible only for the smaller radius of curvature
(R
i
/d
n
=4.16). For other plates, the arc extended over an angle of φ
max
that is less
than π/2. For flat disk, s was same as its radius. It can be noticed that decreasing
the inner plate radius of curvature to nozzle diameter ratio (R
i
/d
n
) decreases the
dimensionless interface temperature.
Figure 4.8 Dimensionless Interface Temperature for Silicon Curved Plate for Different
Inner Plate Radius of Curvature to Nozzle Diameter Ratio (R
i
/d
n
) for Water as
the Cooling Fluid (β =2.5 , H
n
= 0.30 cm Re=750, d
n
=1.2mm, b=0.60 mm,
q=250 kW/m
2
)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
R /d = 4.16
R /d =6.66
R /d =10
R /d =16.66
Flat
i n
i
i n
i n
n
75
This is because the fluid particles are subjected to a larger gravitational
force that increases the local fluid velocity near the solidfluid interface and
results in larger rate of convective heat transfer. Figure 4.9 plots the local Nusselt
number distribution for different inner plate radius of curvature to nozzle diameter
ratio (R
i
/d
n
) including a flat plate as a function of dimensionless distance (s/d
n
). A
higher Nusselt number is encountered when R
i
/d
n
diminishes increasing the
cooling capacity of the fluid. For all cases, the maximum local Nusselt number is
encountered at s/d
n
≈0.23 and after that it decays monotonically. The trend in
Figure 4.9 agrees with the results reported by Martin [39].
Figure 4.9 Local Nusselt Number Distribution for Silicon Curved Plate for Different Inner
Plate Radius of Curvature to Nozzle Diameter Ratio (R
i
/d
n
) for Water as the
Cooling Fluid (β =2.5 , H
n
= 0.30 cm Re=750, d
n
=1.2mm, b=0.60 mm, q=250
kW/m
2
)
0
5
10
15
20
25
30
35
40
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
R /d = 4.16
R /d =6.66
R /d =10
R /d =16.66
Flat
i n
i
i n
i n
n
76
Figure 4.10 plots the average Nusselt number (Nu
av
) and the maximum
temperature within the solid (θ
max
) as a function of the inner plate radius of
curvature to nozzle diameter ratio (R
i
/d
n
). As expected, the average Nusselt
number decreases with increasing inner radius of curvature and maximum
temperature increases.
Figure 4.10 Average Nusselt Number and Maximum Temperature in the Solid
Variations (Θ
max
) for Silicon Curved plate for Different Inner Plate Radius of
Curvature to Nozzle Diameter Ratio (R
i
/d
n
) for Water as the Cooling Fluid (β
=2.5 , H
n
= 0.30 cm Re=750, d
n
=1.2mm, b=0.60 mm, q=250 kW/m
2
)
Figure 4.11 presents the local heat transfer coefficient and Nusselt
number distributions for different nozzle diameters (d
n
) for a constant flow rate
using water as the working fluid. When flow rate (or Reynolds number) is kept
constant, a reduction in nozzle slot opening requires a proportional increase of jet
20
20.5
21
21.5
22
22.5
23
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0
R
i
/d
n
N
u
a
v
0.145
0.15
0.155
0.16
0.165
0.17
0.175
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
i
n
S
o
l
i
d
,
Θ Θ Θ Θ
m
a
x
Nu
Θ
av
max
77
inlet velocity. A higher jet velocity essentially contributes to a higher fluid velocity
near the solidfluid interface and higher value of convective heat transfer
coefficient. However, since the Nusselt number used the slot width as the length
scale, increasing (d
n
) results in a higher local Nusselt number along the solidfluid
interface.
Figure 4.11 Heat Transfer Coefficient and Nusselt Number Distributions for Different
Nozzle Slot Widths (Q=0.0006 m
3
/s, β =2.5 , H
n
= 0.30 cm β=2.5, Re=750,
q=250 kW/m
2
)
Figures 4.12 and 4.13 display the effects of varying plate thickness to
nozzle diameter ratio (b/d
n
) for a silicon plate using water as the cooling fluid. As
the thickness increases, the temperature at the solidfluid interface decreases
and becomes more uniform due to a better distribution of heat within the solid by
0
5
10
15
20
25
30
35
40
45
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
2000
9000
16000
23000
30000
37000
44000
51000
58000
L
o
c
a
l
H
e
a
t
T
r
a
n
s
f
e
r
C
o
e
f
f
i
c
i
e
n
t
,
h
N , d =0.3mm
N , d =1.2mm
N , d =2.4mm
h, d =0.3mm
h, d =1.2mm
h, d =2.4mm
Nu, d
n
=0.3 mm
Nu, d
n
=1.2 mm
Nu, d
n
=2.4 mm
h, d
n
=0.3 mm
h, d
n
=1.2 mm
h, d
n
=2.4 mm
78
conduction. It may be also noticed that the temperature is lower at the
stagnation region and it increases as s/d
n
increases all the way to the end of the
plate. It can be also noted that a thinner plate provides a lower Nusselt number
at the stagnation region but increases to provide higher maximum value and
maintains a higher value as the fluid moves downstream along the curved plate.
Thus, a thinner plate contributes to a more effective cooling by convection as
opposed to a thicker plate.
Figure 4.12 Dimensionless Interface Temperature for Different Plate Thicknesses (β
=2.5, Re=750, d
n
=1.2mm, q=250 kW/m
2
)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
b/d =0.08
b/d =0.25
b/d =1.0
b/d =1.5
n
n
n
n
79
Figure 4.13 Local Nusselt Number Distribution for Different Plate Thicknesses (β =2.5,
Re=750, H
n
= 0.30 cm, q=250 kW/m
2
)
Figure 4.14 presents the average Nusselt number (Nu
av)
and maximum
temperature (Θ
max
) encountered within the solid as a function of (b/d
n
). The
maximum temperature within the solid occurs at the inner surface of the curved
plate at its outlet end (maximum s or φ). It can be noticed that both average
Nusselt number and maximum temperature within the solid diminish with
increase of plate thickness. A lower average Nusselt number is a result of
integrated local values seen in Figure 4.14, where a thicker plate provided a
lower Nusselt number over a major portion of s/d
n
. A lower maximum
temperature for a thicker plate is a result of lower total heat input when plate
0
5
10
15
20
25
30
35
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
b/d =0.08
b/d =0.25
b/d =1.0
b/d =1.5
n
n
n
n
80
thickness was higher, as the heat flux at the inner surface of the plate and outer
radius of the plate were kept constants in these simulations.
Figure 4.14 Average Nusselt Number and Maximum Temperature in the Solid
Variations with Plate Thickness (β =2.5, Re=750)
Figure 4.15 compares the dimensionless solidfluid interface temperature
for the present working fluid (water) with two other cooling fluids flouroinert (FC
77) and oil (MIL7808). Water (T
J
=310.15 K) presents the highest dimensionless
interface temperature when compared to FC77 (T
J
=310.15 K) and MIL7808
(T
J
=375 K). This trend is well correlated with Prandtl number. A higher Prandtl
number results in lower dimensionless temperature at the solidfluid interface. In
addition, due to lower thickness or smaller growth rate of thermal boundary layer
at higher Prandtl number, a low temperature is maintained over a larger portion
20.6
20.8
21
21.2
21.4
21.6
21.8
22
22.2
22.4
0 0.2 0.4 0.6 0.8 1 1.2
b/d
n
N
u
a
v
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
i
n
S
o
l
i
d
,
Θ Θ Θ Θ
m
a
x
Nu
Θ
av
max
81
of the plate resulting in better cooling performance. Figure 4.16 displays the
local heat transfer coefficient and local Nusselt number for these three fluids. A
higher Nusselt number distribution is encountered when MIL7808 is used as the
working fluid in comparison with FC77 (q=50 kW/m
2
) and water.
Figure 4.15 Dimensionless Interface Temperature for Different Cooling Fluids
(β=2.5, Re=750)
As expected, a higher Nusselt number is obtained for a higher Prandtl
number fluid. However, when dimensional heat transfer coefficient is considered,
the Prandtl number alone cannot correlate the trend. The highest heat transfer
coefficient is still provided by MIL7808 at the impingement region. However, the
curves for MIL7808 and water cross each other and at large values of s, the
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 0.5 1 1.5 2 2.5 3 3.5 4
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
H O(Pr=4.45)
FC77(Pr=16.58)
MIL7808(Pr=159.32)
2
82
highest heat transfer coefficient is obtained for water. FC77 (k
f
=0.06299 W/m
K) gives much lower heat transfer coefficient compared to water or MIL7808
because of its lowest thermal conductivity. These results were obtained for a
constant Reynolds number of 750.
Figure 4.16 Heat Transfer Coefficient and Nusselt Number Distributions for Different
Cooling Fluids (Re=750, β=2.5)
One of the papers used for the validation of this numerical study was the
experimental work carried out by Bartoli and Faggiani [21] where water was used
for the cooling of a stainless steel hollow cylinder by jet impingement. The
numerical simulation attempted to duplicate the exact conditions of this
experiment. A steady state mixing length turbulence model was employed for
0
20
40
60
80
100
120
140
160
180
200
0 0.5 1 1.5 2 2.5 3 3.5 4
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
0
5000
10000
15000
20000
25000
L
o
c
a
l
H
e
a
t
T
r
a
n
s
f
e
r
C
o
e
f
f
i
c
i
e
n
t
,
h
Nu, H 0(Pr=4.45)
Nu, FC77(Pr=16.58)
Nu, MIL7808(Pr=159.32)
h, H O(Pr=4.45)
h, FC77(Pr=16.58)
h, MIL7808(Pr=159.32)
2
2
83
this simulation. The parameters used and their corresponding values according
to Bartoli and Faggiani [21] were the following: Re=3500–20,000, T
J
=293.15–
323.15 K, Pr=3.57–6.99, q=50 kW/m
2
, φ=0°90°, β=5.5, and b=0.5 mm. Figure
4.17 compared the results obtained in the numerical simulation with the
correlation proposed by Bartoli and Faggiani [21] for Nu/Pr
0.4
. The difference is
3.60% at Re= 3,500 and 1.88% at Re= 15,000.
Figure 4.17 Local Nusselt Number Compared with Bartoli and Faggiani [21] at Different
Reynolds Numbers (φ=90°, d
n
=2.0 mm, b=0.50 mm, q=50 kW/m
2
)
0
20
40
60
80
100
120
3000 8000 13000 18000
Reynolds Number, Re
N
u
/
P
r
0
.
4
Present Values
Bartoli and Faggiani
84
Numerical simulations were also carried out to compare with the results
obtained by Gori and Bossi [29] on the cooling of a hollow stainless steel circular
cylinder by a turbulent flow of air from a slot nozzle. The numerical model tried to
mimic the experimental set up that the authors used to obtain the desired
outcome for a constant heat flux condition. In addition, a correlation suggested
by Whitaker [40] for uniform fluid flow was also used for comparison.
Figure 4.18 Average Nusselt Number Compared with Gori and Bossi [29] and Whitaker
[40] for Different Reynolds Numbers (d
n
=2.5 mm, b = 0.2 mm, q=2.35
kW/m
2
)
Figure 4.18 illustrates the numerical results obtained in the simulation
compared to the experimental data obtained by Gori and Bossi [29] along with
0
20
40
60
80
100
120
140
0 5000 10000 15000 20000 25000
Reynolds Number, Re
N
u
a
v
Numerical Values
Gori and Bossi
Whitaker
85
the correlation proposed by Whitaker [40]. The values used during the simulation
included a Reynolds numbers range of 400020,000, a fluid inlet (jet)
temperature (T
J
) of 293.15–296.15 K, the fluid Prandtl number (Pr) of 0.711–
0.712, a constant heat flux (q) of 2.35 kW/m
2
, an azimuthal angle (φ) οf 0°90°, a
fixed dimensionless nozzle to target spacing ratio (β) of 8.0, and a fixed thickness
(b) of 0.2 mm as specified by the authors.
The difference of the numerical simulation and the experimental results
obtained by Gori and Bossi [29] was between 6.80%23.77% with an average
difference of 15.92%. The correlation proposed by Whitaker compared within a
range of 1.76%15.82% with an average difference of 9.31%. Considering the
uncertainly of experimental data and the discretization and roundoff errors
inherent in numerical simulation, both of these comparisons may be considered
to be quite satisfactory.
4.2 Transient Heating
The local heat flux variation along the solidfluid interface for different time
intervals is shown in Figure 4.19. Since an initial isothermal condition was
assumed at the beginning of the process, the interfacial heat flux is zero at t=0 s.
As expected, the heat flux at the solidfluid interfaces increases with time. A
much larger heat flux is seen at the stagnation region since the cold fluid at the
jet strikes that region and keeps the minimum temperature at that location. This
behavior occurs due to the constant renewal of cold fluid to dissipate the heat.
The heat dissipated is utilized to rise the temperature of the solid as well as the
86
fluid and reduces thermal storage within the solid due to convective heat transfer.
Another maximum heat flux is encountered around s/d
n
=0.5, and then it
decreases downstream. This can be explained as the transition of the fluid from
the vertical impingement to a thin film flow along the curved surface where the
boundary layer starts to develop. This crest can be associated with the start of
the thermal boundary layer in the thin film structure. Figure 4.19 also presents
q
av,int
r
o
2
/q
w
r
i
2
which is the ratio of the energy transmitted to the fluid to the energy
input at the bottom face of the plate. The interfacial heat flux reaches within 1%
of the steadystate equilibrium condition at F
o
= 0.103.
Figure 4.19 Dimensionless Local Heat flux Variation at the SolidFluid Interface for
Silicon Plate at Different Time Instants (Re=750, b/d
n
=0.5, ß=2.5)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
f
a
c
i
a
l
H
e
a
t
F
l
u
x
,
(
q
i
n
t
/
q
)
Fo=0.0013,=0.177
Fo=0.0051, =0.405
Fo=0.0107,=0.619
Fo=0.0211, =0.837
Fo=0.051, =1.023
Fo=0.103, =1.057
Fo=0.231, =1.060
F
o
=0.0013, q
av,int
r
o
2
/q
w
r
i
2
=0.177
F
o
=0.0051, q
av,int
r
o
2
/q
w
r
i
2
=0.405
F
o
=0.0107, q
av,int
r
o
2
/q
w
r
i
2
=0.619
F
o
=0.0211, q
av,int
r
o
2
/q
w
r
i
2
=0.837
F
o
=0.051, q
av,int
r
o
2
/q
w
r
i
2
=0.987
F
o
=0.103, q
av,int
r
o
2
/q
w
r
i
2
=0.999
F
o
=0.231, q
av,int
r
o
2
/q
w
r
i
2
=0.999
87
Figure 4.20 shows the dimensionless interface temperature for different
time instants. As can be observed at the very beginning of the heat transfer
process, the solidfluid interface keeps a uniform temperature compared to that
when approaching steadystate equilibrium condition. This pattern is due to the
thermal storage in the fluid necessary to develop the thermal boundary layer
since an isothermal condition was present at the beginning of the problem. As
time goes on, the thickness of the thermal boundary layer increases and
therefore the temperature rises. Figure 4.20 also illustrates the difference of
maximumtominimum temperature at the interface being the maximum situated
at the outer edge of the plate and the minimum at the stagnation point of the
plate due to the constant renewal of cold fluid from the nozzle. Thus, such
temperature difference increases with time as more heat flows throughout the
solid and transmitted to the fluid. The range of temperature encountered at the
solidfluid interface increases with time and reaches a constant value at the
steady state. Figure 4.21 plots the local Nusselt number variation along the
solidfluid interface at different time instants. The local Nusselt number is
controlled by local temperature and heat flux at the solidfluid interface. This
means that all heat reaching the solidfluid interface via conduction through the
solid is more efficiently convected out as the local fluid temperature is low
everywhere at the interface.
88
Figure 4.20 Dimensionless Interface Temperature Variation for Silicon Plate at Different
Time Instants (Re=750, b/d
n
=0.5, ß=2.5)
The local Nusselt number decreases with time until it reaches the steady
state equilibrium distribution. Figure 4.22 also provides the integrated average
Nusselt number for the entire cylindrical shaped plate surface. As expected, the
average Nusselt number is large at the early part of the transient and
monotonically decreases with time all the way to the steady state condition. At
this point, equilibrium conditions with the surroondings have been achieved
between the plate and the fluid and the temperature of the plate along with the
fluid becomes uniform. That is, T
f
=T
s.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
D
i
m
e
n
s
i
o
n
l
e
s
s
I
n
t
e
r
f
a
c
e
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
Fo=0.0013, ΘmaxΘmin (int)=5.36E03
Fo=0.0051, ΘmaxΘmin(int)=5.49E02
Fo=0.0107, ΘmaxΘmin(int)=7.58E02
Fo=0.021, ΘmaxΘmin(int)=8.45E02
Fo=0.051, ΘmaxΘmin(int)=8.78E02
Fo=0.103, ΘmaxΘmin(int)=8.62E02
Fo=0.231, ΘmaxΘmin(int)=8.76E02
89
Figure 4.21 Local Nusselt Number Distribution for Silicon Plate at Different Time
Instants (Re=750, b/d
n
=0.5, ß=2.5)
Figure 4.22 shows results for dimensionless maximum temperature at the
interface, maximum temperature in the solid, and maximumtominimum
difference temperature at the interface for two different Reynolds numbers. The
maximum temperature within the solid was found at the outlet plane next to the
heated surface (z=r
o
, r=r
i
). As expected, the temperature begins to rise with
time as the solid plate begins to store heat starting after the specified initial
condition (T
J
=310 K) showing a rapid response at the earlier part of the heating
process until its thermal storage capacity reduces up to its limit (steadystate).
0
10
20
30
40
50
60
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Dimensionless Distance, s/d
n
L
o
c
a
l
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
Fo=0.0013, Nuav=43.94
Fo=0.0051, Nuav=24.41
Fo=0.0107, Nuav=22.21
Fo=0.021, Nuav=21.45
Fo=0.051, Nuav=21.13
Fo=0.103, Nuav=21.11
Fo=0.231, Nuav=21.09
F
o
=0.0013, Nu
av
=43.94
F
o
=0.0051, Nu
av
=24.41
F
o
=0.0107, Nu
av
=22.21
F
o
=0.0211, Nu
av
=21.45
F
o
=0.051, Nu
av
=21.13
F
o
=0.103, Nu
av
=21.11
F
o
=0.231, Nu
av
=21.09
90
Figure 4.22 Distribution of Dimensionless Maximum Temperature at the SolidFluid
Interface, within the Solid, and MaximumtoMinimum Temperature
Difference with Time for Two Reynolds Numbers (Silicon Plate, b/d
n
=0.5,
ß=2.5)
It is significant to point out that the time necessary to reach steadystate
depends on the Reynolds number. Figure 4.22 also proves that a higher
Reynolds number increases convective heat transfer and therefore lowers the
plate’s temperature. Figure 4.23 plots the average Nusselt number variation
along the solidfluid interface for two different Reynolds numbers over the entire
transient startup of the heat transfer process. As expected, the average Nusselt
number becomes larger as the Reynolds number increases due to higher velocity
of the fluid particles moving along the plate’s outer surface, hence, increasing the
rate of heat transfer.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0 0.05 0.1 0.15 0.2
Fourier Number, F
o
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m

t
o

M
i
n
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
D
i
f
f
e
r
e
n
c
e
a
t
t
h
e
I
n
t
e
r
f
a
c
e
,
Θ Θ Θ Θ
m
a
x

Θ Θ Θ Θ
m
i
n
Θmax(interface), Re=500
Θmax(interface), Re=1000
Θmax(solid), Re=500
Θmax(solid), Re=1000
ΘmaxΘmin(interface), Re=500
ΘmaxΘmin(interface), Re=1000
91
Figure 4.23 Variation of Average Nusselt Number with Time For Silicon Plate at Two
Different Reynolds Numbers (b/d
n
=0.5, ß=2.5)
The time required to reach steadystate for different Reynolds numbers is
presented in Figure 4.24. Fo
ss
was defined as the Fourier number at which the
solidfluid interface temperature everywhere on the solid plate reached within
0.001% of the steady equilibrium distribution. The time to reach thermal
equilibrium condition decreases as the Reynolds number increases in value.
This is due to more fluid flow rate available to carry away the heat and faster
development of thermal boundary layer that is smaller in thickness.
0
30
60
90
120
150
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Fourier Number, F
o
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
Series1
Series2
Nu
av
, Re=500
Nu
av
, Re=1000
92
Figure 4.24 Time Required to Reach SteadyState for Silicon Plate at Different
Reynolds Numbers (b/d
n
=0.5, ß=2.5)
The effects of varying the thickness in the cylindrical plate on maximum
temperature at the interface, maximum temperature within the solid, and
maximumtominimum temperature difference at the interface can be seen on
Figure 4.25. The plate thickness significantly affects the temperature distribution.
It may be note that as the thickness of the solid plate increases, the time needed
to achieve steadystate conditions increases. This is due to more storage
capacity of heat within the solid. Similarly, the temperature at the solidfluid
interface remains lower due to higher thermal resistance of the solid to the path
of heat flow.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
400 600 800 1000 1200 1400 1600
Reynolds Number, Re
F
o
u
r
i
e
r
N
u
m
b
e
r
,
F
o
s
s
93
Figure 4.25 Distribution of Dimensionless Maximum Temperature at the SolidFluid
Interface, within the Solid, and MaximumtoMinimum Temperature
Difference with Time for Different Plate Thicknesses (Silicon Plate Re=750,
ß=2.5)
Figure 4.26 shows the average Nusselt number variation as a function of
time for three distinct plate thicknesses using Constantan as the solid material.
The average Nusselt number is higher for higher plate thickness. The maximum
temperature at the solidfluid interface, maximum temperature within the solid,
and maximum to minimum temperature difference (as a measure for temperature
nonuniformity) for different solid materials is presented in Figure 4.27.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 0.1 0.2 0.3 0.4 0.5
Fourier Number, F
o
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m

t
o

M
i
n
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
D
i
f
f
e
r
e
n
c
e
a
t
h
e
I
n
t
e
r
f
a
c
e
,
Θ Θ Θ Θ
m
a
x

Θ Θ Θ Θ
m
i
n
Θmax(interface), b/d =0.08
Θmax(interface), b/d =1.0
Θmax(solid), b/d =0.08
Θmax(solid), b/d =1.0
ΘmaxΘmin(interface), b/d =0.08
ΘmaxΘmin(interface), b/d =1.0
n
n
n
n
n
n
94
Figure 4.26 Variation of Average Nusselt Number with Time for Constantan Plate at
Three Different Thicknesses (Re=750, ß=2.5)
As expected, the temperature changes occur faster at the beginning of the
heating process for all materials and the slope gradually decays when the
steadystate approaches. It can be observed that materials with a very low
thermal conductivity such as Constantan maintains a higher temperature at the
plate’s outer surface and within the solid as the thermal conductivity controls how
effective the heat flows and distributes throughout the material. The thermal
diffusivity of the material also contributes to the transient behavior of the solid.
As noticed, Silicon and Copper reach the steadystate faster than Constantan
due to their higher thermal diffusivity, which controls the rate of heat being
transferred through the solid material.
0
20
40
60
80
100
120
140
160
0 0.01 0.02 0.03 0.04 0.05 0.06
Fourier Number, F
o
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
Nuav, b/dn=0.08
Nuav, b/dn=1.5
, b/d
n
=0.08
, b/d
n
=1.5
95
Figure 4.27 Distribution of Dimensionless Maximum Temperature at the SolidFluid
interface, within the Solid, and MaximumtoMinimum Temperature
Difference with Time for Different Materials (Silicon Plate, Re=750, ß=2.5)
Figure 4.28 plots the distribution of average Nusselt Number with time for
the three materials used in this study. Constantan shows a higher average heat
transfer coefficient compared to Silicon or Copper over the entire transient
process. A significant difference is seen at the earlier part of the transient and
the curves come close together as the steadystate approaches. It will be also
important to know how the materials responded in reaching thermal equilibrium
based on their thickness.
0
0.05
0.1
0.15
0.2
0.25
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45
Fourier Number, Fo
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
,
Θ Θ Θ Θ
i
n
t
0
0.02
0.04
0.06
0.08
0.1
0.12
D
i
m
e
n
s
i
o
n
l
e
s
s
M
a
x
i
m
u
m

t
o

M
i
n
i
m
u
m
T
e
m
p
e
r
a
t
u
r
e
D
i
f
f
e
r
e
n
c
e
a
t
t
h
e
I
n
t
e
r
f
a
c
e
,
Θ Θ Θ Θ
m
a
x

Θ Θ Θ Θ
m
i
n
Θmax(interface), Constantan
Θmax(interface), Silicon
Θmax(interface), Copper
Θmax(solid), Constantan
Θmax(solid), Silicon
Θmax(solid), Copper
ΘmaxΘmin(interface),
Constantan
96
Figure 4.28 Variation of Average Nusselt Number with Time for Different Materials
(Re=750, b/d
n
,=0.5, ß=2.5)
Figure 4.29 presents Fo
ss
for these materials for different plate
thicknesses. As the thickness increases in value, the time to reach steadystate
also increases. Constantan takes longer in reaching steadystate due to its lower
thermal diffusivity compared to Copper and Silicon. Also, the property of the
solid plays more significant role in determining the duration of the transient heat
transfer process when the thickness is increased.
0
20
40
60
80
100
120
140
160
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Fourier Number, F
o
A
v
e
r
a
g
e
N
u
s
s
e
l
t
N
u
m
b
e
r
,
N
u
a
v
Nuav, Constantan
Nuav, Copper
Nuav, Silicon
, Constantan
, Silicon
, Copper
97
Figure 4.29 Time Needed to Reach SteadyState for Different Cylindrical Plate
Thicknesses and for Different Materials (Re=750, ß=2.5)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Dimensionless Plate Thickness, b/d
n
F
o
u
r
i
e
r
N
u
m
b
e
r
,
F
o
s
s
Constantan
Copper
Silicon
98
Chapter 5 Discussion and Conclusion
Local and average Nusselt number and heat transfer coefficient
distributions showed a strong dependence on the impingement velocity or
Reynolds number. As the velocity increases, the boundary layer or film thickness
decreases and Nusselt number increases over the entire solidfluid interface. A
lower thermal conductivity material showed higher local maximum Nusselt
number as well as higher average Nusselt number among all studied materials.
On the other hand, materials with higher thermal conductivity maintained a more
uniform temperature distribution throughout the solidfluid interface and facilitated
a higher heat transfer rate, lowering the maximum temperature inside the
hemisphere and at its interface. Also, increasing the thickness of the
hemispherical plate proved to decrease the solidfluid interface temperature due
to the resistance of the material to heat flow. The Nusselt number, however, did
not change much with thickness variation. The impingement height affected the
dimensionless interface temperature as well as the Nusselt number. A higher jet
impingement height provided a lower dimensionless interface temperature over
the entire hemisphere and a higher Nusselt number at the stagnation region.
Temperature and heat flux at the solidfluid interface rise with time
whereas the average heat transfer coefficient decreases with time. A larger heat
99
flux is seen near the stagnation region because of the larger temperature
difference between the water jet and the plate. The maximum dimensionless
interface temperature was encountered at the outer edge of the cylindrical plate
while the minimum was encountered at the stagnation region. Local Nusselt
number is high near the stagnation region where convective heat transfer rate is
more effective and then it decays monotonically along the remaining portion of
the plate. The Reynolds number was found to be an essential parameter in
controlling the transient process since the time required to reach steady state
diminished as the Reynolds number increased. Also, the maximum temperature
at the solidfluid interface as well as the temperature inside the solid decreased
as the Reynolds number increased while the maximum to minimum temperature
difference a the interface decreased as the Reynolds number decreased.
Increasing the plate thickness decreased the capacity of the plate of being
effectively cooled (lower average Nusselt number) and decreased dimensionless
maximum temperature at the interface and within the solid. The time required to
reach steadystate became larger as the thickness of the plate increased.
Materials with a higher thermal conductivity maintained a lower dimensionless
solidfluid interface temperature as well as dimensionless maximum temperature
and reached steadystate faster. Nevertheless, Constantan demonstrated to
have a higher average Nusselt number in comparison to Copper and Silicon. The
isothermal lines within the solid demonstrated the transition from a conduction
only at the very early part of the transient to conductionconvection equilibrium
heat transfer as the steady state was arrived.
100
References
[1] Rahman M.M., Dontaraju P., and Ponnappan R., 2002, “Confined jet
impingement thermal management using liquid ammonia as the working
fluid,” In: Proceedings of ASME IMECE2002, November 1722, New
Orleans, Louisiana 33033.
[2] Stevens J., and Webb B.W., 1989, “Local heat transfer coefficients under an
axisymmetric, singlephase liquid jet,” Heat Transfer in Electronics, 111,
pp. 113–119.
[3] Garimella S.V., and Rice R.A., 1995, “Confined and submerged liquid jet
impingement heat transfer,” Journal of Heat Transfer, 117, pp. 871877.
[4] Gomi T., and Webb B.W., 1997, “Local characteristics of impingement heat
transfer with oblique round freesurface jets of large prandtl number
liquid,” International Journal of Heat and Mass Transfer, 40 (10), pp.
22492259.
[5] Lee D.H., Chung Y.S., and Kim D.S., 1997, “Turbulent flow and heat transfer
measurements on a curved surface with a fully developed round
impinging jet,” International Journal of Heat and Fluid Flow, 18, pp. 160
169.
[6] Kornblum Y., Goldstein R.J., 1997, “Jet impingement on semicylindirical
concave and convex surfaces: part II  heat transfer,” International
Symposium on Physics of Heat Transfer in Boiling and Condensation,
pp. 597602.
[7] Lee D.H., Chung Y.S., and Kim M.G., 1999, “Turbulent heat transfer from a
convex hemispherical surface to a round impinging jet,” International
Journal of Heat and Mass Transfer, 42, pp. 11471156.
[8] Tong A.Y., 2003, “A numerical study on the hydrodynamics and heat transfer
of a circular liquid jet impinging onto a substrate,” Numerical Heat
Transfer, 44, pp. 119.
[9] Cornaro C., Fleischer A.S., and Goldstein R.J., 1999, “Flow visualization of a
round jet impinging on cylindrical surfaces,” Experimental Thermal and
Fluid Science, 20, pp. 6678.
101
[10] Cornaro C., Fleischer A.S., Rounds M., and Goldstein R.J., 2001,“Jet
impingement cooling of a convex semicylindrical surface,” Journal of
Thermal Science, 40, pp. 890898.
[11] Fleischer A.S., Kramer K., and Goldstein R.J., 2001, “Dynamics of the
vortex structure of a jet impinging on a convex surface,” Experimental
Thermal and Fluid Science, 24, pp. 169175.
[12] Baonga J. B., LouahliaGualous H., and Imbert M., 2006, “Experimental
study of the hydrodynamic and heat transfer of free liquid jet impinging a
flat circular heated disk,” Applied Thermal Engineering, 26, pp. 1125
1138.
[13] Inada S., Miyasaka Y., and Izumi R., 1981, “A study on the laminarflow
heat transfer between a twodimensional water jet and a flat surface with
constant heat flux,” In: Bulletin of the JSME─1982 JSME 24.
[14] Carper H.J., Jr., 1989, “Impingement cooling by liquid jet,” In: Heat Transfer
Division─1989 ASME HTD 117.
[15] Liu X., and Lienhard V., 1989, “Liquid jet impingement heat transfer on a
uniform flux surface,” In: Heat Transfer Phenomena in Radiation,
Combustion and Fires─1989 ASME HTD 106.
[16] Wadsworth D.C., and Mudawar I., 1989, “Cooling of a multichip electronic
module by means of confined twodimensional jets of dielectric liquid,”
Heat Transfer in Electronics, 111, pp. 7987.
[17] Gau C., and Chung C.M., 1991, “Surface curvature effect on slotairjet
impingement cooling flow and heat transfer process,” Journal of Heat
Transfer, 113, pp. 858864.
[18] Stevens J., and Webb B.W., 1992, “Measurements of the free surface flow
structure under an impinging, free liquid jet,” Journal of Heat Transfer,
114, pp. 7984.
[19] Teuscher K.L., Ramadhyani S., and Incropera F.P., 1993, “Jet impingement
cooling of an array of discrete heat sources with extended surfaces,”
ASME HTD, 263, pp. 110.
[20] Ma C.F., Zhuang Y., Lee S.C., and Gomi T., 1997, “Impingement heat
transfer and recovery effect with submerged jets of large Prandtl number
liquid—II. Initially laminar confined slot jets,” International Journal of
Heat and Mass Transfer, 40, pp. 14911500.
102
[21] Bartoli C., and Faggiani S., 1998, “Local nusselt number at a cylinder cooled
by a slot jet of water,” Heat and Technology, 16 (2), pp.3337.
[22] McDaniel C.S., and Webb B.W., 2000, “Slot jet impingement heat transfer
from circular cylinders,” International Journal of Heat and Mass Transfer,
43, pp. 19751985.
[23] Gori F., and Bossi L., 2000, “On the cooling effect of an air jet along the
surface of a cylinder,” International Communications in Heat and Mass
Transfer, 27, pp. 667676.
[24] Kayansayan N., and Küçüka S., 2001, “Impingement cooling of a semi
cylindrical concave channel by confined slotairjet,” Experimental
Thermal and Fluid Science, 25, pp. 383396.
[25] Shi Y.L., Ray M.B., and Mujumdar A.S., 2003, “Effects of prandtl number on
impinging jet heat transfer under a semiconfined laminar slot jet,”
International Communications Heat Mass Transfer, 30 (4), pp. 455464.
[26] Olsson E.E.M., Ahrné L.M., and Trägårdh A.C., 2004, “Heat transfer from a
slot air jet impinging on a circular cylinder,” Journal of Food Engineering,
63, pp. 393401.
[27] Chan T. L., Leung C.W., Jambunathan K., AshforthFrost S., Zhou Y., and
Liu M.H., 2002, “Heat transfer characteristics of a slot jet impinging on a
semicircular convex surface,” International Journal of Heat and Mass
Transfer, 45, pp. 9931006.
[28] Chan T.L., Zhou Y., Liu M.H., and C.W. Leung, 2003, “Mean flow and
turbulence measurements of the impingement wall jet on a semicircular
convex surface,” Experiments in Fluids, 34, pp. 140–149.
[29] Gori F., and Bossi L., 2003, “Optimal slot height in the jet cooling of a
circular cylinder,” Applied Thermal Engineering, 23, pp. 859870.
[30] Rahimi M., Owen I., and Mistry J., 2003, “Heat transfer between an under
expanded jet and a cylindrical surface,” International Journal of Heat and
Mass Transfer, 46, pp. 31353142.
[31] Yang Y.T., and Hwang C.H., 2004, “Numerical simulations on the
hydrodynamics of a turbulent slot jet impinging on a semicylindrical
convex surface, Numerical Heat Transfer, 46, pp. 9951008.
103
[32] Chen Y.C., Ma C.F., Qin, M., and Li, Y.X., 2005, “Theoretical study on
impingement heat transfer with singlephase freesurface slot jets,”
International Journal of Heat and Mass Transfer, 48, pp. 33813386.
[33] Zuckerman N., and Lior N., 2005, “Jet impingement heat transfer on a
circular cylinder by radial slot jets,” In: Proceedings of ASME
IMECE2005, November 511, Orlando, Florida USA 79565.
[34] White F. M., 2003, “Fluid mechanics 5
th
ed.,” McGrawHill, New York.
[35] Stevens J., and Webb B.W., 1992, “Measurements of the free surface flow
structure under an impinging, free liquid jet,” Journal of Heat Transfer,
114, pp. 7984.
[36] Liu X., Lienhard J.H., and Lombara J.S., 1993, “Convective heat transfer by
impingement of circular liquid jets,” Journal of Heat Transfer, 13, pp.
571–582.
[37] Scholtz M.T., and Trass O., 1970, “Mass transfer in a nonuniform impinging
jet,” AIChE Journal, pp. 8296.
[38] Nakoryakov V.E., Pokusaev B.G., and Troyan E.N., 1978, “Impingement of
an axisymmetric liquid jet on a barrier, International Journal of Heat and
Mass Transfer, 9, pp. 11751184.
[39] Martin H., 1977, “Heat and mass transfer between impinging gas jets and
solid surfaces,” Advances in Heat Transfer, 13, pp. 160.
[40] Whitaker S., 1972, “Forced convection heat transfer correlations for flow in
pipes, past flat plates, single cylinders, single spheres, and flow in
packed beds and tube bundles,” AIChE Journal, 18, pp. 361371.
104
Bibliography
White F. M., 2003, “Fluid mechanics 5
th
ed.,” McGrawHill, New York.
Bejan A., 1995, “Convection heat transfer 2
nd
ed.,” John Wiley & Sons, New
York.
Özisik M.N., 1993, “Heat conduction 2
nd
ed.,” John Wiley & Sons, New York.
105
Appendices
106
Appendix A: CFD Code for Axisymetric Model (FIDAP)
/ File opened for write Mon Oct 2 16:06:01 2006.
/ File opened for write Thu Jun 15 16:15:56 2006.
TITLE( )
FREE SURFACE JET ON CYLINDER
FIGEN( ELEM = 1, POIN = 1, CURV = 1, SURF = 1, NODE = 0, MEDG = 1,
MLOO = 1,
MFAC = 1, BEDG = 1, SPAV = 1, MSHE = 1, MSOL = 1, COOR = 1, TOLE =
0.0001 )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.000000 0.000000 0.000000 1.000000
10.00000 10.00000 7.50000 7.50000 7.50000
7.50000
//POINTS
POINT( ADD, COOR )
0, 0
0.3, 0
0.36, 0
0.86, 0
0.86, 0.5
0.86, 0.56
0.86, 0.59
0.273, 0.06
0, 0.06
0, 0.59
//LINES
POINT( SELE, ID )
4, 5
3
CURVE( ADD, ARC, CENT )
POINT( SELE, ID )
4
6
2
CURVE( ADD, ARC, CENT )
POINT( SELE, ID )
4
7, 8
CURVE( ADD, ARC, CENT )
POINT( SELE, ID )
8, 9
1, 3
CURVE( ADD, LINE )
POINT( SELE, ID )
5, 7
CURVE( ADD, LINE )
//CREATE SURFACE
POINT( SELE, ID )
10
7
107
Appendix A: (Continued)
1
4
SURFACE( ADD, POIN, ROWW = 2, NOAD )
//MESH EDGES
CURVE( SELE, ID )
1, 3
MEDGE( ADD, SUCC, INTE = 62, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 4 )
MEDGE( ADD, SUCC, INTE = 70, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 5 )
MEDGE( ADD, SUCC, INTE = 20, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 6 )
MEDGE( ADD, SUCC, INTE = 70, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID )
7, 8
MEDGE( ADD, SUCC, INTE = 20, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 9 )
MEDGE( ADD, SUCC, INTE = 20, RATI = 0, 2RAT = 0, PCEN = 0 )
//MESH LOOPS
/LOOP 1
CURVE( SELE, ID )
5
4
3
9
2
6
MLOOP( ADD, MAP, VISI, NOSH, EDG1 = 1, EDG2 = 2, EDG3 = 1, EDG4 = 2 )
/LOOP 2
CURVE( SELE, ID )
7
2
8
1
MLOOP( ADD, MAP, VISI, NOSH, EDG1 = 1, EDG2 = 1, EDG3 = 1, EDG4 = 1 )
//MESH FACES
/FACE 1
SURFACE( SELE, ID = 1 )
MLOOP( SELE, ID = 1 )
MFACE( ADD )
/FACE 2
SURFACE( SELE, ID = 1 )
MLOOP( SELE, ID = 2 )
MFACE( ADD )
//MESH FACE ENTITIES
ELEMENT( SETD, QUAD, NODE = 4 )
MFACE( SELE, ID = 1 )
MFACE( MESH, MAP, ENTI = "fluid" )
MFACE( SELE, ID = 2 )
MFACE( MESH, MAP, ENTI = "solid" )
//MESH EDGE ENTITIES
ELEMENT( SETD, EDGE, NODE = 2 )
MEDGE( SELE, ID = 1 )
108
Appendix A: (Continued)
MEDGE( MESH, MAP, ENTI = "bottom" )
MEDGE( SELE, ID = 2 )
MEDGE( MESH, MAP, ENTI = "interface" )
MEDGE( SELE, ID )
3, 4
MEDGE( MESH, MAP, ENTI = "free" )
MEDGE( SELE, ID = 5 )
MEDGE( MESH, MAP, ENTI = "finlet" )
MEDGE( SELE, ID )
6, 7
MEDGE( MESH, MAP, ENTI = "syme" )
MEDGE( SELE, ID = 8 )
MEDGE( MESH, MAP, ENTI = "swall" )
MEDGE( SELE, ID = 9 )
MEDGE( MESH, MAP, ENTI = "fout" )
END( )
FIPREP( )
//Fluid and solid properties
/B=2.5, D=0.12
/FLUID
DENSITY( ADD, SET = "water", CONS = 0.996 )
CONDUCTIVITY( ADD, SET = "water", CONS = 0.0014699 )
VISCOSITY( ADD, SET = "water", CONS = 0.00798 )
SPECIFICHEAT( ADD, SET = "water", CONS = 0.998137 )
SURFACETENSION( ADD, SET = "water", CONS = 73 )
/
/SOLID
DENSITY( ADD, SET = "silicon", CONS = 2.33 )
CONDUCTIVITY( ADD, SET = "silicon", CONS = 0.334608 )
SPECIFICHEAT( ADD, SET = "silicon", CONS = 0.17006 )
/
/ENTITIES
ENTITY( ADD, NAME = "fluid", FLUI, PROP = "water" )
ENTITY( ADD, NAME = "solid", SOLI, PROP = "silicon" )
ENTITY( ADD, NAME = "bottom", PLOT )
ENTITY( ADD, NAME = "syme", PLOT )
ENTITY( ADD, NAME = "swall", PLOT )
ENTITY( ADD, NAME = "finlet", PLOT )
ENTITY( ADD, NAME = "fout", PLOT )
ENTITY( ADD, NAME = "interface", PLOT, ATTA = "solid", NATT = "fluid" )
ENTITY( ADD, NAME = "free", SURF, ATTA = "fluid", DEPT = 21, SPIN,
STRA,
ANG1 = 80, ANG2 = 180 )
/
/SPECIFY BOUNDARY CONDITIONS
BCNODE( ADD, COOR, NODE = 190 )
BCNODE( ADD, SURF, NODE = 190, ZERO )
BCNODE( ADD, VELO, ENTI = "bottom", ZERO )
BCNODE( ADD, UY, ENTI = "finlet", ZERO )
BCNODE( ADD, UX, ENTI = "finlet", CONS = 33.38353 )
BCNODE( ADD, UY, ENTI = "syme", ZERO )
BCNODE( ADD, VELO, ENTI = "interface", ZERO )
BCNODE( ADD, VELO, ENTI = "swall", ZERO )
109
Appendix A: (Continued)
BCNODE( ADD, TEMP, ENTI = "finlet", CONS = 37 )
BCFLUX( ADD, HEAT, ENTI = "bottom", CONS = 5.971 )
BCNODE( ADD, VELO, ENTI = "solid", ZERO )
/
/PROBLEM DEFINITION
PROBLEM( 2D, LAMI, NONL, NEWT, MOME, ENER, FREE, TRAN, SING )
BODYFORCE( ADD, CONS, FX = 981, FY = 0, FZ = 0 )
PRESSURE( ADD, MIXE = 1e11, DISC )
DATAPRINT( ADD, CONT )
EXECUTION( ADD, NEWJ )
PRINTOUT( ADD, NONE )
OPTIONS( ADD, UPWI )
UPWINDING( ADD, STRE )
/
/SOLUTION ALGORITHM
SOLUTION( ADD, N.R. = 50, KINE = 25, VELC = 0.0001, RESC = 0.0001,
SURF = 0.001 )
TIMEINTEGRATION( ADD, BACK, NSTE = 1000, TSTA = 0, DT = 1e07, VARI,
WIND = 9,
NOFI = 10 )
POSTPROCESS( NBLO = 2 )
1, 201, 1
201, 1000, 1
/
/INITIAL CONDITIONS
ICNODE( ADD, UY, ENTI = "fluid", CONS = 18.38353 )
ICNODE( ADD, UX, ENTI = "fluid", CONS = 23.38353 )
ICNODE( ADD, TEMP, ENTI = "fluid", CONS = 37 )
END( )
CREATE( FISO )
RUN( FISOLV, BACK )
/ File closed at Mon Oct 2 16:06:12 2006.
/ File opened for append Tue Oct 3 12:46:08 2006.
FIPOST( )
TIMESTEP( STEP = 1 )
TIMESTEP( STEP = 322 )
CONVERGENCE( ALL, SOLU, LOG )
VECTOR( VELO )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.37148 0.15563 0.36300 0.17218 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
110
Appendix A: (Continued)
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.42935 0.31361 0.30826 0.20659 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.43717 0.33551 0.29887 0.20972 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.29797 0.04771 0.36300 0.14089 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
111
Appendix A: (Continued)
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.11340 0.41840 0.06401 0.33460 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.31204 0.52320 0.16255 0.35024 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
CONTOUR( TEMP, AUTO )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
CONTOUR( TEMP, AUTO = 200 )
MESH( NNUM )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
112
Appendix A: (Continued)
0.19004 0.00547 0.35049 0.17687 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.13912 0.09816 0.30546 0.26914 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
CONTOUR( TEMP, AUTO = 200 )
LINE( TEMP, ENTI = "interface" )
PRINT( TEMP, ENTI = "interface", SCRE )
PRINT( TEMP, NODE, NOD1 = 2913, NOD2 = 4053, NGEN = 19, SCRE )
PRINT( TEMP, NODE, NOD1 = 349, SCRE )
PRINT( TEMP, NODE, NOD1 = 386, SCRE )
END( )
END( )
/ File closed at Tue Oct 3 12:52:48 2006.
/ File opened for append Wed Oct 4 19:38:28 2006.
FIPOST( )
TIMESTEP( STEP = 1 )
TIMESTEP( STEP = 322 )
VECTOR( VELO )
GROUP( ENTI = "fluid" )
VECTOR( VELO )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.29640 0.52477 0.15316 0.35650 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
113
Appendix A: (Continued)
45.000000 45.000000 45.000000 45.000000
REDO
CONTOUR( STRE, AUTO )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.27138 0.47315 0.17819 0.35650 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
CONTOUR( TEMP, AUTO )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.23227 0.48723 0.12344 0.34868 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.430000 0.295000 0.000000 1.000000
0.00000 0.00000 0.00000 0.00000 0.00000
0.00000
45.000000 45.000000 45.000000 45.000000
REDO
VECTOR( VELO )
DEVICE( POST, FILE = "500" )
VECTOR( VELO )
CONTOUR( STRE, AUTO )
CONTOUR( TEMP, AUTO )
END( )
END( )
/ File closed at Wed Oct 4 19:48:27 2006.
114
Appendix B: CFD Code for 2D Model
/ File opened for write Fri Sep 22 20:31:22 2006.
// File opened for write Wed Mar 1 02:13:54 2006.
TITLE( )
FREE SURFACE JET IMPINGMENT
FIGEN( ELEM = 1, POIN = 1, CURV = 1, SURF = 1, NODE = 0, MEDG = 1,
MLOO = 1,
MFAC = 1, BEDG = 1, SPAV = 1, MSHE = 1, MSOL = 1, COOR = 1, TOLE = 1e
05 )
WINDOW(CHANGE= 1, MATRIX )
1.000000 0.000000 0.000000 0.000000
0.000000 1.000000 0.000000 0.000000
0.000000 0.000000 1.000000 0.000000
0.000000 0.000000 0.000000 1.000000
10.00000 10.00000 7.50000 7.50000 7.50000
7.50000
WINDOW( CHAN = 1, MATR )
1, 0, 0, 0
0, 1, 0, 0
0, 0, 1, 0
0, 0, 0, 1
10, 10, 7.5, 7.5, 7.5
7.5
//POINTS
POINT( ADD, COOR, X = 0, Y = 0 )
POINT( ADD, COOR, X = 0.06, Y = 0 )
POINT( ADD, COOR, X = 0.3, Y = 0 )
POINT( ADD, COOR, X = 0.3, Y = 0.06 )
POINT( ADD, COOR, X = 0.16, Y = 0.2 )
POINT( ADD, COOR, X = 0.1, Y = 0.31 )
POINT( ADD, COOR, X = 0.03, Y = 0.47 )
POINT( ADD, COOR, X = 0.27, Y = 0.62 )
POINT( ADD, COOR, X = 0.5, Y = 0.66 )
POINT( ADD, COOR, X = 0.5, Y = 0.56 )
POINT( ADD, COOR, X = 0.5, Y = 0.5 )
POINT( ADD, COOR, X = 0.319, Y = 0.466 )
POINT( ADD, COOR, X = 0.1516, Y = 0.359 )
POINT( ADD, COOR, X = 0.0565, Y = 0.231 )
POINT( ADD, COOR, X = 0.021, Y = 0.1425 )
POINT( ADD, COOR, X = 0.0036, Y = 0.06 )
POINT( ADD, COOR, X = 0.06, Y = 0 )
POINT( ADD, COOR, X = 0.0596, Y = 0.0672 )
POINT( ADD, COOR, X = 0.041, Y = 0.16 )
POINT( ADD, COOR, X = 0.0033, Y = 0.2587 )
POINT( ADD, COOR, X = 0.112, Y = 0.399 )
POINT( ADD, COOR, X = 0.299, Y = 0.52 )
POINT( ADD, COOR, X = 0.3, Y = 0.66 )
POINT( ADD, COOR, X = 0.5, Y = 0 )
//LINES (1,2,3,4)
POINT( SELE, ID = 1 )
POINT( SELE, ID = 2 )
CURVE( ADD, LINE )
115
Appendix B: (Continued)
POINT( SELE, ID = 3 )
POINT( SELE, ID = 4 )
CURVE( ADD, LINE )
POINT( SELE, ID = 4 )
POINT( SELE, ID = 5 )
POINT( SELE, ID = 6 )
CURVE( ADD, ARC )
//LINES (5,6)
CURVE( SELE, ID = 4 )
POINT( SELE, ID = 5 )
CURVE( SPLI )
//LINES (7,8,9)
POINT( SELE, ID = 6 )
POINT( SELE, ID = 7 )
POINT( SELE, ID = 8 )
CURVE( ADD, ARC )
CURVE( SELE, ID = 7 )
POINT( SELE, ID = 7 )
CURVE( SPLI )
CURVE( SELE, ID = 9 )
CURVE( DELE )
POINT( SELE, ID = 7 )
POINT( SELE, ID = 8 )
POINT( SELE, ID = 9 )
CURVE( ADD, ARC )
//LINES (10,11,12,13)
POINT( SELE, ID = 9 )
POINT( SELE, ID = 10 )
CURVE( ADD, LINE )
POINT( SELE, ID = 10 )
POINT( SELE, ID = 11 )
CURVE( ADD, LINE )
POINT( SELE, ID )
11
12
13
CURVE( ADD, ARC )
POINT( SELE, ID )
13
14
15
CURVE( ADD, ARC )
CURVE( SELE, ID = 13 )
POINT( SELE, ID = 14 )
CURVE( SPLI )
//LINES (14,15,16,17,18)
POINT( SELE, ID )
15
16
1
CURVE( ADD, ARC )
116
Appendix B: (Continued)
//LINES (19,20,21,22)
POINT( SELE, ID )
2
18
19
CURVE( ADD, ARC )
CURVE( SELE, ID = 19 )
POINT( SELE, ID = 18 )
CURVE( SPLI )
POINT( SELE, ID )
19
20
21
CURVE( ADD, ARC )
//LINES (23,24,25,26)
CURVE( SELE, ID = 22 )
POINT( SELE, ID = 20 )
CURVE( SPLI )
POINT( SELE, ID )
21
22
10
CURVE( ADD, ARC )
POINT( SELE, ID )
20
6
CURVE( ADD, LINE )
/SURFACES
POINT( SELE, ID = 24 )
POINT( SELE, ID = 3 )
POINT( SELE, ID = 9 )
POINT( SELE, ID = 23 )
SURFACE( ADD, POIN, ROWW = 2, NOAD )
//MESH EDGES
CURVE( SELE, ID = 1 )
MEDGE( ADD, SUCC, INTE = 10, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 2 )
MEDGE( ADD, SUCC, INTE = 30, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 3 )
MEDGE( ADD, SUCC, INTE = 15, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 5 )
MEDGE( ADD, SUCC, INTE = 22, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 6 )
MEDGE( ADD, SUCC, INTE = 16, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 8 )
MEDGE( ADD, SUCC, INTE = 22, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 9 )
MEDGE( ADD, SUCC, INTE = 44, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 10 )
MEDGE( ADD, SUCC, INTE = 30, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 11 )
117
Appendix B: (Continued)
CURVE( SELE, ID = 14 )
MEDGE( ADD, SUCC, INTE = 22, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 15 )
MEDGE( ADD, SUCC, INTE = 16, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 17 )
MEDGE( ADD, SUCC, INTE = 22, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 18 )
MEDGE( ADD, SUCC, INTE = 15, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 20 )
MEDGE( ADD, SUCC, INTE = 15, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 21 )
MEDGE( ADD, SUCC, INTE = 22, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 23 )
MEDGE( ADD, SUCC, INTE = 16, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 24 )
MEDGE( ADD, SUCC, INTE = 22, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 25 )
MEDGE( ADD, SUCC, INTE = 44, RATI = 0, 2RAT = 0, PCEN = 0 )
CURVE( SELE, ID = 26 )
MEDGE( ADD, SUCC, INTE = 30, RATI = 0, 2RAT = 0, PCEN = 0 )
//MESH LOOPS
//LOOP 1
CURVE( SELE, ID = 2 )
CURVE( SELE, ID = 3 )
CURVE( SELE, ID = 5 )
CURVE( SELE, ID = 6 )
CURVE( SELE, ID = 26 )
CURVE( SELE, ID = 23 )
CURVE( SELE, ID = 21 )
CURVE( SELE, ID = 20 )
MLOOP( ADD, MAP, VISI, NOSH, EDG1 = 1, EDG2 = 3, EDG3 = 1, EDG4 = 3 )
//LOOP 2
CURVE( SELE, ID = 26 )
CURVE( SELE, ID = 8 )
CURVE( SELE, ID = 9 )
CURVE( SELE, ID = 10 )
CURVE( SELE, ID = 25 )
CURVE( SELE, ID = 24 )
MLOOP( ADD, MAP, VISI, NOSH, EDG1 = 1, EDG2 = 2, EDG3 = 1, EDG4 = 2 )
//LOOP 3
CURVE( SELE, ID = 1 )
CURVE( SELE, ID = 20 )
CURVE( SELE, ID = 21 )
CURVE( SELE, ID = 23 )
CURVE( SELE, ID = 24 )
CURVE( SELE, ID = 25 )
CURVE( SELE, ID = 11 )
CURVE( SELE, ID = 12 )
CURVE( SELE, ID = 14 )
CURVE( SELE, ID = 15 )
CURVE( SELE, ID = 17 )CURVE( SELE, ID = 18 )
118
Appendix B: (Continued)
SURFACE( SELE, ID = 1 )
MLOOP( SELE, ID = 1 )
MFACE( ADD )
SURFACE( SELE, ID = 1 )
MLOOP( SELE, ID = 2 )
MFACE( ADD )
SURFACE( SELE, ID = 1 )
MLOOP( SELE, ID = 3 )
MFACE( ADD )
// MESHING
MFACE( SELE, ID = 3 )
ELEMENT( SETD, QUAD, NODE = 4 )
MFACE( MESH, MAP, NOSM, ENTI = "silicon" )
MFACE( SELE, ID = 1 )
MFACE( SELE, ID = 2 )
ELEMENT( SETD, QUAD, NODE = 4 )
MFACE( MESH, MAP, ENTI = "water" )
//MESH MAP ELEMENT ID
ELEMENT( SETD, EDGE, NODE = 2 )
MEDGE( SELE, ID )
1
9
MEDGE( MESH, MAP, ENTI = "sides" )
MEDGE( SELE, ID = 2 )
MEDGE( MESH, MAP, ENTI = "axis" )
MEDGE( SELE, ID = 3 )
MEDGE( MESH, MAP, ENTI = "inlet" )
MEDGE( SELE, ID )
4
5
6
7
MEDGE( MESH, MAP, ENTI = "surface" )
MEDGE( SELE, ID = 8 )
MEDGE( MESH, MAP, ENTI = "outlet" )
MEDGE( SELE, ID )
10
11
12
13
14
MEDGE( MESH, MAP, ENTI = "bottom" )
MEDGE( SELE, ID )
15
16
17
18
19
MEDGE( MESH, MAP, ENTI = "interface" )
END( )
FIPREP( )
//Fluid and solid properties
DENSITY( ADD, SET = "water", CONS = 0.996 )
119
Appendix B: (Continued)
CONDUCTIVITY( ADD, SET = "water", CONS = 0.0014699 )
VISCOSITY( ADD, SET = "water", CONS = 0.00798 )
SPECIFICHEAT( ADD, SET = "water", CONS = 0.998137 )
SURFACETENSION( ADD, SET = "water", CONS = 73 )
DENSITY( ADD, SET = "silicon", CONS = 2.33 )
CONDUCTIVITY( ADD, SET = "silicon", CONS = 0.334608 )
SPECIFICHEAT( ADD, SET = "silicon", CONS = 0.17006 )
ENTITY( ADD, NAME = "water", FLUI, PROP = "water" )
ENTITY( ADD, NAME = "inlet", PLOT )
ENTITY( ADD, NAME = "outlet", PLOT )
ENTITY( ADD, NAME = "surface", SURF, DEPT = 31, SPIN, STRA, ANG1 = 45,
ANG2 = 300 )
ENTITY( ADD, NAME = "bottom", PLOT )
ENTITY( ADD, NAME = "axis", PLOT )
ENTITY( ADD, NAME = "silicon", SOLI, PROP = "silicon" )
ENTITY( ADD, NAME = "sides", PLOT )
ENTITY( ADD, NAME = "interface", PLOT, ATTA = "silicon", NATT = "water"
)
BODYFORCE( ADD, CONS, FX = 981, FY = 0, FZ = 0 )
PRESSURE( ADD, MIXE = 1e11, DISC )
DATAPRINT( ADD, CONT )
EXECUTION( ADD, NEWJ )
PRINTOUT( ADD, NONE, BOUN )
OPTIONS( ADD, UPWI )
UPWINDING( ADD, STRE )
RELAXATION( )
0.5, 0.5, 0.5, 0, 0.05, 0.1
BCNODE( ADD, COOR, NODE = 42 )
BCNODE( ADD, SURF, NODE = 42, ZERO )
BCNODE( ADD, VELO, ENTI = "bottom", ZERO )
BCNODE( ADD, URC, ENTI = "inlet", ZERO )
BCNODE( ADD, UZC, ENTI = "inlet", CONS = 50.0753 )
BCNODE( ADD, URC, ENTI = "axis", ZERO )
BCNODE( ADD, VELO, ENTI = "interface", ZERO )
BCNODE( ADD, VELO, ENTI = "sides", ZERO )
BCNODE( ADD, UT, NODE = 42, ZERO )
BCNODE( ADD, TEMP, ENTI = "inlet", CONS = 37 )
BCFLUX( ADD, HEAT, ENTI = "bottom", CONS = 5.971 )
BCNODE( ADD, VELO, ENTI = "silicon", ZERO )
PROBLEM( ADD, CYLI, INCO, TRAN, LAMI, NONL, NEWT, MOME, ENER, FREE,
SING )
SOLUTION( ADD, N.R. = 50, KINE = 25, VELC = 0.0001, RESC = 0.0001,
SURF = 0.001 )
TIMEINTEGRATION( ADD, BACK, NSTE = 1000, TSTA = 0, DT = 1e05, VARI,
WIND = 9,
NOFI = 12 )
POSTPROCESS( NBLO = 2 )
1, 301, 10
301, 1000, 1
CLIPPING( ADD, MINI )
1e20, 1e20, 1e20, 1e11, 37, 0
ICNODE( ADD, URC, ENTI = "water", CONS = 20 )
120
Appendix B: (Continued)
CREATE( FISO )
RUN( FISOLV, IDEN = "t06", BACK, AT = "", TIME = "NOW", COMP )
/ File closed at Fri Sep 22 20:31:32 2006.
Numerical Analysis of Heat Transfer During Jet Impingement on Curved Surfaces
by
Cesar F. HernandezOntiveros
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Mechanical Engineering Department of Mechanical Engineering College of Engineering University of South Florida
Major Professor: Muhammad Mustafizur Rahman, Ph.D. Frank Pyrtle III, Ph.D. Autar Kaw, Ph.D.
Date of Approval: March 30, 2007
Keywords: steady state, transient analysis, hemispherical plate, cylindrical plate, heat flux © Copyright 2007, Cesar F. HernandezOntiveros
Dedication
To
God
My Father and Mother
Without them this could not have been possible
To
My Advisor Professor Muhammad Mustafizur Rahman
Thank you for your immense patience and guidance
Jorge C.Acknowledgments To My Friends: Son H. Ho. and Phaninder Injeti for their help and support throughout this very long journey! . Lallave.
3 Numerical Computation 2.4 Mesh Independence and Time Step Study 2.2.1 Governing Equations: Steady State Heating 2.1 Cylindrical Coordinates 2.3 Thesis Aim Chapter 2 Mathematical Models and Computation 2.1.2 Transient Heating iii x xiii 1 2 2 4 8 8 10 10 12 12 14 15 16 17 17 19 19 21 21 23 24 24 26 29 29 52 i .4.3 Governing Equations : Transient Heating 2.2 Cylindrical Plate Model 2.1.2.3 Governing Equations : Transient Heating 2.2 Cartesian Coordinates Chapter 3 Hemispherical Model Results 3.Table of Contents List of Figures List of Symbols Abstract Chapter 1 Introduction 1.4 Boundary Conditions : Transient Heating 2.4 Boundary Conditions : Transient Heating 2.1 Governing Equations : Steady State Heating 2.2 Overview of Literature 1.3.2.1 Steady State Process 2.1 Hemispherical Model 2.4.2 Boundary Conditions : Steady State Heating 2.2 Transient Process 2.1 Steady State Heating 3.2 Boundary Conditions : Steady State Heating 2.1.3.1.1 Round Jet Impingement 1.1.1.1 Literature Review 1.2 Slot Jet Impingement 1.2.
1 Steady State Heating 4.Chapter 4 Cylindrical Plate Model Results 4.2 Transient Heating Chapter 5 Discussion and Conclusions References Bibliography Appendices Appendix A: CFD Code for Axisymetric Model Appendix B: CFD Code for 2D Model 66 66 85 98 100 104 105 106 114 ii .
b/dn=0.5. Hn = 0. Hn = 0. q=250 kW/m2) 11 Figure 2.30 cm. Hn = 0. q=250 kW/m2) SolidFluid Interface Dimensionless Maximum Temperature Variation for Silicon Hemisphere at Different Time Increments (Re=750.0.6 26 Figure 2.2 11 Figure 2.8 28 iii .7 27 Figure 2.1 Three Dimensional Schematic of a Hollow Hemisphere During an Axisymetric Liquid Jet Impingement CrossSectional View of a Hollow Hemisphere During an Axisymetric Jet Impingement Schematic View of a Curved Plate During a 2D Symmetric Liquid Jet Impingement Mesh Plot of a Curved Plate During a 2D Symmetric Liquid Jet Impingement Local Dimensionless Interface Temperature for Different Number of Elements in z (or r) and Φ Directions for Water as Fluid and Silicon as Solid (Re=750. ß=2. ß=2.4 16 21 Figure 2.5. b/dn=0.5.5.5 25 Figure 2.3 Figure 2.List of Figures Figure 2.5. b/dn=0.30 cm) SolidFluid Interface Dimensionless Maximum Temperature Variation for Silicon Plate at Different Time Increments (Re=750. ß=2.0.30 cm. and y Directions (Re=750. q=250 kW/m2) Local Dimensionless Interface Temperature for Different Number of Elements in x or φ. ß=2.
Figure 3.1
Velocity Vector Distribution for Silicon Hemisphere with Water as the Cooling Fluid (Re = 750, b = 0.60 mm, Q = 5.665x107 m3/s, β = 2.0, Hn = 0.30 cm, q = 250 kW/m2) Pressure Distribution for Silicon Hemisphere with Water as the Cooling Fluid (Re = 750, b = 0.60 mm, Q = 5.665x107 m3/s, β = 2.0, Hn = 0.30 cm, q = 250 kW/m2) Temperature Distribution for Silicon Hemisphere with Water as the Cooling Fluid (Re = 750, b = 0.60 mm, Q = 5.665x107 m3/s, β = 2.0, Hn = 0.30 cm, q = 250 kW/m2) Free Surface Height Distribution for Different Reynolds Numbers and Water as the Cooling Fluid (b=0.60 mm, β=2.0, Hn=0.30 cm, q=250 kW/m2) Dimensionless Interface Temperature Distribution for a Silicon Hemisphere at Different Reynolds Numbers, and Water as the Cooling Fluid (β=2.0, Hn= 0.30 cm, b=0.60 mm, To=373 K) Local Nusselt Number Distribution for a Silicon Hemisphere at Different Reynolds Numbers, and Water as the Cooling Fluid (β=2.0, Hn=0.30 cm, b=0.60 mm, To=373 K) Dimensionless Interface Temperature Distribution for a Silicon Hemispherical Plate at Different Reynolds Numbers, and Water as the Cooling Fluid (β=2.0, Hn= 0.30 cm, b=0.60 mm, q=250 kW/m2) Local Nusselt Number Distribution for a Silicon plate at Different Reynolds Numbers, and Water as the Cooling Fluid (β =2.0 , Hn = 0.30 cm, b=0.60 mm, q=250 kW/m2) Average Nusselt Number and Heat Transfer Coefficient Variation for Different Reynolds Numbers for Constant Heat Flux (q=250kW/m2) and Isothermal (To = 373 K) Boundary Conditions (β=2.0, Hn = 0.30cm)
30
Figure 3.2
31
Figure 3.3
32
Figure 3.4
33
Figure 3.5
35
Figure 3.6
35
Figure 3.7
36
Figure 3.8
37
Figure 3.9
38
Figure 3.10 Dimensionless Interface Temperature for Silicon Hemisphere at Different Nozzle to Target Spacing Ratio (β) for Water as the Cooling Fluid (Re=500, Q=3.776x107m3/s, dn=1.2mm, b=0.60 mm, q=250 kW/m2)
39
iv
Figure 3.11
Local Nusselt Number for Silicon Hemisphere at Different Nozzle to Target Spacing Ratio (β) for Water as the Cooling Fluid (Re=500, Q=3.776x107 m3/s, dn=1.2mm, b=0.60mm, q=250kW/m2) Dimensionless Interface Temperature for Silicon Hemisphere at Different Thickness (b) for Water as the Cooling Fluid (Re=750, Q=5.665x107m3/s, dn=1.2mm, b=0.60 mm, q=250 kW/m2) Local Nusselt Number for Silicon Hemisphere at Different Thickness (b) for Water as the Cooling Fluid (Re=750, Q=5.665x107 m3/s, dn=1.2mm, b=0.60mm, q=250kW/m2) Dimensionless Interface Temperature for Silicon Plate at Different Thickness (b) for Water as the Cooling Fluid (Re=750, Q=5.665x107m3/s, dn=1.2mm, b=0.60 mm, To=373 K) Local Nusselt Number for Silicon Hemisphere at Different Thickness (b) for Water as the Cooling Fluid (Re=750, Q=5.665x107 m3/s, dn=1.2mm, b=0.60mm, To=373 K) Dimensionless Interface Temperature for Silicon Hemispherical Plate for Different Cooling Fluids (Re=750, Q=5.665x107m3/s, dn=1.2mm, b=0.60 mm, To=373 K) Local Nusselt Number for Silicon Hemisphere for Different Cooling Fluids (Re=750, Q=5.665x107 m3/s, dn=1.2mm, b=0.60mm, To=373 K) Local Nusselt Number and Dimensionless Interface Temperature for Different Cooling Fluids (Re = 1500, Q=1.133x106m3/s, β=2.0, Hn=0.30 cm, b=0.60 mm, q=250 kW/m2) Local Nusselt Number and Dimensionless Interface Temperature for Different Hemisphere Materials with Water as the Cooling Fluid (Re=1000, β=2.0, b/dn=0.5, q=250 kW/m2) Dimensionless Interface Temperature for Silicon Hemispherical Plate for Different Materials (Re=750, Q=5.665x107m3/s, dn=1.2mm, b=0.60 mm, To=373 K) v
40
Figure 3.12
41
Figure 3.13
42
Figure 3.14
43
Figure 3.15
44
Figure 3.16
45
Figure 3.17
45
Figure 3.18
46
Figure 3.19
48
Figure 3.20
49
Figure 3.21
Local Nusselt Number Distribution for Silicon Plate for Different Solid Materials (Re=750, Q=5.665x107 m3/s, dn=1.2mm, b=0.60mm, To=373 K) Stagnation Nusselt Number Compared with Liu et al. [36], Scholtz and Trass [37], and Nakoryakov et al. [38] with Actual Numerical Results under Different Reynolds Numbers
50
Figure 3.22
51
Figure 3.23
Dimensionless Local Heat Flux Variation at the SolidFluid Interface for Silicon Hemisphere at Different Time Instants 53 (Re=750, b/dn=0.5, ß=2.5) Dimensionless Interface Temperature Variation for Silicon Hemisphere at Different Time Instants (Re=750, b/dn=0.5, ß=2.5)
Figure 3.24
54
Figure 3.25.
Local Nusselt Number Distribution for Silicon Hemisphere at Different Time Instants (Re=750, b/dn=0.5, ß=2.5) 55 Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface, within the Solid, and MaximumtoMinimum Temperature Difference with Time for Two Reynolds Numbers (Silicon Hemisphere, b/dn=0.5, ß=2.5) Variation of Average Nusselt Number with Time For Silicon Hemisphere at Two Different Reynolds Numbers (b/dn=0.5, ß=2.5) Time Required to Reach SteadyState for Silicon Hemisphere at Different Reynolds Numbers (b/dn=0.5, ß=2.5) Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface, within the Solid, and MaximumtoMinimum Temperature Difference with Time for Different Plate Thicknesses (Silicon Hemisphere Re=750, ß=2.5) Variation of Average Nusselt Number with Time for Constantan Hemisphere at Three Different Thicknesses (Re=750, ß=2.5)
Figure 3.26
56
Figure 3.27
57
Figure 3.28
58
Figure 3.29
59
Figure 3.30
60
vi
r=0) Velocity Vector Distribution for a Silicon Curved Plate with Water as the Cooling Fluid (Re = 750.5) Isotherms Countour Lines at Varius Time Instants for Silicon Hemisphere Plate of b/dn=0. ß=2.5.5 (Re=750.5.5.35 65 Figure 4.Figure 3. within the Solid.5.60 mm. q=250 kW/m2) Local Nusselt Number Distribution for a Silicon Curved Plate at Different Reynolds Numbers and Water as the Cooling Fluid (β =2.31 Distribution of Dimensionless Maximum Temperature at the SolidFluid interface.0. 68 Hn=0.4 70 Figure 4.60 mm.=0. ß=2. Top Left Corner (z=0. q=250 kW/m2) Dimensionless Interface Temperature Distribution for a Silicon Curved Plate at Different Reynolds Numbers and Water as the Cooling Fluid (β=2.34 64 Figure 3. Q = 5.3 cm. b=0. ß=2. b/dn.32 62 Figure 3.1 67 Figure 4.3 69 Figure 4.30 cm. Hn = 0. r=0)) Isotherms Countour Lines at Varius Time Instants for Silicon Hemisphere Plate of b/dn=1. β=2.2 Free Surface Height Distribution for Different Reynolds Numbers and Water as the Cooling Fluid (b=0.60 mm. Hn=0. q = 250 kW/m2) 61 Figure 3.5) Time Needed to Reach SteadyState for Different Hemispherical Plate Thicknesses and for Different Materials (Re=750.5.30 cm. and MaximumtoMinimum Temperature Difference with Time for Different Materials (Re=750.33 63 Figure 3. q=250 kW/m2) Average Nusselt Number and Maximum Temperature Within the Solid for Different Reynolds Numbers for Constant Heat Flux Conditions (q = 250 KW/m2. b= 0. ß=2.665x107 m3/s.5) Variation of Average Nusselt Number with Time for Different Materials (Re=750. β = 2. Top Left Corner (z=0.30cm) Figure 4. ß=2.60 mm.5. b=0. β=2. Hn= 0.5. Hn = 0.30 cm.5 (Re=750.5 71 vii .
Re=750. b=0. dn=1. q=250 kW/m2 Dimensionless Interface Temperature for Different Plate Thicknesses (β =2. q=250 kW/m2) Heat Transfer Coefficient and Nusselt Number Distributions for Different Nozzle Slot Widths (Q=0.2mm. b=0.60 mm.5.2mm.2mm.60 mm.5. Re=750.15 81 viii .2mm.8 74 Figure 4.2mm. Re=750) Dimensionless Interface Temperature for Different Cooling Fluids (β=2. b=0.30 cm. Hn = 0.30 cm Re=750. q=250kW/m2) Dimensionless Interface Temperature for Silicon Curved Plate for Different Inner Plate Radius of Curvature to Nozzle Diameter Ratio (Ri/dn) for Water as the Cooling Fluid (β=2. β =2. q=250 kW/m2) Local Nusselt Number Distribution for Silicon Curved Plate for Different Nozzle to Target Spacing Ratio (β) and Water as the Cooling Fluid (Re=750. dn=1.5. q=250 kW/m2) Local Nusselt Number Distribution for Silicon Curved Plate for Different Inner Plate Radius of Curvature to Nozzle Diameter Ratio (Ri/dn) for Water as the Cooling Fluid (β =2.60 mm.9 75 Figure 4. Hn = 0.11 77 Figure 4.5 .30 cm Re=750. dn=1. q=250 kW/m2) Local Nusselt Number Distribution for Different Plate Thicknesses (β =2.5.30 cm β=2.14 80 Figure 4.60mm. Hn = 0. q=250 kW/m2) Average Nusselt Number and Maximum Temperature in the Solid Variations with Plate Thickness (β =2.0006 m3/s. Re=750) 72 Figure 4.5.Figure 4. q=250 kW/m2) Average Nusselt Number and Maximum Temperature in the Solid Variations (Θmax) for Silicon Curved plate for Different Inner Plate Radius of Curvature to Nozzle Diameter Ratio (Ri/dn) for Water as the Cooling Fluid (β =2.2mm. Hn = 0.5 . Re=750. dn=1.5 .6 Dimensionless Interface Temperature for Silicon Curved Plate for Different Nozzle to Target Spacing Ratio (β) and Water as the Cooling Fluid (Re=750.10 76 Figure 4.13 79 Figure 4. b=0. dn=1. Hn = 0.5. b=0. dn=1.7 73 Figure 4.60 mm.30 cm Re=750.12 78 Figure 4.
b/dn=0.24 92 Figure 4.Figure 4.5) Local Nusselt Number Distribution for Silicon Plate at Different Time Instants (Re=750.17 83 Figure 4. b/dn=0.5. ß=2. and MaximumtoMinimum Temperature Difference with Time for Two Reynolds Numbers (Silicon Plate.23 91 Figure 4. β=2.35 kW/m2) Figure 4.19 Dimensionless Local Heat flux Variation at the SolidFluid Interface for Silicon Plate at Different Time Instants (Re=750. within the Solid. ß=2.2 mm. dn=2.20 88 Figure 4.5. ß=2.5) Dimensionless Interface Temperature Variation for Silicon Plate at Different Time Instants (Re=750.50 mm.26 94 ix . q=2. ß=2.5) Local Nusselt Number Compared with Bartoli and Faggiani [21] at Different Reynolds Numbers (φ=90°.5) Variation of Average Nusselt Number with Time for Constantan Plate at Three Different Thicknesses (Re=750. ß=2. b/dn=0.25 93 Figure 4. q=50 kW/m2) 82 Figure 4. b/dn=0.16 Heat Transfer Coefficient and Nusselt Number Distributions for Different Cooling Fluids (Re=750. ß=2.5) Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface.0 mm.5. ß=2.21 89 Figure 4.5. within the Solid. b=0.22 90 Figure 4.5) Variation of Average Nusselt Number with Time For Silicon Plate at Two Different Reynolds Numbers (b/dn=0.5) 84 86 Figure 4.5.5.5 mm. b = 0.5) Time Required to Reach SteadyState for Silicon Plate at Different Reynolds Numbers (b/dn=0.5) Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface. and MaximumtoMinimum Temperature Difference with Time for Different Plate Thicknesses (Silicon Plate Re=750.18 Average Nusselt Number Compared with Gori and Bossi [29] and Whitaker [40] for Different Reynolds Numbers (dn=2. ß=2.
5) 95 Figure 4.5.28 96 Figure 4. Re=750. ß=2. ß=2. within the Solid. b/dn.Figure 4.29 97 x .=0.5) Variation of Average Nusselt Number with Time for Different Materials (Re=750. and MaximumtoMinimum Temperature Difference with Time for Different Materials (Silicon Plate.27 Distribution of Dimensionless Maximum Temperature at the SolidFluid interface.5) Time Needed to Reach SteadyState for Different Cylindrical Plate Thicknesses and for Different Materials (Re=750. ß=2.
Ri. (hav⋅dn)/kf [m] Numax Maximum Nusselt number for the entire surface. Diameter of the nozzle [m] Fourier number. qint/(TintTj) Average heat transfer coefficient [W/m2K]. αf t/ dn2 Acceleration due to gravity [m/s2] Heat transfer coefficient [W/m2K].(14) Distance of the nozzle from the point of impingement [m] Thermal conductivity [W/m K] Nusselt number. (h⋅dn)/kf Average Nusselt number for the entire surface. (hmax⋅dn)/kf n p Pr q Q r ri ro Coordinate normal to the surface Pressure [Pa] Prandtl number. defined by Eqn. ro.List of Symbols b dn Fo g h hav Hn k Nu Nuav Plate thickness.ri. νf/αf Heat flux [W/m2] Fluid flow rate [m3/s] Radial coordinate [m] Inner radius of hemisphere [m] Outer radius of hemisphere [m] xi . Ro.
z Vx. Ro Φ. (VJ⋅dn)/νf Inner radius of curvature of the plate [m] Outer radius of curvature of the plate [m] Coordinate along the arc length. y x y z Time [s] Jet velocity [m/s] Velocity component in the r.R Re Ri Ro s T To Radius of curvature [m] Reynolds number. zdirection [m/s] Velocity component in the x. Hn/dn Liquid film thickness [m] Kinematic viscosity [m2/s] Angular coordinate [rad] Dimensionless temperature (constant heat flux boundary condition). ro Φ Temperature [K] Constant Temperature at the inner surface of hemisphere (isothermal case) [K] [m] t VJ Vr. 2⋅kf ⋅(Tint TJ)/ (q⋅dn) xii . ydirection [m/s] coordinate along xaxis [m] coordinate along yaxis [m] Axial coordinate [m] Greek Symbols: α β δ ν θ Θ Thermal diffusivity [m2 /s] Dimensionless nozzle to target spacing.
(TintTJ)/ (ToTJ ) Subscripts: atm av f i int j max n s SS w Ambient Average Fluid Initial Condition Solidfluid Interface Jet or inlet Maximum Nozzle solid Steady State Inner surface of hemisphere xiii .Φ Φ ρ σ Angular coordinate for curved plate [rad] Azimuthal coordinate for hemisphere [rad] Density [kg/m3] Surface tension [N/m] Dimensionless temperature (isothermal boundary condition).
Computations for the round nozzle impinging jet on the hemispherical plate and cylindrical plate were done for jet Reynolds number (ReJ) ranging from 500 to 2000.5. Results are presented for dimensionless solidfluid interface temperature. and different nozzle diameters (dn). The simulation model incorporated the entire fluid region and the solid hemisphere or curved plate.1616. Also. plate thickness to nozzle diameter ratio (b/dn) of 0.0. Solution was done for both isothermal and constant heat flux boundary conditions at the inner surface of the hemispherical plate and the constant heat flux boundary condition at the inner surface of the cylindrical shaped plate. computations for the slot nozzle impinging jet on the cylindrical plate were done for inner plate radius of curvature to nozzle diameter ratio (Ri/dn) of 4. dimensionless maximum temperature in the solid.0831.Numerical Analysis of Heat Transfer During Jet Impingement on Curved Surfaces Cesar F.75 to 3. and for various dimensionless plate thicknesses to diameter nozzle ratio (b/dn) from 0. dimensionless nozzle to target spacing ratio (β) from 0.66.081. xiv . HernandezOntiveros ABSTRACT The flow structure and convective heat transfer behavior of a free liquid jet ejecting from a round nozzle impinging vertically on a hemispherical solid plate and a slot nozzle impinging vertically on a cylindrical curved plate have been studied using a numerical analysis approach.
Decreasing the nozzle to target spacing favors the increasing of the Nusselt number. flouroinert (FC77). A higher Reynolds number increased the Nusselt number over the entire solidfluid interface. silver. copper. Increasing the nozzle diameter decreases the temperature at the curved plate outer surface and increases the local Nusselt number. local and average Nusselt number was enhanced by decreasing plate thickness. the local Nusselt number increases over the entire solidfluid interface. Numerical simulation results are validated by comparing with experimental measurements and related correlations. Similarly.local and average Nusselt numbers using the following fluids: water (H2O). Constantan. and silicon. Materials with higher thermal conductivity maintained a more uniform temperature distribution at the solidfluid interface. xv . Local and average Nusselt number and heat transfer coefficient distributions showed a strong dependence on the impingement velocity or Reynolds number. As the velocity increases. and oil (MIL7808) and the following solid materials: aluminum.
These include cooling and heating of metal plates. cleaning of industrial equipment. The structure of the free surface depends on surface tension. Freesurface jets are created when a liquid discharges into ambient air or other type of gaseous environment [1]. cleaning of iced aircraft wings. The advantage of using jets translates into reduced operational cost by improving the cooling efficiency of the system or equipment and reduces early failure of it. and cooling of electrical equipment. Enhancement of such heating/cooling processes will require information on the effects of jet velocity. Thus. then. gravitational and pressure forces. the need for heating/cooling processes able to transfer or remove very high heat fluxes appeals to liquid jet impingement as a choice. cooling of turbine blades. thermophysical properties of fluid and solid. The free surface.Chapter 1 Introduction The implementation of jet impingement due to its high convective heat transfer rate in mechanical and chemical processes has led to many industry applications. begins to form instantly at the nozzle exit and remains throughout the impingement section and wall region as the fluid moves downstream along the plate after impingement. and height from nozzle to impingement surface. drying of wetted surfaces. thickness of the solid. 1 .
00050. singlephase liquid jet impinging on a flat uniformly heated surface. and jet diameter.000. This experimental study investigated the effects of Reynolds number.1 Literature Review 1. Garimella and Rice [3] experimentally measured the local heat transfer from a small heat source to a normally impinging. Kornblum and Goldstein [6] analyzed the flow of circular jets impinging on semicylindrical surfaces (convex and concave) for relatively small jet to semicylindrical diameter ratios ranging 2 . [5] studied heat transfer from a convex surface with low curvature using liquid crystals to measure the local surface temperature at Reynolds numbers ranging from 11.1.000 and nozzle to heat source spacing ratios of 1 to 14. They concluded that the maximum heat transfer coefficient was found to decrease with increasing jet inclination. Lee et al. Gomi and Webb [4] performed an investigation on heat transfer from a vertical heated surface to an obliquely impinging circular freesurface jet of transformer oil for various Reynolds number (2351. The experimental study concluded that the stagnation point Nusselt number increases with increasing surface curvature. They concluded that secondary peaks in the local heat transfer coefficient resulted form increasing nozzle diameters for a given Reynolds number.1 Round Jet Impingement An early investigation involving liquid jet impingement using round nozzles was carried out by Stevens and Webb [2] who considered an axisymmetric.00023. axisymmetric and submerged liquid (FC77) jet from a round nozzle for a range of Reynolds number of 4.1. nozzle to plate spacing.746) with jet angles from 45° to 90°.
[12] experimentally studied the hydrodynamic and thermal characteristics of a free round liquid jet impinging into a heated disk for nozzle to plate spacing of 312 times nozzle diameter and for Reynolds number of 6009.from 0. Cornaro et al. 3 .0394 using a flow visualization technique.00087. The effects of Reynolds number. jet to surface distance. [10] showed the effect of increasing Reynolds number on the local Nusselt number for different types of curvature.01970. and relative curvature were studied. and 0.0726. and air blower) to target a convex hemispherical surface for Re=11.000 at different dimensionless nozzletosurface distance and at constant hemispheretonozzle diameter ratio to obtain information on the stagnation and local Nusselt numbers.000. Baonga et al. In addition.0986 m with Reynolds numbers ranging from 6.000. [11] used a smokewire flow visualization technique to investigate the behavior of a round jet impinging on a convex surface. [9] visualized fluid flow using a smokewire technique on concave and convex surfaces to examine the effects of curvature.0472. They showed the initiation and development of ring vortices and their interaction with the cylindrical surface. nozzle to surface distance. Cornaro et al. Fleischer et al.00015. [7] employed an apparatus consisting of various components (heat exchanger. Tong [8] numerically studied convective heat transfer of a circular liquid jet impinging onto a substrate to understand the hydrodynamics and heat transfer of the impingement process using the volumeoffluid method to track the free surface of the jet. The convex surface was examined for jet diameters of 0. flow meter. and Reynolds number. Lee et al. 0.
Also. Similar analytical solution to the momentum equation using the first order CauchyEuler ordinary differential equation and the integral equation approach to solve for the energy equation was employed by Carper [14]. He attempted to obtain a solution for the ratio of thermal to hydrodynamic boundary layer thickness for a certain range of combinations of jet Prandtl number and jet Reynolds number for an axisymmetric liquid jet impinging on a flat surface.1. Solutions to the thermal boundary layer and film thickness were approximated using the integral method to obtain information about the local Nusselt number and temperature distribution.2 Slot Jet Impingement A number of past studies have considered liquid jet impingement from a twodimensional slot nozzle to a heated flat plate. One of the early investigations on jet impingement on a flat plate was conducted by Inada et al. Wadsworth and Mudawar [16] performed an experiment to investigate single phase heat transfer from a smooth simulated chip to a twodimensional jet of dielectric liquid FC72 delivered from a very thin rectangular slot jet into a 4 . Liu and Lienhard [15] analytically and experimentally investigated convective heat transfer to an impinging liquid jet from a heated flat surface kept at a uniform heat flux. [13] who studied laminar flow between a plane surface and a twodimensional water jet with constant heat flux using the RungeKutta method to obtain solutions for the boundary layer momentum and energy equations. Also.1. the average Nusselt number for a plane jet is generally larger than that for an axisymmetric jet. They concluded that the average Nusselt number for a constant heat flux is larger than for a constant wall temperature.
Teuscher et al.channel confined between the chip surface and nozzle plate. To achieve it. [20] measured heat transfer coefficients resulting from the impingement of transformer oil jet issuing from various slot widths for fluid Prandtl numbers of 200270 and Reynolds number between 55 and 415. they used a novel laserDoppler velocimetry technique to capture the fluctuations of the free surface for Reynolds numbers ranging from 16. their research only focused in measuring the velocity of the free surface of the jet. Experimental results of convective heat transfer were obtained for various slot widths (0.00035. [19] studied jet impingement cooling of an inline array of discrete heat sources (fins of different configurations) by a submerged slot jet of FC77. The experiment studied the effects of jet width. Slot 5 .0220. Stevens and Webb [18] experimentally characterized the flow structure under an impinging liquid jet hitting a flat surface. The authors observed rotating vortices at the stagnation line on the convex surface which increased the heat transfer characteristics of the flow.000. Bartoli and Faggiani [21] experimentally studied heat transfer form a circular cylinder to a slot jet of water at different Prandtl numbers for 3500≤Re≤20000 for tests performed at different angles 0°≤ θ≤ 180° in order to obtain the local and average Nusselt number.000100.2540. Gau and Chung [17] reported heat transfer for air jet impingement on semicylindrical curved surfaces at low and high Reynolds numbers (6.000) by varying slot widths to surface diameter ratios (0. However. Ma et al. confinement channel height.000 to 47.000. and impingement velocity for Reynolds number ranging from 1.125) using smoke particles as means to visualize the air flow.508 mm) and Reynolds number of 10500.
[1] numerically studied the conjugate heat transfer during impingement of a confined liquid jet. Gori and Bossi [23] investigated heat transfer for the flow over the surface of a cylinder for Reynolds number (400020. The results of the experimental investigation concluded that for a range of Reynolds numbers from 200 to 11.000 and varying jet to cylinder distances and cylinder curvatures. [25] carried out a numerical study to examine the effects of Prandtl number (0.2 of the slot width the heat transfer rates at the surface of the concave channel are improved due to the curvature of the channel.771). Olsson et al.01100 on heat transfer under a semiconfined laminar slot jet. and plate thickness for Reynolds number ranging from 4451. By varying slot width. 1 and 2. Rahman et al.000 and slot to surface spacing from 2. and for jet exit to nozzle width spacing from 1 to 11.2 to 4. They stated that the flow characteristics and the heat 6 . impingement height. Kayansayan and Küçüka [24] performed an experimental and numerical study of jet impingement cooling on a concave channel.jet impingement involving heat transfer from circular cylinders using air as the cooling fluid was experimentally studied by McDaniel and Webb [22] for Reynolds numbers in the range 6008000 varying cylinder diameter to jet width spacings of 0.66. [26] simulated heat transfer from a slot air jet impinging on a cylinder at various Reynolds numbers ranging from 23.000100. nozzle to target spacing (210) and Reynolds number ranging from 0. Shi et al.000) and determined that the distance between the exit of the slot jet and the cylinder surface affects the local Nusselt number.545.
transfer distribution around cylinders are found to be dependent on the distance and the opening between the jets. Chen et al.000) of the inlet flow and by varying slot jet width to jet to surface distance ratio. [30] investigated an underexpanded jet impinging on a heated cylindrical surface varying the nozzle to surface spacing (310) for various Reynolds numbers. and slot nozzle width to circumferential distance on the local heat transfer for an air impinging slot jet on a semicircular convex surface. These models 7 .00020. [28] measured the mean flow and turbulence of a turbulent air slot jet impinging on two different semicircular convex surfaces at Reynolds number 12. [27] employed a liquid crystal thermographic system to experimentally determine the effects of jet Reynolds number. Chan et al. Gori and Bossi [29] experimentally determined the optimal height in the jet cooling of an electrically heated circular cylinder for various Reynolds numbers measuring mean and local Nusselt numbers. [32] performed a theoretical analysis to characterize heat transfer from horizontal surfaces to single phase free surface laminar slot jets using a heat flux condition for different working fluids and different nozzle sizes. Yang and Hwang [31] carried out a numerical simulation of flow of a turbulent slot jet impinging on a semicylindrical convex surface for Reynolds number (6. dimensionless slot nozzle width to impingement surface distance ratio. Chan et al.000 using a hotwire Xprobe anemometer. Rahimi et al. Zuckerman and Lior [33] employed numerical models to understand the heat transfer behavior on circular cylinders cooled by radial slot jets.
Therefore. and ammonia (NH3) as working fluids were carried out for different flow configurations. flouroinert (FC77). convex. a through study of liquid jet impingement on a curved surface is needed as this is encountered in many industrial processes.000 and different target diameter to nozzle hydraulic diameter. or flat surfaces taking into account the transient nature of the problem. Computations using water (H2O). A few attempted to obtain local heat transfer distribution of concave. software package focused in computational fluid dynamics (CFD) of 8 . 1. plate configurations. They concluded that the highest average Nusselt number would occur when having a lower number of jets. 1. The computations were carried out using the finite element method using FIDAP version 8. oil (MIL7808).000 to 80.2 Overview of Literature The literature mentioned above describes most studies to date on jet impingement over a curved or flat surface using air or liquid as the working fluid and most of them have considered the steady state condition only.6.attempted to simulate a cylinder exposed to a radial array of slot jets (28) for Reynolds number ranged from 5.3 Thesis Aim The present study attempts to carry out a comprehensive numerical investigation of steady and transient local conjugate heat transfer for laminar free surface jet impingement over a hollow hemisphere and a curved cylindrical shaped plate. and different plate materials.
the FLUENT corporation. That is the reason why the FEM was chosen as the preferred method to solve for the mathematical model presented in this study. or rendering the PDE into an equivalent ordinary differential equation. Also. 9 . the finite difference method is an approximation to the differential equation whereas the finite element method is an approximation to its solution. as appropriate. This approach differs from the finite element method (FEM) such that it uses finite differences schemes (forward. The system of equations is solved for unknown values using the techniques of linear algebra or nonlinear numerical schemes. which is then. being the central difference the one that provides the more accurate approximation. The solution method uses a mesh domain (elements grid) to solve the differential equation completely by applying to each element a system of simultaneous equations. The finite element method (FEM) is used for finding approximate solution of complex partial differential equations (PDE) as well as of integral equations. and central) to approximate derivatives. One of the most attractive features of the FEM over the FDM is its ability to handle complex geometries and boundaries (moving boundaries) with relative ease. solved using finite difference. backward.
1 Hemispherical Model The physical hemispherical model studied here is threedimensional (3D) by nature as depicted in Figure 2.1. The fluid is Newtonian and the flow is incompressible and axisymmetric.2 shows a 2D crosssectional view of the system including the origin and axes used to write the boundary conditions. However. the free jet discharges from the round nozzle and impinges perpendicularly at the top center of the hemisphere while the hemisphere is dissipating heat from within. the problem can be greatly simplified into an axisymmetric liquid jet model that impinges on the outer surface of a hollow hemisphere subjected to an isothermal or constant heat flux boundary condition at its inner surface.Chapter 2 Mathematical Models and Computation 2. 10 . Figure 2. Thus.
Figure 2.2 CrossSectional View of a Hollow Hemisphere During an Axisymetric Jet Impingement 11 .1 Three Dimensional Schematic of a Hollow Hemisphere During an Axisymmetric Liquid Jet Impingement Figure 2.
z = ro : p = patm (2.1. and energy can be written as [34]: r ∇⋅V = 0 (2. 0 ≤ z ≤ Hn : Vr = 0. − b ≤ z ≤ 0 : ∂Ts =0 ∂r (2.2.3) r (V ⋅ ∇ )T f = α f ∇ 2 Tf (2. f = 0 ∂r ∂r (2. z = ro : ∂TS =0 ∂z (2.1.1. ∂VZ ∂T = 0.1 Governing Equations: Steady State Heating The equations describing the conservation of mass.4) The conservation of energy inside the solid can be characterized by the following equation: ∇ 2 Ts = 0 (2.1.2 Boundary Conditions: Steady State Heating The following boundary conditions were used: At r = 0.1.1. momentum (r and z directions respectively).10) At z = Hn . Tf = TJ 2 12 .1.9) (2.8) At ro ≤ r ≤ (ro + δ).1) (2.1.1.6) At r = 0.1. 0 ≤ r ≤ dn : VZ = −VJ .1.2) r r (V ⋅ ∇)Vr = − ρ1 ∂p + ν∇2Vr − V2 ∂r f r r (V ⋅ ∇)V = −g − ρ1 ∂p + ν∇ V ∂z 2 z f z (2. Vr = 0.5) 2.1.7) At ri ≤ r ≤ ro .
Nu = h ⋅ dn kf 13 (2. Similarly. The kinematic condition related velocity components to local slope of the free surface.15) . The normal stress balance took into account the effects of surface tension. the shear stress encountered at the free surface is essentially zero.14) where T int is the average temperature at the solidliquid interface. a negligible heat transfer at the free surface results in zero temperature gradient.1. ks S = k f f S f ∂n 2 ∂n π dn ≤ Φ ≤ 2 2 (2.T = T .1. In the absence of any significant resistance from the ambient gas.12) The boundary condition at the free surface included the kinematic condition and balance of normal and shear stresses. The local and average Nusselt numbers are calculated according to the following expressions.At inner surface of hemisphere: 0≤Φ≤ ∂T π : −k s S = q or 2 ∂r T = To (2. The local and average heat transfer coefficients can be defined as: h= q (Tint − TJ ) π (2.1.1.13) h av = (T 1 int − TJ ) 0 ∫ h(T 2 int − TJ ) ⋅ sin ΦdΦ (2.11) At outer surface of hemisphere: 0≤Φ≤ ∂T ∂T π : Vr = VZ = 0.1.
momentum (r and z directions respectively).1.20) The conservation of energy inside the solid can be characterized by the following heat conduction equation: α∇ 2Ts = ∂Ts ∂t (2.1. Therefore.18) r (V ⋅ ∇)Vr = − ρ1 ∂p + ν ∇2Vr − V2r ∂r f r r (V ⋅ ∇)V z = −g − 1 ∂p + ν∇ 2 Vz ρ f ∂z (2. the equations describing the conservation of mass. and energy [34]: r ∇⋅V = 0 (2.17) (2. At t=0 the power source is turned on and heat begins to flow only after an initially isothermal fluid flow has been established on the hemisphere.1.19) r ∂Tf + V ⋅ ∇ Tf = α f ∇ 2 Tf ∂t ( ) (2.3 Governing Equations: Transient Heating A liquid jet axially discharging from a round nozzle impinges on the outer surface of a hollow hemisphere subjected to a constant heat flux boundary condition at its inner surface.21) 14 .1.1.1.1.16) 2.Nuav = h av ⋅ dn kf (2. The fluid is Newtonian and the flow is incompressible and axisymmetric.
1. Tf = TJ 2 (2. z = ro : p = patm At z = Hn .29) 15 .27) At outer surface of hemisphere: 0≤Φ≤ ∂T ∂T π : Vr = VZ = 0. 0 ≤ r ≤ dn : VZ = −VJ . − b ≤ z ≤ 0 : ∂Ts =0 ∂r (2. z = ro : ∂TS =0 ∂z (2.2.T = T .1.1.24) At ro ≤ r ≤ (ro + δ). The velocity field at this condition was determined by solving only the continuity and momentum equations in the fluid region.ks S = k f f ∂n f S 2 ∂n (2. At t = 0: Tf = Ts = Tj .22) At r = 0.1.1. Vr = 0. Thus. ∂VZ ∂T = 0.28) The plate was assumed to be at thermal equilibrium with jet fluid before the transient heating of the plate was turned on. Vi = Vr. 0 ≤ z ≤ Hn : Vr = 0. z (2.26) At inner surface of hemisphere: 0≤Φ≤ ∂T π : −k s S = q .25) (2. f = 0 ∂r ∂r (2.1.1.23) At ri ≤ r ≤ ro . t ≥ 0 2 ∂r (2.1.1.4 Boundary Conditions: Transient Heating The following boundary conditions were used: At r = 0.
Figure 2. The fluid is Newtonian and the flow is incompressible and symmetric about the midplane.2 Cylindrical Plate Model The physical model corresponds to a twodimensional symmetric liquid jet that impinges on the outer surface of a curved hollow cylindrical shaped plate subjected to a uniform heat flux boundary condition at the inner surface as shown in Figure 2.2. Figure 2.3 also shows the origin and axes used to write the boundary conditions.3.3 Schematic View of a Curved Plate During a 2D Symmetric Liquid Jet Impingement 16 .
2 Boundary Conditions: Steady State Heating The following boundary conditions were used: At x = 0.2.7) ∂ Ts =0 ∂y (2.9) 17 .1 Governing Equations: Steady State Heating The equations describing the conservation of mass.2.6) At x = 0. At R i ≤ x ≤ R o . momentum (x and y directions respectively).2. ∂Tf =0 ∂x (2.2.2. y = − x o ∂Vy ∂x : = 0.2.2.3) (2.1) (2.2.2. and energy using a 2D coordinate system can be written as [34]: r ∇⋅ V = 0 r (V ⋅ ∇)⋅ Vy = −gy − ρ1 ∂p + µ∇2V ∂y f y (2. . y = x o : p = patm (2.2.8) At R o ≤ x ≤ (R o + δ ).2.2) r (V ⋅ ∇)⋅ Vx = − ρ1 ∂p + µ∇2Vx ∂x f r (V ⋅ ∇ )⋅ T f = α ∇ 2 Tf f (2. 0 ≤ y ≤ Hn : Vx = 0.5) 2.2.b ≤ y ≤ 0 : ∂ Ts =0 ∂x (2.4) The conservation of energy inside the solid can be characterized by the following equation: ∇ 2 Ts = 0 (2.
The kinematic condition related velocity components to local slope of the free surface.2.2. The normal stress balance took into account the effects of surface tension. Vx = 0. Ts = T .2. In the absence of any significant resistance from the ambient gas. h= q (Tint − TJ ) (2.14) where T int is the average temperature at the solidliquid interface.16) .At y = Hn .2.15) Nu av = (2. Tf = TJ 2 (2. 0 ≤ x ≤ At inner surface of plate: dn : Vy = −VJ .12) The boundary condition at the free surface included the kinematic condition and balance of normal and shear stresses. Similarly.11) (2. the shear stress encountered at the free surface is essentially zero.13) φmax h av = π T int − TJ ) ( 4 ∫ h(T 0 int − TJ ) ⋅ sinφdφ (2.2. The local and average Nusselt numbers are calculated according to the following expressions Nu = h ⋅ dn kf h av ⋅ dn kf 18 (2.10) ∂T 0 ≤ s ≤ Riφmax: −ks s = q ∂n At outer surface of plate: ∂T ∂T 0 ≤ s ≤ Roφmax: Vx = Vy = 0.2.2. ks s = k f f f ∂n ∂n (2. a negligible heat transfer at the free surface results in zero temperature gradient.
. The fluid is Newtonian and the flow is incompressible and symmetric about the midplane.2. and energy in the fluid region can be written as [34]: r ∇⋅ V = 0 r (V ⋅ ∇)⋅ Vy = −gy − ρ1 ∂p + µ∇2V ∂y f y (2.17) (2.3 Governing Equations: Transient Heating The physical model corresponds to a twodimensional symmetric liquid jet that impinges on the outer surface of a curved hollow cylindrical shaped plate subjected to a uniform heat flux boundary condition at the inner surface.2.22) .b ≤ y ≤ 0 : ∂ Ts =0 ∂x 19 (2. The equations describing the conservation of mass.18) r (V ⋅ ∇)⋅ Vx = − ρ1 ∂p + µ∇2Vx ∂x f (2.2.2.19) ∂T r f + V ⋅ ∇ ⋅ T = α ∇ 2T f f f ∂t ( ) (2.2.2.20) The conservation of energy inside the solid can be characterized by the following equation: α s ∇ 2 Ts = ∂T s ∂t (2.21) 2.2. At t=0 the power source is turned on and heat begins to flow only after an initially isothermal fluid flow has been established on the plate.2.2.4 Boundary Conditions: Transient Heating The following boundary conditions were used: At x = 0. momentum.
At t = 0: Tf = Ts = Tj .2. y = x o : p = patm (2. At R i ≤ x ≤ R o . Ts = T .27) (2. y (2.25) (2.2.29) 20 . Vi = Vx.2. Thus.26) At y = Hn . y = − x o ∂Vy ∂x : = 0. t ≥ 0 ∂n At outer surface of plate: ∂T ∂T 0 ≤ s ≤ Roφmax: Vx = Vy = 0.2.2. 0 ≤ y ≤ Hn : Vx = 0. ks s = k f f f ∂n ∂n (2.24) At R o ≤ x ≤ (R o + δ ). The velocity field at this condition was determined by solving only the continuity and momentum equations in the fluid region. Tf = TJ 2 ∂T 0 ≤ s ≤ Riφmax: −ks s = q. Vx = 0.2.At x = 0. ∂Tf =0 ∂x (2.2. 0 ≤ x ≤ At inner surface of plate: dn : Vy = −VJ .28) The plate was assumed to be at thermal equilibrium with jet fluid before the transient heating of the plate was turned on.23) ∂ Ts =0 ∂y (2.
Four node quadrilateral elements were used as shown in Figure 2.3 Numerical Computation 2. the velocity. The method to solve the set of the resulting nonlinear equations was the NewtonRaphson algorithm due to its 21 . Figure 2. pressure.4.1 Steady State Process The governing equations in conjunction with the boundary conditions described above were solved using the Galerkin finite element method employed by FIDAP. and temperature fields were approximated until convergence was achieved.2.4 Mesh Plot of a Curved Plate During a 2D Symmetric Liquid Jet Impingement For every element.3.
The solution was considered converged when relative change in field values from a particular iteration to the next. and the sums of the residuals for each variable became less than 106. Since the solution of the momentum equation required only two out of the three boundary conditions at the free surface. an iterative procedure was used to converge at the solution for the velocity and temperature fields. The height of the free surface was adjusted after each iteration by satisfying the kinematic condition that relates the slope of the free surface to local velocity components at the free surface. 22 . The NewtonRaphson solver employed spines to track the free surface and rearranged grid distribution with the movement along the free surface. The movement of the free surface affected only the nodes along the spine.coupled nature for two dimensional problems and its ability to solve all conservation equations in a simultaneous manner. The approach used to solve the free surface problem described here was to introduce a new degree of freedom representing the position of the free surface. A scaled dense grid distribution was used to adequately capture large variations near the solidfluid interface of the meshed domain. Due to nonlinear nature of the governing transport equations. This degree of freedom was introduced as a new unknown into the global system of equations. The spines are straight lines passing through the free surface nodes and connect the remaining nodes underneath the free surface. the third condition was used to upgrade the position of the free surface at the end of each iteration step.
2 Transient Process In order to determine the initial velocity field (Vi). because of large changes at the outset of the transient and very small changes when the solution approached the steadystate condition. The solution was considered converged when relative change in field values from a particular iteration to the next. 23 . and energy were solved simultaneously as a conjugate problem. the power of the heat source was turned on and heat began to flow. The computation domain included both solid and fluid regions. Once the initial free surface height distribution and the flowfield for the isothermal equilibrium condition were satisfied. and continuity. momentum. and the sums of the residuals for each variable became less than 106. the equations for the conservation of mass and momentum were solved. The computation continued towards the steady state condition. however. a variable time step was used for the computation.3. The height of the free surface was adjusted after each iteration by satisfying the kinematic condition that relates the slope of the free surface to local velocity components at the free surface.2.
1) C. The numerical solution becomes grid independent when the number of divisions equal to 40x119 in the axial (z) (radial in thin film after impingement) and arc (Φ) directions respectively is used. the value of e was calculated to be 8.2. D. A quantitative difference in grid independence was calculated using the following expression: D Tint = C + e N (2.4.1) has three unknowns at three sets of interface temperatures taken at three different grid sizes.5.3.1 Cylindrical Coordinates Several grids or combinations of number of elements were used to determine the accuracy of the numerical solution as shown in Figure 2.3. it was substituted into the system of equations and the values of C and D were calculated to be 318.72205E19 respectively.4 Mesh Independence and Time Step Study 2.3. In addition.9112 and 0. equation (2.001276 m. and e are constants to be evaluated.554. Numerical results for a 40x119 grid gave almost identical results compared to those using 34x79 and 40x87 grids. This results in a set of nonlinear equations with three variables.2) 24 . At s or Φ=0. the percentage error is calculated used the following expression: Tint − C x100 C (2. N represents the number of divisions along a chosen axis. The value of e was obtained based on doing a particular number of iterations. Tint is the solidfluid interface temperature at a given Φlocation of the hemispherical plate. Thus. Once the value of e had been obtained.
However.5 nzxnΦ = 22x57 nzxnΦ = 26x57 nzxnΦ = 30x79 nzxnΦ = 34x79 nzxnΦ = 40x119 nzxnΦ = 40x87 Dimensionless Distance.5 6 6.987% for 40x119 grids and 4.5 Local Dimensionless Interface Temperature for Different Number of Elements in z (or r) and Φ Directions for Water as Fluid and Silicon as Solid (Re=750. For this numerical study.The error at Φ=0.5 1 1. Notice how the maximum dimensionless 25 . a very small time increment (0. s/dn Figure 2.01 s was selected in order to accomplish a smooth variation. b/dn=0.5 7 7.5 2 2.5 3 3.001276 m was found to be 3.5 5 5.00001 s) ensures the initial condition to be revealed. q=250 kW/m2) Figure 2.26% for 48x87 grids. an appropriate time increment of 0. 355 SolidFluid Interface Temperature.5. Observe that the simulation is not very susceptible to the time increments chosen. K 350 345 340 335 330 325 320 0 0. ß=2.6 plots the transient variation of the maximum dimensionless temperature encountered at the hemisphere’s outer surface (solidfluid interface) for different time increments as a function of time using the Fourier number (Fo) as a dimensionless number to represent time.0.5 4 4.
7.0 0.04 0.Θ max 26 .30 cm.5.1 s Time Increment = 0. 0.14 0.02 0 0.00001 s Time Increment = 0.1 Cartesian Coordinates Several grids or combinations of number of elements were used to determine the accuracy of the numerical solution as shown in Figure 2.16 0.05 s Time Increment = 0. q=250 kW/m2) 2.01 s Figure 2.12 0.1 0.5 0.2 s Time Increment = 0.08 s Time Increment = 0. The numerical solution becomes grid independent when the number of divisions equal to 30x132 in the y and x or φ (in thin film after impingement) directions is Dimensionless Maximum Temperature at the Interface.06 0.4.0.4 0. b/dn=0.08 Time Increment = 0.6 Fourier Number.temperature increases rapidly as the time increases all the way to the steadystate condition.6 SolidFluid Interface Dimensionless Maximum Temperature Variation for Silicon Hemisphere at Different Time Increments (Re=750. ß=2.03 s 0.1 0.3 0. Hn = 0. Fo Time Increment = 0.2 0.
0 3. Numerical results for a 30x132 grid gave almost identical results compared to those using 32x132 and 26x92 grids.0 0.554. Once the value of e had been obtained.63% for 26x92 grids at φ=0. The error percentage is calculated from equation (2.001277 m.30 cm) SolidFluid Interface Temperature.3. At s or φ=0.5.0 Dimensionless Distance. a quantitative difference in grid independence was carried on using equation (2.5 3.0466%. K 27 .7070E19 respectively.5 1. Hn = 0. s/dn nyxnx or nΦ=10x32 nyxnx or nΦ=16x32 nyxnx or nΦ=20x92 nyxnx or nΦ=26x92 nyxnx or nΦ=30x132 nyxnx or nΦ=32x132 Figure 2. it was substituted into the system of equations and the values of C and D were calculated to be 338. Also.5 4.51% for 32x132 grids and 1. ß=2. and y Directions (Re=750. 342 340 338 336 334 332 330 328 326 0. The average difference was 0.7 Local Dimensionless Interface Temperature for Different Number of Elements in x or φ.2391 and 0.0 1. the value of e was calculated to be 8.used.2) and found to be 1.1).001277 m.0 2.5 2.3.
a time increment of 0.2 0.1 Time Increment = 0.1 s Time Increment = 0.03 s Time Increment = 0.3 Figure 2.8 shows the transient variation of the maximum dimensionless temperature found at the plate’s outer surface (solidfluid interface) for different time increments as a function of time using the Fourier number (Fo) as a dimensionless number to represent time.08 0. Dimensionless Maximum Temperature at the Interface.05 0.Figure 2.14 0.08 s Time 0.30 cm.01 s Time Increment = 0. For this study.05 s Chapter 3 Increment = 0.5. Note that the numerical simulation is not very susceptible to the time increments chosen. q=250 kW/m2) 28 .02 0 0 0.2 s 0.00001 s Time Increment = 0.04 Time Increment = 0.8 SolidFluid Interface Dimensionless Maximum Temperature Variation for Silicon Plate at Different Time Increments (Re=750. ß=2.06 0.15 Fourier Number.01 s was chosen in order to obtain a smooth variation for temperature. Hn = 0.25 0.1 0. Θ int 0.5. Notice how the maximum dimensionless temperature at the transient increases rapidly all the way to the steadystate condition.16 0. Fo 0.12 0. b/dn=0.
1 Steady State Heating A typical velocity vector distribution is shown in Figure 3. That is the beginning of the boundary layer zone. the boundary layer thickness increases and the frictional resistance from the wall is eventually transmitted to the entire film thickness where the fully viscous zone develops. The direction of motion of the fluid particles shifts along the angle of curvature. the fluid accelerates creating a region of minimum sheet thickness. However.Chapter 3 Hemispherical Model Results 3. the velocity decreases and the fluid jet diameter increases as the fluid gets closer to the surface during the impingement process. It can be observed that the velocity remains almost uniform at the potential core region of the jet. 29 . afterwards. It can be noticed that as the fluid moves downstream along the convex surface.1.
1 Velocity Vector Distribution for Silicon Hemisphere with Water as the Cooling Fluid (Re = 750.0.2. As it is seen. A pressure gradient distribution is more uniform. however.665x107 m3/s. 30 . β = 2. Hn = 0. q = 250 kW/m2) A pressure distribution plot is shown in Figure 3.60 mm. the pressure at the impingement regions is higher due to the fluid impacting the hemisphere outer surface in comparison with the remaining portion of the hemisphere. Q = 5. along the arc length of the solid.30 cm.Figure 3. b = 0.
0. increasing the velocity boundary layer and temperature boundary layer thickness.30 cm.Figure 3. Hn = 0.60 mm. hence. Higher temperatures are encountered along the end of the arc length of the solidfluid interface whereas lower temperatures are encountered at the stagnation region. After impact has occurred.3 depicts temperature distribution along the solid and fluid region.2 Pressure Distribution for Silicon Hemisphere with Water as the Cooling Fluid (Re = 750. Q = 5. 31 . the particles start to accelerate towards the remaining portion of the hemisphere. b = 0.665x107 m3/s. q = 250 kW/m2) Figure 3. β = 2.
These observations concur with the experimental work of Stevens and Webb [35] for free jet impingement on a flat surface.4 shows the free surface height distribution for different Reynolds numbers when the jet strikes the hemisphere’s surface. Q = 5. As the Reynolds number increases. For the conditions considered in the present investigation. β = 2.60 mm. b = 0.Figure 3.30 cm. the film diminishes in thickness due to a larger impingement velocity that translates to a higher fluid velocity in the film. Hn = 0. It was observed that the velocity remains 32 .0.3 Temperature Distribution for Silicon Hemisphere with Water as the Cooling Fluid (Re = 750. the flow was supercritical and a hydraulic jump did not occur within the computation domain. q = 250 kW/m2) Figure 3.665x107 m3/s. It can be seen that the fluid spreads out moving along the outer surface of the hemisphere.
06 0.32 0.46 0. The direction of motion of the fluid particles shifted along the angle of curvature.48 0. β=2.3 0. However.18 0. afterwards. That was the beginning of the boundary layer zone.4 Free Surface Height Distribution for Different Reynolds Numbers and Water as the Cooling Fluid (b=0. 0.6 Re=500 Re=750 Re=1000 Re=1500 Re=2000 Axial Distance (cm) Radial Distance (cm) Figure 3.almost uniform at the potential core region of the jet. the velocity decreases and the fluid jet diameter increases as the fluid get closer to the surface during the impingement process.4 0.54 0. q=250 kW/m2) 33 .08 0.0. the fluid accelerated creating a region of minimum sheet thickness.04 0.24 0.28 0.34 0.26 0. It was noticed that as the fluid moves downstream along the convex surface.12 0.52 0 0.60 mm. the boundary layer thickness increases and the frictional resistance from the wall is eventually transmitted to the entire film thickness where the fully viscous zone developed.22 0.1 0.2 0.36 0. Hn=0.02 0.30 cm.14 0.16 0. The three different regions observed in the present investigation agreed with the experiments of Liu et al.42 0. [36] for jet impingement over a flat surface.
Figures 3. The curves in Figure 3. 34 .5 and 3.5 reveal that the dimensionless interface temperature decreases with jet velocity or (Reynolds number). reaching a maximum around s/dn = 0.5 and 3. The local Nusselt number depicted in Figure 3. The dimensionless interface temperature has the lowest value at s/dn = 0.6 show dimensionless interface temperature and local Nusselt number distributions as a function of dimensionless distance (s/dn) along the solidfluid interface at different Reynolds numbers under an isothermal heating condition.5.4 and increases along the arc length (s) reaching the highest value at the end of it.6 confirm how an increasing Reynolds number contributes to a more effective cooling by the enhancement of convective heat transfer coefficient. The location of the maximum Nusselt number can be associated with the transition of the flow from the vertical impingement to horizontal displacement where the boundary layer starts to develop. and then decreases along the radial distance as the boundary layer develops further downstream. Figures 3.6 increases rapidly over a small distance (core region) measured from the stagnation point.
1.90 0.5 7 7.5 4 4.60 0 0.Nu 90 75 60 45 30 15 0 0 0.5 1 1.6 Local Nusselt Number Distribution for a Silicon Hemisphere at Different Reynolds Numbers.30 cm.60 mm.5 Dimensionless Distance.5 5 5.5 3 3.5 Dimensionless Interface Temperature Distribution for a Silicon Hemisphere at Different Reynolds Numbers. and Water as the Cooling Fluid (β=2.5 4 4.0.5 3 3.5 7 7. To=373 K) 35 .5 6 6. s/dn Re=500 Re=750 Re=1000 Re=1500 Re=2000 Figure 3.5 6 6.5 2 2. and Water as the Cooling Fluid (β=2.5 5 5.80 Re=500 0.5 2 2.5 Dimensionless Distance.00 Dimensionless Interface Temperature. b=0.70 Re=750 Re=1000 Re=1500 Re=2000 0. To=373 K) 135 120 105 Local Nusselt Number.60 mm.5 1 1. b=0. Hn= 0. s/dn Figure 3. Ω int 0.30 cm. Hn=0.0.
The dimensionless interface temperature has the lowest value at the stagnation point (underneath the center of the axial opening) and increases along the arc length (s) reaching the highest value at the end of it. Hn= 0. and Water as the Cooling Fluid (β=2. s/dn Figure 3.5 Dimensionless Distance.165 Re=1500 Re=2000 0. b=0.Figures 3.5 1 1. The curves in Figure 3. 0.60 mm.30 cm.7 and 3.7 Dimensionless Interface Temperature Distribution for a Silicon Hemispherical Plate at Different Reynolds Numbers.5 6 6.055 0 0 0.5 2 2.Θint Re=750 Re=1000 0.5 4 4.0.5 3 3.5 5 5.7 reveal that the dimensionless interface temperature decreases with jet velocity or (Reynolds number). q=250 kW/m2) 36 .8 show dimensionless interface temperature and local Nusselt number distributions as a function of dimensionless distance (s/dn) along the solidfluid interface at different Reynolds numbers under a constant heat flux condition.22 Re=500 Dimensionless Interface Temperature.5 7 7.11 0.
S/dn Figure 3.5 1 1.The local Nusselt number shows a very similar kind of distribution as the case where the bottom surface of the hemispherical plate is kept at an isothermal condition.5 6 6. q=250 kW/m2) Figure 3.5 5 5. It may be noted that average Nusselt number increases with Reynolds number. It may be also noted 37 .9 plots the average Nusselt number and average heat transfer coefficient as a function of Reynolds number.5 Dimensionless Distance.0 . the magnitude of fluid velocity near the solidfluid interface that controls the convective heat transfer rate increases. and Water as the Cooling Fluid (β =2. As the flow rate (or Reynolds number) increases. b=0.8 Local Nusselt Number Distribution for a Silicon plate at Different Reynolds Numbers.5 4 4. 150 Re=500 Re=750 125 Re=1000 Re=1500 100 Re=2000 Local Nusselt Number.5 3 3.30 cm. Hn = 0.60 mm.5 2 2. This also shows that an increasing Reynolds number contributes to a more effective cooling by the enhancement of convective heat transfer coefficient.Nu 75 50 25 0 0 0.5 7 7. The values are however different.
10 and 3.30cm) The solidfluid dimensionless interface temperature and local Nusselt number distributions for six different nozzletotarget spacing for water as the coolant and Reynolds number of 500 are shown in Figures 3. hav (W/m2 K) Average Nusselt Number.that effects of boundary condition is more at lower Reynolds number and the curves come closer as the Reynolds number increases.0. Nuav . Re Figure 3. Hn = 0. 38 Average Convective Heat Transfer Coefficient. 30 25 20 25000 15 20000 10 5 0 400 600 800 1000 1200 1400 1600 1800 Isothermal Nu number Constant heat flux Nu number Isothermal h Constant heat flux h 35000 30000 15000 10000 2000 Reynolds Number.11 respectively.9 Average Nusselt Number and Heat Transfer Coefficient Variation for Different Reynolds Numbers for Constant Heat Flux (q=250kW/m2) and Isothermal (To = 373 K) Boundary Conditions (β=2.
60 mm.100 0.5 2 2. s/dn Figure 3.085 0. dn=1.2mm.Θint 0.5 β=3.0. Q=3.5 5 5.175 Dimensionless Interface Temperature.0 6. b=0.0 β=1.5 Dimensionless Distance.130 0.115 0.75 β=1.0 β=2.070 0 0.160 0.5 β=2.10 Dimensionless Interface Temperature for Silicon Hemisphere at Different Nozzle to Target Spacing Ratio (β) for Water as the Cooling Fluid (Re=500.776x107m3/s. 39 .5 1 1.145 0. q=250 kW/m2) It may be noticed that the impingement height quite significantly affects the dimensionless interface temperature as well as the Nusselt number only at the stagnation region and the early part of the boundary layer region.5 4 4.5 6 β=0.5 7 7.5 3 3.
5 β=2.11 Local Nusselt Number for Silicon Hemisphere at Different Nozzle to Target Spacing Ratio (β) for Water as the Cooling Fluid (Re=500. It is quite expected since the impingement height essentially controls the change in velocity the fluid particles encounter during the free fall from nozzle exit to target hemisphere’s surface and therefore affects areas controlled by direct impingement.5 3 3. It can be observed that there is a larger variation of interface temperature at a smaller thickness.80 70 Local Nusselt Number. the temperature at the solidfluid interface becomes more uniform due to higher distribution of heat within the solid by conduction. Q=3. s/dn Figure 3.5 5 5.0 β=2.12 shows dimensionless interface temperature for a constant q.776x107 m3/s. 40 .5 β=3. Figure 3. b=0. q=250kW/m2) At larger arc length the values get closer for all impingement heights. As the thickness increases. dn=1.5 2 2.0 β=1.5 Dimensionless Distance.75 β=1.2mm.5 6 β=0.60mm.Nu 60 50 40 30 20 10 0 0 0.0 6.5 4 4.5 7 7.5 1 1.
5 7 7.5 3 3.08 0.2mm. 0. the outer radius of the hemisphere was kept constant ( to preserve same β ) while the inner radius was varied to get different thicknesses.5 Dimensionless Distance s/dn b/dn=0. b=0. Since the heat flux imposed in the inner surface of the hemisphere was also kept constant.5 b/dn=1.5 2 2.It may be also noticed that the average temperature at the solidfluid interface decreases when the thickness of the solid hemisphere increases. a smaller inner radius resulted in smaller heat input rate at that boundary.2 Dimensionless Interface Temperature. In addition.04 0.665x107m3/s.60 mm.18 0.14 0. Θint 0.5 6 6.25 b/dn=0. This decrease in temperature is due to lower heat load at increased thickness.02 0 0 0.0 b/dn=1. Q=5.5 1 1. q=250 kW/m2) 41 .06 0.5 Figure 3.1 0. the resistance of the material to the path of heat flow increases with thickness. dn=1.083 b/dn=0.12 Dimensionless Interface Temperature for Silicon Hemisphere at Different Thickness (b) for Water as the Cooling Fluid (Re=750.5 5 5.16 0.12 0.5 4 4. In the present investigation.
5 5 5. Nusselt number changes by only a small amount over the thicknesses considered in the present investigation when silicon is used as the solid material.13 plots the distribution of local Nusselt number along the surface of the hemisphere for different values of wall thickness. s/dn Figure 3. A larger peak Nusselt number is obtained at a smaller thickness.5 1 1.60mm. Nu 40 b=0.Figure 3. A larger thickness offers a larger 42 .5 Dimensionless Distance. Figure 3. 60 50 Local Nusselt Number.13 Local Nusselt Number for Silicon Hemisphere at Different Thickness (b) for Water as the Cooling Fluid (Re=750.5 7 7.18 30 20 10 0 0 0. It can be seen that increasing the thickness contributes to a lower interface temperature at the outer surface of the hemispherical plate.12 b=0.5 4 4.5 3 3.01 b=0. dn=1.06 b=0.665x107 m3/s. q=250kW/m2) A somewhat similar numerical results were obtained for the isothermal boundary condition of the hemisphere at different thicknesses.2mm.03 b=0. Q=5.5 6 6. b=0.5 2 2.14 shows the dimensionless interface temperature for the isothermal case.
5 4 4.06 b=0.2mm.18 7 7.5 6 6.5 Figure 3. After that.03 b=0. To=373 K) 43 .00 Dimensionless Interface Temperature. dn=1. the variation of interface temperature is more for a thicker plate.90 0.5 1 1.14 Dimensionless Interface Temperature for Silicon Hemisphere at Different Thickness (b) for Water as the Cooling Fluid (Re=750.665x107m3/s.9 as the plate became thinner.01 b=0. Unlike the constant flux case.70 0 0.5 Dimensionless Distance s/dn b=0. Ω int 0.12 b=0.15) along the plate’s interface increased up to s/dn=0. Q=5. b=0.80 0. The local Nusselt number (Figure 3.5 2 2.5 3 3.5 5 5.60 mm. Nusselt number became almost independent of thickness variation and continued decaying along the hemisphere’s outer surface. 1. It may be also noticed that temperature at the stagnation region is very significantly affected by plate thickness.thermal resistance between inner and outer surface of the solid and therefore results in larger temperature drop.
The highest Nusselt (Figure 3.665x107 m3/s. b=0.5 4 4.12 b=0.15 K).2mm. It may be noticed that water (TJ=310 K) presents the lowest interface temperature when compared with FC77 (TJ=273 K) and MIL7808 (TJ=371 K).5 3 3.5 7 7. 44 . To=373 K) Figure 3. Nu 40 b=0.5 2 2.5 1 1. dn=1. These results were obtained for a constant Reynolds number of 750.60 50 Local Nusselt Number.15 Local Nusselt Number for Silicon Hemisphere at Different Thickness (b) for Water as the Cooling Fluid (Re=750.60mm.17) number is obtained when FC77 is used as the working fluid (Tj=273. s/dn Figure 3.5 6 6.5 Dimensionless Distance.16 compares the solidfluid interface temperature for the isothermal boundary condition for the present working fluid (water) with two other coolants that have been considered in previous thermal management studies.06 b=0. namely flouroinert (FC77) and oil (MIL7808).01 b=0. Q=5.18 30 20 10 0 0 0.03 b=0.5 5 5.
1. s/dn H20(Pr=5.5 2 2.94 0.MIL7808(Pr=124. Nu Nu.44) Figure 3.5 Dimensionless Distance.5 4 4.60 mm.5 3 3.17 Local Nusselt Number for Silicon Hemisphere for Different Cooling Fluids (Re=750.5 5 5. dn=1.5 6 6.60mm.88 0 0.5 7 7. b=0.00 0.5 6 6. H20(Pr=5.80) 100 Local Nusselt Number.16 Dimensionless Interface Temperature for Silicon Hemispherical Plate for Different Cooling Fluids (Re=750.5 5 5.665x107m3/s. To=373 K) 45 .49) FC77(Pr=31.2mm.665x107 m3/s.02 Dimensionless Interface Temperature.5 Dimensionless Distance.2mm. Q=5.90 0. To=373 K) 120 Nu.49) Nu.5 1 1.FC77(Pr=31.5 4 4.96 0. b=0. Q=5.5 7 7.80) MIL7808(Pr=124. dn=1. s/dn Figure 3.5 2 2.92 0.5 1 1.98 0.44) 80 60 40 20 0 0 0.5 3 3. Ω int 1.
Θ int Local Nusselt Number.5 4 4.66) Nu.Nu 150 125 100 75 50 25 0 Nu.11 0 0 0.5 7 7.5 3 3.49) Nu. These results were obtained for a constant Reynolds number of 1500.22 0.Figure 3.5 Dimensionless Distance.19 shows the dimensionless interface temperature and local Nusselt number distribution plots as a function of a dimensionless radial distance (s/dn) for different solid materials with water as the working fluid. H20(Pr=5. It may be noticed that water presents the lowest interface temperature and highest Nusselt number distribution in comparison with FC77 and MIL7808.33 0. 200 175 0.18 compares the hemisphere’s solidfluid interface temperature and local Nusselt number results of water with flouroinert (FC77) and oil (MIL7808) for the constant heat flux boundary condition.133x106m3/s. The studied 46 .5 1 1. Q=1.0.60 mm.18 Local Nusselt Number and Dimensionless Interface Temperature for Different Cooling Fluids (Re = 1500.55 0. β=2. The lowest Nusselt number is obtained when FC77 is used as the working fluid. FC77(23.30 cm. s/dn 6 6. MIL7808(Pr=124.66 0.88 0.44 0.5 Dimensionless Interface Temperature.77 0.44) H20 Temp FC77 Temp Mil7808 Temp Figure 3. Hn=0.5 2 2.5 5 5. b=0. q=250 kW/m2) Figure 3.
Constantan shows the lowest dimensionless temperature at the impingement zone and the highest at the outlet in comparison with other solid materials. The crossover of curves for all five materials occurred due to a constant fluid flow and heat flux rates that provide a thermal energy balance. and therefore will result in a lower maximum temperature at the solidfluid interface and within the hemispherical plate. silver. aluminum. The choice of material is also crucial in determining the magnitudes of these temperatures. 47 . A material with larger thermal conductivity will facilitate a faster rate of heat transfer. copper.materials were silicon. The dimensionless temperature and local Nusselt number distributions of these two materials are almost identical due to their similar thermal conductivity values. The temperature difference at the interface is an indication of the level of temperature nonuniformity at the impingement surface. Copper and silver show a more uniform distribution and higher temperature values at the impingement zone due to their higher thermal conductivity. and Constantan. Solid materials with lower thermal conductivity show higher maximum local Nusselt number. having different thermophysical properties.
04 0.5 4 4.02 0 Figure 3.19 Local Nusselt Number and Dimensionless Interface Temperature for Different Hemisphere Materials with Water as the Cooling Fluid (Re=1000. Figure 3.5 2 2. β=2.Nu 120 100 80 60 40 20 0 0 0.0. In Figure 3.14 0.5 5 5. Aluminum Temp.21.5 Dimensionless Distance.16 0. q=250 kW/m2) Dimensionless interface temperature and local Nusselt number were also investigated for various solid plate materials such as silicon.06 0. it can be observed that the local 48 . Also. Silicon Temp. silver. copper.160 140 0. b/dn=0.5 3 3.08 0.5.20 demonstrates that materials with higher thermal conductivity such as silver and copper maintain a higher and more uniform temperature distribution along the solidfluid interface transferring heat faster towards the hemisphere’s outer surface. aluminum. Silver Temp. the interface temperature varies over a larger range due to lower rate of conduction heat transfer within the solid.5 0.12 Silver Copper Aluminum Silicon Constantan Temp. and Constantan using the same working fluid (water) for the isothermal boundary condition. Θ int Local Nusselt Number.5 7 7.5 1 1. Constantan has significantly lower interface temperature due to higher thermal resistance offered by it. s/dn 6 6.1 0. Constantan Dimensionless Interface Temperature. Copper Temp.
40 0 0.60 mm.60 0.5 2 2. dn=1.5 3 3. b=0.5 5 5.80 0.20 Dimensionless Interface Temperature for Silicon Hemispherical Plate for Different Materials (Re=750. To=373 K) 49 .2mm.Nusselt number changes only slightly with solid properties and curves are very close to each other.5 1 1.00 0.665x107m3/s.5 Figure 3.5 6 Dimensionless Distance.5 7 7. s/dn Silver Copper Aluminum Silicon Constantan 6.20 Dimensionless Interface Temperature. 1. Q=5. Ω int 1.5 4 4.
76% of Liu et al. and Nakoryakov et al. 750. 1250 and 1500 50 . To=373 K) Three of the papers used for the validation of this numerical study were the analytical works carried out by Liu et al. The results shown at Figure 21 compared within 7. The graphical representation of actual numerical Nusselt number results at the stagnation point at different Reynolds number are shown in Figure 21.5 1 1.21 Local Nusselt Number Distribution for Silicon Plate for Different Solid Materials (Re=750.5 4 4. 1000. Q=5.5 2 2. [36]. s/dn Figure 3. b=0.48% of Nakoryakov et al.5 6 6. within 8.5 5 5. [38] using fluid for Prandtl number greater than unity (Pr>1) as coolants. dn=1. The fluids were tested for heat removal under free liquid jet impingement on a heated flat surface maintained at uniform heat flux. Nu Silver Copper Aluminum Silicon Constantan 40 30 20 10 0 0 0.5 3 3.665x107 m3/s.60mm. The local Nusselt number under Reynolds number at 500.5 Dimensionless Distance. Scholtz and Trass [37]. [38].20% of Scholtz and Trass [37]. and within 7.5 7 7.2mm.60 50 Local Nusselt Number. [36].
Re Figure 3.65%. Considering the errors inherent in any experimental study as well as discretization and roundoff errors in the simulation. 65 60 Stagnation Nusselt number. and 11. Scholtz and Trass Nakoryakov et al. q=250kW/m2) 51 .correlates with a margin of 12. In addition.22 Stagnation Nusselt Number Compared with Liu et al. Scholtz and Trass [37].17%. the surface curvature may have affected the heat transfer characteristics near the impingement region that was not present in experimental studies. Nu 55 50 45 40 35 30 25 20 15 250 Present Values Liu et al. 500 750 1000 1250 1500 1750 Reynolds Number. 6. this comparison is quite satisfactory.85% respectively.6mm.35%.2mm. and Nakoryakov et al.00%. 3. b=0. [36]. 5. [38] with Actual Numerical Results under Different Reynolds Numbers (dn = 1.
and then it decreases downstream.3. Since an initial isothermal condition was assumed at the beginning of the process. it is necessary to analyze the transient heat transfer process when a local heat flux travels throughout the entire solid up to its outer surface and to the cooling fluid. A much larger heat flux is seen at the stagnation region since the cold fluid at the jet strikes that region and keeps the minimum temperature at that location.23 also presents qav. This peak can be associated with the start of the thermal boundary layer in the thin film structure. This behavior occurs due to the constant renewal of cold fluid to dissipate the heat. Figure 3. the heat flux at the solidfluid interfaces increases with time. As expected.23. It may be noticed that a large amount of energy is absorbed by the solid at the early part of the transient and more and more 52 . the interfacial heat flux is zero at t=0 s.2 Transient Heating In order to understand the thermal response of solid materials to the flowing of heat when the power source is turned on (t > 0). This is due to the transition of the fluid from the vertical impingement to a thin film flow along the curved surface where the boundary layer starts to develop. The local heat flux variation along the solidfluid interface for different time intervals is shown in Figure 3.int ro2/qwri2 which is the ratio of the energy transmitted to the fluid to the energy input at the bottom face of the hemisphere. Another maximum heat flux is encountered around s/dn=0.5. The heat dissipated is utilized to rise the temperature of the solid as well as the fluid and reduces thermal storage within the solid due to convective heat transfer.
the thickness of the thermal boundary layer increases and therefore the temperature rises.5 4 4.995 2 2 Fo=0.0054.372.23 Dimensionless Local Heat Flux Variation at the SolidFluid Interface for Silicon Hemisphere at Different Time Instants (Re=750.103.energy is dissipated to the fluid as the transient progresses.5 1 1.5) Figure 3.546 2 2 Fo=0.5 2 2.5 7 7.957 2 2 2 Fo=0. ß=2.5 Dimensionless Distance s/dn Figure 3. qav. qav. b/dn=0.231. 2. qav. qav.int ro /qwri =0.int ro /qwri =0.00154.5 5 5. As time goes on.int ro /qwri =0.990 1.24 also illustrates the difference of 53 .199 2 2 Fo=0. Figure 3.int ro /qwri =0.24 illustrates the dimensionless interface temperature for different time instants.5 3 3.5 0 0 0.int ro /qwri =0.103.056. qav.010. The interfacial heat flux reaches within 1% of the steadystate equilibrium condition at Fo= 0. the solidfluid interface maintains a uniform temperature compared to that when approaching the steadystate condition. qav.int ro /qwri =0. qav.int ro /qwri =0.5 Dimensionless Interfacial Heat Flux (qint/q) 2 2 Fo=0.5 6 6.386 2 2 Fo=0. As can be observed at the very beginning of the heat transfer process.5 2 2 Fo=0.5. This pattern is due to the thermal storage in the fluid necessary to develop the thermal boundary layer since an isothermal condition was present at the beginning of the problem.995 1 0.
0054.026.08 0.25 shows the variation of local Nusselt number along the solidfluid interface at different time instants.02 0 0 0. such temperature difference increases with time as more heat flows throughout the hemispherical solid and transmitted to the fluid.5 4 4.14 0. ΘmaxΘmin(int) = 3. Θ int 0.5 7 7.5 3 3.44E03 Fo = 0. ΘmaxΘmin (int) = 5.0015.70E02 Fo = 0.372. ΘmaxΘmin (int) = 5.056.46E02 Fo = 0. ΘmaxΘmin(int) = 7.27E04 Fo = 0. ΘmaxΘmin (int) = 5.5 1 1.5. Both of these 54 .5) Fig.5 2 2.04 0. Thus.5 Dimensionless Distance (s/dn) Fo = 0.16 Dimensionless Interface Temperature. 3.24 Dimensionless Interface Temperature Variation for Silicon Hemisphere at Different Time Instants (Re=750. The local Nusselt number is controlled by local temperature and heat flux at the solidfluid interface. ΘmaxΘmin (int) = 1.maximumtominimum temperature at the interface being the maximum located at the outer edge of the hemisphere and the minimum at the stagnation point of the hemisphere due to the constant renewal of cold fluid from the nozzle.5 5 5. b/dn=0.231.1 0. ß=2.06 0. 0. ΘmaxΘmin(int) = 6.50E02 Fo = 0. The range of temperature encountered at the solidfluid interface increases with time and reaches a constant value at the steady state.5 6 6.51E02 Figure 3.99E02 Fo = 0.12 0.010. ΘmaxΘmin (int) = 5.17E02 Fo = 0.103.
20 Fo = 0. the average Nusselt number is large at the early part of the transient and monotonically decreases with time ultimately reaching the value for the steady state condition.quantities increase with time. As expected.5 2 2. ß=2.5 1 1.103.5 5 5.25 Local Nusselt Number Distribution for Silicon Hemisphere at Different Time Instants (Re=750. Nuav = 18.09 50 Local Nusselt Number.72 Fo = 0. This essentially means that all heat reaching the solidfluid interface via conduction through the solid is more efficiently convected out as the local fluid temperature is low everywhere at the interface.0055. Nuav = 18.0015. 60 Fo = 0.5.372. Nuav = 24. The local Nusselt number decreases with time until it reaches the steadystate equilibrium distribution.5 3 3.231. Nu 40 30 20 10 0 0 0. Nuav = 20. Nuav = 18.010.26 Fo = 0.02 Fo = 0.5 4 4. Figure 3. Nuav = 18.056.19 Fo = 0. b/dn=0.5 7 7. Nuav = 36.25 also provides the integrated average Nusselt number for the entire hemispherical surface.51 Fo = 0.5 Dimensionless Distance.5) 55 .5 6 6. s/dn Figure 3. The local Nusselt number shows a higher value at early stages of the transient process due to smaller temperature difference between the jet and the outer surface of the hemisphere.
Re=500 Θmax(interface).16 Dimensionless Maximum Temperature.06 0. the temperature begins to rise with time as the hemispherical solid begins to store heat starting after the specified initial condition (TJ=310 K) showing a rapid response at the earlier part of the heating process until its thermal storage capacity reduces up to its limit (steadystate).02 0 0 0.02 0 Figure 3. r=ri).26 Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface. within the Solid.5) 56 Dimensionless MaximumtoMinimum Temperature Difference at the Interface. ß=2.Figure 3. and maximumtominimum difference temperature at the interface for two different Reynolds numbers.15 Fourier Number.04 0. As noticed.04 0. Re=500 Θmax(solid). maximum temperature in the solid. The maximum temperature within the solid was encountered at the outlet plane next to the heated surface (z=ro.3 0.05 0. Re=1000 ΘmaxΘmin(interface).1 0.26 presents results for dimensionless maximum temperature at the interface.08 0.06 0. Θ int 0. Θ maxΘ min . Re=500 ΘmaxΘmin(interface).07 0.05 0.14 0.03 0. 0.25 0. It is important to mention that the time necessary to reach steadystate depends strongly on the Reynolds number.18 0. and MaximumtoMinimum Temperature Difference with Time for Two Reynolds Numbers (Silicon Hemisphere. Fo 0.1 0. Re=1000 0.12 0. Re=1000 Θmax(solid).5. b/dn=0.01 Θmax(interface).2 0.
Re=500 Nuav .07 0.03 0.26 also confirms that a higher Reynolds number increases convective heat transfer and therefore lowers the hemisphere’s temperature. Fo Figure 3.5) 57 .Figure 3.04 0. the average Nusselt number becomes larger as the Reynolds number increases due to higher velocity of the fluid particles moving along the hemisphere’s outer surface.06 0. 120 100 Average Nusselt Number. Re=1000 80 60 40 20 0 0 0. Nu av Nuav . ß=2.27 plots the average Nusselt number variation along the solidfluid interface for two different Reynolds numbers over the entire transient startup of the heat transfer process.02 0. The control of maximum temperature is important in many critical thermal management applications including electronic packaging.5.27 Variation of Average Nusselt Number with Time For Silicon Hemisphere at Two Different Reynolds Numbers (b/dn=0. hence. Figure 3.08 Fourier Number.05 0. As expected. increasing the rate of heat transfer.01 0.
5) The effects of varying the thickness in the hemispherical plate on maximum temperature at the interface.14 0.16 Fourier Number.06 0. maximum temperature within the solid.2 0. ß=2.001% of the steady equilibrium distribution.1 0. Foss was defined as the Fourier number at which the solidfluid interface temperature everywhere on the hemispherical plate reached within 0.12 0. and maximumtominimum temperature difference at the interface can be seen 58 .02 0 400 700 1000 1300 1600 Reynolds Number.18 0.08 0.5.28.04 0. This is due to more fluid flow rate available to carry away the heat and faster development of thermal boundary layer that is smaller in thickness. Re Figure 3.The time required to reach steadystate for different Reynolds numbers is presented in Figure 3. 0. FoSS 0.28 Time Required to Reach SteadyState for Silicon Hemisphere at Different Reynolds Numbers (b/dn=0. The time to reach thermal equilibrium condition decreases as the Reynolds number increases in value.
12 0.02 Θmax(solid).05 0. Fo 0. Θ maxΘ min 0.03 0. b/dn=1. the time needed to achieve steadystate conditions increases. b/dn=1. Also.Θ max 0. and MaximumtoMinimum Temperature Difference with Time for Different Plate Thicknesses (Silicon Hemisphere Re=750.5 0. It may be note that as the thickness of the hemispherical plate increases. This is due to more storage capacity of heat within the solid.01 0 0. 0.1 0.5 Θmax(interface).1 0.5) 59 .29 Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface.3 Fourier Number.08 0.2 0.04 0. The plate thickness significantly affects the temperature distribution.06 Dimensionless MaximumtoMinimum Temperature Difference at the Interface. ß=2.02 0.0 ΘmaxΘmin(interface).0 0 0 0.0 0.5 Θmax(solid). b/dn=0. b/dn=0. within the Solid.04 0. b/dn=0.16 0.6 Figure 3.on Figure 3.06 Θmax(interface).5 ΘmaxΘmin(interface).29. the temperature at the solidfluid interface remains lower due to higher thermal resistance of the solid to the path of heat flow.14 Dimensionless Maximum Temperature.4 0. b/dn=1.
01 0.05 0.04 0.5 b/d n=1. maximum temperature within the solid.06 0.083 b/d n=0. The average Nusselt number is higher for higher plate thickness.02 0.5) The maximum temperature at the solidfluid interface. the temperature changes occur faster at the beginning of the heating process for all materials and the slope gradually decays when the steadystate approaches. ß=2.31. It can be observed that materials 60 . and maximum to minimum temperature difference (as a measure for temperature nonuniformity) for different solid materials is presented in Figure 3.0 Figure 3. Fo b/d n=0.30 Variation of Average Nusselt Number with Time for Constantan Hemisphere at Three Different Thicknesses (Re=750. As expected.07 0.03 0.Figure 3. 200 175 Average Nusselt Number.08 Fourier Number. Nuav 150 125 100 75 50 25 0 0 0.30 shows the average Nusselt number variation as a function of time for three distinct plate thicknesses using Constantan as the solid material.
4 0. Copper 0.06 Θmax(solid). As noticed. Copper Θmax(interface). Constantan ΘmaxΘmin(interface).02 0 0.06 0.1 0. within the Solid. Silicon 0. The thermal diffusivity of the material also contributes to the transient behavior of the solid. Silicon and Copper reach the steadystate faster than Constantan due to their higher thermal diffusivity.2 ΘmaxΘmin(interface).14 0.04 0.08 0. Copper Θmax(solid).04 0.31 Distribution of Dimensionless Maximum Temperature at the SolidFluid interface. Constantan Θmax(solid). Silicon Θmax(interface).1 0. Θ max 0.1 0. Silicon ΘmaxΘmin(interface).3 0.18 Dimensionless Maximum Temperature. 0.12 0.02 0 0 0.12 Dimensionless MaximumtoMinimum Temperature Difference at the Interface Θ maxΘ min 0.16 0.2 0.with a very low thermal conductivity such as Constantan maintains a higher temperature at the hemisphere’s outer surface and within the solid as the thermal conductivity controls how effective the heat flows and distributes throughout the material. ß=2.5 Fourier Number. and MaximumtoMinimum Temperature Difference with Time for Different Materials (Re=750. Fo Figure 3. Constantan Θmax(interface).5) 61 . which controls the rate of heat being transferred through the solid material.08 0.
Constantan shows a higher average heat transfer coefficient compared to Silicon or Copper over the entire transient process.32 shows the distribution of average Nusselt Number with time for the three materials used in this study.05 Fourier Number.04 0.03 0.02 0. Fo Constantan Copper Silicon Figure 3. ß=2. A significant difference is seen at the earlier part of the transient and the curves come close together as the steadystate approaches. Constantan takes longer in reaching steadystate due to its lower thermal diffusivity compared to Copper and Silicon. b/dn.5. 62 . 90 80 Average Nusselt Number. Nu av 70 60 50 40 30 20 10 0 0 0.5) It will be also important to know how the materials responded in reaching thermal equilibrium based on their thickness. the time to reach steadystate also increases. Figure 3.32 Variation of Average Nusselt Number with Time for Different Materials (Re=750. As the thickness increases in value.=0.33 presents Foss for these materials for different plate thicknesses.Figure 3.01 0.
5) The development of isothermal lines within the solid can be observed in Figure 3.5.4 0. F o 0. 0. they start to form 63 .2 0.5 0.34 at different time instants for b/dn=0. It is important to notice that at early stages of the transient heat transfer process. ß=2.2 0.6 Fourier Number. the isothermal lines start moving upward toward lower temperature regions until they reach the solidfluid interface.1 0 0 0.4 1. the property of the solid plays more significant role in determining the duration of the transient heat transfer process when the thickness is increased.3 0. b/dn Constantan Copper Silicon Figure 3.Also. the isothermal lines grow parallel to the inner (bottom) heated surface of the hemispherical plate. As time goes on.6 0.4 0.2 1.33 Time Needed to Reach SteadyState for Different Hemispherical Plate Thicknesses and for Different Materials (Re=750. After that.8 1 1.7 0.6 Dimensionless Plate Thickness.8 0.
93 K Tmin=310.65 K t=0.1 s Tmax=345.concentric lines near the stagnation point and expand further down into the solid until a steadystate condition is achieved.14 K t=0.5 s t=1.99 K t=0.57 K t=0.05 K Tmin=324. ß=2. Top Left Corner (z=0.14 K t=0.2 s t=2. r=0)) 64 .15 K Tmax=310.5 (Re=750.91 K Tmin=342.05 s Tmax=344.00028 s Tmax=319. Tmax=310.5.15 K t=0.002 s Tmax=326.01 s Tmax=343.39 K Tmin=318.34 Isotherms Countour Lines at Varius Time Instants for Silicon Hemisphere Plate of b/dn=0.60 K Tmin=340.45 K Tmin=312.48 K Tmax=313.2 s Figure 3.24 K Tmin=310.20 K Tmin=343.
Figure 3.35 shows the same phenomenon for b/dn=1.5. The temperatures inside the solid are much lower as compared to a thinner plate and the time to reach steadystate equilibrium thermal condition increases. Tmax=310.43 K Tmin=310.15 K Tmax=310.94 K Tmin=310.19 K Tmax=311.68 K Tmin=310.58 K
t=0.0002 s Tmax=313.72 K Tmin=312.30 K
t=0.002 s Tmax=316.13 K Tmin=314.97 K
t=0.01 s Tmax=329.57 K Tmin=327.05 K
t=0.05 s Tmax=330.75 K Tmin=329.64 K
t=0.1 s Tmax=330.98 K Tmin=329.87 K
t=1.03 s
t=2.06 s
t=2.80 s
Figure 3.35 Isotherms Countour Lines at Varius Time Instants for Silicon Hemisphere Plate of b/dn=1.5 (Re=750, ß=2.5, Top Left Corner (z=0, r=0)
65
Chapter 4 Cylindrical Plate Model Results
4.1 Steady State Heating
A velocity vector distribution is shown in Figure 4.1. It can be observed that the velocity remains almost uniform at the potential core region of the jet. However, the velocity decreases and the fluid jet diameter increases as the fluid gets closer to the surface during the impingement process. The direction of motion of the fluid particles shifts along the angle of curvature, afterwards, the fluid accelerates creating a region of minimum sheet thickness. That is the beginning of the boundary layer zone. It can be noticed that as the fluid moves downstream along the convex surface, the boundary layer thickness increases and the frictional resistance from the wall is eventually transmitted to the entire film thickness where the fully viscous zone develops.
66
Figure 4.1 Velocity Vector Distribution for a Silicon Curved Plate with Water as the Cooling Fluid (Re = 750, b= 0.60 mm, Q = 5.665x107 m3/s, β = 2.5, Hn = 0.3 cm, q = 250 kW/m2)
Fig. 4.2 shows the free surface height distribution for different Reynolds numbers when the jet strikes the curved plate’s surface. It can be seen that the fluid spreads out moving along the outer surface of the curved plate. As the Reynolds number increases, the free surface increases in thickness due to higher flow rate. In addition, a higher Reynolds number causes a higher rebounding velocity of the fluid particles at the impact that also contributes to a higher film thickness, particularly near the impingement region. For the conditions considered in the present investigation, the flow was supercritical and no hydraulic jump was present.
67
2 Free Surface Height Distribution for Different Reynolds Numbers and Water as the Cooling Fluid (b=0.24 0. β=2.36 0.48 Radial Distance (cm) Figure 4. q=250 kW/m2) Figures 4.4 show dimensionless interface temperature and local Nusselt number distributions as a function of dimensionless distance (s/dn) along the solidfluid interface at different Reynolds numbers.30 cm.70 0.86 0.18 0.66 0.3 reveal that the dimensionless interface temperature decreases with jet velocity or Reynolds number.82 0..42 0.50 0.06 0.60 mm. This also shows that an increasing Reynolds number contributes to a more effective cooling by the enhancement of convective heat transfer coefficient.62 0.3 and 4. 68 .78 0.74 Axial Distance (cm) 0.54 0.30 0. 0. The dimensionless interface temperature has the lowest value near the stagnation point (underneath the jet opening) where the fluid particles start to move along the disk surface (after the impingement process) and increases along the arc length reaching the highest value at the end of it.42 0. The curves in Figure 4.12 0.0. Hn=0.00 Re=500 Re=750 Re=1000 Re=1200 Re=1500 Re=1800 0.58 0.46 0.
Θ int Re=500 Re=750 Re=1000 Re=1200 Re=1500 Re=1800 0.5 4. Hn= 0.5.0 0. Figure 4. and then decreases along s as the boundary layer develops further downstream.4 also confirms that an increasing Reynolds number contributes to a more effective cooling by the enhancement of convective heat transfer coefficient. The location of the maximum Nusselt number can be associated with the transition of the flow from the vertical impingement to displacement along the disk surface where the boundary layer starts to develop. reaches a maximum around s/dn = 0.5 2. s/dn Figure 4.5 1.30 cm.0 3.1 0 0.0 Dimensionless Distance.2 Dimensionless Interface Temperature. q=250 kW/m2) The local Nusselt number shown in Figure 4.5 3.0.3 Dimensionless Interface Temperature Distribution for a Silicon Curved Plate at Different Reynolds Numbers and Water as the Cooling Fluid (β=2. The local Nusselt number increased very significantly all along the plate surface with the increase of Reynolds number.60 mm.0 2.0 1. b=0.23. 69 .4 increases rapidly over a small distance measured from the stagnation point (core region).
30 cm.60 mm.5 1 1. That results in lowering of the maximum temperature.5 plots the average Nusselt number (Nuav) and maximum temperature in the solid (Θmax) as a function of Reynolds number.5 4 Dimensionless Distance. q=250 kW/m2) Figure 4. As the flow rate (or Reynolds number) increases. s/dn Figure 4. the magnitude of fluid velocity near the solidfluid interface that controls the convective heat transfer rate increases. b=0.5 2 2. It may be noted that average Nusselt number increases with Reynolds number and maximum temperature within the solid decreases with increasing Reynolds number. 70 . The maximum temperature happens at the outer edge of the plate adjacent to the inner surface which is heated.5. The control of maximum temperature may be crucial in the thermal management of electronic equipment.60 50 Local Nusselt Number. Nu 40 Re=500 Re=750 Re=1000 Re=1200 Re=1500 Re=1800 30 20 10 0 0 0.5 3 3.4 Local Nusselt Number Distribution for a Silicon Curved Plate at Different Reynolds Numbers and Water as the Cooling Fluid (β =2. Hn=0.
5.02 0 2000 600 800 1000 1200 1400 1600 1800 Reynolds Number. Hn = 0. the temperature at the solid. As the curved plate surface gets closer to the nozzle (smaller nozzletotarget spacing ratio). Re Figure 4.fluid interface decreases due to higher jet momentum at impingement that causes higher velocity of fluid particles adjacent to the plate enhancing the heat transfer. Nu av 30 25 0.18 Dimensionless Maximum Temperature in Solid.30cm) The solidfluid dimensionless interface temperature and local Nusselt number distributions for six different nozzletotarget spacing for water as the cooling fluid and Reynolds number of 750 are shown in Figures 4.14 0.06 10 5 0 400 Nuav Θ max 0.1 20 0.5 Average Nusselt Number and Maximum Temperature Within the Solid for Different Reynolds Numbers for Constant Heat Flux Conditions (q = 250 KW/m2.04 0.7 respectively. Θ max 0. β=2.08 15 0. 71 .12 0.16 0.40 35 Average Nusselt Number.6 and 4.
02 0 0.2mm.0 β=2.5 4.16 Dimensionless Interface Temperature. It can be also noticed that no significant change is seen at heights higher than β=2.5 1.0 β=0. b=0.5 β=3.06 0. in Figure 4.75 β=1.0 2.14 0.5 β=2. a higher Nusselt number is seen all along the arc length.6 Dimensionless Interface Temperature for Silicon Curved Plate for Different Nozzle to Target Spacing Ratio (β) and Water as the Cooling Fluid (Re=750.0 0. It can be noticed that the impingement height affects the Nusselt number more at the stagnation region and the early part of the boundary layer region.5 2.1 0. This is quite expected since the jet momentum more strongly affects the areas subjected to direct impingement. dn=1.7.5 3.0 β=1. 72 .0 3.0 Dimensionless Distance.60 mm. s/dn Figure 4. Θint 0. where impingement height does not play a strong role in determining the convective heat transfer process.08 0. q=250 kW/m2) Therefore.0 1.04 0.12 0.0. At larger arc length the values are in a small range for all impingement heights.
The effects of varying the inner radius of curvature of the impingement plate are demonstrated in Figures 4.8 shows the dimensionless 73 .0 1.5 3. q=250kW/m2) Therefore. s/dn Figure 4. Figure 4.75 β=1.0 β=2. b=0.0 Dimensionless Distance.2mm. At larger arc length the values are in a small range for all impingement heights.0 2.0 3. Nu 40 β=0.10.5 1.7. in Figure 4.5 2.5 β=2. a higher Nusselt number is seen all along the arc length. This is quite expected since the jet momentum more strongly affects the areas subjected to direct impingement. where impingement height does not play a strong role in determining the convective heat transfer process. It can be noticed that the impingement height affects the Nusselt number more at the stagnation region and the early part of the boundary layer region.60 50 Local Nusselt Number.84.7 Local Nusselt Number Distribution for Silicon Curved Plate for Different Nozzle to Target Spacing Ratio (β) and Water as the Cooling Fluid (Re=750.5 4. dn=1.5 β=3.0 0.0 β=1.0 30 20 10 0 0. It can be also noticed that no significant change is seen at heights higher than β=2.60mm.
30 cm Re=750.16 n R /dn =6.5 2.14 0.1 0.2 Dimensionless Interface Temperature.0 2.5 .66 Flat Figure 4. b=0.60 mm. It can be noticed that decreasing the inner plate radius of curvature to nozzle diameter ratio (Ri/dn) decreases the dimensionless interface temperature. dn=1. the arc extended over an angle of φmaxthat is less than π/2. Therefore.02 0 0. half cylindrical plate (shown in Fig.06 0.04 0.0 Dimensionless Distance.0 1.5 3.12 0. The corresponding distribution for a flat plate (infinite radius of curvature) is also shown for comparison.66 i Ri/d =10 n Ri/dn =16.16 0. s was same as its radius. For other plates.0 3. 0. Hn = 0.8 Dimensionless Interface Temperature for Silicon Curved Plate for Different Inner Plate Radius of Curvature to Nozzle Diameter Ratio (Ri/dn) for Water as the Cooling Fluid (β =2.5 1. In order to keep the same energy input to the plate in these runs.5 4.16). q=250 kW/m2) 74 . 1) was possible only for the smaller radius of curvature (Ri/dn=4. Θ int 0.18 0. the value of arc length s was kept constant. s/dn Ri/d = 4.0 0.solidfluid interface temperature of water flowing on the curved plate with different Ri/dn as a function of dimensionless distance (s/dn). For flat disk.2mm.08 0.
16 i n R /dn =6. The trend in Figure 4.9 agrees with the results reported by Martin [39]. b=0.5 1.2mm.0 0.23 and after that it decays monotonically.66 i n Flat 3.5 . s/dn Figure 4.0 R /d =10 i n R /d =16. A higher Nusselt number is encountered when Ri/dn diminishes increasing the cooling capacity of the fluid.60 mm.5 2.66 i Local Nusselt Number. For all cases. the maximum local Nusselt number is encountered at s/dn≈0.0 2.9 plots the local Nusselt number distribution for different inner plate radius of curvature to nozzle diameter ratio (Ri/dn) including a flat plate as a function of dimensionless distance (s/dn).0 1.9 Local Nusselt Number Distribution for Silicon Curved Plate for Different Inner Plate Radius of Curvature to Nozzle Diameter Ratio (Ri/dn) for Water as the Cooling Fluid (β =2. Nu 30 25 20 15 10 5 0 0.0 Dimensionless Distance. 40 35 R /d = 4. q=250 kW/m2) 75 .5 3.5 4. Hn = 0.This is because the fluid particles are subjected to a larger gravitational force that increases the local fluid velocity near the solidfluid interface and results in larger rate of convective heat transfer. dn=1. Figure 4.30 cm Re=750.
0 16.0 2.0 Figure 4. Hn = 0. When flow rate (or Reynolds number) is kept constant.60 mm.5 0.5 . As expected.0 8.11 presents the local heat transfer coefficient and Nusselt number distributions for different nozzle diameters (dn) for a constant flow rate using water as the working fluid.0 Ri/dn 0. b=0.0 12.17 22 Nuav Θmax Nuav 0.16 21 0.175 Dimensionless Maximum Temperature in Solid.0 6.5 0. dn=1.5 0. a reduction in nozzle slot opening requires a proportional increase of jet 76 .0 4.15 20 0.0 10.155 20.Figure 4.30 cm Re=750.10 plots the average Nusselt number (Nuav) and the maximum temperature within the solid (θmax) as a function of the inner plate radius of curvature to nozzle diameter ratio (Ri/dn).2mm. q=250 kW/m2) Figure 4. the average Nusselt number decreases with increasing inner radius of curvature and maximum temperature increases.145 18.10 Average Nusselt Number and Maximum Temperature in the Solid Variations (Θmax) for Silicon Curved plate for Different Inner Plate Radius of Curvature to Nozzle Diameter Ratio (Ri/dn) for Water as the Cooling Fluid (β =2.165 21.0 14. 23 0. Θ max 22.
2mm N . s/dn 16000 9000 2000 16.5.3mm N .3mm h. ddn=1.4 mm h. Re=750. ddn=2.30 cm β=2.11 Heat Transfer Coefficient and Nusselt Number Distributions for Different Nozzle Slot Widths (Q=0. increasing (dn) results in a higher local Nusselt number along the solidfluid interface. Nu 30 25 30000 20 23000 15 10 5 0 0. since the Nusselt number used the slot width as the length scale.2 mm Nu. =1. ddn=0. A higher jet velocity essentially contributes to a higher fluid velocity near the solidfluid interface and higher value of convective heat transfer coefficient.0006 m3/s. However. 45 40 35 Local Nusselt Number. As the thickness increases. d n=0.5 .12 and 4.4mm 51000 .3 mm Nu. =0.0 Dimensionless Distance.13 display the effects of varying plate thickness to nozzle diameter ratio (b/dn) for a silicon plate using water as the cooling fluid.0 12. d n=2.2 mm h. d n=1. Hn = 0.0 2.3 mm h. the temperature at the solidfluid interface decreases and becomes more uniform due to a better distribution of heat within the solid by 77 Local Heat Transfer Coefficient. d =2.2mm h.0 8.0 58000 44000 37000 Figure 4. q=250 kW/m2) Figures 4. h Nu.0 6.inlet velocity.0 10. d =1.0 14.4mm N . d =0.4 mm h. =2.0 4. β =2.
08 b/dn =0.5 0.0 1. It can be also noted that a thinner plate provides a lower Nusselt number at the stagnation region but increases to provide higher maximum value and maintains a higher value as the fluid moves downstream along the curved plate.0 b/dn =1. dn=1.5.16 Dimensionless Interface Temperature.14 0. a thinner plate contributes to a more effective cooling by convection as opposed to a thicker plate.5 4. It may be also noticed that the temperature is lower at the stagnation region and it increases as s/dn increases all the way to the end of the plate.conduction.06 0. q=250 kW/m2) 78 .25 b/dn =1.0 Dimensionless Distance.0 2.12 0. Thus.2mm.04 0.0 b/dn =0.5 1.02 0 0.0 3.1 0.12 Dimensionless Interface Temperature for Different Plate Thicknesses (β =2. Re=750.5 3. s/dn Figure 4. 0.08 0. Θ int 0.5 2.
14 presents the average Nusselt number (Nuav) and maximum temperature (Θmax) encountered within the solid as a function of (b/dn). Hn = 0. A lower average Nusselt number is a result of integrated local values seen in Figure 4. A lower maximum temperature for a thicker plate is a result of lower total heat input when plate 79 . where a thicker plate provided a lower Nusselt number over a major portion of s/dn.08 b/dn =0. Re=750.5 3.5 1.13 Local Nusselt Number Distribution for Different Plate Thicknesses (β =2.0 2. Nu 25 20 15 10 5 0 0. s/dn Figure 4.5 2. It can be noticed that both average Nusselt number and maximum temperature within the solid diminish with increase of plate thickness.0 1.30 cm.0 b/dn =1.0 0.5 4. The maximum temperature within the solid occurs at the inner surface of the curved plate at its outlet end (maximum s or φ).0 b/dn =0.5.25 b/dn =1.0 Dimensionless Distance.5 3.35 30 Local Nusselt Number. q=250 kW/m2) Figure 4.14.
5.2 22 21.18 0.1 21.thickness was higher.15 K) and MIL7808 (TJ=375 K).12 Figure 4.14 Average Nusselt Number and Maximum Temperature in the Solid Variations with Plate Thickness (β =2.8 20.2 0.2 21 20.6 0.15 compares the dimensionless solidfluid interface temperature for the present working fluid (water) with two other cooling fluids flouroinert (FC77) and oil (MIL7808).6 0 0.04 0.4 0. Re=750) Figure 4.8 1 1.14 0.4 0.02 0 Nuav Θmax 0.2 Dimensionless Maximum Temperature in Solid.15 K) presents the highest dimensionless interface temperature when compared to FC77 (TJ=310.08 21. A higher Prandtl number results in lower dimensionless temperature at the solidfluid interface.06 0.16 0. Θ max 0. In addition.6 b/dn 0.4 22. This trend is well correlated with Prandtl number. a low temperature is maintained over a larger portion 80 .2 0. due to lower thickness or smaller growth rate of thermal boundary layer at higher Prandtl number.8 Nuav 21. Water (TJ=310. 22. as the heat flux at the inner surface of the plate and outer radius of the plate were kept constants in these simulations.
02 0 0 H2O(Pr=4. the curves for MIL7808 and water cross each other and at large values of s.14 0. However. s/dn Figure 4. Θ int 0.5 3 3. 0.16 Dimensionless Interface Temperature. Re=750) As expected.5 2 2.5 1 1.04 0.08 0.1 0.06 0.5 4 Dimensionless Distance. However. A higher Nusselt number distribution is encountered when MIL7808 is used as the working fluid in comparison with FC77 (q=50 kW/m2) and water. The highest heat transfer coefficient is still provided by MIL7808 at the impingement region. when dimensional heat transfer coefficient is considered.15 Dimensionless Interface Temperature for Different Cooling Fluids (β=2. the 81 .5. Figure 4.32) 0.of the plate resulting in better cooling performance.16 displays the local heat transfer coefficient and local Nusselt number for these three fluids.45) FC77(Pr=16.12 0. a higher Nusselt number is obtained for a higher Prandtl number fluid. the Prandtl number alone cannot correlate the trend.58) MIL7808(Pr=159.
5 4 Dimensionless Distance.5 3 3. FC77 (kf =0. H20(Pr=4. MIL7808(Pr=159.45) Nu. H2O(Pr=4.32) 25000 15000 10000 5000 0 Figure 4. 200 180 160 Local Nusselt Number. β=2.32) h. FC77(Pr=16. The numerical simulation attempted to duplicate the exact conditions of this experiment.5 2 2. h 20000 .06299 W/m K) gives much lower heat transfer coefficient compared to water or MIL7808 because of its lowest thermal conductivity. Nu 140 120 100 80 60 40 20 0 0 0.58) h. These results were obtained for a constant Reynolds number of 750. MIL7808(Pr=159.highest heat transfer coefficient is obtained for water.45) h.5 1 1.58) Nu. FC77(Pr=16.16 Heat Transfer Coefficient and Nusselt Number Distributions for Different Cooling Fluids (Re=750. s/dn Nu. A steady state mixing length turbulence model was employed for 82 Local Heat Transfer Coefficient.5) One of the papers used for the validation of this numerical study was the experimental work carried out by Bartoli and Faggiani [21] where water was used for the cooling of a stainless steel hollow cylinder by jet impingement.
Re Figure 4.15 K. TJ=293.500 and 1.15– 323.5 mm.99. β=5.0 mm.50 mm. Pr=3.4 60 40 Present Values 20 Bartoli and Faggiani 0 3000 8000 13000 18000 Reynolds Number.000. dn=2. The parameters used and their corresponding values according to Bartoli and Faggiani [21] were the following: Re=3500–20. Figure 4.88% at Re= 15.17 compared the results obtained in the numerical simulation with the correlation proposed by Bartoli and Faggiani [21] for Nu/Pr0. The difference is 3.5.60% at Re= 3. b=0. q=50 kW/m2) 83 .4.this simulation. 120 100 80 Nu/Pr 0.000. q=50 kW/m2.57–6. φ=0°90°. and b=0.17 Local Nusselt Number Compared with Bartoli and Faggiani [21] at Different Reynolds Numbers (φ=90°.
a correlation suggested by Whitaker [40] for uniform fluid flow was also used for comparison.18 illustrates the numerical results obtained in the simulation compared to the experimental data obtained by Gori and Bossi [29] along with 84 .35 kW/m2) Figure 4.18 Average Nusselt Number Compared with Gori and Bossi [29] and Whitaker [40] for Different Reynolds Numbers (dn=2. 140 Numerical Values 120 100 80 Nuav 60 40 20 0 0 5000 10000 15000 20000 25000 Reynolds Number.2 mm. q=2.5 mm.Numerical simulations were also carried out to compare with the results obtained by Gori and Bossi [29] on the cooling of a hollow stainless steel circular cylinder by a turbulent flow of air from a slot nozzle. In addition. b = 0. Re Gori and Bossi Whitaker Figure 4. The numerical model tried to mimic the experimental set up that the authors used to obtain the desired outcome for a constant heat flux condition.
76%15. Since an initial isothermal condition was assumed at the beginning of the process. the interfacial heat flux is zero at t=0 s.77% with an average difference of 15. 4.19. This behavior occurs due to the constant renewal of cold fluid to dissipate the heat.711– 0. The values used during the simulation included a Reynolds numbers range of 400020.82% with an average difference of 9.2 mm as specified by the authors. the heat flux at the solidfluid interfaces increases with time.35 kW/m2.0. a fluid inlet (jet) temperature (TJ) of 293. The correlation proposed by Whitaker compared within a range of 1.712.the correlation proposed by Whitaker [40].15–296.92%. and a fixed thickness (b) of 0. both of these comparisons may be considered to be quite satisfactory.31%. As expected. The heat dissipated is utilized to rise the temperature of the solid as well as the 85 . The difference of the numerical simulation and the experimental results obtained by Gori and Bossi [29] was between 6.2 Transient Heating The local heat flux variation along the solidfluid interface for different time intervals is shown in Figure 4.15 K.000. an azimuthal angle (φ) οf 0°90°. a fixed dimensionless nozzle to target spacing ratio (β) of 8.80%23. a constant heat flux (q) of 2. A much larger heat flux is seen at the stagnation region since the cold fluid at the jet strikes that region and keeps the minimum temperature at that location. Considering the uncertainly of experimental data and the discretization and roundoff errors inherent in numerical simulation. the fluid Prandtl number (Pr) of 0.
and then it decreases downstream.int ro2/qwri2 =0. qav.0107.051. qav.837 2/qwri2 =0.999 Figure 4.fluid and reduces thermal storage within the solid due to convective heat transfer.231. qav. 1.405 2/qwri2 =0.6 Dim ensionless Intefacial H eat Flux.103.int ro2/qwri2 which is the ratio of the energy transmitted to the fluid to the energy input at the bottom face of the plate. qav.0 2.=0. This can be explained as the transition of the fluid from the vertical impingement to a thin film flow along the curved surface where the boundary layer starts to develop. ß=2.0013. =1.103.619 Fo=0.5 4. Another maximum heat flux is encountered around s/dn=0.231.5.0 1.int ro2/qwri2 =0.int ro2/qwri2 =0.5 3.int ro Fo=0.4 1.5.0107.0 0.2 0 0.0211.405 Fo=0. qav.8 0. =1.2 1 0.023 Fo=0.19 also presents qav.19 Dimensionless Local Heat flux Variation at the SolidFluid Interface for Silicon Plate at Different Time Instants (Re=750.060 Fo=0. Figure 4. =1. qav. The interfacial heat flux reaches within 1% of the steadystate equilibrium condition at Fo= 0.0051. b/dn=0. (qint/q) 1. =0.103.999 Fo=0.0211.4 0.6 0.057 Fo=0.051.837 Fo=0. This crest can be associated with the start of the thermal boundary layer in the thin film structure.177 Fo=0.0 3.int ro Fo=0.int ro2/qwri2 =0.=0.0 Dimensionless Distance. s/dn Fo=0.int ro2/qwri2 =0.5 1.987 Fo=0.5 2.0051.5) 86 . =0.619 Fo=0.0013. qav.177 Fo=0.
As time goes on. Figure 4.20 also illustrates the difference of maximumtominimum temperature at the interface being the maximum situated at the outer edge of the plate and the minimum at the stagnation point of the plate due to the constant renewal of cold fluid from the nozzle. the thickness of the thermal boundary layer increases and therefore the temperature rises.Figure 4. The range of temperature encountered at the solidfluid interface increases with time and reaches a constant value at the steady state. The local Nusselt number is controlled by local temperature and heat flux at the solidfluid interface. This pattern is due to the thermal storage in the fluid necessary to develop the thermal boundary layer since an isothermal condition was present at the beginning of the problem. This means that all heat reaching the solidfluid interface via conduction through the solid is more efficiently convected out as the local fluid temperature is low everywhere at the interface. As can be observed at the very beginning of the heat transfer process.21 plots the local Nusselt number variation along the solidfluid interface at different time instants. 87 . such temperature difference increases with time as more heat flows throughout the solid and transmitted to the fluid. Figure 4. Thus.20 shows the dimensionless interface temperature for different time instants. the solidfluid interface keeps a uniform temperature compared to that when approaching steadystate equilibrium condition.
Figure 4. b/dn=0.0107.0 2. Θ int 0.231.5 1.20 Dimensionless Interface Temperature Variation for Silicon Plate at Different Time Instants (Re=750.0 1.12 0. the average Nusselt number is large at the early part of the transient and monotonically decreases with time all the way to the steady state condition. That is.45E02 Fo=0. Tf=Ts. ΘmaxΘmin(int)=8.1 0.58E02 Fo=0.5.5) The local Nusselt number decreases with time until it reaches the steadystate equilibrium distribution. ß=2.0013.103. equilibrium conditions with the surroondings have been achieved between the plate and the fluid and the temperature of the plate along with the fluid becomes uniform.62E02 Fo=0.49E02 Fo=0. ΘmaxΘmin (int)=5. At this point.0 Figure 4.06 0.0 Fo=0. s/dn 3. ΘmaxΘmin(int)=5.18 Dimensionless Interface Temperature.14 0.5 2. ΘmaxΘmin(int)=8.0051. As expected.5 3.5 4.76E02 0.78E02 Fo=0. ΘmaxΘmin(int)=8.16 0. 88 . ΘmaxΘmin(int)=8.36E03 Fo=0.0 Dimensionless Distance. ΘmaxΘmin(int)=7.22 also provides the integrated average Nusselt number for the entire cylindrical shaped plate surface.0.08 0.02 0 0.04 0.051.021.
103. and maximumtominimum difference temperature at the interface for two different Reynolds numbers.5 3.231.45 Fo=0.13 Fo=0.Nuav=43. Nuav=21.Nuav=21. The maximum temperature within the solid was found at the outlet plane next to the heated surface (z=ro.Nuav=24.94 Fo=0.231.21 Local Nusselt Number Distribution for Silicon Plate at Different Time Instants (Re=750. Nuav=21.09 Fo=0.09 0. Nuav=22.60 50 Local Nusselt Number.5.22 shows results for dimensionless maximum temperature at the interface.051.5 1.21 Fo=0. As expected.051. Nuav=21. maximum temperature in the solid. Nu 40 30 20 10 0 0.0 Fo=0.021. r=ri). b/dn=0.0107. ß=2.0 1. Nuav=21.5 4. Nuav=24.0211.5) Figure 4.0013. Nuav=21.103. the temperature begins to rise with time as the solid plate begins to store heat starting after the specified initial condition (TJ=310 K) showing a rapid response at the earlier part of the heating process until its thermal storage capacity reduces up to its limit (steadystate).Nuav=22.41 Fo=0.94 Fo=0. 89 .41 Fo=0.0051.11 Fo=0.0 Dimensionless Distance.13 Fo=0.45 Fo=0. s/dn Figure 4.0051.11 Fo=0. Nuav=43.0 3.21 Fo=0.0013.5 2.Nuav=21.0 2.Nuav=21.0107.
90 Θ min . Re=500 0. b/dn=0.02 0 0 0.Θ max0. Re=500 0.05 0.01 0 Figure 4. Re=1000 Θmax(solid).1 Dimensionless MaximumtoMinimum Temperature Difference at the Interface. As expected.14 0.04 0.1 0.18 0.22 also proves that a higher Reynolds number increases convective heat transfer and therefore lowers the plate’s temperature. increasing the rate of heat transfer.03 0.08 0. Re=1000 0.5) It is significant to point out that the time necessary to reach steadystate depends on the Reynolds number.22 Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface.07 0. hence.2 ΘmaxΘmin(interface).06 Θmax(solid). ß=2.5. within the Solid. Figure 4.09 0.16 Dimensionless Maximum Temperature.06 0.05 Θmax(interface). Θ int 0.1 Fourier Number.08 0. Re=1000 Θmax(interface).0. Re=500 ΘmaxΘmin(interface). and MaximumtoMinimum Temperature Difference with Time for Two Reynolds Numbers (Silicon Plate.15 0.04 0.23 plots the average Nusselt number variation along the solidfluid interface for two different Reynolds numbers over the entire transient startup of the heat transfer process. Figure 4.12 0. the average Nusselt number becomes larger as the Reynolds number increases due to higher velocity of the fluid particles moving along the plate’s outer surface. Fo 0.02 0.
23 Variation of Average Nusselt Number with Time For Silicon Plate at Two Different Reynolds Numbers (b/dn=0. Re=500 Nuav. ß=2.07 0. This is due to more fluid flow rate available to carry away the heat and faster development of thermal boundary layer that is smaller in thickness.24.5. Re=1000 Series2 90 60 30 0 0 0.001% of the steady equilibrium distribution. Nuav 120 Series1 Nuav. Foss was defined as the Fourier number at which the solidfluid interface temperature everywhere on the solid plate reached within 0.01 0.02 0.03 0. The time to reach thermal equilibrium condition decreases as the Reynolds number increases in value.5) The time required to reach steadystate for different Reynolds numbers is presented in Figure 4.05 0.08 Fourier Number.150 Average Nusselt Number. Fo Figure 4.06 0. 91 .04 0.
Re Figure 4. the time needed to achieve steadystate conditions increases.02 0 400 600 800 1000 1200 1400 1600 Reynolds Number. the temperature at the solidfluid interface remains lower due to higher thermal resistance of the solid to the path of heat flow.08 0. maximum temperature within the solid.06 0. and maximumtominimum temperature difference at the interface can be seen on Figure 4. Similarly. 92 . Foss 0.12 0. It may be note that as the thickness of the solid plate increases. The plate thickness significantly affects the temperature distribution.04 0.1 0.14 Fourier Number. This is due to more storage capacity of heat within the solid.18 0.0.25.24 Time Required to Reach SteadyState for Silicon Plate at Different Reynolds Numbers (b/dn=0. ß=2.5.5) The effects of varying the thickness in the cylindrical plate on maximum temperature at the interface.16 0.
1 0.06 0.07 Θmax(interface).18 Dimensionless Maximum Temperature. ß=2. and maximum to minimum temperature difference (as a measure for temperature nonuniformity) for different solid materials is presented in Figure 4.06 0.01 0 Figure 4.04 0.4 0. b/dn =1.02 0. b/dn =0.0. b/dn=1. b/dn =1.03 0. The maximum temperature at the solidfluid interface. 93 .0 Θmax(solid). b/dn =0. Θ int 0.14 0.08 ΘmaxΘmin(interface).1 0.08 Θmax(solid).02 0 0 0.05 0.26 shows the average Nusselt number variation as a function of time for three distinct plate thicknesses using Constantan as the solid material. and MaximumtoMinimum Temperature Difference with Time for Different Plate Thicknesses (Silicon Plate Re=750. maximum temperature within the solid.08 0.16 0.27. within the Solid.08 0. b/dn=0.04 0.2 0.25 Distribution of Dimensionless Maximum Temperature at the SolidFluid Interface.5 Fourier Number.09 Dimensionless MaximumtoMinimum Temperature Difference a the Interface.0 0. The average Nusselt number is higher for higher plate thickness.12 0.5) Figure 4.3 0. Fo Θmax(interface).Θ maxΘ min 0.08 0.2 0.0 ΘmaxΘmin(interface).
5) As expected. the temperature changes occur faster at the beginning of the heating process for all materials and the slope gradually decays when the steadystate approaches. The thermal diffusivity of the material also contributes to the transient behavior of the solid.01 0.02 0. Fo . As noticed.03 0.04 0. which controls the rate of heat being transferred through the solid material.08 .26 Variation of Average Nusselt Number with Time for Constantan Plate at Three Different Thicknesses (Re=750.08 b/dn=0.5 b/dn=1.5 Figure 4. b/dn=1.05 0. Nuav. Silicon and Copper reach the steadystate faster than Constantan due to their higher thermal diffusivity. 94 . Nuav 120 100 80 60 40 20 0 0 0. It can be observed that materials with a very low thermal conductivity such as Constantan maintains a higher temperature at the plate’s outer surface and within the solid as the thermal conductivity controls how effective the heat flows and distributes throughout the material.160 140 Average Nusselt Number. Nuav.06 Fourier Number. b/dn=0. ß=2.
5) Figure 4.1 0.2 0. Θ maxΘ min .35 0.0. Constantan 0. ß=2.25 0. Silicon Θmax(solid).02 0 0 0. A significant difference is seen at the earlier part of the transient and the curves come close together as the steadystate approaches.15 0.27 Distribution of Dimensionless Maximum Temperature at the SolidFluid interface.1 Dimensionless Maximum Temperature. Constantan shows a higher average heat transfer coefficient compared to Silicon or Copper over the entire transient process.08 0.04 0.4 Fourier Number. within the Solid.05 0. It will be also important to know how the materials responded in reaching thermal equilibrium based on their thickness.12 0.15 0. and MaximumtoMinimum Temperature Difference with Time for Different Materials (Silicon Plate.3 0. Silicon Θmax(interface).06 Θmax(interface).05 Θmax(solid).45 Figure 4. Constantan 0.28 plots the distribution of average Nusselt Number with time for the three materials used in this study.2 ΘmaxΘmin(interface). Fo 0 0. Copper Θmax(solid). Copper 0.25 0. Re=750. Constantan 0. Θ int 0. 95 Dimensionless MaximumtoMinimum Temperature Difference at the Interface.1 Θmax(interface).
006 0. Fo Nuav.003 0.009 0.=0.007 0. 96 . Also. Silicon Figure 4. the time to reach steadystate also increases. the property of the solid plays more significant role in determining the duration of the transient heat transfer process when the thickness is increased.5. Copper Nuav. Silicon .28 Variation of Average Nusselt Number with Time for Different Materials (Re=750. Nuav 120 100 80 60 40 20 0 0 0.29 presents Foss for these materials for different plate thicknesses.01 Fourier Number. As the thickness increases in value.160 140 Average Nusselt Number. Copper .5) Figure 4. ß=2.002 0. Constantan takes longer in reaching steadystate due to its lower thermal diffusivity compared to Copper and Silicon.004 0. Constantan .001 0. b/dn.008 0. Constantan Nuav.005 0.
b/dn Constantan Copper Silicon Figure 4.4 0.1 0 0 0.2 0.2 1.29 Time Needed to Reach SteadyState for Different Cylindrical Plate Thicknesses and for Different Materials (Re=750.8 0.6 Fourier Number.5) 97 .4 0.8 1 1.6 0.2 0.3 0. Foss 0.0.5 0.7 0.6 Dimensionless Plate Thickness. ß=2.4 1.
As the velocity increases. however. materials with higher thermal conductivity maintained a more uniform temperature distribution throughout the solidfluid interface and facilitated a higher heat transfer rate. The Nusselt number. did not change much with thickness variation. increasing the thickness of the hemispherical plate proved to decrease the solidfluid interface temperature due to the resistance of the material to heat flow. Also. A larger heat 98 . the boundary layer or film thickness decreases and Nusselt number increases over the entire solidfluid interface. On the other hand. lowering the maximum temperature inside the hemisphere and at its interface. The impingement height affected the dimensionless interface temperature as well as the Nusselt number. A lower thermal conductivity material showed higher local maximum Nusselt number as well as higher average Nusselt number among all studied materials. Temperature and heat flux at the solidfluid interface rise with time whereas the average heat transfer coefficient decreases with time. A higher jet impingement height provided a lower dimensionless interface temperature over the entire hemisphere and a higher Nusselt number at the stagnation region.Chapter 5 Discussion and Conclusion Local and average Nusselt number and heat transfer coefficient distributions showed a strong dependence on the impingement velocity or Reynolds number.
The maximum dimensionless interface temperature was encountered at the outer edge of the cylindrical plate while the minimum was encountered at the stagnation region. The time required to reach steadystate became larger as the thickness of the plate increased. 99 . Constantan demonstrated to have a higher average Nusselt number in comparison to Copper and Silicon. Local Nusselt number is high near the stagnation region where convective heat transfer rate is more effective and then it decays monotonically along the remaining portion of the plate.flux is seen near the stagnation region because of the larger temperature difference between the water jet and the plate. Also. the maximum temperature at the solidfluid interface as well as the temperature inside the solid decreased as the Reynolds number increased while the maximum to minimum temperature difference a the interface decreased as the Reynolds number decreased. Materials with a higher thermal conductivity maintained a lower dimensionless solidfluid interface temperature as well as dimensionless maximum temperature and reached steadystate faster. Increasing the plate thickness decreased the capacity of the plate of being effectively cooled (lower average Nusselt number) and decreased dimensionless maximum temperature at the interface and within the solid. The Reynolds number was found to be an essential parameter in controlling the transient process since the time required to reach steady state diminished as the Reynolds number increased. The isothermal lines within the solid demonstrated the transition from a conduction only at the very early part of the transient to conductionconvection equilibrium heat transfer as the steady state was arrived. Nevertheless.
heat transfer. 111.H.. “Jet impingement on semicylindirical concave and convex surfaces: part II . 18.” Experimental Thermal and Fluid Science....H. 117. 160169. 1995. 1997.. [5] Lee D. 113–119.S. pp. Dontaraju P. [8] Tong A. Chung Y. 6678. 11471156.M. “Local characteristics of impingement heat transfer with oblique round freesurface jets of large prandtl number liquid....” International Symposium on Physics of Heat Transfer in Boiling and Condensation. 1999. 42.S. [4] Gomi T. and Goldstein R.. pp. and Rice R. 22492259. pp. 44. singlephase liquid jet.” International Journal of Heat and Mass Transfer. pp. 2003. “Turbulent heat transfer from a convex hemispherical surface to a round impinging jet. [6] Kornblum Y. Goldstein R. 119.” Journal of Heat Transfer. 40 (10).” Numerical Heat Transfer.A. [9] Cornaro C..W.. 2002. “Confined jet impingement thermal management using liquid ammonia as the working fluid..” In: Proceedings of ASME IMECE2002..S. 1989. Louisiana 33033. and Kim D.. pp. and Webb B. New Orleans.” International Journal of Heat and Mass Transfer. 1997. and Webb B. pp.. pp. 1997.V. Chung Y. “A numerical study on the hydrodynamics and heat transfer of a circular liquid jet impinging onto a substrate.” International Journal of Heat and Fluid Flow..J.” Heat Transfer in Electronics. 597602. November 1722. and Ponnappan R.S.J.G.W.Y.References [1] Rahman M. “Confined and submerged liquid jet impingement heat transfer... [7] Lee D. [2] Stevens J. 1999. “Turbulent flow and heat transfer measurements on a curved surface with a fully developed round impinging jet. 871877. and Kim M. Fleischer A. pp.. 100 . 20.. [3] Garimella S. “Flow visualization of a round jet impinging on cylindrical surfaces.. “Local heat transfer coefficients under an axisymmetric.
[10] Cornaro C., Fleischer A.S., Rounds M., and Goldstein R.J., 2001,“Jet impingement cooling of a convex semicylindrical surface,” Journal of Thermal Science, 40, pp. 890898. [11] Fleischer A.S., Kramer K., and Goldstein R.J., 2001, “Dynamics of the vortex structure of a jet impinging on a convex surface,” Experimental Thermal and Fluid Science, 24, pp. 169175. [12] Baonga J. B., LouahliaGualous H., and Imbert M., 2006, “Experimental study of the hydrodynamic and heat transfer of free liquid jet impinging a flat circular heated disk,” Applied Thermal Engineering, 26, pp. 11251138. [13] Inada S., Miyasaka Y., and Izumi R., 1981, “A study on the laminarflow heat transfer between a twodimensional water jet and a flat surface with constant heat flux,” In: Bulletin of the JSME─1982 JSME 24. [14] Carper H.J., Jr., 1989, “Impingement cooling by liquid jet,” In: Heat Transfer Division─1989 ASME HTD 117. [15] Liu X., and Lienhard V., 1989, “Liquid jet impingement heat transfer on a uniform flux surface,” In: Heat Transfer Phenomena in Radiation, Combustion and Fires─1989 ASME HTD 106. [16] Wadsworth D.C., and Mudawar I., 1989, “Cooling of a multichip electronic module by means of confined twodimensional jets of dielectric liquid,” Heat Transfer in Electronics, 111, pp. 7987. [17] Gau C., and Chung C.M., 1991, “Surface curvature effect on slotairjet impingement cooling flow and heat transfer process,” Journal of Heat Transfer, 113, pp. 858864. [18] Stevens J., and Webb B.W., 1992, “Measurements of the free surface flow structure under an impinging, free liquid jet,” Journal of Heat Transfer, 114, pp. 7984. [19] Teuscher K.L., Ramadhyani S., and Incropera F.P., 1993, “Jet impingement cooling of an array of discrete heat sources with extended surfaces,” ASME HTD, 263, pp. 110. [20] Ma C.F., Zhuang Y., Lee S.C., and Gomi T., 1997, “Impingement heat transfer and recovery effect with submerged jets of large Prandtl number liquid—II. Initially laminar confined slot jets,” International Journal of Heat and Mass Transfer, 40, pp. 14911500.
101
[21] Bartoli C., and Faggiani S., 1998, “Local nusselt number at a cylinder cooled by a slot jet of water,” Heat and Technology, 16 (2), pp.3337. [22] McDaniel C.S., and Webb B.W., 2000, “Slot jet impingement heat transfer from circular cylinders,” International Journal of Heat and Mass Transfer, 43, pp. 19751985. [23] Gori F., and Bossi L., 2000, “On the cooling effect of an air jet along the surface of a cylinder,” International Communications in Heat and Mass Transfer, 27, pp. 667676. [24] Kayansayan N., and Küçüka S., 2001, “Impingement cooling of a semicylindrical concave channel by confined slotairjet,” Experimental Thermal and Fluid Science, 25, pp. 383396. [25] Shi Y.L., Ray M.B., and Mujumdar A.S., 2003, “Effects of prandtl number on impinging jet heat transfer under a semiconfined laminar slot jet,” International Communications Heat Mass Transfer, 30 (4), pp. 455464. [26] Olsson E.E.M., Ahrné L.M., and Trägårdh A.C., 2004, “Heat transfer from a slot air jet impinging on a circular cylinder,” Journal of Food Engineering, 63, pp. 393401. [27] Chan T. L., Leung C.W., Jambunathan K., AshforthFrost S., Zhou Y., and Liu M.H., 2002, “Heat transfer characteristics of a slot jet impinging on a semicircular convex surface,” International Journal of Heat and Mass Transfer, 45, pp. 9931006. [28] Chan T.L., Zhou Y., Liu M.H., and C.W. Leung, 2003, “Mean flow and turbulence measurements of the impingement wall jet on a semicircular convex surface,” Experiments in Fluids, 34, pp. 140–149. [29] Gori F., and Bossi L., 2003, “Optimal slot height in the jet cooling of a circular cylinder,” Applied Thermal Engineering, 23, pp. 859870. [30] Rahimi M., Owen I., and Mistry J., 2003, “Heat transfer between an underexpanded jet and a cylindrical surface,” International Journal of Heat and Mass Transfer, 46, pp. 31353142. [31] Yang Y.T., and Hwang C.H., 2004, “Numerical simulations on the hydrodynamics of a turbulent slot jet impinging on a semicylindrical convex surface, Numerical Heat Transfer, 46, pp. 9951008.
102
[32] Chen Y.C., Ma C.F., Qin, M., and Li, Y.X., 2005, “Theoretical study on impingement heat transfer with singlephase freesurface slot jets,” International Journal of Heat and Mass Transfer, 48, pp. 33813386. [33] Zuckerman N., and Lior N., 2005, “Jet impingement heat transfer on a circular cylinder by radial slot jets,” In: Proceedings of ASME IMECE2005, November 511, Orlando, Florida USA 79565. [34] White F. M., 2003, “Fluid mechanics 5th ed.,” McGrawHill, New York. [35] Stevens J., and Webb B.W., 1992, “Measurements of the free surface flow structure under an impinging, free liquid jet,” Journal of Heat Transfer, 114, pp. 7984. [36] Liu X., Lienhard J.H., and Lombara J.S., 1993, “Convective heat transfer by impingement of circular liquid jets,” Journal of Heat Transfer, 13, pp. 571–582. [37] Scholtz M.T., and Trass O., 1970, “Mass transfer in a nonuniform impinging jet,” AIChE Journal, pp. 8296. [38] Nakoryakov V.E., Pokusaev B.G., and Troyan E.N., 1978, “Impingement of an axisymmetric liquid jet on a barrier, International Journal of Heat and Mass Transfer, 9, pp. 11751184. [39] Martin H., 1977, “Heat and mass transfer between impinging gas jets and solid surfaces,” Advances in Heat Transfer, 13, pp. 160. [40] Whitaker S., 1972, “Forced convection heat transfer correlations for flow in pipes, past flat plates, single cylinders, single spheres, and flow in packed beds and tube bundles,” AIChE Journal, 18, pp. 361371.
103
1995. Özisik M.” McGrawHill. 1993.” John Wiley & Sons. New York. M... 2003...Bibliography White F. New York.” John Wiley & Sons. Bejan A. New York.. “Heat conduction 2nd ed.N.. 104 . “Convection heat transfer 2nd ed. “Fluid mechanics 5th ed.
Appendices 105 .
000000 0. LINE ) POINT( SELE.0001 ) WINDOW(CHANGE= 1. 5 3 CURVE( ADD. BEDG = 1. ID ) 5. 0 0. MFAC = 1. POIN = 1.50000 7. 3 CURVE( ADD.86.Appendix A: CFD Code for Axisymetric Model (FIDAP) / File opened for write Mon Oct 2 16:06:01 2006.000000 0. SURF = 1. 9 1. ARC.50000 7.36.06 0. ID ) 4. CURV = 1.86. 0 0. / File opened for write Thu Jun 15 16:15:56 2006.000000 0. CENT ) POINT( SELE.06 0.000000 0.000000 0.00000 7. SPAV = 1.000000 10. 0.000000 0. 0. COOR = 1. ARC.000000 0.86. NODE = 0.000000 0.00000 10. 0 0.50000 //POINTS POINT( ADD.59 //LINES POINT( SELE. ID ) 10 7 106 . ID ) 8. ID ) 4 7.5 0.000000 1. CENT ) POINT( SELE.000000 1. CENT ) POINT( SELE. MLOO = 1. ID ) 4 6 2 CURVE( ADD. 0.000000 0.000000 0.56 0. MATRIX ) 1.273. MEDG = 1. 0. LINE ) //CREATE SURFACE POINT( SELE. ARC.3.86. MSHE = 1.000000 0. 7 CURVE( ADD. COOR ) 0. TITLE( ) FREE SURFACE JET ON CYLINDER FIGEN( ELEM = 1. 8 CURVE( ADD.000000 1. MSOL = 1.50000 7. 0.59 0. 0 0.000000 0. 0. TOLE = 0.
MAP.Appendix A: (Continued) 1 4 SURFACE( ADD. NODE = 2 ) MEDGE( SELE. ID = 2 ) MFACE( MESH. EDG1 = 1. VISI. INTE = 70. POIN. SUCC. EDG3 = 1. CURVE( SELE. QUAD. ID = 2 ) MFACE( ADD ) //MESH FACE ENTITIES ELEMENT( SETD. /LOOP 2 CURVE( SELE. ENTI = "fluid" ) MFACE( SELE. CURVE( SELE. NOSH. EDG1 = 1. PCEN = 0 ) 2RAT = 0. PCEN = 0 ) 2RAT = 0. INTE = 20. //MESH FACES /FACE 1 SURFACE( SELE. INTE = 62. NOAD ) //MESH EDGES CURVE( SELE. ID = 5 ) MEDGE( ADD. ID ) 7 2 8 1 MLOOP( ADD. RATI = 0. 3 MEDGE( ADD. EDG4 = 2 ) EDG2 = 1. MAP. CURVE( SELE. ID = 1 ) MLOOP( SELE. NODE = 4 ) MFACE( SELE. MAP. SUCC. INTE = 70. RATI = 0. EDG3 = 1. RATI = 0. INTE = 20. ID = 9 ) MEDGE( ADD. PCEN = 0 ) 2RAT = 0. CURVE( SELE. ID = 4 ) MEDGE( ADD. ID ) 5 4 3 9 2 6 MLOOP( ADD. CURVE( SELE. ID = 1 ) MLOOP( SELE. ID = 1 ) 2RAT = 0. EDGE. 8 MEDGE( ADD. VISI. SUCC. INTE = 20. EDG4 = 1 ) 107 . ID ) 7. ID = 6 ) MEDGE( ADD. SUCC. //MESH LOOPS /LOOP 1 CURVE( SELE. PCEN = 0 ) 2RAT = 0. ID = 1 ) MFACE( MESH. ROWW = 2. ENTI = "solid" ) //MESH EDGE ENTITIES ELEMENT( SETD. RATI = 0. RATI = 0. ID = 1 ) MFACE( ADD ) /FACE 2 SURFACE( SELE. SUCC. RATI = 0. NOSH. SUCC. PCEN = 0 ) 2RAT = 0. PCEN = 0 ) EDG2 = 2. MAP. ID ) 1.
NAME = "bottom". MAP. SET = "water". ID = 2 ) MEDGE( MESH. ZERO ) BCNODE( ADD. UY. ID ) 3. ENTI = "free" ) MEDGE( SELE. PLOT ) ENTITY( ADD. PLOT ) ENTITY( ADD. CONS = 73 ) / /SOLID DENSITY( ADD. ZERO ) BCNODE( ADD. ID = 8 ) MEDGE( MESH. ENTI = "syme" ) MEDGE( SELE. STRA. ZERO ) 108 . PLOT ) ENTITY( ADD. MAP. ENTI = "interface" ) MEDGE( SELE. NAME = "interface". ENTI = "swall". SET = "water". ENTI = "interface". MAP. VELO. ENTI = "finlet". UX. UY. CONS = 33. ID = 9 ) MEDGE( MESH. CONS = 2. ANG1 = 80. SPIN. NODE = 190 ) BCNODE( ADD. ENTI = "syme". NODE = 190. ID = 5 ) MEDGE( MESH.38353 ) BCNODE( ADD.12 /FLUID DENSITY( ADD. DEPT = 21. SURF. ENTI = "finlet". PROP = "silicon" ) ENTITY( ADD. SET = "silicon".00798 ) SPECIFICHEAT( ADD. MAP. SET = "water". SET = "water". ENTI = "swall" ) MEDGE( SELE. NAME = "fluid".334608 ) SPECIFICHEAT( ADD. MAP.998137 ) SURFACETENSION( ADD. NAME = "solid".Appendix A: (Continued) MEDGE( MESH. MAP. CONS = 0. ENTI = "bottom" ) MEDGE( SELE. FLUI. VELO. ENTI = "finlet" ) MEDGE( SELE. VELO. PLOT ) ENTITY( ADD. NAME = "swall". CONS = 0. SURF. NATT = "fluid" ) ENTITY( ADD. CONS = 0. PLOT ) ENTITY( ADD. ENTI = "fout" ) END( ) FIPREP( ) //Fluid and solid properties /B=2. NAME = "fout". D=0.996 ) CONDUCTIVITY( ADD. COOR. ENTI = "bottom".0014699 ) VISCOSITY( ADD. ATTA = "solid". PLOT.17006 ) / /ENTITIES ENTITY( ADD. ZERO ) BCNODE( ADD. NAME = "free". CONS = 0. SOLI.5. ZERO ) BCNODE( ADD. ANG2 = 180 ) / /SPECIFY BOUNDARY CONDITIONS BCNODE( ADD. CONS = 0. ATTA = "fluid". SET = "silicon". NAME = "syme". SET = "silicon". 7 MEDGE( MESH. MAP. PROP = "water" ) ENTITY( ADD. ID ) 6. ZERO ) BCNODE( ADD. CONS = 0. NAME = "finlet".33 ) CONDUCTIVITY( ADD. SET = "water". 4 MEDGE( MESH.
0001. NONL. CONS = 23. FY = 0.000000 1.000000 0.00000 0.000000 1.000000 0. NEWJ ) PRINTOUT( ADD. UY. ENTI = "fluid".38353 ) ICNODE( ADD. CONT ) EXECUTION( ADD.430000 0. SING ) BODYFORCE( ADD. N. SURF = 0. DT = 1e07.000000 REDO WINDOW(CHANGE= 1.0001.000000 1. ENTI = "fluid". TEMP.Appendix A: (Continued) BCNODE( ADD. MATRIX ) 1.000000 109 . CONS = 5. ZERO ) / /PROBLEM DEFINITION PROBLEM( 2D. MIXE = 1e11. ENER.000000 0.000000 0.001 ) TIMEINTEGRATION( ADD.000000 45.000000 0. 1 / /INITIAL CONDITIONS ICNODE( ADD. UPWI ) UPWINDING( ADD. NONE ) OPTIONS( ADD. HEAT.37148 0. LAMI. MATRIX ) 1. FIPOST( ) TIMESTEP( STEP = 1 ) TIMESTEP( STEP = 322 ) CONVERGENCE( ALL.38353 ) ICNODE( ADD. BACK ) / File closed at Mon Oct 2 16:06:12 2006. NSTE = 1000.R. WIND = 9. ENTI = "bottom". VARI. 1000.000000 0.000000 0.000000 0. VELO. ENTI = "solid". TSTA = 0. TEMP.971 ) BCNODE( ADD.000000 0. = 50. FX = 981. MOME.000000 0. TRAN. UX. RESC = 0. / File opened for append Tue Oct 3 12:46:08 2006.00000 45.15563 0. CONS = 37 ) BCFLUX( ADD.000000 0. FZ = 0 ) PRESSURE( ADD.000000 0. DISC ) DATAPRINT( ADD.36300 0. KINE = 25. SOLU.000000 0.000000 0.000000 0. ENTI = "finlet". NOFI = 10 ) POSTPROCESS( NBLO = 2 ) 1.000000 0. BACK. ENTI = "fluid". LOG ) VECTOR( VELO ) WINDOW(CHANGE= 1. FREE.000000 0.000000 0. CONS = 37 ) END( ) CREATE( FISO ) RUN( FISOLV. 1 201.000000 0.000000 1.000000 0.000000 45. CONS. 201.430000 0.17218 0. VELC = 0.000000 1. CONS = 18. STRE ) / /SOLUTION ALGORITHM SOLUTION( ADD.000000 0.000000 45.295000 0.000000 1.295000 0. NEWT.
295000 0.00000 0.000000 0.000000 0.000000 0.14089 45.000000 0.000000 0.295000 0. MATRIX ) 0.000000 0.000000 0.000000 0.000000 0.00000 45.000000 0.00000 45. MATRIX ) 0.430000 0.000000 0.00000 45.000000 1.000000 REDO WINDOW(CHANGE= 1.000000 0.00000 0.000000 0.295000 0.000000 0.000000 0.000000 0.000000 0.000000 REDO WINDOW(CHANGE= 1. MATRIX ) 0. MATRIX ) 0.000000 0.000000 0.000000 1.000000 0.000000 0.000000 1.000000 1.000000 0.000000 0.000000 0.000000 0.430000 1.000000 0.000000 0.000000 0.00000 45.000000 1.000000 0.00000 0.000000 0.04771 45.00000 45.000000 0.000000 0.000000 1.000000 0.000000 0.295000 0.000000 1.000000 0.33551 45.000000 1.000000 REDO WINDOW(CHANGE= 1.000000 0.30826 45.31361 45.00000 0.000000 1.000000 0.000000 0.000000 1.000000 0.000000 0.000000 0.00000 45.36300 45.295000 0.000000 0.000000 0.00000 45.29887 45.000000 0.000000 REDO WINDOW(CHANGE= 1.000000 0.000000 0.000000 0.00000 45.000000 0.000000 1.430000 0.00000 45.000000 0.000000 0.000000 0.42935 0.000000 0.000000 0.000000 1.00000 45.000000 0.00000 0.000000 0.000000 0.000000 0.000000 1.000000 0.000000 0.000000 1.000000 0.000000 0.295000 0.430000 0.000000 0.00000 0.000000 1.000000 0.000000 0.Appendix A: (Continued) WINDOW(CHANGE= 1.20972 45.20659 45.29797 0.000000 1.000000 1.430000 0.43717 0.000000 REDO WINDOW(CHANGE= 1. MATRIX ) 0.000000 0.000000 0.000000 1.00000 45.000000 0.430000 0.000000 0.000000 0.000000 110 .000000 0.000000 1.000000 0.000000 1. MATRIX ) 0.000000 1.000000 0.000000 1.000000 0.000000 0.00000 0.000000 0.000000 1.
295000 0.00000 45.35024 45.000000 0.41840 45.000000 0. WINDOW(CHANGE= 1.000000 0.000000 1.00000 45.000000 0.000000 0.000000 0.00000 45.000000 0.00000 0.000000 0.000000 0.00000 45.000000 0.000000 0.000000 0.00000 0.000000 0.295000 0.000000 0.00000 45.430000 0.000000 1.000000 0.000000 0.295000 0.000000 1.000000 0. MATRIX ) 0.000000 0.000000 1.000000 0.000000 AUTO = 200 ) 1.000000 0.000000 0.000000 REDO CONTOUR( TEMP.000000 0.00000 45.000000 0.000000 0.000000 0.000000 0.00000 45.000000 0.000000 0.000000 0.000000 0.000000 0.00000 0.31204 0.000000 0. MATRIX ) 0.000000 0.000000 REDO WINDOW(CHANGE= 1.000000 0.000000 0.00000 45.000000 111 .000000 0.000000 0.000000 0.000000 1. MATRIX ) 0.000000 0.000000 0.295000 0.000000 0.000000 1.000000 0.00000 45.00000 0.295000 0.000000 0.430000 1.000000 0.000000 0.000000 1.000000 0.000000 REDO WINDOW(CHANGE= 1.000000 0.430000 0.00000 45.000000 0.000000 AUTO ) 1.000000 REDO CONTOUR( TEMP.000000 1.000000 0.000000 0.000000 1.000000 0.000000 1.000000 0.52320 45.000000 1.00000 45.00000 45.000000 0.Appendix A: (Continued) WINDOW(CHANGE= 1.000000 0.430000 0.000000 REDO WINDOW(CHANGE= 1.000000 0.000000 0.000000 0.000000 1.16255 45.00000 0.000000 1.000000 0.33460 45.11340 0.000000 1.430000 0.000000 0.430000 0.000000 1.00000 0.000000 0.000000 0.000000 0.00000 0.000000 0.00000 0.00000 45. MESH( NNUM ) WINDOW(CHANGE= 1.000000 0.000000 0.000000 1. MATRIX ) 0.000000 0.00000 45.000000 0.000000 0.000000 0.000000 0.06401 45.295000 0.000000 0.000000 0.000000 1. MATRIX ) 0.000000 1.000000 0.000000 1.000000 0.000000 1.000000 0.000000 1. MATRIX ) 0.
NOD1 = 349.000000 45.000000 1.000000 0.000000 1.26914 0. ENTI = "interface".430000 0. NOD2 = 4053.19004 0.000000 45.000000 0.000000 0. SCRE ) PRINT( TEMP. MATRIX ) 1.000000 1.00000 0.295000 0.000000 0. NOD1 = 386.000000 0.000000 0.000000 0.000000 45.000000 1.000000 0.000000 1.000000 0.000000 0.000000 0.000000 1.00000 0. AUTO = 200 ) LINE( TEMP.Appendix A: (Continued) 0.000000 0.295000 0.000000 45.00000 45.00000 45.000000 0.13912 0.000000 45.52477 0. FIPOST( ) TIMESTEP( STEP = 1 ) TIMESTEP( STEP = 322 ) VECTOR( VELO ) GROUP( ENTI = "fluid" ) VECTOR( VELO ) WINDOW(CHANGE= 1.000000 0.000000 1.000000 0.000000 0.000000 0.000000 45. MATRIX ) 1.000000 0.00000 0.00000 0.35049 0. NOD1 = 2913.000000 0.29640 0.000000 0. ENTI = "interface" ) PRINT( TEMP. MATRIX ) 1.000000 1. NGEN = 19.000000 45.000000 0. / File opened for append Wed Oct 4 19:38:28 2006.17687 0.00000 0.000000 0.000000 0. SCRE ) PRINT( TEMP.000000 45.000000 0.000000 112 .430000 0. NODE.295000 0.000000 0. SCRE ) PRINT( TEMP.000000 1.30546 0.000000 0.00000 0. NODE.000000 0.000000 0.00547 0.000000 45.000000 REDO CONTOUR( TEMP.000000 0.000000 0.000000 0.00000 0.000000 0.000000 0.000000 REDO WINDOW(CHANGE= 1.000000 0. MATRIX ) 1.000000 0.000000 0.35650 0.000000 45.000000 REDO WINDOW(CHANGE= 1.00000 45.000000 0.09816 0.000000 1.00000 45.15316 0.000000 0.000000 45.430000 0.000000 0.000000 45.000000 1.000000 0.00000 0. NODE.000000 0. SCRE ) END( ) END( ) / File closed at Tue Oct 3 12:52:48 2006.000000 0.000000 REDO WINDOW(CHANGE= 1.
295000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 45.000000 REDO WINDOW(CHANGE= 1.00000 0.00000 0. AUTO ) WINDOW(CHANGE= 1.000000 REDO CONTOUR( STRE.000000 0.000000 0.000000 0.00000 0.12344 0.000000 1.000000 0.000000 0.000000 0.000000 0.295000 0.00000 0.000000 0.000000 0.000000 1.000000 45.000000 0. MATRIX ) 1.000000 0.000000 0.00000 0.000000 45.295000 0.00000 0.000000 REDO CONTOUR( TEMP.17819 0.000000 1.00000 113 .000000 0.000000 0.000000 0.00000 45.000000 0.000000 0.000000 0.34868 0.430000 0.000000 REDO WINDOW(CHANGE= 1.00000 45.000000 45.000000 0.000000 1.000000 0. AUTO ) CONTOUR( TEMP.000000 45.000000 1.000000 1.000000 0.000000 0.430000 0.430000 0.000000 45.000000 45.00000 0.000000 45.Appendix A: (Continued) 45.000000 0.000000 0. 0.000000 1. AUTO ) END( ) END( ) / File closed at Wed Oct 4 19:48:27 2006.000000 0.000000 1.00000 0.000000 REDO VECTOR( VELO ) DEVICE( POST.000000 0.000000 45.48723 0.000000 1.000000 45. MATRIX ) 1. AUTO ) WINDOW(CHANGE= 1.000000 45. MATRIX ) 1.000000 0.00000 0.000000 0.35650 0.000000 0.000000 0.00000 45.000000 0.00000 45.000000 45.000000 1. MATRIX ) 1.000000 45.23227 0.000000 0.00000 0.000000 0.47315 0.295000 0.000000 45.000000 0.000000 45.000000 0.430000 0.000000 0. FILE = "500" ) VECTOR( VELO ) CONTOUR( STRE.000000 0.000000 0.00000 0.000000 1.000000 1.27138 0.
X = 0.5. COOR. TOLE = 1e05 ) WINDOW(CHANGE= 1. X = 0. Y = 0. X = 0. COOR. Y = 0. NODE = 0. SURF = 1.000000 1. 10.000000 0. Y = 0 ) POINT( ADD.5 ) POINT( ADD. COOR = 1. 0. Y = 0. X = 0. MATRIX ) 1. X = 0. X = 0. COOR. COOR. Y = 0.5. X = 0.000000 0. 7. COOR.06 ) POINT( ADD. X = 0. COOR. 0 0. Y = 0.16 ) POINT( ADD. X = 0. Y = 0. COOR. COOR.52 ) POINT( ADD. X = 0.112.359 ) POINT( ADD. COOR.66 ) POINT( ADD. Y = 0. 7. // File opened for write Wed Mar 1 02:13:54 2006. 7.4) POINT( SELE.3. MATR ) 1.62 ) POINT( ADD. MLOO = 1. X = 0. COOR. Y = 0 ) //LINES (1.0033. X = 0.299. COOR.2 ) POINT( ADD. Y = 0 ) POINT( ADD. MSHE = 1.27.000000 0.5.1425 ) POINT( ADD. X = 0. 0. X = 0. X = 0. Y = 0.Appendix B: CFD Code for 2D Model / File opened for write Fri Sep 22 20:31:22 2006.399 ) POINT( ADD.1516.2. X = 0.041.06 ) POINT( ADD. COOR. MFAC = 1. BEDG = 1.50000 7.66 ) POINT( ADD.5. COOR. LINE ) 114 .5 7. COOR.06. Y = 0. 0 0.000000 1. Y = 0.000000 0.56 ) POINT( ADD. Y = 0 ) POINT( ADD. X = 0.50000 WINDOW( CHAN = 1.3. Y = 0.50000 7.2587 ) POINT( ADD. COOR. 0 0.466 ) POINT( ADD. COOR.5.3.000000 0. Y = 0. Y = 0. 0.000000 0.000000 0. SPAV = 1. MSOL = 1.000000 1. Y = 0. COOR.16.000000 0. MEDG = 1. COOR. CURV = 1. ID = 1 ) POINT( SELE. COOR. Y = 0 ) POINT( ADD. COOR.0036.000000 0.47 ) POINT( ADD.0672 ) POINT( ADD. 1.0596.000000 0. X = 0.31 ) POINT( ADD. X = 0. X = 0. 0.00000 7. COOR.319. COOR.1. 0. Y = 0.000000 0. Y = 0.00000 10. X = 0. Y = 0. X = 0.03.50000 7.5. 0. TITLE( ) FREE SURFACE JET IMPINGMENT FIGEN( ELEM = 1. Y = 0.3. COOR. X = 0.231 ) POINT( ADD. COOR. 1.000000 10.0565.000000 0. POIN = 1.06.021. 1 10.5 //POINTS POINT( ADD. X = 0. Y = 0. ID = 2 ) CURVE( ADD.
ID = 7 ) CURVE( SPLI ) CURVE( SELE. ARC ) CURVE( SELE. ID = 5 ) CURVE( SPLI ) //LINES (7. ID ) 11 12 13 CURVE( ADD.Appendix B: (Continued) POINT( SELE. ID = 14 ) CURVE( SPLI ) //LINES (14. ID = 13 ) POINT( SELE. ID = 7 ) POINT( SELE. ID = 8 ) CURVE( ADD. ID = 11 ) CURVE( ADD. ARC ) //LINES (10. ARC ) 115 . ID = 9 ) POINT( SELE. ID = 6 ) CURVE( ADD. ARC ) POINT( SELE.9) POINT( SELE. LINE ) POINT( SELE. ID = 7 ) POINT( SELE. ID = 9 ) CURVE( ADD. LINE ) POINT( SELE. ID = 4 ) CURVE( ADD.13) POINT( SELE. ID = 4 ) POINT( SELE. ID = 10 ) POINT( SELE.12. ID = 7 ) POINT( SELE. ARC ) //LINES (5. ID = 6 ) POINT( SELE.17. ID = 3 ) POINT( SELE.11.15. ARC ) CURVE( SELE. LINE ) POINT( SELE.8. ID = 8 ) POINT( SELE. ID ) 15 16 1 CURVE( ADD.16.6) CURVE( SELE.18) POINT( SELE. ID = 4 ) POINT( SELE. ID ) 13 14 15 CURVE( ADD. ID = 5 ) POINT( SELE. ID = 9 ) CURVE( DELE ) POINT( SELE. ID = 10 ) CURVE( ADD.
ID = 20 ) CURVE( SPLI ) POINT( SELE. POIN. SUCC.20. PCEN = 0 ) 30. 2RAT = 0. ID ) 2 18 19 CURVE( ADD. PCEN = 0 ) 16. RATI = 0. ID ) 21 22 10 CURVE( ADD. PCEN = 0 ) 30. INTE = CURVE( SELE. 2RAT = 0. PCEN = 0 ) 44. PCEN = 0 ) 22. NOAD ) 10.22) POINT( SELE. ID = 9 ) POINT( SELE. ID = 23 ) SURFACE( ADD. ID = 22 ) POINT( SELE. SUCC. SUCC.25. RATI = 0.26) CURVE( SELE. ID = 8 ) MEDGE( ADD. ID = 1 ) MEDGE( ADD. RATI = 0. ID = 10 ) MEDGE( ADD. INTE = CURVE( SELE. ID = 18 ) CURVE( SPLI ) POINT( SELE. INTE = CURVE( SELE. PCEN = 0 ) 116 . SUCC. ARC ) POINT( SELE. ROWW //MESH EDGES CURVE( SELE. ID = 5 ) MEDGE( ADD. INTE = CURVE( SELE. ARC ) //LINES (23. ID = 9 ) MEDGE( ADD. 2RAT = 0. INTE = CURVE( SELE. SUCC. LINE ) /SURFACES POINT( SELE. 2RAT = 0.24. SUCC. RATI = 0. ID = 11 ) = 2. 2RAT = 0. PCEN = 0 ) 15. ID = 3 ) MEDGE( ADD.Appendix B: (Continued) //LINES (19. INTE = CURVE( SELE. PCEN = 0 ) 22. ID ) 19 20 21 CURVE( ADD. ARC ) CURVE( SELE. RATI = 0. 2RAT = 0.21. ID = 6 ) MEDGE( ADD. ID = 19 ) POINT( SELE. RATI = 0. SUCC. INTE = CURVE( SELE. RATI = 0. SUCC. ID = 24 ) POINT( SELE. INTE = CURVE( SELE. 2RAT = 0. ID ) 20 6 CURVE( ADD. 2RAT = 0. ID = 2 ) MEDGE( ADD. RATI = 0. ID = 3 ) POINT( SELE.
ID = 17 ) ADD. 2RAT = 0. PCEN = 0 ) RATI = 0. ID = 26 ) SELE. ID = 24 ) ADD. VISI. 2RAT = 0. ID = 21 ) SELE. 2RAT = 0. ID = 5 ) SELE. PCEN = 0 ) RATI = 0. ID = 9 ) SELE. ID = 25 ) SELE. EDG2 = 2. EDG3 = 1. PCEN = 0 ) RATI = 0. ID = 3 ) SELE. 2RAT = 0. ID = 20 ) SELE. 2 SELE. ID = 25 ) SELE. ID = 20 ) ADD. SELE. PCEN = 0 ) EDG1 = 1.Appendix B: (Continued) CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( CURVE( MEDGE( //MESH //LOOP CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( MLOOP( //LOOP CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( MLOOP( //LOOP CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( CURVE( SELE. ID = 12 ) SELE. NOSH. PCEN = 0 ) RATI = 0. EDG3 = 1. ID = 14 ) ADD. ID = 25 ) ADD. INTE = 44. ID = 23 ) SELE. ID = 26 ) SELE. 2RAT = 0. ID = 1 ) SELE. EDG4 = 3 ) EDG1 = 1. SELE. SUCC. PCEN = 0 ) RATI = 0. SUCC. MAP. INTE = 15. ID = 21 ) ADD. ID = 23 ) SELE. ID = 26 ) ADD. PCEN = 0 ) RATI = 0. INTE = 22. INTE = 15. ID = 18 ) ADD. ID = 18 ) 117 . PCEN = 0 ) RATI = 0. ID = 6 ) SELE. ID = 20 ) ADD. ID = 15 ) SELE. EDG2 = 3. SUCC. ID = 21 ) SELE. INTE = 30. INTE = 22. LOOPS 1 SELE. PCEN = 0 ) RATI = 0. SELE. 2RAT = 0. VISI. INTE = 16. ID = 14 ) SELE. 2RAT = 0. 2RAT = 0. ID = 24 ) SELE. SELE. 2RAT = 0. NOSH. ID = 24 ) ADD. 2RAT = 0. ID = 15 ) ADD. ID = 8 ) SELE. INTE = 16. MAP. SUCC. ID = 2 ) SELE. SUCC. SUCC. SELE. SELE. ID = 23 ) ADD. PCEN = 0 ) RATI = 0. SUCC. ID = 11 ) SELE. SELE. INTE = 22. SUCC. SUCC. SELE. EDG4 = 2 ) SELE. 3 SELE. ID = 17 )CURVE( RATI = 0. SELE. SUCC. ID = 10 ) SELE. INTE = 22.
ID = 3 ) MFACE( ADD ) // MESHING MFACE( SELE. NODE = 2 ) MEDGE( SELE. ID = 1 ) MLOOP( SELE. ID = 8 ) MEDGE( MESH. ID ) 4 5 6 7 MEDGE( MESH. ID = 2 ) ELEMENT( SETD. QUAD. ENTI = "water" ) //MESH MAP ELEMENT ID ELEMENT( SETD. ENTI = "interface" ) END( ) FIPREP( ) //Fluid and solid properties DENSITY( ADD. MAP. SET = "water". MAP. ENTI = "bottom" ) MEDGE( SELE. MAP. NODE = 4 ) MFACE( MESH. ID ) 10 11 12 13 14 MEDGE( MESH. ID = 1 ) MLOOP( SELE. ID = 1 ) MFACE( SELE. ENTI = "sides" ) MEDGE( SELE. ID = 3 ) MEDGE( MESH. ENTI = "axis" ) MEDGE( SELE. EDGE.Appendix B: (Continued) SURFACE( SELE.996 ) 118 . MAP. QUAD. NOSM. ENTI = "inlet" ) MEDGE( SELE. ID = 1 ) MFACE( ADD ) SURFACE( SELE. MAP. MAP. ID = 3 ) ELEMENT( SETD. ID = 2 ) MFACE( ADD ) SURFACE( SELE. CONS = 0. ID = 2 ) MEDGE( MESH. MAP. NODE = 4 ) MFACE( MESH. ENTI = "outlet" ) MEDGE( SELE. ID = 1 ) MLOOP( SELE. ENTI = "silicon" ) MFACE( SELE. ID ) 1 9 MEDGE( MESH. MAP. MAP. ID ) 15 16 17 18 19 MEDGE( MESH. ENTI = "surface" ) MEDGE( SELE.
= 50. STRE ) RELAXATION( ) 0. NEWT.Appendix B: (Continued) CONDUCTIVITY( ADD. NAME = "inlet". 1000. NAME = "bottom".00798 ) SPECIFICHEAT( ADD. NAME = "surface". ZERO ) BCNODE( ADD.05. SPIN. FX = 981. COOR. ZERO ) BCNODE( ADD. SET = "water". PLOT ) ENTITY( ADD. 1e20. URC.334608 ) SPECIFICHEAT( ADD. PROP = "water" ) ENTITY( ADD. PLOT. ENTI = "bottom". SET = "silicon". DISC ) DATAPRINT( ADD. 0. VELO. NAME = "axis". SOLI. NAME = "sides". TEMP. CONS = 0. NATT = "water" ) BODYFORCE( ADD. 37. FY = 0. N. TSTA = 0.0001. NAME = "water".1 BCNODE( ADD. SURF. BOUN ) OPTIONS( ADD. NONL. URC.998137 ) SURFACETENSION( ADD. NODE = 42 ) BCNODE( ADD. CONS = 20 ) 119 . CYLI. CONS = 0. NOFI = 12 ) POSTPROCESS( NBLO = 2 ) 1. NODE = 42. CONS = 5. ENTI = "axis".971 ) BCNODE( ADD.001 ) TIMEINTEGRATION( ADD. 0. FLUI.0014699 ) VISCOSITY( ADD. KINE = 25. CONT ) EXECUTION( ADD. PLOT ) ENTITY( ADD. VELO. NEWJ ) PRINTOUT( ADD. CONS = 2. PLOT ) ENTITY( ADD. FREE. NODE = 42. ENTI = "sides". 0 ICNODE( ADD. WIND = 9. SURF = 0.0753 ) BCNODE( ADD. PROP = "silicon" ) ENTITY( ADD. ENER. ZERO ) BCNODE( ADD. URC. SET = "water". BACK. 0.33 ) CONDUCTIVITY( ADD. DEPT = 31. NSTE = 1000. INCO. 0. UZC. NAME = "interface". VELC = 0. ZERO ) PROBLEM( ADD. STRA. UT. RESC = 0. ENTI = "inlet". 1e11. FZ = 0 ) PRESSURE( ADD. PLOT ) ENTITY( ADD. 301. MINI ) 1e20.0001. CONS = 73 ) DENSITY( ADD. ZERO ) BCNODE( ADD. MOME.R.17006 ) ENTITY( ADD. CONS. DT = 1e05. ANG2 = 300 ) ENTITY( ADD. ATTA = "silicon". ENTI = "water". CONS = 0. SET = "water". PLOT ) ENTITY( ADD. 1e20. HEAT. NONE. CONS = 37 ) BCFLUX( ADD. SET = "water". CONS = 0. ZERO ) BCNODE( ADD. 1 CLIPPING( ADD. ENTI = "inlet". VELO. MIXE = 1e11. NAME = "silicon". ZERO ) BCNODE( ADD. SET = "silicon". ENTI = "silicon". LAMI.5. SET = "silicon". NAME = "outlet". ZERO ) BCNODE( ADD. TRAN. CONS = 50. ENTI = "bottom". ENTI = "inlet". 0. VARI. ANG1 = 45. 10 301. VELO.5. CONS = 0. SING ) SOLUTION( ADD. ENTI = "interface". SURF. UPWI ) UPWINDING( ADD.5.
120 .Appendix B: (Continued) CREATE( FISO ) RUN( FISOLV. IDEN = "t06". COMP ) / File closed at Fri Sep 22 20:31:32 2006. BACK. TIME = "NOW". AT = "".
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.