Heat Transfer Engineering

ISSN: 0145-7632 (Print) 1521-0537 (Online) Journal homepage: http://www.tandfonline.com/loi/uhte20

Numerical investigation of flow and heat transfer

Neeraj Sinai Borker & Chakravarthy Balaji

To cite this article: Neeraj Sinai Borker & Chakravarthy Balaji (2017): Numerical investigation
of flow and heat transfer from impinging jets on a target surface with protrusions, Heat Transfer
Engineering, DOI: 10.1080/01457632.2017.1320172

To link to this article: http://dx.doi.org/10.1080/01457632.2017.1320172

Accepted author version posted online: 21

Submit your article to this journal

Article views: 7

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

Download by: [The UC San Diego Library] Date: 05 May 2017, At: 00:13
ACCEPTED MANUSCRIPT

Numerical investigation of flow and heat transfer from impinging jets on a target surface with

protrusions

Neeraj Sinai Borker, Chakravarthy Balaji

Department of Mechanical Engineering, Indian Institute of Technology Madras, Chennai, India

Address correspondence to Professor Chakravarthy Balaji, Department of Mechanical

Engineering, Indian Institute of Technology Madras, Chennai 600 036, India. E-mail:

[email protected] Tel: 91-44-2257 4689; Fax: 91-44-2257 4652

Abstract

Results of numerical simulation of flow field and heat transfer due to submerged multiple

circular jets impinging on a target surface with protrusions, at orifice plate to target

spacing smaller than twice the jet diameter, are presented. The spent fluid is either locally

extracted using effusion holes or allowed to flow outwards through the side walls. The

study establishes a methodology to systematically modify the impinging surface based on

the local fluid-flow structures. The effect of protrusions on the flow field and heat

transfer is studied for varying protrusion sizes and locations at Reynolds numbers in the

range of 1,000 to 10,000. It was found that addition of protrusions improves the heat

transfer characteristics of the system, lowering the average heater side-temperature as

well as the energy required to pump the fluid per unit amount of heat removed.

1 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Neeraj Sinai Borker is a Ph.D. student in the Sibley School of Mechanical and Aerospace

Engineering at Cornell University, Ithaca, New York, USA. He obtained his B.Tech. (Honours)

in Mechanical Engineering from Indian Institute of Technology Madras in 2014. His research

interests include fluid mechanics, computational fluid mechanics, heat transfer, micro-scale

fabrication and sustainable energy systems.

Chakravarthy Balaji is a Professor in the Department of Mechanical Engineering at the Indian

Institute of Technology (IIT) Madras. He graduated in Mechanical Engineering from Guindy

Engineering College, Chennai (1990), and obtained his M.Tech. (1992) and Ph.D. (1995) both

from IIT Madras. His research interests include computational and experimental heat transfer,

optimization in thermal sciences, inverse heat transfer, satellite meteorology, and numerical

weather prediction.

2 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Introduction

Impinging liquid jets are an effective means of providing high heat and mass transfer

rates, which find use in cooling of high-heat-dissipation electronic devices, internal combustion

engines and in the thermal treatment of metals. In many applications, heat fluxes approach values

of 5 MWm2 and above for electronic devices and around 20 MWm2 for semiconductor lasers

(Hannemann [1]). According to Kiper [2], the reliability of electronic components on a heat

dissipating chip decreases nearly by 10%, for every 2℃ rise in temperature, and these

components need nearly a constant spatial temperature. Single phase liquid jet impingement can

provide chip-to-coolant thermal resistances as low as 0.3 ℃cm1W1 with a modest coolant flow

rate of 1 l min1 , as reported by Incropera [3].

One way of enhancing the impinging jet heat transfer rates is to modify surface properties

of the impinging surface. Gabour and Lienhard [4] reported a dramatic increase in stagnation

point heat transfer with wall roughness for free water jets. Fu and Huang [5] and de Lemos and

Dórea [6] studied the effect of porous medium on the thermal performance of an impinging jet.

Kanokjaruvijit and Martinez-botas [7] and Ekkad and Kontrovitz [8] investigated the effect of

dimpled surfaces with cross-flow arrangement. Spring et al. [9], Andrews et al. [10] and Zhang

et al. [11] have given some insight into the effect of protrusions on the heat transfer

characteristics of an impinging jet system. In the current study, protrusions on the impinging

surface were considered to enhance heat transfer between the coolant and the heater surface.

Although the above studies have a discussion about the effect of the flow structures in their

respective systems, there was no-attempt made to approach an optimal arrangement of

protrusions on the impinging surface. The current study utilizes the knowledge of these previous

3 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

studies to give the geometry of protrusions that will lead to enhanced performance. Fitzgerald

and Garimella [12] articulated the effect of flow structure and turbulence on the radial variation

of heat transfer rates. They give insight into the origins of the secondary maximum of the heat

transfer coefficient observed in an impinging jet flow. O'Donovan and Murray [13,14] elaborated

on the evolution of vortex structures and its corresponding impact on the local heat transfer.

These flow-structures observed for a single jet and a flat impinging surface are used to design the

layout of protrusions. Huber and Viskanta [15,16] reported the effect of spent air exits in the

nozzle plate. Hoberg et al. [17] studied the effect of effusion holes on the local heat transfer

coefficients. In this study, the effect of effusion on the location of protrusions is also described,

to design a cooling block that can be placed directly above a heater surface.

From a review of the pertinent literature, it is clear that multiple jets impinging on a

target surface with protrusions has not been utilized to its full potential. In consideration of the

above, in this study the effect of protruding surfaces on the flow and heat transfer characteristics

of confined and submerged circular impinging jets is numerically investigated. A methodology

to obtain a layout of protrusions is elucidated based on the flow-structures that arise at these

orifice-plate-to-target spacing. The effect of protrusions on the pressure drop and heater side

temperature is sought to be established. A model system of multiple jets with and without cross-

flow is also studied to show the effectiveness of the analysis.

Mathematical model and solution procedure

Jet impingement heat transfer for both single and multiple circular jets is numerically

investigated. A three dimensional model, discretized using a finite volume approach, is used. The

4 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

following section elucidates the numerical procedure used for the present study. The conjugate

heat transfer problem is formulated with the following assumptions: the fluid is incompressible,

the fluid flow is steady, the fluid properties are constant, radiation and natural convection are

neglected, and viscous dissipation is absent.

Geometry

Figure 1 shows a typical cross-sectional view of an impinging single jet on a target plate

with constant heat flux boundary condition. The target surface is an aluminum plate of size

35 35 mm2 and water is used as the fluid.

Governing equations

The governing equations for steady state, three dimensional, incompressible flow and

heat transfer are as follows.

Continuity equation

ui  0 (1)
xi 
Momentum equations

  ui u j  1 P    u u j 
   ν  νt   i    (2)
x j ρ xi x j   x j x i 

Energy equation

   ν νt  T 
x j
 u j T     
x j  Pr σT  x j 
(3)

5 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Turbulence closure equations

To close the problem mathematically, the κ- ω turbulence model proposed by Menter

[18], as reported in Zuckerman and Lior [19] is used. The κ and ω transport equations are given

as follows.

κ and ω transport equations

   κ 
x j
 ρκu j   Γ   Gκ  Yκ  Sκ (4)
x j  κ x j 

   k 
x j
 ρωu j   Γ   Gω  Yω  Sω (5)
x j  ω x j 

In equation (4) Gk represents the generation of turbulent kinetic energy due to mean

velocity gradients. G κ is defined in terms of the modulus of the mean rate of the strain tensor in a

manner consistent with the Boussinesq hypothesis. Similarly Gω represents the generation of ω .

The effective diffusivities, Γκ and Γω can be defined in terms of turbulent Prandtl numbers and

turbulent viscosity. Yκ and Yω represent the dissipation of κ and ω due to turbulence Sω and

Sκ are source terms. The details of the scheme can be found in the work of Menter [18].

Boundary Conditions

The boundary conditions of the present study are

 No-slip boundary conditions at the walls.

(6)
 Flow inlet conditions at the nozzle inlet

(7)

6 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(8)

 Entrainment boundary condition at the pressure outlet,

(9)

 Conjugate boundary condition at the impingement surface, i.e. temperature and heat flux

continuity at the impinging surface,

(10)

(11)

 Constant heat flux condition at the heater surface

(12)

 Adiabatic conditions at the nozzle walls and side wall of impingement surface

(13)

Numerical scheme

The governing equations are discretized on a uniform hexagonal mesh using a finite

volume approach. The velocities and the pressure are calculated using the semi implicit pressure

linked equation (SIMPLE) algorithm. The second order upwind scheme is used for the

interpolation of the gradients of velocities and temperature. Ansys 14.5 is used to solve the

equations (1-5) subject to boundary conditions given by equations (6-13). The solution is

considered to converge, when the residual is of the order of 105 for the continuity, momentum

and turbulence equations and 107 for the energy equation. Further, the area weighted average of

temperature of the impingement surface is continuously examined, so that the variation is within

0.5% for 1000 consecutive iterations. Convergence is declared if both the above conditions are

satisfied.

7 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Grid independence study

The computational domain is divided into sub-domains and grid independence of each of

these sub-domains ensures global grid independence for the whole assembly. Following Roy

[20], Richardson extrapolation is used for obtaining a grid with low discretization error at each of

these sub-domains. There are three major zones: bulk fluid above the protrusions, fluid near the

protrusions and wall near the protrusion, and conduction in the solid below the protrusions.

The last two zones are clubbed together as the grid is uniform in the impingement

direction. Figure 2 shows a plot of the average Nusselt number against the reciprocal of the total

number of nodes. The grid independent Nusselt number is found out to be 90.6 and corresponds

to a computational grid with infinite number of nodes. A computational grid with 2.27 million

nodes gave a relative error of 1% for all the zones and is used for all subsequent calculations.

Validation of numerical study

The experimental results presented by Tie et al. [21] are used for validating the numerical

study for multiple jets with cross-flow. The overall heat transfer coefficient is used as the

parameter and matches closely with the reported experimental results. The Nusselt number

variation with Reynolds number ( ReD  UD / ν ) is regressed, using the power dependence of

0.51 suggested by Tie et al. [21]. The definitions of the average heat transfer coefficient and the

average Nusselt number are given in equations (14-15). The predictions are within 3% of the

reported results, as shown in Figure 3.

q
αavg  (14)
Tavg  Tin

8 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

αavg D
Nuavg  (15)
k

Results and discussion

Generally, jet impingement heat transfer coefficients radially deteriorate after the

stagnation zone for a single impinging jet. However, for multiple jets with low orifice plate-to-

target spacing abrupt changes in local heat transfer coefficients are usually observed and these

are mainly attributed to flow separation and formation of wall eddies and secondary stagnation

points. Protruding surfaces are expected to reduce the chip temperature, due to an increase in the

surface area. For the case with local extraction of spent fluid, it is possible to maintain the overall

drop in pressure if the protrusions are positioned away from the global flow-direction. The height

of protrusions is chosen based on the boundary layer thickness given in equations (16) and (17),

following Incropera [3]. Equations (16) and (17) are valid for 1  H / D  10 and give an

approximate value for the thickness of the boundary layer in the stagnation zone. This was

chosen so that the slowly moving fluid in the boundary layer can effectively transfer heat with

the cooler convecting fluid in the free jet stream above and take maximum advantage of the

increase in surface area near the protrusions. The fluid trapped under the protrusion moves very

slowly compared to the free stream motion. The main mode of heat transfer is conduction and

convection due to eddies in the region. The width and spacing between protrusions is chosen to

be of the same order as the height of the protrusion due to the eddy structures that lead to heat

transfer. An extreme value of height in comparison to the spacing will lead to a pool of stagnant

fluid at the bottom of the space whose temperature will now increase, thus deteriorating local

heat transfer. The associated increase in the drag with increasing protrusion heights needs to be

9 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

taken into account, which is yet another reason to maintain low protrusion heights. A symmetry

element surrounding the central jet is studied assuming net zero cross-flow in the system and the

top view of the geometry is shown in Figure 4. The location of the protrusion ( r1 ) is defined as

the shortest distance of the leading edge of the first protrusion from the stagnation point. The

effect of height of protrusions is studied for heights up to 200 μm , in steps of 50 μm and is

presented in the following section.

δ 1.95
 (16)
D  G ReD 0.5

H
G  1.04  0.034   (17)
 D

Effect of location of and size of protrusions for a symmetry element

A term called performance factor ( η ) is introduced for taking the pressure drop into

consideration. This is defined as the amount of heat removed by the coolant for a given rise in

heater surface temperature per unit amount of work input, and is mathematically given by

equation (18). Since the heat flux varies linearly with the chip surface temperature, for a constant

convection resistance, it is customary to solve for a fixed heat flux and convert it based on the

reference temperature chosen. A reference temperature ( Tref ) of 320K is chosen which

corresponds to an excess temperature with respect to the inlet fluid temperature of 20K (

ΔTref  Tref  Tin  20K), although the comparison is independent of this choice. The pressure

required to pump the fluid Psolved is evaluated for a given volume flow rate ( V ). Ah is the area

of the heater through which heat is being supplied.

10 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

q  Ah qo ΔTref  Ah
η  (18)
PV ΔTsolved Psolved V

A parametric study was conducted on three types of protrusions depending on the location

and orientation with the flow direction for a fixed Reynolds number of 6000. The geometries are

shown in Figures 4-6.

Type 1. Protrusions in the region of interaction wall jets of secondary jets: aligned normal

to the local flow direction as shown in Figure 4.

Type 2. Protrusions in the region of interaction wall jets of secondary jets: aligned inline

to the local flow direction as shown in Figure 5.

Type 3. Protrusions surrounding the stagnation zone under the outlet aligned in the global

flow direction as shown in Figure 6.

Figures 7 and 8 show the overall picture of the effect of protrusions on Type-1 and Type-

2 configurations. The pumping pressure remains nearly constant for all the cases, with the overall

deviation remaining within 5% . On the other hand, the average heat transfer coefficient shows

an increasing trend. The average temperature of the impinging surface steadily decreases with an

increasing protrusion size. This is mainly due to the increase in area of the impinging surface.

Nearly 70% increase in performance factor is observed. The reason for the increase in heat

transfer coefficient is attributed to a decrease in the average surface temperature as well as an

increase in the overall surface area. The key point, however, is to look at heater side wall

temperature which also shows a maximum decrease of over 30%. Thus, the addition of

protrusions in this zone enhances the heat transfer in the system with negligible effect on the

11 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

pumping pressure. As the protrusion height is much smaller than the orifice to target spacing

(h/H ~ 0.1), the flow speeds near the protrusion are significantly small. There is 11-21%

reduction in the average heater side temperature with increasing size of protrusions as seen in

Figure 9. However, the pumping pressure also increases with protrusion size thus reducing the

performance factor at larger size of protrusions. The increase in pumping power is mainly

attributed to the location of protrusions that lie along the global flow direction increasing the

flow resistance. This is seen for the protrusions of Type-3 where the required pumping pressure

increases sharply as shown in Figure 9a.

Effect of protrusions for a multiple jet configuration

Introduction

From the results of the previous study, protrusions of Type-1, with r1 = 1.5D and

protrusion size of 200 μm is used for analysing the complete model for cooling a 35  35 mm2

aluminium heater using jet impingement with local extraction of spent fluid (Model-1). Jet

impingement with cross-flow of spent fluid at low jet-to-jet spacing of a staggered array (Model-

2) is also analyzed.

Geometry

The jet diameter is fixed to 2 mm and the orifice plate to target spacing equals the jet

diameter (H/D = 1). The thickness of the impinging plate is fixed as 2mm. The jets are arranged

in a staggered array. The protrusion size is 200 μm , pitch is twice the protrusion height (p = 2h)

and r1  1.5D. Water is used as the coolant and aluminium is chosen as the heater material. Two

models as shown in Figure 10) are considered: Model-1 - Local extraction of spent fluid via

12 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

effusion holes in the nozzle plate, and Model-2 - Cross-flow of spent fluid at low jet-to-jet

spacing.

The jets in Model-1 are arranged in a 3  3 linear array with a jet-to-jet spacing of 5

diameters (S/D = 5). The jets in Model-2 are arranged in a 5  5 staggered array at S/D = 3.54.

The study of both of these models is independent of each other. The Reynolds number is varied

from 1000 to 10,000.

Multiple-jet heat transfer results

Heat transfer data from Model-1 and Model-2 with and without modified impingement

surface, for Re = 6000, is summarized in Figures 11 and 12 along the radial direction from the

centre jet parallel to one of the edges. The effect of protrusion on the two models is summarized

below.

The stagnation zone heat transfer coefficient for both Model-1 and Model-2 is similar to

that of a single impinging jet, at the same Reynolds number, within r/D<1 from the center jet.

The effect of cross-flow and effusion holes is felt only after this distance. The stagnation zone

heat transfer becomes asymmetric for perimeter jets when there is cross-flow in the system,

which increases downstream. This is mainly attributed to deflection of jets caused due to strong

cross-flow velocities at low values of H/D. Although the impingement core is disrupted, the

increased level of turbulence and low velocity top wall jet, help to compensate for this loss. The

corresponding streamline contours shown in Figure 13 suggest that at low values of H/D, the

separated flow impinges on the top wall to form a low velocity wall jet. This stream which has

high levels of turbulence, as seen in Figure 18, entrains into the main jet stream of the perimeter

13 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

jet, leading to the enhancement of heat transfer on that side. As there is no entrainment of fluid

on the opposite side, the heat transfer coefficient is similar to that of the center jet. The average

stagnation zone heat transfer coefficient is observed to be higher for perimeter jets than the

center jet. This is contrary to the result observed at H/D = 5, a study by Tie et al. [21], where the

up wash flow does not impinge on the top wall. Figure 14 shows a comparison for the present

case with that of a simulated case (similar to the results of Tie et al. [21]) at H/D = 5 and S/D =

7. For Model-1 the heat transfer coefficient in the stagnation zone is almost similar in all jets due

to the addition of the effusion holes that minimizes the cross-flow in the system. The valleys

observed in Figure 11 correspond to the flow separation points and the secondary peaks observed

correspond to the location of wall eddies formed. Figure 15) shows the variation of Nusselt

number with Reynolds number. The Nusselt number variation with Reynolds number for Model-

2 matches closely with the power law fit, Nu  Re0.51

error. However, the functional relationship with Reynolds number for Model-1 increases to the

2/3 power, similar to the one given by Martin [22] for gas jets and suggested by Huber and

Viskanta [15,16] for air jets with local extraction of spent fluid. Although the average Nusselt

number is greater for Model-2, due to higher pressure drop, the performance factor is lower in

comparison to Model-1. The jet-to-jet interaction is strong due to the lower values of orifice

plate-to-target spacing (H/D = 1) as well as low jet-to-jet spacing along the diagonal path (S/D =

3.54) for Model-2.

The stagnation zone heat transfer for both models is altered by the addition of

protrusions. The heater side temperature as well as the temperature variance (equation (19))

comes down with the inclusion of protrusions with a meagre change in the pumping pressure.

14 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

βT   T  Tavg  dx1  dx2  Σ Ti  Tavg  ΔAi

The pumping pressure reduces due to earlier flow separation that reduces the overall skin

friction coefficient. The increase in heat transfer in the wall jet interaction zones of adjacent jets

is mainly attributed to the relative increase in turbulence which leads to an increase in the heat

transfer locally, although the velocities are sufficiently small. The secondary stagnation zone

created under the outlet is mainly responsible for the increase in heat transfer coefficient locally

that takes in huge amounts of heat along with the escaping fluid.

For Model-1 with protrusions, the stagnation zone heat transfer is similar for all the 9

impinging jets, as expected. The protrusions enhance the stagnation zone heat transfer compared

to a case with no protrusions, and the stagnation point heat transfer is nearly 60% higher in

magnitude. Due to the large size of protrusions, in comparison to the stagnation zone boundary

layer thickness, the wall jet flow separates at the leading edge of the first protrusion. The

separated flow, that has high levels of turbulence, as seen in Figure16, impinges on the top wall.

This low velocity wall jet entrains into the potential core of the mainstream jet thus leading to an

enhancement of heat transfer. Near the protrusions due to lower magnitude of velocity, the

turbulence level is far lower than that in the wall jet, as seen in Figure 17. The local fluid

temperature rises, but the increase in area compensates for this loss and allows for lower heater

temperature. The excess impinging surface temperature decreases by 17% and the heater

temperature by 12 % for nearly the same pumping power as seen in Table 1.

For Model-2 with protrusions, the jet deflection is observed to reduce compared to a case without

any protrusions of the top wall jet in the mainstream leads to an increase in heat transfer.

15 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

However, the average heat transfer coefficient is similar to the case without any protrusions. The

main advantage of this model is the decrease in spatial temperature variance, defined in equation

(19), which is 60% lower than the one with the protrusion as shown in Table 2. The heater

temperature also reduces by 18%. The overall heat transfer coefficient remains nearly the same

and is mainly attributed to a decrease in the average surface temperature. The flow is observed to

be heavily disturbed by the addition of protrusions due to flow separation at the leading edge of

the closest protrusion. The consequential effect is the reduction in the overall pumping pressure

due to lowering of skin friction coefficient. The variation in local heat transfer coefficient, as in

Figure 11, confirms the effect of cross-flow on the thermal transport. Due to the cross flow the

formation of wall eddies is suppressed and one can find a more uniform variation near the wall

jet interaction zone (r/D = 2.5). The local peaks in heat transfer coefficient near the protrusions

are mainly due to the sudden increase in area, decrease in surface temperature and sharp edges.

The overall results for Re = 6000 are summarized in Tables 1 and 2. The average surface

temperature at the heater side comes down by 5 ℃ for the modified surface which corresponds to

a 12-18% reduction in excess temperature ( Th  Tin ) . This can be seen as the effect of either

increase in heat transfer coefficient or increase in surface area. The variation of average heat

transfer coefficient suggests that the increase in area is the main reason for improving the heat

transfer and reducing the overall temperature of the chip. Since more area is in contact with the

cooler fluid, the average impinging surface temperature also comes down. Hence the protrusions

act as extended surfaces.

The performance factor as defined earlier is used to judge how much heat can be removed, given

the limit to maximum chip temperature and input power. The performance factor of various

16 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

models is shown for the Reynolds number of 6000 and reference temperature excess of 20℃.

The increase in performance factor is around 16% for Model-1 and around 40% for the Model-2.

The increase suggests a better cooling performance of the unit itself. The heater temperature

comes down by 12% for Model-1 and by 18% for Model-2.

The variation of the Nusselt number with Reynolds number follows the 2/3 power

variation as suggested by Martin [22] for both Model-1 and Model-2 within 1% error. This

implies that the protrusions enhance heat transfer rates for Model-1 at higher Reynolds number,

while improving the heat transfer coefficient for Model-2 at all Reynolds numbers.

Conclusions

In this study flow and heat transfer characteristics of multiple impinging jets was

numerically investigated for low orifice plate to target spacing with and without protruding

surfaces. The shear stress transport κ - ω turbulence model, developed by Menter [18], was used

to solve the Reynolds Average Navier Stokes Equation and the κ and ω transport equations.

Water was chosen as the coolant and aluminum or copper as the target surface material.

The numerical study was able to accurately predict the heat transfer in the wall jet region.

However there was a slight over prediction in the stagnation zone. Initially a symmetry element

was analyzed to understand the effect of protrusions on the heat transfer characteristics and the

most effective type of protrusions was then used to cool a 35 35 mm2 heater plate. The flow-

field around the protrusions was used as a basis to design a topology that was found to work best

amongst the geometries studied. Two cooling modules namely, one in which the spent fluid is

extracted through effusion holes in the orifice plate (Model-1) and the second with cross-flow of

17 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

the spent fluid (Model-2) were analyzed. The effect of cross-flow was analyzed for jets of 2 mm

diameter and an orifice plate-to-jet spacing of 2mm (H/D = 1). The effect of protruding

surfaces, in the wall jet region, on the overall heat transfer and input power was studied for both

models. The following salient conclusions were obtained from the investigations.

Protruding surfaces in the wall jet region of Type-1 are most effective in cooling the heater

surface when fluid is locally extracted using effusion holes. The protrusions reduce the heater

temperature and increase the performance factor, both improving with an increase in the

protrusion height till 200 μm . For multiple jets with cross-flow type arrangement, jet deflection

disrupts the stagnation zone of outer jets that leads to a deterioration of heat transfer rates at high

values of H/D while enhancement at low values of H/D and S/D. The protrusions of size

200 μm reduced the excess heater surface temperature by 12% and increased the performance

factor by 16% for Model-1 for Re = 6000. Under the same conditions for Model-2, an 18%

decrease in the excess heater surface temperature and 40% increase in performance factor was

observed. The Nusselt number variation with Reynolds number for Model-1 without protrusions

matched closely with the power law fit, Nu  ReD0.51 . However the exponent becomes 0.67 for

the modified surface.

Nomenclature

Ah area of the heater, [ m2 ]

dxi elemental area in the ith-direction, [m]

D jet diameter, [m]

18 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Gκ generation of turbulent kinetic energy, [ kg m1s3 ]

Gω generation of ω , [ kg m3s2 ]

H orifice to target spacing, [m]

h protrusion height, [m]

k thermal conductivity, [ Wm1K1 ]

N1 number of mesh elements in the fluid domain away from the impinging surface

N2 number of mesh elements near the wall and the heater surface

Nu0 Nusselt number at the stagnation point of the centre jet

Nuavg Nusselt number averaged over the impinging surface

NuD Nusselt number based on diameter

P fluid pressure, [Pa or kg m1s2 ]

p protrusion pitch, [m]

Pr Prandtl number

q heat flux, [ Wm2 ]

r radial distance from the stagnation point (or stagnation point of the centre jet in

case of multiple jets) on the impinging surface, [m]

19 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

r1 shortest distance of the leading edge of the first protrusion from stagnation point, [m]

ReD Reynolds number based on diameter

S jet-to-jet spacing, [m]

Sκ source of turbulence kinetic energy, [ kg m1s3 ]

Sω source of specific rate of dissipation, [ kg m3s2 ]

T temperature, [K]

Ti temperature of the ith mesh element, [K]

ui velocity field in the ith direction, [ ms1 ]

V inlet jet velocity, [ ms1 ]

V volume flow rate, [ m3s1 ]

w protrusion width, [m]

xi spatial position in the ith direction, [m]

Yκ dissipation of turbulence kinetic energy ( κ),  [ kg m1s3 ]

Yω dissipation of specific rate of dissipation ( ω),  [ kg m3s2 ]

20 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Greek Symbols

α heat transfer coefficient, [ Wm2K1 ]

βT temperature variance, [K]

ΔAi elemental area of the ith mesh element, [ m2 ]

δ boundary layer thickness, [ m ]

η performance factor

κ turbulent kinetic energy, [ m2s2 ]

ν kinematic viscosity, [ ms1 ]

νt turbulent kinematic viscosity, [ ms1 ]

ρ fluid density, [ kg m3 ]

σT turbulent Prandtl number

ω specific rate of dissipation,[ s 1 ]

Subscripts

avg average value on the impinging surface

f fluid

h heater wall

21 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

i ith dimension

in inlet conditions

o stagnation point (or stagnation point of the centre jet in case of multiple jets)

out outlet conditions

ref reference

s solid

T temperature

22 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

References

[1] Hannemann, R. J., Thermal Control of Electronics: Perspectives and Prospects,

Proceedings of the Rohsenow Symposium on Future Trends in Heat Transfer, 2003.

[2] Kiper, A. M., Impinging Water Jet Cooling of VLSI Circuits, Int. Comm. Heat Mass

Transfer, vol. 11, pp. 517-526, 1984.

[3] Incropera, F., Liquid Cooling of Electronic Devices by Single-Phase Convection, Wiley,

1999.

[4] Gabour, L. A., and Lienhard, J. H., V., Wall Roughness Effects on Stagnation-Point Heat

Transfer Beneath an Impinging Liquid Jet, ASME. Journal of Heat Transfer, vol.116, no.

1, pp. 81-87, 1994.

[5] Fu, W.S., and Huang, H.C., Thermal performances of different shape porous blocks

under an impinging jet, Int. J. Heat Mass Transfer, vol. 40, pp. 2261–2272, 1997.

[6] de Lemos, Marcelo J. S., and Dórea, Felipe T., Simulation of a Turbulent Impinging Jet

into a Layer of Porous Material Using a Two–Energy Equation Model, Numerical Heat

Transfer, Part A: Applications, vol. 59, no. 10, pp. 768-798.

[7] Kanokjaruvijit, K., and Martinez-botas., R. F., Jet impingement on a dimpled surface

with different crossflow schemes, International Journal of Heat and Mass Transfer, vol.

48, no. 1, pp. 161-170, 2005.

[8] Ekkad, S. V., and Kontrovitz, D., Jet impingement heat transfer on dimpled target

surfaces, International Journal of Heat and Fluid Flow, vol. 23, no. 1, pp. 22-28, 2002.

23 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

[9] Spring, S., Xing, Y., and Weigand, B., An experimental and numerical study of heat

transfer from arrays of impinging jets with surface ribs, ASME, Journal of Heat Transfer,

vol. 134, no. 8, pp. 082201-1 – 082201-11, 2012.

[10] Andrews, G. E., Hussain, R. A., and Mkpadi, M. C., Enhanced impingement heat

transfer: the influence of impingement x/d for interrupted rib obstacles (rectangular pin

fins), Journal of Turbomachinery, vol. 128, no. 2, pp. 321-331, 2006.

[11] Zhang, D., Qu, H., Lan, J., Chen J., and Xie, Y., Flow and heat transfer

characteristics of single jet impinging on protrusioned surface, International Journal of

Heat and Mass Transfer, vol. 58, no. 1, pp. 18-28, 2013.

[12] Fitzgerald, J. A., and Garimella, S. V., A study of the flow field of a confined and

submerged impinging jet, International Journal of Heat and Mass Transfer, vol. 41, no.

8, pp. 1025-1034, 1998.

[13] O'Donovan, T. S., and Murray, D. B., Jet impingement heat transfer, Part I: A

temporal investigation of heat transfer and local fluid velocities. International Journal of

Heat and Mass Transfer, vol. 50, no. 17-18, pp. 3291-3301, 2007.

[14] O'Donovan, T. S., and Murray, D. B., Jet impingement heat transfer, Part II: A

temporal investigation of heat transfer and local fluid velocities, International Journal of

Heat and Mass Transfer, vol. 50, no. 17-18, pp. 3302-3314, 2007.

[15] Huber, A. M., and R., Viskanta., Convective heat transfer to a confined impinging

array of air jets with spent air exits, ASME Journal of Heat Transfer, vol. 116, no. 3, pp.

24 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

570-576, 1994.

[16] Huber, A. M., and R. Viskanta, Effect of jet-jet spacing on convective heat

transfer to confined, impinging arrays of axisymmetric air jets, International Journal of

Heat and Mass Transfer, vol. 37, no. 18, pp. 2859-2869, 1994.

[17] Hoberg, T. B., Onstad, A. J., and Eaton, J. K., Heat transfer measurements for jet

impingement arrays with local extraction, International Journal of Heat and Fluid Flow,

vol. 31, no. 3, pp. 460-467, 2010.

[18] Menter, F. R., Two-equation eddy-viscosity turbulence models for engineering

applications, AIAA Journal, vol. 32, no. 8, pp. 1598-1605, 1994.

[19] Zuckerman, N. and Lior, N., Jet impingement heat transfer: physics, correlations,

and numerical modelling, Advances in Heat Transfer, vol. 39, pp. 565-631, 2006.

[20] Roy, C. J., Review of code and solution verification procedures for computational

simulation, Journal of Computational Physics, vol. 205, no. 1, pp. 131–156, 2005.

[21] Tie, P., Li, Q., and Xuan, Y., Investigation on the submerged liquid jet arrays

impingement cooling, Applied Thermal Engineering, vol. 31, no. 14, pp. 2757-2763,

2011.

[22] Martin, H., Heat and mass transfer between impinging gas jets and solid surfaces,

Advances in Heat Transfer, vol. 13, pp. 1-60, 1977.

25 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Table 1. Results for Model-1 at Re = 6000.

h Tavg Pumping power havg  T

[ m ] [K] [mW] [ Wm2 K 1 ] [ K2 ]

0 332.53 338 22554 1337 5.90

200 328.65 331 20340 1551 7.08

26 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Table 2. Results for Model-2 at Re = 6000.

h Tavg Pumping power havg  T

[ m ] [K] [mW] [ Wm2 K 1 ] [ K2 ]

0 327.31 608 27983 885 0.33

200 322.434 531 27923 1235 0.13

27 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 1. Cross section of the geometry of a single impinging jet with the boundary

conditions.

28 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

29 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(c)

Figure 2. Grid independence study for (a) zone-1, (b)zone-2 and zone-3, (c) Mesh used in

the analysis.

30 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 3. A comparison between the present work and the results reported by Tie et al. [21]

shows good agreement of the results presented in current study.

31 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

Figure 4. (a) Top view of a symmetry element of the geometry, Type-1, the location of the

protrusion defined by (b) Cross sectional view (side-view) of protrusions and its

32 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

position relative to the centre of the impinging jet defining the height(h), width(w) and

pitch(p) of the protrusion. The local flow-direction over the protrusion is on an average

transverse to the protrusions.

33 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 5. Protrusion in the wall jet region in line with the flow direction (Type-2). The local

flow-direction over the protrusion is on an average along the protrusions.

34 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 6. Protrusion surrounding the stagnation zone around the outlet (Type-3). The

protrusions surround the region near the outlet-holes. The local flow-direction over the

protrusion is on an average transverse to the protrusions.

35 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

Figure 7. Variation of (a) pumping pressure, (b) performance factor with the protrusion

height for Type-1 and Type-2. The inlet pressure is almost constant with the position of

36 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

the protrusion, as it does not directly come in the path of the global flow direction, which

is from the impingement point beneath the inlet and the outlet effusion holes. The

performance factor increases with the protrusion height.

37 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

Figure 8. (a) Average heat transfer coefficient and (b) average surface temperature with the

protrusion height for Type 1 and Type 2.

38 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

39 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(c)

(d)

Figure 9. Variation of (a) pumping pressure, (b) performance factor, (c) average heat

transfer coefficient and (d) average temperature with the protrusion height for Type 3.

40 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

41 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(c)

Figure 10. (a) Geometry with local extraction of fluid (Model-1); which consists of 16 inlets

and 9 outlet ports on the orifice plate. (b) Geometry with cross-flow of fluid (Model-2)

with 25 inlets arranged in a staggered array, with outlet on the side-wall. (c) Top-view of

the protrusions for Model-2, Type-1. Line segments AB and AC on the impingement

surface are used to provide heat transfer coefficient in later figures.

42 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 11. Heat Transfer coefficient for Model-1 and Model-2 with and without protrusion

along line segment AB shown in Figure 10. The abrupt change in heat transfer-coefficient

is due to rapid change in temperature and heat flux at the protrusion surface.

43 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 12. Temperature variation for Model-1 and Model-2 with and without protrusion

along line segment AB shown in Figure 10.

44 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

45 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(c)

(d)

Figure 13. Streamlines along a plane normal to the impingement surface passing through AC

for (a) Model-1 without protrusions; (b) Model-2 without protrusions; (c) Model-1 with

protrusions; (d) Model-2 with protrusions.

46 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 14. Comparison of the radial variation of normalized Nusselt number with the study

of Tie et al. [21]. The Nusselt number is normalized using its value at the centre of the

impinging surface ( ). The radial distance is normalized with respect to the jet-to-jet

spacing.

47 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figure 15. Nusselt number variation with Reynolds number. Model-1 has a power law

exponent of (2/3) as suggested by Martin [22] and Model-2 has a power law fit of 0.51

similar to Tie et.al [21].

48 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

Figure 16. (a) Pressure variation and (b) turbulence kinetic energy contours for Model-2,

without protrusions along a plane normal to the impingement surface passing through

AC.

49 ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

(a)

(b)

Figure 17. Turbulence kinetic energy contours for Model-1 with protrusions along a plane

normal to the impingement surface passing through AB (b) along a plane normal to the

impingement surface passing through AC.

50 ACCEPTED MANUSCRIPT