International Journal of Heat and Mass Transfer 127 (2018) 413–425

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

Transient heat transfer characteristics of array-jet impingement

on high-temperature flat plate at low jet-to-plate distances
Keqian Zhu a,b, Pingping Yu b, Ningyi Yuan b,⇑, Jianning Ding a,b,⇑
a
School of Mechanical Engineering, Micro/Nano Science and Technology Center, Jiangsu University, Zhenjiang 212013, China
b
School of Mechanical Engineering, Jiangsu Collaborative Innovation Center for Photovoltaic Science and Engineering, Jiangsu Province Cultivation Base for State Key Laboratory
of Photovoltaic Science and Technology, Changzhou University, Changzhou 213164, China

a r t i c l e i n f o a b s t r a c t

Article history: Numerical studies on the transient heat transfer characteristics of air-array-jet impingement, for small
Received 7 November 2017 jet-to-plate distances and a large temperature difference between the nozzle and plate, are presented.
Received in revised form 8 June 2018 The total mass flow rate of the jets (mÞ _ is constant at 30.34 kg/h. The nondimensional jet-to-plate
Accepted 19 July 2018
distance (H/D) for a nozzle diameter (D) of 5 mm is varied from 0.2 to 1. The nondimensional hole-to-
hole spacing is S/D = 5, 7, and 10, respectively. The variations in the transient heat transfer characteristics
and flow velocity at different values of H/D and S/D, as a function of the cooling time, are discussed. It is
Keywords:
found that there exists a turning point H/D = 0.4 in the effect of the transient heat transfer. As H/D is
Transient heat transfer
Glass tempering
decreased, the quenching time shrinks quickly. The velocity field is proposed as an explanation for the
Array jet impingement observed transient heat transfer. In addition, an appropriate proposal is presented for designing equip-
Low jet-to-plate distances ments of tempering ultra-thin glass.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction
The thinning and miniaturization of components, such as
liquid-crystal displays and solar cells, have increased the market
demand for ultra-thin glass. The glass tempering process is based
on a sudden cooling process using air jets, in which the glass is
heated close to its melting temperature, strengthening the glass
[1,2]. Ultra-thin glass needs a shorter cooling time and faster sur-
face cooling compared to ordinary 4–6-mm-thick glass [3–6].
Jet-impingement heat transfer is used for the sudden cooling of
thin glass, during the glass tempering process. From the resent lit-
erature, a single jet [7–14] is generally used for the local cooling of
a small surface area. Choo et al. [11,12] investigated the heat trans-
fer characteristics of impinging jets at low nozzle-to-plate dis-
tances (H/D = 0.125–1.0). They found that the Nusselt number is
independent of H/D at a fixed pumping power. In addition, Kuraan
et al. [10] studied the heat transfer and hydrodynamics of free
water jet impingement at low nozzle-to-plate spacing (H/D =
0.08–1.0). They showed that the normalized Nusselt number, pres-
sure, and hydraulic jump diameter are divided into two regions
⇑ Corresponding authors at: School of Mechanical Engineering, Jiangsu Collabo-
rative Innovation Center for Photovoltaic Science and Engineering, Jiangsu Province
Cultivation Base for State Key Laboratory of Photovoltaic Science and Technology,
Changzhou University, Changzhou 213164, China (J. Ding).
E-mail addresses: nyyuan@cczu.edu.cn (N. Yuan), dingjn@cczu.edu.cn (J. Ding).
over the entire H/D range. With respect to array-jet impingement,
while a certain amount of experimental [15–23] and numerical
[24–28] studies on the multiple impinging air-jets have been done,
few works [29–36] focused on the transient heat transfer charac-
teristics. Glass tempering is performed at a large temperature dif-
ference between the inlet and glass plate with a small jet-to-plate
distances (H), which is very different from the normal operating
conditions. However, most of the above mentioned studies concen-
trated on steady-state conditions and large jet-to-plate distances.
Several empirical correlations were suggested for air jet at a large
H/D for single and array-jet impingement, but the transient heat
transfer characteristics at small H/D are still limited. An investiga-
tion of the transient heat transfer characteristics for air-jet
impingement is necessary to ensure the tempering quality of large
ultra-thin glass plates. More importantly, the differences in the
transient heat transfer characteristics from the normal conditions
should be elucidated.
The purpose of this study is to investigate the transient heat
transfer characteristics of air-array impinging jets for a small H/D
and a large temperature difference at a constant total mass flow
_ for air jets. A simulation is conducted with an m
rate (m) _ value
of 30.34 kg/h and the results are compared and discussed in
detailed for the different nondimensional H/D varied from 0.2 to
1, the different nondimensional hole-to-hole spacing (S/D = 5, 7,
and 10), respectively. The results will be useful for understanding
the transient heat transfer characteristics a under a confined space,

https://doi.org/10.1016/j.ijheatmasstransfer.2018.07.099
0017-9310/Ó 2018 Elsevier Ltd. All rights reserved.
414 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425

Nomenclature

D nozzle diameter (mm) u fluid velocity (m=s)

Dx cross-diffusion term W the width of the impingement plate (mm)
Gj mass flow rate of coolant per unit area (kg=m2  s)
H jet to impingement plate distance (mm) Greek symbols
hc convection heat transfer coefficient q density ðkg=m3 Þ
hext external heat transfer coefficient k thermal conductivity ðW=K  mÞ
I Enthalpy (J=kg) s stress tensor
L the length of impingement plate (mm) l dynamic viscosity ðkg=m  sÞ
m_ total mass flow rate of jets (kg/h) m kinematic viscosity (m2 =s)
N number of jet nozzle f quenching times reduction ratio
P static pressure (Pa) g temperature difference enlargement ratio
Pamb ambient pressure
Re Reynolds number Subscripts
S jet to jet spacing (mm) j Jet
Sk user-defined source term
p target plate
Sx user-defined source term air air
T temperature (K) ext external
DT temperature difference (K)

and for designing an array-jet impingement system with certain _ =

theoretical significance.
2. Geometry and numerical analysis
2.1. Geometry
Fig. 1 shows the basic geometry of the computational domain.
Air through circular tubes vertically impinges on a glass plate with
a length of 100 mm, width of 100 mm, and thickness of 2 mm. The
circular jet diameter, D, is constant at 5 mm. The origin of the coor-
dinate system is at the center of upper surface of the impingent
plate. The distance between the nozzle and glass plate (H) is varied
from 1 to 5 mm. For S/D = 5, 7 and 10, respectively, the jet nozzle
plates are all velocity-inlet nozzles. Only for the case of S/D = 5, 8
nozzles with reversed flow are used, and the number of the jet noz-
zles are the same as that for S/D = 5; for the jet nozzle plate, the
velocity-inlet nozzles and pressure-outlet nozzles are arranged in
a staggered manner. The position of the jet is symmetrical at the
center of the target plate. For the case of S/D = 10, the two rows
of jet nozzles are psitioned at y ¼ 25 mm, respectively. When
S/D = 7, the jet position is located at y ¼ 17:5 mm, respectively.
When S/D = 5, the four rows of jet nozzles are positioned at
y ¼ 12:5 mm and y ¼ 37:5 mm, respectively.
The jet inlet velocity is varied to set the total mass flow rate, m
30.34 kg/h, which corresponds to jet-inlet Reynold numbers of
30,000 for the four inlet nozzles, respectively. The distance
between the nozzles (S) is varied to set the nondimensional hole-
to-hole spacing, S/D = 5, 7 and 10, respectively. The detailed
array-jet impingement nozzle arrangements are depicted in
Fig. 2. The parameters used for different array-jet arrangements
are listed in Table 1.
In the process of glass tempering, not only is the dependence of
the surface temperature on time nonlinear, but the material prop-
erties also change significantly with temperature. In our study, the
density of glass is fixed at 2500 kg/m3, while the other thermal
characteristics vary with the temperature. Moreover, the thermal
characteristics of the fluid flow also vary with temperature. The
initial temperature of glass plate is set at 953 K, and we consider
the heat-transfer characteristics through conjugate heat transfer
between the fluid and solid domains. In Fig. 1, the top part is con-
sidered as the fluid domain, whereas the bottom is the solid
domain, in which the material is glass.
The flow is assumed to be a three-dimensional, incompressible
turbulent flow. A structural mesh was created for the entire
domain using the O-block option, with the ANSYS ICEM CFD 15.0
package. In order to improve the grid quality for capturing the
near-wall flow phenomenon, and the region near the target plate,

Fig. 1. (a) Basic geometry and (b) the genmetry of array-impingement-jet with reversed flow.
 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425 415

(a) S/D = 10, 4 nozzles (b) S/D = 7, 4 nozzles

Inlet Outlet

(c) S/D = 5, 16 nozzles (d) S/D = 5, 8 nozzles with reversed flow

Fig. 2. Detailed array jet impingement nozzle arrangements.

Table 1 In our study, the SST K-x turbulent model was employed. This
Parameters used for array jet impingement nozzle arrangements. model utilizes the K-x model in the near-wall region and switches
S/D N u (m/s) Re _ (kg/h)
m to the k-e model in the far field. The equations for the turbulence
10 4 87.64 30,000 30.34
kinetic energy, k, and the specific dissipation rate, x, are,
7 4 87.64 30,000 30.34  
@ @ @ @k
5 8 43.82 14,999 30.34 ðqkÞ þ ðqkui Þ ¼ Ck þ G k  Y k þ Sk ð4Þ
5 16 21.91 7499 30.34 @t @xi @xj @xj
 
@ @ @ @k
ðqxÞ þ ðqxui Þ ¼ Cx þ Gx  Y x þ Dx þ Sx ð5Þ
@t @xi @xj @xj
where the flow and temperature gradients are expected to be the
highest, the grid was refined near the glass plate. Some of the other Here C, G and Y are the effective diffusivity, the generation, and
similar studies [13,37] have used the same mesh resolution as the dissipation of the corresponding variables, respectively. Dx is
ours. An exemplary grid used in circular air array-jet impingement the cross-diffusion term. Sk and Sx are user-defined source terms.
with a jet-to-plate distance, H = 2 mm, and nondimensional jet-to-
jet spacing, S/D = 7, is shown in Fig. 3. 2.3. Boundary conditions

2.2. Governing equations (1) Air jet nozzle: The nozzle is specified as a velocity inlet, the
air was injected with a uniform velocity at a different total
ANSYS fluent was used for solving the governing equations of mass flow rate, with the turbulence intensity of the flow
mass, moment, turbulence and energy, which are, being 5%. The inlet temperature was at a constant tempera-
@q ture of 298 K.
þ r  ðquÞ ¼ 0 ð1Þ In most studies of array jet impingement, the impingement
@t
Reynolds number is determined on the bulk-average of multiple
@ jets, and defined as:
ðquÞ þ r  ðquuÞ ¼ rðs
 Þ  rp ð2Þ
@t uj dj _
4m
Rej ¼ ¼ ð6Þ
@ m Npldj
ðqIt Þ þ r  ðuðqIt þ pÞÞ ¼ r  ðkrT þ ðs
  UÞÞ ð3Þ
@t where d is the jet diameter, m is the kinematic viscosity of the jet, uj
Here q represents the density, P is the pressure, k is the thermal is the velocity of the jet inlet, l is the dynamic viscosity of the jet, m
_
conductivity, u is the velocity, s
 is the stress tensor, and It is the is the total mass flow rate of the jet and N is the number of nozzles
total enthalpy. in the orifice plate.
416 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425

Fig. 3. Mesh of the computational domains.

(2) Upper wall: The wall in the fluid domain is adiabatic and has included in the simulation to obtain more accurate results for
no slip. describing the jet dynamics.
At Z = H, - 2L 6 X 6 2L and - W
2
6Y6W2

@T 3. Grid independence analysis

¼0 ð7Þ
@z
The near wall resolution guaranteed the dimensionless wall dis-
(3) Flow outlet boundary: Flow outlet is specified as a pressure
tance yþ of the first node to less than 1 due to the requirements of
outlet. The pressure of the flow outlet is the same as the
the SSTK  x turbulent model, and the minimum size near the wall
ambient pressure.
was 1.52621e06m, and the minimum size varied with the jet
At X ¼  W2 , 0 6 Z 6 H, X ¼ W2 , 0 6 Z 6 H, Y ¼  2L, 0 6 Z 6 H and
Reynolds number for the other cases. Based on the surface temper-
Y ¼ 2L, 0 6 Z 6 H _ =30.34 kg/h and
ature variation as a function of cooling times for m
P ¼ Pamb ð8Þ H/D = 0.4, the results of grid independence study is shown as an
example in Fig. 4. A careful grid independence check has been done
where P amb is the ambient pressure. by considering several grid system with nodes ranging from
(4) Flow top outlet boundary: For the case of S/D = 5, the top 602,780 to 1,174,620. The nodes ranging from 602,780 to
flow outlet is also a pressure outlet and the pressure is the 1,174,620 corresponding to the jet-to-plate distance (H) of 2 mm
same as the ambient pressure. where the boundary layer is used 15-layer cells to 30-layer cells,
(5) Impingement plate: The upper part of the impingement respectively. It can be seen that when the nodes number is larger
plate is specified as fluid–solid coupling boundary condition. than 602,780, the differences of the surface temperature is less
The remainder is specified as a convective heat transfer coef- than 1%. As a result, a nodes number of 820,060 was used in our
ficient boundary condition. simulations to ensure the grid independence results and fine reso-
lution of the heat transfer distribution.
At Z = 0, - L
2
6 X 6 2L and - W
2
6Y6W
2

@T 4. Comparison of turbulent models

kp ¼ hc ðT p  T air Þ ð9Þ
@z
where kp is the thermal conductivity of the target plate, T p is the Hofmann et al. [38] concluded that SSTk  x gave satisfactory
surface temperature of the target plate, T air is the air temperature, results for the small nozzle to plate distances for modeling single
and hc is the convection heat transfer coefficient. jet impingement heat transfer. Xing et al. [39] also reported that
At Z ¼ Hp ,  2L 6 X 6 2L and  W 6Y6W , x ¼  W2 , 0 6 Z 6 Hp , SST k-w model is the best suitable model for modeling the jet
2 2
x ¼ 2 , 0 6 Z 6 Hp , y ¼  2, 0 6 Z 6 Hp and y ¼ 2L , 0 6 Z 6 Hp
W L

@T
kp ¼ hext ðT ext  T p Þ ð10Þ
@z 1000

where Hp is the height of the target plate, hext is the external heat 950 602780
transfer coefficient defined by us, T ext is the external heat-sink tem-
Surface temperature (K)

900 820060
perature defined by us. 850 1174620
(6) Initial condition: When cooling time t = 0, the air flow began
800
impinging and the plate surface temperature decreased. The
initial temperature of the impingement plate (total solid 750
domain) is set to a value of 953 K. 700
650
2.4. Simulation
600
Numerical simulations were performed with the use of the 550
commercial computational fluid dynamics (CFD) solver Fluent 500
15.0; the flow, turbulence and energy equations were solved, and 450
the SIMPLEC algorithm was applied to the pressure velocity cou- 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
pling. The two-order upwind discretization scheme was used for Cooling time (s)
the pressure, momentum, turbulent kinetic specific dissipation
rate and energy. In our study, the SSTK  x turbulent model was _ = 30.34 kg/h, H/D = 0.4.
Fig. 4. Grid independence at m
 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425 417

impingement phenomenon for an inline and staggered array of jet
with constant wall temperature. Most of the numerical simulations
have been carried out for a steady state, but our study focuses on a
transient state. Therefore, it is important to study how different
turbulence models predict the heat transfer characteristics in an
1000
900
Surface temperature (K)
array for a transient state. We simulated the surface temperature
variation as a function of cooling time using different turbulent
models to compare with available experimental data [29,30].
Fig. 5 shows the comparison between numerical results based on
different turbulence models and experimental results at Re =
30,000, H/D = 1. It was found that the SSTk  x turbulence model
performed better than others. Thus, the SSTk  x turbulence model
is selected for the followed simulations.
Experimental
RNG k- 5. Results and discussion

SST k-
Realizable k- 5.1. Effects of H/D and S/D on transient heat transfer
800
For the simulation, we set the initial temperature of the glass
700 plate to 953 K. Following exposure to a sudden cooling process
by air-jet impingement, its surface temperature rapidly decreases
to 473 K. We define the time taken for the surface temperature
600 to decrease from 953 to 473 K as the quenching time. Figs. 6 and
7 show the dependence of the average impingement-surface tem-
500 perature and the glass-plate-center average surface temperature
on the cooling time, for different values of H/D and S/D
0 2 4 6 8 10 12 14 16 18 20 22 24 (D = 5 mm) at a constant m,_ respectively.
cooling time (s) From the results in Figs. 6 and 7, the quenching time first
increases and then decreases, as H/D decreases, and there is a turn-
Fig. 5. Comparison between different turbulence model and experiment. ing point, which does not change with S/D. The maximum quenching

1000 1000
950 S/D=10 4 nozzles 950 S/D=7 4 nozzles
H/D=0.2 H/D=0.2
Surface temperature (K)
Surface temperature (K)

900 900 H/D=0.4

H/D=0.4
850 850 H/D=0.6
H/D=0.6
800 H/D=0.8 800 H/D=0.8
750 H/D=1 750 H/D=1
700 700
650 650
600 600
550 550
500 500
450 450
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Cooling time (s) Cooling time (s)
(a) (b)

1000 1000
950 950
S/D=5 16 nozzles H/D=0.2 S/D=5 with reversed flow
Surface temperature (K)

Surface temperature (K)

900 900
H/D=0.4 H/D=0.2
850 850
H/D=0.6 H/D=0.4
800 H/D=0.8 800
H/D=0.6
750 750 H/D=0.8
700 700
650 650
600 600
550 550
500 500
450 450
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30
Cooling time (s) Cooling time (s)
(c) (d)
_ = 30.34 kg/h.
Fig. 6. The dependence of average impingement surface temperature on cooling time for different H/D and S/D at m
418 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425

1000 1000

950 950
H/D=0.2 H/D=0.2
S/D=10 4 nozzles S/D=7 4 nozzles

Plate center surface temperature (K)

Plate center surface temperature (K)

900 H/D=0.4 900 H/D=0.4

850 H/D=0.6 850 H/D=0.6
800
H/D=0.8 800
H/D=0.8
H/D=1 H/D=1
750 750

700 700

650 650

600 600

550 550

500 500

450 450
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Cooling time (s) Cooling time (s)
(a) (b)

1000 1000

950 950 S/D=5 with reversed flow

H/D=0.2 H/D=0.2
S/D=5 16 nozzles
Plate center surface temperature (K)

Plate center surface temperature (K)

900 H/D=0.4 900 H/D=0.4
850 H/D=0.6 850 H/D=0.6
H/D=0.8 H/D=0.8
800 800

750 750

700 700

650 650

600 600

550 550

500 500

450 450
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30
Cooling time (s) Cooling time (s)
(c) (d)
Fig. 7. The dependence of glass plate center average surface temperature on cooling time for different H/D and S/D at a constant mass flow rate.

time is always obtained at H/D = 0.4, regardless of the S/D. It was also is increasingly obvious, increasing the flow velocity of the nearby
revealed that the value of H/D corresponding to the minimum impingement surface. This phenomenon is explained in our previ-
quenching time varies with S/D. When S/D = 7 and 10, the minimum ous article [40]. Due to this, when S/D is smaller, the shortest
quenching time was obtained at H/D = 1, whereas when S/D = 5, the quenching time is obtained at the least H/D. However, when S/D
minimum quenching time was obtained at H/D = 0.2. This indicates is larger, jet deflection behavior also exists, and a larger jet inlet
that for the densest array arrangement, the least value of H/D obtains velocity leads to greater fluid velocity in the nearby impingement
the shortest quenching time. However, this shortest quenching time surface. When H/D is smaller, the space for the jet fluid is com-
obtained for S/D = 5 and H/D = 0.2 is the maximum quenching time pressed, impeding jet-to-jet interaction; therefore, heat transfer
for all of the array-impingement arrangements [22]. This is because is further deteriorated for the least H/D.
the quenching time initially decreases and then increases, as Comparing (c) and (d) in Fig._6 and 7, we observe that the
the S/D ratio increases, and the maximum quenching time is quenching time obviously decreases for S/D = 5, which includes
obtained at S/D = 5. reversed-flow nozzle. This phenomenon occurs because the exist-
The variation in the quenching time for different array arrange- ing reversed-flow nozzles decrease the number of inlet nozzles
ments is due to the nozzle inlet velocity, ratio of the overall and increase the inlet velocity, enhancing convective heat transfer
impingement jet area to the entire impingement plate, and the at constant S/D and m. _ Moreover, from Figs. 6 and 7, we can
effect of the jet deflection, which is mainly caused by small jet- observe that the tendency of the dependence of the glass-plate
to-plate distances. For the same H/D, the smaller the S/D, the lesser center average surface temperature and the average impingement
is the jet inlet velocity, while the ratio of the impingement area to surface temperature on cooling time is consistent. Besides, the
the entire impingement plate is larger. The inlet velocity for S/D = 7 average temperature of the glass-plate center is higher than the
is increased by a factor of four, relative to that of S/D = 5, for a con- impingement surface, mainly because the cooling of the impinge-
stant m._ Therefore, for different S/D, the nozzle inlet velocity is a ment plate depends on convective heat transfer, while the glass-
dominant factor for the quenching time, whereas the change in plate center depends on heat conduction and natural convection;
the ratio is a secondary factor. The smaller the S/D, the lesser is therefore, the average surface temperature of the impingement
the inlet velocity. On decreasing H/D, the jet deflection behavior surface is lower.
 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425 419

35 To further verify the impact of different nozzle arrangements on

S/D=7 the quenching times, compared to S/D = 5, we define the quenching
Quenching times reduction ratio (%)

S/D=10
30 time reduction ratio, f as shown in Fig. 8.
S/D=5 with reversed flow

25 f ¼ ðtother  tS=D¼5 Þ=tS=D¼5 ð11Þ

20 It can be observed that, when reversed-flow nozzles are

included in the array-jet impingement, f apparently increases.
15
31.74
29.78 When H/D increases from 0.2 to 0.8, for each nozzle arrangement,
27.63 26.94 27.78 the reduction in the quenching time is 21.19–31.74% for S/D = 7,
23.42
10 21.19 21.58
20 20.51 16.42–29.78% for S/D = 10, and 16.42–21.58% for S/D = 5, for which
16.42 16.42 reversed flow nozzles exist.
5
5.2. Surface and internal heat-transfer distribution
0
0.2 0.4 0.6 0.8 Uniform heat transfer on the glass-plate surface is more impor-
H/D tant for obtaining better glass tempering. Figs. 9 and 10 display the
contours of the transient surface temperature for different inlet
Fig. 8. The dependence of quenching times reduction ratio on H/D for different nozzle arrangements, and the dependence of the center of the
nozzle arrangements.
impingement plate (glass plate) surface-temperature contour on

_ = 30.34 kg/h, t = 12 s, H/D = 0.2), (a) S/D = 10, (b) S/D = 7, (c) S/D = 5 with reversed flow, (d) S/D = 5.
Fig. 9. Transient temperature contour (m
420 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425

the cooling time, respectively. From the results of Fig. 9, the perature region at the center of the plate surface. Moreover, on
temperature of the stagnation point region is the least. For inline decreasing S/D, the highest-temperature-region area decreases,
array-jet impingement, there always exists a highest surface tem- the highest surface temperature decreases, and the least surface

(a) t = 1 s (b) t = 4 s

(c) t = 8 s (d) t = 12 s

(e) t = 16 s (f) t = 20 s
Fig. 10. The dependence of the center of the impingement plate (glass plate) surface temperature contour on cooling time for S/D = 5 with reversed flow.
 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425 421

(g) t = 24 s (h) t = 28 s

Fig. 10 (continued)

600 600
Surface temperature difference (K)
Surface temperature difference (K)

550 550

500 500

450 450

400 400

350 350

300 300
S/D=7 4 nozzles
250
H/D=0.2 250 H/D=0.2
S/D=10 4 nozzles H/D=0.4
200 H/D=0.4 200
H/D=0.6 H/D=0.6
150 150
H/D=0.8 H/D=0.8
100
H/D=1
100 H/D=1
50 50
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Cooling time (s) Cooling time (s)

600 600
Surface temperature difference (K)

Surface temperature difference (K)

550 550
H/D=0.2 H/D=0.2
500 S/D=5 16 nozzles 500
H/D=0.4 S/D=5 with reversed flow H/D=0.4
450 H/D=0.6 450 H/D=0.6
400 H/D=0.8 400 H/D=0.8
350 350

300 300

250 250

200 200

150 150

100 100

50 50
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30
Cooling time (s) Cooling time (s)

_ = 30.34 kg/h).
Fig. 11. Impingement plate surface temperature difference on cooling times for different S/D and H/D (m
422 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425

temperature increases, causing a smaller temperature gradient and the impingement-plate-surface average temperature. Figs. 11
along the entire impingement plate. and 12 shows the dependence of surface-temperature difference
When reversed-flow nozzles are included in the array-jet and the internal-temperature difference on cooling time, respec-
impingement, wherein the nozzle plate varies from the inline pat- tively. For different array-impingement arrangement, the surface-
tern to a similarly staggered mode, the highest surface- temperature difference always increases as H/D decreases. During
temperature region at the center of the plate surface vanishes the entire quenching process, the surface-temperature difference
and heat-transfer uniformity is obtained, which is as good as that initially increases and then decreases, and the cooling time corre-
under a small S/D condition. This occurs mainly because, when sponding to the maximum surface-temperature difference varies
the impingement jet is in an inline mode, fluid flow will be accu- for different array-arrangements. The surface-temperature differ-
mulated at the central region of the impingement surface, deterio- ence decreases monotonously with the decrease in S/D; when
rating heat transfer. The greater the fluid velocity, the more reversed-flow nozzles are included in the array arrangement, the
obvious is the heat transfer deterioration. temperature difference increases slightly. When S/D = 7, the mini-
From Fig. 10, it is revealed that with the increase in cooling mum surface-temperature difference is obtained at H/D = 1; how-
time, the glass-plate-center average surface temperature decreases ever, this obtained minimum temperature difference is still
continuously; moreover, the least temperature is always obtained larger than the maximum temperature difference obtained at
at the stagnation-point region. As the quenching process proceeds, S/D = 5 and H/D = 0.2. The main reason for this phenomenon is
the flow pattern of the impingement surface spreads from the the change in the jet inlet velocity.
stagnation-point region to the surrounding region, shaped like a From the results of Fig. 12, the internal temperature difference
jet orifice. After t = 16 s, the flow pattern does not change with still increases initially, and then decreases; however, the maximum
the cooling time. temperature difference is always obtained at a constant value of t
The surface temperature difference (DT) is defined as the differ- = 4 s. When jet impingement commences, the internal temperature
ence between the maximum and minimum surface temperatures. difference initially decreases and then increases, as H/D decreases,
The internal temperature difference is defined as the difference and the minimum temperature difference is always obtained at
between the center of the glass-plate-surface average temperature H/D = 0.4. With the increase in cooling time, the dependence of

60
60
55 H/D=0.2
Inside temperature difference (K)

H/D=0.2
Inside temperature difference (K)

55
50 H/D=0.4 H/D=0.4
H/D=0.6 50 H/D=0.6
45
H/D=0.8 H/D=0.8
40 H/D=1 45
H/D=1
35 40

30 S/D=10 4 nozzles S/D=7 4 nozzles

35
25
30
20

15 25

10 20
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Cooling time (s) Cooling time (s)

50 40

H/D=0.2 H/D=0.2
Inside temperature difference (K)
Inside temperature difference (K)

45 35
H/D=0.4 H/D=0.4
H/D=0.6 H/D=0.6
40 H/D=0.8 30 H/D=0.8

35 25

30 20
S/D=5 16 nozzles

25 S/D=5 with reversed flow 15

20 10
0 5 10 15 20 25 30 0 5 10 15 20 25 30 35 40
Cooling time (s) Cooling time (s)

Fig. 12. Temperature difference distribution internal the impingement plate (glass plate) on cooling times for different S/D and H/D.
 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425 423

Temperature difference enlargement ratio (%)

Temperature difference enlargement ratio (%)

80 110
76.19
73.29
69
H/D=0.2 70.52 100
100.23
97.76
97.56 H/D=0.4 S/D=7
S/D=10
70
66.22 92.24 S/D=5 with reversed flow
64.58 89.39
62 61.97 S/D=7 90 86.24
60.73 84.4
60 57.43 S/D=10 80.11 80.81
S/D=5 with reversed flow 80 77.79
75.24
51.45
50 70
46.13
44.61 44.54 63.19
43.25 42.1
40 60 55.73
50.98 50.12
50
44.99 44.85
30
38.87 40.08
22.48 40
33.67
20 18.55 29.79
30 28.63
14.97 26.6 26.86
12.34 12.39 12.19 12.8 12.62
10 20
0 4 8 12 16 20 24 0 4 8 12 16 20 24
Cooling time (s) Cooling time (s)
Temperature difference enlargement ratio (%)

Temperature difference enlargement ratio (%)

90 87.75 87.88 100 99.91

82.6783.38
81.51 H/D=0.6 S/D=7
S/D=10
H/D=0.8 S/D=7
S/D=10
80 90 87.76
86.64 89.07
78.89 S/D=5 with reversed flow 85.57 S/D=5 with reversed flow
85.66
70.95
70 80 78.25
66.56
67.72 77.71
62.87
60.23
60 61.35 57.73 70 68.19 67.89

55.85 62.6
63.06
62.17
50 50.63 60
53.83
45.42
53.9
40 50
37.28
35.38 35.92 45.89

42.51
30 32.03
30.22
40 42.07
29.02
27.43 38.01 38 37.65 37.8
26 35.93
33.47
20 30
0 4 8 12 16 20 24 0 4 8 12 16 20 24
Cooling time (s) Cooling time (s)

Fig. 13. The dependence of temperature difference enlargement ratio on cooling times for different nozzle arrangements.

the internal-temperature difference on H/D becomes increasingly 5.3. Instantaneous flow field
lesser. Moreover, this tendency becomes increasingly obvious as
S/D decreases. Particularly, when S/D = 5, the internal temperature The instantaneous contours of the air-impinging jet velocity
difference does not vary with H/D. In addition, for a denser array fields are used for identifying the transient heat transfer character-
arrangement, the smaller is the internal temperature difference istics. The contours of the velocity magnitude are presented for dif-
and the temperature gradient along the thickness of the glass plate. ferent values of H/D, and for the same value of m _ (m_ = 30.34 kg/h)
To further clarify the effect of nozzle arrangements on the heat- and S/D (Fig. 14).
transfer uniformity, compared to S/D = 5, we define the tempera- Fig. 14 shows that the magnitude of the overall velocity is effec-
ture difference enlargement ratio, g; as shown in Fig. 13. tively zero at the stagnation point. When the spacing of the fluid
flow is compressed, the local flow velocity is increased, and heat
g ¼ ðDT other  DT S=D¼5 Þ=DT S=D¼5 ð12Þ transfer is enhanced in the stagnation point region. Along the
downstream flow, the spacing of the fluid flow is enlarged, and
It can be observed that g increases as S/D increases. For S/D = 5, the local flow velocity is decreased, resulting in heat-transfer
with and without reversed-flow nozzles, the variation in g reduction.
increases slowly as H/D increases. However, g for S/D = 5, with The area of the impinging jet for the same value of S/D is the
reversed-flow nozzles, is still significantly lesser than that for S/ same in the stagnation-point region. Moreover, the velocity in
D = 7 and 10. the stagnation-point region and wall-jet region increases as H/D
Comparing Figs. 6 and 13, we can find that, on the one hand, the decreases, enhancing convective heat transfer.
addition of reversed-flow nozzles in the nozzle plate decreases the Fig. 15 shows the variations in the maximum flow velocity as a
heat-transfer uniformity, and the decrease is not obvious, com- function of H/D, for the same values of m _ (m_ = 30.34 kg/h) and S/D.
pared to the S/D = 7 and 10; On the other hand, the addition of From Fig. 11, we can see that the maximum velocity of the fluid
reversed-flow nozzles significantly enhances convective heat flow dramatically increases for H=D 6 0:4; as H/D decreases. More-
transfer and decrease the quenching time obviously. Therefore, over, when H/D changes from 0.4 to 0.2, the maximum fluid veloc-
the addition of reversed-flow nozzles is beneficial for transient ity increases more than 1.5 times; when H/D changes from 0.4 to 1,
heat transfer. the maximum velocity remains the same. Therefore, we conclude
424 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425

(a) = 30.34 kg/h, S/D = 10, t = 12 s, H/D = 0.2

(b) = 30.34 kg/h, S/D = 10, t = 12 s, H/D = 0.4

(c) = 30.34 kg/h, S/D = 10, t = 12 s, H/D = 0.6

(d) = 30.34 kg/h, S/D = 10, t = 12 s, H/D = 0.8

(e) = 30.34 kg/h, S/D = 10, t = 12 s, H/D = 1

_ = 30.34 kg/h, S/D = 10 and t = 12 s for different H/D.
Fig. 14. Contours of velocity magnitude of the flow of m

that a turning point exists at a certain value of H/D over the entire
range of jet-to-plate distances, which is consistent with [10]. This
175 is the reason for the existence of transient heat transfer in two
regions.
Maximum velocity (m/s)

150 6. Conclusions

The transient heat transfer characteristics for small nozzle-to-

125 plate spacing and a high temperature difference between the noz-
zle inlet and a 100 mm  100 mm glass plate were obtained
numerically. The variations in the transient heat transfer charac-
teristics at different values of H/D and S/D, as a function of the cool-
100 ing times, were studied in detail.
When S/D is large, for a larger H/D (H/D > 0.4), the heat-transfer
and uniformity are better, while the dependence of the internal-
temperature difference on H/D is negligible.
75 When S/D is small, the best heat-transfer rate and internal uni-
0.2 0.4 0.6 0.8 1.0
formity are obtained at the least H/D (H/D = 0.2), and the surface-
H/D
temperature difference is the largest; however, this maximum
Fig. 15. Variations of maximum flow velocity (m/s) as a function of H/D of internal- temperature difference is still considerably less than that
m_ = 30.34 kg/h, S/D = 10, and t = 12 s. of the minimum surface temperature for a large S/D.
 K. Zhu et al. / International Journal of Heat and Mass Transfer 127 (2018) 413–425 425

The inclusion of reversed-flow nozzles in the nozzle plate sig- [17] A. Terzis, On the correspondence between flow structures and convective heat
transfer augmentation for multiple jet impingement, Exp. Fluids 59 (2016)
nificantly increase the heat-transfer rate, whereas the surface
146–160.
and internal temperature difference increases only slightly. [18] E.Y. Jung, H. Chung, S.M. Choi, T. Woo, H.H. Cho, Conjugate heat transfer on
_ when S is constant, the heat-transfer rate is
For a constant m, full-coverage film cooling with array jet impingements with various Biot
numbers, Exp. Therm. Fluid Sci. 83 (2017) 1–8.
effectively improved by increasing the jet-nozzle diameter (D),
[19] U. Ansu, S.C. Godi, A. Pattamatta, C. Balaji, Experimental investigation of the
and the surface and internal heat transfer uniformity will also be inlet condition on jet impingement heat transfer using liquid crystal
better. When reversed-flow nozzles are included herein, the thermography, Exp. Therm. Fluid Sci. 80 (2017) 363–375.
heat-transfer rate can be improved more significantly. [20] P. Singh, S.V. Ekkad, Effects of spent air removal scheme on internal-side heat
transfer in an impingement-effusion system at low jet-to-plate spacing, Int. J.
Heat Mass Transf. 108 (2017) 998–1010.
Conflict of interest [21] H.A. Hasan, K. Sopian, A.H. Jazz, A.N. Al-Shamani, Experimental investigation of
jet array nanofluids impingement in photovoltaic/thermal collector, Solar
Energy 144 (2017) 321–334.
The authors declared that there is no conflict of interest. [22] Y. Shan, J.Z. Zhang, G.N. Xie, Convective heat transfer for multiple rows of
impinging air jets with small jet-to-jet spacing in a semi-confined channel, Int.
J. Heat Mass Transf. 86 (2015) 832–842.
Acknowledgements
[23] S. Caliskan, S. Baskaya, T. Calisir, Experimental and numerical investigation of
geometry effects on multiple impinging air jets, Int. J. Heat Mass Transf. 75
This study was supported by the National Natural Science (2014) 685–703.
Foundation of China (Grant No. 51335002) and Jiangsu Province [24] X.C. Li, J.L. Gaddis, T. Wang, Multiple flow patterns and heat transfer in
confined jet impingement, Int. J. Heat Fluid Flow 26 (2005) 746–754.
Strategic Emerging Industry Development Project (2015). [25] R. Gharraei, A. Vejdani, S. Baheri, A.A. Davani, Numerical investigation on the
fluid flow and heat transfer of non-Newtonian multiple impinging jets, Int. J.
References Thermal Sci. 104 (2016) 257–265.
[26] L.B.Y. Aldabbagh, A.A. Mohamad, A three-dimensional numerical simulation of
impinging jet arrays on a moving plate, Int. J. Heat Mass Transf. 52 (2009)
[1] A. Büyükyildiz, Automation of glass tempering furnace by using PLC,
4894–4900.
Pamukkale Univ. J. Eng. Sci. 13 (2007) 247–256.
[27] M. Draksler, B. Končar, L. Cizelj, B. Ničeno, Large Eddy Simulation of multiple
[2] S. Türkbas, Ö.E. Ataer, Numerical modelling of heating and cooling processes in
impinging jets in hexagonal configuration – flow dynamics and heat transfer
glass tempering with mixed boundary conditions, J. Fac. Eng. Arch. Gazi Univ.
characteristics, Int. J. Heat Mass Transf. 109 (2017) 16–27.
22 (2007) 727–738.
[28] P.S. Penumadu, A.G. Rao, Numerical investigations of heat transfer and
[3] F. Monnoyer, D. Lochegnies, Heat transfer and flow characteristics of the
pressure drop characteristics in multiple jet impingement system, Appl.
cooling system of an industrial glass tempering unit, Appl. Therm. Eng. 28
Therm. Eng. 110 (2017) 1511–1524.
(2008) 2167–2177.
[29] H. Yazici, M. Akcay, M. Golcu, M.F. Koseoglu, Y. Sekmen, Experimental
[4] M. Golcu, H. Yazici, M. Akcay, M.F. Koseoglu, Y. Sekmen, Experimental
investigation of the transient cooling characteristics of an Industrial Glass
investigation of cooling with multiple air jets on auto glass tempering, J. Fac.
tempering unit, World Acad. Sci. Eng. Technol. 61 (2012) 207–211.
Eng. Arch. Gazi Univ. 27 (2012) 775–783.
[30] H. Yazici, M. Akcay, M. Golcu, M.F. Koseoglu, Y. Sekmen, Experimental
[5] J.Y. San, M.D. Lai, Optimum jet-to-jet spacing of heat transfer for staggered
investigation of the transient temperature distribution and heat transfer by
arrays of impinging air jets, Int. J. Heat Mass Transf. 44 (2001) 3997–4007.
jet impingement in glass tempering processing, IJST Trans. Mech. Eng. 39
[6] H. Yazici, Determination of optimum cooling unit configuration in automobile
(2016) 337–349.
glass tempering process by different Reynolds number, PhD Thesis, Karabuk
[31] S.J. Yi, M. Kim, K. Dong, H.D. Kim, K.C. Kim, Transient temperature field and
University, Turkey.
heat transfer measurement of oblique jet impingement by thermographic
[7] R.B. Kalifa, S. Habli, N.M. Saïd, H. Bournot, G.L. Palec, Parametric analysis of a
phosphor, Int. J. Heat Mass Transf. 102 (2016) 691–702.
round jet impingement on a heated plate, Int. J. Heat Fluid Flow 57 (2016) 11–
[32] Q. Guo, Z. Wen, R. Dou, Experimental and numerical study on the transient
23.
heat-transfer characteristics of circular air-jet impingement on a flat plate, Int.
[8] M.J. Rau, E.M. Dede, S.V. Garimella, Local single and two-phase heat transfer
J. Heat Mass Transf. 104 (2017) 1177–1188.
from an impinging cross-shaped jet, Int. J. Heat Mass Transfer 79 (2014) 432–
[33] S. Luhar, D. Sarkar, A. Jain, Steady state and transient analytical modeling of
436.
non-uniform convective cooling of a microprocessor chip due to jet
[9] X.T. Trinh, M. Fénot, E. Dorignac, The effect of nozzle geometry on local
impingement, Int. J. Heat Mass Transf. 110 (2017) 768–777.
convective heat transfer to unconfined impinging air jets, Exp. Therm. Fluid Sci
[34] M. Aamir, L. Qiang, W. Hong, Z. Xun, J.Q. Wang, M. Sajid, Transient heat
70 (2016) 1–16.
transfer performance of stainless steel structured surfaces combined with air-
[10] A.M. Kuraan, S.I. Moldovan, K.S. Choo, heat transfer and hydrodynamics of free
water spray evaporative cooling at high temperature scenarios, Appl. Therm.
water jet impingement at low nozzle-to-plate spacings, Int. J. Heat Mass
Eng. 115 (2017) 418–434.
Transf. 108 (2017) 2211–2216.
[35] L. Yang, Y. Li, P.M. Ligrani, J. Ren, H. Jiang, Unsteady heat transfer and flow
[11] K.S. Choo, S.J. Kim, Heat transfer characteristics of impinging air jets under a
structure of a row of laminar impingement jets, including vortex development,
fixed pumping power condition, Int. J. Heat Mass Transf. 53 (2010) 3366–3371.
Int. J. Heat Mass Transf. 88 (88) (2015) 149–164.
[12] K.S. Choo, T.Y. Kang, S.J. Kim, the effect of inclination on impinging jets at small
[36] T.L. Fu, Z.D. Wang, X.T. Deng, J. Hun, G.D. Wang, Surface heat transfer study of
nozzle-to-plate spacing, Int. J. Heat Mass Transf. 55 (2012) 3327–3334.
ultra heavy plate during quenching process of jet array impingement, Appl.
[13] V. Jensen, H. Walther, Numerical analysis of jet impingement heat transfer at
Therm. Eng. 117 (2017) 522–533.
high jet Reynolds number and large temperature difference, Heat Transf. Eng.
[37] M.L. Hosain, R.B. Fdhila, A. Daneryd, Heat transfer by liquid jets impinging on a
34 (2013) 801–809.
hot flat surface, Appl. Energy 164 (2016) 934–943.
[14] T.T. Zhou, D. Xu, J. Chen, C.M. Cao, T.H. Ye, Numerical analysis of turbulent
[38] H.M. Hofmann, R. Kaiser, M. Kind, H. Martin, Calculations of steady and
round jet impingement heat transfer at high temperature difference, Appl.
pulsating impinging jets—an assessment of 13 widely used turbulence models,
Therm. Eng. 100 (2016) 55–61.
Numer. Heat Transf. Fundament. 51 (2007) 565–583.
[15] Y. Ichikawa, M. Motosuke, Y. Kameya, M. Yamamoto, S. Honami, Three-
[39] Y.F. Xing, S. Spring, B. Weigand, Experimental and numerical investigation of
dimensional flow characterization of a square array of multiple circular
heat transfer characteristics of inline and staggered arrays of impinging jets, J.
impinging jets using stereoscopic PIV and heat transfer relation, J. Visualiz. 19
Heat Transf. 132 (2010) 53–58.
(2016) 89–101.
[40] P.P. Yu, K.Q. Zhu, Q. Shi, N.Y. Yuan, J.N. Ding, Transient heat transfer
[16] H. Shariatmadar, S. Mousavian, M. Sadoughi, M. Ashjaee, Experimental and
characteristics of small jet impingement on high-temperature flat plate, Int.
numerical study on heat transfer characteristics of various geometrical
J. Heat Mass Transf. 114 (2017) 981–991.
arrangement of impinging jet arrays, Int. J. Therm. Sci. 102 (2016) 26–38.