Research Article

Advances in Mechanical Engineering

2021, Vol. 13(12) 1–13
Ó The Author(s) 2021
Anti-icing hot air jet heat transfer DOI: 10.1177/16878140211066212
journals.sagepub.com/home/ade
augmentation employing inner
channels

Farooq Saeed1 , Kamran Z Ahmed2, Amro OE Owes1 and

Ion Paraschivoiu3

Abstract
Many approaches exist today that employ hot-air from aircraft compressor bleed for anti-icing critical aircraft surfaces.
This paper introduces and numerically analyzes the novel application of an inner or etched channel to augment heat
transfer from a hot-air jet impinging on a curved surface representing the inner surface of an aircraft wing’s leading edge
or slat. The study shows that proper positioning, geometry, and flow characteristics of a channel along the inner surface
of the leading edge can significantly enhance heat transfer, boost the anti-icing system performance, and greatly enhance
flight safety during critical icing weather conditions. Commercially available CFD software, ANSYS Fluent is used to
model and analyze the effect of different geometric and flow parameters typical of those found in small to medium
category commercial transport aircraft to help determine the optimum arrangement. These parameters include: (1) jet
nozzle height-to-slot diameter ratios from 4 to 8, (2) channel width-to-slot diameter ratios from 0.4 to 1.8, and (3)
inner-channel inlet location angles from 10° to 60°. Each configuration resulting from a combination of the above para-
meters was simulated at Reynolds numbers based on jet-slot diameter of 30,000, 60,000, and 90,000. Empirical relations
based on available experimental data are used to validate the results. The main findings of the study reveal that the jet
height-to-slot diameter ratio of 6, inner channel height-to-slot diameter ratios of 1.8, and inner-channel inlet angular
locations of 10° combination resulted in the highest heat transfer at all Reynolds number as well as higher at increased
Reynold numbers.

1
Mechanical and Energy Engineering Department, Imam Abdulrahman Bin
Faisal University, Dammam, Saudi Arabia
2
Aerospace Engineering Department, King Fahd University of Petroleum
and Minerals, Dhahran, Saudi Arabia
3
Mechanical Engineering Department, Polytechnique Montreal, Montreal,
QC, Canada

Corresponding author:
Farooq Saeed, Mechanical and Energy Engineering Department, Imam
Abdulrahman Bin Faisal University, King Faisal Street, Dammam 31441,
Saudi Arabia.
Email: [email protected]

Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License
(https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work
without further permission provided the original work is attributed as specified on the SAGE and Open Access pages
(https://us.sagepub.com/en-us/nam/open-access-at-sage).
2 Advances in Mechanical Engineering

Graphical abstract

Description of figures: (a) RAE 2822 airfoil with a modified leading edge to incorporate, (b) a typical wing leading slat, (c)
internal layout of the piccolo tube inside a typical edge slat (Courtesy of Bombardier Aerospace), (d) numerical simula-
tion of heat transfer from the hot-air jet from a piccolo tube impinging on the inner surface of the slat, and (e) numerical
model with etched channel (novel idea) for enhanced heat transfer being investigated in current study.

Keywords
Hot-air anti-icing system, jet impingement heat transfer, aircraft icing and anti-icing, CFD with heat transfer, 2D jet flows,
channel flow

Date received: 20 August 2021; accepted: 23 November 2021

Handling Editor: Chenhui Liang

Introduction its mitigation using a host of deicing and anti-icing

techniques is a focus of many experimental and numeri-
Ice accretion on critical aircraft or rotorcraft surfaces is cal simulation techniques to this day. In particular, the
a serious flight and passenger safety concern for aero- anti-icing hot-air jet impingement technique using
space community, industry, and regulators alike1 and
Saeed et al. 3

compressor bleed air is the most common system weather conditions. The hot air is drawn into the pic-
employed in many of the aircrafts to prevent ice accre- colo tube centered on the inside of the aircraft wing
tion on critical aircraft surfaces such as wings. leading edge. Holes in the piccolo tube then direct the
Extensive research has been carried on fluid jet impin- hot air to impinge upon the inner surface of the wing
gement with broad applications ranging from heat or leading edge to heat it and, therefore, prevent ice accu-
mass transfer to ultra-precision surface polishing.2–8 mulation but at the cost of engine performance.
For the past several decades, extensive research has Extensive literature exists on jet-impingement heat
been carried out with a focus on jet impingement heat transfer to cool flat as well as curved surfaces such as
transfer on concave surfaces.2–4 The current study the leading edge of a wing or turbine rotor blade.
introduces and numerically analyses the novel applica- Lately, the focus has moved to include surface heating
tion of an inner or etched channel to augment heat applications to aid in better understanding and design
transfer from a hot-air jet impinging on a curved sur- of hot-air based anti-icing systems.19–24
face representing the inner surface of an aircraft wing’s In order to optimize the use of jet engine
leading edge or slat. A recent similar study with focus compressor-bleed air, there is a renewed interest in
on the effect of external flow parameters on thermal improving the anti-icing system performance. Many
performance of an inner-liner anti-icing system with jet studies focus on the interaction between a hot jet and a
impingement heat transfer has been reported in litera- curved surface typical of an aircraft wing leading edge
ture.4 The present study, however, examines the effects to gain insight into the physical mechanism and draw
of internal hot-air flow as well as geometric parameters heat transfer correlations to better predict the heat
on thermal performance of an inner-liner based anti- transfer as a function of key geometric and flow para-
icing system with jets impingement heat transfer. meters. These key parameters include jet diameter, jet
The steady rise in global aviation traffic means an flow Reynolds number based on jet diameter, jet
increased likelihood of operation in and encounter with nozzle-to-surface height, jet nozzle-to-nozzle distance
natural icing conditions. Studies9,10 of the National and distribution, to name a few. This has been the com-
Transportation Safety Board (NTSB) online database mon practice in industrial applications involving jet
and NASA Aviation Safety Reporting System (ASRS) impingement heat transfer on flat surfaces.25
reports revealed 2235 icing related accidents and inci- In 1977, Martin26 performed a broad study to derive
dents in the US from 1978 to 2020. A significant num- empirical correlations for diverse combinations of sin-
ber of the icing related accidents resulted in 100% gle and arrays of hot-air jet impinging on a flat surface.
fatality and the cause was attributed to ice build-up on One such empirical correlation for single hot-air slot
critical aircraft surfaces such as wings and tails. In (2D) jet heat transfer in terms of average Nusselt num-
keeping pace with the advancement in technologies ber distribution as a function of Reynolds number
related to operation of aircraft in all-weather condi- based on slot jet diameter Re, slot nozzle-to-target sur-
tions, regulatory authorities are continuously trying to face dimensionless height z/s, and dimensionless dis-
improve in-flight safety standards as well as certifica- tance measured from stagnation point x/s, is given by
tion criteria to ensure safety. It was also documented as the correlation:
one of the main tasks in the FAA In-Flight Aircraft
Icing Plan11 to ensure the reliability and validation of Nu 1:53
simulation methods of icing and anti-icing being = Rem ð1Þ
Pr0:72 x=s + z=s + 1:39
used in the certification of aircraft. Research into simu-
lation of hot-air anti-icing systems continue to draw h i1
where m =0:695  x=s + ðz=sÞ1:33 +3:06 and s =2d:
interest12–18 to this day not only in the aerospace field
but also in renewable energy field where ice throw miti- The above empirical correlation offers good agreement
gation from wind turbine blades is a major safety issue. with the experiments in that 95% of the experimental
The research reported here is part of that greater effort data lay within 10% of the above correlation with a
in that it is aimed at greater understanding as well as range of validity: 3000 ł Re ł 90,000, 4 ł x/s ł 50,
better prediction of hot-air based anti-icing heat trans- and 4 ł z/s ł 20. In the absence of similar correlations
fer requirements. The results presented in this paper for curved surfaces, numerical predictions reported in
will aid in developing numerical correlations for hot-air the current study are intentionally shown in compari-
jet heat transfer on typical aircraft wing/slat leading son with the above empirical correlation to highlight
edge surfaces where effectiveness of anti-icing systems the differences both quantitatively and qualitatively.
is of prime concern. Many experimental and numerical studies have
Hot-air anti-icing systems draw hot air from the investigated various techniques to enhance heat trans-
compressor bleed-port of the engine to heat the leading fer so as to increase the efficiency of the jet impinge-
edge of the wings of an aircraft to prevent formation ment based anti-icing system. Use of water jet to study
and accumulation of ice during operation in icing the effect of intermittency on convective heat transfer
4 Advances in Mechanical Engineering

was investigated by Zumbrunnen and Aziz.27 Surface have also been experimentally studied42 on leading hot-
pressure and heat transfer characteristics of a round jet air heating effectiveness of engine inlet guide strut. The
impinging normal on isothermal plate were experimen- improvement in the overall heating effectiveness was
tally studied by Tawfek.28 Many studies have shown attributed to the reduction of normal jet impinging dis-
that various mechanisms can be used to enhance heat tance and the formation of recirculation flow structure
transfer over a larger area such as jet intermittency29 or inside the concave cavity. It was also indicated that hot-
increased turbulence.30,31 air heating effectiveness was effected by the normal jet-
Zhou et al.32 suggest use of mesh screens while Hee to-leading edge distance.
Lee et al.33 propose use of perforated plates in between Numerical simulation of internal flow in aircraft
the jet nozzle-exit and impingement surface to enhance anti-icing systems with and without inner channel can
heat transfer through increased turbulence. Papadakis provide greater insight into the physics and further
and Wong34 investigated the effect of different para- enhance understanding of the mechanism for improved
meters on performance of bleed air ice protection sys- system performance and enhanced flight safety.
tems. Patel and Roy,35 Zhou and Lee,36 Saeed,37 and Recently, a similar study with focus on the effect of
Fregeau et al.38 have looked at the effect of different external flow parameters on thermal performance of an
arrangements for an array of hot-air jets. Saeed37 and inner-liner anti-icing system with jets impingement heat
Fregeau et al.38 have suggested development of an transfer was reported in literature.4 The present study,
extensive data base for numerical correlations to aid in however, examines the effects of internal hot-air flow
design of anti-icing system39 through the Kriging-inter- as well as geometric parameters on thermal perfor-
polation technique to alleviate the huge computational mance of an inner-liner or channel based anti-icing sys-
costs associated with the optimum design of anti-icing tem with jets impingement heat transfer
systems. The research reported here is therefore also a The objective of the current study is to study the
part of that greater effort to develop a cost effective effect of various flow and geometry related parameters
design methodology for optimizing hot-air anti-icing and identify the ones that are more effective in enhan-
systems. Use of multiple jets certainly enhances surface cing heat transfer. Therefore, a numerical investigation
heat transfer over a greater area but at the cost of was carried out using commercial CFD software,
engine performance. The focus of the present study is ANSYS Fluent,43 with the objective to study the effect
to use single jet to enhance heat transfer over a greater of jet-to-surface height, channel width, channel location
area by investigating the effect of different geometric as well as jet Reynolds number on heat transfer charac-
and flow parameters. teristics so that the best possible arrangement config-
Two-dimensional heat transfer via hot-jet impinge- uration can be identified. In this regards, a curved
ment on a flat plate and the inner leading-edge surface surface similar to that of a typical wing leading edge, as
of a modified RAE 2822 airfoil was simulated by Saeed illustrated in Figure 1, was selected for the numerical
et al.39 using the commercial CFD software ANSYS analysis. The computational domain comprises of a
Fluent. Numerical simulation of a pair of slot jets piccolo tube with a slot from which issues a hot-jet and
impinging on an inclined surface was performed by impinges upon the curved surface representing the wing
Patel and Roy35 to study the effect of jet angle and leading edge.
Reynolds number on the local and average Nusselt.
Numerical investigation of convective heat transfer
from slot-jets impinging on concave cylindrical surfaces Numerical modeling using CFD
under turbulent and constant heat flux conditions was Modeling of the flow in the internal hot-air region as
carried out by Sharif and Mothe.30 well as heat transfer simulation was conducted on the
Numerical study of jet nozzle characteristics on commercial CFD software ANSYS Fluent.43 ANSYS
impingement cooling of gas turbine blade leading Fluent numerically solves the conservation of mass,
edge40 indicated that the average Nusselt number at the momentum and energy equations in a segregated or
blade leading edge increases with jet speed but decreases coupled manner. The mass, momentum, and energy
with the distance between the jet nozzle and the pres- conservation equations for a steady, two-dimensional
sure side. It was recommended to have a side entry jet flow are given by:
which will improve the performance of impingement
cooling on turbine leading edge. Fluid flow and thermal  * 
Mass : r  rV V = 0 ð2Þ
analysis for impingement cooling system with a series
of air jets have also been reported41 for varying func-  * * 
tional parameters. The results indicated that overall Momentum : r  rV V V V =  rp + r  m
Nusselt number and total entropy generation increases   ð3Þ
* * T 2 * *

with Reynold number however velocity ratio decreases rV V + rV V  r  V V I + rV g

3
with channel height. The offset-jet arrangement effects
Saeed et al. 5

Figure 1. Sketch illustrating the physical (3D) and the computational (2D) domains.

n* o
Energy : r  V V ðrE + pÞ
!
X * * ð4Þ
=r  keff rT  hj V J j + t eff  V V
j

where the first three terms on the right-hand side of

equation (4) represent energy transfer due to conduc-
tion, species diffusion, and viscous dissipation,
respectively.
For the current study, the 2D segregated implicit sol- Figure 2. The computational domain for jet impingement on a
flat surface.
ver was chosen which sequentially and iteratively solves
the conservation equations: first, the momentum equa-
tion, then the pressure corrections by continuity equa- combinations of three different jet-to-surface spacings
tion considering the current pressure and mass flow (z/d), five different channel heights (h/d), and five dif-
rate values, and finally the remaining equations includ- ferent channel inlet location angles (u) with respect to
ing energy, turbulence, and scalar equations are solved the piccolo tube center. The select combinations were
till the desired convergence criteria are met. For closure analyzed at two different Reynolds numbers. Initially a
of the turbulence model, additional transport equations flat plate computational model, shown in Figure 2, was
are employed. For solid walls, for example the wing analyzed and the results were compared with
surface or impingement wall, only the heat conduction Martin’s26 empirical correlation. The flat plate compu-
through the wall was calculated. tational domain was discretized using a 160 3 100
ANSYS Fluent’s Design Modeler was used to build structured grid as shown in Figure 3.
the physical model of the piccolo tube and associated
geometry. Fluent has the capability of generating vari-
ous types of meshes such as unstructured, structured,
Turbulence model study
and combined meshes in 2- and 3-dimensional The flat plate model was simulated with various turbu-
domains. For this study, hybrid meshes which involved lence models like one-equation Spalart-Allmaras, two-
structuring around the piccolo tube and near the wall, equation Standard K-Epsilon, RNG K-Epsilon,
and unstructured mesh in the rest part of the region Realizable K-Epsilon, and Standard K-Omega models.
were used. In order to substantiate the numerical study, The results were compared with the Martin’s empirical
grid independence studies were carried out and are pre- data and are presented in Figure 4 where the average
sented in a later section. However, an unstructured Nusselt number (Nuavg) distribution is plotted along
mesh of approximately 340k cells was found to yield the non-dimensional distance parameter x/s. The para-
good agreement with the experimental data of meter x/s represents the non-dimensionalized form of
Martin.26 In this study, a total of 48 cases were numeri- the horizontal distance x along the impingement sur-
cally simulated to study the effect of the jet-to-surface face (bottom surface in Figure 2) non-dimensionalized
height, channel width, channel location as well as jet by the hot-jet slot dimension (s = 2d, where d is the jet
Reynolds number. These cases included select slot diameter). The average Nusselt number is
6 Advances in Mechanical Engineering

Figure 3. Domain discretized using a structured grid for a flat

plate model.

Figure 5. Grid independence study.

Figure 4. Comparison of different turbulence models run for

flat plate.

determined from the local Nusselt number through the

following relation.

ðx
1
 uavg ð xÞ =
N Nulocal ðx0 Þdx0
x
0
Figure 6. The curved wall model illustrating the geometric
Comparison of the results predicted by the different parameters used in this study.
turbulence models indicate that two-equation
K-Epsilon turbulence models yield higher heat transfer (240 3 150 quadrilateral cells), and fine (320 3 200
values, with the standard K-omega turbulence model quadrilateral cells). The results in terms of average
giving the highest heat transfer near the stagnation Nusselt number distribution along the surface are
region. The results of the Spalart-Allmaras turbulence plotted in Figure 5 which shows that the fine mesh gave
model were found to be in good agreement with the highest heat transfer while the coarse mesh gave the
Martin’s empirical data. It is a simplified one-equation lowest. The medium mesh results were just slightly
model which was mainly developed for wall bounded lower than the finer one; as a result the finer mesh was
flows in aerospace applications. Being one-equation chosen so as to improve the accuracy of results. The
turbulence model, it allowed simulation of a large num- finer mesh gave a y + value of 46 which was accepta-
ber of cases due to its reduced computational time. ble since Fluent recommends a value of y + should be
Thus, the Spalart-Allmaras turbulence model was used either less than 1 or greater than 30 for the Spalart-
in the current study due to the accuracy of its predic- Allmaras turbulence model.
tion and reduced computational time. Next a curved wall model (Figures 1 and 6) was ana-
lyzed with a leading edge profile of the NACA 23012
airfoil. Figure 6 also shows the different non-
Grid independence study dimensional variables used to construct the different
As mentioned earlier, three meshes were examined: configurations namely (a) jet-to-target spacing (z/d) of
coarse (160 3 100 quadrilateral cells), medium 4, 6, and 8. (b) channel heights (h/d) of 0.4, 0.7, 1.0,
Saeed et al. 7

Figure 9. The boundary condition for the curved wall model.

Figure 7. The complete meshed domain for the curved wall simulations and on an average each case took approxi-
model. mately 36 h to converge. A mesh size of 300 3 490
quadrilateral cells was used to analyze all the cases. In
addition, several passes of grid adaption were also used
to ensure that the wall y + was within the values rec-
ommended by Fluent.
Figure 9 shows the different boundary conditions
imposed on the curved wall model. A stationary iso-
thermal wall boundary condition was imposed on the
impinging wall with a temperature (Twall) of 260 K such
as that experienced by aircraft operating under icing
weather conditions and exposed to ambient tempera-
ture (much lower than the wall temperature) on the
Figure 8. Close up view of the mesh with and without the other side. The piccolo wall is kept at a constant heat
etched channel. flux with a thickness of 0.002 m. Pressure outlet bound-
ary condition is used to define the static pressure at
flow outlets. An added advantage of this type of outlet
1.4, and 1.8. (c) channel inlet location angle (u) of 10°, boundary condition is that it defines scalar variables in
20°, 40°, and 60° with respect to the piccolo tube cen- case of back flow. The pressure outlet boundary condi-
ter. The channel inlets were placed at an angle of 10° tion requires the specification of static (gauge) pressure
with respect to the hot-air jet axis and care was taken at the outlet boundary. Since the outlet conditions are
so that as the jet-to-target spacing was varied, the angle ambient, zero gauge is defined as the pressure for pres-
remained fixed at 10°. Figures 7 and 8 show the details sure outlet boundary condition. Ambient pressure is
of the computational domain discretized using a combi- defined as the operating pressure. The mass flow rate
nation of structured and unstructured grids. Figure 8 inlet boundary condition is used to model the flow inlet
shows the close-up views of the computational mesh boundary. Air acting as ideal gas is considered to flow
around the hot-jet slot with and without the presence out from a jet with Reynolds number 60,000
of the etched channel. Again, three meshes were exam- (V = 34.959 m/s) and viscosity (v = 1.789 3 1025 kg/
ined: coarse (150 3 246 quadrilateral cells), medium m s) at 400 K (Tin). Appropriate reference values were
(225 3 372 quadrilateral cells), and fine (300 3 490 set for post processing as shown in Table 1.
quadrilateral cells). The fine mesh size of 300 3 490
quadrilateral cells was used to analyze the different
geometric and flow parameters. Results and discussions
The two-dimensional double precision solver in The results shown in Figure 10 are the validation results
Fluent was used to simulate all cases. The convergence for a flat plate, curved, and channeled baseline model
criterion for the residuals was set at 1026 for continu- with Martin’s correlation. Since a 6 15% spread in
ity, momentum, and energy terms. A 2.80 GHz Intel i5 experimental data was observed by Martin, both the
Processor desktop with 4 GB RAM was used for all upper and lower bounds of the empirical results are
8 Advances in Mechanical Engineering

Table 1. Reference values used for post processing.

Property (units) Value

Area (m2) 1
Density (kg/m3) 0.887502
Depth (m) 0.025
Enthalpy (J/kg) 102,885.3
Length (m) 0.045
Pressure (pascal) 0
Temperature (K) 400
Velocity (m/s) 33.959 m/s
Viscosity (kg/m s) 1.789 3 1025
Ratio of specific heat 1.4
Area (m2) 1
Density (kg/m3) 0.887502

Figure 11. Results for z/d = 8 for different channel heights.

Figure 10. Comparison of different numerical models with

empirical data.

plotted for comparison with numerical results. In addi- Figure 12. Results for z/d = 6 for different channel heights.
tion, the average Nusselt number (Nuavg) distribution is
plotted as a function of the x/s distance along the lower
curved surface of the wing profile (the impingement sur- which results in an enhanced heat transfer. Lastly, as a
face). It is mentioned here that since the wing operates baseline case, the results of the curved model with an
at a positive angle of attack during flight and as such inner channel simulated under same conditions is also
the ice accretes on the lower surface of the wing starting plotted for reference. The results predict a significant
around the stagnation point. Therefore, the average increase in heat transfer up to 20% compared to the
Nusselt number (Nuavg) results in this study are pre- curved model and significantly more compared to the
sented as a function of the non-dimensional curvilinear flat plate model because of the channel placed at an
distance x/s, starting from the stagnation point where angle of 10° from the center of the piccolo tube. For
the hot-jet strikes the curved wall or the impingement the channel case, again increase in the heat transfer can
surface. be attributed to increased wall jet momentum close to
The validation results show that there is an increase the inlet of the channel. The increased heat transfer
in heat transfer as we move from a flat plate model to a around the stagnation point region can be attributed to
curved surface model, this is due to the effect of curva- the local confinement and circulation due to curvature.
ture which has also been proven by previous research- The results for the different jet-to-target spacing
ers like Sharif and Mothe30 who have suggested that (z/d) are shown in Figures 11 to 13 respectively. The
there is a 20% increase in heat transfer in curved model results show that heat transfer increases when the jet-
impingement compared to a flat plate model. The cur- to-target spacing is lowered from 8 to 6 and decreases
vature is seen to increase the momentum of the wall jet jet-to-target spacing is lowered from 6 to 4 heat
Saeed et al. 9

Figure 13. Results for z/d = 4 for different channel heights.

Figure 14. Results for channel inlet placement angle of 10° for
different channel heights.
transfer. Thus, the results indicate that a jet-to-target
spacing of 6 is the optimum in case of impingement on
a curved surface. The results also show a peak in heat
transfer value around x/s = 2 followed by a gradual fall
off with distance from the stagnation point. This fact is
attributed to the start of the channel inlet at that loca-
tion which is seen to impart more momentum to the
wall jet region as a result of increased turbulence and
shear stress along the impinging wall. The validity of
this reasoning is confirmed by the fact that the peak
does not appear in the simulation case without a
channel.
Livingood and Hyrcak31 conducted an extensive lit-
erature survey of experimental and numerical studies
on heat transfer due to jet impingement on flat sur-
faces. They reported that the optimum jet-to-target
spacing for a single circular jet is around 6–7 and for a Figure 15. Results for channel inlet placement angle at 20° for
single slot jet is 8. The present study suggests that the different channel heights.
optimum jet-to-target spacing for enhanced heat trans-
fer is closer to 6 for curved surfaces. Moreover, it has
been observed that the optimum jet-to-target spacing
depends on the presence of a channel and the channel
height. The channel height (h/d) of 1.8, the maximum
value investigated in this study, at the optimum jet-to-
target spacing was observed to yield the highest heat
transfer. Further study is needed to determine the opti-
mum value of this non-dimensional channel height
(h/d). Based on these results, the jet-to-target spacing of
6 was fixed for all later simulations to determine the
optimum channel inlet location angle and the Reynolds
number.
Figures 14 to 17 present the effect of channel inlet
location angle (u) on the average Nusselt number Nuave
distribution. The simulations were performed for u val-
ues of 10°, 20°, 40°, and 60°. Amongst the 4 channel
inlet location angles investigated, the inlet location Figure 16. Results for channel inlet placement angle of 40° for
angle 10° yielded the highest overall heat transfer different channel heights.
10 Advances in Mechanical Engineering

Figure 19. Results for Reynolds number 90,000 for jet-to-

Figure 17. Results for channel inlet placement angle of 60° for target spacing of 6.
different channel heights.

Figure 18. Results for Reynolds number 30,000 at a jet-to- Figure 20. Effect of channel angle (u).
target spacing of 6.

empirical correlation for flat plate for Re = 30,000 and

distribution. The results suggest that the channel 90,000 are shown for comparison. The result shows that
should be placed as close to the stagnation region as an increase in the jet flow Reynolds number results in a
possible. All the angles were investigated at a jet-to- substantial overall increase in heat transfer distribution.
target spacing z/d of 6 since it gave the highest overall Figure 20 shows the effect of channel angle for fixed
heat transfer distribution. The simulations predicted values of channel height-to-slot diameter value of
that as the channel inlet location angle was increased h/d = 1.0, a jet-to-surface spacing of z/d = 6.0, and a
there was a decrease in the value of heat transfer. This jet Reynolds number of 60,000. The results indicate
can be attributed to the fact that as the distance of the that as the channel angle is increased, there is an overall
channel inlet increases from the stagnation region, the decrease in the surface heat transfer from the impinge-
flow and its momentum entering the channel decreases. ment surface. This can be attributed to the fact that the
A jet Reynolds study was conducted the results of wall jet region tends to increase away from the stagna-
which are shown in Figures 18 and 19. The effect of jet tion region and therefore less amount of hot-air enters
Reynolds number on Nuavg was simulated for Re values the channel resulting in a decrease in the heat transfer
of 30,000 and 90,000 in order to understand how the from the impingement wall.
change in Reynolds number affects the heat transfer on Figure 21 shows the effect of channel height-to-slot
the impinging wall and inside the channels. Martin’s diameter for fixed values of channel angle u= 10°, a
Saeed et al. 11

Conclusions
In this study, a two-dimensional numerical simulation
study on heat transfer from a hot-air jet impinging on a
curved surface resembling the leading edge of an air-
craft wing or the slat of an aircraft wing was success-
fully carried out using the commercially available CFD
software, ANSYS Fluent. The simulation is used to
model the different hot-air jet configurations typical of
those found in small to medium category commercial
transport aircraft employing hot-air anti-icing systems
in order to identify key parameters that can yield
enhanced heat transfer not only around the stagnation
region but also further downstream. The major find-
ings of this study can be summarized as follows:

 Use of inner channels in aircraft anti-icing sys-

Figure 21. Effect of channel height (h/d). tems employing hot-air jets can greatly enhance
heat transfer as compared to the conventional
systems without the inner channels thereby
improve flight safety
 The optimum hot-air jet-to-target spacing is
found to be six times the hot-air jet with and
without the inner channel.
 Critical inner channel heights should be identi-
fied for practical considerations. In the present
investigation, the best channel height is found to
be h/d = 1.8 for the optimum jet-to-target spac-
ing of z/d = 6. The results indicate that a smaller
jet-to-target spacing yields better overall heat
transfer for smaller channel heights (h/d = 0.4
and 0.7). Similarly, a higher jet-to-target spacing
of z/d = 6 gave higher overall heat transfer dis-
tribution for higher channel heights (h/d = 1.0,
1.4, and 1.8). Further study is needed to deter-
mine the optimum value of this non-dimensional
Figure 22. Effect of jet-to-target spacing (z/d). channel height (h/d).
 The best location of the inlet to the inner channel
jet-to-surface spacing of z/d = 6.0, and a jet Reynolds with respect to the hot-air jet exit in terms of the
number of 60,000. The results indicate that a change in channel inlet location angle for the cases investi-
the channel height-to-slot diameter for the optimum gated suggest that smaller inlet location angle
jet-to-surface spacing of z/d = 6.0 does not affect the leads to an increase in the heat transfer. Further
surface heat transfer from the impingement surface. investigations are needed to determine the opti-
This can be attributed to the fact that the wall jet region mum location of the channel inlet angle as well
is the same for the different channel height-to-slot dia- as for a better understanding of the underlying
meter ratios close to the jet slot, that is at the 10° chan- physics. This is currently being investigated.
nel angle and hence yields the same of heat transfer for  The channel inlet location angle which gave
these select channel height-to-slot diameter ratios. highest overall heat transfer distribution is
Figure 22 shows the effect of jet-to-surface spacing u = 10°. In addition, the results indicate that as
z/d for fixed values of channel angle u= 10°, a channel the channel inlet location angle is increased, heat
height-to-slot diameter value of h/d = 1.0 and a jet transfer decreases due to the fact that the dis-
Reynolds number of 60,000. The results indicate that a tance from the jet inlet to the channel increases
jet-to-surface spacing of z/d = 6.0 yields the optimum and as a result less amount of flow is seen enter-
surface heat transfer from the impingement surface. ing the channel.
This fact has been recorded for jet impingement on flat  Increased heat transfer was observed at higher
surfaces as well. Reynolds number and vice versa.
12 Advances in Mechanical Engineering

Author contributions science meeting and exhibition – January 2006, Reno,

Farooq Saeed: Conceptualization, methodology, simulation NV.
setup, validation. Kamran Z Ahmed: Literature review and 10. Philip-Appiah K. U.S Inflight Icing Accidents and Inci-
numerical simulations. Amro OE Owes: Write-up and editing. dents, 2006 to 2010. MS Thesis, The University of Ten-
Ion Paraschivoiu: Reviewing and editing. nessee, Knoxville, 2011.
11. United States Department of Transportation and Federal
Aviation Administration. FAA in-flight icing plan, 1997.
Declaration of conflicting interests 12. Bu X, Lin G, Shen X, et al. Numerical simulation of air-
The author(s) declared no potential conflicts of interest with craft thermal anti-icing system based on a tight-coupling
respect to the research, authorship, and/or publication of this method. Int J Heat Mass Transf 2020; 148: 119061.
article. 13. Hann R, Hearst RJ, Sætran LR, et al. Experimental and
numerical icing penalties of an S826 airfoil at low Rey-
nolds numbers. Aerospace 2020; 7: 46.
Funding 14. Lawrence Raj P, Jeong H, Roy R, et al. Computational
The author(s) received no financial support for the research, simulation of an electrically heated ice protection system
authorship, and/or publication of this article. for composite leading edges of aircraft. SAE technical
paper 2019-01-2041, 2019.
15. Takahashi T, Fukudome K, Mamori H, et al. Effect of
Research ethics characteristic phenomena and temperature on super-
The authors declare that no human or animal subjects were cooled large droplet icing on NACA0012 airfoil and axial
involved in this work. fan blade. Aerospace 2020; 7: 92.
16. Ibrahim GM, Pope K and Muzychka YS. Effects of blade
design on ice accretion for horizontal axis wind turbines.
ORCID iD J Wind Eng Ind Aerodynamics 2018; 173: 39–52.
Farooq Saeed https://orcid.org/0000-0002-6910-3003 17. Hannat R and Morency F. Numerical validation of con-
jugate heat transfer method for anti-/de-icing piccolo sys-
tem. J Aircr 2014; 51: 104–116.
References
18. Mu Z, Lin G, Shen X, et al. Numerical simulation of
1. Cao Y, Tan W and Wu Z. Aircraft icing: an ongoing unsteady conjugate heat transfer of electrothermal dei-
threat to aviation safety. Aerosp Sci Technol 2018; 75: cing process. Int J Aerosp Eng 2018; 2018: 1–16.
353–385. 19. Thomas SK, Cassoni RP and MacArthur CD. Aircraft
2. Lyu Y, Zhang J, Liu X, et al. Experimental study of anti-icing and deicing techniques and modeling. J Aircr
single-row chevron-jet impingement heat transfer on con- 1996; 33: 841–854.
cave surfaces with different curvatures. Chin J Aeronaut 20. Cornaro C, Fleischer AS and Goldstein RJ. Flow visuali-
2019; 32: 2275–2285. zation of a round jet impinging on cylindrical surfaces.
3. Fregeau M, Saeed F and Paraschivoiu I. Numerical heat Exp Therm Fluid Sci 1999; 20: 66–78.
transfer correlation for array of hot-air jets impinging on 21. Metzger DE, Yamashita T and Jenkins CW. Impinge-
3-dimensional concave surface. J Aircr 2005; 42: ment cooling of concave surfaces with high velocity
665–670. impinging air jets. J Eng Power Trans ASME 1969; 91:
4. Guo Z, Zheng M, Yang Q, et al. Effects of flow para- 149–158.
meters on thermal performance of an inner-liner anti- 22. Dyban EP and Mazur AI. Heat transfer for a planar air
icing system with jets impingement heat transfer. Chin J jet striking a concave surface. J Eng Phys 1969; 17:
Aeronaut 2021; 34: 119–132. 1321–1325.
5. Granados-Ortiz F-J and Ortega-Casanova J. Passive 23. Al-Khalil KM. Numerical simulation of an aircraft anti-
mechanical heat/mass exchange enhancement by semi- icing system incorporating a rivulet model for the runback
confined laminar parallel wall jets. Int J Mech Sci 2021; water. PhD thesis, University of Toledo, Ohio, USA,
206: 106623. 1991.
6. Wang CJ, Cheung CF and Liu MY. Numerical modeling 24. Al-Khalil KM and Potapczuk MG. Numerical modeling
and experimentation of three dimensional material of anti-icing systems and numerical comparison to test
removal characteristics in fluid jet polishing. Int J Mech results on a NACA 0012 airfoil. In: 31st aerospace
Sci 2017; 133: 568–577. sciences meeting and exhibit, AIAA paper 93-0170, Reno,
7. Guo ZQ, Zheng M, Dong W, et al. Influence of surface NV, 1993.
convex on heat transfer enhancement of wing hot air 25. Jambunathan K, Lai E, Moss MA, et al. A review of heat
anti-icing system. Acta Aeronaut Astronaut Sin 2017; 38: transfer data for single circular jet impingement. Int J
520709. Heat Fluid Flow 1992; 13: 106–115.
8. Zhou Y, Lin G, Bu X, et al. Experimental study of cur- 26. Martin H. Heat and mass transfer between impinging gas
vature effects on jet impingement heat transfer on con- jets and solid surfaces. Adv Heat Trans 1977; 13: 1–60.
cave surfaces. Chin J Aeronaut 2017; 30: 586–594. 27. Zumbrunnen DA and Aziz M. Convective heat transfer
9. Green SD. A study of U.S. inflight icing accidents, 1978 enhancement due to intermittency in an impinging jet. J
to 2002. In: AIAA paper 2006-0082, 44th AIAA aerospace Heat Transf 1993; 115: 91–98.
Saeed et al. 13

28. Tawfek AA. Heat transfer and pressure distributions of of engine inlet guide strut. Appl Therm Eng 2018; 128:
an impinging jet on a flat surface. Heat Mass Transf 357–372.
1996; 32: 49–54. 43. Swanson JA, and Chang N. ANSYS fluent user’s guide,
29. Haneda Y, Tsuchiya Y, Nakabe K, et al. Enhancement 2019R1. Lebanon, NH: ANSYS Inc, 2019.
of impinging jet heat transfer by making use of mechano–
fluid interactive flow oscillation. Int J Heat Fluid Flow
1998; 19: 115–124. Appendix
30. Sharif MA and Mothe KK. Parametric study of turbulent
slot-jet impingement heat transfer from concave cylindri- Notations
cal surfaces. Int J Therm Sci 2010; 49: 428–442.
A surface area (m2)
31. Livingood JNB and Hyrcak P. Impingement heat transfer
from turbulent air jets to flat plates: a literature survey. cp specific heat (J/kg K)
Washington, DC: Lewis Research Center, NASA, 1973. d* slot (jet) diameter (m)
32. Zhou DW, Lee SJ, Ma CF, et al. Optimization of mesh Vg gravitational acceleration (m/s2)
screen for enhancing jet impingement heat transfer. Heat h inner-channel width (m)
Mass Transf 2006; 42: 501–510. hc heat-transfer coefficient (W/m2 K)
33. Hee Lee D, Min Lee Y, Taek Kim Y, et al. Heat transfer h/d dimensionless channel width-to-slot
enhancement by the perforated plate installed between an diameter ratio
impinging jet and the target plate. Int J Heat Mass Transf I unit tensor
2002; 45: 213–217. k thermal conductivity (W/m K)
34. Papadakis M and Wong SHJ. Parametric investigation L length (m)
of bleed air protection system. In: 44th AIAA aerospace
Nu Nusselt number based on length, hcL/k
sciences meetings and exhibit, Reno, Nevada, USA, 2006,
p pressure (Pa)
pp.9–12.
35. Patel P and Roy S. Heat transfer for a pair of rectangular
Pr Prandtl number, n/a
jets impinging upon an inclined surface. In: 40th aero- R slot (jet) radius (m)
space sciences meeting and exhibit, AIAA paper 2002- Re Reynolds number based on slot-jet
0210, Reno, NV, 2002. diameter, Vd/n
36. Zhou DW and Lee SJ. Forced convective heat transfer s slot dimension, 2d (m)
with impinging rectangular jets. Int J Heat Mass Transf T temperature (K)
2007; 50: 1916–1926. Tin hot-air jet inlet temperature (K)
37. Saeed F. Numerical simulation of surface heat transfer Tw impingement-side wall temperature (K)
from an array of hot-air jets. J Aircr 2008; 45: 700–714. V jet velocity (m/s)
*
38. Fregeau M, Gabr M, Paraschivoiu I, et al. Simulation of VV velocity vector
heat transfer from hot-air jets impinging a three- x distance measured from stagnation point
dimensional concave surface. J Aircr 2009; 46: 721–726.
on impingement surface (m)
39. Saeed F, Morency F and Paraschivoiu I. Numerical
simulation of a hot-air anti- icing simulation. In: 38th
x/s dimensionless distance measured from
aerospace sciences meeting and exhibit, AIAA paper 2000- stagnation point along the lower surface
0630, Reno, NV, 2000. y+ dimensionless wall distance
40. Liu Z and Feng Z. Numerical simulation on the effect of z jet nozzle-to-target surface height (m)
jet nozzle position on impingement cooling of gas turbine z/d dimensionless jet height to slot diameter
blade leading edge. Int J Heat Mass Transf 2011; 54: ratio (jet-to-target spacing)
4949–4959. a thermal diffusivity, k/rcp (m2/s)
41. Lam PAK and Prakash KA. A numerical investigation m dynamic viscosity (Pa s)
and design optimization of impingement cooling system n kinematic viscosity, m/r (m2/s)
with an array of air jets. Int J Heat Mass Transf 2017; u inner-channel inlet angle location (°)
108: 880–900. r fluid density (kg/m3)
42. Guan T, Zhang JZ and Shan Y. Effect of offset-jets
arrangement on leading edge hot-air heating effectiveness