43rd Aerospace Sciences Meeting and Exhibit AIAA 2005-1374

10-13 January 2005, Reno, Nevada

Simulation of an Airfoil Electro-Thermal Anti-Ice System

Operating in Running Wet Regime

Guilherme Araújo Lima da Silva*

Empresa Brasileira de Aeronáutica – Embraer S. A, São José dos Campos, SP, 12227-900, Brazil.

Otávio de Mattos Silvares§

Escola Politécnica da Universidade de São Paulo, São Paulo, SP, 05508-900, Brazil.
Instituto Mauá de Tecnologia, São Caetano do Sul, SP, 09580-900, Brazil.

Euryale Jorge G. J. Zerbini‡

Escola Politécnica da Universidade de São Paulo, São Paulo, SP, 05508-900, Brazil.

An airfoil electro-thermal anti-ice system operating in a running wet regime was

mathematically modeled and simulated. The results were compared to experimental data
and results from other numerical codes. For the regions wetted by the water film, the
present model predicted airfoil solid surface temperatures within experimental error margin
and smaller deviation from experimental data than the other codes. The heat transfer
characteristics, runback water flows and water film ending point position were evaluated
and compared to other numerical codes results and experimental data. The model predicts
satisfactorily the convection heat and mass transfer effects, i.e., wet and dry convection
parcels, between the non-isothermal surface and the gaseous flow around the airfoil for
engineering purposes as well as the influence of the runback water flow on the anti-ice
system thermal performance.

Nomenclature
A = finite volume upper surface [m ], A = ∆s ⋅ 1
2

m′′ G
bh = blowing parameter, bh = evap
St
Bm = mass transfer driving force
c = airfoil chord [m]
LWC = liquid water content [g/m3]
m = mass flow [kg/s]
m′′ = mass flux [kg/(s⋅m2)]

mH 2O = mass fraction of water vapor in gaseous flow

p = pressure [Pa]
s = curvilinear distance on the airfoil surface from stagnation point [m]
sm = transition region mean position location [m]
T = temperature [K]
Ttot = total air temperature [K]
T = gaseous flow mean temperature for properties evaluation [K]
U = overall heat transfer coefficient [W/(m2⋅K)]
V = velocity [m/s]

*
M.Sc., Thermal Analysis Team Leader, Environmental Systems Engineering, Avenida Brigadeiro Faria Lima,
2170 - PC439, AIAA member
§ Ph.D., Associate Professor, Mechanical Engineering Department, Av. Prof. Mello Moraes, 2231.
Dean, Mauá University, Praça Mauá, 1.
‡ Ph.D., Doctor, Mechanical Engineering Department, Av. Prof. Mello Moraes, 2231.

1
American Institute of Aeronautics and Astronautics
Copyright © 2005 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
v(s,y) = water film velocity [m/s]
y = distance normal to the airfoil surface [m]

Greek Letters
β(s) = local water collection efficiency
σ = transition region standard deviation length [m]

Subscripts
air = mean condition at boundary layer flow
∞ = condition at free stream
d = condition at impinging water droplet
e = condition at the edge of the boundary layer
evap = water evaporation
G = condition in the gaseous flow
imp = water impingement
in = water entering the finite volume
int = interface with boundary layer condition; if it is wet, it will indicate water-boundary layer interface, if
it is fully dry, it will indicate solid wall-boundary layer interface
mixt = condition at air-steam mixture
out = water leaving the finite volume
rec = kinetic recovery for high speed heat transfer
ref = reference temperature, Tref= 273 K
S = condition near the water film surface
stag = condition at stagnation point
vap = condition at steam partial component
wall = condition at airfoil solid surface
water = mean condition at water film flow

I. Introduction

T he ice accretion on aircraft wings and stabilizers can cause some aerodynamic performance degradation, weight
increase, control and maneuver abilities difficulties that may reduce the aircraft operational safety margin. When
the aircraft is flying through a supercooled water-droplet cloud, which is in a meta-stable thermodynamic
equilibrium below water freezing temperature, the ice accretion on some aerodynamic surfaces will occur if they are
not protected adequately.
In order to protect the airfoils and guarantee safe flight in icing conditions, commercial and some military
aircraft have ice protection systems, which can be classified in de-ice and anti-ice types. The de-ice system
cyclically operates to remove the ice layer accreted after some exposition time. When the system is not actuated, the
ice is allowed to build up on the airfoil; when it is actuated, the system removes the ice from the airfoil. On the other
hand, the anti-ice system prevents the ice accretion on airfoils and continuously works if the aircraft is flying in
icing condition. Most commercial large aircraft have thermal anti-ice systems that use engine hot bleed air or
electrical heaters. The electro-thermal anti-ice system is composed of a set of electrical heaters distributed spanwise
and streamwise on the airfoil, mainly at the leading edge region.
Basically, the anti-ice systems can operate in fully evaporative, evaporative or running wet regimes depending
on the system design parameters (such as magnitude and distribution of power density, size of protected area,
external flow characteristics and airfoil geometry) and aircraft operational condition. In the former regime, the
impinging water droplets are vaporized close to the impingement region. On the other hand, in evaporative regime
operation, the runback water flows over the airfoil leading edge and evaporates in a position upstream to the end of
protected area. In the latter, the running wet regime occurs when the water runback flows downstream the end of the
protected area. Consequently, if the runback water flows to regions downstream the thermally protected zone, the
water will freeze and form runback ice. Depending on the amount, height, shape and roughness size, this residual ice
may degrade the airfoil aerodynamic characteristics and aircraft operational performance.
A thermal anti-ice numerical code can be used for conception, integrated optimization of aircraft systems,
architecture definition, ice protection system sizing and development. In the certification phase, a simulation
numerical tool can be used to define a critical cases matrix, to plan tests points and to predict system performance

2
American Institute of Aeronautics and Astronautics
within the icing operational envelope. If the software results are consistently validated and tests planned to have
only some spot checks, the icing tunnel and natural icing flight tests duration and costs can be minimized. In
addition, the numerical tool can be used to identify the actual operational regime (full evaporative, evaporative or
running wet) within the entire icing envelope. The anti-ice system may behave either as evaporative regime in some
conditions or as running wet in other conditions within icing envelope.

II. Previous Works

A comprehensive bibliographic research, performed by Silva1, covered major works published since late 40’s,
including USA´s NACA (National Advisory Committee for Aeronautics) works, UK’s RAE (Royal Aircraft
Establishment) reports and some early AIAA (American Institute for Aeronautics and Astronautics) conference
papers.
Silva and Silvares2 described the thermal anti-ice mathematical model in details and briefly summarized Silva’s
thesis results1 and conclusions. Silva, Silvares and Zerbini 3 described the mathematical model, presented the
numerical code results, validation with experimental data and compared with other codes results for an anti-ice
system operating in an evaporative regime.
Al-Khalil et al.4 implemented the numerical code ANTICE to predict parameters of a thermal anti-ice system
operation. It was developed to work together the NASA´s LEWICE code, which adopts Messinger5 thermal balance
model and was originally developed to predict ice accretion. ANTICE does not solve the hot air impinging jets flow
inside the “D” bay; however, it requires some inputs like heat flux, internal leading edge surface convection heat
transfer coefficient or temperature distributions. An improved model of water runback behavior, which considers
surface tension effects and rivulets formation, was implemented in ANTICE by Al-Khalil as a result of his previous
work6.
Henry7 developed an ice protection numerical simulation code for ONERA (Office National D'etudes et de
Recherches Aérospatiales), which can predict both anti-ice (steady state) and de-icing (transient) systems operation.
The author implemented a bi-dimensional prediction of the water freezing on the airfoil surface. In addition, a
boundary layer calculation considering the non-isothermal surface effects in the air thermodynamic properties was
implemented.
Morency, Tezok and Paraschivoiu8 implemented the numerical code CANICE for anti-ice simulation and
validated its results with Al-Khalil et al4 experimental data. The authors published results of two versions of the
main code: CANICE A, which uses a experimental heat transfer coefficient; CANICE B, which uses a convection
heat transfer coefficient predicted by thermal boundary layers equations in the integral form.
At Royal Aircraft Establishment, Cansdale and Gent9 implemented one of pioneers works regarding thermal
balance around an non-heated airfoil surface under icing conditions, which was after improved by Gent10 with
development numerical code TRAJICE. Further, the research has branched to a rotorcraft de-icing code11.
Downs and James12 and after Riley13 implemented a anti-ice simulation numerical code for engine nacelle
equipped with hot air impinging jets.

III. Objective
The model presented in this paper intends to predict electro-thermal anti-ice system operational parameters of an
airfoil with similar geometry and conditions of the reference case
presented hereafter.
s/c position Silva, Silvares and Zerbini3 presented the parameters prediction for an
Heater evaporative operation of the system. However, it is necessary to extend the
Element start end
results beyond the evaporative conditions regime because the running wet
F -0.1024 -0.0607
D -0.0607 -0.0329 regime has important particular characteristics. In relation to the
B -0.0329 -0.0051 evaporative operation: 1) the running wet regime is affected more
A -0.0051 0.0157 significantly by the coupling of convection heat and mass transfer; 2) the
C 0.0157 0.0435 film of liquid water is thicker; 3) the streamwise thermal conduction in the
E 0.0435 0.0713 solid surface is less important; 4) the operation demands smaller amount of
G 0.0713 0.1129 electrical power.
Table 1: Electrical heater elements, Therefore, the objective of the present paper is to demonstrate the
Al-Khalil et al.4 (Negative values simulation of the thermal anti-ice mathematical model3 extended to
refers to lower surface) running wet operational conditions.

3
American Institute of Aeronautics and Astronautics
 IV. Reference Case
A reference case is necessary to determine the scope of the mathematical modeling and validate the numerical
code results. The reference case was chosen based on the existence of reliable experimental data in open literature
and on the similitude of the experimental data with an actual anti-ice operation.
Al-Khalil et al.4 performed anti-icing experiments at Icing Research Tunnel at NASA Glenn Research Center
facilities, Cleveland, Ohio, USA. The authors measured surface
temperature and overall heat transfer coefficient in order to Heater Case 67A Case 67B
validate LEWICE/ANTICE numerical code results. Several cases Element [kW/m2] [kW/m2]
were run at icing tunnel using an electrically heated airfoil for
F 20.15 8.37
anti-icing system operation simulation. The airfoil was 72 inches
D 21.70 11.94
span by 36 inches chord NACA 0012 profile with electronically
B 32.55 10.85
controlled heaters. Each heater element in streamwise direction
had one thermocouple, one thermo-resistor sensor and one heat A 43.40 15.19
flux gauge installed. There were two sets of electrical heaters C 26.35 9.92
placed side by side spanwise and each heater set contained seven E 18.60 12.87
individually controlled heaters. The total protected area was 36 G 18.60 8.68
inches spanwise (2 heater sets of 18 inches) by 7.75 inches Table 2: Anti-ice electrical heaters power
streamwise (upper and lower and surfaces at leading edge density distribution (Al-Khalil et al.4)
region). The heaters length streamwise are presented in Table 1.
For running wet regime representation, the cases 67A and 67B from Al-Khalil et al.4 experimental data sets were
selected. Test case 67A is considered a limit-case between running wet and evaporative condition because the liquid
water film covered the entire protected region length though with a negligible amount runback water ice formed
downstream the heated area. Despite test case 67B was run in the same icing conditions as 67A, the running wet
condition was more severe in the former due to the different power density distribution as presented in Table 2.
The total electrical power provided in case 67A was 2.21 kW per each heater set while for the case 67B the value
was 0.96 kW per each heater
set.
Cases 67A and 67B were
Flight
Flight conditions
conditions Atmospheric conditions
Atmospheric conditions ProfileGeometry
Profile Geometry
run at V∞=200 mph (true air
speed, TAS), angle of attack
α=0° and icing condition
defined by
Ttot=-21,7 °C, MVD=20 µm
ONERA2D
ONERA2D Pre-processor and LWC=0.55 g/m3.
Boundary
Anti -iceLayer
andSystem
Anti-ice Input
Data Pressure Field V. Numerical Code
Data Architecture
Velocity Field
Basically, the present
Water Collection Efficiency
numerical code predicts
Main code operational parameters like
solid surface temperatures,
Thermodynamic
Thermodynamic runback water and
Solver Converter
Converter
convection heat transfer
Solver
coefficient distributions
along the airfoil solid
surface.
Temperature and
The development and
Numerical Results Validation
with Experimental Data implementation of a
Runback Water Distributions
thermodynamic solver for
airfoil anti-ice system
Figure 1: Numerical Code Architecture
simulation by Silva, Silvares
and Zerbini1, 2, 3 comprised
the following: 1) integration with a recognized flow and droplet trajectory solver by a new converter routine; 2)
development of a new and original Thermodynamic solver; 3) development of new dynamic and thermal boundary

4
American Institute of Aeronautics and Astronautics
layer evaluation routines; 4) development of new graphical routines for results post-processing; 5) validation of
numerical results with recognized experimental data and other numerical codes results availavle in open literature.
Briefly, the numerical code architecture is presented in Figure 1. The main code is a thermodynamic solver,
which receives input data directly from user regarding anti-ice, icing conditions and boundary layer, and from the
converter code, which analyzes the pre-processor results and tranforms them into the thermodynamic solver input
format.

VI. Pre-Processor
A pre-processor code is required to both solve the flow field around airfoil and calculate the water droplets
trajectories. For a given two-dimensional airfoil profile, atmospheric and flight condition, it provides the pressure,
velocity and local water catch efficiency distributions around the airfoil. Similar strategy was implemented by Al-
Khalil et al.4 that uses LEWICE as a pre-processor to provide data to ANTICE code. However, the thermodynamic
solver presented herein does not require the pre-processor to solve either dynamic or thermal boundary layers.
The ONERA2D was chosen to serve as pre-processor code, for it was validated at a broad range of conditions
with experimental data14 and had its results extensively compared with other numerical codes15, 16. Therefore, the
pre-processor results are considered acceptable, not requiring further validation for the objectives of the present
paper.
The code ONERA2D is only necessary to prepare input data for the thermodynamic solver. From the pressure
and velocity fields and water collection efficiency distribution provided by pre-processor (through the converter
code as shown in Figure 1), the thermodynamic solver (main code) estimates the surface temperature and runback
liquid water flux distributions. The pre-processor modules used are: 1) Flow Solver Module, which was originally
developed and validated by Bredif17, 18 and later included in ONERA2D by Guffond and Brunet14. By using finite
elements, it solves the Full Potential flow around profile and predicts the pressure and velocity fields; 2) Water
Droplets Trajectories Module, which was developed by Guffond and Brunet14 to predict the local collection
efficiency distribution on the airfoil surface.
Other remaining modules of ONERA2D, heat transfer (boundary layer routine based on Makkonnen19) and
thermal balance (ice accretion), are not used because they are replaced by a totally new thermodynamic solver (main
code), which was specially developed for simulation of anti-ice systems1, 2, 3.

VII. Mathematical Model Description

Figure 2 shows the coordinates system and the five domains used in the present mathematical model, which are
the following: I) free stream flow; II) gaseous flow; III) boundary layer; IV) water film flow; V) solid surface. The
First Law of Thermodynamics applied to solid surface (domain V) results:


+ (1 − F ) ⋅ [− hair ⋅ (Twall − Trec )] = 0

dTwall
− F ⋅ h water ⋅ (Twall − Twater ) + Q anti
d 

"
k wall (1)


− ice
ds ds

Trec = (1 − r ) ⋅ Te + r ⋅ Tstag (2)

Tstag
Te = (3)
γ −1
1+ ⋅Me
2

The recovery factor is assumed to be r= Pr 1 2 in laminar regime and r= Pr 1 3 in turbulent regime.

A type of wetness factor F is defined in order to represent the wetted area fraction in the finite volume (F=1
when fully wet, 0<F<1 partially wet, F=0 when fully dry). The wetness factor is used at the finite volume where the
water film disappears. It represents the wetted area fraction of that finite volume, which is partially wet.
The last finite volume at the trailing edge on upper or lower airfoil surface is considered to be adiabatic.
According Eckert20 definition, the air thermodynamic properties for high-speed flow are calculated considering film
mean temperatures defined by:

Tair = Tint + 0.5 ⋅ (T∞ − Tint ) + 0.22 ⋅ (Trec − Tint ) (4)

5
American Institute of Aeronautics and Astronautics
 In order to compare the present model numerical results with Al-Khalil et al.2 experimental data, an overall heat
transfer coefficient Uair was defined taking into account the effects of convection (both hair and hwater), runback water
mass flow, impingement and evaporation.
By applying the First Law of Thermodynamics to the water film flow (domain IV), it is possible to obtain the
following equation:

( ) (
F ⋅ A ⋅ hwater ⋅ (Trec − Twater ) + F ⋅ A ⋅ hair ⋅ (Twall − Twater ) + min ⋅ c p , water ⋅ Tin − Tref + mout ⋅ c p , water ⋅ Tref − Tout
 

)
[ ) ]


( ) (
Vd2 (5)
+ F ⋅ mimp ⋅ c p , water ⋅ Td − Tref + + F ⋅ mevap ⋅ c p , water ⋅ Tref − Tout − ilv = 0

 

2

Tin + Tout
T water = (6)
2
y
Y
T∞, P∞ s
V∞, M ∞
Gaseous Flow
LWC, MVD
(II)
Airfoil solid surface
Free stream flow (I) (V)
X

u∞(s)
p∞(s), Trec(s) -s
Mlocal (s)

Tf,ar(s)
Dynamic and Thermal Boundary
Layers (III)
Vl(s), τ0(s)

Vm(s),
Tagua(s) Water film flow (IV)
Airfoil solid surface Tp(s)
(V)

s
Figure 2: Domains of the mathematical model
The convection heat transfer coefficient hwater, between water film (IV) and solid surface (V), is calculated using
Reynolds Analogy (St = 0.5 ⋅ C f ⋅ Prwater
−2 3
) . The water thermodynamics properties are calculated according to Eckert14
definition for low speed flows at mean film temperature:

T water = Twall + 0.5 ⋅ (T water − T wall ) (7)

By applying the Mass Conservation to the water film flow (domain IV), we have:

min + mimp = mout + mevap

    (8)

According to Spalding21, the water evaporation flux is calculated by:

′′ = g m ⋅ Bm
mevap

(9)

6
American Institute of Aeronautics and Astronautics
 2
ln(1 + Bm )
g m = St⋅ G ⋅ Le 3 ⋅ (10)
Bm

The driving force Bm for mass transfer is calculated by the following expressions21, 22:

m H 2 O , S − m H 2 O ,G
Bm = (11)
mH 2O , S − 1

pvap ,G
m H 2 O ,G = (12)
1.61⋅ p mixt ,G − 0.61⋅ pvap ,G

pvap , S
m H 2O ,S = (13)
1.61⋅ p mixt ,G − 0.61 ⋅ pvap , S

It is assumed that the gaseous flow is saturated mixture at temperature Trec, i.e., the relative humidity is 100% at
gaseous flow around the airfoil. Despite the blowing velocity caused by evaporation not being considered in
boundary layer equations, the effect of evaporation enthalpy in the convection heat transfer coefficient hair is taken
*
into account by the present model. A corrected hair value is calculated from a previous iteration hair value and
*
blowing parameter. The convergence will be considered satisfactory when both hair and bh satisfies the convergence
criteria. For the mass flux levels seen in the present reference case, an acceptable approximation is:

*
hair b
= b h (14)
hair e − 1 h

From the water droplet local collection efficiency definition, the impinging water flow is given by:

mimp = V∞ ⋅ β (s ) ⋅ LWC ⋅ ∆s ⋅ 1 (15)

At stagnation point, it is assumed that no runback water enters the finite volume and the Eq. (8) is solved from
the stagnation point to downstream direction for both lower and upper airfoil surfaces.
At the volume finite where the Twater reaches 0°C, it was assumed that there is no outlet runback water flow
m out = 0 , which denotes the beginning of the water freezing. In e Eq. (5), it was not considered the effect of

solidification enthalpy. If the warming effect due to enthalpy releasing had been considered, the model would have
indicated the water freezing beginning in a further position than predicted by this model. Al-Khalil6 (basic model)
used similar assumptions for freezing process. Finally, the present model is applied to anti-ice system, which is
designed to keep the solid surface free of ice during its operation.
The Momentum Conservation equation for the water film in the present case is:

1 ∂p∞ ∂ 2 v(s, y )
⋅ = υ water ⋅ (16)
ρ water ∂s ∂y 2

The boundary conditions at the water film flow (domain IV) are assumed to be:

dv(s, y )
µ water = τ + m imp
"
⋅ud (17)


dy y =δ f

v(s,0) = 0 (18)

7
American Institute of Aeronautics and Astronautics
 With the assumption of a parabolic velocity profile for the water film flow, the solution of Eq. (16) is:

∂p ∂p


y2
v(s, y ) = ⋅ ′′ ud − δ water (s ) ∞ ⋅
y
+ τ +mimp


(19)


∂s 2 ⋅ µ water ∂s µ water

The water film thickness δwater can be calculated from the mean water film velocity v (s, y ) , which is obtained
with Eq. (21), is given by:


m out
δ water = (20)
1 ⋅ ρ water ⋅ v (s, y )

The evaluation of dynamic and thermal boundary layers is performed following the mathematical models and
procedures described by Silva, Silvares and Zerbini in a previous paper3. The authors implemented a completely
1 0.7

0.8 0.6

0.6 0.5

0.4
0.4
Cp

0.3
beta

0.2
0.2
0
0.1
-0.2
0
-0.4 -0.1
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
0 0.2 0.4 0.6 0.8 1
s/c
s/c
evaluated with ONERA2D
evaluated with ONERA2D Al-Khalil et al.(4) – LEWICE/ANTICE

Figure 3: (a) Pressure Coefficient Cp (b) Local Efficiency Collection β(s), Cases 67A and 67B

new routine in order to calculate the boundary layer in integral form at laminar, transitional and turbulent regimes.
Such routine evaluates the thermal
boundary layer over a non-isothermal
0.35 0.06
surface with effects of coupled mass
0.3 0.05 transfer due to water evaporation where
the liquid water exists. The
impingement [g/(s*m)]

0.25
0.04 thermodynamic solver considers that
runback [g/(s*m)]

0.2 there is a transition region with a defined

0.03 length where the flow goes from laminar
0.15 to fully turbulent regime. A mean
0.1
0.02 position sm and a standard deviation
length σ describe statistically the
0.01
0.05 laminar-turbulent transition region.
0 0
-0.15 -0.1 -0.05 0 0.05 0.1 0.15 VIII. Simulation Results
s/c
The reference cases 67A and 67B
were simulated in pre-processor
Present model – water runback flow
Present model – water impingement
ONERA2D and coefficient Cp and local
Al-Khalil et al. (4) – runback flow predicted by LEWICE/ANTICE
Figure 4: Runback and Impinging water distribution, Case 67A efficiency collection β(s) were obtained
(Figure 3). For case 67A, it was assumed
that the transition region is statistically described by mean position sm/c=0.14 and standard deviation σ c =0.035.
For the case 67B, the mean position was sm/c=0.048 and standard deviation was σ c =0.012. Both values are
normalized by airfoil chord c. Al-Khalil et al.4 informed that the transition was manually tripped at str/c=0.0556 in

8
American Institute of Aeronautics and Astronautics
case 67A and at str/c=0.035 in case 67B, however, they have not published details about the method used to
determine the abrupt transition location and extension. The abrupt transition point assumed by Al-Khalil et al.2 is
within the transition region defined in the
present work.
Figures 4 and 5 compare the runback 0.6 0.06

water distribution with ANTICE numerical

results for cases 67A and 67B. The water 0.5 0.05
impingement flux is presented in Figures 4
and 5 for reference. The solid surface 0.4 0.04

impingement [g/(s*m)]
temperature distribution calculated with

runback [g/(s*m)]
the present model is shown in Figure 6 and
0.3 0.03
7 for cases 67A and 67B. They are
4
compared with Al-Khalil et al. numerical
and experimental results. 0.2 0.02

In case 67A, as indicated in Figure 6,

the water film stop flowing due to freezing 0.1 0.01
at position s/c=0.1269 at upper airfoil
surface and s/c=-0.1108 at lower airfoil 0 0
surface. The estimated freezing rate was -0.15 -0.1 -0.05 0 0.05 0.1 0.15
0.17E-3 g/(s.m) and 0.071E-3 g/(m.s) for s/c
upper and lower surface, respectively.
This rate values are negligible when present model - runback water flow
present model - water impingement
compared to total water impingement mass
Al-Khalil et al. (4) - LEWICE/ANTICE
flux mimp =1,0 g/(s.m). On the other
Figure 5: Impinging water distribution, Case 67B
hand, in case 67B the water film froze at
s/c=0.1108 and s/c=-0.0987 and the initial
freezing rate was 0.367 g/(m.s), which is significant higher than case 67A values.
Figure 7 represents the temperature distribution of the solid surface obtained by the present model for case 67A
and the experimental data4 and those numerical results of CANICE A8.

IX. Conclusions
A mathematical model was implemented and the numerical results were considered satisfactory, since all
proposed objectives were reached for
engineering purposes. From flow field
50 and water droplets collection efficiency
40 results provided by a pre-processor, the
present numerical code is capable to
30
predict the main parameters of an
20 airfoil electro-thermal anti-ice system
Tp [C]

10 (solid surface temperature, runback

water and convection heat transfer
0 coefficient distributions) at reference
-10 case conditions with acceptable
deviations from experimental results.
-20
The solid surface temperature
-30 distribution obtained with the present
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
model was clearly within experimental
s/c
error margin for most wet region. The
Present model highest deviations from experimental
experimental Al-Khalil et al.(4)
Al-Khalil et al. (4) - LEWICE/ANTICE
results were found near freezing
Present model – end of liquid water film temperature (0° C) and at the end of
Figure 6: Solid Surface Temperature Distribution, Case 67A the liquid water film position, which
Comparison with ANTICE. were located within the predicted
laminar-turbulent transition region.
The present code and its new thermodynamic solver had smaller deviation from experimental data than ANTICE and

9
American Institute of Aeronautics and Astronautics
 much more smaller deviation when compared with the simulator CANICE. The runback water numerical results
have an acceptable and reasonable trend but are lower than mass flow predicted by ANTICE. In addition, the local
collection efficiency β(s) provided by the pre-processor ONERA2D presented deviations when compared to
ANTICE, mainly at impingement limit position (Figure 3b).
In Figure 8, it is clear that at temperatures near 0° C, the new thermodynamic solver results obtained larger
deviation in relation to experimental data and ANTICE numerical results. It may be due to the fact that the Eq. (5)
does not take into account the
solidification enthalpy release effect.
However, the 67B case is very severe 60
and may not be realistic for a well- 50
designed anti-ice system, which 40
usually maintains temperatures within
30
protected area above the case 67B

Tp [C]
mean temperature level 20
(~5 °C). 10
As expected, the present 0
thermodynamic solver predicted a the -10
water film freezing in a position more
-20
upstream than ANTICE code due to
differences in freezing process and -30
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
model assumptions.
s/c
The present numerical code has
shown that the momentum and Present model
experimental Al-Khalil et al.(4)
thermal boundary layer equations in Morency et al(8) - CANICE A
integral form can be appropriate for Present model – end of liquid water film
anti-ice simulation in running wet Figure 7: Solid Surface Temperature Distribution, Case 67A
regime. In general, the mathematical Comparison with CANICE A
model predictions for running wet
condition were closer to experimental data than for evaporative condition presented in previous paper3.
Some factors allowed a better application of boundary layer equations in integral form than previous anti-ice
simulation works4, 8: 1) a better evaporation model by the use of convective driving force mass transfer model; 2)
coupling between convection heat
10 transfer coefficient and solid surface
temperature due to evaporation enthalpy
5
effects in thermal boundary layer; 3)
0 streamwise surface temperature variation
-5
were taken into account; 4) transition
from laminar to turbulent flow modeled
Tp -10 as occurring in a region instead of
[C]
-15 abruptly in a point; 5) the enthalpy and
momentum thicknesses are continuous
-20
functions at transition point.
-25

-30
Acknowledgements
-35
The authors would like to acknowledge
-0.15 -0.1 -0.05 0 0.05 0.1 0.15 Prof. Dr. Marcos de Mattos Pimenta for
s/c
his contributions in laminar-turbulent
transition subject, and the Brazilian jet
present model aircraft manufacturer Embraer, which is
experimental Al-Khalil et al.(4) represented by its Environmental
Al-Khalil et al. (4) - LEWICE/ANTICE
present model - end of liquid water film
Systems Engineering Manager, Eng.
Rodrigo Persico Oliveira, and the
Figure 8: Solid Surface Temperature Distribution, Case 67B
Technology Development Manager, Dr.
Comparison with ANTICE.
Hugo Borelli Resende, for all the support
that allowed the publishing of this work.

10
American Institute of Aeronautics and Astronautics
 References
1
Silva, G. A. L., “Modelagem e simulação da operação de sistema antigelo eletrotérmico de um aerofólio”, M.Sc. Thesis,
Mechanical Engineering Department, Escola Politécnica, Universidade de São Paulo, São Paulo, Brazil, 2002
2
Silva, G. A. L.; Silvares, O. M., “Airfoil anti-ice system thermal simulation”, Universidade de São Paulo, Escola Politécnica
Technical Bulletin, BT / PME / 0207, São Paulo, Brazil, 2002
3
Silva, G. A. L.; Silvares, O. M.; Zerbini, E. J. G. J., “Airfoil anti-ice system modeling and simulation”, AIAA Paper, AIAA
2003-0734, 2003
4
Al-Khalil, K. M. et al., “Validation of NASA thermal ice protection computer codes. Part 3 – Validation of ANTICE”,
AIAA Paper, AIAA 97-0051, 1997
5
Messinger, B. L., “Equilibrium temperature of an unheated icing surface as a function of air speed”, Journal of Aeronautical
Sciences, v. 20, n. 1, 1953.
6
Al-Khalil, K. M., “Numerical simulation of an aircraft anti-icing system incorporating a rivulet model for the runback
water”, Ph.D. Dissertation, College of Engineering, University of Toledo, Toledo, OH, 1991
7
Henry, R., “Development of an electrothermal de-icing/anti-icing model”, Office National d'Etudes et de Recherches
Aérospatiales, Rapport ONERA TAP-92005, Chatillon Cedex, France, 1992
8
Morency, F.; Tezok, F.; Paraschivoiu, I., “Heat and Mass transfer in the case of an anti-icing system modelisation”, AIAA
Paper, AIAA-99-0623, 1999
9
Cansdale, J.T.; Gent, R.W., “Ice Accretion on Aerofoils in Two-dimensional Compressible Flow – A Theoretical Model”,
Royal Aircraft Establishment, RAE TR 82128, Farnborough, UK, 1983”.
10
Gent R.W., “TRAJICE2 – A Combined Water Droplet Trajectory and Ice Accretion Prediction Program For Aerofoils”,
Royal Aerospace Establishment, RAE TR 90054, Farnborough, UK, 1990.
11
Gent, R.W., Moser, R., Cansdale, J.T., Dart, N.P. “The Role of Analysis in the Development of Rotor Ice Protection
System”, SAE Paper, SAE-2003-01-2090
12
Downs, S. J.; James, E. H. “Heat transfer characteristics of an aero-engine intake fitted with a hot air jet impingement anti-
icing system”. Proceedings of 25th National Heat Transfer Conference, V. 1, American Society of Mechanical Engineers,
Houston, 1988, pp.163, 170
13
Riley, S. J., “Investigation relating factors influencing the effectiveness of an aero-engine intake thermal anti-icing system”,
PhD Dissertation, Loughborough University of Technology, Loughborough, UK, 1991
14
Guffond, D.; Brunet, L.,“Validation du programme bidimensionnel de captation”. Office National D’Études et de
Recherches Aérospatiales, ONERA RP 20/5146 SY, Chátillon Cedex, France, 1988
15
Wright, W.B.; Gent, R.W.; Guffond, D. “DRA/NASA/ONERA Collaboration on Icing Research Part II - Prediction of
Airfoil Ice Accretion” NASA CR-202349, Cleveland, OH, 1997“
16
Kind, R.J. “Ice Accretion Simulation Evaluation Test”, NATO Research and Technology Organization, RTO-TR-038,
Cedex, France, 2001
17
Bredif, M., “A fast finite element method for transonic potential flow calculations”. AIAA Paper, AIAA-83-16755, 1983.
18
Bredif, M., “Finite element method calculation of potential flow around wings”, Lecture Notes on Physics, v.218, 1985, pp.
109,114.
19
Makkonnen, L., “Heat Transfer and icing of a rough cylinder”, Cold Regions and Technology, V. 10, 1985, pp. 105,116
20
Eckert, E. R. G., “Engineering Relations for Friction and Heat Transfer to Surfaces in High Velocity Flow”, Journal of the
Aeronautical Sciences, v. 22, 1955, pp. 585,587
21
Spalding, D. B., “Convective mass transfer, an introduction”, New York, McGraw-Hill, 1963.
22
Kays, W. M.; Crawford, M. E., “Convective heat and mass transfer”, New York, McGraw-Hill, 1993

11
American Institute of Aeronautics and Astronautics