Special Issue Article

Advances in Mechanical Engineering

2016, Vol. 8(5) 1–15
Ó The Author(s) 2016
Numerical simulation and DOI: 10.1177/1687814016646976
aime.sagepub.com
experimental validation of aircraft
ground deicing model

Bin Chen1,2, Li-wen Wang1,2, Rui Gong3 and Shou-kun Wang3

Abstract
Aircraft ground deicing plays an important role of guaranteeing the aircraft safety. In practice, most airports generally
use as many deicing fluids as possible to remove the ice, which causes the waste of the deicing fluids and the pollution of
the environment. Therefore, the model of aircraft ground deicing should be built to establish the foundation for the sub-
sequent research, such as the optimization of the deicing fluid consumption. In this article, the heat balance of the deicing
process is depicted, and the dynamic model of the deicing process is provided based on the analysis of the deicing
mechanism. In the dynamic model, the surface temperature of the deicing fluids and the ice thickness are regarded as
the state parameters, while the fluid flow rate, the initial temperature, and the injection time of the deicing fluids are
treated as control parameters. Ignoring the heat exchange between the deicing fluids and the environment, the simplified
model is obtained. The rationality of the simplified model is verified by the numerical simulation and the impacts of the
flow rate, the initial temperature and the injection time on the deicing process are investigated. To verify the model, the
semi-physical experiment system is established, consisting of the low-constant temperature test chamber, the ice simula-
tion system, the deicing fluid heating and spraying system, the simulated wing, the test sensors, and the computer mea-
sure and control system. The actual test data verify the validity of the dynamic model and the accuracy of the simulation
analysis.

Keywords
Aircraft ground deicing, deicing fluids, dynamic modeling, simulation, experiment validation

Date received: 2 December 2015; accepted: 6 April 2016

Academic Editor: Zheng Chen

Introduction (Figure 1).1,2 Besides, aircraft ground ice is the direct

cause of the flight delay, cancelation, and the airport
Aircraft ground ice refers to the verglas, ice, snow, and closure in winter, which can cause enormous economic
frost accumulated on the surface when aircrafts are on
the ground, which might change the aerodynamic con- 1
figuration of the aircraft, reduce the lift force, and School of Mechanical Engineering, Tianjin University, Tianjin, P.R. China
2
Ground Support Equipments Research Base, Civil Aviation University of
damage the airframe components. Therefore, it brings China, Tianjin, P.R. China
the hidden danger to air safety. Moreover, the winter 3
Automation Department, Beijing Institute of Technology, Beijing, P.R.
flight safety is greatly influenced by aircraft ground ice. China
According to statistics, there have been approximately
Corresponding author:
20 accidents caused by aircraft ground ice in the past
Bin Chen, School of Mechanical Engineering, Tianjin University, Tianjin
30 years. Some typical accidents include Baotou air 300072, People’s Republic of China.
crash in 2004 and Colorado air crash in 1987 Email: chenbindavid@163.com

Creative Commons CC-BY: This article is distributed under the terms of the Creative Commons Attribution 3.0 License
(http://www.creativecommons.org/licenses/by/3.0/) 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
Figure 1. Typical accidents caused by aircraft ground ice: (a)
Colorado air crash site and (b) Baotou air crash site.
losses as well as have a strong impact on the public
transportation.
Aircraft ground deicing or anti-icing has been an
important problem drawing great attention of interna-
tional civil aviation departments. Federal Aviation
Administration (FAA), European Aviation Safety
Agency (EASA), and Civil Aviation Administration of
China (CAAC) have stringent regulations on the opera-
tion of the aircraft whose surface is attached by the
ice.3,4 How to deice safely and efficiently is one of the
most crucial issues in the civil aviation field.
Meanwhile, interrelated techniques and theories have
also received a great deal of attentions.
There are many kinds of advanced and potential air-
craft deicing methods emerging, the typical of which
include infrared aircraft deicing,5 microwave deicing,6
and hot air deicing.7,8 But when it comes to aircraft
ground deicing or anti-icing, the aircraft deicing fluids
are mainly adopted: the specific deicing facilities or sys-
tems are utilized to spray the deicing fluids to the sur-
face of aircraft at the appointed time. Then, the deicing
fluids guarantee that the ice on the surface melts and
prohibit the aircraft from icing up within a certain
period of time, ensuring the safety of taking off.
Concerning the aircraft deicing fluids, the investiga-
tion mainly involves deicing mechanism, deicing model,
deicing efficiency, deicing parameter optimization, dei-
cing multiparameter control characteristic method, dei-
cing operation optimal dispatch, and so on. Previous
studies on deicing model and deicing parameter optimi-
zation are explained as follows.
SE Ornitz9 studied the dependence relation between
the deicing fluid concentration, the ambient tempera-
ture, the wind velocity, and the deicing fluid holding
period. The optimization model that the deicing fluid
concentration varied with the environment was also
built. The mixed integer nonlinear programming
(MINP) was adopted to carry out online intelligent
optimization of the amount that the deicing fluids were
used, overcoming the uncertain problems caused by the
aircraft model, temperature, and facilities during the
process of deicing.
The research workers of APS Aviation, Inc.10 made
an experimental study on the routine and the technique
of using deicing or anti-icing fluids. The deicing para-
meters were also optimized to reduce the environmental
pollution caused by the deicing fluids. Through the
experimental research at the airport, the deicing fluid
holding periods at different fluid concentrations were
verified. In the case that the deicing fluids were used to
clear the frost with centralized deicing technologies, the
impacts of the parameters such as deicing fluid tem-
perature and concentration on the amount of deicing
fluids used were also studied through the test.
A Zouzou et al.11 considered the change in deicing
fluid holding periods as the variation of the tempera-
ture. The model that deicing fluid holding periods var-
ied with the temperature was built. The numerical
simulation method was adopted to describe the relation-
ship between the deicing fluid holding period and the
temperature, the result of which was compared with the
deicing fluid holding period at constant temperature.
Besides, the model took into account the effect of some
parameters such as the flow rate of deicing fluids, the
deicing fluid concentration, the shear rate, the ice thick-
ness, and the surface angle of the attachment to the ice.
However, these works mainly focused on the proper-
ties of the deicing fluids, such as the concentration and
the holding period. Unfortunately, to the best of our
knowledge, there is no adequate research on the deicing
process.12 The study results provide basic operating
rules and technical standards for the deicing fluid pro-
ducers. However, from the perspective of consumers,
there has been little investigation on the cost, the equip-
ment parameters, and the deicing efficiency. A small
Chen et al.
number of previous studies determine the parameters
based on the experiments and the project experience,
which are complicated for the operation and do not
consider the cost and the deicing efficiency. From the
green and environment-friendly point, the waste and
the pollution of the deicing fluids should be taken into
consideration. Therefore, this article establishes the
model to study the impacts of the equipment para-
meters on the deicing process, which provides the basis
for the parameter optimization.
In contrast to most of the available results, this article
investigates the mechanism and the physical model of
the deicing process. The main contributions of this arti-
cle lie in the following: the dynamic model of the deicing
process is depicted as a system of differential equations,
and the model is further simplified under the reasonable
assumption that the heat exchange between the deicing
fluids and the ambient environment is ignored. The fluid
flow rate, the temperature, and the injection time are
treated as control parameters, while the surface tem-
perature of the deicing fluids and the thickness of the ice
are state parameters in the dynamic model. In addition,
the simplified model and the original model are all veri-
fied through the comparison of the numerical simula-
tion13–15 data and the semi-physical experiment data.
Moreover, the equilibrium temperature is proposed for
the first time, which suggests that there exists a constant
temperature stage during the deicing process. Therefore,
the effects of the parameters such as the fluid flow rate,
the temperature, and the injection time on the deicing
process are investigated through the simulation and the
test. Both the flow rate and the initial temperature can
change the rate and the volume of the ice melting by
changing the equilibrium temperature, while the injec-
tion time can affect the volume of the ice melting by
changing the duration of the equilibrium temperature
but has no effect on the ice melting rate.
This article is organized as follows: we introduce the
deicing mechanism of the deicing fluids in section
‘‘Mechanism of aircraft ground deicing.’’ In section
‘‘Dynamic model of the deicing process,’’ the dynamic
model of the deicing process is investigated, and the
simplified model is derived. Section ‘‘Numerical simula-
tion’’ is devoted to verifying the model and analyzing
the effect of the parameters such as the fluid flow rate,
the temperature, and the injection time on the deicing
process through the numerical simulation. In section
‘‘Experimental studies,’’ the semi-physical experiment
system is built, and the experiments are carried out
under different parameters to verify the dynamic model
and the correctness of the numerical simulation.
Mechanism of aircraft ground deicing
In the aircraft ground deicing process, the deicing
fluids, heated to a specific temperature, are sprayed on
3
the ice-forming parts of the aircraft surface when some
certain conditions are satisfied such as the requirements
of the flow rate and the nozzle moving speed.16 Deicing
fluids spread out quickly on the aircraft surface, form-
ing a slow-moving and high-temperature liquid film.
The ice, snow, and frost melt depending on the heat
carried by the liquid film. Due to the properties of dei-
cing fluids such as the high viscosity and the low freez-
ing point, the freezing point of the ice liquid mixture is
decreased in a certain period of time. Meanwhile, the
melting ice, snow, frost, and the redundant deicing
fluids flow to the ground. In terms of aircraft ground
deicing, airfoil ground deicing is the most common; so,
this article regards airfoil ground deicing as the main
research object.
The airfoil ground deicing process17 is depicted in
Figure 2, and the abstract model is shown in Figure
2(c). As shown in Figure 2(c), the rectangle represents
the equivalent airfoil, L is the width of the airfoil, W
expresses the length of the airfoil, and V is the nozzle
moving speed of the deicing facilities. The deicing time
which is equal to the moving time of the nozzle along
the direction of the length of the airfoil is expressed as
Dt, Dt = W =V .
The airfoil ground deicing process is a typical
liquid–solid transition process. It is also a quintessential
process of heat transfer and conservation. The deicing
fluids, ice, and aircraft wing constitute a thermody-
namic system. The internal energy of the system
changes only through the heat transfer. The increment
or decrement of the system internal energy is equal to
the amount of the heat absorbed from or released to
the outside.
If the impacts of the wind speed, the vaporization
heat, and the exchange of other energy between the sys-
tem and the environment are all ignored,18,19 the heat
transfer in the aircraft ground deicing process is
depicted in Figure 3.20,21 As shown in Figure 3, Q0 rep-
resents the energy of the deicing fluids on exiting the
spray gun, Q1 is the convection between the liquid film
and the environment, Q2 is the energy lost as thermal
radiation to the environment, Q3 expresses the energy
required to heat the ice from ambient temperature to
the melting point, Q4 is the energy required to melt the
ice on the wing surface, Q5 represents the heat trans-
ferred between the top and the bottom of the liquid
film, and Q6 is the energy required to increase the tem-
perature of the wing material.
In order to make it easy to model and analyze, the
whole process of deicing is divided into two stages. The
first stage is the ice melting stage, where the heat of the
deicing fluids makes the ice melt and flow off the wing,
and this stage is a stage of heat transfer and a process
of phase transition. In the first stage, the thickness of
the ice tends to thin until the thickness is zero. The sub-
stance on the surface of the wing is the mixture of ice,
4 Advances in Mechanical Engineering

Figure 3. Heat transfer in the aircraft ground deicing process.

(a) concentration and the holding period. This article

mainly discusses the ice melting stage and makes mod-
eling study for it.

Dynamic model of the deicing process

Heat balance analysis
The input energy of the thermodynamic system is the
heat released from the deicing fluids with the drop in
temperature. The input energy depends on the flow
rate, the temperature, and the injection time of the dei-
cing fluids,22,23 which is expressed as Q0
   
(b) Q0 = mf cf Tf  Ts = rf Vf cf Tf  Ts ð1Þ

V where rf is the consistency, cf is the specific heat, Tf is

the initial temperature, Vf is the volume, and Ts is the
surface temperature of the deicing fluids. It is assumed
that the flow rate of the deicing fluids is a constant in
the spraying process. Therefore, Vf is a step function
varying with the time, which can be formulated as equa-
L
tion (2)

 ðt 
F0 t\t0 F0 t t\t0
W Ff ð t Þ = , Vf ðtÞ = Ff dt =
0 t  t0 F0 t0 t  t0
(c) 0

ð2Þ
Figure 2. Aircraft ground deicing process: (a) panorama of
aircraft ground deicing, (b) scene of airfoil ground deicing, and where Ff is the function of the deicing fluid flow rate,
(c) abstract model. F0 is the constant flow rate of the spray gun, and t0 is
the injection time.
The consumed energy of the thermodynamic system
deicing fluids, and water. The second stage is the ice mainly consists of the convection between the liquid
inhibiting stage, where the temperature of the wing film and the environment, the thermal radiation to the
material increases and the anti-icing heat is transferred. environment, and the ice melting heat. The convection
Compared with the first stage, the second stage does between the liquid film and the environment can be
not involve the process of ice melting and phase transi- expressed as Q1
tion. Also, the first stage mainly depends on the flow
rate, temperature, and injection time of the deicing ðt
fluids, while the second stage mainly depends on the Q1 = Ah4 ðTs  Ta Þdt ð3Þ
properties of the deicing fluids, such as the 0
Chen et al.
where h4 represents the coefficient of energy convec-
tion.24 The theoretical calculation of h4 is so complex
that it is generally obtained through engineering mea-
surement, and h4 can be expressed as a function of the
surface temperature of the deicing fluids Ts . A is the
area of the airfoil, and Ta is the ambient temperature.
For the temperature difference between the deicing
fluids and the surroundings, deicing fluids release some
heat to the environment in the form of radiation. The
radiation capacity positively correlated with the tem-
perature of the deicing fluids. The radiation energy is
expressed as Q2
ðt
 
Q2 = Aes Ts4  Ta4 dt ð4Þ
0
where e is the radiation rate of the deicing fluid film
and s is the Boltzmann constant, which is equal to
5:67 3 108 W m2 K4 .25,26
Ice melting on the surface of the wing is a complex
process, which can be divided into two stages. The first
stage is the temperature increasing stage where the tem-
perature of the ice increases from ambient temperature
to the melting point. After the temperature of the ice
increases to the melting point, the process enters the
second stage. The second stage is the ice melting stage.
To simplify the study, the assumption is made that the
temperature of the ice on the surface of the wing is
equal to that of the environment. On this basis, the
energy consumed in the temperature increasing stage is
Q3
ðt
Q3 = AHr2 c2 ð273  Ta Þdt ð5Þ
0
where H is the thickness of the melting ice, c2 is the spe-
cific heat of the ice, and r2 is the density of the ice.
In the ice melting stage, the heat required to melt the
ice at 0°C to the water at 0°C is Q4
ðt
Q4 = AHr2 h5 dt ð6Þ
0
where h5 is the latent heat of the ice melting.27
The liquid film is attached to the surface of the wing,
and the bottom of the film is directly in contact with
the wing. In the ice melting process, the temperature of
the bottom of the liquid film is the temperature of the
ice. Therefore, there is a temperature difference
between the top and the bottom of the liquid film. The
heat transferred between the top and the bottom of
the liquid film can be obtained on the basis of the
definition of the heat conduction, which is expressed as
Q5
5
ðt
lf AðTs  273Þ
Q5 = dt ð7Þ
d
0
where lf is the energy conduction coefficient of the dei-
cing fluids and d is the thickness of the deicing fluids
per unit area, which can be regarded as a constant in a
very short time.
In the ice melting stage, the assumptions are as fol-
lows. The kinetic energy of the deicing fluids and the
mass loss due to evaporation of the deicing fluids are
ignored, and the effect of the collision of the deicing
fluids with the ice is also ignored. Besides, the deicing
fluids on the surface of the wing are hypothetical in the
steady laminar conditions. According to the conserva-
tion of energy, the total energy of the thermodynamic
system can be formulated as equation (8)
Q0 = Q1 + Q2 + Q3 + Q4 ð8Þ
Equation (8) shows that the partial energy of the
deicing fluids is consumed in the ambient environment,
while other energy is consumed as the heat required to
increase the temperature of the ice and melt the ice.
Based on the heat transfer in the ice melting process,
the ice melting heat derives from the heat transferred
between the top and the bottom of the liquid film,
which can be formulated as equation (9)
Q5 = Q3 + Q4 ð9Þ
Substituting equation (9) into equation (8), equation
(10) is obtained
Q0 = Q1 + Q2 + Q5 ð10Þ
Dynamic modeling
Many parameters are involved in the process of aircraft
ground deicing, such as the wind speed, the environ-
ment temperature, the air humidity, and the amount of
the ice on the wing. These parameters are nonlinear
and changeable.28 To study the dynamic changes of the
ice in the deicing process, the fluid flow rate Ff , the ini-
tial temperature Tf , and the injection time t0 are treated
as input (control) parameters, while the surface tem-
perature of the deicing fluids Ts and the thickness of
the ice H are output (state) parameters of the system.
In this case, the deicing process is simplified as a con-
trol object with three input parameters and two output
parameters, which establishes the foundation for the
control29,30 and the optimization of the deicing process.
The model of the aircraft ground deicing process is
shown in Figure 4.
After substituting equations (3)–(7) into equation (9)
and differentiating the obtained equation with respect
to the time, the dynamic relationship between the ice
6 Advances in Mechanical Engineering

 
rf F0 cf Tf  Tb = Ah4 ðTb  Ta Þ
  lf AðTb  273Þ ð14Þ
+ Aes Tb4  Ta4 +
d
Equation (14) shows that the equilibrium temperature
Tb depends on the area of the airfoil A, the initial tem-
Figure 4. Dynamic deicing model.
perature of the deicing fluids Tf , the ambient tempera-
ture Ta , and the constant flow rate of the spray F0 . The
third stage is the slow cooling stage. In this stage, the
thickness H and the surface temperature of the deicing
spray gun stops spraying the deicing fluids, resulting in
fluids Ts can be formulated as equation (11)
the shutoff of the heat input and the break of the
 dynamic balance of the heat transfer. The surface tem-
0 lf ðTs  273Þ d
H = ð11Þ perature of the deicing fluids Ts begins to drop from
r2 c2 ð273  Ta Þ + r2 h5 the equilibrium temperature Tb until it reaches the
freezing point, and the cooling rate tends to decrease
where the minus means that the ice thickness tends to
with the drop in the Ts . The above three stages of the
decrease if Ts is higher than the freezing point.
variation of Ts can be formulated as equation (15)
Equation (11) shows that the melting rate of the ice has
a positive linear relation with the surface temperature 8
> h4 ðTf Ta Þ + esðTf4 Ta4 Þ + lf ðTf 273Þ=d
of the deicing fluids. Therefore, in order to increase the >
< A rf F0 Dtcf t  Dt
speed of deicing, the temperature should be maintained 0
Ts = Tb Dt  t  t0
>
>
: A h4 ðTs Ta Þ + esðTs Ta Þ + lf ðTs 273Þ=d
4 4
as a relatively high level. t  t0
r F0 t 0 c f
After substituting equations (2)–(4) into equation f

(10) and differentiating the obtained equation with ð15Þ

respect to the time, the dynamical change in the surface
temperature of the deicing fluids Ts can be formulated
as equation (12) Model simplifying
If the radiation energy from the surface of the wing to
Ts0 =
     the environment and the convection between the liquid
rf Ff cf Tf  Ts  Ah4 ðTs  Ta Þ  Aes Ts4  Ta4  lf AðTs  273Þ d
film and the environment are ignored, which means
rf Vf cf
Q1 = Q2 = 0, all of the heat of the deicing fluids is con-
ð12Þ sumed to melt the ice. In this case, the simplified model
of the ice melting process can be formulated as equa-
In the derivation process of equation (12), the differ-
tion (16)
ential form of the heat released from the deicing fluids
is formulated as equation (13) 
0 lf ðTs  273Þ d
  H =
Q00 = rf Ff cf Tf  Ts  rf Vf cf Ts0 ð13Þ r2 c2 ð273  Ta Þ + r2 h5
   ð16Þ
0
r f f cf Tf  Ts  lf AðTs  273Þ d
F
Ts =
Equations (11) and (12) are the nonlinear differential rf Vf cf
equations depicting the dynamic changes in the ice
thickness H and the surface temperature of the deicing In the following simulation study, the rationality of
fluids Ts . Therefore, the system of nonlinear equations ignoring Q1 and Q2 will be illustrated further, and the
builds the mathematical model of the ice melting validity of the simplified model will be verified.
process. On the basis of the simplified model, the explicit
In the above model, the initial value of the surface expression of the equilibrium temperature is formulated
temperature of the deicing fluids Ts is Tf . On the basis as equation (17)
of equation (12), the variation of Ts can be divided into
three stages. The first stage is the rapid cooling stage, rf F0 cf Tf + lf A=d kT Tf + 1 r f F 0 cf d
Tb = = , kT =
where Ts drops dramatically from the initial tempera- rf F0 cf + 273lf A=d kT + 273 lf A
ture within a very short period of time Dt until the pro-
ð17Þ
cess reaches the dynamic equilibrium. The second stage
is the equilibrium temperature stage, where the heat where kT is the variation factor of the equilibrium tem-
transfer is in a dynamic balance, and the equilibrium perature Tb . The factor depends on the flow rate F0 and
temperature Tb can be formulated as equation (14) the area of the airfoil A. Equation (17) shows that Tb
Chen et al. 7

Table 1. Simulation parameters. In the case that the flow rate is 1.0 L/min, the initial
temperature of the deicing fluids is 60°C and the injec-
Parameters Symbol Unit Value tion time is 40 s, the dynamic simulation curve of the
Wing length W m 0.6 deicing process is shown in Figure 5. As is shown by
Wing width L m 0.6 the dynamic change curve of the surface temperature of
Deicing fluid density r1 kg=m3 1113 the deicing fluids Ts , Ts drops dramatically from the ini-
Deicing fluid specific c1 kJ=kg K 2.35 tial temperature 60°C to the equilibrium temperature
heat 10.9°C within a short period of time after the beginning
Ice density r2 kg=m3 913
Ice specific heat c2 kJ=kg K 1.93 of the injection, and Ts is at the equilibrium tempera-
Latent heat h5 kJ=kg 33.5 ture constantly in the spraying process. After the spray
Liquid film radiation rate e 0.28 gun stops spraying the deicing fluids, Ts starts to drop
Boltzmann constant s W(m2 K4 ) 5:673108 from the equilibrium temperature, and the cooling rate
Thermal conductivity lf W=m K 0.3 tends to decrease until it reaches the freezing point after
Liquid film thickness d m 0.001
Environment Ta °C 210 70 s. For the dynamic change curve of the ice thickness,
temperature it indicates that the ice thickness H decreases linearly in
Energy convection h4 0:02733 the equilibrium temperature stage. After the spray gun
coefficient Ts  3:8909 stops spraying in 40 s, the falling rate of H tends to
decrease. H stabilizes after 70 s, and the total melting
thickness is 3.88 mm.
In order to verify the validity of the simplified model
has a positive linear relation with Tf , which means Tb through the comparison, the dynamic change curve of
increases with the increase in Tf . the simplified model is also depicted in Figure 5. It is
observed that the curve of the simplified model is
almost coincided with that of the original model.
Numerical simulation Because the radiation energy from the surface of the
wing to the environment and the convection between
Simplified model verifying
the liquid film and the environment are ignored, the
In order to accomplish the simulation analysis of the equilibrium temperature falls by 0.15°C, and the melt-
dynamic model, MATLAB or Simulink is adopted to ing thickness decreases by 0.025 mm. But the error
solve the system of nonlinear differential equations caused by the simplification is negligible for the whole
through the numerical calculation. The simulation deicing process.
parameters are listed in Table 1. The parameters are set The energy consumed in the deicing process mainly
based on practical system and are in conformity with consists of the convection energy Q1 , the thermal radia-
the basic parameters of the semi-physical experiment tion energy Q2 , ice heating energy Q3 , and the ice melt-
system. The coefficient of energy convection h4 in Table ing energy Q4 . The proportion of each part is depicted
1 is formulated as a linear function of Ts , obtained by in Figure 6. As shown in Figure 6, the convection
fitting the curve to the experimental data. energy Q1 and the thermal radiation energy Q2 account

Figure 5. Dynamic simulation curves: (a) surface temperature change curves and (b) melting thickness change curves.
8 Advances in Mechanical Engineering
Figure 6. Proportion of each part of the total energy
consumed.
for just 4% of the total energy. Based on the above
analysis, the energy of Q1 and Q2 can be ignored, and
the assumption of simplifying the dynamic model is
reasonable.
Impacts of parameters on the deicing process
The input parameters are the fluid flow rate Ff , the ini-
tial temperature Tf , and the injection time t0 , which can
be controlled by the deicing facilities. In the following
part, the impacts of the parameter variations on the
dynamic deicing process will be analyzed through
simulation.
The initial temperature of the deicing fluids is 50°C,
60°C, 70°C, and 80°C, respectively. The flow rate is
fixed to 1.0 L/min, and the injection time is 40 s. The
simulation curves of the deicing process under different
initial temperature conditions are depicted in Figure 7.
As shown in Figure 7, the equilibrium temperature
increases with the increasing initial temperature of the
Figure 7. Simulation curves under different initial temperatures: (a) surface temperature change curves and (b) melting thickness
change curves.
deicing fluids. Meanwhile, the change rate of the ice
thickness in the heat balance stage and the total melting
thickness also increase with the increase in the initial
temperature. This indicates that the rate and the vol-
ume of the ice melting can be effectively increased by
increasing the initial temperature of the deicing fluids.
The flow rate of the deicing fluids is 0.8, 1.0, 1.2,
and 1.4 L/min, respectively. The initial temperature of
the deicing fluids is 60°C, and the injection time is 40 s.
The simulation curves of the deicing process under dif-
ferent flow rate conditions are depicted in Figure 8. As
shown in Figure 8, the equilibrium temperature
increases with the improved flow rate of the deicing
fluids. Meanwhile, the change rate of the ice thickness
in the heat balance stage and the total melting thickness
also increase with the increase in the flow rate. This
indicates that increasing the flow rate of the deicing
fluids can obtain the same control effect as increasing
the initial temperature of the deicing fluids.
The injection time is 20, 30, 40, and 50 s, respectively.
The flow rate of the deicing fluids is 1.0 L/min, and the
initial temperature of the deicing fluids is 60°C. The
simulation curves of the deicing process under different
injection time conditions are depicted in Figure 9. As
shown in Figure 9, the equilibrium temperature remains
unchanged with the increasing injection time, and the
change rate of the ice thickness in the heat balance stage
also remains the same. But the duration of the heat bal-
ance stage increases linearly, resulting in the increase in
the total melting thickness. This indicates that increas-
ing the injection time cannot change the equilibrium
temperature. It can only increase the total volume of
the ice melting by changing the duration of the heat bal-
ance stage.
The above simulation analysis shows that the flow
rate, the initial temperature, and the injection time of
Chen et al.
Figure 8. Simulation curves under different flow rates: (a) surface temperature change curves and (b) melting thickness change
curves.
Figure 9. Simulation curves under different injection time conditions: (a) surface temperature change curves and (b) melting
thickness change curves.
the deicing fluids have significant impacts on the dei-
cing process. The flow rate and the initial temperature
almost have the same effect, both of which can change
the rate and the volume of the ice melting by changing
the equilibrium temperature. The injection time can
change the duration of the equilibrium temperature
stage. But it cannot change the value of the equilibrium
temperature while it can only affect the volume of the
ice melting and has no effect on the rate.
Experimental studies
Semi-physical experiment system
To verify the accuracy of the above dynamic model
and the match between the model and the actual
9
process of ice melting, the semi-physical experiment
system is established (Figure 10). The semi-physical
experiment system consists of the low-constant tem-
perature test chamber, the ice simulation system, the
deicing fluid heating and spraying system, the simu-
lated wing, the test sensors, and the computer measure
and control system.
The low-constant temperature test chamber is shown
in Figure 11, which is used to simulate different experi-
mental environments under the different climatic condi-
tions. The simulated temperature ranges from 240°C
to 100°C. The simulated humidity ranges from 30% to
98% relative humidity (RH). The cooling power is
1100 W, and the capacity of the test chamber is
800 mm 3 1000 mm 3 1000 mm. There is a damper in
the side of the test chamber, which can simulate the
10 Advances in Mechanical Engineering

Figure 10. Semi-physical experiment system.

s1 water tank; s2 ventilation door; s
3 fiber-optic sensor; s 4 PT100 temperature sensor; s 5 sensor signal conditioning device; s
6 PC; s
7 and
s
13 pump; s s
8 and 12 valve; s9 water nozzle; s
10 deicing fluid nozzle; s
11 flow meter; s
14 deicing fluid heating cabinet.

Figure 11. Low-constant temperature test chamber. Figure 12. Heating and spraying system.

natural wind and humidity. The ice simulation system

consists of the water tank, the pump, the water nozzle,
and the valve. The liquid water is sprayed on the sur-
face of the simulated wing and then freezes at different
temperatures, forming the ice layers of different
thicknesses.
The deicing fluid heating and spraying system is
shown in Figure 12, consisting of the heating cabinet,
the pump, the nozzle, the electric valve, and the flow
meter. The device can make accurate control of
the temperature, the flow rate, and the injection time of
the deicing fluids. The deicing fluids are heated to the
designed temperature, the highest of which is 84°C, and
the error is \2%. The electric valve is used to adjust
the flow rate and the injection time of the deicing Figure 13. Sensors embedded in the aluminum plate.
fluids. The flow rate is adjusted to range from 0 to
100 L/min, and the flow meter is installed between the
nozzle and the electric valve, which is used to measure
the actual flow rate of the deicing fluids. 3 2 mm. There are two fiber-optic sensors and several
The 7075 aviation aluminum plate is used to simulate thermocouple temperature sensors embedded in the alu-
the wing skin, the size of which is 600 mm 3 600 mm minum plate, which are shown in Figure 13. The
Chen et al. 11

Figure 14. Fiber-optic icing detection process.

simulated wing is placed in the low-constant temperature

test chamber, keeping an angle of 5° to simulate the
angle of the wing when the aircraft is on the ground.
Prior to the experiment, the surface of the aluminum
plate is polished to be even, bright, and smooth, guaran-
teeing that there is no dirt attached to the surface. There
are infrared temperature sensors installed above the alu-
minum plate, which are used to measure the surface tem-
perature of the deicing fluids sprayed on the ice.
The fiber-optic icing sensor is a kind of nonfunction-
type fiber-optic sensor. The basic operation principle is
as follows: The light source emits the light. Then, the
light is transmitted by the fiber optic and reaches the
detection surface. If there is not any ice on the detection
surface, the light will emit into the air, and the fiber
Figure 15. Data display and processing application interface.
receiving side will not detect any emitted light. In con-
trast, if there is ice on the detection surface, the light
will do a series of optical actions such as reflection,
experimental data. The data display and processing
scattering, transmission, and absorption in the ice. The
application interface compiled with Visual Studio C# is
light scatters in the ice and reflects on the interface of
shown in Figure 15. Installed in the low-constant tem-
the ice and the air. Finally, the light will enter into the
perature test chamber, the CCD is used to collect and
receiving fiber. The process is shown in Figure 14. The
transmit video images to the computer in order to
end of the receiving fiber is connected with the photo-
remotely monitor the experimental process.
triode and signal processing circuit, which are utilized
to detect and receive the intensity and the change in the
optical signal to measure the state of the ice. The coax-
Dynamic deicing model verifying
ial fiber-optic sensor and the wedgy fiber-optic sensor
are adopted to measure the ice thickness. The accuracy In order to verify the dynamic deicing model and the
of the measurement is improved through the data correctness of the simulation analysis, the parameters
fusion. of the semi-physical experiment system are basically in
The computer measure and control system consists conformity with the simulation parameters. The experi-
of industrial personal computer (IPC), signal condi- mental data are compared with the simulation data
tioning module, data acquisition card, and charge under the same condition. The deicing fluids are heated
coupled device (CCD). The system is used to set test to 60°C. The flow rate is adjusted to 1.0 L/min by con-
parameters of the deicing fluids and collect signals of trolling the electric valve, and the injection time is set
different sensors. It also achieves the real-time display, as 40 s. The comparison between the simulation curve
calculation, processing, storage, and printing of the and the test curve is shown in Figure 16.
12
Figure 16. Comparison between the simulation curves and the test curves: (a) surface temperature change curves and (b) melting
thickness change curves.
The contrast curves of the surface temperature
between the test and the simulation are depicted in
Figure 16(a). As shown in Figure 16(a), the overall var-
iation tendency of the test curve is almost the same as
that of the simulation curve. The test curve shows that
there is a relatively stable equilibrium temperature
stage during the injection process, and the equilibrium
temperature is slightly lower than the simulation curve,
which can be construed as the neglect of some heat
loss. The test curve obviously lags behind the simula-
tion curve, and there exits some fluctuation in the equi-
librium temperature stage, which is because that the
measurement of the temperature has a certain lag, and
the continuous jet of the high-temperature deicing
fluids can partly interfere with the measurement accu-
racy of the infrared temperature sensor.
The contrast curves of the melting thickness between
the test and the simulation are shown in Figure 16(b).
It is observed that the overall variation tendency and
the stable value of the test curve and the simulation
curve are almost the same. During the deicing fluid
injection process, the test curve is apparently ahead of
the simulation curve. It is because that the central posi-
tion of the injection exactly covers the sensors measur-
ing the ice thickness, accelerating the melting rate of
the ice around the sensors and resulting in the rapid
decline of the measured ice thickness. There exist some
aberrant data points on the test curve in the injection
stage, and the stable value of the ice thickness is slightly
lower than that of the simulation curve, the reason of
which is that the deicing fluids sprayed from the nozzle
have a certain amount of kinetic energy, and the dei-
cing fluids will directly wash some ice away.
Although there is some error between the test data
and the simulation data, these data are still similar, and
Advances in Mechanical Engineering
the overall variation tendency is consistent. Therefore,
the dynamic deicing model is similar to the actual dei-
cing process, and the model can reasonably reflect the
change in the state parameters during the deicing pro-
cess and establish the foundation for the subsequent
research and analysis.
Test analysis of parameter variations
In order to further verify the correctness of the simula-
tion analysis under the condition of the parameters var-
iations, the flow rate, the initial temperature, and the
injection time are changed in the experiment system,
respectively, and the impacts of these parameters varia-
tions on the actual deicing process are analyzed.
The flow rate is set as 1.0 L/min, and the injection
time is fixed to 40 s by controlling the electric valve.
The deicing fluids are heated to 50°C, 60°C, 70°C, and
80°C, respectively. The test curves of the deicing pro-
cess under different initial temperature conditions are
depicted in Figure 17.
The deicing fluids are heated to 60°C, and the injec-
tion time is fixed to 40 s. By controlling the electric
valve, the flow rate is set at 0.8, 1.0, 1.2, and 1.4 L/min,
respectively. The test curves of the deicing process
under different flow rate conditions are depicted in
Figure 18.
The deicing fluids are heated to 60°C, and the flow
rate is set at 1.0 L/min by controlling the electric valve.
The injection time is controlled as 20, 30, 40, and 50 s,
respectively. The test curves of the deicing process
under different injection time conditions are depicted in
Figure 19.
The contrast analysis shows that there exist some
differences between the test data and the simulation
Chen et al. 13

Figure 17. Test curves under different initial temperatures: (a) surface temperature change curves and (b) melting thickness change
curves.

Figure 18. Test curves under different flow rates: (a) surface temperature change curves and (b) melting thickness change curves.

Figure 19. Test curves under different injection time conditions: (a) surface temperature change curves and (b) melting thickness
change curves.
14
data. The data error mainly comes from the following
aspects. First, the deicing fluids sprayed from the nozzle
have a certain amount of kinetic energy, and the deicing
fluids will directly wash some ice away, resulting in the
sudden change in the ice thickness. Second, the central
position of the injection exactly covers the sensors mea-
suring the ice thickness, accelerating the melting rate of
the ice around the sensors. Third, the signal acquisition
of the infrared temperature sensor has a certain lag and
is easy to be influenced by the continuous jet of the
high-temperature deicing fluids. In the following study,
the error between the test and the simulation can be
further reduced by improving the experimental condi-
tions. The improvement of the experimental conditions
involves the change in the spray patterns, the adjust-
ment of the angle and the position of the nozzle, the
collection of the signal in different points, and so on.
These experimental results show that although there
is some error between the test curve and the simulation
curve, the overall variation tendency is still consistent,
and the data error is within the accepted range.
Therefore, the results confirm the rationality and the
accuracy of the dynamic deicing model. The experimen-
tal process verifies the significant impacts of the flow
rate, the initial temperature, and the injection time on
the deicing process. The conclusion of the experiment is
consistent with that of the simulation, which shows
that both the flow rate and the initial temperature can
change the rate and the volume of the ice melting by
changing the equilibrium temperature, while the injec-
tion time can only change the duration of the equili-
brium temperature and only affect the volume of the
ice melting.
Conclusion
Aircraft ground deicing is an important technical means
of removing the aircraft surface ice and ensuring flight
safety. Therefore, a reasonable and valid model of the
deicing process is important to the subsequent research.
However, the previous researches mainly focused on
the properties of the deicing fluids rather than the dei-
cing process. In this article, one crucial idea is that the
dynamic model of the deicing process is established
based on the analysis of the deicing mechanism, and the
influential factors on the deicing process are studied. By
the numerical simulation, the impacts of the flow rate,
the initial temperature, and the injection time on the
deicing process are investigated, and the model suggests
that there exists an equilibrium temperature stage,
where the surface temperature of the deicing fluids
remains unchanged. Combined with the actual process
of aircraft ground deicing and the mathematical model,
the semi-physical experiment system is established. The
experimental results show that the theoretical model
Advances in Mechanical Engineering
presented in this article is effective for the prediction of
the ice thickness and the surface temperature, which
indicates that the dynamic model is valid, and the simu-
lation analysis is reliable. Moreover, the simulation and
experiment show that the changes in the flow rate, the
initial temperature, and the injection time of the deicing
fluids have significant effect on the deicing process.
Both the flow rate and the initial temperature can
change the rate and the volume of the ice melting by
changing the equilibrium temperature, while the injec-
tion time can only change the duration of the equili-
brium temperature and only affect the volume of the ice
melting.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
The author(s) disclosed receipt of the following financial sup-
port for the research, authorship, and/or publication of this
article: This work was financially supported by the National
Key Technology Research and Development Program of
China (grant no. 2012BAG04B02).
References
1. Cole JA and Sand WR. Statistical study of aircraft icing
accidents. 1991, AIAA paper 91-0558, http://
arc.aiaa.org/doi/abs/10.2514/6.1991-558
2. Eugene G, Hill Thomas A, et al. Aerodynamic effects of
aircraft ground deicing/anti-icing fluids. J Aircraft 1993;
30: 24–34.
3. JAR/FAR 25.1419. Ice protection including Appendix C
and ACJ 25.1419, http://www.flightsimaviation.com/
data/FARS/part_25-1419.html
4. Civil Aviation Administration of China. CCAR-121-R4.
Beijing, China, 2010; 3: 182–183.
5. Koenig GG and Ryerson CC. An investigation of infra-
red deicing through experimentation. Cold Reg Sci Tech-
nol 2011; 65: 79–87.
6. Harireche O, Verdin P, Thompson CP, et al. Explicit
finite volume modeling of aircraft anti-icing and de-icing.
J Aircraft 2008; 45: 1924–1936.
7. Saeed F and Paraschivoiu I. The optimization results of
the hot-air based anti-icing system. In: Proceedings of the
41st aerospace sciences meeting and exhibit, Reno, NV,
AIAA, 6–9 January 2003.
8. Song B and Viskantat R. Deicing of solids using radiant
heating. J Thermophys 1990; 4: 311–317.
9. Ornitz SE. A mixed integer nonlinear programming model
to optimize the use of aircraft deicing and anti-icing fluids.
PhD Thesis, Florida Atlantic University, Boca Raton,
FL, 2009.
10. D’Avirro J and Chaput M. Optimizing the use of aircraft
deicing and anti-icing fluids. Montreal, QC, Canada: APS
Aviation, Inc., 2011.
Chen et al.
11. Zouzou A, Louchez PR and Bernardin S. Modeling of
the draining and freezing of aircraft de-icing and anti-
icing fluids. 1998, AIAA paper 98-0576, http://
arc.aiaa.org/doi/abs/10.2514/6.1998-576
12. Morency F, Tezok F and Paraschivoiu I. Heat and mass
transfer in the case of anti-icing system simulation. J Air-
craft 2000; 37: 245–252.
13. Kato S. Solution for aircraft anti-icing system simulation
by a modified perturbation method. J Aircraft 2006; 43:
544–554.
14. Da Silva GAL, De Mattos Silvares O and De Jesus Zer-
bini EJG. Numerical simulation of airfoil thermal anti-ice
operation part 1: mathematical modeling. J Aircraft 2007;
44: 627–633.
15. Louchez PR, Bernardin S and Laforte JL. Physical prop-
erties of aircraft de-icing and anti-icing fluids. In: Pro-
ceedings of the 36th aerospace sciences meeting & exhibit,
Reno, NV, AIAA, 12–15 January 1998.
16. Liu X, Rees SJ and Spitler JD. Modeling snow melting
on heated pavement surfaces. Part I: model development.
Appl Therm Eng 2007; 27: 1115–1124.
17. Fortin G, Laforte J-L and Ilinca A. Heat and mass trans-
fer during ice accretion on aircraft wings with an
improved roughness model. Int J Therm Sci 2006; 45:
595–606.
18. Myers TG. Extension to the messinger model for aircraft
icing. AIAA J 2001; 39: 211–218.
19. Myers TG and Charpin JPF. A mathematical model for
atmospheric ice accretion and water flow on a cold sur-
face. Int J Heat Mass Tran 2004; 47: 5483–5500.
20. Liu Z, Cheng H and Sun J. Numerical and experimental
analysis of solar intrusion for radiometer in geostationary
orbit. Heat Transfer Eng 2010; 31: 1193–1202.
15
21. Al-Khalil K, Ferguson T and Phillips D. A hybrid anti-
icing ice protection system. In: Proceedings of the 35th
aerospace sciences meeting & exhibit, Reno, NV, AIAA,
6–7 January 1997.
22. Huang Y, Gao Q, Ma C, et al. Heat melting characteris-
tic of ice layer in ice-snow melting process on road pave-
ment. J Jilin Univ 2010; 40: 391–396 (in Chinese).
23. Guan S. A simulation study of electric cable heating system
used in de-snowing. Master Thesis, Shanghai Jiao Tong
University, Shanghai, China, 2008 (in Chinese).
24. Fitt AD and Pope MP. De-icing by slot injection. Acta
Mech 2001; 147: 73–86.
25. Thongwik S, Kiatsiriroat T and Nuntaphan A. Heat
transfer model of slurry ice melting on external surface of
helical coil. Int Commun Heat Mass 2008; 35: 1335–1339.
26. Huang Y. Study on ice-snow melting and heat absorption
of solar radiation in road. PhD Thesis, Jilin University,
Jilin, China, 2010 (in Chinese).
27. Roy S, Kumar H and Anderson R. Efficient defrosting of
an inclined flat surface. Int J Heat Mass Tran 2005; 48:
2613–2624.
28. Tao YB, He YL and Qu ZG. Numerical study on perfor-
mance of molten salt phase change thermal energy stor-
age system with enhanced tubes. Sol Energy 2012; 86:
1155–1163.
29. Yao J, Jiao Z and Ma D. Extended-state-observer-based
output feedback nonlinear robust control of hydraulic
systems with backstepping. IEEE T Ind Electron 2014;
61: 6285–6293.
30. Yao J, Deng W and Jiao Z. Adaptive control of hydrau-
lic actuators with LuGre model based friction compensa-
tion. IEEE T Ind Electron 2015; 62: 6469–6477.