In-flight Ice Accretion Prediction Code

Vladimr HORK*, Zdenk CHRA**

*Corresponding author
University of Defence in Brno,
Kounicova 65, 612 00 Brno, Czech Republic
vladimir.horak@unob.cz
**Institute of Hydrodynamics, Academy of Sciences of the Czech Republic,
Pod Patankou 5, 166 12 Praha 6
chara@ih.cas.cz
Abstract: The phenomenon of in-flight icing may affect all types of aircraft. The paper deals with the
development of the computational wing airfoil ice accretion prediction code Ice. Presented code
versions enable computational rime ice and glaze ice accretion prediction on single and multi-element
airfoils in acceptable time of solution. There are presented results of rime ice and glaze ice accretion
predictions and compared with current ice accretion prediction codes. Various cases of predicted ice
shapes are shown in dependence of air temperature. The latest Ice code version enables to solve
system of several airfoils. The example of ice prediction on the wing airfoil with a slotted flap is shown for
two angles of deflection.
Key Words: aircraft icing, ice accretion simulation, icing code.
1. MOTIVATION
The formation of ice on airplane wings occurs when the aircraft flies at a level where
temperature is at, or below freezing point and hits supercooled water droplets. The in-flight
icing may affect all types of aircraft.
Presence of ice on an aircraft surface can lead to a number of performance degradations:
- changes in pressure distribution
- decreased maximum lift and increased drag
- stall occurring at lower angles of attack and increased stall speed
- reduced controllability.
It is important to understand how the different ice shapes affect aircraft aerodynamics. It
can be studied by flight tests, wind tunnel measurements, and computational simulations.
Computational simulation of ice accretion is an essential tool in design, development and
certification of aircraft for flight into icing conditions.
Currently, there exist several approved ice accretion codes:
- LEWICE (LEWis ICE accretion program) is software developed by the Icing Branch
at NASA Glenn Research Center
- CANICE code developed at the Ecole Polytechnique de Montreal
- ONERA (Office National d'Etudes et de Recherches Arospatiales) code in France
- TRAJICE code which was developed by DERA (Defence Evaluation and Research
Agency) in United Kingdom
- CIRA code from Italian Aerospace Research Center.
INCAS BULLETIN, Volume 2, Number 4/ 2010, pp. 119 - 126
DOI: 10.13111/2066-8201.2010.2.4.16
Vladimr HORK, Zdenk CHRA 120

2. ICE ACCRETION PREDICTION CODE
In conjunction with the project of the Czech Ministry of Industry and Trade, was developed
the tool for simulating flight into icing conditions. Presented software was subsequently
developed and improved. Currently, there are three main code versions:
- R-Ice 1.1 Rime ice accretion prediction [1]
- Ice 3.1 Glaze ice accretion prediction [3]
- Ice 4.1 Multi-element airfoils icing [4]
The rime ice is formed if all the impinging water droplets freeze immediately upon
impact. It tends to form at combinations of low ambient temperature, low speed (low kinetic
heating) and a low value of cloud water concentration.
The glaze ice creates at combinations of temperature close to freezing, high speed or
high cloud liquid water content. In that case, not all of the impinging water freezes on
impact, the thin layer of remainder water is flowing along the surface and freeze at other
locations. The process is strongly influenced by the heat transfer.
The latest code version is modified for multi-element airfoils, when the mutual flow
overlap of circumfluent bodies can occur. This modification enables the solution of more
complicated icing simulation cases, e.g. airfoil with slotted flap, wing slot, etc.
3. TRAJECTORIES OF WATER DROPLETS
The potential flow field is calculated using 2-D panel method. The relation for any point
inside the control area is in form
S
r
S
n r
S S
d grad d
1
ln grad grad 2
2 } }
|
.
|

\
|
+
c
c
|
.
|

\
|
=
n r

t
. (1)
The searching solution has to respect the boundary area condition
n
v = n grad , where
v
n
is the normal direction velocity component on the boundary. It is possible to approximate
the velocity potential , or the n c c value respectively, on the body surface and on the
flow field potential discontinuity surfaces by a linear combination of an appropriate function
class. If we choose the needed quantity of control points on the boundary in which we
require to fulfill the integral equation for the gradient of potential and boundary conditions,
the method leads to the set of linear equations solution for unknown coefficients of functions
linear combination that approximate the solution.
Potential flow field is then used to determine the trajectories of water droplets and the
impingement points on the body. Droplets passing through the atmosphere are considered as
spherical elements with a mass m on that the surrounding fluid forces and gravitation F act.
Droplets acceleration and position vector r are given by relations
p
p
t m t
v
v
= =
d
d
,
d
d
r F
. (2)
Typical results of trajectories solution near an airfoil are presented in Fig. 1. It is
perceptible that the small water droplets have trajectories similar to streamlines, vice versa
the large water droplets trajectories are affected by the airfoil inherency only slightly.
INCAS BULLETIN, Volume 2, Number 4/ 2010
121 In-flight Ice Accretion Prediction Code








Fig. 1 Influence of the water
droplets diameter on their trajectories.
Airfoil NACA 0018, chord is 1 m,
free stream velocity is 50 ms
-1
and
angle of attack is 5.





4. R-ICE 1.1 AIRFOIL RIME ICE ACCRETION PREDICTION
The rime ice accretion is the simplest case of ice simulation to predict when impinging
super-cooled water droplets freeze immediately upon impact. Code applies a time-stepping
procedure to calculate the shape of an ice accretion [1]. The new flow field and droplet
impingement recalculations are applied for every procedure step. This procedure is repeated
until the desired icing time is reached.
Results of the airfoil NFL0414 ice accretion prediction for the total icing duration time
1224 seconds in five time steps of solution are presented in Fig. 2. There are presented icing
parameters (the R-Ice code incoming data) in the figure either. Figure also shows the final
ice shape from the in-flight icing experiment [2] at the same conditions by a red color line.



-0,050
-0,025
0,000
0,025
0,050
-0,050 -0,025 0,000 0,025 0,050 x/c [1]
y/c [1]
Chord=0.9144 m; Vext=92.54 m.s
-1
; FluidP=100000 Pa;
FluidT=257.6 K; Alpha=0
o
; PartD=0.00002 m;
PartContent=0.00033 kg.m
-3
, RimeIceRho=900kg.m
-3
;
Steps=5; TimeStep=244.8 s; Time=1224 s.








Fig. 2 Illustration
of the successive
rime ice accretion
for the icing time
1224 sec.


Rime ice accretion prediction provides the comparable results like other current
computational ice-accretion simulation methods [2]. It is evident from the quantitative
comparison plotted in Fig. 3. Icing parameters of solutions are the same as those outlined
above.
INCAS BULLETIN, Volume 2, Number 4/ 2010
Vladimr HORK, Zdenk CHRA 122


-0,050
-0,025
0,000
0,025
0,050
-0,050 -0,025 0,000 0,025 0,050
x/c [1]
y/c [1]
Clean Airfoil
Experimental
Simon
Paraschivoiu
ADSE
Duprat
NASA
R-ICE
ONERA1990
ONERA2000
CANICE
TRAJICE









Fig. 3 Quantitative
comparison of current
computational ice-
accretion simulation
methods for the icing
time 1224 sec.

5. ICE 3.1 GLAZE ICE ACCRETION PREDICTION
Generally, current ice accretion codes give satisfied results of the rime ice simulation, but
glaze icing cases are the most difficult to predict. There is still room for improvement in the
quality of ice-accretion-space predictions [2].
Glaze ice creates at combinations of temperature close to freezing. In that case, not all of
the impinging water freezes on impact. Thin layer of water is flowing very slowly along the
surface and freeze at other locations. The Ice 3.1 code uses so called a shallow water theory
for the solution of the flow of thin water layer on the airfoil surface and gradual freezing.
The conservative equations using for the solution of water flow in open channels are
formally arranged.
Conservative equations written in the general form are

q
x t
S S
F Q
+ =
c
c
+
c
c
. (3)
Vectors of variables Q, flow F and sources S, S
q
are given by relations
|
|
|
.
|

\
|
+ =
|
|
.
|

\
|
=
A EQ
I g A Q
Q
E
Q
A
n 1
2
, | F Q
, (4)

( )
( ) ( )
|
|
|
.
|

\
|
+ +
+ + =
p p e e e e w w w
e e w w t n
cT q o T T o T T o
o o A dx dp A g I g
o o
t t
2
0
S
, (5)
( ) ( )
|
|
|
|
.
|

\
|
+ +


=
ev ev w fr fr w p p e
ev w fr w px p e
ev w fr w p e
q
L q o L q o cT q o
q o q o q o
q o q o q o
cT cT
v v v S
. (6)
INCAS BULLETIN, Volume 2, Number 4/ 2010
123 In-flight Ice Accretion Prediction Code
Where g represents acceleration due to gravity with components g
n
and g
t
. The quantity
is the liquid density, the liquid temperature is denoted by T, channel wall temperature is T
w

and the ambient temperature above water level is T
e
.
Integrals I
1
and I
2
are given by the shape of the channel cross-section

( ) ( ) ( )
( )
q
q
q q q q d
d
, d
, d ,
0
2
0
1
} }
= =
h h
x
x b
h I x b h I
. (7)
The coordinate is measured upwards from the lowest point of channel bed level in the
section x = const. and b (x, ) is the channel width. The channel geometry description is
complemented by the wetted perimeter o
w
and the level width o
e
.
Variable quantities: Q
1
= A represents the local flow cross-section, Q
2
= Q = Av is the
flow volume and Q
3
= E = A

c

T expresses the thermal energy of liquid having specific heat
capacity c. Quantities: F
1
is the mass flux, F
2
is the momentum flux and F
3
is the flux of
energy. Vector S includes sources of mass, momentum and energy. Certain liquid volume
inflows from external sources S
q1
(impacting flux) with the area intensity q
p
[m
3
s
-1
m
-2
]. Some
water can freezes, the freezing fraction could be expressed from the heat balance like the
area intensity q
fr
. Similarly, the area intensity q
ev
represents the quantity of evaporating
water, which is determined by the vapors diffusion from the surface. Source of momentum S
2

includes: hydrostatic pressure (effect of the cross-section change dA/dx), tangential
component of gravity force g
t
(generally external volume forces), friction on the channel bed

w
and on the liquid surface
e
and the momentum component supplying from external
sources with the radial velocity v
px
. Finally, the source of energy S
3
constitutes heat transfer
on the channel wetted perimeter, heat transfer on the liquid level (coefficients
w
and
e
), the
heat supplied from external sources S
q3
by means of liquid and the thermal influence of
freezing and evaporating process, where L
fr
is the latent heat of fusion and L
ev
is the latent
heat of evaporation.
The formulated problem of the thin liquid layer flow solution is solved by a
discontinuous Galerkin method, which could be considered as a generalized finite volumes
classical method. Principles of the Galerkin method, applied to the solution of the flow of a
thin water layer and gradual freezing, are closely described in [3].
Results of the glaze ice accretion prediction for the icing time 300 seconds and airfoil
NACA 0012 are outlined in Fig. 4.
Process of glaze ice accretion is strongly
influenced by the wall temperature T
w
.

Fig. 4 Effect of wall temperature on
glaze ice shapes and comparison with
experiment.
Input data of the solution are airfoil
chord b = 0.45 m, free stream velocity
v

=

77.2 m

s
-1
, angle of attack

=

0,
cloud liquid water content LWC

=

0.32

g

m
-3
, droplets median volume diameter
MVD = 18 m, ambient air temperature
T
e
= 270.5 K and wing surface
temperature T
w
= 273.0 K.
INCAS BULLETIN, Volume 2, Number 4/ 2010
Vladimr HORK, Zdenk CHRA 124

It could be noted that the results outlined above qualitatively correspond to the
experimental observations of glaze ice shapes. Comparison in Fig. 5 acknowledges that the
presented solution could
be considered at least as
a fully comparable with
the current ice accretion
prediction codes.
-0,06
-0,04
-0,02
0,00
0,02
0,04
0,06
-0,02 0,00 0,02 0,04 0,06 0,08 0,10
Clean Airfoil
Experimental
Paraschivoiu
ADSE
Duprat
NASA
y/b
x/b
LEWICE
ONERA
TRAJICE
CANICE



Fig. 5 Quantitative
comparison of current
computational ice accretion
simulation methods from [2].

6. INFLUENCE OF AIR TEMPERATURE ON ICE SHAPES
The glaze ice accretion process is strongly dependent on temperature, besides other icing
parameters like air liquid water content (LWC) and median droplets diameter (MVD).
Influence of air temperature T on iced airfoil shapes predicted by the Ice code, version 3.1, is
shown in Fig. 6, where we can see various cases of glaze ice shapes: stream-wise shape (b),
(c), double-horn shape (d), (e), and span-wise ridge shape (f).







(a) Rime ice (b) Glaze ice: T = 269.65 K (c) Glaze ice: T = 270.15 K








(d) Glaze ice: T = 270.65 K (e) Glaze ice: T = 271.65 K (f) Glaze ice: T = 272.65 K

Fig. 6 Ice code simulation of air temperature influence on iced airfoil shapes for T = T
w
.
Airfoil NFL0414, chord 0.45 m, angle of attack

= 0
o
, free stream velocity v

= 77.2 m s
-1
,
MVD = 18 m, LWC = 0.32 g m
-3
, atmospheric pressure 100 kPa, icing time 900 seconds.
INCAS BULLETIN, Volume 2, Number 4/ 2010
125 In-flight Ice Accretion Prediction Code
7. ICE 4.1 MULTI-ELEMENT AIRFOILS ICING
The latest Ice code version [4] enables to solve system of several airfoils, by default, up to
eight separate parts. Model algorithms have been extended to involve mutual flow overlap
of multi-element airfoils (e.g. overlap between the airfoil and flap). The typical results of
air streamlines of droplet trajectories around an airfoil with a slotted flap are presented in
Fig. 7. There are seen droplet trajectories and impact locations near the airfoil leading
edge. Trajectories of droplets impacting an airfoil surface are depicted by a black square.
The impact locations where droplet trajectories intersect an airfoil surface may be
divided into several separated subsections. It can be seen for the case of airfoil with the
slotted flap in landing position. The flap is not fully overlapped in this case. Then black
squares of impinging droplets trajectories are divided on one impacting the airfoil and
another one impacting the flap surface.


Fig. 7 Droplet trajectories near an airfoil with a slotted flap and for flap in landing position

The ability of the Ice code 4.1 version to predict ice accretion of flapped airfoils is
presented on the case of the wing airfoil with a slotted flap. Example of the ice prediction
on the wing airfoil with the slotted flap for angles of flap deflection 20 and 38 is shown
in Fig. 8



Fig. 8 Example of ice prediction on the wing airfoil with a slotted flap for angles of deflection 20 and 38
Mentioned ice accretion on the flap causes the reduction of the gap size between main
element and flap. Consequently, it can have a large impact on the performance degradation
of iced multi-element airfoils. Lastly, there is a potential mechanical problem in the elevator
mechanism itself.

INCAS BULLETIN, Volume 2, Number 4/ 2010
Vladimr HORK, Zdenk CHRA 126

INCAS BULLETIN, Volume 2, Number 4/ 2010
8. CLOSING REMARKS
The Ice code enables computational rime ice and glaze ice accretion prediction on single and
multi-element airfoils in acceptable time of solution. Mathematical model has been modified
for variable wall temperature along the airfoil surface. The code was also subsequently
improved for the better approximation of transition boundary layer location.
The code is designed for the icing simulation as an aid to the certification process of
small transport aircraft for flight in icing conditions according to international aircraft
standards, where maximum and intermittent maximum icing conditions are specified.
Presented code could be considered at least as a fully comparable with the current ice
accretion prediction codes.
ACKNOWLEDGEMENT
The work presented in this paper has been supported by the Czech Science Foundation
project No. P101/10/0257, by the Czech Ministry of Industry and Trade project FT-TA/044
InICE, and by the Ministry of Defense project No. FVT 0000404.
REFERENCES
[1] V. Hork and B. Hoen. Wing Airfoil Rime Ice Accretion Prediction. In Engineering Mechanics 2006.
Svratka, May 15-18 2006.
[2] R. J. Kind. Ice Accretion Simulation Evaluation Test. RTO Technical Report 38, November 2001.
[3] B. Hoen and V. Hork. Wing Airfoil Glaze Ice Accretion Prediction: Thin Freezing Water Layer. In Sixth
International conference on Mathematical Problems in Engineering and Aerospace Sciences ICNPAA
2006. Cambridge Scientific Publishers, 2006.
[4] B. Hoen, V. Hork, and Z. Chra. Improved Ice Accretion Prediction Code. Advances in Military
Technology, Vol. 3, No. 1, p. 44-50, September 2008.