Progress in Aerospace Sciences 71 (2014) 85–127

Contents lists available at ScienceDirect

Progress in Aerospace Sciences

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

Effects of rainfall on aircraft aerodynamics

Yihua Cao, Zhenlong Wu n, Zhengyu Xu
School of Aeronautic Science and Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China

art ic l e i nf o a b s t r a c t

Article history: Rainfall has been considered as an important meteorological factor to threat aircraft flight safety.
Received 16 May 2014 Adverse effects of rainfall on aircraft aerodynamics have been a constantly hot subject in meteorological
Accepted 23 July 2014 aviation community for decades. This paper presents a systematic and comprehensive overview of the
Available online 4 September 2014
effects of rainfall on aircraft aerodynamics. The overview includes an introduction of rain-induced
Keywords: aviation accidents, a list of the hazards of rainfall to aircraft, the natural characteristics of rain, the
Rainfall existing rain research techniques, some aerodynamic considerations for rainfall simulation and the
Aviation meteorology current state-of-the-art research achievements in the field of effects of rainfall on aircraft aerodynamics.
Aircraft Raindrop impingement, splashback and flow of the formed water film upon lifting surfaces effectively
Aerodynamics
degrade aircraft aerodynamic performance, leading to severe aviation accidents. The previous lessons
Hazard
learned should be disseminated and accepted by later generations to avoid aviation accidents due to
Water film
flight in heavy rain.
& 2014 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
2. Hazards of rainfall to aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3. Natural characteristics of rain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.1. Raindrop size distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.2. The intensity of rain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.3. The terminal velocity of raindrop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.4. The range of rainfall rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.5. The frequency of rainfall occurrence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4. Rain research techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.1. Analytic estimation investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.2. Experimental investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.2.1. Full-scale flight test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.2.2. Scale model test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.3. Computational fluid dynamics numerical simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.4. Scaling considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5. Aerodynamic simulation considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.1. Spray manifold effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.2. Nozzle design effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.3. Cruise configuration lift performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.4. High-lift configuration lift performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.5. Drag characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.6. Surface tension effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6. Effects of rainfall on aircraft aerodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.1. Rain-induced aerodynamic penalties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.2. Raindrop impinging and surface water flow characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.3. Potential mechanisms of rain effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

n
Corresponding author.
E-mail address: jackilongwu@gmail.com (Z. Wu).

http://dx.doi.org/10.1016/j.paerosci.2014.07.003
0376-0421/& 2014 Elsevier Ltd. All rights reserved.
86
7. Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
1. Introduction
As the function of aircraft is expanding, aircraft has also been
widely applied in various domains such as aerobatic flight,
exploration, rescue, private flight, etc., in addition to the primary
commercial and military functions. So it is required that aircraft fly
normally in all kinds of weather conditions to accomplish the
anticipated assignments. Although some measures like hourly
sequence reports, terminal forecasts, low level windshear alert
system (LLWAS), etc. are taken to avoid hazardous weather as
much as possible [1–3], however, the fact is that since manned
aircraft comes into being, flight accidents due to flying in adverse
weather conditions like low-altitude wind shear, thunderstorm,
atmospheric turbulence, frost, hail, lightning, ice accretion and
rainfall have been reported year after year. Over 50% of commer-
cial airline accidents are weather-related [4] and weather remains
the greatest uncertain factor in flight operations, especially incle-
ment weather which can bring numerous problems in aviation
operations [5–9], thus deeply attracting people's attention to deal
with it [10–16]. Over the aforementioned severe weather condi-
tions, rainfall is of particular interest ever since and rain-induced
aviation accidents are reported all over the world. In 1977, the
America Federal Aviation Administration (FAA) conducted a survey
of 25 aviation accidents and incidents occurring from 1964 to
1976, in which low-altitude wind shear was considered a con-
tributing factor [17]. Of the 25 cases (23 were landing and 2 were
take-off) in the survey, ten cases occurred in a rainy condition and
five were identified as intense or heavy rain encounters. Rudich's
statistical analysis of all accidents of all American airline compa-
nies from 1962 to 1984 suggested that in all the weather factors
affecting flight safety, rainfall took up the proportion of 40%,
exceeding wind shear and atmospheric turbulence and ranking
the first [18]. The Eastern Flight 066 accident at Kennedy Airport
(National Transportation Safety Board, NTSB, 1976) [19], the Flight
693 accident at Atlanta International Airport (1979) [20], the Flight
759 accident at New Orleans International Airport (1982) [21], the
Korean Airlines Boeing 747 and the Vietnamese Tupolev accidents
(1997) [22], the Far Eastern Air Flight 066 accident in Taiwan
(2006) [23] and the MI-171 helicopter accident during the strong
earthquake in Wenchuan County of Southwest China's Sichuan
Province (2008) [24] are more or less due to adverse weather
environment of rain associated with wind shears or thunderstorm.
These results led to the consideration of rainfall as a potential
weather-related aircraft safety hazard, particularly in the take-off
and/or approach phases of flight. Even nowadays, people are still
unable to resolve the threat of heavy rain to flight, the best means
is to stop flying in heavily rainy days which is heard by the author
from some airlines.
Although there have been many reviews about other adverse
weather conditions, ice accretion in particular, only several short
review articles involve the effects of rainfall on aviation in the
1980s and 1990s of the last century. Besides, with the computer
technology advancing fast over the last 20 years, a new research
approach for rain named Computational Fluid Dynamics (CFD)
numerical simulation is emerging and has obtained new phenom-
ena that traditional techniques like wind-tunnel experiment did
not observe. This paper presents an overview of the effects of
rainfall on aircraft aerodynamics. This overview includes the
hazards of rainfall to aircraft flight safety, results of existing
attempts to measure the natural characteristics of rain, a review
Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
of the current state-of-the-art methods for evaluating rainfall
effects, some important scaling considerations for extrapolating
model data as well as the latest research achievements on this
subject. A complete understanding of the effects of rain on aircraft
performance will require additional significant effort in both
experimental and analytical ways before an assessment of the
degree of hazard associated with flight operations in a rain
environment can be made.
2. Hazards of rainfall to aircraft
Detrimental effects of rain on aircraft flight were firstly
reported by Rhode in 1941 [25]. It suggested that rainfall could
impart severe aerodynamic penalties to aircraft, but these penal-
ties did not threat aircraft flight safety. It was not until Luers and
Hanies conducted an assessment on the effects of heavy rain for a
couple of flight accidents related to wind shear in 1981 that people
in aviation community began to realize the seriousness of heavy
rain endangering aircraft flight [26].
When an aircraft flies in a rain condition, particularly during
the stages of takeoff and landing, its flight safety is threatened by
massive detrimental effects of rain. The literatures [19,20,27–36]
comprehensively discussed the main detrimental influences of
rain on aviation, here we present a summary of them, which
mainly include the following several aspects:
● Rain-induced visibility reduction is certainly a well-known
phenomenon [37]. Rain can decrease visibility at least in the
following two ways. In one way through the scattering of light
from a large number of liquid droplets, high rainfall rates can
reduce the visibility to less than 400 m. In the other way
through the water film and splashing of raindrops on the
windshield of an aircraft, high-speed windshield wipers will
possibly deteriorate or even lose their effectiveness when
encountering a heavy rain.
● Rain can affect the accuracy of measurement instruments on an
aircraft. For example, it is important for an aircraft stall warning
system to function well. An α-vane instrument is usually
equipped to measure aircraft angle of attack to warn of impend-
ing stall. Because the direction of incoming raindrops that
impact the α-vane can be up to 81 or 101 higher than that of
the freestream air due to the vertical velocity of rain, the angle of
attack of the vane will be influenced by the raindrops to some
extent (as shown in Fig. 1). The bias will lead to a decrease in
measured angle of attack, which will delay the warning of stall
angle and inhibit the activation of the stall warning stick shaker.
From literature [27] it is learnt that rain may also affect the
accuracy of the pitot tube which measures the airspeed of an
aircraft. It is just a speculation, however, since there has not
appeared to be any strong evidence indicating that airspeed
indicators do not work well in a heavy rain environment. Rain
can also influence the accuracy of radar scatterometer measure-
ments on the aircraft [35,38]. The wind sensitivity of the radar
backscatter is known to rely on the resonance with the capillary
waves of wavelength of 1.7 cm [39]. Raindrops can produce
waves of a similar or larger size and disrupt the patterns of the
wind-driven waves. Correlative calculations of wind vector from
the SeaSat Active Scatterometer System (SASS) and related
sensors suggest that measurements can still work well at high
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 1. Possible influence of rain on an angle of attack sensor [27].
Fig. 2. Rain erosion of aluminum alloy on a pitot [43].
wind speeds when rain is present, while at low wind speeds
they become invalid when rain is present [39]. In addition, rain
erosion of materials of inflight aircraft surface [40–46], effects of
rain on weapon projectile [47] and high altitude platform net-
works [48], rain-induced vibration [49] cannot be neglected as
well. A picture of rain erosion damage sustained by the cap of a
pitot boom made of aluminum alloy on a Gloster Javelin aircraft
during trials in the Singapore area is as shown in Fig. 2.
● Standing water on the runway can be a danger to aircraft take-
off. Large ingestions of water that splashed up from the wheels
and undercarriage of an aircraft can result in engine flameout.
For example, rear mounted engines may ingest excess amounts
of water thrown off the fuselage during maneuvering, which
may be a cause of the Southern Airlines DC9 accident [27].
Standing water can also be hazardous to aircraft at the phase of
landing. Large accumulations of water can cause aircraft hydro-
planing, which negates the aircraft braking action and then
leads to longer stopping distance or even sliding off the run-
way. Many aviation accidents are caused by aircraft hydroplan-
ing, for example, October 8, 2004, in Bangladesh, an airplane
rushed off the runway due to the failure for braking when it
was taking off on the rainwater-immerged runway, as shown in
Fig. 3. In addition, standing water can also affect bird-aircraft
collisions at airport. There was evidence suggesting that stand-
ing water increased collision rates at O'Hare International
Airport [36].
● Rain water can also affect aviation engines [50–57]. It is known
that small amounts of water ingested into a jet engine can
enhance the thrust performance of the engine, but extremely
high rain rates will cause engine flame-out thus absolutely
destroying the thrust of the engine. For various jet engines
widely applied in commercial aircraft, the curve of engine
thrust vs water ingestion rate is as shown in Fig. 4. However,
87
Fig. 3. Airplane-hydroplaning-caused accident in Bangladesh.
Fig. 4. Unknown relationship between water ingestion rate and engine thrust [27].
Fig. 5. Rain-induced torque implemented on aircraft and the change in the center
of pressure [27].
no open literatures have been found in revealing any existing
test data accurately relating engine thrust with the amount of
water being ingested. Despite it, the unknown relationship still
has important reference significance in aircraft engine design.
Another rain-induced aircraft engine power loss is that as
raindrops enter the engine, energy is removed to change the
liquid water into stream, thus more fuel is required to keep the
engine at a constant power setting, bringing an accompanying
problem of weight increase of aircraft.
● Raindrops striking the frontal upper surface of the wings and
the frontal segment of the fuselage will impart a downward
and backward momentum as well as a torque on the aircraft
which makes the nose to pitch down and changes the center of
pressure on the airfoil, as shown in Fig. 5. For the acting point
of the torque is distant from the center of gravity, even a
raindrop with a relatively small kinetic energy can bring a
significant torque to the aircraft.
● Water vapor condensation cloud will occur in extensively in the
low-pressure region above the airfoil and release latent heat of
88 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
water drops (see Fig. 6). In the Pan-Am New Orleans accident
[21] a large condensation cloud was observed behind the path
of the flight shortly after takeoff. The addition of heat is known
to destabilize the boundary layer, but the destabilizing effect on
aircraft aerodynamic performance at different angles of attack
still remains unknown and is worthy of further investigations.
● Raindrops accumulated on the surface of the wings will form a
wavy and uneven water film, which can effectively roughen the
wings and increase the aircraft mass. The roughened airfoil
certainly induces a change in the pressure distribution, which
can adversely affect the control input of the aircraft. Besides,
when leading edge slats are extended in takeoff or landing
configurations, a gap exists between the leading edge slat and
the main airfoil section. If the water film present on the leading
edge slat does not reattach to the main airfoil section, then the
existing water sheet may on the one hand clog the gap causing
premature trailing edge separation [58,59] (see Fig. 7) and
aircraft stall or on the other hand interact with the airflow over
the main airfoil section resulting in a change in the
Fig. 6. Possible destabilization effect of water vapor condensation on airfoil
boundary layer [27].
Fig. 7. Premature trailing edge separation for NACA 64-210 airfoil in the rainfall
condition at angle of attack of 141 [59].
Fig. 8. Water film on the slotted airfoil. Left: water sheet clogs the gap between the
leading edge slat and the main airfoil section; right: water film interact with the
airflow over the main airfoil section [27].
aerodynamics (shown in Fig. 8). The premature stall has been
reported by Bezos through wind-tunnel experiment in which
the NACA 64-210 landing configuration underwent a prema-
ture stall shifted from angle of attack of 161 in clear air to 141 in
the rain condition of liquid water content (LWC) of 29 g/m3 and
46 g/m3 when the free-stream dynamic pressure is set to 30 psf
[60]. Another potential detrimental influence of the water film
is the water film separation occurring over the airfoil some-
what aft of the minimum pressure area, as shown in Fig. 9. The
water film separates as a sheet interacting with the airflow over
the airfoil and deflecting the free-stream air up thus encoura-
ging separation.
● Rain will cause severe aerodynamic penalties to airfoils when
an aircraft encounters rain in flight. The seriousness of aero-
dynamic performance degradation depends on the intensity of
rain and the type of airfoil in interest. For example, for a single
rain condition of a liquid water content of 30 g/m3, the NACA
64-210, NACA 0012 and Wortmann FX 67-K170 airfoils respec-
tively had approximately 5%, 15% and 25% lift degradation at
low angles of attack in the wet condition [61,62]. There has
been many researches in this aspect and we will introduce
them in the later chapters of this overview.
However, the research in the rain domain is still in the
exploration stage, so far the existing research results cannot be
extensively applied to engineering practice. Besides, there are
many other uncertain rain effects on aircraft, even nowadays
commercial aircrafts still try to avoid flying in heavy rain environ-
ment. The challenge to relieve or eliminate the detrimental effects
of rain on aircraft is still worthy of more and further efforts.
3. Natural characteristics of rain
Rainfall is a common physical phenomenon in people's life.
Practical observation showed that the maximum raindrop dia-
meter would not exceed 6 mm and the maximum falling velocity
was approximately 9 mm/s. In order to conduct experimental
investigations or develop analytical models on the effects of rain
on aircraft aerodynamics, the characteristics of naturally occurring
rainfall need to be understood. An understanding of the raindrop
size distribution, the intensity of rain, the terminal velocity of
raindrop, the range of rainfall rate and the frequency of rainfall
Fig. 9. Possible influence of water film separation on the airflow above an airfoil
[27].
Table 1
Values of N0 and I for different types of rainfall.
Type of rainfall N0 (m
3 mm
1) I (mm
1)
Drizzle 30,000 I ¼ 5:7  R
0:21
Widespread 7000 I ¼ 4:1  R
0:21
Thunderstorm 1400 I ¼ 3:0  R
0:21
Marshall and Palmer 8000 I ¼ 4:1  R
0:21
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 89

occurrence is required in order that potential hazards of rain can not always increase with the increase of falling distance or time.
be assessed for aircraft flight operation. The raindrop was not only subjected to the force of gravity but also
subjected to the aerodynamic drag and the buoyancy of the air
3.1. Raindrop size distribution during the falling process. Among these forces, the resistance of
the air increases with the increasing raindrop speed until it
The analysis of rainfall data has traditionally assumed that the reaches the maximum velocity named terminal velocity when
raindrop size distribution has exponential form with N(Dp) the force of gravity is equal to the resultant force of the drag and
(m
3 mm
1) equal to the number of raindrops per unit volume the buoyancy. Then the raindrop will fall at a uniform terminal
per unit size interval having equivolume spherical diameter Dp velocity.
(mm), which can be expressed by Research on terminal velocity of raindrop has been a long
history, the most widely used form of terminal velocity is devel-
NðDp Þ ¼ N0 expð
IDp Þ ð0 r Dp rDpmax Þ ð1Þ
oped by Markowitz [72] as a function of the raindrop size and

3
1
where Dpmax is the maximum drop diameter. N 0 (m mm ) and I altitude. At low altitudes
(mm
1) are the parameters of NðDp Þ. This form was originally ( "   #)
Dp 1:147
developed by Marshall and Palmer [63] who also suggested that V T ðDp Þ ¼ 9:58 1
exp
ð7Þ
I varied with rainfall rate R (mm h
1, which will be discussed in 1:77
the next subsection) as I ¼ 4:1  R
0:21 (mm
1) associated with where V T ðDp Þ is the terminal velocity.
the constant N 0 ¼ 8000 (m
3 mm
1). A similar analysis of the A correction for it aloft is given by Markowitz as
raindrop size spectra of Laws and Parsons [64] indicated that their
 0:4
data could also be expressed by the exponential form of Eq. (1) but ρ
V T ðDp Þ ¼ V T0 ðDp Þ 0 ð8Þ
with I ¼ 3:8  R
0:20 and N0 weakly dependent on rainfall rate of ρa
N 0 ¼ 5100R
0:03 . More recent studies [65] showed that the values
where V T0 ðDp Þ is the terminal velocity consistent with the density
of N 0 and I are dependent on different types of rain, as shown in
of air aloft ρ0 .
Table 1. There are many other references available in the raindrop-
related research, such as research papers [66–70], review article
[71], etc. 3.4. The range of rainfall rates

The range of rainfall rates that an aircraft may encounter varies

3.2. The intensity of rain
from light rate of 5 mm/h to very large rainfall rates. The ground-
level world record of short-duration rainfall accumulation of
Two important characteristic quantities generally used to
1.23 in. in one minute was measured in an intense afternoon
describe intensity of rain are rainfall rate (R) and liquid water
thunderstorm in Unionville, Maryland, 4 July, 1956 [32,73,74]. This
content (LWC). Rainfall rate is the linear accumulation depth at
volume of rain in one-minute time interval is equal to a rainfall
ground level per unit time (usually in unit of mm/h) used to
rate of 1875 mm/h (73.8 in./h). The world's greatest annual mean
characterize rainfall at ground level. Liquid water content is the
rainfall precipitation is 624 in., measured in Kauai, Missouri
mass of liquid water per unit volume usually in unit of g/m3. The
between 24 July 1947 and 27 July 1948 [74]. The greatest rainfall
relationship between R and LWC is dependent on the type of
rate record of airborne measurements is approximately 2920 mm/
rainfall. According to the investigation report by Marshall in 1948
h, corresponding to 44 g/m3, measured by Roys and Kessler using
[63], the liquid water content can be expressed in the following
an instrumented F-100 aircraft [32,75].
form:
Z 1
π
LWC ¼ ρw D3p NðDp ÞdðDp Þ ð2Þ 3.5. The frequency of rainfall occurrence
0 6
where ρw is the density of water in unit of g/m3. The majority of the ground-based rainfall database was aver-
Using the aforementioned expressions for the drop size dis- aged over relatively long time using a weighing-bucket, more
tribution for different types of rain in Table 1 and integrating Eq. commonly referred to as a tipping bucket by Jones and Sims
(2), the LWC is related to rainfall rate by [76,77]. They analyzed the data collected by weighing-bucket rain
gages placed all over the world for over one year. These data are
30; 000π10
3
Drizzle : LWC ¼ ¼ 0:08928R0:84 ð3Þ useful for determining the potential of a given rainfall rate
5:74 R
0:84 encountered. Fig. 10 is a summary of the averaged zonal frequency
distribution curves obtained. It indicates that for about two
7000π10
3
Widespread : LWC ¼ ¼ 0:07782R0:84 ð4Þ minutes every year in the maritime subtropical zone, a rainfall
4:14 R
0:84 rate greater than 200 mm/h could be anticipated at any location.
While according to Melson [78], this measurement technique
1400π10
3
Thunderstorm : LWC ¼ ¼ 0:05430R0:84 ð5Þ masks the characteristics of the short-duration, high-intensity rain
34 R
0:84 associated with thunderstorms.
Marshall and Palmer: In the 21st century, weather extremes in particular heavy
rainfall have caused massive losses throughout the world, which
8000π10
3
LWC ¼ ¼ 0:08894R0:84 ð6Þ attracts more attention of meteorology and aviation researchers.
4:14 R
0:84
On 12 August 2002, 312 mm of rain fell in 24 h in Zinnwald-
Georgenfeld and caused the disastrous August 2002 flood in the
3.3. The terminal velocity of raindrop Elbe River basin and parts of the Danube basin which broke the
records of floods observed since the 13th century [79]. During July
Once it was thought that when a raindrop was falling from the 2010, heavy rainfall caused the worst flood of its history in
sky to the ground, its speed could reach hundreds of meters Pakistan [80]. Just for one year of 2011, extreme weather caused
per second, which was of great harm to human beings. In fact, the losses over USD one billion for each of 14 events in United States
motion of raindrop in the sky cannot be a free fall, its speed will [81]. Using statistics, scientist can compare the number of recent
90 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 10. Averaged relationship between rainfall rate and frequency for four rain
climatological zones [76].
Fig. 11. Linear trends in total seasonal rainfall and frequency of heavy rainfall
events for various countries [83].
extreme events with that expected in an unchanging climate [82].
Most countries that underwent a significant change in monthly or
seasonal rainfall also experienced a disproportionate change in the
amount of rainfall during heavy and extreme precipitation events
[83], as shown in Fig. 11.
4. Rain research techniques
Since the eighties in the last century, many researchers of
research institutions and universities at home and abroad have
done a lot of work on the influences of rain on aircraft aero-
dynamics. Generally, there are three broad kinds of approaches for
rainfall-aircraft research, namely analytic estimation investigation,
experimental investigation (including full-scale flight test and
scale model test) and numerical simulation.
4.1. Analytic estimation investigation
In this part, the analytic estimation investigation of rainfall is
specifically a method that estimates the effects of rain on aircraft
aerodynamic performance by establishing analytic models based
on substantial experimental data or empirical formulas, excluding
the following computational-fluid-dynamics Navier–Stokes
numerical simulation. In the process of modeling, a few reasonable
assumptions may be used to simplify the analytic work. For
example, as is known that raindrops striking an aircraft lose
momentum to the aircraft and thus change the velocity of the
aircraft. The vertical component of the raindrop velocity imparts
downward momentum to the aircraft, tending to make it sink, and
the horizontal component of the raindrop velocity decelerates the
aircraft because the aircraft momentum is lost to accelerate the
droplets to the aircraft velocity. When estimating the impacted
momentum, an assumption of inelastic collision between rain-
drops and the aircraft was made by Haines and Luers [19,28],
which underestimates the momentum of raindrops imparting to
the aircraft, because in fact a fraction of incident raindrops will
reflect at an angle greater than 901, the impact force from this
fraction of raindrops with elastic collision is actually greater than
that using the inelastic collision assumption. For another instance,
when calculating film thickness, a numerical model which ignored
pressure gradient, gravity and other effects was adopted by Haines
and Luers, this would certainly cause the calculated water film
thickness less approximate to the realistic one. There is a scarcity
of published literature in analytic research of two-phase boundary
layer around an airfoil. Henry et al. presented an integral approach
to study the two-phase boundary layer problem [84], in which the
main assumption that the existence of zero shear stress at the wall
and on the interface is not physically realistic. Bilanin analytically
investigated the rain effect on airfoil aerodynamic performance
[85] and assumed that the ejecta layer thickness was constant. Hsu
thought this assumption was quite doubtful and improved the
analytic research in [86]. Wan and Wang also made an attempt to
estimate aircraft performance degradation under adverse weather
conditions of ice accretion, low level windshear, heavy rain and
turbulence via a FW-factor analytic approach [87].
Although this method may sometimes get a relatively less
accurate result than experiment as the established assumptions
may underestimate or overestimate the effects of rain, in the years
when experimental facilities were limited and the computer
technology did not exist or was at the starting stage, the analytic
investigation was a necessary method for evaluating the effects of
rain on aircraft aerodynamic performance because of its low cost
and high safety compared with experiment and numerical
simulation.
4.2. Experimental investigation
4.2.1. Full-scale flight test
Full-scale flight test is to test the aerodynamic performance of
full-scale aircrafts under natural or artificial rainfall environment.
Necessary specimens and measuring devices must be equipped on
the aircraft to record the environmental parameters for the pilot to
explore the safe flight envelope of the aircraft. Although there
exist a few literatures available about the capability description of
full-scale flight test on effects of rainfall on aircraft aerodynamics
[88–91], the public testing results of them are scarce. The most
extensive application of flight test is to study rain erosion of
materials [40–45], which is out of the topic of this overview, thus
we will not discuss this method in the later chapters of this review.
Although the result of natural flight test is most approximate to
the actual situation in flight, this method is somewhat risky and
strongly limited by the natural environment. Besides, acquiring
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 91

aerodynamic data of large-scale models to verify the rain influence

on aircraft flight performance under the natural condition of
rainfall is a big challenge to testing technology. Although the most
ideal testing method is to measure aircraft performance under
heavy rain environment, even if not impossible, it is hard to
achieve accurate performance parameters in an environment of
wind, turbulence as well as the same rainfall intensity due to the
variability of natural environment. Therefore, flight test cannot be
widely used in practice.

4.2.2. Scale model test

Scale model test uses test devices to measure the performance
of a scaled model aircraft or sections like wing and fuselage under
simulated scaled rain environment. Currently, there are three
types of scale model tests including a rotating arm in which a
model is set at the end of a counterbalanced rotating beam [85,92],
traditional wind tunnels, and a track in which a model is propelled
down a straight track segment. Rotating arm facilities are quite
useful and effective for studying the complete splashback of a
single droplet impacting normal to a surface. However, airfoil
performance measurements have not been tried with the rotating
arm facility for the reason that the centrifugal effects on the water
film would influence the experimental results. The wind tunnel is
usually applied to study small-scale models in simulated rain
Fig. 12. Configuration of the rotating arm and droplet impacting apparatus [85,92].
environment, the challenge or difficulty lies in obtaining a uniform
size distribution of the raindrops with minimum effects on the
tunnel flow conditions. The track is usually for large-scale ground
test of rain effects on aircraft performance, the water manifolding
and nozzle distribution need to be elaborate and extensive to
cover the test area.

4.2.2.1. Single-drop rotating arm experiment. An experiment of the

complete splashback for a single droplet impacting normal to a
surface using a rotating arm was conducted both by C.D.I personnel
as well as by staff from INTA in 1986 and 1987 [85]. The apparatus
for conducting the impact tests centered around an existing 15 hp
variable speed drive unit and the configuration of the rotating arm is
shown in Fig. 12. Drops are produced periodically, fall under gravity
and impact a NACA 0024 airfoil on the end of the rotating arm.
Analysis of the experimental data suggested that the normal or
near normal impact of raindrops produced an ejecta cloud directed
along the impacting surface. This cloud was consisted of micron-
size droplets splashed back with their total mass less than 30% of
the impacting mass and their velocities several times the incoming
droplet velocity, but the splashed droplets were rapidly deceler- Fig. 13. Photograph of the Rain Simulation System in the NASA Langley Research
Center 14- by 22-Foot Subsonic Tunnel [95,96].
ated by aerodynamic drag forces. The water which did not splash
back formed an irregular film on the surface of the airfoil,
complicating the aerodynamic analysis of the airfoil. Though the
rotating arm test method can reveal some impacting character- rain spray system is set at the upstream position and the spray nozzles
istics of single drops, the experimenter also suggested that the are facing with the testing model. Water is injected from the spray
existing methods of modeling droplet impact and splashback nozzles into the flow field to immerse the testing model. Then the
would not be able to predict airfoil performance degradation aerodynamic data of the testing model are obtained in the simulated
in rain. scaled rain environment. More details about the experimental
apparatus can be referred to in [32,60]. The simulation of natural
4.2.2.2. Small-scale wind-tunnel experiment. The initial develop- rain in a wind tunnel environment must be able to simultaneously
ment of a precipitation environment simulation in a wind tunnel produce a natural raindrop size distribution, alter the intensity level
considered the injection of ice particles or water droplets into the and supply uniform rain spray coverage at the model location with
tunnel airflow and allowed the flow to accelerate them to minimal influences on the tunnel free-stream conditions.
supersonic velocity by aerodynamic drag [93]. The application of As detailed in reference [97,98], when water with a velocity
wind-tunnel experiment approach to simulate rain environment much lower than the air free-stream velocity is injected into a
mainly concentrated in the 1980s and 1990s. So far, most of the high-velocity airstream, the lager raindrops will almost immedi-
experimental data of rainfall are acquired in the NASA Langley ately break up into much smaller drops. Although this trouble can
Research Center 14- by 22-Foot Subsonic Tunnel [94], a be alleviated by increasing the nozzle dynamic pressure in order to
photograph of NACA 64-210 airfoil model installed in this wind make raindrop initial velocity approach the air free-stream velo-
tunnel of the Langley Research Center is shown in Fig. 13 [95,96]. The city, it will result in a too high rain intensity.
92 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Based on the exploratory small-scale tests indicating that the
droplet size distribution and the rain intensity were a function of
the nozzle design, water injection pressure and air free-stream
velocity, an extensive experimental research was conducted by the
Jet Propulsion Laboratory (JPL) to develop a nozzle design that
could produce a wider range of rain intensity levels and a raindrop
Fig. 14. Three types of nozzles used in NASA wind-tunnel experiments [32].
Fig. 15. Experimental apparatus developed by Beitel et al. [93]. (a) Tunnel assembly and (b) Perspective of tunnel test section area.
size distribution including 2 mm and larger [99,100]. The JPL
designed nozzles consisted of two nozzle configurations, a 5-
and a 7-tube (B1N5 and B1N7, respectively) hypodermic nozzles
[32], as shown in Fig. 14. The nozzle on the most right side is a
commercial fan jet nozzle which had an elliptical cross section and
produced the highest rain intensity while the 5-tube nozzle
produced the lowest. All the three types of nozzles had been used
in Bezos wind-tunnel experiments [60] to simulate various levels
of rain intensity.
Beitel and Matthews also developed a precipitation particle
injection system operated in a supersonic aerothermal wind
tunnel to simulate real flight precipitation environment [93], the
test apparatus is illustrated in Fig. 15. The main part is a Mach-4
aerothermal wind tunnel which is a supersonic closed-circuit,
high-temperature, free-jet wind tunnel with an axisymmetric
contoured nozzle and an exit of diameter of 25 in.. The required
quantity of ice particles/water droplets are produced with a
separate system. A high-pressure vessel is developed to transfer
these particles suspending in liquid nitrogen. The vessel is pres-
surized to a specific pressure level to provide a desired particulate
mass flow rate and can be opened at the desired time to initiate
the particulate cloud. The test model is mounted on a sting
support device in an installation tank beneath the tunnel test
section for the particulate flow to immerse. Although this set of
test apparatus is primarily used for ice accretion simulations, it can
also be applied to rain simulations by using ice particles to
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 16. An artist illustration of the large-scale track method developed to test
aerodynamic performance in simulated rain environment [101,102].
represent water droplets. A brief study showed negligible differ-
ence between the ice and water impact craters through limited
data, however, this conclusion cannot be generalized for all
materials of interest and it needs to be evaluated for each
situation.
4.2.2.3. Large-scale track experiment. Due to the aforementioned
technical challenges of full-scale flight test and as an alternative
approach of wind tunnel test, NASA and the Federal Aviation
Administration (FAA, USA) developed a large-scale, ground-based
test capability at the Langley Aircraft Landing Dynamics Facility
(ALDF) in 1987 [101–103]. Fig. 16 illustrates the research method
developed to get aerodynamic performance data on large-scale
wing sections immersed in a simulated natural rain environment.
A large-scale wing section is mounted on a test vehicle which is
propelled by a propulsion system down a track traveling through a
highly-concentrated rain environment which is produced by a
series of spray nozzles suspended above the track.
The ALDF consists of three main components: the test carriage,
the propulsion system and the arrestment system. The concrete
constitution can be referred to in [101,102]. The test carriage is
propelled along the track by a high-pressure jet of water. An open
bay area is designed on the carriage to accommodate tire testing
hardware. The large-scale testing wing section is mounted above
the open bay area. A grooved nose block assembly is equipped at
the front of the carriage to capture the arresting gear cables which
are used to stop the carriage. The ALDF was modified to be a large-
scale aerodynamic performance testing facility with minimum
modifications to the test carriage and track test section. The test
carriage was modified to mount a wing section with a specified
chord length in the open bay area. The track test section was
modified to incorporate an overhead rain simulation system which
could provide rainfall conditions of LWC of 2, 9, 26 and 35 g/m3
and raindrop size distribution ranging from 0.5 mm to 4 mm.
This technique supplies a simulated rain environment which is
more realistic than that produced by a wind tunnel. The water
spray nozzles in the vertical direction can produce raindrops with
the proper drop size distribution and terminal velocities discov-
ered in severe natural thunderstorms. The test carriage is pro-
pelled into a simulated rain shower which exemplifies an aircraft
flying into a rain shower.
4.3. Computational fluid dynamics numerical simulation
With the advent of the computer era, researchers began to pay
much attention to using the tool of Computational Fluid Dynamics
(CFD) to study the rain effects on aircraft aerodynamics and
achieved some important progress. The CFD numerical simulation
approach is a way achieving the goal of studying engineering,
physical and even various kinds of natural problems through the
numerical calculation and image display method based on the
electronic computer. It is first to establish the mathematical model
93
which reflects the nature of the problems (engineering problems,
physical problems, etc.), namely, to establish differential equations
of involving variables and the corresponding definite conditions.
For the field of aircraft aerodynamics, the fundamental mathema-
tical model of flow field can be the well-known Navier–Stokes (N–
S) equations or Reynolds-Averaged Navier–Stokes (RANS) equa-
tions. After the mathematical model is established, it is necessary
to pursue a calculation method with high efficiency and accuracy,
such as discretization and solving methods of the differential
equations, establishment of the body-fitted coordinate, definition
of the boundary conditions, etc. After determining the calculation
method, it is time to start programming and calculation. Practice
shows that the third part is the main body of the whole work,
taking up most of the total time. After the calculation is accom-
plished, large amounts of data are processed through various post-
process technologies.
Numerical simulation of aircraft aerodynamic performance
under rain conditions began in the 1980s, but the development
is relatively slow. The present air-rain issue is exactly a gas-fluid
two-phase flow problem. Currently, there are two numerical
simulation methods applied in gas-fluid two-phase flows in
general, namely the Eulerian two-phase flow method and the
Lagrangian particle tracking method [104,105]. The former method
treats the droplets distributed in the airflow field as quasi fluid
permeating the gas phase and solves Eulerian conservation equa-
tions for each phase. Interphase exchanges of mass, momentum
and energy are included as source terms in the appropriate
conservation equations. This Eulerian approach has been applied
by a huge number of researchers in two-phase flow problems. The
second is a method that firstly solves the Eulerian conservation
equations of the gas phase and then establishes the particle
Lagrangian equations of motion based on Newton's second law
to simulate and track the trajectories of the droplets in the airflow
field. A one-way coupled model assumes that the motion of the
discrete phase is affected by the continuous phase, while the latter
is unaffected by the former [106,107]. A two-way coupled model
assumes that there is a two-way exchange of mass, momentum or
energy between the continuous and dispersed phases [58]. Among
them, the momentum exchange between the two phases is due to
interphase drag forces. The momentum transfer from the contin-
uous phase to the discrete phase acts as a momentum source/sink
in the continuous phase momentum balance in any subsequent
calculations of the continuous phase flow field. Two approaches
have been developed in the Lagrangian two-way coupled models:
one is a non-iterative transient scheme where both phases are
calculated simultaneously [108] and the other is the iterative
particle source in cell (PSI-Cell) method where the two phases
are considered separately and iterated until a steady state is
reached [109].
In terms of numerical simulation technology, the basic steps of
studying the aircraft aerodynamics in rain conditions include the
following aspects: studying the field of the airflow around the
body, raindrop impact characteristics, flow characteristics of water
on the surface of the aircraft and the aerodynamic performance
research in rain conditions. Numerical simulation is an economic
and effective method that can predict the aerodynamics of aircraft
in rain conditions, and can simulate rain conditions beyond the
range of model experiment and flight test. With the continuous
development of computer technology and Computational Fluid
Dynamics, numerical simulation will be an important means of
studying the effects of rain on aircraft aerodynamics.
4.4. Scaling considerations
The scaling laws are necessary to ascertain the utility of the
scale model test and numerical simulation techniques for
94 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127

Table 2 usually somewhat arbitrary. The splashback model for raindrop is

Dimensional variables controlling force N. a very important part of numerical study of rain effects on aircraft,
since it affects the momentum loss of the boundary layer in
ρa Density of air
ρw Density of water accelerating the splashed back droplets. However, the splash is a
va Kinematic viscosity of air quite complex process, the characteristics of the splashed back
vw Kinematic viscosity of water droplets are influenced by many factors such as incidence velocity,
σw,a Surface tension of water/air incidence angle. Besides, many characteristic changes may be
σw,s Surface tension of water/solid
σa,s Surface tension of air/solid
induced during the splash process due to deformation, evapora-
D Volume average drop diameter tion, interaction between each other, etc. Thus there exists no
l Mean spacing between drops sufficient experimental data to accurately identify the drop splash
c Airfoil chord length process, usually simple or simplified models of splashback are
U Flight speed
adopted to study the effect of droplet splashback on the boundary
α Angle of attack
N Aerodynamic force on airfoil layer and aerodynamic characteristics of aircraft.
In spite of the aforementioned difficulties, both the wind-
tunnel experiment and CFD numerical simulation techniques
determining the effects of rain at full-scale conditions. In 1985, appear to provide a valid estimate of the effects of rainfall on
Bilanin determined that variables such as kinematic viscosity, aircraft aerodynamics at full-scale conditions.
surface tension interaction, mean raindrop spacing, air and rain-
drop density, volume average raindrop diameter and velocity are
necessary parameters for scaling rain effects on airfoil aerody- 5. Aerodynamic simulation considerations
namic performance [110]. The normalized force N on the airfoil is a
function of nine nondimensional groups as shown in the following Detrimental aerodynamic effects caused by rainfall which are
equations: of most concern to researchers include the following aspects:

N ● Aerodynamic penalties, including reductions in lifting capacity

¼ f ðπ 1 ; …; π 9 Þ
ρa U 2 c2
where
2
π 1 ¼ cU
νa ; π 2 ¼ cU
νw ; π 3 ¼ ρwσUw;a D
σ w;s σ a;s l
π4 ¼ ; π5 ¼ ; π6 ¼
σ w;a σ w;a c
D ρ
π 7 ¼ ; π 8 ¼ α; π 9 ¼ a
c ρw
where the dimensional variables that control the aerodynamic
force N generated on the airfoil are shown in Table 2.
However, as mentioned in [110,111], difficulties will arise when
trying to preserve all the scaling parameters during considering
testing conditions for model testing. Both the Reynolds number
and Weber number are included in the scaling laws for rain. The
Reynolds number is proportional to the testing velocity, thus as
the model size is decreased the testing velocity must increase to
maintain an identical full-scale Reynolds number. The Weber
number, however, defined as the ratio of inertial forces to surface
tension forces cannot be preserved simultaneously since it varies
as the square of the testing velocity. In addition, there are many
operational difficulties in scale model tests and CFD simulations
which will be presented in the following text.
For the wind tunnel approach, except for the aforementioned
difficulty in producing large size drops typical of natural rain while
at the same time acquiring the desired rainfall intensity and drop
size distribution, another difficulty lies in the measurement error
as mentioned in [112]. For the wind-tunnel test rule, LWC is
identified as a function of the free-stream velocity, the spray area
and the mass flow rate. Due to the dynamic nature of water drops,
the boundaries of the spray region are not adequately straight
lines at any instant in time, there exists an unstable fluctuation in
the width and height boundaries of the spray region. Therefore, it
brings an inevitable inaccuracy in achieving the spray area by
photographic means.
For the CFD numerical simulation approach, there are many
difficulties in modeling the droplet impact, splashback, and the
formed water film on the surface of aircraft. As little literature
exists on the nature of high-speed droplet impacts such as those
occur on an airfoil surface in rain, the impact model used in
numerical simulation of airfoil aerodynamic performance in rain is
ð9Þ
and increases in drag of lifting surfaces, such as main wings and
tail wings, especially loss of control due to reductions in
maximum lift capability, i.e. reductions in stall angle of attack
and associated pitching moment changes.
● Increased roughness of wing surface due to raindrop impinge-
ment and the formed uneven surface water film.
● Complicated mechanisms that dominate the varied character-
istics of the flow field of aircraft in rain and thus deteriorate the
ð10Þ
aerodynamic performance.
For each of these, there are a number of key flow physics
characteristics and testing simulation features which must be
carefully evaluated if existing experimental results for rainfall
effects are to be properly interpreted. These key factors and
considerations are presented in detail in the following subsections.
5.1. Spray manifold effects
The position of the spray manifold is very important for rain
tests in wind tunnels. The position must allow adequate time for
the stabilization of the accelerating water droplets and the devel-
opment of the spray region. The position must also provide an
adequate distance to minimize any manifold-induced wake dis-
turbances on the wing and to account for gravity effects on the
water droplets. A common practice is to keep the manifold and the
test model near the centerline of the tunnel to minimizing wall
interference effects. Drop injection influences on freestream wind
velocities also need to be considered, though they are sometimes
insignificant in relation to the scopes of some investigations.
5.2. Nozzle design effects
Before the rain tests in wind tunnels, it is important to note
that nozzle design can influence the aerodynamics of test models.
Nozzle calibration studies [98] have shown that changing LWC and
nozzle types can bring about differences in spray characteristics.
Among these differences are variations in the volume percentage
mean drop diameter, the arithmetic mean drop diameter and the
dispersal patterns, all of which have been thought to account for
the variation in aerodynamics in rain for the same LWC. The
differences in spray characteristics may produce nonlinear
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
aerodynamics results at low Reynolds numbers [60–62], yet the
full influence of these nonlinearities has not been completely
understood.
5.3. Cruise configuration lift performance
Establishing meaningful incremental lift effects due to rainfall
on this class of geometries is typically more dependent on
achieving representative aerodynamic lift levels and characteris-
tics in the baseline (dry) condition than in a wet one, particularly
when using ground-test facilities. This is because the maximum
lift levels of airfoil, wing, etc. in dry condition are usually quite
sensitive to the leading-edge flow physics, which is particular for
the laminar-flow airfoils where the flow is a strong function of the
model design and installation, the test Reynolds number, etc.
Figs. 17 and 18 show the baseline aerodynamic coefficients of
two 2-D wing sections tested in two pressurized wing tunnels at
different Reynolds numbers [60,113,114]. It can be clearly seen that
for the laminar-flow Wortmann FX63-137 airfoil, the difference of
lift coefficients between Reynolds numbers of 100,000 and
200,000 can be neglected. As the Reynolds number is increased
to 300,000, the deviation of lift coefficient curves is enlarged,
especially at high angles of attack. But for the transport-type NACA
64-210 airfoil, Reynolds number has a negligible effect on lift
coefficient over the range tested. Thus, for different types of
airfoils, Reynolds number has different effects on aerodynamic
forces in dry condition, which is an important factor when
assessing the baseline aerodynamic forces and characteristics.
Another aspect of Reynolds number effect on the baseline
aerodynamic lift levels depends on different types of wind tunnels,
as is noted in [115] that using a lower Reynolds number measure-
ments in a low-turbulence pressure tunnel for baseline aerody-
namic forces ( here the baseline is aerodynamic forces of un-iced
airfoil for ice accretion and clean or dry airfoil for rainfall) would
yield a lower maximum lift level than in a high-turbulence
pressure tunnel. Thus, a lower maximum lift penalty would
certainly appear.
Synthesizing the above two factors of Reynolds number effect
considered, one must be very careful about using low Reynolds
number test results to predict high Reynolds number perfor-
mances. Examples illustrating this concern will be presented in
the following sections. However, low Reynolds number data on
rain may be useful in analyzing high Reynolds number trends if
interpreted carefully with full consideration of the physics of
the flow.
Two other critical factors necessary to be considered when
assessing 2-D airfoil maximum lift degradations caused by rain are
Fig. 17. Lift coefficient of Wortmann FX63-137 airfoil vs angle of attack at various
Reynolds numbers in dry condition, data obtained from [113,114].
95
understanding the different flow physics for different types of
airfoils and confirming the minimum effect of side wall on wind-
tunnel test measurements. It is well known in fluid dynamics that
some airfoil types undergo a leading-edge stall identified by a
rapid loss in lift beyond stall, some like relatively thin airfoils have
a leading-edge stall without the rapid loss in lift beyond stall,
while others have a less abrupt trailing-edge stall, an example of
this case can be referred to in [61]. In addition, as is proved in [61]
that water on the airfoil greatly roughens the airfoil surface and
can cause premature leading-edge transition, which also results in
a loss of lift, but the transition stations and roughening effects are
different for different airfoils. So, it should not be a surprise if the
effects of rain on maximum lift levels are different for these
different types of transition and stall characteristics.
The proper assessment of rain effects on 3-D wings, fuselage,
etc. is somewhat more complex because that, in addition to all the
aforementioned factors which are critical in assessing rain effects
on 2-D airfoils, it is necessary to consider the spanwise variation in
stall and separation initiation and the flow characteristics of water
film on the surface of 3-D configurations, especially for 3-D wings
or rotor blades with taper, sweep or twist. The stall and separation
initiation is critical in determining subsequent maximum lift and
pitching moment changes caused by rain. Since well-designed
wings and other lifting surfaces usually do not happen to stall
across the whole span and regions away from the stall initiation
are less important, it is important to select a particular spanwise
location for achieving anticipated stall characteristics. The span-
wise flow of water film can greatly affect the boundary-layer
transition and separation locations in rain. Though the chordwise
flow characteristics of water film have been explored by some
investigations [61,62,116], the spanwise effects have be little
concerned. Thus, it is an interesting aspect meriting further
research.
5.4. High-lift configuration lift performance
Due to the wider range of geometries, flow physics features and
flow conditions, it is much more complex to appropriately assess
the effects of rain on the lift characteristics of high-lift configura-
tions. It is necessary to understand the relative importance of
interactions between the various elements in different rain con-
ditions. As mentioned in [115] that in no-ice condition, the
maximum lift attainable is limited by an eventual reduction in lift
on the main element, the lift reduction on the main element
beyond maximum lift is caused by an unloading of the aft part of
the main element, etc., which should be the same situations in the
dry condition for rain effect study. Unfortunately, there exists
Fig. 18. Lift coefficient of NACA 64-210 airfoil vs angle of attack at various Reynolds
numbers in dry condition, data obtained from [60].
96 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
neither numerical nor experimental investigation on the effects of
rain on components of a multi-element lifting surface. Thus, this is
also an aspect deserving deep investigation. For example, since
different parts have different impacts on the flow field, it can be a
good perspective to study the lift performance variations of each
components in different flight conditions like takeoff and
approach.
Similar to the situation for single-element lifting-surface con-
figurations, the proper assessment of rainfall effects on the lift
performance of 3-D multi-element lifting surfaces becomes rather
more complicated. It is necessary to determine where stall or
separation on the flap is initiated spanwise on the dry and wet
conditions at takeoff or approach stage. Further, due to the much
higher angles of attack when high-lift devices are extended, wing-
mounted engines can have a more dominant influence on stall or
separation initiation. At this moment, it would then be quite
difficult to acquire any very generally applicable conclusions
regarding rainfall effects on maximum lift characteristics for such
configurations. As aforementioned that for the high-lift configura-
tion, the existing water sheet may clog the gap causing premature
trailing edge separation and stall or interact with the airflow over
the main airfoil section resulting in a change in the aerodynamics,
the spanwise flow characteristics of water film will be much more
complex, thus determining the spanwise stall initiation becomes
more difficult.
Another challenge encountered in numerical simulation is the
grid generation of multi-element configuration. Looking through
all the literatures involving numerical study of rain effects on
airfoils of such complex configuration, almost all adopted the
unstructured grid technique [117–119]. However, as is well known,
unstructured grid cannot reflect the precise information of the
flow field of interest when viscosity is considered. If unstructured
grid is adopted in boundary layer, the size of the mesh will be very
huge. Nowadays, a better method for this is to adopt the hybrid
mesh technique, i.e., first to generate body-fitted quadrilateral or
prism grids which are available for viscous computation, and then
to generate triangular unstructured grids outside the preceding
grids, like in [120]. However, it is very difficult to generate
quadrilateral or prism grids for complex configurations like the
multi-element lifting surfaces here. Another good choice is to
adopt structured grid for mesh generation. Structured grid is very
proper for the viscous computation of boundary-layer flow field.
However, for aircraft with complex configurations, it is also very
hard to generate a single domain of body-fitted structured grid,
sometimes it is even impossible. In practical applications, the most
common measure is using domain-partitioned mesh technique, i.
e., to partition the flow field into several sub-domains based on
the configuration characteristics of the body and then generate
grid in every sub-domains. Fig. 19 shows a domain-partitioned
structured mesh of NACA 64-210 high-lift configuration airfoil,
which is a representative 2-D four-element (slat, main wing
section, vane and flap) geometry in the high-lift configuration
and has been widely used to study the aerodynamic effects of rain
on multi-element lifting surfaces both experimentally and
computationally.
5.5. Drag characteristics
The mechanism of rain-induced drag increase is very complex
in a word. Interactions between raindrops and air, the increased
roughness of wing surface due to the uneven water film and the
increased skin friction because of transition movement are poten-
tial sources of drag penalties by rain. Determining significant drag
increments of aircraft due to impacts of rainfall is likewise
dependent on a proper evaluation of the same or similar testing
considerations as required for the lift characteristics. To start with,
Fig. 19. Structured mesh of NACA 64-210 high-lift configuration airfoil using the
domain-partitioned mesh technique.
flow conditions on the dry baseline must be understood and
accounted for regarding where boundary-layer transition occurs,
as well as any tend towards boundary-layer separation [61,62],
since two potential effects of rainfall on aircraft are causing
premature boundary-layer transition and separation, which are
two important sources of drag penalty due to rainfall. For single-
element or cruise configuration lifting surfaces, the type of airfoil,
testing Reynolds number, intensity of rain, and measurement
accuracy of test facility are all important variables in determining
drag penalty due to rain. Consideration of baseline characteristics
is particularly important when assessing drag penalty due to
relatively light rain where the difference between causing separa-
tion or not can have a great influence on the magnitude of the
incremental drag penalty. Likewise, testing Reynolds number as
well as the type of airfoil, intensity of rain are all logically crucial
variables in determining the incremental drag penalty due to rain.
5.6. Surface tension effects
Droplets impacting aircraft surface at high velocities and the
inertial forces at impact totally dominate the surface tension
forces. References [110,111] identified the effect of surface tension
on the heavy rain phenomenon through the Weber number
squared and conjectured on the differences anticipated at subscale
since it had been noted previously in the said references that the
ratio of inertial to viscous forces (Reynolds number) and inertial to
surface tension forces (Weber number squared) are likely to be
distorted at subscale. Bezos et al. tested the wetted aerodynamic
characteristics of the NACA 64-210 landing configuration with and
without the surface-tension reducing agent added to the water at
the test Reynolds numbers of 2.6  106 (equal to dynamic pressure
of 30 psf) and 3.3  106 ( dynamic pressure of 50 psf) [60]. The
chemical properties of the water was changed by the addition of
the surface-tension reducing agent, which allowed the water
droplets to shatter and spread outward as a thin film more easily.
Data for the treated rain spray showed the same trends in
maximum lift coefficient, maximum-lift angle of attack and drag
coefficient as the data for the untreated case. Although there was a
slight decrease in lift for the treated water spray at both Reynolds
numbers, the maximum lift was nearly the same. This means
small-scale model test of high-lift configurations which have little
laminar flow is not strongly dependent on surface tension effects.
However, the data from reference [121,122] (as shown in the latter
Figs. 67–69) showed that surface tension had a strong impact on
the aerodynamic performance of laminar flow airfoils in the same
treated rain spray. Therefore, airfoils with different types have
different extents of dependence on surface tension effects.
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 97

6. Effects of rainfall on aircraft aerodynamics

6.1. Rain-induced aerodynamic penalties

Hansman and Craig [61,62] conducted wind-tunnel experi-

ments on the Wortmann FX67-K170, NACA 0012 and NACA 64-
210 airfoils (as shown in Fig. 20) at a simulated rain rate of
1000 mm/h and at a low Reynolds number of 3.1  105. The lift
results show that at low angles of attack, the Wortamann airfoil
has the greatest lift degradation (around 25%) and the NACA 64-
210 airfoil has the least (about 5%), as shown in Figs. 21–23. The
erratic lift polar for the dry Wortmann FX67-K170 airfoil at high
angles of attack was thought to be related to the low Reynolds
number boundary-layer behaviors, such as laminar separation
bubbles and trailing-edge separation [123]. In the wet condition,
the Wortmann FX67-K170 airfoil experiences an accompanying
decrease in lift performance compared to the dry condition, but
the polar is much smoother, the higher-angle-of-attack behavior of
the Wortmann airfoil in the wet condition is typical of an airfoil
with a turbulent boundary layer. For the NACA 0012 airfoil, the
stall angle is shifted from 141 in the dry condition to 181 in the wet

Fig. 22. Lift coefficient vs angle of attack for the NACA 0012 airfoil in the dry and
wet conditions [61,62].

Fig. 20. Airfoils tested in the wind-tunnel experiment by Hansman and Craig
[61,62].

Fig. 23. Lift coefficient vs angle of attack for the NACA 64-210 airfoil in the dry and
wet conditions [61,62].

Fig. 21. Lift coefficient vs angle of attack for the Wortmann FX67-K170 airfoil in the
dry and wet conditions [61,62].
condition, while the stall angle for the NACA 64-210 airfoil in the
wet condition is approximately the same as that in the dry
condition. At high angles of attack, the two NACA-type airfoils
experienced improved aerodynamic performance in rain condi-
tions due to a reduction of boundary-layer separation.
Marchman et al. [113,114]conducted wind-tunnel tests to
determine the effects of rain on the aerodynamic performance of
the laminar Wortmann FX63-137 wing with aspect ratio of 6 at
low Reynolds numbers of 100,000 to 300,00. For all the three
Reynolds numbers, rain caused a reduction in lift coefficient CL at
most angles of attack as well as the maximum lift coefficient CLmax
and a more gentle stall and even a gentle secondary stall with no
hysteresis. The major difference between different Reynolds
98 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
numbers lies in the effects of rain on drag characteristics. Although
at the lowest Reynolds number, the very low drag values result in
considerable scatter in the data, it appears that the presence of
rain has no adverse effect on drag relative to the dry case, which is
of the same conclusion for the case of the moderate Reynolds
number. However, the drag is definitely increased by the rain at
the highest Reynolds number. The presence of water droplets on
the wing significantly increases turbulence and skin-friction drag
at moderate angles of attack, which has much the same influence
as enhanced surface roughness. Tests were also conducted using a
waxed wing surface to see if the enhanced beading effect of a
waxed surface would alter the effects of rain. Comparisons
between the unwaxed and waxed cases show that the dry cases
are nearly identical, whereas the waxed wing resulted in increased
beading of the water droplets, increasing the surface roughness
and further reducing the lift and increasing the drag over most the
range of angle of attack in rain. The effect of aspect ratio on the
aerodynamic performance in rain was also tested. The test results
of an aspect-ratio-4 wing at Reynolds number of 200,000 show
that the lift coefficient results are quite similar to that for the
Fig. 24. Details of the cruise and high-lift configurations of the NACA 64-210 airfoil
[60,124,125].
Fig. 25. Aerodynamic coefficients of NACA 62-210 cruise configuration in no-rain and rain conditions at two Reynolds numbers [60]. (a) Re =2.6e06 and (b) Re =3.3e06.
higher-aspect-ratio case. However, the lower-aspect-ratio tests
show a definite reduction in drag in rain at most angles of attack
compared to little change in the aspect-ratio-6 case in the same
condition.
Campbell and Bezos [112] conducted an investigation to deter-
mine the steady-state and transitional effects of simulated heavy
rain on the subsonic aerodynamic characteristics of a wing with a
NACA 23015 airfoil section in the Langley 14- by 22-Foot Subsonic
Tunnel. First, the effects of spray manifold, Reynolds number and
nozzle design effects were tested and discussed. It suggested that
the effects of the manifold wake were overall very small, especially
for higher velocities. At very low velocities, the manifold wake
resulted in an increase of the stall angle of attack as well as a slight
increase in drag. As to the Reynolds number effect, it was quite
small for all configurations in the range of Reynolds number
tested. In addition, results for the high-lift configurations appeared
to be less sensitive to the nozzle differences, which was encoura-
ging since the largest aerodynamic degradations often occur for
these configurations in rain. Next, the steady-state and transitional
performance data were measured and showed that heavy rain
resulted in large degradations in aerodynamic performance for
high-lift configurations. A 27% decrease in lift and a 39% increase
in drag were observed for the largest flap deflection of 201 at the
highest test speed. The largest performance losses occurred at
angles of attack near maximum lift for the wet wing, which were
several degrees below that for the dry wing. The aerodynamics of
the wing underwent both linear and highly nonlinear character-
istics as it entered the rain environment. Nonlinear characteristics
were more apparent for higher flap deflections and higher speeds.
Bezos et al. spent a long time to study the aerodynamic penalty
of the NACA 64-210 cruise and high-lift configurations (shown in
Fig. 24) in simulated rain environment in the Langley 14- by 22-
Foot Subsonic Tunnel [60,124,125]. From the aerodynamic data of
the NACA 64-210 cruise configuration from reference [60] shown
in Fig. 25 we can see significant reductions in the maximum lift
coefficient and slope of the lift curve as LWC is increased. An
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 99

Fig. 26. Aerodynamic coefficients of NACA 62-210 landing configuration in no-rain and rain conditions at two Reynolds numbers [60]. (a) Re =2.6e06 and (b) Re =3.3e06.

increase in the drag coefficient (CD) was also obtained for both
Reynolds numbers at low and moderate angles of attack. The effect
of rain on pitching moment for the cruise configuration is
negligible for both Reynolds numbers prior to stall, while past
stall there appears to be an enhancement in the pitching-moment
performance for both Reynolds numbers in rain. In addition, an
about 11 premature stall angle of attack was also indicated for
Reynolds number of 3.3  106 and the rain condition of LWC of
30 g/m3. The effects of rain on the NACA 64-210 landing config-
uration for both Reynolds numbers are shown in Fig. 26 (data
obtained from reference [60]). As we can see that the stall angle of
attack moved forward as LWC increased for both Reynolds num-
bers and the greatest reduction in stall angle of attack emerged at
the higher Reynolds number and LWC of 39 g/m3 (about 81).
Similar to the data of the cruise configuration, data also show a
progressive reduction in the slope of the lift curve as LWC
increases. The drag at a constant lift condition appear sensitive
to LWC and test velocity. For example, the drag data for Reynolds
number of 2.6  106 show increases in drag coefficient of 23% and
46% for LWC of 29 and 46 g/m3 at a constant lift coefficient of
CL ¼ 2.2, respectively. At the higher Reynolds number and the same
lift coefficient, the obtained increases in drag coefficient are nearly
the same (around 2%) for LWC of 16 and 36 g/m3, respectively. As
to the effect of rain on pitching moment for the landing config-
uration, we can acquire that the slope of the pitching moment (Cm)
curve increases progressively with increasing LWC for both Rey-
nolds numbers prior to stall. Past stall, rain continued to decrease
the pitching-moment performance at both Reynolds numbers,
which was completely different from the effect of rain on the
pitching moment of the cruise configuration. In addition to these,
a phenomenon of premature stall is observed on this high-lift
configuration for all rain intensities of interest of both Reynolds-
number cases. In Bezos's previous investigations [124,125], the
effects of partial-span and full-span spray coverage on the aero-
dynamic performance of the NACA 64-210 high-lift configuration
was studied in simulated rain environment. The lift data obtained Fig. 27. Lift and drag coefficients vs angle of attack given by Hsu [126].
100 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
with partial-span spray coverage showed the same incremental
decrease in lift with increasing LWC at a constant angle of attack as
the data from full-span spray coverage, indicating the gross effect
of the water spray on the model had notably greater influence on
the airfoil lift performance than any three-dimensional effects
caused by the water spray partially covering the model span. The
drag data obtained with partial-span spray coverage showed the
same incremental increase in drag with increasing LWC at con-
stant lift as the full-span-coverage drag data. The trends in the lift
and drag measurements were overall the same for either partial-
span or full-span spray coverage, i.e., a progressive reduction in
maximum lift and the angle of attack for maximum lift.
Hsu [126] used the CFD numerical simulation approach to
study the two-phase flow around the laminar Wortmann FX67-
K170 airfoil in simulated heavy rain environment. The following
assumptions were made to simplify the problem: first, the fluid
flow was nonsteady, viscous and incompressible; second, laminar
boundary existed in the flow region; third, the airfoil was
represented by a flat plate. Computed lift and drag coefficients of
the airfoil in rain condition were compared with that obtained by
Hansman's experimental results [61,62] and are presented in
Fig. 27. The trends of the computed lift- and drag-coefficient
curves are generally consistent with those obtained by the
experiment, though there is a quite large difference between the
computed and the experimental results of the drag coefficients in
rain. Hsu's effort is highly appreciated and can be viewed as an
early attempt to study the adverse effects of rainfall on aircraft
aerodynamics by CFD simulation approach.
Fig. 28. Effect of rain on the pressure distribution [127]. (a) Pressure coefficient for
NACA 0012 airfoil in fine rain (b) Pressure coefficient for NACA 0012 airfoil in coarse
rain.
Calaress et al. [127] numerically analyzed rain effects on a
NACA 0012 airfoil. The full 2-D Navier–Stokes equations were
solved. In that literature, three categories of forces were identified
as the affecting forces of rain on airfoils, i.e., impact forces, surface
forces and volume or body forces. The impact of raindrops on the
surface imparts a momentum exchange, the surface forces effec-
tively increase surface roughness and the volume forces change
the airflow field by the droplet drag acting as a body force in the
Navier–Stokes equations of the air phase. Only the effects of the
third forces were analyzed in that paper. The rain conditions of
LWC of 0, 10, 100 and 1000 g/m3 were adopted. Through the
calculation, it was concluded that for the fine rain with very small
droplets (the droplet parameter Re⪡1), the pressure distribution
changes drastically with LWC. The pressure coefficient increases
on the bottom surface and decreases on the top surface with
increasing LWC, as shown in Fig. 28. Meanwhile, the lift and drag
coefficients increase with increasing LWC for the fine rain (see
Fig. 29), but the lift-to-drag-ratio decreases with increasing LWC
(see Fig. 30), indicating that the increase in drag coefficient is
somewhat higher than the increase in lift coefficient. However, for
the coarse rain with large droplets (the droplet parameter Re⪢1),
the calculation indicated no evident changes with LWC.
Valentine and Decker [58] developed a two-way coupled
Lagrangian–Eulerian scheme to numerically simulate the flow
around a NACA 64-210 airfoil in rain. The air phase was treated
as continuum and solved with a thin layer incompressible Navier–
Stokes code [128,129], while the raindrop particle phase was
treated as discrete and was solved by integrating Lagrangian
Fig. 29. Lift coefficient (CL) and drag coefficient (CD) vs liquid water content (LWC)
[127].
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 30. Lift-to-drag-ratio (L/D) vs liquid water content (LWC), data obtained from [127].
Fig. 31. Splashback model used by Valentine and Decker [58].
equations of motion and tracking individual particles through the
flow field. An appropriate splashback model is very critical for the
numerical scheme, since it directly influences the momentum loss
by the boundary layer to accelerate the splashed back droplets. A
splashback model was used by Valentine et al. in [58], as shown in
Fig. 31. It was assumed that the splashback occurred over a range
of angle θ centered at about the normal direction of the surface.
When the angle of incidence β¼901, 50% of the incident mass was
assumed to be splashed back over a range of θ¼1201 with an
initial velocity equal to the incident velocity; when β¼01, all the
aforementioned parameters of the incident droplets went to zero;
between the two extreme cases, a linear distribution of splashback
parameters was assumed according to the angle of incidence. The
simulation results of Valentine and Decker are shown in Fig. 32
which were compared with the experimental results of Bezos et al.
[60]. The experimental results showed a decrease in lift and an
increase in drag at most angles of attack with a maximum lift
degradation at stall. However, the numerical results showed no
change in airfoil aerodynamic performance until stall was reached.
A rain-induced more severe stall angle of attack was simulated by
Valentine.
Wan et al. have been studying on effects of rain on aircraft
aerodynamic performance since 2003 [130–132]. Airfoils such as
NACA 4412, NACA 0012 and NACA 64-210 airfoils, a 3-D Blended-
Wing-Body aircraft and helicopter rotor blades were selected to
investigate their aerodynamic performance in heavy rain via a
two-phase flow approach. Generally, it was also concluded that
heavy rain can cause severe aerodynamic penalties, i.e., lift loss
and drag gain, to all the research objects of interest. Besides, this
influence of rain becomes increasingly detrimental as the rain
intensity, LWC, increases. These work is highly appreciated that it
opens a wide perspective of understanding both the approach and
the application prospect of studying rain effects on aircraft aero-
dynamic performance.
Yeom et al. [133] numerically investigated the aerodynamic
characteristics of the NACA 0012 airfoil in a transonic air-water
101
Fig. 32. Comparison of lift and drag coefficients of NACA 64-210 airfoil between
experimental and Valentine's numerical results [58].
droplet airflow. A compressible Eulerian two-fluid model was
solved by the WAF-HLL scheme developed earlier by them [134].
In that study, various gas–droplet interactions such as heat
transfer, phase change, droplet breakup were involved, but neither
droplet–droplet interaction nor droplet–wall interaction was taken
into consideration. The effects of the initial droplet volume
fraction and size on the chordwise wall pressure coefficient and
the scaled aerodynamic coefficients of the airfoil in the air-water
flow were examined and elaborated, as respectively shown in
Fig. 33 through Fig. 36. From Fig. 33 it can be clearly seen that the
shock exists at the trailing edge of the airfoil in the gas-only flow
and when the droplet volume fraction αd equals 0.01%. As the
droplet volume fraction continues to increase, the wall pressure
increases rapidly ahead of the shock but slightly decreases aft of
the shock, resulting in the disappearance of the shock for the
higher droplet volume fraction. In Fig. 34, the lift coefficient
increases and the drag coefficient decreases with the increasing
angle of attack. The slopes of the lift curves initially decrease to the
minimum value at the droplet volume fraction of 0.1% and rise
afterwards. In contrast, the slopes of the drag curves increase
monotonically as the droplet volume fraction increases for all the
angles of attack of interest. In addition, it can be concluded that
the aerodynamic performance of the airfoil degrades seriously at
high density of water drops (i.e. the droplet volume fraction). For
example, at angle of attack of 11, the scaled lift-to-drag-ratio drops
from 0.625 (¼ 0.75/1.2) at droplet volume fraction of 0.01% to
0.185% (¼ 0.37/2), 0.100 (¼ 0.55/5.5) and 0.096 (¼1.25/13) at
droplet volume fraction of 0.1%, 1% and 10%, respectively. In terms
of the effect of the initial droplet size, shock wave exists for the
smallest droplet size of 1 μm but disappears for higher droplet
sizes (see Fig. 35). Besides, it can be seen from Fig. 36 that the
scaled lift coefficient increases while the scaled drag coefficient
102 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127

Fig. 33. Wall pressure coefficient for (a) the upper surface and (b) the lower surface at angle of attack of 11. The droplet diameter d is initially set to 100 μm [133].

Fig. 34. Scaled (a) lift coefficient and (b) drag coefficient vs the initial droplet volume fraction, the superscript 0 stands for the gas-only state. The droplet diameter d is
initially set to 100 μm [133].

decreases with the angle of attack. For angle of attack of 11, the
scaled lift-to-drag-ratio rises from 0.110 ( ¼0.44/4) at initial
droplet diameter of 1 μm to 0.119 (¼0.43/3.6) and 0.18 (¼0.36/2)
at droplet volume fraction of 10 μm and 100 μm, respectively.
Therefore, the droplets of a smaller initial size implement a more
adverse effect on the aerodynamic performance of the airfoil in
the wet condition.
Ismail and Cao et al. studied the aerodynamic efficiency of 2-D
NACA 0012 cruise and landing configuration airfoil and 3-D NACA
0012 rectangular wing in heavy rain via two-phase flow approach
[135,136]. The discrete phase model (DPM) was used to simulate
the discrete raindrop particles suspending in the continuous air
phase. For the NACA 0012 cruise configuration at Reynolds
number of 3.1  105 and LWC of 39 g/m3, lift decreased and drag
increased significantly, the lift-to-drag-ratio (L/D) degradation
reached up to 25% for angle of attack of 61 through 101. For the
NACA 0012 landing configuration at Reynolds number of 1.7  106
and LWC of 22 g/m3, the lift-to-drag-ratio decreased up to 20%
relative to the no-rain case for angle of attack of 61 to 121 at which
airplanes usually perform takeoff and landing. So the degradation
in aerodynamic efficiency at the moment of takeoff and landing
becomes critical. For the 3-D NACA 0012 rectangular wing at
Reynolds number of 7  106 and LWC of 32 g/m3, the lift-to-drag-
ratio degradation reached up to 10%. Fig. 37 shows the pressure
and velocity distributions around different spanwise sections of
the 3-D NACA 0012 rectangular wing for LWC of 0 and 32 g/m3. It
can be clearly seen that the pressure difference between the upper
and lower surfaces of the wing leading edge decreases with rain
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 103

Fig. 35. Wall pressure coefficient for (a) the upper surface and (b) the lower surface at angle of attack of 11. The droplet volume fraction αd is initially set to 0.1% [133].

Fig. 36. Scaled (a) lift coefficient and (b) drag coefficient vs the initial droplet diameter, the superscript 0 stands for the gas-only state. The droplet volume fraction αd is
initially set to 0.1% [133].

case, thus the lift and lift-to-drag-ratio of the wing decrease in
heavy rain conditions
Recently, Wu and Cao [137] developed a two-way momentum
coupled Eulerian–Lagrangian two-phase flow approach to study
the heavy rain effects on the aerodynamic performance of a 2-D
NACA 64-210 cruise and landing configuration airfoil and a 3-D
NACA 64-210 cruise configuration rectangular wing. The steady-
state incompressible Reynolds-averaged Navier–Stokes equations
for the continuous phase and the Langrangian equations of motion
with a developed algorithm for the discrete phase were solved
alternately by incorporating an interphase momentum exchange
term to couple the interphase effects. Scaling laws for raind-
rops and the Spalart–Allmaras (S–A) turbulence model were
implemented for all the test cases. The thin liquid film model
[138] was adopted to model droplet–wall interactions as shown in
Fig. 38. The criteria by which the regimes are partitioned are based
on the raindrop Weber number defined using the relative velocity
and drop diameter as
ρp u2r Dp
We ¼ ð11Þ
σp
where ρp is the density of raindrop particle, ur is the raindrop
relative velocity in the frame of the wall ( i.e. ur ¼ up
uwall ), and
σ p is the particle surface tension. As shown in Fig. 38, the stick
regime occurs when an impinging drop with extremely low
impact energy adheres to the film surface in approximately a
104 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127

Fig. 37. Pressure and velocity distributions around the 3-D NACA 0012 rectangular wing in the dry and wet conditions [135]. (a) Pressure distribution, no rain, (b) Pressure
distribution, rain, (c) Velocity distribution, no rain and (d) Velocity distribution, rain.

spherical form and the wall temperature is below the pure
adhesion temperature Tpa; the rebound regime occurs when the
impinging drop with low impact energy bounces off the film using
the lost energy of the air layer trapped between the drop and
liquid film; the third regime, spread, occurs at higher W e where
the drop merges with the liquid film upon impact; the final
regime, splash, occurs if the particle impacting the surface has
an effectively high energy, and then it splashes and breaks up into
a cloud of secondary droplets or a parcel and a crater is formed
with a crown at the periphery. The transition criteria for these
four impingement regimes are listed in Table 3, where f is the
frequency of the impinging drops.
For the simulation results of the cruise configuration airfoil, the
maximum percentage decrease in lift coefficient and increase in
drag coefficient reached respectively up to 19% and 41% for AOA 4–
161 at LWC of 39 g/m3, and the maximum L/D percentage degrada-
tion reached by 39%. For the simulation results of NACA 64-210
landing configuration, the maximum percentage decrease in lift
coefficient and increase in drag coefficient reached respectively up
to 15% and 35% for AOA 4–201 at LWC of 29 g/m3, and the
maximum L/D percentage degradation reached by 37%. A 11 rain-
induced more severe stall was also observed. The numerical
results of the two cases were compared and were in good
agreement with the experimental data. Finally for the 3-D case,
the maximum percentage decrease in lift and increase in drag
reached respectively up to 33% and 24% at LWC of 50 g/m3, and the
maximum L/D percentage degradation by 36% of the dry value.
In addition, a phenomenon, i.e. premature boundary-layer
separation, was observed on the NACA 64-210 airfoil by Valentine
et al. [58] and by Wu et al. [137] at the rain conditions of LWC of 25
and 39 g/m3, respectively. Fig. 39 shows the pressure distribution
of the air flow field and streamlines around the NACA 64-210
airfoil for LWC of 0 and 39 g/m3 at AOA of 121 and 131. Here the
reference pressure is 101,325 Pa. Generally, the pressure difference
between the upper and lower surfaces of the airfoil leading edge
decreases with rain case for both angles of attack, so the lift and
lift-to-drag-ratio decrease. Another obvious phenomenon is that at
AOA of 121, rain has caused no obvious change in the airflow.
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 105

Fig. 38. (a) Mechanisms of mass, momentum and energy transfers for the thin liquid film; (b) a typical wall film cell used in the formulation of the film model; (c) the four
impingement regimes included in the film model; (d) velocity of a drop rebounding from the film [138].

Table 3
Transition criteria for the four droplet impingement
regimes.
Regime Transition criteria
Stick We o5
Rebound 5o W e o 10
Spread 10o W e o 18:02 Dp ðρp =σ p Þ1=2 up 1=4 f
3=4
Splash 18:02 Dp ðρ =σ p Þ1=2 up 1=4 f
3=4
oWe
p momentum sink around a NACA 64-210 airfoil in simulated rain
While at AOA of 131, rain induces severe separation at the trailing
edge of the upper surface of the airfoil, suggesting a rain-induced
premature separation occurs at this moment.
6.2. Raindrop impinging and surface water flow characteristics
Two physical phenomena have been hypothesized to contri-
bute to the degradation of airfoil performance in rain. As raindrops
strike an airfoil, some fraction of the incident mass is splashed
back and forms an ejecta fog near the leading edge, while the
remainder forms a thin water film upon the airfoil surface [60], as
shown in Fig. 40. The acceleration of the splashed-back droplets by
the air is hypothesized to act as a momentum sink, decelerating
the boundary-layer airflow field. As subsequent raindrops impact
the water film, many craters are formed on it and make it uneven.
The uneven water film can effectively roughen the airfoil surface,
resulting in reduction of lift and increase in drag. Thus, in this part,
two aspects will be discussed, i.e., raindrop impinging and splash-
ing characteristics and surface water flow behaviors. Due to the
fact that these two water behavior processes occur simultaneously,
therefore they are put into the same subsection.
As to the raindrop impinging characteristics, Valentine and
Decker developed a one-way coupled Lagrangian particle tracking
scheme to evaluate droplet concentrations and the accompanying
environment [106,107]. A particle splash model similar to that
used in [58] was employed to simulate the impinging character-
istics of raindrops approaching the wall surface, as shown in
Fig. 41, the difference between these two splash models lies in
the value of the parameters selected to determine each model.
Droplet concentrations around the airfoil at angle of attack of
4.0391 is shown in Fig. 42. As to the largest rain rate of 500 mm/h,
the corresponding LWC is about 10 g/m3 according to Eq. (5). The
minimum average diameter of raindrop is about 0.45 mm obtained
by Eqs. (14) and (15) in [106,107] with the first diameter interval
0 oDP o1 mm, thus the maximum mass of individual raindrop is
less than 0.00038 g and the maximum freestream raindrop num-
ber density is of the order of 104 raindrops/m3, well below the
values of the first contour levels of all the three cases. Therefore, it
can be assumed that the majority of droplets near the airfoil
are formed by the secondary splashed-back droplets. A high-
concentration region similar to the ejecta fog and water bow wave
shown in Fig. 40 can be seen. As rain density is increased, the
106 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127

Fig. 39. Pressure and streamline distributions around the NACA 64-210 airfoil [137].

water layer is more dense. Due to the stagnation point being
slightly below the leading edge at this angle of attack, more
raindrops impacting the leading edge are carried over the upper
surface of the airfoil than the lower surface. Also, the downward
component of raindrop velocity causes more raindrops impact the
upper surface, especially for those larger raindrops with higher
vertical velocities.
Yeom et al. also studied the variation of the thickness of the
droplet breakup layer [133], the maximum and minimum values of
the thickness of the droplet breakup layer for different initial
droplet volume fractions at the three different angles of attack are
shown in Fig. 43. A water bow wave near the airfoil can be clearly
seen. As to the thickness of the droplet breakup layer, the
maximum thickness δmax gradually increases with the angle of
attack while the minimum thickness δmin shows little dependence
on the angle of attack.
In Hansman's photographic observations of the top surface of
the test airfoil in rain [121,122], three regions were observed with
significantly different water flow characteristics, as shown in
Fig. 44. The droplet impingement zone is where the splash craters
discussed by Haines and Luers [19,28] occur. A high-speed sha-
dowgraph of a droplet impact crater occurring in this zone was
taken by Hansman in this experiment, along with a cloud of small
residual raindrops extending forward from the leading edge, as
shown in Fig. 45. In the forward runback zone which located just
aft of the impingement region, shear from the external flow
dominated the water behavior. In the aft runback zone, as the
external shear was reduced, the water tended to stagnate and
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 107

Fig. 40. Rain impacts on airfoil forming the droplet ejecta fog and surface water
film [60].

Fig. 43. Maximum and minimum thickness of the droplet breakup layer around
the airfoil [133].

Fig. 41. Another splashback model used by Valentine and Decker [106,107].

Fig. 44. Three zones with different water flow behaviors for the Wortmann FX67-
K170 airfoil [121].

Fig. 42. Droplet concentrations (drops/m3) around the NACA 64-210 airfoil for
rainfall rates of R ¼(a) 100 mm/h, maximum¼ 1.2e9; (b) 300 mm/h, max-
imum ¼ 3.6e9; (c) 500 mm/h, maximum ¼5.1e9 [106,107].

other factors like surface tension became dominant. The behavior

of water in the runback zones was also discussed by Hansman and
Barsotti [121]. The water on the gel coated surface behaved in a
similar way to that observed by Hastings [116,139,140] with a
smooth water film in the forward runback zone and rivulets in the
aft zone, while the continuous water film on the soap coated
surface was remarkably roughened by irregular waves and the
runoff streams were much broader and more irregular. In the
waxed case where the largest aerodynamic performance loss
occurred, water was clearly observed to bead in both the forward Fig. 45. A typical crater formed by drop impacts and ejecta fog formed by the
and aft zone, resulting in a rougher surface than the other two residual droplets near the leading edge of the Wortmann FX67-K170 airfoil [121].
108 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127

Fig. 46. Typical water film pattern for the clean NACA 0012 wing at θ ¼41 (LWC ¼ Fig. 49. Typical water film pattern for the clean NACA 64-210 wing at α ¼141
20 g/m3, Re¼ 8.9  106) [116,140]. (LWC ¼39 g/m3, Re¼ 3.8  106) [139].

Fig. 47. Typical water film pattern for the clean NACA 0012 wing at θ ¼9.31 Fig. 50. Typical water film pattern for the flapped NACA 64-210 wing at α¼ 81
(LWC ¼ 20 g/m3, Re¼8.9  106) [116,140]. (LWC ¼20 g/m3, Re¼ 2.5  106) [139].

cases. Therefore, the roughness appeared to its greater perfor-

Fig. 48. Typical water film pattern for the clean NACA 64-210 wing at α ¼01
(LWC ¼ 39 g/m3, Re¼3.8  106) [139].
mance degradation.
Hastings and Manuel [116,139,140] discussed some of the
surface water characteristics obtained in two rain experiments
including a clean NACA 0012 wing section (i.e. no flaps) [140] and
a NACA 64-210 wing section with and without flaps at the Langley
Research Center [124]. Photographs of typical water film patterns
on the upper surfaces of both clean wing sections showed that at
relatively low angles of attack, a smooth continuous film formed
on the forward section of the surface. After that location, the film
broke down into many individual runoff streams, extending to the
leading edge (see Figs. 46–48). At large angles of attack, when
separation occurred, a large region of apparently separated flow
between breakdown of the continuous sheet and the trailing edge
(see Fig. 49), which was thought to be a potential source of
aerodynamic penalties. However, for the flapped NACA 64-210
wing, significant water could be seen near the trailing-edge flaps
at angle of attack α¼81, and at α ¼201 a large area of flow
separation appeared near the trailing-edge flaps as well as a large
amount of water accumulated in and around the flap gaps, as seen
in Figs. 50 and 51.
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Campbell and Bezos [112] investigated the water flow patterns
existing on the wing surfaces. At low to moderate angles of attack
below angle of stall at which flow is attached on the wing surface,
the water adheres to the wing surface and forms the surface water
patterns of droplet impacting and splashing, water film and rivulet
runoff as mentioned in [110,111,116,139], as shown is Fig. 52(a). It
is the development of this water film layer, coupled with raindrops
impacting the wing surface that interacts with the air boundary
layer that is suspected of inducing the aerodynamic performance
losses of the wing. Similar water flow characteristics were also
depicted by Bilanin for high-lift configurations, as shown in Fig. 53.
As angle of attack is increased but not yet to stall, the detrimental
effects of rain continue to develop, as shown in Fig. 52(b). The
separated airflow above the upper surface of the wing results in
regional pooling of the water as well as a breakdown of the flow
patterns in Fig. 52(a). However, the lower-surface water film
behavior appears to be little affected by the upper-surface separa-
tion, therefore the rain-induced aerodynamic degradations due to
continuous raindrop impact still exist for the lower surface. Similar
surface water flow characteristics was simulated by Wu et al. [137]
for the NACA 64-210 rectangular wing at angle of attack of 141 and
LWC of 50 g/m3, the uneven water film formed upon the wing
Fig. 51. Typical water film pattern for the flapped NACA 64-210 wing at α ¼201
(LWC ¼ 20 g/m3, Re¼2.5  106) [139].
Fig. 52. Water flow patterns about the tested wing [112]. (a) Attached wing water flow and (b) Stalled wing water flow.
109
surface are shown in Fig. 54. It can be observed that thin water
film forms on both the upper and lower surfaces of the wing.
However, due to the relatively high angle of attack, more raindrops
concentrate upon the lower surface and form thicker water film.
Besides, thanks to the craters developed by raindrop impingement,
instability, breakup into rivulets and interaction with the air
boundary layer, the water film distributes unevenly upon both
surfaces.
Thompson et al. [141] identified four surface-water flow
regions over a NACA 4412 wing with aspect ratio of 6 tested in
simulated moderate rain environment, as shown in Fig. 55. The
droplet-impact region extends from the leading edge downstream
to about 6% chord on the suction side of the airfoil. Large craters
and waves were observed in the surface water film in this region.
“Ejecta-fog” described by Bilanin in [110,111] was also observed to
appear in this region, near the leading edge. This “ejecta fog” is
virtually the scattering of smaller droplets splashed back into the
boundary layer flow and drain energy from the boundary layer to
accelerate these splashed droplets. This loss of the boundary-layer
energy affects the development of the downstream boundary layer
with premature boundary-layer separation and accounts for aero-
dynamic penalties of airfoil in rain. Rivulet-formation region is
characterized by convection of a surface-water film covering the
airfoil surface like a sheet. Surface water in this region flows
downstream and convects with some surface waves and thus can
be turbulent, which will probably shift the location of mean-
streamline detachment aft and cause loss in lift and gain in drag.
In the rivulet-formation region, separate rivulets and beads form.
Fig. 53. Water film pattern above a flapped wing depicted in [111].
110 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 54. Water-film mass distribution upon the upper surface of the NACA 64-210 rectangular wing [137]. (a) Water-film mass distribution on the upper surface and
(b) Water-film mass distribution on the lower surface.
Fig. 55. Identification of the surface-water flow regions: 1, droplet-impact region; 2,
film-convection region; 3, rivulet-formation region; 4, droplet-convection region [141].
The flow characteristics of the rivulets in this region in the said
experiment are similar to those found in [142–144]. The last is the
droplet-convection region in which beads coalescing with other
beads and unstable rivulets flow downstream to the trailing edge
and are torn away by the external flow. One year later, Thompson
and Jang obtained measurements of lift, drag and moment
coefficients for the same rectangular planform NACA 4412 wing
section with wettable and nonwettable surface coatings in mod-
erate rain [145]. Their testing results show that aerodynamic
efficiency caused by rain depended on the location of rivulet
formation and the diameter of the rivulets. Lift forces decrease
with longer film-convection regions and drag forces increase with
increasing diameter of rivulets. Thinner films, flatter thinner
rivulets and flatter droplets were found on wing surfaces with
wettable coatings, which resulted in the least loss in lift-to-drag
ratio induced by rain. Also, droplet and rivulet diameter were
found to be more critical to aerodynamic performance in rain than
the location of rivulet formation. In 1999, Thompson et al. [146]
used numerical simulation approach to calculate the onset loca-
tion of rivulet formation in the surface-water flow over a NACA
4412 airfoil in rain at Reynolds numbers of 100,000, 250,000 and
400,000 and compared with his experimental results in [145].
A model where rivulets formed when the surface-tension shear
stress between the airfoil surface and the liquid equals the inter-
face shear stress due to the aerodynamic forces on the liquid by air
motion was used. The former shear stress was obtained from the
measured values of surface tension and contact angle at the wall-
water interface in the experiment and the latter shear stress was
calculated by solving the Reynolds-averaged Navier–Stokes equa-
tions. Fig. 56 compares the onset locations of rivulets formed on
the wettable and nonwettable surfaces of the NACA 4412 airfoil
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 111

Fig. 57. Time-dependent lift and water runback position for the Wortmann FX67-
K170 airfoil [61,62]. (a) 2-deg angle of attack and (b) 15-deg angle of attack.

Fig. 56. Onset locations of the rivulets formed on the two types of NACA 4412
airfoil in rain [146]. (a) Wettable surface and (b) nonwettable surface.

numerically and experimentally. Generally, both the calculated

and experimental results suggest that the onset of the formed
rivulets on the wettable and nonwettable surfaces moves
upstream as the freestream velocity and the angle of attack
increase. Measured and calculated locations of rivulet onset are
sensibly within about 2% and 3% of chord for the wettable and
nonwettable surfaces in the range of angle of attack from 51 to 121,
although agreement is poorer at lower angles of attack. The
maximum difference between the calculated and experimental
onset locations occurs on the nonwettable surface of the airfoil at
the lowest Reynolds number about 4% of chord downstream for all
angles of attack, which corresponds to an underprediction of about
0.001 in the calculated skin-friction coefficient at almost angles of
Fig. 58. Time-dependent lift and water runback position for the NACA 0012 airfoil
attack in the present calculation. [61,62]. (a) 2-deg angle of attack and (b) 15-deg angle of attack.
Hansman and Craig [61,62] also recorded the time-dependent
lift output and water runback behavior, as shown in Figs. 57–59 for
the Wortmann FX67-K170, NACA 0012 and NACA 64-210 airfoils, airfoils. At approximately 0.9 s, water appeared at the trailing
respectively. For the Wortmann FX67-K170 airfoil at angle of edge and progressed forward. This was an indication of a trailing-
attack of 21, within the initial 0.2 s, there was an evident loss of edge separation as the water was forced from the lower surface
lift, which appeared to be a consequence of premature boundary- around the trailing edge to the upper surface. At the higher angle
layer transition near the leading edge, as can be seen in Fig. 57(a). of attack, the lift loss of the Wortmann airfoil occurred much more
After the initial loss of lift, an extra lift change occurred at times slowly than compared to the low-angle-of-attack lift behavior, as
scales consistent with the water runback time at the full chord shown in Fig. 57(b). For the NACA 0012 airfoil at angle of attack of
positions. This behavior was observed for each of the tested 21, the magnitude of the initial lift loss was much less than that of
112 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 59. Time-dependent lift and water runback position for the NACA 64-210
airfoil [61,62]. (a) 1-deg angle of attack and (b) 15-deg angle of attack.
Fig. 60. Computer-generated flowfield and pressure distribution for the Wortmann
FX67-K170 airfoil at 2-deg angle of attack for (a) natural transition and (b) forced
transition at 5% chord on the top and bottom surfaces. Boundary-layer separation
(S), transition (T), reattachment (R), and forced transition (FT) locations are
illustrated [61,62].
the Wortmann airfoil (seen in Fig. 58(a)), but the mechanisms
appeared to be the same. At the higher angle of attack, however,
this airfoil experienced a rapid increase in lift when rain was
Fig. 61. Film thickness for the clean NACA 0012 wing at θ¼ 41 [116].
Fig. 62. Effect of pitch attitude on film thickness for the clean NACA 0012 wing
(LWC ¼12 g/m3 and Re¼ 7.2  106) [140].
Fig. 63. Film thickness for the clean NACA 64-210 wing at α ¼41 [116].
initiated (seen in Fig. 58(b)). There was also a secondary but slow
lift loss corresponding to the time scale of the water runback. The
water runback became almost stagnant downstream of 60% chord,
indicating a trailing-edge separation at this point. From the lift
performance of the NACA 0012 airfoil at high angles of attack in
Fig. 22 we can see that, this trailing-edge separation in the wet
condition did not decrease the airfoil performance as much as
the leading-edge separation occurring in the dry condition. For the
NACA 64-210 airfoil at angle of attack of 11 (seen in Fig. 59(a)), the
initial lift loss occurred within a much longer time of approxi-
mately 2 s compared to the aforementioned two airfoils, but the
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 113

Fig. 64. Film thickness for the clean NACA 64-210 wing (LWC ¼39 g/m3 and
Re ¼3.8  106) [116].

Fig. 66. Lift coefficient vs angle of attack for different surfaces of the Wortmann
FX67-K170 airfoil [121,122].

Fig. 65. Film thickness for the flapped NACA 64-210 wing (LWC ¼ 20 g/m3 and
Re ¼2.5  106) [116].

magnitude of the total lift loss was nearly equivalent to that of the
NACA 0012 airfoil. At the higher angle of attack, the lift increased
within the first 0.3 s after entering the rain condition. The water
film layer was present only within the upstream 10% chord and the
water runback on the upper surface did not begin until about 2 s
after the rain was initiated (seen in Fig. 59(b)), therefore indicating
the performance enhancement was resulted from a leading-edge
behavior. A CFD code that predicts airfoil boundary-layer char-
acteristics [147,148] was also used and showed that for the
Wortmann airfoil at 2-degree angle of attack when boundary-
layer transition was allowed to occur naturally, the CFD code
predicted a boundary-layer transition at 65% in the dry condition
and a trailing-edge boundary-layer separation at approximately
80% chord when transition was forced at the leading edge (as
shown in Fig. 60), which are consistent with the positions
observed by the flow visualization technique in the said experi-
ment. In addition, both the flow visualization technique and the Fig. 67. Drag coefficient vs angle of attack for different surfaces of the Wortmann
CFD code indicated a boundary-layer transition at respectively FX67-K170 airfoil [121,122].
about 55% and 75% for the NACA 0012 and NACA 64-210 airfoils in
the dry condition, while the trailing-edge boundary-layer separa- of 50.4% chord on the lower surface where continuous film
tion behaviors in the wet condition were not mentioned in the breakdown might be occurring (see Figs. 63–64). As for the
report. flapped NACA 64-210 wing, data showed that the film was
In terms of the film thickness, different airfoils and even the relatively independent of angle of attack and was thinner than
same airfoil with different configurations had different dependen- that on the clean configuration (see Fig. 65). The decrease in water
cies on angle of attack and LWC. For the clean NACA 0012 airfoil, film thickness on the upper surface of this high-lift or landing
data indicated that the film thickness increased with increasing configuration was explained by droplet trajectory analyses [149].
LWC and pitch attitude θ (see Figs. 61–62), while for the clean Since data had shown that for a clean airfoil, increasing angle of
NACA 64-210 airfoil, LWC and angle of attack α nearly had no attack resulted in a decrease in the impingement efficiency on the
effects on the water film thickness at all locations except the one upper surface, deflecting flaps and slats had the same effect.
114 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 68. Lift-to-drag-ratio vs angle of attack for different surfaces of the Wortmann
FX67-K170 airfoil [121,122].
Fig. 69. Lift coefficient vs angle of attack for the Wortmann FX67-K170 airfoil with
boundary layer intentionally tripped [121].
The decrease in water film thickness on the lower surface was
thought to be probably due to the effect of the leading slat in
catching and deflecting approaching raindrops. Also, they
Fig. 70. Drag coefficient vs angle of attack for the Wortmann FX67-K170 airfoil
with boundary layer intentionally tripped [121].
Fig. 71. Dry, wet and intentionally tripped lift polars for the Wortmann FX67-K170
airfoil [61,62].
addressed that the probable causes of aerodynamic performance
decrements of the multi-element NACA 64-210 wing in simulated
rain environment resulted from water in and around the flap gaps
at large angles of attack. Wind-tunnel tests showed that the lift
and drag of flapped wings were very sensitive to the geometry of
flap gaps [150,151]. Surface water flowing from the lower to the
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 115

Fig. 74. Near mesh around the NACA 64-210 airfoil with steps placed at 5% chord
positions of the upper and lower surfaces [152].

1.4

1.2

1.0
Lift Coefficient
0.8

0.6 3
Original airfoil(LWC=25g/m )
Original airfoil(dry)
0.4 75% chordwise tripped airfoil(dry)
Fig. 72. Dry, wet and intentionally tripped lift polars for the NACA 0012 airfoil 50% chordwise tripped airfoil(dry)
[61,62]. 25% chordwise tripped airfoil(dry)
0.2
5% chordwise tripped airfoil(dry)

0.0
-2 0 2 4 6 8 10 12 14 16 18
Angle of Attack/Degree
Fig. 75. Numerical lift coefficient vs angle of attack for different tripped positions
on the NACA 64-210 airfoil [152].

upper surface through these gaps could effectively change the

geometry of these flap gaps and thus affect the aerodynamic
characteristics. Besides, as with the single-element configuration,
the roughness and waviness of water film in an adverse pressure
gradient on the aft upper surface might induce local flow separa-
tion and further affect the aerodynamic characteristics. However,
for both airfoils, measurements showed that the film was generally
thicker on the lower surface than that on the upper surface and
the point of maximum thickness was further aft on the lower
surface.

6.3. Potential mechanisms of rain effects

In 1985, Hansman and Barsotti [121,122] assessed the effect of

Fig. 73. Dry, wet and intentionally tripped lift polars for the NACA 64-210 airfoil
[61,62].
surface wettability on the performance degradation of a natural
laminar-flow Wortmann FX67-K170 airfoil in simulated heavy rain
environment. The airfoil was tested with three different surfaces.
The baseline surface was a clean epoxy gel coat that had been
carefully sanded with 660 grit paper before testing. Epoxy paint is
commonly used for wing surface coatings thanks to durability,
affordability and smoothness, even if the contact angle for rain-
drops impacting the surface is about 901. Gel coat can be used to
make the epoxy-painted surface smoother and more wettable. An
incompletely wettable surface was got with a soap-coated airfoil. A
nonwettable surface was obtained by employing a wax coating to
116 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 76. Effect of water spray on canard lift and pitching moment coefficients [153].
Fig. 77. Effect of water spray on canard drag polar [153].
Fig. 78. Effect of fixed transition on canard chordwise pressure distribution, α ¼81
[153].
the airfoil. Results for the lift coefficient (CL), drag coefficient (CD)
and lift-to-drag-ratio (L/D) of the airfoil with different surfaces are
shown from Fig. 66 through Fig. 68. As can be seen that, all the
surfaces experienced performance degradation in the rain condi-
tion, but the difference in the magnitude of the performance loss
for the different surfaces was large. The wax surface had the
largest aerodynamic degradation of 75% reduction in maximum
L/D and the clean gel coat had the least of 45%, which conclusion is
consistent with the test conclusion drawn by Thompson et al., as
Fig. 79. Effect of fixed transition on elevator control effectiveness [153].
seen from Fig. 7 of reference [145]. The zero CL angle increased for
all the different surfaces in the wet condition with the maximum
value of 21 for the wax surface. In the drag data, both the effects of
the surface roughening due to water and momentum transfer from
the raindrops to the airfoil can be seen. The large surface-
wettability effect implied that the performance degradation was
a primary consequence of the roughening effect of the water. The
fact that the minimum degradation observed on the gel coated
surface indicates that a smooth wettable surface may possibly
alleviate the adverse effects of heavy rain.
In Hansman's view of point, there are two primary mechanisms
responsible for the aerodynamic performance degradation in rain.
The initial effect was to induce premature boundary-layer transi-
tion at the leading edge. The second was consistent with water
runback on the top surface which occurred at time scales and
could alter the airfoil geometry effectively [61,62]. The second
effect has been depicted in detail in the second subsection of this
chapter. Here we will discuss some explorations of the first effect.
In Hansman's earlier experimental investigation [121,122], forced
measurements were made on the test airfoil with the boundary
layer tripped at 1/2-, 1/4- and 1/16-chord positions by applying a
thin (2 mm) strip of sand grains with average size of approxi-
mately 0.3 mm. From a comparison of the lift and drag coefficients
shown in Figs. 69 and 70 we can see that, the slop of the lift
coefficient curve decreases as the boundary-layer transition point
moves forward, which correlates with the reduction in slope
observed in the wet condition (see Fig. 69), while the drag curve
tends to increase. The change in drag coefficient as the transition
point moves from the 1/2 to 1/16 chord position is similar to the
observed difference between the gel coat and the wax surface
under the wet condition (see Fig. 70), which indicates that the
drag increase due to rain is accounted for by the rain-induced
premature boundary-layer transition near the leading edge of the
airfoil.
This technique was also employed in Hansman's later experi-
ment to more deeply investigate the mechanisms of rain effects on
the aerodynamic performance of the Wortmann FX67-K170, NACA
0012 and NACA 64-210 airfoils [61,62]. The results of the forced
boundary-layer transition tests are shown in Fig. 71–73. The
behaviors of the Wortmann FX67-K170, NACA 0012 and NACA
64-210 airfoils in rain were best emulated with strips placed at
25%, 5% and 5% chord on the upper surface respectively and 5%
chord on the lower surface for all airfoils. In general, the low-
angle-of-attack performance degradation of all the three airfoils in
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 80. Aerodynamic coefficients of the clean, grit- and wire-tripped wings with a NACA 4412 airfoil section in moderate rain [141].
the wet condition was well emulated by placing trip strips near
the leading edge of each airfoil. However, the high-angle-of-attack
behavior of the two NACA-type airfoils was not emulated by this
technique, thus the trip strips were suspected of being aft of the
leading-edge separation point by them. The technique of emulat-
ing rain-induced performance degradation with trip strips placed
on the various airfoil chord positions indicate that the aerody-
namic performance degradation for airfoils in rain environment is
caused by premature boundary-layer transition at low angles of
attack.
Wu et al. [152] used a Computational Fluid Dynamics (CFD)
approach to emulate the forced boundary-layer transition, that is,
to give a normal and outward increment as modeling the trip strip,
like a step, to the specified position of the airfoil, as shown in
Fig. 74. Trips were placed separately at 5%, 25%, 50%, 75% chord
stations on the top surface and 5% on the bottom surface of the
NACA 64-210 airfoil. This new method was named as numerical
boundary-layer-tripped technique by Wu and Cao. The computa-
tional conditions are consistent with that used by Hansman et al.
[61,62]. Fig. 75 shows the numerical lift coefficient comparison for
the original airfoil in dry and wet (LWC 25 g/m3) conditions and
the tripped airfoil with different tripping positions in the dry
condition. Evidently, trip strips at 5% chord on the top surface best
modeled the wet conditions. However, the high-angle-of-attack
behavior of the NACA 64-210 airfoil is not emulated by the current
tripping technique, just like the results of Hansman's wind-tunnel
experiment. The overall ability to model aerodynamic perfor-
mance degradation in heavy rain condition with our numerically
tripped boundary-layer technique suggests that the aerodynamic
degradation results from premature boundary-layer transition at
low angles of attack, which is also consistent with the conclusion
drawn by Hansman et al. [61,62].
To simulate the underlying mechanisms of rain effects, a wind
tunnel experiment was conducted by Yip to investigate the rain-
induced premature boundary-layer transition on a full-scale
117
canard-configured general aviation airplane in simulated rain
environment [153]. First, tests showed by photographs that
transition occurred at the 55% chord position on the canard and
at the 65% chord location for the wing when the free model was at
an angle of attack of 1.51 and at a Reynolds number of 1.6  10
6
with transition free in the dry condition. Then, carborundum grit
was applied at the 5% chord location of the canard to trip the
boundary layer to transition and aerodynamic data were measured
with forced transition in the dry condition. At last, a water spray
system was located upstream of the model and covered only one
side of the model. The spray rate was approximately 1 gal/min.
Results from free and forced transition of the boundary layer as
well as water-spray tests of the canard shown in Figs. 76 and 77
indicate that the water-spray data show the same trends as the
data of fixed transition on the canard with grit, i.e., a nose-down
pitching-moment increment, a reduction in the slope of the
canard lift curve and an increase in the drag. Besides, Yip
examined the chordwise pressure distribution on the canard and
found that the loss of lift of the canard with a fixed transition was
due to premature separation at the trailing edge probably caused
by the thickened turbulent boundary layer which had to overcome
a sharp pressure recovery near the trailing edge, as shown in
Fig. 78. This premature separation could also cause a decreased
elevator control effectiveness as indicated by the curve of canard
elevator angle to trim vs trim lift coefficient for the free and fixed
transitions in Fig. 79. Therefore, it can be deduced that rain can
have the same effect of premature trailing-edge separation as the
fixed transition with grit. Yip also pointed out in his technical
paper that if the canard were fully immersed in water spray,
the data would be in closer agreement with that with forced
transition.
Thompson et al. also determined whether this technique of
transition fixing for wings in rain would significantly affect the test
chord Reynolds number [141]. Two types of transition strips, i.e., a
grit-trip and a wire-trip were used and placed on the suction side
118 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 81. magnitude of the dimensionless airflow momentum source/sink due to
raindrop particle drag at the leading edge of the NACA 64-210 airfoil for rain rates
of R ¼(a) 100 mm/h, maximum¼ 0.085; (b) 300 mm/h, maximum¼ 0.189;
(c) 500 mm/h, maximum ¼ 0.267. 1/16 Chord position marked for reference
[106,107].
Fig. 82. Extent of the airflow momentum source/sink for rain rate of R ¼500 mm/h.
1/2 Chord position marked for reference [106,107].
of the NACA 4412 wing at 2.5% chord position. Increments in lift,
drag, and moment coefficients of the clean and tripped wings
were obtained at angles of attack from
21 through stall and for
rainfall rates of 77, 100 and 137 mm/h, as shown in Fig. 80. These
increments are the difference between the measurements of
forces coefficients in rain minus that for the airfoil in the dry
condition at the same angle of attack. Seen from the first column,
the lift loss for the clean wing keeps constant at about 15% in the
range of incidence of 51 to 121. At angles of attack above the
maximum lift, separation in the surface-water flow patterns is
well developed and the lift increment becomes positive, suggest-
ing a surface-water-delayed contributing to an enhancement in lift
performance. However, the lift for the grit-tripped wing keeps
about 5% below that for the clean wing from incidence of zero to
the maximum-lift incidence of 121 and then about 5% above until
stall. Increments in lift for the wire-tripped wing appear the least
dependent on the rainfall rate and are smaller than that for the
clean wing but larger than that for the grit-tripped wing. The lift
decrements are about 60% smaller than that for the clean wing.
Seen from the second column, the drag for the clean wing clearly
increases with increasing rainfall rate at all angles of attack. Drag
increases nearly linearly with incidence between incidences of 51
to 121 where the lift loss keeps constant. At angles of attack below
maximum lift, the drag increments remain almost constant at
0.008 and 0.012 for the grit- and wire-tripped wings, respectively.
At higher angles of attack above maximum lift, drag decreases
with increasing angle of attack for both the two tripped wings.
Seen from the last column, the change in moment coefficient for
the clean wing seems negligible from angles of attack from zero to
maximum lift, while moment generally decreases with incidence
for the grit-tripped wing at the same region. At higher angles of
attack, moment coefficient approximately increases linearly with
increasing incidence for both the clean and grit-tripped wing.
Generally, the moment change with wire strips shows an approxi-
mately identical trend to that for the clean wing but a less
dependence on the rainfall rate. Thompson et al. also divided the
range of the test angles of attack into four regimes, i.e., incidences
between
51 and 51, 51 and 121, 121 and 181 and incidences above
181 and discussed the different behaviors of surface-water flow by
the four flow regions shown in the second subsection of this
section. Combining the differences between the behaviors of the
clean, grit- and wire-tripped wings suggest that transition fixing
for wings in rain does not lead to an effective increase in chord
Reynolds number, thus would not improve the correlation
between flight and wind-tunnel data obtained in moderate rain
conditions.
Valentine and Decker also presented the magnitude of the
dimensionless airflow momentum source/sink due to particle drag
at the leading edge of the NACA 64-210 airfoil for three rainfall
rates in their papers [106,107], as shown in Figs. 81 and 82. For the
lowest rain rate, the extent of momentum sink is restricted to a
small region near the leading edge of the airfoil and it becomes
increasingly large as well as the maximum value as the rain
intensity increases. When the rain rate is boosted to 500 mm/h,
the extent of momentum sink extends to approximately one-half
chord position, and the maximum value reaches 0.267 assuming
that the dominant terms of the nondimensionalized Navier–Stokes
equations are of the order of magnitude of 1, which may have a
pronounced influence on the characteristics of the boundary-layer
flow, i.e., a significant loss in boundary-layer momentum, resulting
in lift degradation and possible premature boundary-layer separa-
tion and stall as mentioned above.
To reveal the detrimental effect of the above raindrop particle
drag on the airflow boundary layer momentum, the boundary-
layer velocity profiles at six chordwise positions on the top surface
of the NACA 64-210 airfoil simulated by Valentine and Decker [58]
is presented, as shown in Fig. 83. The deceleration of the boundary
layer near the leading edge by the splashback raindrops compared
to the boundary layer in the no-rain condition is clearly indicated.
The boundary layer appears to recover downstream, but separa-
tion appears to be imminent at the 75% chord position. Similar
conclusion was also drawn by Hsu [126] with the Wortmann
FX67-K170 airfoil and Wu et al. [137] with the same NACA 64-210
airfoil in another rain condition, as shown in Figs. 84–86. The
acceleration of these splashed back droplets by the airflow field
acts as a momentum sink for the boundary layer, resulting in a
decrease in the air velocity field. Deceleration of the boundary
layer induces losses in lift and lift-to-drag-ratio, increase in drag,
premature stall and separation. The effect of uneven water film on
the roughness of the wing surface was also studied by Wu et al.
[137]. Fig. 87 shows the chordwise skin friction coefficient (Cf)
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 119

Fig. 83. Boundary layer velocity profiles at six chordwise positions on the upper surface of the airfoil at angle of attack (α) of 121. Position is measured from the leading edge.
Dimensionless velocity v is plotted vs dimensionless distance x from the airfoil surface [58].

distribution upon the upper and lower surfaces of the wing section
located at z¼0 position for angle of attack of 141 and LWC of 50 g/m3.
Generally, the overall roughness of the lower surface increases
massively due to the presence of the uneven water film, while it
only increases greatly at the leading edge for the upper surface.
Analytically speaking, this may be because of the stall and separation
of the boundary layer at such a relatively high angle of attack of 141,
the downward component of raindrop velocity leads to more impacts
on the leading edge of the upper surface and water on the front of
the upper surface runs back to incorporate into the water film upon
the lower surface, while the rest water moving downstream to the
back of the upper surface forms rivulets, which then disappear and
pool in the separated portion. Therefore, less water will be present on
the back of the upper surface. Another phenomenon seen from the
two pictures is that a rain-induced premature boundary layer
transition occurs at approximately 4% chord of both the upper and
lower surfaces. The similar phenomenon has been observed at 5%
chord of the same airfoil for a lighter rain of LWC of 30 g/m3 in
Hansman's experiment [61,62]. The uneven water film is also
hypothesized to be accountable for the degradation of aircraft
aerodynamic performance in rain, since it effectively roughens the
airfoil surface and causes significant changes to the geometry of
the wing.
The 3-D effect of rain on the spanwise separation character-
istics on the NACA 0012ngular wing was studied by Ismail et al.
[136], as shown in Figs. 88–90. In the dry condition, boundary-
layer separation initiates at the inboard position and is suppressed
at the intermediate and outboard positions, and no apparent
separation at the outboard position, especially. In the rain condi-
tion, the complex 3-D effect of rain can be clearly seen. At the
inner position, rain appears to promote the boundary-layer
separation, which is the main cause of aerodynamic performance
degradation in rain conditions. While at the intermediate position,
rain seems to suppress the boundary-layer separation. At the
outboard position, though there is no clear separation in both
the dry and wet conditions, the boundary layer is suppressed by
rain at this moment too. The rain-suppressed boundary-layer
separation is thought to account for the recovery of the high-
angle-of-attack aerodynamic performance of this airfoil in rain
[61,62], also referred to in Fig. 22. In addition, Ismail also reminded
120 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 84. Dimensionless tangential velocity component profile near the surface of the Wortmann FX67-K170 airfoil [126].
that rain and wing tip vortices are believed to cause premature
boundary-layer transition, which is thought to increase the mixing
and energize the boundary layer, making it less susceptible to
separation.
7. Summary and conclusions
The foregoing content is a systematic and comprehensive
review, correlation and analysis of experimental and computa-
tional measurements available in the public domain which
addresses the current state-of-the-art research on the effects of
rainfall on aircraft aerodynamics. The primary intent has been to
define the range of possible consequences which can occur at
natural flight conditions, especially at the most heavy rain condi-
tions which could be encountered. The full-scale flight test
technique has not been used herein to either promote or help
correlate existing test results regarding the scarce testing results
and the intrinsic limitations of this method such as risk and
inaccuracy acquisition of rainfall parameters due to the variability
of natural environment.
Generally, three parts have been made to identify current state-
of-the-art research achievements of rain effects on aircraft aero-
dynamics in this review. The first part discusses the aerodynamic
penalties caused by rain for various types of airfoils and wings
either by analytic estimation, wind-tunnel experiment or CFD
numerical simulation approaches. The second part presents the
characteristics of raindrop impact, splashback and formed water
film flow. Some numerical simulation results show consistent
phenomena with those observed by experiment. The last synthe-
sizes the current research achievements on the potential mechan-
isms of rainfall effects on aircraft aerodynamics. Some experiment
conclusions about the rain-influenced mechanics are also drawn
by CFD numerical simulation. The following summarizes some
important conclusions by the aforementioned three categories:
2) Raindrop impinging and surface water flow characteristics:
1) Rain-induced aerodynamic penalties:
● In general, rain can impart various extents of decrease in lift,
slope of lift curve and lift-to-drag-ratio and increase in drag
and slope of drag curve to aircraft, which is critically
hazardous to aircraft flight safety, especially at the stages
of take-off and land. At the same condition, laminar flow
airfoils usually suffer greater aerodynamic penalties than
that imparting the other types of airfoils like some sym-
metric and transport-type airfoils. However, it should be
borne in mind that not all types of airfoils will undergo
aerodynamic performance degradations by rainfall for all
rain conditions and flight parameters. For example, effects
of rain on the aerodynamic performance of a wing with the
same airfoil section sometimes may be absolutely different
for different intensities of rainfall, different angles of attack,
different Reynolds numbers, different aspect ratio, etc.
Furthermore, some airfoils may experience lift increase in
rain, though it is usually the opposite case for most airfoils.
Even for one airfoil, different configurations may sustain
completely disparate influences of rain at the equivalent
conditions, meaning that an aircraft may feel nothing when
cruising in rain but suddenly lose control when performing
take-off or land in a rain encounter.
● Rain can change the stall characteristics of wings, which
contains both beneficial and detrimental effects. Most air-
foils have premature stall by several degrees of stall angle
ahead in rain relative to the dry case, while some airfoils
have just approximately the same stall angle in rain as that
in dry condition. Furthermore, some airfoils even obtain
enhanced stall characteristics in rain.
● The characteristics of the air flow field around an aircraft are
usually changed in rain. The pressure difference between
the upper and lower surfaces of the airfoil/wing leading
edge decrease with rain, and thus resulting in the degrada-
tions of lift and lift-to-drag-ratio. Besides, rain can induce
premature boundary-layer transition and separation for an
airfoil, which also accounts for the aerodynamic penalties of
aircraft flying in rain environment.
● As raindrops impact an airfoil surface, some of them are
split into massive secondary smaller droplets and splashed
back, forming an ejecta fog near the leading edge and a
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 121

Fig. 85. Boundary layer velocity profiles at six chordwise positions on the upper surface of NACA 64-210 airfoil at angle of attack (α) of 101 [137].
122 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127

Fig. 86. Boundary layer velocity profiles at six chordwise positions on the upper surface of NACA 64-210 airfoil at angle of attack (α) of 131 [137].
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 123

Fig. 87. Chordwise skin friction coefficient (Cf) plot for a section of the 3-D NACA 64-210 wing [137]. (a) Upper surface and (b) Lower surface.

Fig. 88. Velocity magnitude and streamlines around the z¼ 0 m (symmetric axis) cross section of the 3-D NACA 0012 rectangular wing at angle of attack of 161 [136].
(a) No rain and (b) Rain (LWC¼39 g/m3).

Fig. 89. Velocity magnitude and streamlines around the z ¼1 m (symmetric axis) cross section of the 3-D NACA 0012 rectangular wing at angle of attack of 161 [136]. (a) No
rain and (b) Rain (LWC ¼ 39 g/m3).
124 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
Fig. 90. Velocity magnitude and streamlines around the z¼ 2 m (symmetric axis) cross section of the 3-D NACA 0012 rectangular wing at angle of attack of 161 [136]. (a) No
rain and (b) Rain (LWC = 39g/m3)
water bow wave surrounding the airfoil, while the remain-
der forms a thin and uneven water film upon the airfoil
surface. As subsequent raindrops impact the water film,
many craters are formed near the leading edge of the airfoil,
making the film more uneven.
● The characteristics of raindrop impingement change with
angle of attack and intensity of rainfall. Given a positive
angle of attack, more raindrops will be accumulated over
the upper surface and the maximum thickness of the
droplet breakup layer (i.e., the water bow wave) gradually
increases as the angle of attack is increased. The droplet
concentration around the airfoil also increases with increas-
ing rainfall rate.
● The characteristics of surface water flow depend on many
factors, such as the wettability of the airfoil surface, the type
and configuration of the airfoils (single-element or multi-
element), flow location on the airfoil surface, angle of attack
and intensity of rainfall R or LWC, etc. Generally, the more
nonwettable the airfoil surface is, the rougher, broader and
more irregular the water film will be. Different types and
configurations of airfoils may determine absolutely different
surface-water flow characteristics. Airfoils with flapped
configurations may have more irregular but thinner water
films and depend more on the above factors. At low angles
of attack, a smooth continuous film forms on the forward
section of the airfoil surface and the film breaks into many
rivulets afterwards. At high angles of attack with separation,
a large region of separated flow appears between the
breakdown of the continuous sheet and the trailing edge.
In addition, it can be easily imagined that the surface-water
flow characteristics are more evident and drastic as the
rainfall intensity increases.
3) Potential mechanisms of rain effects:
● The surface wettability greatly influences the rain-induced
aerodynamic performance degradations. Generally, the
smoother and the more wettable the airfoil surface is, the
less aerodynamic performance degradation of the airfoil will
occur. Thus, a smooth wettable surface may possibly alle-
viate the adverse effects of heavy rain, which is significant in
practice.
● Rain can cause premature boundary-layer transition at low
angles of attack to an airfoil, which is accountable for the
aerodynamic performance degradation for airfoils in rain
environment.
● The splashed-back droplets are accelerated by the raindrop
particle drag which acts as a momentum source or sink to
the airflow, resulting in an effective decrease in the
boundary-layer airflow velocity. Deceleration of the bound-
ary layer can bring losses in lift and lift-to-drag-ratio,
increase in drag, premature boundary-layer separation
and stall.
● The uneven water film formed on the surface can effectively
increase the surface roughness and change the geometric
configuration of an airfoil, thus it is also accountable for the
aerodynamic degradation of airfoils in rain.
Lastly, after reviewing many test results of the adverse rain
effects, a great number of which have been available for decades, it
is concluded that important lessons usually learned by one
generation with aspect to the various aerodynamic consequences
of heavy rainfall have not been well disseminated or accepted by
the later generations, resulting in many rain-induced aviation
accidents which seemingly could have been avoided if enough
attention had been paid to. The review presented here is intended
to attract people's eyes to the detrimental effects of rainfall on
aircraft aerodynamics since it is inevitable for an airplane to
encounter rainfall in flight, especially a short-duration shower at
the stage of take-off or approach. Some measures are necessary to
deal with problems of aircraft flight safety in rain, especially heavy
rain environment. In addition, continuing education is necessarily
conducted to broadly utilize the profound lessons in the past and
spread the knowledge of rainfall on aircraft aerodynamics in the
future.
References
[1] Gorney JL. Avoidance of hazardous weather in the terminal area. AIAA-88-
0679, AIAA 26th aerospace sciences meeting. Reno (NV); 1988.
[2] Newton D. Weather accident prevention using the tools that we have. AIAA-
89-0707, AIAA 27th aerospace sciences meeting. Reno (NV); 1989.
[3] Doviak RJ, Mazur V, Zrnic DS. Aviation weather surveillance systems:
advanced radar and surface sensors for flight safety and air traffic manage-
ment, vol. 8, let; 1999.
[4] Klehr JT. The simulation of hazardous flight conditions. AIAA-84-0278, AIAA
22nd aerospace sciences meeting. Reno (NV); 1984.
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
[5] Sullivan J. The effects of inclement weather on airline operation. AIAA-89-
0797, AIAA 27th aerospace sciences meeting. Reno (NV); 1989.
[6] Cornwall D. Impact of severe weather on aviation: pilot's viewpoint. AIAA-
89-0798, AIAA 27th aerospace sciences meeting. Reno (NV); 1989.
[7] David S. Aspects of severe weather on usaf and army aviation. AIAA-88-0678,
AIAA 26th aerospace sciences meeting. Reno (NV); 1989.
[8] Patnoe Michael W, Moravec Bradfor A. Rain and hail threat to aviation: a
probability analysis of extreme rain/hail concentrations aloft. SAE Trans
1996;105:1586–603.
[9] Vaughan W, Anderson B. Environmental effects consideration: a case study-
lessons learned. In: Proceedings of AIAA-92-5075, AIAA 4th international
aerospace planes conference. Orlando (FL); 1992.
[10] Mandel Eric. Severe weather: impact on aviation and FAA programs in
response. AIAA-89-0794, AIAA 27th aerospace sciences meeting. Reno (NV);
1989.
[11] Forbes G, Hosler C, Klemp J, Krider E, McGinley J, Ray P, et al. Weather
support for the space program. AIAA-89-0747, AIAA 27th aerospace sciences
meeting., Reno (NV); 1989.
[12] Nicholson JR, Jafferis W. Atmospheric sciences program at nasa Kennedy
space center. AIAA-88-0197, AIAA 26th aerospace sciences meeting. Reno
(NV); 1988.
[13] Mclean JC. Determining the effects of weather in aircraft accident investiga-
tions. AIAA-86-0323, AIAA 24th aerospace sciences meeting. Reno (NV);
1986.
[14] Elshamy M. Parametric analysis of atmospheric processes. AIAA-94-0479,
32nd aerospace sciences meeting and exhibit. Reno (NV); 1994.
[15] Elshamy M. Effect of atmospheric processes on launch decisions. J Spacecr
Rockets 1995;32(5):801–5.
[16] Ferguson D, Radke J. System for adverse weather landing. AIAA-93-3980,
AIAA aircraft design, systems and operations meeting. Monterey (CA); 1993.
[17] Townsend J. Low-altitude wind shear and its hazard to aviation. National
Academy Press. Library of Congress Catalog #83-63100; 1983.
[18] Rudich RD. Weather-involved US Air carrior accidents 1962–1984. A com-
pendiumand brief summary. AIAA-83-0327; 1983.
[19] Haines PA, Luers JK. Aerodynamic penalties of heavy rain on a landing
aircraft. NASA CR-156885; July 1982.
[20] Luers JK, Haines PA. Heavy rain influence on airplane accidents. J Aircr
1983;20(2):187–91.
[21] Dietenberger MA, Haines PA, Luers JK. Reconstruction of Pan Am New
Orleans accident. J Aircr 1985;22(8):719–28.
[22] 〈http://www.smartcockpit.com/download.php?path=docs/
&file=Hazards_Of_Flight_In_Heavy_Rain.pdf〉.
[23] Wan T, Pan SP. Aerodynamic efficiency study under the in-fluence of heavy
rain via two-phase flow approach. In: Proceedings of 27th international
congress of the aeronautical sciences; 2010.
[24] 〈http://news.xinhuanet.com/english/2008-06/01/content_8292980.htm〉.
[25] Rhode RV. Some effects of rainfall on flight of airplanes and on instrument
indications. NASA TN-803; 1941.
[26] Luers JK, Haines PA. The effect of heavy rain on wind shear attributed
accidents. AIAA-81-0390, 19th aerospace sciences meeting. St. Louis (MO);
1981.
[27] James K. Luers. Heavy rain effects on aircraft. AIAA paper 83-0206; 1983.
[28] Haines PA, Luers JK. Aerodynamic penalties of heavy rain on landing
airplanes. J Aircr 1983;20(2):111–9.
[29] Luers JK. Rain effects on the aerodynamics of general aviation aircraft. In:
Proceeedings of AIAA conference on design for the eightites. Dayton (OH);
1981.
[30] Haines PA, Luers JK. Heavy rain penalties for flight simulators. AIAA-82-0213,
AIAA 20th aerospace sciences meeting. Orlando (FL); 1982.
[31] Flight in an adverse environment. AGARD lecture series 1994; 197.
[32] Reinmann JJ. Effects of adverse weather on aerodynamics. In: AGARD
conference proceedings 1991; 496.
[33] Luers JK. Rain influences on a wind turbine—a comparison of theory to
measurements. AIAA-85-0256, AIAA 23rd aerospace sciences meeting. Reno
(NV); 1985.
[34] Krozel Jimmy, McNichols William M, Prete Joseph. Causality analysis for
aviation weather hazards. AIAA 2008-8914, the 26th congress of interna-
tional council of the aeronautical sciences (ICAS). Anchorage (AK); 2008.
[35] Moore R, et al. Preliminary study of rain effects on radar scattering from
water surfaces. Oceanic Eng, IEEE J Ocean Eng 1979;4(1):31–2.
[36] Gabrey SW, Dolbeer RA. Rainfall effects on bird-aircraft collisions at two
united states airports,. Wildl Soc Bull 1996:272–5.
[37] Huffman PJ, Haines PA. Visibility in heavy precipitation and its use in
diagnosing high rainfall rates. AIAA-84-0541, AIAA 22nd aerospace sciences
meeting. Reno (NV); 1984.
[38] Trammell A. Three unconventional airborne radar clues to severe convective
storms. AIAA-89-0705, AIAA 27th aerospace sciences meeting. Reno (NV);
1989.
[39] Chan HL, Fung AK. A theory of sea scatter at large incident angles. J Geophys
Res 1977;82(24):3439–44.
[40] Methven TJ, Fairhead B. A correlation between rain erosion of perspex
specimens in flight and on a ground rig. HM stationery office; 1960.
[41] Schmitt Jr GF. Flight test-whirling arm correlation of rain erosion resistance
of materials. AIR force materials lab Wright–Patterson AFB OH; 1968.
125
[42] Busch H, Hoff G, Langbein G, et al. Rain erosion properties of materials [and
discussion]. Philos Trans R Soc Lond Ser A, Math Phys Sci 1966;260
(1110):168–81.
[43] Fyall AA. Practical aspects of rain erosion of aircraft and missiles. Philos Trans
R Soc Lond Ser A Math Phys Sci 1966;260(1110):161–7.
[44] Hurley CJ, Schmitt Jr GF. Development and calibration of a Mach 1.2 rain
erosion test apparatus. Dayton Univrsity of OH Research Institution; 1970.
[45] Déom A, Gouyon R, Berne C. Rain erosion resistance characterizations: link
between on-ground experiments and in-flight specifications. Wear
2005;258(1):545–51.
[46] Schmitt GF. Influence of rainfall intensity on erosion of materials at super-
sonic velocities. J Aircr 1975;12(10):761–2.
[47] Clayton RM, Cho YI, Shakkottai P, et al. Rain simulation studies for high-
intensity acoustic nose cavities. J Aircr 1988;25(3):281–4.
[48] Spillard CL, Gremont B, Grace D, et al. The performance of high altitude
platform networks in rainy conditions. AIAA 2004-3220, 22nd AIAA inter-
national communications satellite systems conference and exhibit. Monterey
(CA); 2004.
[49] Rodeman R, Longcope DB. Rain-induced vibration. AIAA J 1983;21
(3):459–65.
[50] Murthy SNB. Effect of heavy rain on aviation engines. AIAA-89-0799, AIAA
27th aerospace sciences meeting. Reno (NV); 1989.
[51] Murtht SNB. Effect of atmospheric water ingestion on the performance and
operability of flight gas turbines. In: Proceedings of 32nd AIAA, ASME, SAE
and ASEE joint propulsion conference and exhibit. Lake Buena Vista (FL);
1996.
[52] Rajen Gaurav. Comment on aerodynamic penalties of heavy rain on landing
airplanes. J Aircr 1984;21(1):92.
[53] Devine Kirk. Inclement weather induced aircraft engine power loss. In:
Proceedings of AIAA-90-2169, AIAA/SAE/ASME/ASEE 26th joint propulsion
conference Orlando (FL); 1990.
[54] Russell RE, Victor IW. Evaluation and correction of the adverse effects of (i)
inlet turbulence and (ii) rain ingestion on high bypass engines. AIAA-84-
2486, AIAA/AHS/ASEE aircraft design systems and operations meeting. San
Diego (CA); 1984.
[55] Venkataramani KS, McVey LJ, Holm RG, et al. Inclement weather considera-
tions for aircraft engines. AIAA 2007-695, 45th AIAA aerospace sciences
meeting and exhibit. Reno (NV); 2007.
[56] Murthy SNB, Ehresman CM. Effects of water ingestion into jet engine. AIAA-
84-0542, AIAA 22nd aerospace sciences meeting. Reno (NV); 1984.
[57] Tsuchiya T, Murthy SNB. Water ingestion into jet engine axial compressors.
AIAA-82-0196, AIAA 20th aerospace sciences meeting. Orlando (FL); 1982.
[58] Valentine JR, Decker RA. A Lagrangian–Eulerian scheme for flow around an
airfoil in rain. Int J Multiph Flow 1995;21(4):639–48.
[59] Ruimin Zhang, Yihua Cao. Study of aerodynamic characteristics of an airfoil
in rain. J Aerosp Power 2010;25(9):2064–9.
[60] Bezos GM, Dunham RE Jr, Garl L Gentry Jr, Edward Melson W Jr Wind tunnel
aerodynamic characteristics of a transport-type airfoil in a simulated heavy
rain environment. NASA TP-3184; 1992.
[61] Hansman R John, Craig Anthony P. Low Reynolds number tests of NACA 64-
210, NACA 0012, and Wortmann FX67-K170 airfoils in rain. J Aircr 1987;24
(8):559–66.
[62] Hansman R John, Craig Anthony P. Comparative low Reynolds number tests
of NACA 64-210, NACA 0012, and Wortmann FX67-K170 airfoils in heavy
rain. AIAA-87-0259, AIAA 25th aerospace sciences meeting. Reno (NV); 1987.
[63] Marshall JS, Palmer WMcK. The distribution of raindrops with size. J
Meteorol 1948;5(4):165–6.
[64] Laws JO, Parsons DA. The relation of raindrop-size to intensity. Trans Am
Geophys Union 1943;24:452–60.
[65] Joss J, Waldvogel A. Raindrop size distribution and sampling size errors. J
Atmos Sci 1969;26(3):566–9.
[66] Ulbrich CW. Natural variations in the analytical form of the raindrop size
distribution. J Clim Appl Meteorol 1983;22(10):1764–75.
[67] Best AC. The size distribution of raindrops. Q J R Meteorol Soc 1950;76
(327):16–36.
[68] Smith James A, Hui Eric, Steiner Matthias, et al. Variability of rainfall rate
and raindrop size distribution in heavy rain. Water Resour Res 2009;45(4).
[69] Hess CF, Li F. Optical technique to characterize heavy rain. AIAA-86-0292,
AIAA 24th aerospace sciences meeting. Reno (NV); 1986.
[70] Kissel Glen J. Rain and hail extremes at altitude. J Aircr 1980;17(7):464–8.
[71] Uijlenhoet R, Stricker JNM. A consistent rainfall parameterization based on
the exponential raindrop size distribution. J Hydrol 1999;218(2):101–27.
[72] Markowitz Allan H. Raindrop size distribution expressions. J Appl Meteorol
1976;15(9):1029–31.
[73] Riordan P. Weather extremes around the world. Earth Sciences Laboratory.
TR-70-45-ES; January 1970.
[74] Riordan P. Weather extremes around the world. No. ETL-TR-74-5. Army
engineer topographic labs. Fort Belvoir (VA); 1974.
[75] Roys GP, Kessler E. Measurements by aircraft of condensed water in great
plains thunderstorms. National Severe Storms Laboratory Publications. TN-
49-NSSP-19; 1966.
[76] Jones DMA, Sims AL. Climatology of instantaneous rainfall rates. J Appl
Meteorol 1978;17:1135–40.
[77] Jones Douglas M, Sims Arthur L. Climatology of instantaneous precipitation
rates. Illinois State Water Survey Div Urbana 1971.
126 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127
[78] Melson WE Jr. Heavy rain characteristics and ground measurement compar-
isons. Presented at the WHO IAHS 8th international workshop on precipita-
tion measurement; 1989.
[79] Becker A, Grünewald U. Flood risk in central Europe. Science 2003;300:1099.
[80] Webster PJ, Toma VE, Kim HM. Were the 2010 Pakistan floods predictable?
Geophys Res Lett 2011;38:L04806.
[81] World meteorological organization provisional statement on the status of
the global climate (WMO; 2011). Available at: 〈http://www.wmo.int/pages/〉.
[82] Coumou D, Rahmstorf S. A decade of weather extremes. Nat Clim Change
2012;2(7):491–6.
[83] Easterling David R, et al. Climate extremes: observations, modeling, and
impacts,. Science 2000;289.5487:2068–74.
[84] Henry R, Pasamehmetoglu K, Eno B, Anderson L. A Von Karman integral
approach to a two phase boundary layer problem. AIAA-87-0256, AIAA 25th
aerospace sciences meeting. Reno (NV); 1987.
[85] Bilanin AJ, Quackenbush TR, Feo A. Feasibility of predicting performance
degradation of airfoils in heavy rain. NASA CR 181842; 1989.
[86] Hsu YK. An analytic study of nonsteady two-phase laminar boundary layer
around an airfoil. NASA conference paper; 1989.
[87] Wan T, Wang CM. A Study of Aircraft performance parameter under adverse
weather conditions. AIAA 2006-234, 44the AIAA aerospace sciences meeting
and exhibit. Reno (NV); 2006.
[88] Hermanspann Fred. Rain effects on natural laminar flow airfoils. AIAA-96-
0899, AIAA 34th aerospace sciences meeting and exhibit. Reno (NV); 1996.
[89] Adams KJ. The air force flight test center palletized airborne water spray
system. AIAA-83-0030, AIAA 21st aerospace sciences meeting, Reno (NV);
1983.
[90] Adams KJ. The air force flight test center artificial icing and rain testing
capability. In: Proceedings of AIAA-83-2688, AIAA/AHS/IES/SETP/SFTE/DGLR
2nd flight testing conference. Las Vegas (NV); 1983.
[91] Russel Ashenden, Marwitz John D. A comparison of the air force water spray
tanker artificial drizzle cloud distributions to the natural environment. AIAA-
96-0632, AIAA 34th aerospace sciences meeting and exhibit. Reno (NV);
1996.
[92] Feo A. Rotating arms applied to studies of single angular drop impacts. AIAA-
87-0257, AIAA 25th aerospace sciences meeting. Reno (NV); 1987.
[93] Beitel GR, Matthews RK. A New rain/ice test capability to simulate flight. In:
Proceedings of AIAA-86-9824, IAA/AHS/CASI/DGLR/IES/ISA/ITEA/SETP/SFTE
3rd flight testing conference. Las Vegas (NV); 1986.
[94] Bower RE. Current wind tunnel capability and planned improvements at the
NASA Langley research center. AIAA paper 86-0727; 1986.
[95] Dunham RE. Potential influences of heavy rain on general aviation airplane
performance. AIAA-86-2606. In: Proceddings of AIAA general aviation
technology conference. Anaheim (CA); 1986.
[96] Dunham RE. The potential influence of rain on airfoil performance. Von
Karman institute for fluid dynamics lecture series 1987-03; February 16–20,
1987.
[97] Booker JD., Nash MJB. Facilities at R.A. E. for the simulation of flight through
rain. RAE-TR-67245. British Royal Aircraft Establ.; Oct. 1967.
[98] Aaron KM, Hernan MA, Parikh P, et al. Simulation and analysis of natural rain
in a wind tunnel via digital image processing techniques. AIAA-86-0291,
AIAA 24th aerospace sciences meeting. Reno (NV); 1986.
[99] Hernan MA, Gharib M, Parikh P, Sarohia V. Simulation and analysis of natural
rain in a wind tunnel via digital image processing techniques-interim report.
JPL D-2231 (contract NAS7-918, task order RE 65/amendment 459). Califor-
nia Institute of Technology; October 31, 1984.
[100] Aaron KM, Hernan MA, Parikh P, Sarohia V. Simulation and analysis of
natural rain in a wind tunnel via digital image processing techniques. AIAA
1986:86–0291.
[101] Taylor JT, Moore CT, Campbell BA, et al. The development of a facility for full-
scale testing of airfoil performance in simulated rain. AIAA-88-0055, AIAA
26th aerospace sciences meeting. Reno (NV); 1988.
[102] Bezos, GM, Campbell, BA, Melson WE. The development of a capability for
aerodynamic testing of large-scale wing sections in a simulated natural rain
environment. AIAA-89-0762, AIAA 27th aerospace sciences meeting. Reno
(NV); 1989.
[103] Bezos GM, Campbell BA. Development of a large-scale outdoor, ground-based
test capability for evaluating the effect of rain on airfoil lift. NASA TM 4420;
1993.
[104] Durst F, Miljevic D, Schonung B. Eulerian and Lagrangian predictions of
particulate two-phase flows: a numerical study. Appl Math Model
1984;8:101–15.
[105] Decker R, Shafer CF. Mixing and demixing processes in multiphase flows
with application to propulsion systems. NASA CP-3006; 1989.
[106] Valentine JR, Decker RA. Tracking of raindrops in flow over an airfoil. AIAA-
93-0168, 31st aerospace sciences meeting and exhibit. Reno (NV); 1993.
[107] Valentine JR, Decker RA. Tracking of raindrops in flow over an airfoil. J Aircr
1995;32(1):100–5.
[108] Dukowicz JK. A particle-fluid numerical model for liquid sprays,. J Comput
Phys 1980;35(2):229–53.
[109] Crowe CT, Stock DE, Sharma MP. The particle-source-in-cell (PSI-cell) model
for gas-droplet flows. J Fluids Eng 1977;99:325–32.
[110] Bilanin AJ. Scaling laws for testing of high lift airfoils under heavy rainfall.
AIAA-85-0257, AIAA 23rd aerospace sciences meeting. Reno (NV); 1985.
[111] Bilanin AJ. Scaling laws for testing airfoils under heavy rainfall. J Aircr
1987;24(1):31–7.
[112] Campbell BA, Bezos GM. Steady-state and transitional aerodynamic charac-
teristics of a wing in simulated heavy rain. NASA TP 1989;2932.
[113] Marchman JF, Robertson Edward A, Emsley Howard T. Rain effects at low
reynolds number. AIAA-87-0258, AIAA 25th aerospace sciences meeting.
Reno (NV); 1987.
[114] Marchman JF, Robertson Edward A, Emsley Howard T. Rain effects at low
Reynolds number. J Aircr 1987;24(9):638–45.
[115] Lynch Frank T, Khodadoust Abdollah. Effects of ice accretion on aircraft
aerodynamics. Prog Aerosp Sci 2001;37:669–767.
[116] Hasting EC, Manuel GS. Measurements of water film characteristics on airfoil
surface from wind-tunnel tests with simulated heavy rain. AIAA-85-0259,
AIAA 23rd aerospace sciences meeting. Reno (NV); 1985.
[117] Tung Wan Szu-Peng Pan. Aerodynamic efficiency study under the influence
of heavy rain via two-phase flow approach. In: Proceedings of 27th interna-
tional congress of the aeronautical sciences; 2010.
[118] Tung Wan, Hei Yang. Aerodynamic performance investigation of a modern
blended-wing-body aircraft under the influence of heavy rain condition. In:
Proceedings of 27th international congress of the aeronautical sciences; 2010.
[119] Wan Tung, Lin Jing-Xuan, Kuan Hsiang-Chun. Aerodynamic analysis of
helicopter rotor blades in heavy rain condition. AIAA 2013-0653, 51st AIAA
aerospace sciences meeting. Grapevine (TX); 2013.
[120] Wan Tung, Chou Chi-Ju. Reinvestigation of high lift airfoil under the influence
of heavy rain effects. AIAA 2012-1202, 50th AIAA aerospace sciences meet-
ing. Nashville (TN); 2012.
[121] Hansman R John, Barsotti Martitia F. Surface wetting effects on a laminar
flow airfoil in simulated heavy rain. J Aircr 1985;22(12).
[122] Hansman R John, Barsotti Martitia F. The aerodynamic effect of surface
wetting characteristics on a laminar flow airfoil in simulated heavy rain.
AIAA-85-0260, AIAA 23rd aerospace sciences meeting. Reno (NV); 1985.
[123] Muller, TJ., Low Reynolds number vehicles. AGARD AG-288; 1985.
[124] Dunham RE, Bezos GM, Gentry GL, Melson E. Two-dimensional wind tunnel
tests of a transport-type airfoil in a water spray. AIAA paper 85-0258, 23rd
aerospace sciences meeting. Reno (NV); 1985.
[125] Bezos GM, Dunham RE, Gentry GL, Melson WE. Wind tunnel test results
of heavy rain effects on airfoil performance. In: Proceedings of the AIAA 25th
aerospace sciences meeting. AIAA-87-0260, Reno (NV) 1987.
[126] Hsu YK. An analytic study of a two-phase laminar airfoil in simulated heavy
rain. NASA conference paper; 1993.
[127] Calaress W, Hankey WL. Numerical analysis of rain effects on an airfoil. AIAA-
84-0539, AIAA 22nd aerospace sciences meeting. Reno (NV); 1984.
[128] Hartwich PM, Hsu CH. High resolution upwind schemes for the three-
dimensional Navier–Stokes equations. AIAA J 1988;26:1321–8.
[129] Hartwich PM, Hall RM. Navier–Stokes solutions for vertical flows over a
tangent-ogive cylinder,. AIAA J 1990;28:1171–9.
[130] Wan T, Wu SW. Aerodynamic analysis under influence of heavy rain. In:
Proceedings of 24th international congress of the aeronautical sciences;
2004.
[131] Wan T, Song BC. Aerodynamic performance study of a modern blended-
wing-body aircraft under severe weather situations. AIAA 2012-1037, 50th
AIAA aerospace science meeting including the new horizons forum and
aerospace exposition. Nashville (TN); 2012.
[132] Wan T, Lin JX, Kuan HC. Aerodynamic analysis of helicopter rotor blades in
heavy rain condition. AIAA 2013-0653, 51st AIAA aerospace science meeting
including the new horizons forum and aerospace exposition. Grapevine (TX);
2013.
[133] Yeom GS, Chang KS, Baek SW. Numerical prediction of airfoil characteristics
in a transonic droplet-laden air flow. Int J Heat Mass Transf 2012;55
(1):453–69.
[134] Yeom GS, Chang KS. Dissipation of shock wave in a gas-droplet mixture by
droplet fragmentation. Int J Heat Mass Transf 2012;55(4):941–57.
[135] Ismail M, Cao YH, Bakar A, Wu ZL. Aerodynamic efficiency study of 2D airfoils
and 3d rectangular wing in heavy rain via two-phase flow approach. Proc Inst
Mech Eng Part G: J Aerosp Eng 2014;228(7):1141–55.
[136] Ismail M, Cao YH, Wu ZL, Sohail MA. Numerical study of aerodynamic
efficiency of a wing in simulated rain environment. J Aircr 2014 (published
on line).
[137] Wu ZL, Cao YH. Aerodynamic study of aerofoil and wing in simulated rain
environment via a two-way coupled eulerian-lagrangian approach. Aeronaut
J 2014;118(1204):643–68.
[138] Stanton DW, Rutland CJ. Multi-dimensional modeling of thin liquid films and
spray-wall interactions resulting from impinging sprays. Int J Heat Mass
Transf 1998;44:3037–54.
[139] Hasting EC, Manuel GS. Scale-model tests of airfoil in simulated heavy rain. J
Aircr 1985;22(6):536–40.
[140] Hastings EC, Weinstein LM. Preliminary indications of water distribution and
thickness on an airfoil in a water spray. NASA TM 85796; 1984.
[141] Thompson BE, Jang J, Dion JL. Wing performance in moderate rain. JAircr
1995;32(5).
[142] Towell GD, Rothfeld LB. Hydrodynamics of rivulet flow. AIChE J 1966;12
(5):972–80.
[143] Nakagawa T, Scott JC. Stream meanders on a smooth hydrophobic surface. J
Fluid Mech 1984;149(1):89–99.
[144] Schmuki P, Laso M. On the stability of rivulet flow. J Fluid Mech 1990;215
(1):125–43.
[145] Thompson BE, Jang Juneho. Aerodynamic efficiency of wings in rain. J Aircr
1996;33(6):1047–53.
 Y. Cao et al. / Progress in Aerospace Sciences 71 (2014) 85–127 127

[146] Thompson BE, Marrochello MR. Rivulet formation in surface-water flow on [151] Riebe JM. A correlation of two-dimensional data on lift coefficient available
an airfoil in rain. AIAA J 1999;37(1):45–50. with blowing-, suction-, slotted-, and plain-flap high-lift devices. NACA RM
[147] Drela M. Two-dimensional transonic aerodynamic design and analysis using L55D29a; 1955.
the euler equations. (Ph.D. thesis). Cambridge (MA): Department of Aero- [152] Wu ZL, Cao YH, Ismail M. Numerical simulation of airfoil aerodynamic
nautics and Astronautics, Massachusetts Institute of Technology; 1985. penalties and mechanisms in heavy rain. Int J Aerosp Eng 2013;2013.
[148] Giles MB. Newton solution of steady two-dimensional transonic flow. (Ph.D. [153] Yip Long P.Wind-tunnel investigation of a full-scale canard-configured
thesis). Cambridge (MA): Department of Aeronautics and Astronautics, general aviation airplane. NASA TP 2382; 1985.
Massachusetts Institute of Technology; 1985.
[149] Sragg MB, Gregorek CM, Shaw RI. An analytical approach to airfoil icing. AIAA
paper 81-0403; 1981.
[150] Riebe JM, McLeod RC. Low-speed wind-tunnel test of a thin 60” delta wing with
double-slotted, single-slotted, plain, and split flaps. NACA RM L52j29; 1953.