AIRCRAFT ENVIRONMENTAL CONTROL

Environmental control system in aircraft ...................................................................................................... 1

Introduction ............................................................................................................................................... 1
Cabin air conditioning ............................................................................................................................... 3
Pressurization and oxygen control ........................................................................................................ 4
Air source. Bleed air circuit .................................................................................................................. 7
Humidity control ................................................................................................................................... 8
Thermal loads ...................................................................................................................................... 10
Mass and energy balances ................................................................................................................... 12
Possible solutions: HVAC systems ..................................................................................................... 12
The air cycle machine (ACM) ............................................................................................................ 13
ACM working cases ............................................................................................................................ 15
ACM size and efficiency (energy cost)............................................................................................... 16
Future ECS .......................................................................................................................................... 16
Aircraft environmental protection system................................................................................................... 17
Against ice formation .............................................................................................................................. 17
Icing problems in aviation................................................................................................................... 17
Anti-icing (prevention) and de-icing (fight) systems:......................................................................... 19
Icing tests ............................................................................................................................................ 19
Ice accretion modelling ....................................................................................................................... 20
Against rain ............................................................................................................................................. 24
Against dust............................................................................................................................................. 25
Against high temperature ........................................................................................................................ 25
Supersonic aircraft ...................................................................................................................................... 25
References ................................................................................................................................................... 26

ENVIRONMENTAL CONTROL SYSTEM IN AIRCRAFT

Introduction
In general, the environmental control system (ECS) refers to equipment in charge of maintaining a
comfortable close environment for a given payload (goods, living matter, and people), i.e. keeping
temperature, pressure, and composition, within acceptable limits, usually by circulating a fluid for thermal
control and for life-support, if required (the term ECLSS, for environmental control and life support
system, is also used to make the latter explicit). The ECS for vehicles in hostile environments is most
demanding: submarines, aircraft and spacecraft. The ECS usually focuses on the inside part of the
vehicle, whereas the environmental control of the outer side is usually named environmental protection
system (EPS).
Aerospace engineering (aeronautical and space) is high-tech transport engineering, involving vehicles,
infrastructures and payloads. Aviation refers to activities involving man-made flying devices (manned
and unmanned aircraft), and the people, organizations, and regulatory bodies involved in their use.
Vehicles (aircraft and spacecraft; aircraft=aerostats (balloons and airships) and aerodynes
(airplanes and rotorcraft), manned or unmanned).
o Structure (frame): fuselage, wings and appendices.

o Propulsion: engines and propellers.

o Other systems
Energy systems for propulsion and control: fuel, mechanical, hydraulic, pneumatic,
and electrical systems.
Flight control systems: sensors and actuators, electronic systems (avionics).
Communications: air traffic control (ATC), radio-navigation, intercoms.
Start and stop systems (for engines and other systems): undercarriage (landing
gear), APU, ground starter, ram air turbine.
Environmental control system (ECS, internal)
Cabin air conditioning: pressure, temperature, ventilation, humidity (e.g.
windows defogging), and fire protection.
Water and sanitation. Flexible hoses are used, and cold water pipes must
incorporate thermal sensors and electrical strip heaters to guarantee Tw>0 C
even with Text=40 C, in spite of the hose being resistant to water freezing.
Hot water is electrically heated to Tmax=50 C.
Food, and solid waste.
Others: fuel tank inertization, cabin furniture ergonomics, noise (70 dB
typical in a cruising airliner), lighting, entertainment
Environmental protection system (EPS, external)
Against high temperatures (rarely against cryogenic temperatures).
Against high winds or turbulences.
Against water and ice.
Against radiations and electrical shock.
Others: against biological attack (from micro-organisms to beasts) either
through openings or by impact in flight.
Infrastructures: base, tracks, navigation aids, telecommunications, operations and payload
management.

We here focus on aircraft thermal systems aside of propulsion, i.e. ECS and EPS. Engine thermal
problems are basically different to ECS because of the very hot working parts involved (e.g. turbine blade
refrigeration to keep T<1700 K). There are other environmental impacts on flight that cannot be fight out,
like jetlag (discomfort by distortion of circadian rhythm). Others, like mobility restriction, may be
alleviated within large planes.
In addition to providing thrust, the engines also have to deliver electrical, pneumatic and hydraulic
secondary power to the systems in the aircraft. Among these systems, the ECS is the main power
consumer (75 % of non-propulsive power on cruise, which is 1 % of propulsion), and it generates the
most important extra consumption of fuel. With engines off on ground, the auxiliary power unit (APU, a
smaller turboshaft engine of some 1 MW) provides electrical power, and bleed air for air conditioning and
the start of main-engines.
The pneumatic system of an aircraft includes:
Bled air at 250 kPa and 450 K for:
o The internal ECS (cabin air).
o The external EPS (anti-icing, anti-rain windscreen air-curtain).
o The pressurization of some hydraulic systems (non-power, like in lavatories).
o Cross-engine coupling for start-up.
Additionally, it may include a pneumatic power system, at some 2 MPa (obtained on an additional
compressor), but this is only used by some European manufacturers (Americans use hydraulic or
electric systems).

Cabin air conditioning

Cabin air conditioning must provide comfort conditions (i.e. some 222 C, 90..100 kPa, and 50..70 %
RH) within a closed container (the cabin), under all foreseeable circumstances (60..+50 C, 10..100 kPa,
0..100 % RH, ozone, etc.), i.e. it must provide ventilation, pressurization, heating, cooling,
humidification, deshumidification (demisting), and disinfection. One may split air conditioning factors in:
physical, chemical, and biological. Besides, cabin air monitoring provides the basic smoke detection
means for fire warning.
Air conditioning is the second power-consuming system, after propulsion; e.g. for an A320, propulsion
takes Qprop m f LHV 40 MW and ECS QECS m f LHV 200 kW ; all the rest (instruments, lighting,
entertainment, kitchen) account only for a few tens of kW.
Conditioned air enters the cabin from distribution ducts by wall-floor and ceiling grilles and directional
outlets above the seats, and goes out through other grilles and collecting ducts (Fig. 1). About half of this
exiting air is exhausted from the airplane through an outflow valve in the underside, and the other half is
drawn by fans through special filters (for trapping microscopic particles, bacteria and viruses) and then
recirculated On some flight decks there is no recirculation, to have more margin for avionics cooling.

Fig. 1. Air flow inside an aircraft cabin.

Ventilation governs air supply. Air quality (as well as blood oxygenation) is primarily measured by CO 2
concentration (instead of oxygen availability), according to ANSI-ASHRAE Standard 62-1989:
Ventilation for Acceptable Indoor Air Quality, Appendix D:
Vair,STP

VCO2 ,gen
xCO2 ,cabin xCO2 ,outside

4% 0.5 L/(breath) 1/ 3 breath/s

1700 380 10

L/s
pax

Normal breathing. An adult breathe some 0.5 L of air in normal conditions 20 times per minute (up to 2 L
in deep gasps, and the rate varies from 12 breath/min at rest to 120 breath/min on panting). Intake
composition (fresh air) is 77 % N2 + 21 % O2 + 1 % H2O + 1 % Ar + 0.04 % CO2, and exhalation is 74 %
N2 +17 % O2 +4 % H2O +1 % Ar +4 % CO2. Maximum environmental CO2 for comfort: 0.17 %vol (1700
ppm). Design condition for any habitable space is to guarantee an air flow of 10 L/(scap), to have enough
oxygen (0.15 L(scap) would be enough: 0.5 (L/breathe)(1/3 breathe/s)), but mainly to sweep odours,
microorganisms, and heat release.

A person metabolism processes a minimum of 0.110-3 kg/s of air in respiration, consuming a minimum
of 510-6 kg/s de O2 and generating a minimum of 710-6 kg/s de CO2 and a minimum of 310-6 kg/s de
H2O. Skin transpiration (dry) and perspiration (sweat) contributes in an even greater amount of H 2O
vapour release.
Disinfection may range from applying an aerosol repellent to insects, to inertization with CO2 (fumigation
is banned) to kill rodents and reptiles that enter with the payload (in the pallets) or doors. An A340
requires 2500 kg de CO2 each time. Rodents die in half an hour, but reptiles may take half a day, and
cockroach larva even longer).
In days gone by, it was not unusual for people to smoke in any public place and a commercial airplane
was no exception. At that time, the ECS was additionally required to get rid of tobacco smoke present in
the cabin.
Exercise 1. How long does it takes for the ECS to change all the air in the cabin?
Sol.: In a full loaded plane, each passenger may be allocated some, let say, 1 m 3 of air, so that, at 5
L/(spax) of minimum fresh-air ventilation rate, it takes 1000/5=200 s to renovate the air. A further check
is to divide actual cabin-air volume by the air-supply for a given aircraft; e.g. for B747-400, Vair=886 m3,
416..524 pax, and 5 L/(spax), the typical renewal period is 886/(5000.005)=350 s (i.e. a few minutes).

Pressurization and oxygen control

Typical ventilation. Before cabin pressurization in the 1950s, ventilation was by infiltration as for
buildings. Pilots used oxygen mask since WWWI reconnaissance flights (pressure suits were
developed in the 1930s). Early jet liners (in the 1940s) pressurized the cabin with 10 L(spax) of
fresh air, but modern jets (since 1970) only supply and renew 5 L/(spax), forcing another 5
L/(spax) of cabin air recirculation (otherwise, the primary jet engine would deteriorate too much in
large bypass turbofans; e.g. each of the two A320 engines uses 80 kg/s primary air, and cabin air is
0.006200=1.2 kg/s minimum, up to 2 kg/s usual, 1 kg/s from each engine). According to the
European Aviation Safety Agency (EASA), each passenger and crew compartment must be
ventilated with enough fresh air (but not less than 0.3 m3/min STP = 6 g/(spax)), with CO<50 ppm,
CO2<0.5 % (it was 3 % up to 1997), O3<0.25 ppm, and particle filter for >10 nm (for virus; tobacco
particles are much larger, some 10 m). Cabin air is renewed every 2 or 3 min. The cockpit has a
larger air supply to keep a little overpressure against the main cabin (to avoid gases in). Air speed
flow near people must be in the range 0.05..0.3 m/s.
Oxygen concentration needs. Both, pressure and oxygen-fraction are important because what
matters for the flow of matter is chemical potential, depending on the product xO2p=pO2 (oxygen
partial pressure). Safe range for prolonged exposure is pO2=18..40 kPa at the intake. As atmospheric
pressure decreases with altitude, either cabin pressure or oxygen fraction should be increased to
keep pO2 within range. Although we just focus on oxygen needs, breathable air quality imposes
many other constrains on air composition: <0.5 % CO2 mole fraction (i.e. <5000 ppm; up to 1997
the limit was 3 %), <50 ppm CO, <0.25 ppm ozone, limits in volatile organic compounds
(lubricating oils and hydraulic fluids can enter the aircraft cabin via the ECS system), limits in
particle concentration, etc. Air composition control is also known as air revitalization.
Emergency oxygen can be supplied from:
Gas bottled at 15 MPa (heavy, and might be dangerous).
Liquid cryogens at cabin pressure (continuous loss, and might be dangerous).
Solid candles of lithium perchlorate (LiClO4=LiCl+2O2) that releases O2 when ignited
(locally heating above 400 C) by pulling a pin on the cartridge; in an airliner, when oxygen
masks drop down, you have to give a pull on it to actuate the igniter pin. Lithium perchlorate
is most used in aerospace applications as a last resource since it gives the highest quantity of
oxygen for its weight and for its size, in spite of its high price (cheaper sodium perchlorate,

NaClO4, and even sodium chlorate, NaClO3, are used in mines and submarines). A typical 2.2
kg cartridge used in spacecraft supplies enough oxygen for one person per day. Some metal
peroxide is added to the perchlorate (about 5 % by mass) to scavenge possible chlorine gas
formed (Cl2+Li2O2=2LiCl+O2). Oxygen candles have a long shell life, but they cannot be
stopped after ignition, and the heat release must be properly evacuated.
Gas separation from ambient air in a molecular sieve (the lightest and most modern way).
Pressurisation needs. Ambient air at 10 km altitude is at 26 kPa (Table 1) and 50 C (too thin for
oxygenation, and too cold for thermal comfort). Airliners fly up to 12 km altitude (above that, air
density falls faster), where p=19.4 kPa pO2=4 kPa 50 % people would die in 5 minutes.
Aircraft must keep pO2>16 kPa (ICAO), what implies pair>75 kPa zcabin<2.5 km (EASA CS-25
requires zcabin2440 m); the tendency is to decrease zcabin<2 km. The larger the cabin overpressure,
the heavier the structure must be; that is another reason why airliner cannot fly over 12 km; e.g.
when flying at 12 km with 20 kPa outside pressure and 80 kPa cabin pressure, a typical 0.50.5 m 2
aircraft skin panel must withstand a force, when the 50 % safety factor is included (fs=1.5),
F=Apfs=0.50.5(80-20)1031.5=22.5 kN. Although pressure may have no influence on inert
payloads, cargo compartments are usually pressurised (although perhaps at a smaller overpressure,
e.g. at 50 kPa, 5000 m boot-altitude). Combat aircraft keep a lower cabin overpressure: 35 kPa
instead of 50 kPa. Only the landing-gear boot and some technical-system boots nearby or at the tail
are un-pressurised. The pressurised cabin is not airtight (there are too many through-wall openings),
but uncontrolled air loss is small compared to 5 L/(s/pax) renewal. Cabin pressurization cycling may
cause structural fatigue; many early pressurised aircraft, like piston-engine Avro Tudor in the 1940s,
and Comet in the 1950s (the first commercial jet), cracked in flight. Controlling the rate of change
in pressure to avoid ear pain is particularly important during ascent and descent.
Pressurisation control. Cabin air is just compressed outside air. In the conventional air-cycle airconditioning system, the whole aircraft is pressurised by bleed air (bled at 250 kPa from engine
compressor, upstream of combustion chambers) supplied to the conditioning packs at some 450 K
(180 C) with a pre-cooler. Cabin pressure is controlled by outflow valves (controlled by redundant
electric motors) to keep cabin pressure above 75 kPa (equivalent to 2400 m altitude), and never
below 5 kPa underpressure or 500 kPa overpressure to outside air. Pressure control is based on
outflow valves (Fig. 2) and safety valves:
o Safety overpressure is limited by a positive pressure-relief valve to ppintpext=65 kPa for
structural reasons.
o Overpressure during normal operation is limited to p<55 kPa by the main outflow valves
(i.e. SLP cannot be maintained at z>6 km).
o Sonic valves at lavatories and kitchens provide a constant independent outflow at those
locations, avoiding odours to mix with cabin air.
o Safety underpressure is limited by a negative pressure-relief valve to p=7 kPa to avoid
structural collapse (controlled by a spring loaded flapper valve).
o A dump valve quickly releases cabin pressure difference when landed.
o A sudden change in cabin pressure p>3 kPa (equivalent to z=300 m near sea level)
cause pain and vertigo.
o Rate of change in cabin pressure is limited to p/t=1 kPa/min (equivalent to z/t=1.5
m/s) to avoid ear problems (one may relieve it by swallowing).
o The pressurisation control system is all pneumatic for reliability.

Table 1. Pressure-altitude correspondence in ISA model (with this model, temperature at sea level is 15
C, drops 6.5 C/km up to 11 km, and remains at 56.5 C up to 20 km).
z [km] p [kPa]
p [kPa] z [km]
0
101.32
100. 0.111
0.5
95.46
95. 0.539

1.
1.5
2.
2.5
3.
3.5
4.
4.5
5.
5.5
6.
6.5
7.
7.5
8.
8.5
9.
9.5
10.
10.5
11.
11.5
12.
12.5

89.87
84.55
79.50
74.69
70.12
65.78
61.66
57.75
54.04
50.53
47.21
44.07
41.10
38.29
35.65
33.15
30.80
28.58
26.50
24.54
22.70
20.98
19.40
17.93

90. 0.986
85. 1.454
80. 1.945
75. 2.461
70. 3.006
65. 3.584
60. 4.198
55. 4.855
50. 5.563
45. 6.331
40. 7.171
35. 8.101
30. 9.146
25. 10.343
20. 11.805

Fig. 2. Cabin pressure regulation as a function of altitude, and a photo of an outflow valve (below the
fuselage) that regulates it.
Pressure sensors are most important in aviation. Two air data computers (ADC) receive total and static
pressure from independent pitot probes (Fig. 3) and static ports on each side of the fuselage (must be
placed where free-stream disturbance is minimum), and the aircraft's flight data computer compares the
information from both computers, checks one against the other, and compute pressure altitude, Mach
number, and vertical speed.

Fig. 3. A flight-officer side pitot probe in Airbus A380, and generic diagram for a pitot system.

Air source. Bleed air circuit

The air source in aircraft ECS is ultimately outside air (in spacecraft it must be produced). Most ECS, up
to now, take compressed air from main engines (before the combustion chambers, of course), instead of
having dedicated compressors to pump outside air (but B787 is based on the latter).
Bleed air systems should bleed at compressor interfaces (HP, LP-HP or LP-MP-HP) because bleeding at
intermediate stages may cause unwanted flow deflections.
One bleed, from HP (exit).
Two bleeds, from HP and LP, for better regulation.
Three bleeds, from HP, LP and fan (the latter to cool the hot air (which is at >200 C if
compressed from 24 kPa to >250 kPa). This is the usual choice in airliners (Fig. 4).
On ground, bleed from APU (must be closed for main engine start-up), or direct air supply from
ground-based air-conditioning units.

Fig. 4. Pressure (in kPa) and temperature (in K) at engine inlet (e.g. 100/288 means 100 kPa and 288 K),
at the end of the low pressure compressor (around the 6th stage), and at the end of compression
(around the 16th stage). The LP and HP bleedings from main-engine compressor are blended to
provide the nominal 250 kPa air-conditioning input pressure, and cooled, if needed, by an additional
pre-cooler, to the 450 K nominal input temperature. A typical 30:1 pressure-ratio main-compressor
is assumed, with a 5:1 bypass ratio. The primary flow follows into the combustion chambers and
then drives the turbine (of some 5 or 6 stages) before exiting the engine.
A nominal bleed-air condition is defined to facilitate interfacing with the ECS.
Bleed pressure is regulated to some 250 kPa (from 200 kPa to 300 kPa, depending on
manufacturer). Lower values may be insufficient to force the air through the two heat exchangers
in the air conditioning machine, and higher values may be a waste of resources (the higher the
bleed-pressure, the higher the cost on fuel).

Bleed temperature is regulated to some 450 K (some 180 C) by cooling the bleed air with ram air
(or better with fan air, to avoid a fan at the heat exchanger). The air bled from the compressor
may be very hot (e.g. at >200 C if compressed from 24 kPa to >250 kPa). Lower temperature
values (<180 C) may be insufficient for the anti-icing system, and higher values may cause
damage to composite structures and plastic materials in the neighbourhood of ECS ducts (the
ducts are made of heat-resistant fibreglass or aluminium tubes). To notice also that the airconditioning packs are always close to fuel tanks.

Most often, there are two air-conditioning packs for safety (3 packs on B747 and DC10), nominally
supplying 50 % of air needs, but able to operate at their 180 % nominal flow rate in case of one failures.
On twin-engine aircraft, each engine bleeding is designed to supply half the total air flow (although the
two bleedings are connected). On three-engine aircraft, the third engine bleed is on stand-by for
redundancy. On 4-engine aircraft each bleeding only supplies of the total design flow.
Control valves in all the bleed system (Fig. 5) protect against flow reversal, and maximum bleeding (e.g.
in case of a break).

Fig. 5. Some of the control valves in the bleed system (from Boeing).
Flow rate multiplier. If only hot air is needed, a small amount of bled air at 250 kPa and 180 C may be
used to pump the necessary total flow rate (5 L/(spax)) from outside air (at 25 kPa, 250 K) with:
A jet pump.
A compressor, driven by a turbine in the bled stream.

Humidity control
Human comfort demands relative humidity in the range 50..70 %, but this is not maintained aboard
aircraft to avoid condensation problems on cold walls and equipment, which tends to corrode metals and
boost micro-organism growth. Relative humidity is kept in the very low range, 10..20 % RH, in spite of

some discomfort (it is highly recommended to drink during flight, to compensate for evapotranspiration in
such a dry environment) and electrostatic build-up (a person with leather shoes and wool suits may grow
15 kV of electrostatic charge on carpets without carbon treatment). At cruise altitudes there is practically
no water in the air, so it is natural to have dry cabin air (but see the example below to account for
passenger evapotranspiration). On ground and during ascent, the air conditioning system must dehumidify
the fresh air introduced into the cabin, particularly in tropical climates when doors are still open.
There are water sinks at the cabin wall bottom to cope with liquid condensate dripping down the external
envelop walls. The condensate frosts on the very cold walls on cruise (at some 30 C), but it is being
melted by circulating air between the outer skin of aluminium and the insulating wall that protects
passengers and cargo (cabin wall usually has three layers. fuselage, insulation panel, and lining). Notice
that the air flow from the cabin interior to this surrounding air curtain (drained at the bottom) is a
continuous source of humidity, whereas if the air conditioning system directly discharges the dry air in
this envelop, it acts as a humidity barrier.
Outside humidity is very small at high altitudes, and is kept very low inside the plane to minimize
problems of condensation and frost on the cold structure (there are water sinks at the cabin wall bottom).
Water vapour released by passengers, some mw,gen=50 g/(hpax) (a third or so by respiration and the rest
by transpiration), and recirculation already yields some 10 % RH within the cabin. New aircraft using
more corrosion-resistant composites in their construction, are being designed to operate with a cabin
relative humidity of 15..20 %, providing more comfort on long flights.
Cabin relative humidity may vary from 90 % (on tropical zones on ground, with doors open) to less than
2 % (in flight with low passenger occupation). In some cases, mist may form inside windows, and hot air
is used to demist the windshields, as in cars (outside defrosting is covered below under Environmental
Protection System). After 3 or 4 hours of exposure to relative humidity in the 510 % range, some
passengers experience discomfort, such as dryness of the eyes, nose, and throat, but there is no conclusive
evidence of extensive or serious adverse health effects of low relative humidity on the flying population
that would justify recommending a regulation to add supplementary humidification systems to aircraft.
Sometimes, the injection of cold air over passenger seats on ground or at low altitudes may yield small
condensation clouds by the mixing of cold dry air with warm humid air in the cabin (not on cruise).
For humidity calculations, one must recall Clapeyrons equation of liquid-vapour equilibrium of a pure
substance for the vapour pressure of a pure substance, pv*:

dpv*
dT

LVE

hLV
TvLV

ln

pv* T
*
v ,0

hLV 1 1

Rv T T0

(with pv,0*=0.6 kPa for T0=0 C, hLV=2.5 MJ/kg, and Rv=462 J/(kgK)), and Raoults equation of liquidvapour equilibrium of an ideal binary mixture:

xv , LVE
xl , LVE

pv* T
p

l , LVE

pv , LVE xv , LVE p pv* T

where xv and xl stand for the molar fraction of water in the vapour phase and the liquid phase (the latter
assumed unity because of the low solubility of air in water), and pvxvp is the partial pressure of water in
the vapour phase. Relative humidity, , is defined as xv/xv,LVE=pv/pv,LVE=pv/pv*, and absolute humidity
by:

mv
M va
M va
M p* (T )

va v
1
p
ma
p
1
1
xv
pv* (T )

Dew point temperature, Tdew, is defined by pv*(T)=pv*(Tdew), i.e. the condensation temperature when a
given unsaturated mass of humid air is cooled at constant pressure and constant absolute humidity. For
solid-vapour equilibrium, the only change is to substitute hLV=2.5 MJ/kg by hSV=0.33 MJ/kg, and the dew
point is then called frost point.
Example 2. Find the condensate production by the air conditioning system on a hot and humid ground
place, 35 C, 80 % RH, to supply cabin air at 22 C and 50 % RH.
Sol.
Absolute humidity values
are
wext=Mvap*(T)/p=0.6220.85.6/100=28
g/kg and
win=0.6220.52.6/100=8 g/kg, so that, for a 2 kg/s ECS flow rate, the required condensation is
ma w 2 28 8 40 g s (i.e. 2.4 kg/h).
Example 3. Find the cabin humidity at cruise, if air renewal is dry (it is always <5 % HR).
Sol.: For an air renewal of 5 L/(spax), 30 g/(hpax). Mean cabin humidity
w=mw,gen/mout=(0.03/3600)/0.005=1.7 g/kg (=1.7/16=11 % HR). Recommendation: drink, and dont
wear contact lens.
Note: (0.5 L/breath)(1/3 breath/s)(0.0012 kg/L)(0.040 kg/kg)=810-6 kg/s=30 g/h. wsat37=40 g/kg.

Thermal loads
There are two basic thermal control requirements:
People comfort demands air supply at T=18..26 C depending on heating or cooling needs, without
thermal chocks, i.e. |TTcabin|<5 C, and without drafts (i.e. v<0.2 m/s close to people and <2 m/s
anywhere). Floor, lateral walls and ceiling are also heated/cooled to enhance temperature
uniformity and help comfort (eliminate high radiative couplings).
Avionics cabinets must be kept within T=30..70 C for proper operation.
The thermal loads on cabin air may come from outside (positive or negative heat flows depending on
environmental conditions, Fig. 5), and from inside (heat dissipation only (from people and equipment).
Besides the basic ISA model, other temperature profiles are considered in aircraft design, like ISA20
(i.e. 20 C below ISA values), ISA+20, or the bounds in Fig. 6.

Fig. 6. Outside air temperature range to be considered in aircraft design.

The thermal loads are:
Heat flow through the fuselage: heat loos on cruise (Tmax=(60)20=80 C, but we must take
into account the dynamic temperature recovery of almost 30 C at M=0.85, so that
Tmax,sub=(30)20=50 C), but heat gain on ground (Tmax=5020=30 C). Taking for the

fuselage 3 mm aluminium pressure shell (not structural: frames and beams), 20 mm insulation
(k=0.1 W/(mK)), and 2 mm coating (k=0.1 W/(mK)), the overall transmittance is K=2.5
W/(m2K).
K

1
1
W

2.5 2
1
0.003 0.02 0.002 1
1 Al in co 1
m K

0.1
1
5
he kAl kin kco he 100 200

Heat release inside: avionics (10 kW), passengers and crew (200*0.1=20 kW), kitchen (10 kW),
lighting and entertainment (1 kW).
Solar heating. May reach 1 kW/m2 on lit surfaces, and it may significantly contribute to heating on
cruise (but worsens the cooling problem on ground on summer).
Indirect heat input from engines through inside piping.

Outside air temperature is measured with a Pt-100 probe extending beyond the boundary layer and
measuring total air temperature (TAT); static temperature can be computed with T=TTAT/(1+0.194M2) in
K (T=TTAT/1.124 for Mach=0.85). Notice that flying through clouds decrease ram heating because part of
it goes to evaporation (e.g. at M=0.7, T=20 C in dry air, but T=10 C in middle clouds, and T=15 C
in high-middle clouds at some 6 km altitude).
The heat sink is ultimately the outside air, but in some circumstances the fuel in the reservoirs is used as
cold fluid (e.g. to cool the lubrication oil, and the avionics in supersonic flight); this fuel heating is
beneficial to its piping and combustion, particularly on cold climates during the early phases of flight
(fuel temperature may be <40 C on ground with cold weather, close to the freezing point of 50 C for
Jet A1, although it gets hot by cooling the lubricating oil and the fuel pump, up to 150 C during descent).
Thermal design is driven by the need to match the following extreme requirements:
Worst cooling case (hottest case):
o Cooling on ground at a hot and humid place, aircraft full, doors closed, APU power.
Refrigeration must be able to cool from 47 C to 21 C in <30 min. In A320, of fuselage
area 350 m2 (4 m diameter and 40 m length), for the 2 kg/s ECS-air-flow-rate to evacuate
the Qwall,max KAT 2.5 350 47 22 22 kW plus some 25 kW from people and
lights, the supply temperature to cabin must be Ti=22(22+25)/2=2 C on hot ground (in
practice always >0 C to avoid frost). Maximum refrigeration efficiency:
(22+25)/150=0.31.
o Cooling on cruise is only applicable to supersonic flight,
where
T=v2/(2cp)=(2300)2/2000=180 C and Twall=220+180=400 K (some 350 K in reality
because of the lower recovery factor).
Worst heating case (coldest case):
o Heating on ground at a cold and humid place, aircraft empty, doors closed, APU power.
Heating must be able to heat from 40 C to 24 C in <30 min.
o Heating on cruise at night. In A320, for the 2 kg/s ECS-air-flow-rate to supply the
Qwall,max KAT 2.5 350 22 (35) 50 kW minus some 30 kW from people and
avionics, or just the 10 kW in the worst case of low occupancy, the supply temperature to
cabin must be Ti=22(5010)/2=42 C on (cold) cruise. Maximum heating efficiency is
then 40/150=0.30.
The fact that cabin air is recirculated in 50 % of the mass flow rate has no influence on the thermal loads
because the entry rate is doubled and the entry temperature difference halved.

A characteristic of aircraft air conditioning is the quick cycling between heating and cooling needs,
associated to rapid changes from ground temperature (that may be >40 C), to cruise temperature (that
may be <60 C). Only spacecraft thermal control has to deal with such environmental jumps (of order
102 K in some 103 s).

Mass and energy balances

A balance equation is a relation between the change with time of some magnitude in a system, and the
causes that motivate it. According to the type of interactions between a system and its surroundings, the
balance equation for a generic magnitude can be layout like that (leaving aside radiation exchages):
For an isolated system:
Accumulation = Production
For a control-mass system:
Accumulation = Production + Diffusion
For a control-volume system: Accumulation = Production + Diffusion + Convection
The mass and energy balances may be written as:

d me
dmi
Q W me hte
mi , gen mi ,e ,
dt
dt
e
e

( Qpax & mw,pax hvap depend on system choice)

Some nomenclature details follow. An isolated system has no interaction with the environment. The term
'accumulation' is better than its equivalent 'storage increment' (notice that both terms convey the idea of a
relative measure, not an absolute measure as the term 'store' implies). All terms in the balance should be
understood as net contributions, i.e., net accumulation (which may be negative if the store decreases), net
production (i.e. production minus consumption), and net fluxes for diffusion or convection (and not flux
in minus flux out).
Example 4. Find the mass of CO2 required to fill an A340 (5.2 m diameter by 70 m length) to kill living
beings for disinfection, on ground.
Sol.: The cabin air cannot be extracted and then the CO2 fed because the structure would collapse beyond
a 7 kPa vacuum pressure. The common procedure is to pump CO2 in through the air-conditioning groundport, while letting the air escape through the outflow valves, until the composition at the exit is largely
CO2. As a minimum the mass of CO2 to fill the aircraft is m=pV/(RT)=1051400/(189288)=2600 kg,
where volume has been estimated as V=LD2/4=7035.22/4=1400 m3, and R=Ru/MCO2=8.3/0.044=189.
As a first approximation, mass and energy balances for cabin air are:
Dry air balance: min mout .
Water balance: min win mw, gen mout wout , where w stands for absolute humidity, and mw, gen is the
water vapour released by passengers (50 g/(hpax).
Energy balance: min hin mw,gen hw,gen Qpax,gen Qequip,gen Qwall mout hout , where h stands for humidair enthalpy (h=cp(TT0)+whLV), hw,gen can be neglected if the metabolic heat release is fully
accounted in Qpax,gen (100 W/pax), Qequip,gen accounts for avionics and lighting dissipation (and
kitchen ovens), and Qwall is the heat flow through the cabin wall (positive when entering, negative
if leaving)

Possible solutions: HVAC systems

Cabin air heating may be achieved by:
Electrical power (resistive or heat pumps). Used in helicopters and some small planes.
Mechanical power (heat pumps directly driven by the engine, instead of by an electrical motor).
Burners. Used on large aircraft in the past. Breathable air (from outside or recirculated) is forced
to pass through a heat exchanger around the burner exhaust.
Exhaust heat recovery. Used in some small aircraft. Similar to a car heating system: a flow of
breathable air (from outside or recirculated) is forced to pass through a heat exchanger that takes

residual heat from the engine (from the exhaust gases here, since aircraft engine is not watercooled like in cars).
Compressor air bleeding (adiabatic heating). This is the system used on most large aircraft.

Cooling below ambient temperature is more difficult to achieve, and it impose peculiar problems: cooling
fluids must pass through walls, water condensate must be eliminated, temperature jumps and efficiencies
cannot be high, etc. Cooling may be achieved by:
Vapour cycles (mechanical or absorption): high efficiency, but do not provide air renovation and
pressurization. Components: compressor (driven by the engine or a motor), heat exchangers
(vaporiser and condenser; fan assisted), expansion valve, piping, insulation, and controls.
Air cycle (inverse Brayton): low efficiency (worst on ground), but lighter, more compact, reliable,
cheap, and provides full solution to cooling/heating plus ventilation and pressurization. This is the
system used on most large aircraft.
Other refrigeration means are only used in special applications: thermoelectric cooling, Ranque
tube, etc.
Temperature control is best achieved by partial mixing of an untreated stream with a heated or cooled
stream, by using a bypass-control valve governed by a downstream temperature sensor (a single valve
shaft may be closing a butterfly valve in one stream while opening another butterfly valve in the other
stream, if they are 90 out of phase).
To satisfy the air-conditioning needs in flight and on ground (we want a cabin air renovation flow of 5
L/(spax) at 75..100 kPa and 22 C), different systems can be used, e.g.:
Up to 1950, only heating was available (from engine heat recovery, electrical heaters, or burners).
A vapour-cycle system like in cars air conditioning (best with heat-pump capability),
supplemented with a separate ventilation system. This is most used in small aircraft.
An air-cycle system that provides both heating/cooling and ventilation. This is the standard in
medium and large aircraft because of its compactness and reliability, in spite of its poor energy
efficiency (its operation cost is typically 1 kW/pax at full capacity; some 350 kW for a 400
passenger airliner, about 75 % of allnon-propulsive energy consumption aboard). The ECS packs
are typically located close to the wing box section (at the belly, in the vicinity of centre fuel tank).
State of the art:
All in one solution used in most aircraft: the air cycle machin (ACM) pack (but
heating/refrigeration efficiency, =qcabin/wmain,comp=0.5). A separate electrically-driven ACM
would be more versatile and efficient, =0.7, but heavier and less reliable.
Vapour-compression heat pump (=3, but no pressurisation, no ventilation) used in many small
aircarft.
On ground, instead of using the APU to feed with bleed air the ECS aboard, high-efficiency ground
refrigeration systems are being used in main airports. A novel ground-ECS service is based on water
chillers with ice storage (ice is made by night, and used on spot to avoid large refrigerators).

The air cycle machine (ACM)

The typical air-cycle conditioner pack works as follows (Fig. 9):
Air is bled from the main engine compressors through a flow control valve at around 250 kPa; it is
very hot (>200 C) because of the adiabatic compression-heating. As pressure and temperature of
bleeding depends on compressor stage and spinning rate, this pressure is regulated by having two
or three bleedings at different stages and control valves (see Fig. 7).

The hot bled air always requires cooling before entering the cabin, but a simple heat exchanger
(HE) with outside air is not efficient (a big HE is needed to cool that amount of air from 200 C to
20 C, particularly at low altitudes). To guarantee that bled air can be safely piped from the engine
through the fuselage to the air-cycle machine and deicing system, a heat-exchanger (pre-cooler) in
the engine cools bled air from the compressor (at >200 C) to some 180 C (450 K).
The air-cycle machine is based on an inverse Brayton refrigeration cycle. Pre-cooled air from the
bleed system is further cooled in a so-called primary heat exchanger (Fig. 8) down to some 110
C, and then enters a compressor with pressure ratio around 12=1.8 that raises the air to 210 C,
and a second heat exchanger lowers the temperature again to some 100 C. Pressure losses in these
compact heat exchangers are high (20..40 kPa). Afterwards, air passes through a turbine and exits
at about 5 C, to be mixed with some hotter bled air at around 100 C to get the 10..35 C (it
depends on operation phase) needed to keep the cabin air at around 22 C at all times, accounting
for internal heat release (passengers and equipment), outside heat gain and loses, and air
recirculation. The thermodynamic process is sketched in Fig. 9. Additional filters/separators are
provided to remove water condensate on ground and low altitudes in humid climates (just after the
ACM turbine in most cases, and before the turbine too when very cold cabin air is needed on
ground, to avoid the risk of icing inside the turbine); the condensate is injected into the ram air
stream to help HE efficiency. Maximum absolute humidity for condenser design is 22 g/kg
(saturated air at 26 C and 100 kPa, or 40 % RH and 40 C).The common-axis turbo-compressor,
of some 0.2 m in size, is made of aluminium, with air bearings.
Ram air is used as a heat sink in both heat exchangers, which are of the compact type (cross-flow
corrugated aluminium-plates). Ram air is captured through an inlet door and a diffuser, and forced
by a fan (driven by the ACM turbine) to go through the two heat exchangers mentioned above, a
exhaust nozzle, and a louvered exit door. The inlet and outlet doors are linked and driven by the
same actuator (one for each pack); these doors remain closed when there is no cooling need (e.g.
when the ACM compressor-outlet temperature is <120 C or so). Typical mass flow rate of ram air
is two or three times greater than bleed air, in each ACM.
Two equal ACM nominal systems are implemented, to allow for one failure (Fig. 10).

Fig. 7. Environmental control system schematics.

Fig. 8. ECS heat-exchanger layout (typical size of each HE is 0.70.50.3 m3, and 10..15 kg). Photography
of one of the heat exchangers.

Fig. 9. ACM thermodynamic diagram and flow schematics for the cooling case.

Fig. 10. Typical location of the ACM and cockpit controls.

The ACM controls at the cockpit are:
A mode knob for each ACM pack with four positions: Auto, Heat, Off, Cool.
A thermostat knob for each zone in the plane with three positions: Normal, Warmer, Cooler.

ACM working cases

Heating at cruise level with low occupancy. The ACM must supply air to the cabin at some
Tcab,in=42 C (315 K) and p>75 kPa, from bled air at regulated values of 450 K and 250 kPa (from
high-pressure compressor), and outside air at 250 K and 22 kPa.

Heating on coldest ground conditions with low occupancy. The ACM must supply air to the cabin
at some Tcab,in=42 C (315 K) and 100 kPa (ambient pressure really), from bled air at a regulated
pressure of 250 kPa (from APU or high-pressure compressor, idle) and available temperature of
some 30 C (300 K), and outside air at 240 K and 100 kPa (ambient pressure really). In this case,
using the overpressure to drive a jet pump with outside air, may alleviate the cooling need in the
ACM, saving on bleed flow rate.
Cooling on hottest ground conditions and at low altitudes. The ACM must supply air to the cabin
at some Tcab,in=2 C (270 K) and ambient pressure, from bled air at a regulated pressure of 250
kPa (from APU or idle conditions on ground, or from low pressure in flight) and available
temperature of some 100 C (370 K), and outside air at 320 K and ambient pressure.

Notice that, because hot bled air (at 350..450 K) is always the input to the air conditioning machine, the
ACM must always cool this input air to feed the cabin at 2 C in the worst cooling case or at 42 C in the
worst heating case, and thus, ACM may by understood as either air-cycle-machine or air-coolingmachine.

ACM size and efficiency (energy cost)

The ECS is usually split in one ACM pack per engine. The ACM size is dictated by the ventilation
requirement of 6 (g/s)/pax minimum (e.g. 1.2 kg/s minimum for the 200 pax capacity of A320; some 2
kg/s is the typical design value). The A340 ECS weights 720 kg, and has four ACM, each fed with 1 kg/s
of air.
On large airliners, some 1 kg/s of compressed air is bled from each engine (from the 80..100 kg/s of
primary air used). Each air pack (ACM) weights some 150 kg. Each ACM takes some 3 kg/s of ram air
per pack (plus 0.5 kg/s of post-fan air for the pre-cooler at each engine).
The operation cost in the nominal ECS is the work spent in compressing the bled air, which is
WECS mbleed c p Tbleed T0 =11000(450-250)=200 kW per pack.

Future ECS
Present ECS design is reliable, but expensive to operate (1 kW/pax). There are also some maintenance
problems due to overheating on ground and during ascent/descent (when ram air flow must be forced with
a fan, the engine fan or a dedicated one). New designs seem to point to a more distributed and
electrically-driven system, but different manufacturers follow different trends:
AIRBUS. For its new born aircraft the A350 XWB, Airbus decided to keep on implementing an
engine bleed air driven ECS but combined with advanced engines, the GE GEnx-1A and the RR
Trent 1700.
BOEING. For its new aircraft, the twin-engine 300 seat B787, Boeing has moved to an all-electric
ECS (AE-ECS), eliminating any engine bleed. Cabin pressure is increased to >82 kPa (1800 m
cabin-altitude). Additional electrical generators on the engines supply the ECS power (hundreds of
kilowatts) to drive dedicated compressors that feed on ram air, and electric heaters for de-icing
(electro-thermal heater-mats attached to the aircraft slats). There are two ECS compressors on
each pack, which are 0.3 m diameter, run at 40..50 thousands rpm on air bearings, and have a total
pressure ratio of =5 (on each pack), taking ram air (at cruise, from 20 kPa to 100 kPa), and
delivering it at 90 C instead of the 180 C in conventional bleed-air ECS. This AE-ECS adds
some 200 kg on airliners (and some maintenance costs), but saves some 5 % fuel (may be
5000 kg). Higher humidity in the passenger cabin is possible because of the use of composites
(which do not corrode). Ozone is removed from outside air; HEPA filters remove bacteria, viruses
and fungi; and a gaseous filtration system removes odours, irritants and gaseous contaminants.

AIRCRAFT ENVIRONMENTAL PROTECTION SYSTEM

Protection against external actions: ice, rain, high temperatures, and cold temperatures, must be
adequately provided. Protection against very low temperatures in aircraft is needed to avoid ice
formation, and to avoid freezing of internal liquids (from fuel to sanitary water). Protection against
unwanted organism invasion (micro-organisms, insects, rodents, and reptiles, treated under ECS, may be
considered EPS too.

Against ice formation

Ice formation may occur:
On ground, by precipitation of ice particles on upper surfaces. Aircraft take-off is forbidden with
ice or snow on wings and controls. They are cleaned by ground means (using propylene glycol).
Icing may cause severe damage also to grid power lines, wind turbines and ship masts and ropes
(the oldest problem).
On flight through clouds (mainly cumuliform), by ice particles and supercooled droplets striking
onto frontal surfaces. The worst case is subcooled liquid drops (in metastable state, favoured by
the absence of nucleation sites). Warning; we use supercooled and subcooled as synonyms,
whereas other authors use subcooled as synonym of unsaturated liquid above its freezing point.

Icing problems in aviation

Falling pellets of ice (hail falling from cumulonimbus clouds with very strong rising air currents) can
damage anything, including aircraft and engines. Heavy snow can cause severe icing depending on its
temperature and state. Heavy rain may impede visibility and soak critical parts. But flying inside clouds
with supercooled liquid droplets (SLD) is most feared because of the degradation it causes on aircraft
flight dynamics.
When flying through subcolled droplet clouds, ice chunks a few centimetres thick may form when flying
through a 50 km span cloud (some millimetres per minute; it grows linearly with distance and with
subcooled water content). An ice thickness around 1 mm near the leading edge of wings and rotors may
decrease lift up to 30 %, particularly at large angles of attack, greatly changing the point at which
aerodynamic stall takes place (it may decrease the stall angle of attack by 6). The lift loss is most critical
during the descent and approach phases, where speed is reduced and angle of attack is large. The worst
occurs when the stall leaves without lift some control surfaces, an a widely unstable flight dynamics
develop (e.g. American Eagle Flight 4184 in 1994, or Colgan Air flight 3407 in 2009)
Icing is one of the most dangerous weather conditions for general aviation (icing is involved in nearly 10
% of all aircraft accidents). All aircraft must be able to heat up leading edges (by bleeding hot air, or
electrically), or de-ice by other means (ancient rubber bands, or modern electromagnetic striction pulses).
Requirements: airliners must fly in all-weather conditions, thence ice protection is needed on:
Structure
o Wings, flaps, tail, and mechanisms (e.g. ram air inlets and outflow valves). In the case of a
multi-element wing airfoil, the anti-icing hot air is used to remove ice from the first
element, i.e. at the slat only, not on the main body or the flap.
o Windscreens. They are laminated compounds made of (from outside-inside): a thin
tempered glass layer, a transparent heater film of SnO2 (or fine-wire mesh), a vinyl layer
(hot bonded), a structural glass layer, a vinyl layer, and a glass layer; some 30..40 mm
thick in total (it must resist the 100 kPa cabin overpressure and the impact of a 2 kg bird at
high speed). They incorporate temperature sensors (termistors), and is an expensive item.

Windscreens wipers must be often cleaned and never operated dry, to avoid scratches on
the outer glass surface.
o Sensors (pitot and static probes, antenna).
o Water discharge (sanitary, condensates).
Engines
o Nose cowling (the streamlined metal covering fitted around an aircraft engine). Ice
particles usually melt on compressor heated air, but sometimes the water freezes again on
metal surfaces, and accumulated ice can either break into chunks that damage turbine
blades, or melt and douse the ignition system. There have been up to 14 reported cases of
in-flight dual-engine flame extinction (all of them were fortunately quickly re-ignited)
attributed to ice ingestion (plus many single-engine cases). Although torrential rain and
hail were known to shut down engines, icing was not considered a big problem before year
2000.
o Fan or guide blades. Besides, tiny ice crystals (undetected by pilots and weather radars)
may be sucked in great amounts inside jet engines and become a hazard on high-altitude
storms because they can coalesce after melting and arrive in liquid form to combustion
chambers and douse the flame.
o Propellers, if any (and old carburettors).

In brief, icing decreases lift, adds resistance, decreases thrust, unbalances rotors, blocks eyesight, sensors,
and actuators, and adds weight.
Solution: stages in solving the problem (of icing):
Detection and warning of risk.
Action (available tools and procedures) to avoid or repel damage.
Analysis to solve (predict and avoid, or minimise unwanted effects of) future problems.
Ice detection
Visual ice detector. Pilot visual clues of ice formation:
1. Light icing: ice under the windscreen wiper blades. Ice accretion may create a problem if flight
is prolonged in this environment over 1 h.
2. Moderate icing: ice on the wiper nut.
3. Severe icing: ice on the central windscreen pillar (Fig. 11).

Fig. 11. Severe icing on the central windscreen pillar of an aircraft, formed in under a
minute flying at 6 km altitude (FL200).

Optical ice detector. Based on a protruding acrylic-glass cylinder (the same material used for
aircraft windshields). The device works as a combined optical spectrometer and optical switch. A
change in opacity registers as rime ice. A change in refractive index registers as clear ice. Ice
thickness down to 0.01 mm can be detected. An internal heater may be added to force the cleanup. The wavelength of the transducer's excitation light is not visible to the human eye so as not to
be mistaken for any kind of navigational running light.

Electrical ice detector. Based on magneto-strictive resonance change in a magnetic cylinder (6

mm diameter and 25 mm long), which is axially vibrated with an oscillating magnetic field, and
the extra ice mass lowers its resonant frequency. When f>150 Hz (ice>0.5 mm), an internal
heater melts the ice and a counter is updated; after two heating cycles, the engine-anti-ice system
is switched on; after 10 cycles, the wings anti-ice system too).
Other kind of detectors (acoustic, nuclear) have been tried too.

Anti-icing (prevention) and de-icing (fight) systems:

Thermal (by hot air). Hot air is bled off the jet engine into tubes routed through wings, tail
surfaces, and engine inlets (the hot air is expelled through holes in the piping, and finally vented
outside). Ice must be completely vaporised because if it is just molten, the run-off water may
freeze downwards; on large airliners it may take up some 100 kW of air bleed. On modern highbypass airliners, primary air is scarce, and they are being designed to tolerate some icing in noncritical areas (e.g. wing tips and roots). On small aircraft, engine-exhaust heat recovery systems
are used.
Electro-thermal (quasi-steady, or short-pulse). Unlike conventional windshield defrosters that rely
on gradual warming to liquefy snow and ice, there are new devices that deliver a pulse (~1 s) of
high-power electricity that immediately melts the ice at its interface with an object's surface, and
the flow sweeps it away. This system demands less energy (some 20 MW/m 2 during a few
seconds instead of some 500 W/m2 during many minutes) because heat is not wasted warming up
the object and the ice.
Mechanical (inflatable rubber boots). These systems require less engine bleed air but are usually
less effective than a heated surface. They must be inflated in small lengths at a time, to avoid large
aerodynamic disturbances.
Chemical (application of a deicing fluid through small holes in leading edges (weeping wing
system).
Electromagnetic (microwaves through composite walls).

Icing tests
Airliners must be able to safely fly through subcooled clouds under the following foreseen extremes
(Table 2):
Continuous regime. Clouds up to 40 km long at any altitude up to 6 km, with subcooled drops
characteristics in the envelop shown in Fig. 12a.
Intermittent regime. In addition to the above continuous regime, the aircraft must cope with
smaller but denser clouds: up to 5 km long, with subcooled drops characteristics in the envelop
shown in Fig. 12b.

Fig. 12. Fly worthiness on subcooled clouds: a) continuous regime, b) intermittent regime.
Table 2. Conditions likely to be encountered in service operation in icing conditions.

Altitude range
Atmospheric
[km]
temperature [C]
3..9
0..20

5..12

20..40

Maximum water content

[g/m3]
5
2
1
5
2
1
0.5

Maximum cloud extent

[km]
5
100
500
5
20
100
500

The main method used by aircraft manufacturers to certify their aircraft is to test a model of the aircraft in
simulated icing conditions using an icing tunnel. Another method is to perform a tanker test, where a
tanker aircraft sprays water droplets onto a following aircraft, at freezing altitudes. Since both of these
tests are expensive and time consuming, computational methods are gaining in popularity.
An icing wind tunnel is a classical closed-loop subsonic wind tunnel, with an air cooler (a refrigeration
heat exchanger to deliver air at 20..0 C). To initiate icing, a droplet sprayer array injects chilled water
(>0 C) upstream of the test section to create the amount of water and drop-size desired. NASAs Icing
Research Program uses an Icing Research Tunnel (IRT) of 23 m 2 test-section, two Icing Research
Aircraft (IRA), and computer codes to predict ice growth.
Those parts of the airframe where the accretion of ice under the conditions of Appendix C is likely to
have an adverse effect on the airworthiness of the aeroplane, should be tested for a period of 30 minutes
duration at each of the conditions specified in Table 3. The encounters considered with Intermittent
Maximum concentrations should include three clouds of 5 km horizontal extent, separated by spaces of
clear air of 5 km. At the end of the tests the total ice accretion should be such as not to adversely affect
the safety of the aeroplane.
Table 3. Test to be performed on critical airframe parts for icing resistant certification.
Atmospheric
Continuous maximum
Intermittent maximum
Mean effective
3
3
temperature [C]
liquid water content [g/m ]
liquid water content [g/m ]
drop diameter (m)
0
0.8
2.5
20
10
0.6
2.2
20
20
0.3
1.7
20
30
0.2
1.0
20

Ice accretion modelling

The first formulation was by B.L. Messinger in 1953 (Equilibrium temperature of an unheated icing
surface as a function of airspeed, J. Aeronaut. Sci., 2942), who established the energy balance on icing.
Subsequent enhancements have included the addition of an energy source at the substrate for deicing, and
a temperature rise due to compressible flow.
Icing models are based on mass and energy balance equations for a heterogeneous flow of air with
subcooled liquid droplets, against the aircraft surface: wing, engine cowling, etc. There are several stages
in modelling the problem:
1. Specifying the configuration (meteorological and aerodynamic)
2. Computing droplet impacts, for catching rate determination.
3. Solving the energy balance, for freezing fraction determination.
4. Ice accretion rate determination in case of no deicing action.
5. Effect of deicing means.

The factors controlling ice formation and accretion rate are:

Meteorological: liquid water content, LWC (LWC stands for liquid water content, in the range
0.2..2 g/m3), droplet size, d (in the range 5..50 m, related to LWC and upward air motion),
environmental temperature, T (in the range 35..0 C), and longitudinal extent of these conditions
(cloud size along trajectory, in the range 5..100 km). Environmental pressure, p (in the range
45..95 kPa) has a minor influence. Satellite images in the 3.9 m band are being used to
investigate potential icing conditions.
Aerodynamic: geometry (e.g. airfoil shape and angle of attack), and flight speed (which
determines the residence time on clouds, and droplet drift through the flow field). For subsonic
aircraft, the flow field at the leading edge can be approximated by the flow field around the front
side of a circular cylinder with the same radius of curvature, but for transonic wings this leadingedge equivalent cylinder is so small that most icing takes place downwards in the airfoil. In the
former case, the potential air flow velocity in body-fixed axes is:

x2 y 2

y
u x, y U 1 R 2

2
2
2
2

2 x y 2
x

2 xy

x
2

v x, y U R

2
2
2
2
2

y
x y

where U is the free stream velocity (i.e. flight speed), R the cylinder radius, and G the circulation
that places the stagnation point at the corresponding point in the leading edge of the wing.
Catching factor
If a uniform distribution of droplets in the cloud is assumed with a density LWC, the undisturbed
incoming water flux is jw,=LWCU (normal flux, [(kg/s)/m2]), and we want to find the impacting water
flux (per body surface area), jw,p (the p refers to a panel at the wall, and the w to water), which is
solved by analysing droplet trajectories.
Droplet trajectories and impact locations are calculated by integrating the momentum equation for each
individual droplet, accounting for air drag (air-to-droplet relative speed is 0 initially and close to U on
impact). Droplets are assumed spherical up to the impact, without gravity, and without influencing the air
flow field (the potential flow solution u x, y , v x, y as given above is good enough, since the
residence time in the boundary-layer flow is minimal). Namely:

1
mxd vr2cd
2
1
myd vr2 cd
2

d2

cos
yd v

4
with

arctan
, and vr

u
d2
d
sin

xd u yd v
2

where vr(x,y) is the relative drop-to-air velocity at each point, related to droplet velocity
xd x, y , yd x, y , and the free air stream velocity u x, y , v x, y .
The output of this trajectory analysis yields the upstream region of droplets impacting on the solid, and
the spatial distribution of impacts, what is evaluated by a catch factor such that jw,p=jw, ( can be
viewed as the ratio dy/ds, i.e. for a small stream-tube of droplets, the quotient between their undisturbed
cross-section, dy, and their striking area at the body surface, ds, Fig. 13).

Fig. 13. Impact of water droplets on the leading edge of an airfoil. a) Simple airfoil, and compound
airfoil (with slat and flap). b) Close-up view and nomenclature. c) A typical impingement
efficiency distribution (catching factor).
Type of ice formed
Rime ice. Frost formed by the freezing of highly-supercooled small water-droplets onto cold solid
objects (T<15 C, above 4..5 km altitude) is called rime, an opaque white form of ice (with
entrapped air) with lower density than glass transparent ice (some 880 kg/m3 instead of 917
kg/m3). Light or moderate icing occurs in clouds of type Ac, As, St, or fog, and moderate to severe
rime in Ns and Cb. Splashing (droplet bouncing on impact) is never considered, but partial droplet
sticking must be considered for large droplets that do not fully freeze, and the remaining liquid
part runs away in a liquid layer over the ice layer formed.
Glaze ice. Under milder conditions (10..5 C, in strong convective clouds at lower altitudes
with larger subcooled drops), glaze ice forms (a transparent wet glossy ice layer under a liquid
water layer). In this case only a fraction of the droplet will freeze instantaneously, the rest of the
droplet can remain liquid for some time and may flow as runback water. Large drops can only
occur in atmospheric upward vertical motion, to compensate terminal velocity (3 mm/s at 10 m,
70 mm/s at 50 m; Stokes law applies up to d<50 m).
The ice shape formed on leading edges is highly irregular, three-dimensional, and bulky (not a smooth
layer).
Thermodynamic analysis. Freezing fraction
Under very cold conditions, all water freezes and dry rime ice forms; the problem reduces then to solving
a simple mass balance for ice accretion. In milder conditions, part of the water freezes, forming glaze ice,
and the rest runs off along the surface, and now the problem is governed by coupled mass and energy
balances, which determine the ice height and water layer thickness. We intend here to model all possible
cases, but we neglect the effect of the thickness of the liquid water film formed, if any.
The energy equation determines the droplet temperature after impact, discriminating between droplets
that remain liquid and are swept away with the flow, fully solidified droplets that stick to the surface and
modify its geometry, or the fraction of droplet mass that freezes (the rest runs away and spreads along the
surface). The quasi-steady energy balance at the air-water interface over the solid is:
H w,in Qp,in fHice,out 1 f H w,out Qp,out

(1)

where H w,in is the total enthalpy of the water flow incoming (mind that there is no air flow through the
interface), Qp,in is the heat flow coming from the panel (p is used instead of wall to avoid confusion
with water, and because numeric aerodynamic simulation is based on the panel method); f is the fraction

of water mass frozen, H ice,out and H w,out are the total enthalpies of ice and liquid water exiting the
interface, and Qp,out the heat flow lost from the panel to the environment. Notice that a possible
evaporation term on the right hand side of (1) has been neglected, in spite of the expected temperature
increase.
In more detail, taking the enthalpy origin H=0 for liquid water at 0 C, and working per unit surface area,
dA,

U 2
U 2
jw cw
qp,in fjw cice p hSL 1 f jw cwp h p
2
2c p ,air

(2)

where jw=jw, is the flux of water through dA (the jw,p calculated above), cw, cice and cp,air are the water,
ice and air thermal capacities, and p the undisturbed temperature and panel temperature (both in
degrees Celsius, since we put the enthalpy origin at 0 C), hSL the enthalpy of fusion, and h the heat
convection coefficient. Notice that a single temperature has been assumed for both the frozen fraction and
the liquid remaining, if any.
For a numerical estimation of (2), we start neglecting the heat flux coming from the panel, qp,in 0 (i.e.
adiabatic wall), and assuming that the water temperature after impact is 0 C for both the ice formed and
the possible run-off liquid; the aim of this exercise is to find the amount of water frozen (i.e. the ice
fraction, f). Consider an aircraft flying at some U=150 m/s (M=0.4..0.5, during ascent or descent),
through a cloud with liquid water content LWC=0.5 g/m3 (typical of continuous icing regime, Fig. 10a);
the undisturbed water flux is jw,=LWCU=0.510-3150=0.075 (kg/s)/m2. Consider a leading-edge
characteristic length of D=0.2 m (which yields a Reynolds number, ReD=UD/=1500.2/(1510-6=2106
), and let assume the catching factor is =1 to have jw=jw,=0.075 (kg/s)/m2 in (2), with cw=4200
J/(kgK), cice=2000 J/(kgK), cp,air=1000 J/(kgK), hSL=334103 J/kg, and h=300 W/(m2K), obtained from
the high-speed forced-convection correlation:

NuD 0.3

0.62 Re1D2 Pr1 3

14

1 0.4 Pr 2 3

ReD 5 8
1

282000

45

(3)

good for 102<ReD<107 and ReDPr>0.2 [Churchill, S.W. & Bernstein, M., 1977]; which, with Pr=0.7 for
air, and ReD=210-6 yields NuD=hD/k=2430, and h=kNuD/D=0.0252430/0.2=303 W/(m2K). Solving (2)
we get f=0.0247Tw0.17, which is plotted in Fig. 14, showing that, below T=47 C, all water in the
droplet freezes on impact, and above T=7 C no ice forms; for ambient temperatures in between, part of
the droplet freezes and the rest runs off as a liquid film.

Fig. 14. Fraction of ice formed, f, as a function of droplet temperature (assumed to equal outside air
temperature, T). Below T=47 C, all water in the droplet freezes on impact, and above
T=7 C no ice forms (water nuns off).
Furthermore, the ice accretion rate would be:

dz fjw

dt ice

(4)

which in our numeric example, in the worst case of f=1 (all incoming water freezes), becomes
dz/dt=(0.075 (kg/s)/m2) / (900 kg/m3)=0.083 mm/s=5.0 mm/min=30 cm/h, or in other words an ice layer
55 mm thick after a 100 km flight on these conditions (at 150 m/s). This exercise has been an upperbound analysis (the catching factor is always <1), and the ice fraction is usually on the lower range too
(f<1), but it is not uncommon that icing becomes dangerous in a few minutes of flight in some subcooled
clouds.
The heat power needed to vaporize the ice being deposited, in the worst case just mentioned (f=1), is:
qp,in jw hSV

(5)

which, in our numeric example becomes qp,in =0.0752.83106=0.21106 W/m2. If a strip 0.1 m wide and
10 m long is to be heated with that flux (only part of the wing span is deiced), the total heat flow needed
is 210 kW, which, as an upper bound found in this example, is of the correct order of magnitude for the
bleed air system used in airliners.
Nowadays, detailed numerical codes are used to find the air flow around the wing, the distribution of the
catching factor, and the fraction of droplet mass frozen. Finally, from the solid accretion rate found and
ice density jump, the solid deposition may be accumulated into subsequent integration steps, and the
transient ice layer growth incorporated to the air flow field around the wing, until wing stall takes place in
the simulation (see Fig. 15 for a comparison of numerical simulation with wind tunnel data).

Fig. 15. Icing predictions on a NACA 0012 airfoil at T=10 C (horizontal and vertical scales in
fractions of the cord) [3].

Against rain
Rain rates are classified by measuring the precipitation collected in one hour interval (Table 4).
Table 4. Rain classification [mm/h].
Very light
<0.5
Light
0.5..2
Moderate
2..15
Heavy
15..30 (d>1 mm)
Very heavy
30..60
*
Extreme
>60 (d>2 mm, v>5 m/s)
*
The record is 700 mm/h during 10 s (NASA-1990).

Torrential rain (>15 mm/h, d~1.5 mm, v~5 m/s) prevents visibility, and may shut down an engine by
flameout. The same malfunction may occur at take-off by water in the runway being thrown up by the
wheels. In all this cases it is convenient to turn the engine igniters on (in some engines this action is
automated). The few cases where flameout has taken place in all engines at a time happen to be near
landing, when the engines were throttled back.
The windscreen wiper must allow reasonable visibility with rain of 40 mm/h, d~2.5 mm, v~8 m/s. An
additional anti-rain system is an external hot-air curtain (bleed air).

Against dust
More dangerous to engines than water precipitation is the effect of clouds of dust or ash created by
volcanic eruptions like Mount St. Helens in 1980 and 2006, or Eyjafjallajkull eruption in 2010, where
most of Europe airspace was closed for nearly a week, the largest air-traffic shut-down since World War
II, causing 10 million passengers to be stranded not only in Europe, but across the world, with over
100 000 flights cancelled and 1300 million losses in the aviation industry.

Against high temperature

Usually due to aerodynamic heating. Total temperature of a (perfect gas) gas flow,
Tt=T+v2/(2cp)=T(1+(1)M2/2). This total value is attained in stagnation points, and even when
decelerated by friction in the boundary layer in the adiabatic-wall case, where some 85 % of the dynamic
temperature is recovered at the wall (the recovery factor is r Pr ).
To compute heat transfer flow by convection from a high-speed vehicle to/from the external fluid (i.e. in
compressible flows), it is this adiabatic wall temperature:

1 2

Taw T 1 r
M , r Pr
2

which must be used in Q hA Twall Tref as reference temperature, Tref (recall that it was simply the
temperature at the far field for external flows, Tref=T, and the fluid bulk mean temperature for internal
flows, Tref=Tbulk).
Airliners flying at Mv/c=0.85 at 10 km altitude (T=50 C, c RT =300 m/s, v=250 m/s)
suffer a dynamic temperature jump Tdyn=v2/(2cp)=32 C.
The Concorde flying at M=2 at 17 km altitude (T=57 C, c RT =295 m/s, v=590 m/s) suffer
a dynamic temperature jump Tdyn=174 C.
Spacecraft during re-entry are exposed to very hot dissociated ionised air that may reach 10 000 K
at the frontal shock wave layer (the rule-of-thumb is to assume the maximum air temperature in
kelvin to be equal to the entry speed in meters per second); the outer skin must withstand up to
2000 K at their nose and other stagnation points, either by suitable refractory materials, or most
often by ablation.

SUPERSONIC AIRCRAFT
Concorde (1976-2003) has been the only successful commercial supersonic aircraft (the only other one,
Tu-144, was cancelled after less than a year in commercial use); it flew at M=2 (v=Mc=2295=590 m/s)
for 2 h, at 17 km altitude. The other type of supersonic aircraft is the combat aircraft, which may keep
supersonic speed only for tens of minutes.

Concorde had unique environmental control and protection needs. To begin with, in spite of the nearly
60 C outside air temperature at 17 km, the aircraft had to be permanently cooled to keep a cabin
temperature of some 20 C, because the dynamic heating is T=v2(2cp)=5902/(21000)=174 C (measured
nose tip temperature reached 400 K, and windows in the cockpit were too hot to touch). However, on
ground and during ascent (subsonic) heating was required on winter.
If the outside air was at >100 C close to the aircraft, how to cool the cabin? Thermodynamics teaches
that one may cool a system in a hot environment, by using a refrigerator, but the solution adopted for
Concorde was to take advantage of the large thermal inertia of the fuel carried onboard (96 000 kg out of
the 187 000 kg at take-off), to use the fuel as the heat sink during the flight for the environmental control
system (cabin air, avionics, hydraulics). At cruise, conditioned air entered the cabin at 25 C.
Concorde went through two cycles of heating and cooling during a flight, first cooling down as it gained
altitude, then heating up after going supersonic. The reverse happened when descending and slowing
down.
Designers wanted to keep to aluminium for the aircraft main structure, and, because highest sustained
working temperature for aluminium is about 400 K, which limited the top speed to the nominal
maximum, Mach 2.02.
The use of aluminium and the supersonic heating caused the 62 m long fuselage to extend some
L=LT=602010-6200=0.25 m, the most obvious manifestation of this being a gap that opened up on
the flight deck between the flight engineer's console and the bulkhead. It is said that on all Concordes that
had a supersonic retirement flight, the flight engineers placed their hats in this gap before it cooled, where
the hats remain to this day.
Cabin pressurization was a big concern, and the fuselage had to be narrow (and windows very small) to
withstand 82 kPa inside (1800 m cabin pressure altitude) with 10 kPa outside. Sudden cabin
depressurization was lethal, so that a compressed air reserve, and a rapid descent manoeuvre was
designed to reach a safe flight level before loosing too much pressure (window holes were small).
Ozone had to be reduced from ambient 6 ppm at 18 km altitude to <0.2 ppm, using a Ni-catalyser in the
heat-exchangers (working at 580 C).

REFERENCES
1. http://en.wikipedia.org/wiki/Environmental_Control_System.
2. http://www.liebherr.com/ae/en/products_ae.asp?menuID=106050!160-0. Liebherr-Aerospace has
designed, developed, manufactured and serviced airplane and helicopter air conditioning systems
for more than 50 years.
3. Neese, B., 1999, Aircraft Environmental Systems, Endeavor Books.
4. Prez-Grande, I., Leo, T., 2002, Optimisation of a commercial aircraft environmental control
system, Applied Thermal Engineering 22 (17), pp. 18852004.
5. Leo, T., Prez-Grande, I., 2005, A thermo-economic analysis of a commercial aircraft
environmental control system, Applied Thermal Engineering 25, pp. 309325.Cebeci, T., Kafyeke,
F., 2003, Aircraft icing, Annu. Rev. Fluid Mech. 35:1121.
6. Fortin, G., Laforte J.L., Ilinca A., 2006, Heat and mass transfer during ice accretion on aircraft
wings with an improved roughness model, International Journal of Thermal Sciences 45, 595606.
7. Saeed, F., 2002, State-of-the-Art Aircraft Icing and Anti-Icing Simulation. ARA Journal, Vol.
2000-2002, No. 25-27.