Structural loads in flight (1/3)

Flight Envelop
Yves GOURINAT
Professeur de Mécanique des Structures, Institut Supérieur de l’Aéronautique et de l’Espace

Abstract
This paper presents the frame and definitions required for the computation of the
mechanical loads applied to the major structural elements of aircraft, in the context of
civil aviation. In view of civil certification, the applications to thin structures for lifting
surfaces and pressurized elements are presented. The flight domain is defined in the
context of civil regulation, in relation with the probabilities of occurrence of static
equivalent loads, and replaced in the global perspective of transient (gust) or
continuous (turbulence) dynamic loads, responding to flexible aircraft model.

6 keywords:
Flight domain, load factor, lifting surface, pressurized shell, limit loads, gust.

Glossary of main abbreviations and notations

a Incidence aérodynamique
Aerodynamic incidence

b Corde moyenne de voilure

Average chord of wing

d Braquage de gouverne
Rudder deflection

EASA European Aviation Safety Agency

CS-23 Certification Specification part 23

CS-25 Certification Specification part 25

CX Coefficient of drag

CY Coefficient of lateral lift

CZ Coefficient of lift

DSP Densité Spectrale de Puissance

Spectral density of power

e envergure
span
FAA Federal Aviation Administration

FAR-23 Federal Aviation Regulation part 23

FAR-25 Federal Aviation Regulation part 25

FL Flight Level

g0 Accélération gravitationnelle de reference

Reference gravitational acceleration
g0 ≅ 9.80665 m/s2

HdV Heure de Vol

Flight time

kgf Kilogramme force

1 kgf ≅ 9.80665 N

LL Limit Load

MTOM Maximal Take-Off Mass

nZ Facteur de charge
Load factor

OEM Operating Empty Mass

pkin / pcin Pression cinétique

Kinetic pressure

SL Safe Load

Sref Surface conventionnelle de reference

Conventional reference surface

UL Ultimate Load

V0 Vitesse cinétique équivalente ramenée au niveau de la mer en

atmosphère standard aéronautique
Equivalent kinetic speed at sea level in aeronautical standard
atmosphere

VC Cruise Velocity

VD Dive Velocity

VS Stall Velocity
INTRODUCTION

In the process of certification of a civil aircraft, the calculation of the loads occupies an
essential place. Indeed, it connects the general specification - architectural - of the
aircraft to the structural dimensioning, the term being taken in the widest sense. The
calculation of structures is fundamentally conditioned by "classical" mechanical loads:
external aerodynamics (manoeuvre, gust, turbulence), internal flight loads
(pressurization, interfaces, thermal loads) and the ground (dynamic landing, taxiing).
The contemporary calculation also incorporates loads due to the integration of
systems with dynamic structures and effects of electric flight controls.

In this context, it is first of all essential to address the fundamental elements of the
regulation, resulting from a long experience of civil aviation, allowing to ensure a robust
calculation of the envelope of the loads in flight - defined by the flight range - allowing
a qualification of structures in static loading and fatigue related to lift. This is the
purpose of this chapter, which presents the general process and definitions related to
regulation. Starting from the general framework of civil certification, the stochastic
definition of limit and extreme loads leads to the concept of failsafe structure. The
parameters defining the flight range of an aircraft are reviewed, and lead to the
definition of the three standard domains, associated with manoeuvring and gusting
cases. These elements are decisive for the design of the primary structures of the
aircraft: aerofoils and interfaces with the rest of the cell.
1 General Definitions - Certification Context

1.1 Structural Certification

1.1.1 Regulatory process

The certification of an aircraft integrates an iterative process, accompanying the

program throughout the life of the product, from market studies, design loops
(feasibility, preliminary, detailed, optimization by loops) to its withdrawal, going
through the setting and monitoring in service.

This process directly involves three types of independent partners: the architect
(or aircraft manufacturer), the operators (or user companies) and the
certification authorities. These stakeholders assume complementary roles and
responsibilities that we do not detail here, the important thing being to locate
the calculations in this context, which results from the dynamics of quality which
implies close cooperation between these partners.

The initial type certification process usually takes 5 to 7 years (until

commissioning) and the following phases 20 to 30 years, including in-service
monitoring (the basics of which have naturally been laid when initial
certification, including traceability and conformity control) and even withdrawal
of service. It is therefore not unreasonable to say that the certification process
for a contemporary aircraft actually extends over a period of 20 to 50 years.

The whole process of structural certification is placed under the aegis of

regulatory texts. For the European continent, under EASA (European Aviation
Safety Agency), these texts are:

- CS-25 (Certification Specification Part 25) for civil aircraft with a

maximum take-off mass greater than 5700 kg (Large Aircraft, Maximum
Take-Off Mass MTOM ≥ 5.7 t);

- CS-23 for MTOM civil aircraft less than 5700 kg.

US certification is the responsibility of the Federal Aviation Administration,

which relies on very similar regulations (FAR-25 and FAR-23).

Finally, it is fundamental to specify that this procedure is not fixed (static) but
evolutive (dynamic); it results from permanent exchanges, feedbacks and
adjustments that guarantee both the quantitative value and the qualitative
robustness, these two assurances being essential.

1.1.2 Structural justification

When the loads are known - under a clean level of precision and robustness -
the structural elements subjected to these loads are justified according to 3
approaches:
 - The analytical approach first. Irreplaceable, it is the most fundamental
because even if its precision is relative (approximate model) the
robustness of the model is often the best possible. Indeed, in the context
of an analytical calculation, the assumptions and approximations are
clearly known, and a good engineer is precisely the one who will know
what can be expected from a simple but relevant model, particularly
valuable not only in pre-dimensioning (this is the main tool) but also in
actual validation;

- The numerical simulation then, precious to refine the model. It should

systematically be subjected to a very critical analysis, and qualitatively
confronted with the analytical results and quantitatively with the results
of tests;

- The material tests finally, on a small scale (validation by coupons, in

very large numbers), average scale (primary structural elements) and
complete (whole cell).

The latter category includes ground tests (static resistance test cell, fatigue test
cell, dynamic prototype identification) and flight tests (the purpose of which is
precisely to explore the flight envelope).

Flight tests usually involve four dedicated aircraft, including the aircraft 1 on
which will be included the initial defects covered by the certification
(delamination, delamination, impacts, wider tolerances on the assembly
dimensions) which will allow, in a process iterative, to define the certified
procedure for monitoring defects (derogations) guaranteeing, as part of the
certification, the nominal level of performance and safety of the aircraft
throughout its life.

It is this complete and integrated process that is certified, and not only the
aircraft itself. Note that among the test aircraft, some can be then marketed (as
"used aircraft") after a major visit, and any special specifications.

Full-cell ground tests typically use one of the test aircraft for the initial
identification of the natural modes of vibration (mode-by-mode harmonic tests,
modal superimposition white noise tests, transient load test). This test requires
a real and complete aircraft, including the equipment (and at least partially the
internal fluids) because the inertial and dissipative elements are necessarily
represented.

For the other full-cell ground tests, two complete aircraft structures - taken from
the assembly lines - are used which are not equipped with all the systems,
however. Indeed, for the static cell test and the cell fatigue test, which operate
in a quasi-static equivalent domain, the embedded systems constitute non-
structural ancillary elements not essential to the representativeness of the tests,
which moreover have a character damaging and destructive (unlike other tests
on the complete cell).
 In total, six complete cells are therefore dedicated to type certification tests: four
complete aircraft for flight tests and ground modal test, and two cells for
destructive certification tests.

1.2 General and local loads

1.2.1 Probabilities of Occurrence - Limit Loads

The life of an airplane represents 100,000 hours of flight (HD) in order of

magnitude. It will be more for a long haul (140000 HdV) and a little less for a
short haul (70000 HdV) but 105 is the right time gauge.

As a result, a relatively rare occurrence, but likely to occur once in the course
of the aircraft's life, may be affected by a probability of occurrence of 10-5 per
flight hour (10-5 / HdV). One can thus legally describe as Charge Limite (Limit
Load, LL) a load likely to occur once in the life of the plane.

The statistical analysis of all the possible loads and their probabilistic
combinations bases the load calculation, and is even a primordial element. This
predictive approach is continuously enriched by databases of flight incidents
that refine the models, both in terms of extrinsic factors to the aircraft (gusts,
turbulence) and intrinsic (maneuvers controlled manually or numerically,
interaction with the systems and their partial or total breakdowns).

It can be noted that the lifetime of the aircraft in flight hours - in the order of 10-
5 - is similar to the total operating time (hours) of a human being during his

career. And so the associated probability is actually related to events unlikely

but of the same order of magnitude that can happen during a human life. This
general definition is important to understand the rule that the certifier sets itself:
to build a system whose architecture and functioning guarantee a risk-taking
compatible with the "natural" risk of human life (risk linked to biology and
accidental risk in everyday life, excluding aeronautics). This will be even more
explicit for critical events, so much rarer than Limit Loads. This remark is placed
in the general context of this chapter, which explains the perfect and logical
coherence of the regulation - resulting from 60 years of civil aviation jet - with
the requirements of performance and safety - both inseparable - which qualify
all system intended for the public.

Indeed, if the Load Limit (LL) is the fundamental starting point of the definition
of the loads in flight, the regulation will naturally seek to widen this definition to
very different levels of probabilities. It focuses first on the costs arising from
events or situations "very unlikely" defined by regulation at 10-9 / HdV. It should
be noted that life-threatening events are far more frequent than this, since
meteorite falls must be reduced to 10-9 / hour (meaning that one inhabitant , in
any place on the surface of the Globe, runs every hour of its existence a risk of
the order of 10-9 to perish by direct impact of a meteorite). This risk is extremely
 low, and is considered for aeronautics as the limit of what should be considered
in the calculation of loads (and sizing of the system more generally).

The regulation imposes to consider the load at 10 -9 / HdV as Ultimate Load

(UL). At the same time, it requires to justify the macroscopic elastic resistance
structure at LL, and the macroscopic rupture resistance at UL, under a loading
applied for a duration of 3 seconds (duration allowing a quasi-static
representation).

For a relatively high probability load configuration - this term means that it can
occur once during the life of the aircraft, and thus covers any event with a
probability of occurrence equal to or greater than 10 -5 / HdV , and therefore in
particular the LL itself - the UL / LL ratio between Extreme Charge and Charge
Limite is statutorily set at 1.5 (3/2).

It should be noted that this ratio decreases when the probability of occurrence
is less than 10-5 / HdV, according to the following logarithmic law given in Figure
1 below:

Figure 1: Definition of the extreme load according to the probability of occurrence

A probability configuration of less than 10-9 / HdV is very unlikely and is not
considered. The limit case at 10-9 / HdV imposes a justification at break at LL,
which amounts to considering the ratio equal to 1. Moreover, a probability
situation greater than 10-5 / HdV is likely (probable event) because before occur
during the life of the aircraft, and in this case the regulatory ratio is 1.5. In
intermediate cases, a semi-logarithmic interpolation is imposed. This means
that a configuration whose probability of occurrence is 10 -7 / HdV induces a
ratio of 1.25.

1.2.2 Local and global models

In the design process, the external loads and the distributed loads will first be
represented on global models based on the theory of thin structures: slender
beams (straight or curved), plates and thin shells (developable or non-
developable). These calculations will therefore have for input the external loads
and for outputs the internal flows of circulation of the forces (elements of
stressors) heard in the current zone. This first calculation is quite fundamental
 because it is related to the overall design of the structure (structural
architecture).

It will then be up to local calculations to refine the local distribution of the

stresses, in the vicinity of particular points and in particular the accidents of
form: fixations, discontinuities, concentrated efforts, connections.

This analysis by separate scales is lawful, because of the St-Venant principle

which "smooths" the effects of local disturbances in distributed flows. The
principle of Saint-Venant - the fundamental principle of the mechanics of
deformable solids - in fact makes it possible to replace any concentrated load
given at a point P by another statically equivalent concentrated load, that is to
say of the same general resultant and even moment resulting in P. This
approximation is valid "far" from the point of application. The concept of
"remoteness" is quite simple to define in a thin structure of aerospace type
(beam or shell). It suffices to consider a distance at least equal to five times the
thickness of the structure (width and thickness of a beam, thickness of a shell).
The St-Venant principle therefore applies to most of the geometrical domain of
a thin structural element.

1.2.3 Hyperstatic concept (Fail-Safe Design)

In general, Class I structural elements, ie those whose ruin leads to a situation
of potential loss of the aircraft during flight (therefore considered an immediate
loss) will necessarily be designed in a hyperstatic configuration, that is
superabundant.

One example is a high-lift flap, which is aerodynamically part of the wing. It is a

rotating structure for which the axis of rotation must be held by three
independent joints. The geometric definition of a straight line in 3D requires only
two points of reference, the regulation requires three of them, which makes it
possible to lose an articulation (obvious failure) while keeping the definition of
the axis, thus a structure operational on the remaining 2 joints. This concept is
valid for any articulated primary structure. Figure 2 below shows the
configuration of a single landing gear door, for which is not redundant not only
the positioning of the axis of rotation (three joints) but also the locked structural
position. Indeed, the natural stop that constitutes the locking latch would be
enough to block the rotation, but we add an additional stop hyperstatic, which
also allows to slightly preload the hatch during closing (thus preventing any
static disruption or dynamic hammering).
 Figure 2: Train hatch on three joints

This concept of triple articulation poses however difficulties of alignment,

because if two points are obviously always aligned, the three points of
articulation will however require to define geometrical tolerances with respect
to the points of attachment on the caisson of wing.

A hyperstatic structure (Fail-Safe) can therefore be the seat of assembly or

assembly constraints, and will also be potentially sensitive to stresses of
thermoelastic origin, while a structure in isostatic configuration retains a
freedom of deformation that allows it to undergo thermal deformation without
constraints.

Primary aeronautical structures have been designed as fail-safe since the

Second World War, specifically since the bombers were designed to be able to
sustain significant damage in flight while being able to return from mission. This
concept of tolerance for major damage has been included in the regulation of
civil aircraft, with the following specification:

- in intact configuration (without obvious rupture) the structure is justified in

macroscopic elastic limit at LL as specified above, in major non-rupture (general
ruin or collapse) three seconds at UL, and for three lives minimum plane fatigue;

- in configuration with obvious rupture of an element (for example with one of

the three joints broken on a shutter, and the potential rupture of each of the
joints must be justified) the structure must be justified in non-major rupture three
seconds to LL. This justification - in which one could say that the LL replaces
the UL (but in degraded structural configuration - means that the plane always
holds its flight range (possibly leaving the elastic domain, unless we remain in
the envelope of SL) and is therefore able to return to the base, so for a civil
aircraft to complete the flight. In contrast, the subsequent life of the structural
part will no longer be justified, and therefore the element subject to obvious
rupture must be regulated by law. replaced the next flight.
There are, however, two notable exceptions to this rule, which are structural
elements whose design must by nature be isostatic:

- an undercarriage, a structural 3D element in a locked configuration - that is to

say on the ground and in flight off maneuver - but which must imperatively
become hypostatic (mechanism) during retraction and deployment; these
changes in configurations must - in the vast majority of cases, for reasons of
reliability of hydraulic systems - be made by disconnecting a single mechanical
connection at a time (latch or cylinder). This implies that the structural
configuration is isostatic, so that only one less link transforms the structural
element into a mechanism and thus allows the overall motion; isostaticity is
here justified (and almost inevitable) for mechanical reasons;

an engine suspension, subjected to significant thermal gradients, applied in

addition in a zone where a wide variety of materials mix (steel, titanium, light
alloys, composites), moreover in a dynamically charged environment in
vibroacoustics, shocks and even fluids ( flight controls, lubricants and fuels);
this suspension is then done by three points (defining exactly and isostatically
a laying plane, blocking the solid motor relative to the solid-built or solid-wing);
here, the isostaticity is justified for reasons of thermomechanical constraints
(the same one which forced the engineers to connect the space shuttle to its
cryogenic tank by only three points - two at the rear and one at the front - see
figure 3 below).

Figure 3: External Tank Fasteners (aft & forward attachment, doc. NASA)

These structural elements assembled in isostatic configuration are called Safe-

Life and are subject to two major obligations, because they are not
superabundant:

- the structural support points must each be considered as unique stress paths
(in the structural sense), and thus be locally fail-safe design; concretely,
 concentrated load recovery screeds are designed in double-screed, with a
nominal half-clevis and a half-screed arranged in parallel, pending (thus
mounted with a game relative to the axis); this cage in reserve is also not
subject to fatigue loads and will load automatically (after catching the game) if
the nominal screed is to give in fatigue.

The globally designed Safe-Life structural element shall be specified to have a

limited service life, significantly less than that of the airplane; in practice, this
translates into the regulatory replacement of all the landing gear structures after
a certain number of landing (and this regardless of the load spectrum of the
actual mission of the aircraft); for example, if the aircraft is sold for a life
expectancy of 90,000 flying hours, new landing gear will be fitted every 30,000
landing or touchdown, well before the end of the aircraft's life. This means that
Safe-Life structures are "consumable" structures like tires. But unlike tires,
which are replaced according to wear (usually after 500 to 1000 impacts) the
trains are automatically changed.

2 Bearing surfaces

2.1 External aerodynamic loads

2.1.1 Aerodynamic result

We first define the aircraft axes (structure axes or fuselage axes as follows (see
figure 4 below):

⃗⃗⃗
𝑋 longitudinal unitary vector fuselage, directed forward

⃗⃗⃗
𝑌 unit vector orthogonal to the previous one, located in the plane of
symmetry right left of the fuselage, and directed downwards

⃗⃗⃗
𝑍 ⃗⃗⃗ , 𝑌
then complete the direct orthonormal basis (𝑋 ⃗ , 𝑍), in the direction of
the right half-wing (see diagram below).

Figure 4: Aircraft axles linked to the fuselage

These axes are related to the plane (in particular they rotate with it) and must
obviously be distinguished from the axes related to the air mass and those
related to the Earth (considered to be approximately Galilean). This system of
axes is convenient to represent the aircraft as a dimensionally stable solid,
whose information is carried by this orthonormal reference useful in quasi-static
calculation of the structure.

The trajectory of the plane is defined by the movement of its center of mass G,
whereas its attitude is defined by the movement of the structure around G, so
⃗⃗⃗ , 𝑌
that the reference which interests us at the structural level and (𝑋 ⃗ , 𝑍). The
angular rotational speeds of the aircraft around these 3 axes are classically
noted:

𝑝 = 𝑟𝑜𝑙𝑙 𝑟𝑎𝑡𝑒 (𝑎𝑟𝑜𝑢𝑛𝑑 𝐺𝑋)

{𝑞 = 𝑝𝑖𝑡𝑐ℎ 𝑟𝑎𝑡𝑒 (𝑎𝑟𝑜𝑢𝑛𝑑 𝐺𝑌⃗)
𝑟 = 𝑦𝑎𝑤 𝑟𝑎𝑡𝑒 (𝑎𝑟𝑜𝑢𝑛𝑑 𝐺𝑍)

The pseudo-vector rotation speed refers to the aerodynamic reference or Earth

(absolute speed of rotation or relative to the air mass) but is vectorially
decomposed (p, q, r) in the axes related to the aircraft.

The Galilean resultant of the external forces applied to the aircraft (reduced to
G) is decomposed into a resultant of the distance forces and the surface forces.
Here, the forces at distance (or volume) are reduced to the weight, the
electromagnetic forces being neglected. Surface (or contact) forces are
reduced to aerodynamic pressures (aerostatic lift is structurally
nondimensional) to which must be added point contact of the undercarriages
with the ground, when the plane rolls. In this paragraph, we limit ourselves to
the case where the airplane is in flight.

Weight of the plane in flight

The resultant 𝑃⃗ = 𝑚𝑔 is reduced and examined in G.

m is the mass of the airplane in flight, nominally OEM ≤ m ≤ MTOM (the

minimum mass in operation is the OEM, Operating Empty Mass)

g is the local acceleration of the gravity directed according to the local

vertical (and not according to 𝑍 previous, related to the plane).

The load calculations consider the international value: g≅g0≡9.80665 m / s2.

In the vast majority of cases, this value is approximated: g≅9.81 m / s2.

It should be noted that the actual value applied locally on the aircraft is slightly
different. In fact, in static ground g varies from 9.79 (at the equator) to 9.83 m /
s2 (pole) and moreover the plane moves very slightly away from the center of
the Earth, and acquires a horizontal velocity which is not totally negligible
compared to the speed of orbitation. Despite this, the value 9.81 remains fully
compatible with load calculations, the actual accuracy of which can not exceed
2%, maneuver and burst being fundamentally dynamic phenomena.

It can also be noted that some of the load files are established with the
approximation g≈10 m / s2, which identifies kgf (kilogram force) at daN. The
danger here is that the error becomes systematic, but it is considered that this
approximation remains admissible.

Aerodynamic result

The aircraft advancing in the atmosphere induces contact forces whose

resultant is decomposed according to X, Y, Z. For the sake of simplification, we
will work in the plane XZ (symmetrical flight), but the definitions obviously
extend to three-dimensional case integrating Y.

The aerodynamic result is a first approximation proportional to the square of the

relative speed V of the aircraft relative to the local air mass. In reality, there is
a V-shaped viscous term, but it is limited to a thin layer where the speed
gradients require a calculation in viscous fluid, called precisely boundary layer,
adhering to the wall.

For an airplane, this viscous effect is second-rate. Indeed, if a force F expresses

F = AV + BV2 (viscous term and kinetic perfect fluid), when the velocity
increases, the term in V2 becomes systematically dominant compared to the
viscous term.

This approximation valid in the mid and high subsonic, clearly differentiates the
dynamics of the aircraft from that of the automobile, which evolves much more
slowly. Another difference comes from the fact that the aircraft is essentially
looking for a high lift (more exactly a great fineness, lift / drag ratio) while the
automobile is looking for a drag ... and a lift as low as possible.

The aircraft undergoes, during its progression in symmetrical flight, the

decomposed aerodynamic resultant as follows:

𝐹𝑍 = −𝑝𝑘𝑖𝑛𝑒𝑡𝑖𝑐 𝑆𝑟𝑒𝑓 𝐶𝑍 ∶ 𝑙𝑖𝑓𝑡

{
𝐹𝑋 = −𝑝𝑘𝑖𝑛𝑒𝑡𝑖𝑐 𝑆𝑟𝑒𝑓 𝐶𝑋 ∶ 𝑑𝑟𝑎𝑔

1
The kinetic pressure 𝑝𝑘𝑖𝑛𝑒𝑡𝑖𝑐 = 2 𝜌𝑉 2 is induced by the speed V of the aircraft
relative to the air mass. The sign - of the drag simply comes from the orientation
of the X axis towards the front of the plane, and from the fact that it is supposed
that the plane progresses according to + X, and therefore that the component
on X of the relative speed is negative. The sign - of the lift comes from the fact
that it is supposed that the lift is nominally directed towards the extrados
(moreover induced by the bend of the air on the upper surface ...).

This kinetic pressure depends on the effective density ρ of the air surrounding
the aircraft. In fact, the Badin (Pitot tube attached to the aircraft) measures the
kinetic pressure, but does not take into account the external density (because
it is calibrated on the ground). Thus, it is customary to reduce the actual speed
of the aircraft relative to the air mass in flight at its equivalent speed V0 under
standard conditions:

1 1
𝑝𝑘𝑖𝑛𝑒𝑡𝑖𝑐 = 𝜌𝑉 2 = 𝜌0 𝑉02
2 2
where ρ0≅1.225 kg / m3 (conventional air density in standard conditions, 15 ° C at
sea level).

In practice, it is this speed (which is also that of the speedometer on the

dashboard of the pilot) that generates the structural loads, not the actual speed
of the aircraft. It is therefore she who will be considered for the calculation of
charges.

In practice, it is this speed (which is also that of the speedometer on the

dashboard of the pilot) that generates the structural loads, not the actual speed
of the aircraft. It is therefore she who will be considered for the calculation of
charges.

The reference surface makes it possible to go from the kinetic pressure to a

reference force. For an airplane, consider the area of the airfoil (wing) including
the part passing through the fuselage, orthogonal projection to Z. The diagram
below (Figure 5) explains this definition which includes the case - rare but
possible - an asymmetrical wing:

Figure 5: Reference Surface for Aircraft Aerodynamics

Only the bearing surface is taken into account. As the CS25 imposes
longitudinal dynamic stability of the aircraft, the center of mass is necessarily in
front of the main focus, and therefore it is necessary to have a plane deporting
to the rear in normal flight to balance the pitch.

The horizontal empennage is therefore not a carrier plane, and therefore does
not fall within the definition of the reference surface, as shown in the diagram.

The coefficients Cz and Cx thus make it possible respectively to define the lift
and the drag with respect to an aerodynamic reference, and therefore
adimensionally. It should be noted, consistent with the previous remark, that for
a terrestrial vehicle, the reference surface will be the master torque, and thus
the total surface this time, and viewed orthogonally to X.
The resulting moment applied around G, denoted (L, M, N) by decomposition
on the XYZ axes, must also be able to be expressed adimensionally. For this,
we must define a reference lever, linked to the reference surface. For an
airplane, we consider the average rope b of the wing, defined from the total
span e:

𝑆𝑟𝑒𝑓
𝑏≡
𝑒
Therefore, the moment of pitching is expressed: M = pkineticSrefbCM

It can thus be seen that the torque of the aerodynamic forces applied to the
aircraft at G can be expressed by 6 coefficients in total (obtained by fluid
calculation or wind tunnel test): CX, CY, CZ; CL, CM, CN. These are the
coefficients relating to the complete aircraft.

In the case of symmetrical flight (GXZ plane, pitch rotation around GY), 3
coefficients are necessary: CZ, CX; CM.

In the established flight domain, the value of Cz evolves with the incidence α;
the incidence being defined as the angle formed by -X and the projection of the
relative wind speed in the XZ plane.

If one linearizes this evolution - which is quite realistic for a conventional plane
- one can thus approximate Cz by an affine function, in which (∂Cz / ∂α) appears
like an additional aerodynamic coefficient related to the incidence:

𝜕𝐶𝑍
𝐶𝑍 ≅ 𝐶𝑍0 + 𝛼
𝜕𝛼

If a movable surface is added on the wing (high lift flap for example) and if δ is
noted its steering angle downwards relative to the wing (δ = 0 if the wing is not
modified, flap retracted) then can generalize this linearization, which leads to a
new coefficient related to this moving surface (or rudder):

𝜕𝐶𝑍 𝜕𝐶𝑍
𝐶𝑍 ≅ 𝐶𝑍0 + 𝛼+ 𝛿
𝜕𝛼 𝜕𝛿

Thus, we can roughly represent the lift variation by a straight line, this one being
limited by the stall appearing at a certain incidence (at which the Coanda effect
on the upper surface is no longer sufficient compared to the dynamic low). The
presence of high lift surfaces (single and slit nozzles and flaps) delays the
appearance of the stall (see Figure 6 below).
 Figure 6: Lift Curve

Finally, these linearized equations are transposable for the variation of the
lateral lift coefficient CY as a function of β, skid and rudder deflection. As for
the wing, the fact of staying in incidences below the stall allows a good
approximation by linearization.

2.1.2 Load factor

Consider the lift, which is the main load that concerns the wing structure since
the drag is very much lower (since the fineness is relatively high).

This ratio also appears very clearly on the polar (Figure 7). This curve, which
accurately represents the lift coefficient CZ as a function of the drag coefficient
CX, is graduated in incidence (represented as curvilinear abscissa along the
curve).

Figure 7: Typical aerodynamic polar

For the curve to be balanced, it is necessary to dilate the scale of C X by a

coefficient 10, precisely because of the important value of the lift,

It is now essential to represent the lift - a major measure of structural wing load
- by a dimensionless number of dynamic nature (while the CZ is aerodynamic).
This will be the load factor nZ. It is defined simply in relation to the conventional
weight of the airplane in flight:
 𝐹𝑍
𝑛𝑍 ≡
𝑚𝑔0

This load factor represents the "apparent gravity on board the aircraft, along
the Z axis". Indeed, what is called gravity actually results from contact forces
balancing the forces of volume of gravity.

The load factor nZ reflecting the surface (aerodynamic) forces according to Z, it

corresponds very exactly to the level of gravity according to Z suffered on board
the aircraft, expressed in g0.

Thus, in horizontal stabilized cruising nZ = 1. In parabolic flight (ballistic

trajectory, marrying very exactly that of an orbit, that is to say a free fall in the
vacuum under the effect of gravity alone) nZ = 0. In horizontal flight stabilized
on the back nZ = -1.

Dimensional load factors (nZ≥1) are induced by manoeuvres and vertical gusts.

2.1.3 Manoeuvre and vertical gust

There may be mentioned two symmetrical or quasi-symmetrical manoeuvres
generating such significant load factors.

The first manoeuvre considered is the symmetrical resource, which we can

represent at its extremal point by the osculating circle at the trajectory (the real
trajectory at constant nZ being of parabolic form) this vertical circle having for
radius R, according to the diagram presented in figure 8 below.

Figure 8: Trajectory in symmetrical pull

The plane is represented by a material point of mass m animated with a

horizontal velocity, so that at this moment its local axis Z coincides practically
with the vertical; the resultant of the applied forces according to Z is
algebraically FZ-mg0, that is (nZ-1) mg0.

According to the fundamental principle of the dynamics applied in the terrestrial

frame supposed Galilean, one thus has:
 𝑉2 𝑉2
(𝑛𝑍 − 1)𝑚𝑔0 = 𝑚 ⇒ 𝑛𝑍 = 1 +
𝑅 𝑔0 𝑅

In this manoeuvre, the load factor is therefore directly related to the speed and
the curvature 1 / R of the trajectory.

We find in particular:

- if the trajectory is flat (cruise, 1 / R = 0) then nZ = 1;

1 𝑔
- if nZ = 0 (ballistic flight) then 𝑅 = − 𝑉02 (curvature in parabolic flight)

In addition, in this symmetrical manoeuvre, as the aircraft remains tangent to

its trajectory, the pitch rate (Galilean rotation speed of the aircraft around GY)
is identical to the angular velocity of the circular motion, namely:

𝑉
𝑞𝑝𝑢𝑙𝑙 =
𝑅

From the previous expression of the load factor, we immediately deduce the
pitch rate induced by the symmetric load factor:

𝑔0
𝑞𝑝𝑢𝑙𝑙 = (𝑛𝑍𝑝𝑢𝑙𝑙 − 1)
𝑉

The second manoeuvre considered relative to the load factor nZ is the stabilized
correct bend (pseudo-symmetric); the manoeuvre takes place at constant
altitude and under inclination θ (Figure 9 below).

Figure 9: Correct Turn Load Factor

1
The vertical equilibrium gives immediately: 𝑛𝑍 = cos 𝜃

The pitch rate induced by the turn is:

1 𝑔0
𝑞𝑡𝑢𝑟𝑛 = (𝑛𝑍𝑡𝑢𝑟𝑛 − )
𝑛𝑍𝑡𝑢𝑟𝑛 𝑉

For vertical gusts, the isolated burst of continuous turbulence is distinguished.

The isolated vertical burst can be considered as an area of local atmospheric
descent (or descent), which the aircraft traverses. It is modelled by a V vertic
vertical velocity profile, generally assimilated to a crenellation or semi-
sinusoidal distribution, characterized by its length λ (in m) and its amplitude
ΔVvertic. (in m / s) according to the schematic models presented in figure 10.

Figure 10 : Profile-types of vertical gusts

The half-sinusoidal model is relatively realistic, and commonly used to

represent the gust as a pulse (Dirac).

Several aircraft models can be applied to this gust, here are the two
Main:

- the rigid airplane model, in which the transient increase ΔN Z of the nZ

is due exclusively to the temporary increment of incidence, in the ratio
Vvertic/V. This model, extremely simple, has the disadvantage of being
conservative, and therefore systematically increase the load factor
obtained;

- a softened aircraft model obtained by a coefficient of reduction (due

mainly to the deformation of the wing) much more realistic.

Extensive analyzes have been conducted for more than 50 years to determine
the statistical distributions of gusts in the atmosphere. These studies
incorporate results from in-flight measurements, tracer observations
(stratospheric balloons) and anemometric measurements in statistical tools.
They converge, for the probability of 10-5 / HdV and in the whole aeronautical
atmosphere, towards the load factor limit values which will be presented in
paragraph 2.2.1.

Continuous turbulence in flight has been modeled in detail by Théodore von

Karman, in relation to the measurements made in flight. Stochastically, the
distribution of turbulence ΔV in the Earth's atmosphere is almost uniform in
cruising altitudes of civil jet aircraft, ie between 25,000 ft and 40,000 ft (FL 250
to FL 400). This qualitative result is, in itself, remarkable. Indeed, the
thermodynamic state of the atmosphere (for example at 37000 ft, that is to say
 11300 m, T = -56.5 ° C, p≈217 hPa) is homogeneous at the scale of the globe,
that the one is in winter or in summer, above the equator or a pole, a continent
or an ocean. This will make it possible to determine fatigue load spectra that
will be valid irrespective of the operation areas, and which will depend only on
the operational altitudes of the aircraft considered.

Quantitatively, the lengths (spatial periods) and amplitudes (ΔV) of the turbulent
granularity of the atmosphere will allow - as for the isolated gust - to define a
statistical frequency spectrum of load factors related to atmospheric turbulence
in flight, in the form of a Spectral Density of Power W (f) (DSP, expressed in g02
/ Hz).

It should be noted that this spectrum is obtained by applying an airplane model,

and transforms the atmospheric representation of von Karman (ΔV / spatial
periods) into a representation nZ / temporal frequencies. The load is then
defined stochastically by a function W (f) scanning frequencies, arbitrarily
limited to the overall structural frequencies of the aircraft (therefore generally
between 1 and 100 Hz). The term "global structural frequencies" means the
frequencies corresponding to eigen modes associated with significant modal
masses (greater than 1% of the mass of the aircraft). In this way, general aircraft
loads are excluded from very local modes, such as those of an auxiliary
antenna which can vibrate independently and which will have to be classified
separately, in particular for the holding of the mounting interfaces.

Finally, the spectrum of nZ associated with turbulence can be associated with

an equivalent input static load, which is computed as follows by statistical
equivalence, for a given frequency f:

𝑛𝑍𝑒𝑞𝑢𝑖𝑣 (𝑊, 𝑓) = 𝛼𝑠𝑡 √𝑓. 𝑊(𝑓)

where αst is a coefficient associated with the confidence level of the calculation:

𝛼 ≅ 3.76 𝑓𝑜𝑟 𝑎 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 "3𝜎" 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 99.7%

{ 𝑠𝑡
𝛼𝑠𝑡 ≅ 2.72 𝑓𝑜𝑟 𝑎 "2𝜎" 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑜𝑛 95% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒

In civil aviation, certification generally requires a "3σ" calculation (3 standard

deviations). For the dimensioning of a space vehicle, we consider rather the
qualification to "2σ". The spacecraft is not certified but qualified, and will
therefore be the subject of a specific recipe by the operator, whose level of
structural risk is an order of magnitude lower than that of a type certification.

2.2 Flight domains

2.2.1 Limit domain

The flight range - in which the airplane must remain in flight - is represented
graphically in the plane "nZ as a function of V0": load factor as a function of the
equivalent speed, see figure 11.
We will consider the domain in quasi-symmetrical flight.

Figure 11: Typical Flight Domain

Flight Domain in smooth flight

Let us first consider the "flaps retracted" domain, representing the airplane in
smooth flight - that is, flaps, spoilers, and retracted trains - and which is the
most important area for the architecture of the aircraft.

This domain is limited to the left by the stall speed (minimum physical V0 to be
in flight) which defines a parabolic boundary (since the kinetic pressure is in
V02).

This curve is not quite symmetrical with respect to the axis V 0 since the offset
(nZ <0) is less effective on an asymmetric wing profile (which is always the case
for civil transport aircraft, symmetrical profiles are rather the lot of aerobatic or
combat aircraft, designed to be able to fly on the back continuously).

The positive branch of this curve makes it possible to define the first
characteristic speed VS1 (Velocity Stall, nZ = 1), that is to say the stall speed in
smooth configuration under unit load factor (symmetrical horizontal stabilized
flight).

The domain is then limited in nZ values by high and low values:

- upwards by the positive limit value at 10-5 / HdV; in the context of CS25
(civil aircraft, with a maximum take-off weight MTOM ≥ 5700 kg) this load
factor is always between 2.5 ("large" aircraft) and 3.8 ("medium" aircraft)
and given in the cases intermediates by:

10886
𝑛𝑍𝐿𝐿+/𝐶𝑆25 = 2.1 +
𝑀𝑇𝑂𝑀 (𝑘𝑔) + 4536
 For a given aircraft, the value according to the above formula is
calculated, the result of which is valid if it lies in the range [2.5; 3.8].

In the case where the calculated value is outside the range, we take the
limit closest to the result of the calculation. Thus, for MTOM airplanes
greater than 22670 kg, the limit value is nZ = 2.5.

- downwards by the negative limit value equal -1 in the context of the

CS25. This does not mean that the airplane is certified to fly backwards,
but that it can structurally withstand (3 seconds, as well as the positive
limit load) a gust or manoeuvre downwards generating a load factor of
negative one (-1).

These limit values of nZ make it possible to define the second characteristic

equivalent speed, namely the stall speed VA under nZLL+.

This value immediately gives a third, really useful in approach; indeed, it is

necessary to take into account, for the stall, a possible rear horizontal gusts.
Indeed, the stall is dangerous at low altitude, so it is the typical horizontal burst
that should be considered. In practice, VB = VA + 17m / s.

The flight envelope is finally limited to the right, in equivalent speed, by the
maximum permissible speed VD (Dive) which can be defined as the maximum
dive speed allowed in service.

The VD is the diving limit speed corresponding to rare but nevertheless possible
situations in the life of the aircraft. It thus perfectly meets the definition of the
Limit Load (LL), and therefore of the limit domain. This is the speed adopted for
example in case of cruising depressurization.

The VD is significantly higher than the VNE (Never Exceed) which corresponds
to the red line on the pilot's speedometer, the VNE being itself naturally higher
than the maximum cruising speed VC (Cruise). It should be noted that between
VC and VD, the regulation admits a progressive non-accumulation with a
negative load factor. Thus, the smooth plane flight domain is dehorned lower
right, as shown in the previous diagram.

The limit domain is therefore completed; the point representing the

configuration of the airplane in flight may and must remain within the curve
throughout its flight duration.

Flight Domain in flaps deployed

This flight range covers take-off and landing configurations, and is most often
referred to as "Extended Layers, Spouts, and Spacers". It is fundamental for
the calculation of loads on moving surfaces, hatches and undercarriages in
flight.

Given the high-lift devices, the stall speed is significantly reduced (and therefore
the parabolic curve limiting the domain on the left is steeper, as can be seen in
 the diagram). Of course, this difference is only valid for the positive n Z, the high-
lift devices being by nature asymmetrical.

Moreover, since this area covers a short phase of the flight, and to a lesser
extent, CS25 admits that the nZLL+ is reduced to 2 in this case, which limits the
domain from above.

Finally, the speed will be limited much more severely than for the smooth plane,
the speed allowed in "extended" configuration; the speed thus delimited is the
VF (Flaps). We also observe that the VSE (Stall Extended) and the VF delimit the
white arc flaps released on the pilot speed indicator. We will thus have a smaller
flight area. The two boundary domains are schematically represented in the
previous diagram.

2.2.2 Extreme Domain

The extreme domain is extrapolated from the limit domain:

-the parabolic curve associated with the stall is simply prolonged, its
shape remaining unchanged;

- the upper and lower load factors are multiplied by 1.5;

- the maximum speed is increased by a factor 1.5 (pkin is multiplied by

1.5).

If we consider the smooth plane domain, we can give a representation of the

extreme domain according to the diagram presented in figure 12 below.

Figure 12: Limit domain and extreme domain (smooth flight)

2.2.3 Envelope of the Fatigue Domain

After considering the domain of the limit loads (LL, P ~ 10-5 / HdV) and extreme
(UL, P ~ 10-9 / HdV) the safe load area (SL) represents the envelope of the
standard flights (P ~ 1 / HdV), which constitutes a third domain included in the
two previous ones. This makes it possible to complete the scheme of smooth
aircraft loads, according to the diagram presented in figure 13 below.

Figure 13: Flight Areas with Safe Loads, Limits and Extremes

The field of fatigue in safe load is really an envelope, because the justification
to the 3 lives of the plane requires to simulate the history of the solicitations, all
included in this field. That's why it was represented in dotted green on the flying
field. The fatigue load cycles will all be included in this envelope (bounded to
the right by the maximum cruising speed and to the left by the stall curve) and
will represent all the loads sustained by the aircraft throughout its life.

Conclusion

The three flight domains thus defined constitute the reference for regulatory
dimensioning in civil aviation. They have the advantage of giving loads of static
type, while the phenomena at the origin of these loads are fundamentally
dynamic. They can therefore be supplemented by a dynamic parallel approach
of the "soft plane" type which will integrate the structural modes (vibrations and
response following a transient event) and the aeroelastic interactions. They are
nevertheless sufficient to qualify the structures and prepare the certification
campaigns, with primary tests in static and fatigue faced with the calculation
file.