Chapter 1

Introduction

The first wind turbines designed to produce electricity were realized in 1886 and 1887 (see Chapter 5). The
development of turbines has continued throughout the twentieth century, somewhat intermittently. The oil
crisis in 1973 revived the interest in wind turbines, and led to major opportunities of development starting
from the 1980’s. Over the past three decades, wind energy has undergone a very impressive growth: its
annual growth rate, averaged between 1990 and 2014, has been 24.8% (IEA 2015 report on renewables).
In numerous places in the European countryside (Figure 1.1), wind turbines are visible in the fields, raising
public debates on their impacts. Penetration of the wind energy sector is particularly strong in certain other
European countries, Germany and Denmark in particular (see Chapters 5 and 11). Offshore wind turbines
are also growing and developing, with many wind farms now operating in the North Sea for instance. In
essentially all cases, the turbines look alike: tall white structures with three blades upwind from the tower.
Yet wind energy has only barely started to account for a significant share in the ’production’1 of energy:
together, geothermal, solar and wind energies accounted for only 3.7% of world electricity production in
2013, and only 1.3% of the world total primary energy supply (IEA 2015 report on renewables).
The main motivation for the development of wind energy today is climate change, which constitutes an
unprecedented challenge for humankind. The agreement achieved by the COP21 meeting in Paris in 2015
officially aims at keeping the increase of global temperature below 2◦ C. Projections using climate models
show that global emissions of greenhouse gases (GHG) need to be reduced by 40 to 70% by 2050 in order
to achieve this goal (IPCC 5th report, 2013).
In a summary published in November 2015, prior to the COP21 meeting, the International Energy
Agency described the world energy outlook and expected energy trends to 2040:

• ’energy demand grows by nearly one-third between 2013 and 2040, with all of the net growth coming
from non-OECD countries and OECD demand ending 3% lower.’

• ’despite signs that a low-carbon transition is underway, energy-related CO2 emissions are projected
to be 16% higher by 2040.’

The above paragraphs provide a concise statement for the need and motivation to develop renewable
energy sources, and suggest that we are not yet taking the measures necessary to change our track consis-
tently with claimed objectives. A more recent report from the IEA (Perspectives for the energy transition,
2017) has investigated scenarios for the energy sector that would allow, with a probability of 2/3, to remain
below 2◦ C warming: ’this would require an energy transition of exceptional scope, depth and speed.’ The
concluding chapter will come back in more detail on this report. Let us retain, at this point, that renewables
are one of the most significant levers with which significant reductions of green house gas emissions can
be achieved.
Among renewables, wind energy is a mature technology, and already represents a significant share of
the electricity production in some countries. It will continue to grow, develop, and to require innovations
as new territories (in particular offshore installations) expand. A full description of wind energy requires
inputs from many disciplines (several branches of mechanics, electrical engineering, economics. . . ). The
1
energy is not ’produced’, it is converted from one form to another form, which is usable by humans. In the process, there
is unavoidably some loss.

1
 Figure 1.1: Turbines seen along the road from Aachen to Jülich, Germany, July 2018.

purpose of this course is to introduce the basics from mechanics to understand how wind turbines operate
and what drives the wind (the resource). Before delving into these topics, some preliminary considerations
are exposed below

• to quantify how much energy is used today (section 1.1), and in doing so to discuss some fundamentals
on energy, and in particular the units used for energy;

• to illustrate how wind energy has grown extraordinarily in the last decades to constitute a finite share
of the world energy supply (section 1.2),

• to document how energy consumption and production has changed in the most recent years (section
1.3).

1.1 Energy consumption

Energy is a fundamental concept in physics, and one difficulty comes from its relevance for a great variety
of phenomena, on a very wide range of scales. Correspondingly, there is a lengthy catalog of units that
are available for quantifying energy in different contexts. A first challenge is to become familiar with the
orders of magnitude and the units that are relevant for describing wind energy.

1.1.1 Units for energy

In the examples below, different units for energy and for power are presented; the illustrations naturally lead
to introduce different forms of energy. Different energy units are adapted to different scales and different
forms of energy. This multiplicity is part of the difficulty in developing our intuition regarding energy.
 • a Joule (J): it is the energy transferred to an object (or the work done) by a force of 1 N acting on
this object in the direction of its motion over a distance of 1 m. The Joule is part of the International
System of Units, and it is named after the English physicist, James Prescott Joule (1818-1889). 1J
can also be written 1 kg m s−2 .

• a Watt (W): it is the power corresponding to an energy of 1 J per second: 1 W = 1 J s−1 . It is

named after the Scottish engineer James Watt (1736-1819), who brought crucial improvements to
the steam engine, which were major contributions to the industrial revolution.

• Calories: the calorie was defined by Nicolas Clément (1779-1841) in 1824, as a unit for heat. There
are in fact several definitions, falling into two broad families. The small calorie (cal), or gram calorie,
is the amount of heat necessary to raise the temperature of 1 g of water by 1 degree Celsius at
standard pressure. The large calorie or kilogram calorie (Cal or kcal) is the amount of heat necessary
to raise the temperature of 1 kg of water by 1 degree Celsius. This is not unit of the International
System, but it is still widely used in nutrition.

• ton of oil equivalent (toe): this is the amount of energy released by burning one tonne of crude oil.
Different crude oils have different calorific values, so this may vary. It is defined by the IEA as 41.868
GJ. One can retain that it is approximately 42 GJ.

The total world annual energy consumption in 2010 was estimated2 to be 12002.4 Mtoe (1 Mtoe = 106
toe). According to the Key World Statistics of the IEA (2017 report), the World Total Primary Energy
Supply in 2015 was 13 647 Mtoe.

1.1.2 Energy consumption per capita

Following the suggestion of David Mackay’s remarkable book (Sustainable energy - without the hot air’3 ,
it is worthwhile trying to have numbers to quantify energy that bear some meaning relative to our everyday
experience. We will follow his suggestion and use kWh (kilowatt-hour) as the energy unit. It will often be
more relevant to discuss an energy used (or extracted) over a certain time, hence a power. A meaningful
time unit relative to our intuition and to consumption patterns is the day, so that the unit for power will
be the kWh/d (kilowatt-hour per day).

1.2 Wind energy today

Wind energy has been growing spectacularly over the past two decades. The present section provides a
few figures to both quantify this growth, and put it in perspective.

1.2.1 Growth of wind energy in recent decades

Report from the International Energy Agency titled ’Renewables Information’ of 2015: from 1990 to 2013,
the production of wind energy has gown annually at an average rate of 25%. For comparison, solar energy
has grown annually at an average rate of 47% over the same period.
Figure 1.2 shows the time evolution of the cumulative installed wind capacity. In 2017, global wind
energy installed capacity has gone over 500 GW. It is essential to keep in mind that this is installed capacity,
and that the intermittency of the wind will imply a charge factor that is typically of 20 to 30%. The curve
describing this growth suggests an exponential growth. The evolution of the new capacity installed each
year is shown in figure 1.3. This annual installed capacity indeed grows every year until 2009, again
suggestive of an exponential growth, but since 2009 it varies from year to year between roughly 40 and 60
GW.
2
Statistical review of world energy 2011, BP.
3
available online: https://www.withouthotair.com/
Figure 1.2: Evolution of the cumulative wind capacity over the past two decades, following the 2017 annual
report of the Global Wind Energy Council [18].

Figure 1.3: Evolution of the cumulative wind capacity over the past two decades, following the 2017 annual
report of the Global Wind Energy Council [18].

1.2.2 Present share of wind energy

The preceding section describes an amazing growth, but this in part reflects the absence of wind energy prior
to the 1990’s. According to the IEA’s 2015 report (Renewables Information), renewable energies accounted
for 13.5% of the world total primary energy (TPE) supply. However, this is mostly accounted for by biofuels
and waste (73.4% of renewables) and by hydroelectricity (17.8% of renewables). Geothermal accounted for
3.6% of renewables and wind for 3.0%. Although in five years these figures have likely changed, it remains
that wind energy remains a minor contribution to the world TPE. Its rapid growth raises challenges in
terms of energy transport and management.

1.3 Energy and CO2 emissions in 2018

The present section documents the changes in CO2 emissions and in energy production in 2018, based on
the report from the International Energy Agency (Global Energy and CO2 Emissions report, 2018, IEA).
The figures presented reaffirm the urgency to increase substantially the share of renewables in the energy
mix, and sets the stage for a more general discussion of the energy transition in chapter 11.
1.3.1 Growth in energy demand and in emissions
The global energy demand has been continuously growing in recent years (Figure 1.5). After a slowdown
in 2015 and 2016, growth has picked up again in 2017 and 2018. In 2018, energy consumption grew by 2.3
%, nearly twice the average rate of growth since 20104 . This has been driven by a robust global economy,
as well as enhanced heating and cooling needs in some parts of the world.
The demand in energy worldwide grows faster than what has been offset by strategies to decrease
Greenhouse Gas (GHG) emissions, resulting in a 1.7 % increase in energy related CO2 emissions which
reached 33.1 GtCO2 (Figure 1.6). Almost a third (more than 10 GtCO2) of emissions results from coal
use, mostly in Asia. The major part of the net increase in emissions is accounted for by China, India and
the United States.
Energy-related emissions had also increased in 2017, but had stagnated or slightly decreased for the
previous two years (Figure 1.7). This results from several factors: economic growth, as mentionned above,
but also enhanecd needs for heating and cooling in particular in the United States and slower gains in
energy efficiency (Figure 1.8).
This increase Given the economic growth, the increase in global emissions would have been significantly
greater than observed without improvements in energy efficiency, the development of renewables and other
practices for decrbonizing energy (nuclear, fuel switching from coal to gas. . . ). Figure 1.6 displays the
share of each in the avoided emissions in 2018.

1.3.2 Evolution of the electricity generation mix

The electricity generation mix in 2018 is depicted in Figure 1.9. Despite the formidable growth of renewables
(section 1.2), the mix evolves only slowly because of the amounts of energy involved, because of continued
growth of generation from fossil fuels, and of the unavoidable inertia of the sector: the investments involved
are colossal and energy-related assets pay for themselves on timescales of decades.
The electricity demande worldwide has grown by 4 % in 2018, nearly twice the rate of growth for the
global energy demand. A significant part of the growth in electricity demand has been met by the growth of
renewables. The growth of renewable electricity generation in 2018 by region and by technology is depicted
in Figure 1.10. The growth is well balanced between solar, wind and hydropower, each accounting for
about 30 %. Regionally, the growth of renewables has been led by China and Europe.
Nonetheless, this growth in renewable elecrticity, complemented by some growth of nuclear generation of
electricity, has not sufficed to account for the 4 % growth of the electricty demand. Figure 1.11 shows how
growth in coal and gas generation of electricity has complemented the growth in renewables, accounting
for a 2.5 % of energy related emissions for 2018.

1.4 Scope and organization of this course

This course aims at providing an introduction to the mechanics needed to understand the basics of wind
turbines, but also of winds and our current modelling and forecasting capacity of winds. The main funda-
mental notions of fluid mechanics necessary for this are reviewed in chapter 2. Chapter 3 then introduces
the aerodynamics of wind turbines, from simpler to more complex models. Aerodynamics is essential
to understand how winds turbine work, and to guide their optimal design. Considerations of structural
mechanics will however be central to design appropriately machines that are robust enough to last and
produce electricity for many years. Chapter 4 prodivdes some elements to understand the considerations
on structural mechanics relevant for wind turbines. The constraints and guiding principles for the design of
wind turbines are then sketched in chapter 5, along with a concise historical account on the development
of wind turbines. This brings to light in particular some instructive errors.
In order to sketch the fundamentals of atmospheric dynamics and to understand what drives the wind
and its variability, additional notions and results of fluid dynamics need to be introduced (chapter 6). In
particular, the notion of vorticity is introduced, which is central to an alternative description of the flow
around wings or blades. The aerodynamics of wind turbines is hence revisited in chapter 7. Ultimately,
4
According to the BP Statistical Review of World Energy, 2019, the growth has been 2.9 %. This is mentionned simply to
recall the difficulty and uncertainty in estimating precisely these quantities, despite their globality.
the source of energy of course is the wind: two chapters are devoted to the description of winds and the
modeling of these winds: chapter 8 winds near the surface, means of measuring them and simple models to
approximate their vertical variations. It ends with a case study illustrating from observations the variability
of surface winds and its relation to the larger-scale flow above. Chapter 8 moves on to describe the main
balances which constrain the evolution of the atmosphere, focusing on mid-latitudes. It aims at giving a
flavour of some of the ingredients of dynamic meteorology, the general circulation of the atmosphere, some
of the reasons for its variability, and describing the components of the models that provide us today with
forecasts of winds, on timescales from hours to seasons. Building on these two chapters of atmospheric
science, chapter 10 explores the wind as a resource for energy, from the local to the global scale, explaining
the different sources of information that can be tapped for resource estimation.
Finally, chapter 11 attempts to bring these different threads together, and to give an overview of wind
energy in the context of the energy transition called for by the Paris agreement of 2015.

1.A Appendix: Resources

Online resources that are relevant for an assessment of the share and growth of wind energy and renewables
in general:

• International Energy Agency (IEA)5 , providing reference data and analyses on energy issues worldwide.
It is also committed to exploring scenarios for energy compatible with climate change.

• Global Wind Energy Council (GWEC)6 , presenting data and reports on wind energy, from the point
of view of the wind energy industry.

• Wind Europe, formerly European Wind Energy Association (EWEA)7 , providing data and analyses
from the wind energy sector for Europe.

5
https://www.iea.org/
6
https://gwec.net/
7
https://gwec.net/
Figure 1.4: Distribution of installed wind capacity in December 2017, following the annual report of the
Global Wind Energy Council.
Figure 1.5: Growth of the annual energy demand worldwide in recent years. For 2018, the share of different
fuels is indicated by colors.

Figure 1.6: Growth of energy related emissions of CO2, to 33.1 GtCO2 in 2018, a historic high.
Figure 1.7: Annual change in energy-related emissions in recent years, from the IEA Global Energy and
CO2 Status Report.

Figure 1.8: Annual average change in energy intensity for recent years, from the IEA Global Energy and
CO2 Status Report. Also shown as the last bar is the target proposed by IEA’s Sustainable Development
Scenario.
 Figure 1.9: Electricity generation mix in 2018, from the IEA Global Energy and CO2 Status Report.

Figure 1.10: Growth of renewable electricity generation in 2018, from the IEA Global Energy and CO2
Status Report: by region (left) and by technology (right).
Figure 1.11: Change in electricty generation in 2018, from the IEA Global Energy and CO2 Status Report:
although renewables and nuclear accounted for more than half of the increase in electricity generation,
generation from coal- and gas-fired plants also increased considerably.
Chapter 2

Preliminaries in fluid mechanics

This is not intended as a course in fluid mechanics, but rather as a concise reminder or introduction of
the main notions that will be necessary to describe the aerodynamics of turbines. We will use notations
that are more common in textbooks: for example, boldface u for a vector. For furhter reading on fluid
mechanics, readers are referred to the clear textbook of P. Kundu1 , or the more concise text of P. Bernard2 .

2.1 Navier-Stokes equations

Although the fundamental laws of mechanics were laid down by Isaac Newton at the end of the seventeenth
century, more than a century was required to obtain the proper form for the equations of motion of a fluid.
A major difficulty has been to understand and express mathematically how adjacent regions of the fluid
act on each other. A minor difficulty concerns the derivative in time. Newton’s laws were expressed for a
given system. For a fluid, this would require to follow fluid ’parcels’. This is called Lagrangian description
of a fluid: u(a, b, c, t) designates the velocity of the fluid parcel which initially (at a reference time t0 ) was
located at position (a, b, c). The Lagrangian description generally becomes quickly untractable because
of the complexity of fluid. Much more convenient is the Eulerian description of a fluid, where u(x, y, z, t)
corresponds to the velocity at time t of the fluid at position (x, y, z). However, if one considers the
variation in time of a property F (x, y, z, t) of the fluid which is at (x, y, z) at time t, it can not simply be
the partial derivative in time of that property3 . If one follows the fluid parcel, the rate of change of F is
DF ∂F dx ∂F dy ∂F dz ∂F
= + + + ,
Dt ∂t dt ∂x dt ∂y dt ∂z
∂F ∂F ∂F ∂F
= +u +v +w , (2.1)
∂t ∂x ∂y ∂z
where we have introduced a new notation (D/Dt is the Lagrangian derivative, or time derivative following
the fluid motion) and we have simply applied the chain rule since the position (x(t), y(t), z(t)) of the fluid
parcel depends on time (see [Bat67]).
For a Newtonian fluid, the Navier-Stokes equations can be written:
Du
ρ = ρ f − ∇p + div τ (2.2)
Dt
where τ is that part of the stress tensor not related to pressure: τ = σ − pI, with I the identity tensor. It
is also called the viscous stress tensor.
Continuity equation (conservation of mass) can be written:
∂ρ
+ div(ρ u) = 0 (2.3)
∂t
1
Fluid Mechanics, P.K. Kundu, I.R. Cohen and D.R. Dowling, Academic Press
2
Fluid Dynamics, P.S. Bernard, 2015, Cambridge University Press, 252pp.
3
Think for instance of water flowing from the tap in a quiet, steady, laminar flow. The water is accelerated by gravity,
its velocity increases down, as evidenced by the narrowing of the flow. It accelerates downward. However, the partial time
derivative of its velocity is zero because the flow is steady.

13
or alternatively
Dρ
+ ρ div(u) = 0. (2.4)
Dt
Air is compressible of course. For ’weak velocities’, it can nonetheless be treated as incompressible. By
’weak velocities’, we mean velocities that are small relative to the speed of sound. This is quantified by
the Mach number:
U
Ma = (2.5)
cs
where cs is the speed of sound. Consideration of some thermodynamics andpassuming that air is an ideal
gas allows to obtain the following expression for the speed of sound: cs = γ P/ρ, where γ = 1.4, P is
pressure and ρ is density.

2.2 A little kinematics

Contrast between Lagrangian and Eulerian description of the flow:
• Eulerian: spatial coordinates describe a fixed position in space, at which fluid is passing;
• Lagrangian: spatial coordinates (labels) identify a fluid parcel by its original position in a reference
configuration at an initial time.
The Eulerian viewpoint is generally considered and is of much greater simplicity for practical implemen-
tation. Nonetheless, the Lagrangian viewpoint is very relevant, eg for transport, and is used for specific
applications.

2.2.1 Different families of lines

Trajectories
Trajectories correspond to the successive positions of a given ’fluid parcel’:

dx
= u(x, y, z, t) , (2.6)
dt
dy
= v(x, y, z, t) , (2.7)
dt
dz
= w(x, y, z, t) . (2.8)
dt

Streamlines
Streamlines are lines which, at a given time, are everywhere parallel to the velocity field:

dl u = 0 , (2.9)
at every point along the curve, where dl is a line element along the curve (along the streamline).

Streaklines
Position of all fluid particles that have passed through a given point in the past: trace left by injecting dye
(or a pollutant) from a given location in the flow.

Confusion between these different families of lines

It is difficult to conceive truly time-dependent, three dimensional flows, and we more easily imagine and
think about simple stationary flows. However, stationary flows are degenerate as the three families of lines
then coincide. It hence requires some effort to clearly distinguish between these three families of lines in
more complex flows. Thinking about the likely aspect of these lines in a flow around a turbine is a useful
exercise.
2.2.2 Streamfunction
Although the flow around a wind turbine will be truly three-dimensional and complex, our fundamental
understanding and modelling is based on consideration of two-dimensional flows: locally, the flow around
a blade (far enough from the hub) can be approximately and usefully described as two-dimensional. It
is hence essential to consider two-dimensional flows around an obstacle. A fundamental tool to describe
these flows is the streamfunction.
For two-dimensional incompressible flow, the continuity relation
∂u ∂v
+ = 0,
∂x ∂y
provides a diagnostic relation between the two scalar variables u and v. It is convenient to summarize the
flow with just one scalar variable, a streamfunction ψ(x, y, t) such that
∂ψ ∂ψ
u=− , v= . (2.10)
∂y ∂x
Note that the streamfunction is defined up to a constant. (Only the derivatives matter.) Note also
that the streamfunction is more generally defined with the opposite sign convention (ie u = ∂y ψ and
v = −∂x ψ). This is somewhat arbitrary, and the above choice is motivated by geophysical applications to
be considered later.
The streamfunction has the advantage of summarizing the flow with
An interesting property of the streamfunction is to provide information on the volume flux through a
given surface. Assuming a unit length in the z direction in all the calculations that follow, let’s consider
the mass flux through a curve C which goes from A to B:
Z
FC = ρu (k × dl) , (2.11)
C

where k is the unit vector in the z direction, × denotes the vectorial product, dl is an element vector along
the curve C which goes from A to B. Hence (k × dl) is normal to the curve, and its norm is the length of
dl. This element vector can be written dl = dx i + dy j so that, with u = u i + v j the above mass flux
becomes
Z
FC = ρ(u i + v j) (−dy i + dx j) , (2.12)
Z C

= ρ (−∂y ψ i + ∂x ψ j) (−dy i + dx j) , (2.13)

CZ Z
= ρ ∂y ψ dy + ∂x ψ dx = dψ = ψB − ψA , (2.14)
C C

where ψB denotes the value of the streamfunction at point B, and ψA at A, and where we have used (2.10)
to express the velocity.

2.3 Bernoulli theorem

For simplicity we restrict to the case of an incompressible, inviscid fluid. Then the equation for motion can
be written
Du 1
= − ∇P − ∇φ (2.15)
Dt ρ
where φ = g z is geopotential (as an example of an external, conservative force). Multiply equation (2.15)
by u and note that
Du Dkuk2
u =
Dt Dt
and that
Dφ
u∇φ =
Dt
one obtains
Dkuk2 1 Dφ
= − u∇P − (2.16)
Dt ρ Dt
Now, taking advantage of incompressibility we have u∇P = ∇(u)P . Noting that
 
D P 1 ∂P 1
= + u∇P
Dt ρ ρ ∂t ρ

So that we obtain

kuk2
 
D P 1 ∂P
+ gz + = (2.17)
Dt 2 ρ ρ ∂t
Hence, for an incompressible and inviscid fluid, and for the special case of steady flow, we have along
a streamline:
kuk2 P
+ gz + = cst . (2.18)
2 ρ
More general versions can be derived, but require some more work. This form of the Bernoulli theorem
is sufficient to describe common flows such as:

• flow from a pitcher of water, in which the above expression simply translates the conversion of
potential to kinetic energy,

• the initial speed of a liquid flowing out of a tank through a hole: consider a large tank water with a
small hole at the bottom, on the side. The outflow is sufficiently weak that the surface is approximated
as being at constant height H above the hole. In the weak, steady flow that sets up in the tank,
one can draw a streamline from the hole up to the surface. A key observation is that at both ends
of the streamline (at the surface and right outside the hole, in the small outgoing jet of water), the
pressure is equal to the atmospheric pressure. The Bernoulli theorem along that streamline then tells
that the potential energy gH has been converted to kinetic energy, 1/2 u2 , yielding the velocity of
the outgoing flow.

• the physical principle of a Pitot tube, a device commonly used in aircrafts to measure the speed of
the aircraft relative to the air around. It consists of a probe at the tip of which the flow is brought
to a stop (converting all the kinetic energy of the incoming flow into a pressure anomaly) and of two
measurements of pressure (one at the tip of the probe, one for the environmental pressure).

The Bernoulli theorem expresses also the Venturi effect, which is counter-intuitive: for steady, incom-
pressible flow in a pipe, the pressure decreases where the pipe is constricted. Indeed, the flow has to be
faster in order to keep the mass flux constant, and consequently the pressure decreases to compensate the
increase in kinetic energy.

2.4 Flow past an obstacle

Much of fluid dynamics has been concerned with flow past an obstacle, the latter being for instance a
sphere, a bullet, or a wing. Questions of interest are the description of the flow and the force exerted by
the fluid on the object. Concerning the latter, dimensional considerations can yield the general form to be
expected for the force. First, we need to list the relevant quantities of the problem considered:

• incoming velocity, U , in m s−1

• density of the fluid, ρ, in kg m−3 ,

• section of the object or obstacle, A, in m2 .

 We are interested on the force on the object:
force F in N or kg m s−2 . Given the quantities involved, there is only one combination which yields the
right dimensions: hence this force should scale with

F ∝ ρ U2 A . (2.19)

This does not provide the answer for any given problem, but it does imply that the force exerted by a fluid
on an obstacle can be non-dimensionalized with ρ U 2 A (or 1/2 ρ U 2 A). This is quite a powerful result:
for a given object (say a ball) of known shape (and cross-section A1 ), if we measure the force (F1 ) exerted
by a fluid flowing around it with a certain fluid (ρ1 ) coming at a certain speed (U1 ), then the force on
an object having the same shape and surface but different size (A2 ), when a different fluid (ρ2 ) has the
same flow (same streamlines) but at a different speed (U2 ) around it, well that force can be obtained from
knowledge of the first situation:
ρ2 U22 A2
F2 = F1 . (2.20)
ρ1 U12 A1
We will come back to the above and its implications in section 2.5, but beforehand we can it is useful to
describe a simple reference flow: that of an incompressible, inviscid fluid around a cylinder having a circular
cross-section. Using a streamfunction to describe the two-dimensional flow around the cylinder, it is left
as an exercise to show that a simple expression may be obtained for the streamfunction. The symmetric
flow around a cylinder, for incompressible, inviscid fluid, is shown in figure 2.1. The lines depicted are
streamlines: where they are more closely packed together, the flow is faster. As the Bernoulli theorem
implies, the pressure in such region is weaker.

Streamfunction for idealized, inviscid flow past a cylinder

1.5

1.0

0.5

0.0
y

−0.5

−1.0

−1.5

−2.0
−3 −2 −1 0 1 2
x

Figure 2.1: Analytical solution for the flow around a cylinder, considering that the fluid is incompressible
and inviscid. This last assumption in particular implies that the fluid velocity on the surface of the object is
non-zero (there is of course no velocity normal to the object, but the tangential velocity is not constrained).
 This example serves several purposes: it constitutes a fundamental example of flow around an obstacle,
with a simple, compact analytical solution. The description of such solutions, by Jean le Rond d’Alembert
(1717-1783) led him to highlight the paradox which bears his name: the flow described in the solution
exerts no force on the object (as the streamlines are symmetric the pressure anomalies in front and behind
are symmetric, hence the integral of the pressure anomalies along the whole surface yields zero. This is
contrary to intuition and common experience, ie that the fluid exerts a force on the object, in the direction
of the flow. It took many decades to understand this discrepancy between the exact solution displayed
in Figure 2.1 and the evidence that the flow exerts a force on ’obstacles’. Pieces necsessary to solve this
puzzle will be introduced in chapter 6.

2.5 Drag and lift coefficient

2.5.1 Definitions
For the design of wind turbines it will be essential to know and quantify the force exerted by a fluid on
an object. The discussion above has prepared the introduction of the drag coefficient. Indeed, consider a
sphere of cross-section A, in a fluid of density ρ coming with a uniform speed U and flowing around it.
The force on the sphere, in the direction of the incoming flow, will be

FD = CD ρ U 2 A . (2.21)

Hence it is sufficient in principle to measure once the value CD , and then the force for any flow that is
geometrically the same4 (same streamlines, allowing for a homothetic transformation) is known.
The coefficient CD that has naturally been introduced in equation 2.21 is the drag coefficient. It is a
non-dimensional coefficient, and relates to the drag, ie the force in the direction of the incoming flow.
Now consider an elongated object such as a cylinder, or an infinite wing. Because of the geometry, we
will not be aiming for a force but for a force per unit length. This will be proportional to ρ U 2 l, where l
is a length relevant for the section of the object considered. Now a second difference comes to mind when
considering a wing: the force is not necessarily oriented in the direction of the flow, it can form an angle
relative to the incoming flow. (This was hinted at in equation 2.20 by using vectors.) In other words, as
we consider a two-dimensional problem, the force has two components: drag, parallel to the incoming flow,
and lift, normal to the incoming flow.
Hence the force on an object such as a wing of infinite span is characterized by two numbers, cd and
cl , such that the drag force and lift force, per unit length, are expressed as:

Fd = cd ρ U 2 c , Fl = cl ρ U 2 c . (2.22)

Two points are worth noting in the above expression: first, following common usage, we have switched to
small letters for coefficients for the force per unit length spanwise. Second, the length scale retained to
characterize the object has been noted c, with the chord in mind. This is explained and justified in the
next section as we consider airfoils.

2.5.2 Airfoils
An airfoil is the cross-sectional shape of a body such as a wing, a blade or a sail. These are designed
so that the flow produces on them an aerodynamic force of interest. For example, the symmetric airfoil
sketched in figure 2.2 allows the fluid to flow around it smoothly, thus offering little resistance to the flow.
In more precise terms, the curvature of the airfoil is such that the viscous boundary layer remains attached
to the body from the leading edge to the trailing edge. Only very little fluid is affected by the presence of
the body, and hence, following Newton’s third law, the fluid exerts only a weak force on the body. In more
precise terms, there is only a weak drag.
To describe airfoils more generally it is necessary to introduce and define some characteristics of airfoils:
4
This is an essential condition which should not be forgotten. Even for the simple case of a sphere, as the speed changes,
the nature of the flow around the object can change.
 • the leading edge is the point, at the front of the airfoil, where the curvature is maximum,
• the trailing edge is the point, at the rear of the airfoil, where the curvature is maximum,
• the chord is the line joining the leading and trailing edge. The chord length, c, is the reference
dimension for the airfoil (see its use in equation 2.22).

Figure 2.2: Sketch of the flow around a symmetric airfoil, with the flow aligned with the airfoil’s chord line
(in other words, with an angle of attack equal to zero). from [BJSB11].

The angle of attack is the angle between the incoming flow and the chord line of an airfoil. When
an airfoil is slightly tilted relative to the incoming flow, resulting in a small angle of attack, the flow may
remain smooth around it. The streamlines are parallel at the trailing edge, as illustrated in Figure ??.

Figure 2.3: Sketch of the steady flow around a symmetric airfoil for a small angle of attack.

It has been explained in section ?? that the force on a body due to flow around it will have two
components, lift and drag. Wings are designed in such way that the flow results in a strong lift (supporting
the weight of the plane, or glider, or bird), but only a weak drag. Hence a glider or a soaring bird needs to
move forward at a speed sufficient that the flow around its wings produces the lift sufficient to support its
weight. Its forward speed decreases only slowly because the drag on the wings is weak.
These properties of airfoils are illustrated for an example in figure 2.5. Points to note from this figure
are:
• the lift increases with angle of attack, essentially in a linear way, up to a certain angle (here around
12◦ ). Beyond that angle, an abrupt change in the nature of the flow takes place, the flow is stalled.
 • For weak angles of attack, the drag force remains nearly constant and very weak.

• The ratio of lift to drag, for weak angles of attack, car reach formidable values (more than 50).

• In the stalled regime, drag dramatically increases, and the ratio of lift to drag becomes much weaker.

Figure 2.4: Sketch of stalled flow around an airfoil, when the angle of attack becomes too large.

Drag and lift coefficients are illustrated in Figure 2.5 for a symmetric profile, NACA0012. This is an
opportunity to mention the considerable work necessary to document airfoils. It is essential, for designing
objects that require airfoils, to precisely document the lift and drag coefficients. This requires conceiving
and referencing airfoils, to start. The NACA family of airfoils is a famous and well documented family,
with each number in the airfoil name quantifying a characteristic of the airfoil. The first two digits inform
about the camber (amount of camber as a percentage of the chord length for the first digit, position where
the maximum camber occurs, in units of 10 % of the chord length). The last two digits give the ratio of
the maximum thickness to chord length, as a percentage. In this family of airfoils, the maximum thickness
occurs at 30 % of the chord.
Families of profiles with varying characteristics have been documented. Need to quantify lift and drag
for a range of angles of attack, but also for different Reynolds numbers.
The above explanations were for symmetric airfoils. However, airfoils very commonly have a camber. A
cambered airfoil generally has a nonzero lift for a zero angle of attack. To explore the variations of lift and
drag with regards to angle of attack and camber, one can use the online tool made available by NASA5 :

https://www.grc.nasa.gov/WWW/K-12/airplane/foil3.html

The tool allows an exploration of the theoretical variations of lift and drag for different shapes, cambers,
thicknesses, and angles of attack. One should be careful with the number of parameters taht can be
changed, with the scaling of the graph displaying lift and drag, and with the units used.

5
Douglas Keller is gratefully acknowledged for pointing out this tool.
Figure 2.5: Schematic (top) of the drag (red arrow) and lift (green arrow) exerted by the incoming wind
(black arrows) on an airfoil with an incident angle α. Example of the lift and drag coefficients for a given
airfoil (NACA0012) and for a given Reynolds number (106 ) are displayed below; lift on the left, and drag ’à
droite. The ratio of the lift and drag coefficients is shown at the bottom. Graphs are taken from [BJSB11].