Airborne Wind Energy

Basic mechanics of crosswind kites

Roland Schmehl

26 September 2025

CC BY 4.0

Outline

Kitepower

Learning objectives

  • Derive a basic theory of tethered flight and energy harvesting with kites
  • Identify all relevant problem parameters
  • Introduce a consistent set of non-dimensional numbers
  • Illustrate idealized flight scenarios practical examples

Relevant literature sources are Loyd (1980) and Diehl (2013).

Definitions and assumptions

Ni Yan / AWESCO doctoral training network

Terminology

  • Velocity is a vector property.
  • Speed is a scalar property, the magnitude
    of the velocity vector.
  • A kite can be of fixed-wing of soft-wing type.


Terminology

  • Velocity is a vector property.
  • Speed is a scalar property, the magnitude
    of the velocity vector.
  • A kite can be of fixed-wing of soft-wing type.


Mathematical notation

  • Scalars are set in italic: \(a\), \(b\), \(\rho\), \(\phi\)
  • In digital resources, vectors are set in bold roman: \(\vec{v}\), \(\vec{L}\), \(\vec{D}\)
  • In handwritten resources, vectors are overset with an arrow: \(\overset{→}{v}\), \(\overset{→}{L}\), \(\overset{→}{D}\)
  • Units are set in roman: \(\rm m\), \(\rm kg\)
  • Mathematical functions are set in roman: \(\sin\), \(\cos\), \(\max\)
  • Product of scalars: \(a b\)
  • Scalar product of vectors: \(\vec{v}\cdot\vec{v} = v^2\)
  • Cross product of vectors: \(\vec{v} = \boldsymbol\omega\times\vec{r}\)
  • Cartesian unit vectors: \(\vex\), \(\vey\), \(\vez\)

Environmental properties

Ambient wind velocity
\[\vvw, \quad {\rm m/s}\]

Air density
\[\rho, \quad {\rm kg/m}^3\]

Dynamic wind pressure
\[q = \frac{1}{2}\rho\vw^2, \quad {\rm N/m}^2\]

Wind power density
\[\Pw = \frac{1}{2}\rho\vw^3, \quad {\rm W/m}^2\]

Gravitational acceleration
\[\vec{g}, \quad {\rm m/s}^2\]

Kite properties

Flat wing surface area (unrolled)

Wing planform area (projected)
\[S, \quad {\rm m}^2\]

Wing span
\[b, \quad {\rm m}\]

Wing side area (projected)

Wing maximum chord
\[\cmax, \quad {\rm m}\]

Airfoil geometry

Kite mass¹
\[\mk, \quad {\rm kg}\]

1 Often also just denoted as \(m\), without subscript.

Example: TU Delft V3 kite

Kite kinematics

Friedl (2015)

Kite kinematics

Kite velocity
\[\vvk\]

Apparent wind velocity
\[\vva = \vvw - \vvk\]

Resultant aerodynamic force
\[\Fa = \frac{1}{2}\rho\CR S \vaexp{2}, \quad \text{with} \quad \vaexp{2} = \vva\cdot\vva\]

Apparent wind velocity

Kite velocity
\[\vvk\]

Apparent wind velocity
\[\vva = \vvw - \vvk = \vvw + (- \vvk)\]



Relative flow

Kite aerodynamics

Poland et al. (2025)

Aerodynamic forces

Resultant aerodynamic force can be decomposed into lift and drag force
\[\vFa = \vec{L} + \vec{D}\]

Drag force component is defined as
\[\vec{D} \parallel \vva\]

Lift force component is defined as
\[\vec{L} \perp \vva\]

Aerodynamic side force \(\vFas \sim \betas\) neglected for performance modeling.

\[\begin{align} \vec{L} & = \frac{1}{2}\rho\CL S \vaexp{2}\\ \vec{D} & = \frac{1}{2}\rho\CD S \vaexp{2} \end{align}\]

Determining aerodynamic coefficients

Wind tunnel measurement

  • Rigid scale model with 1.28 m span
  • Unloaded design shape
  • Variable angle of attack & side slip angle
  • Reynolds number lower than real kite

Computational Fluid Dynamics

  • CAD model before meshing
  • Unloaded design shape
  • Variable angle of attack & side slip angle
  • Uncertainty about captured physics

Aerodynamic characteristics

  • Data for TU Delft V3 kite with empirical corrections for different flight phases: powered, depowered, steered.

Source: Cayon et al. (2025), with Vortex Step Method (VSM) as implemented in Cayon et al. (2023).

compare_polars_v3.py

Aerodynamic lift devices

Airfoil

  • Full range of \(L/D\)
  • \(D\) can be minimised
  • Control by angle of attack

Rotating cylinder

  • Limited \(L/D\)
  • \(D\) can not be modified
  • Control by angular velocity

Tethered flight

  • Aerodynamic forces
  • Aerodynamic lift and drag

Approach

Start from the minimal setup, the untethered wing.
Step by step add physical components,
add motion components,
and compute forces acting on the system.

Analytic modelling framework
\[\zeta = \frac{P}{\Pw S} = \CR\left[1+\left(\frac{L}{D}\right)^2\right] f \left(\cos\beta\cos\phi - f\right)^2\]




Steady gliding flight

TobiWanKenobi/Wikipedia

Stagnant wind

Kite velocity
\[\vvk = \vvkx + \vvkz = \const\]

Glide ratio
\[E = \frac{\vkx}{\vkz} = \frac{1}{\tan\gamma} \quad ⤳ \quad \text{glide angle } \gamma\]

Apparent wind velocity
\[\vva = \vvw -\vvk = -\vvk, \quad \text{with} \qquad \vw=0\]

Steady force equilibrium
\[m\vg + \vFa = 0\]

Geometric similarity of kinematic and force triangles because \(\vvkz \parallel \vFa\) and \(\vvk \parallel \vec{D}\)
\[E = \frac{\vkx}{\vkz} = \frac{L}{D}\]





Geometric similarity




Same shape but different scale, in fact, different units (N vs m/s).

Uniform and constant wind

Effect of additional wind velocity \(\vw > 0\)

If wing flies against wind, glide angle \(\gamma\) gets steeper

Apparent wind velocity
\[\begin{align} \vva & = \vvw -\vvk\\ \vax & = \vw -\vkx \\ \vaz & = -\vkz \end{align}\]

Geometric similarity of kinematic and force triangles
\[E = \frac{\vax}{\vaz} = \frac{L}{D}\]

Application in AWE:
limiting case of retraction in pumping cycle with
vanishing tether tension.






Tethered static flight

Library of Congress / NPS.gov

Dan Tate, left, and Wilbur Wright, right, flying
the 1902 glider as a kite, on 19 September 1902.

Static massless kite

Kite velocity
\[\vvk = 0\]

Apparent wind velocity
\[\vva = \vvw - \vvk = \vvw\]

Static force equilibrium
\[\vFt + \vFa = 0\]

\[\Ft = \Fa = \sqrt{L^2 + D^2}\]

\[\begin{align} \tan\beta & = \frac{L}{D} \\ \beta & = \arctan\frac{L}{D} \end{align}\]

\[\lim_{\frac{L}{D}⤳\infty} \beta = \frac{\pi}{2}\]




Effect of gravity

Static force equilibrium
\[\vFt + \vFa + m\vg = 0\]

\[\Ft = \sqrt{\left(L-mg\right)^2 + D^2}\]

\[\begin{align} \tan\beta & = \frac{L-mg}{D} \\ & = \frac{L}{D} \left[1-\frac{mg}{L}\right] \\ & = E\,\left[1-\frac{2gm}{\rho\CL S \vwexp2}\right] \end{align}\]

\[\begin{align} \phantom{\tan\beta} & = E\,\left[1-\frac{\color{red}{\frac{2g}{\rho\CL}\frac{m}{S}}}{\vwexp2}\right] \end{align}\]

But what is the physical meaning of \(\color{red}{\frac{2g}{\rho\CL}\frac{m}{S}}\)?

Static take-off limit

SkySails’ static mast launch

Generated lift just balances the effect of gravity:
\[\frac{1}{2}\rho\CL S \vwexp2=mg\]

Equivalent to \(\beta = 0\) (horizontal tether).

Solving for \(\vw\) leads to static take-off wind speed:
\[\vwsto = \sqrt{\frac{2g}{\rho\CL}\frac{m}{S}}\]

Mass-to-area ratio \(m/S\) is an important scaling parameter affecting take-off limit

Definition identical to the stall speed \(\vstall\) of an aircraft (Anderson 2016).

Effect of gravity

Static force equilibrium
\[\vFt + \vFa + m\vg = 0\]

\[\Ft = \sqrt{\left(L-mg\right)^2 + D^2}\]

\[\begin{align} \tan\beta & = \frac{L-mg}{D} \\ & = \frac{L}{D} \left[1-\frac{mg}{L}\right] \\ & = E\,\left[1-\frac{2gm}{\rho\CL S \vwexp2}\right]\\ & = E\,\left[1-\frac{\color{red}{\frac{2g}{\rho\CL}\frac{m}{S}}}{\vwexp2}\right] \end{align}\]

\[\begin{align} \phantom{\tan\beta} & = E\,\left[1-\left(\frac{\color{red}{\vwsto}}{\vw}\right)^2\right], \quad \text{with} ~ \vwsto = \sqrt{\frac{2g}{\rho\CL}\frac{m}{S}} \end{align}\]

Static take-off wind speed


Kite \(S~\text{(m$^\text{2}$)}\) \(m~\text{(kg)}\) \(\CLmax~\text{(-)}\) \(m/S~\text{(kg/m$^\text{2}$)}\) \(\vwsto~\text{(m/s)}\)
Ampyx AP2 3 35 1.5 11.7 11.3
Mozaero AP3 12 475 2.1 39.6 17.5
Makani MX2 54 1850 2 34.3 16.7
MegAWES 150.45 6885 1.9 45.8 19.8
TU Delft V3 19.75 22.8 0.88 1.15 4.6
Kitepower V9 47 73 1.19 1.55 4.6

Source: Joshi et al. (2024)

Ideal crosswind flight

TU Delft V3 kite flying on launch mast in 2012

Massless crosswind kite

Reeling factor ⤳ degree of freedom
\[f = \frac{\vkx}{\vw}\]

Tangential velocity factor
\[\lambda = \frac{\vky}{\vw}\]

Apparent wind velocity
\(\begin{aligned} \vva & = \vvw - \vvk \\ & = \left( \vw - \vkx \right)\vex + \left( - \vky \right)\vey \end{aligned}\)

Geometric similarity of velocity and force triangles
\[\frac{L}{D} = \frac{\vky}{\vw-\vkx} \qquad ⤳ \qquad \frac{L}{D}\left( \vw - \vkx \right) = \vky \\ \lambda = \frac{L}{D} \left( 1-f \right) = E \left( 1-f \right) \quad \quad \text{dependent variable, i.e. not a degree of freedom!}\]






Massless crosswind kite

Apparent wind velocity
\(\begin{aligned} \vaexp{2} & = \left( \vw-\vkx \right)^2 + \vky^2 \qquad ⤳ \qquad \left(\frac{\va}{\vw}\right)^2 = \left( 1-f \right)^2 + \lambda^2\\ \frac{\va}{\vw} & = \left( 1-f \right) \sqrt{1+E^2} \end{aligned}\)

Aerodynamic force
\[\Fa = \frac{1}{2}\rho S \CR \vaexp{2} = \frac{1}{2}\rho S \CL \sqrt{1+\frac{1}{E^2}} \left(\frac{\va}{\vw}\right)^2 \vwexp{2}\]

Non-dimensional tether force
\[\frac{\Ft}{qS} = \CL \sqrt{1+\frac{1}{E^2}} \left( 1-f \right)^2 \left( 1+E^2 \right), \quad \text{with} \qquad q=\frac{1}{2}\rho\vwexp{2}\]

Power harvesting factor
\[\zeta = \frac{P}{\Pw S} = C_L \sqrt{1+\frac{1}{E^2}} f \left( 1-f \right)^2 \left( 1+E^2 \right), \quad \text{with} \qquad \Pw = \frac{1}{2}\rho\vwexp{3}\]

Optimal reel-out speed

Maximum power harvesting factor at maximum value of
\(f \left( 1-f \right)^2\)

This leads to a condition for the first derivative
\[0 = \frac{\rm{d}}{\rm{df}} \left[ f \left( 1-f \right)^2 \right]\]

from which the optimal reeling factor can be determined
\[\fopt=\frac{1}{3}\]

Optimal harvesting factor using \(\fopt\)
\[\zetaopt = \frac{4}{27} C_L \sqrt{1+\frac{1}{E^2}}\left( 1+E^2 \right)\]

Reality check

The analytic formula for \(\zetaopt\) is based on a number of idealizations:

  • ideal crosswind operation at \(\beta=\phi=0\),
  • neglected reel-in phase,
  • vanishing mass of the airborne system components,
  • dragless tether.

In reality, the achievable power harvesting factor is substantially lower than predicted by \(\zetaopt\).

Compare achievable power harvesting factor of the Kitepower Falcon:

  • rated power of 100 kW at the rated wind speed of 10 m/s,
  • optimal power harvesting factor \(\zetaopt=\) 11.36,
  • achievable power harvesting factor \(\zeta=\) 3.47.
  • This is about 30% of the theoretical optimum value.

Maximum reel-out factor?

  • At \(f=1\) the apparent wind speed vanishes
    ⤳ kite can not produce any lift.
  • But what happens for \(f>1~(\vt>\vw)\)?
    ⤳ gliding in downwind direction,
    ⤳ vanishing tether tension.
  • Practical application is step towing, by
    alternating downwind gliding with \(f>1\), and
    upwind retraction of the kite with \(f<0\).
  • Technique well known from winch
    launching of paragliders.

Note: for \(\beta>0\) and/or \(\phi \ne 0\), \(\va\) vanishes for \(f<1\).


Crosswind kite with mass

Static force equilibrium
\[\vFt + \vL + \vD + m\vg = 0\]

Horizontal lift force component
\[|\vec{L} + m\vec{g}| = \sqrt{L^2 - \left(mg\right)^2}\]

Geometric similarity
\[\frac{\vky}{\vw-\vkx} = \frac{\sqrt{L^2 - \left(mg\right)^2}}{D}\]

Tangential velocity factor
\[\lambda = (1-f)E\sqrt{1 - \left( \frac{mg}{L} \right)^2}\]

2025-04-10-resit



Crosswind kite with mass

Apparent wind speed
\[\begin{align} \left( \frac{\va}{\vw} \right)^2 & = (1-f)^2 + \lambda^2 \\ & = \left( 1 - f \right)^2 \left\{ 1 + E^2\left[ 1-\left(\frac{mg}{L}\right)^2 \right] \right\} \\ & = \left( 1 - f \right)^2 \left\{ 1 + E^2\left[ 1-\left(\frac{\frac{2g}{\rho\CL}\frac{m}{S}}{\vaexp{2}}\right)^2 \right] \right\} \end{align}\]

Substitute
\[\color{red}{x} = \left( \frac{\va}{\vw} \right)^2\quad ⤳\]

\[\color{red}{x} = \left( 1 - f \right)^2 \left\{ 1 + E^2\left[ 1-\left(\frac{\vwsto}{\vw}\right)^4 \frac{1}{\color{red}{x^2}}\right]\right\}\]

2025-04-10-resit

Crosswind kite with mass

\[\color{red}{x} = \left( 1 - f \right)^2 \left\{ 1 + E^2 - E^2\left(\frac{\vwsto}{\vw}\right)^4 \frac{1}{\color{red}{x^2}}\right\}\]

Cubic equation, with missing linear term
\[\color{red}{x^3} - \left( 1 - f \right)^2 \left( 1 + E^2 \right) \color{red}{x^2} + \left( 1 - f \right)^2 E^2 \left(\frac{\vwsto}{\vw}\right)^4 = 0\]

Cubic equation in standard form
\[a x^3 + b x^2 + c x + d = 0, \quad \text{with}\]

\[\begin{align} a & = 1, \\ b & = - \left( 1 - f \right)^2 \left( 1 + E^2 \right), \\ c & = 0, \\[-15px] d & = \left( 1 - f \right)^2 E^2 \left(\frac{\vwsto}{\vw}\right)^4. \\ \end{align}\]

2025-04-10-resit

Number and nature of roots

A cubic equation has three roots. The nature of the three roots depends on the value of the discriminant
\[ \begin{align*} \Delta & = \ccancel{18 abcd} - 4b^3d + \ccancel{b^2c^2} - \ccancel{4ac^3} - 27 a^2d^2 \fragment{0}{, \quad \text{because} \quad \color{red}{c=0}} \\ & \fragment{1}{{} = 4(1-f)^8(1+E^2)^3 E^2 \left( \frac{\vwsto}{\vw} \right)^4 - 27(1-f)^4 E^4 \left( \frac{\vwsto}{\vw} \right)^8,} \\ & \fragment{2}{{} = (1-f)^4 E^2 \left( \frac{\vwsto}{\vw} \right)^4 \left[ 4 (1-f)^4 (1+E^2)^3 - 27 E^2 \left( \frac{\vwsto}{\vw} \right)^4 \right]} \end{align*} \]

For \(\Delta < 0\), the equation has one real root and two non-real complex conjugate roots.

For \(\Delta = 0\), the equation has one real root and two real repeated roots.

For \(\Delta > 0\), the equation has three distinct real roots.

Cubic polynomial

Polynomial function and derivatives
\[\begin{alignat*}{2} &p(x) &&= x^3 - \left( 1 - f \right)^2 \left( 1 + E^2 \right) x^2 + \left( 1 - f \right)^2 E^2 \left(\frac{\vwsto}{\vw}\right)^4, \\[-6mm] &p'(x) &&= 3x^2 - 2\left( 1 - f \right)^2 \left( 1 + E^2 \right) x, \\ &p''(x) &&= 6x - 2\left( 1 - f \right)^2 \left( 1 + E^2 \right). \end{alignat*}\]

⤳ Wind speed \(\vw\) affects vertical offset of polynomial

End behavior
\[\begin{alignat*}{3} &x\to & -\infty &⤳ p(x)\to & ~ -\infty, \\ &x\to & \infty &⤳ p(x)\to & ~ \infty. \end{alignat*}\]

Critical points: local extrema \((p'=0)\)
\[\begin{alignat*}{4} &x && =0 &&⤳ \text{local maximum} && (p'' < 0). \\ &\xc && =\frac{2}{3} \left( 1 - f \right)^2 \left( 1 + E^2 \right) &&⤳ \text{local minimum} \quad && (p'' > 0). \end{alignat*}\]

Polynomials for the TU Delft V3 kite at \(f=0\).

Physical solution space

\(x \ge \xmin = (1-f)^2.\)

This lower limit follows from the definition of the non-dimensional apparent wind speed as
\[x = (1-f)^2+\lambda^2,\]

and the fact that the crosswind contribution is always positive, i.e. \(\lambda^2 \ge 0\).

\(x \le \xmax = \left( 1 - f \right)^2 \left( 1 + E^2 \right).\)

This upper limit is the apparent wind speed for Loyd’s ideal crosswind case, where \(m=0\).

Interesting is also the observation

\[\xc = \frac{2}{3} \xmax.\]

Polynomials for the TU Delft V3 kite at \(f=0\).

Minimum flyable wind speed

  • At \(\vw=\frac{1}{4}\vwsto\), the cubic equation has no real roots in the physical solution space.
  • At \(\vw=\frac{1}{3}\vwsto\), the cubic equation has two real roots in the physical solution space.
  • At what value of \(\vw\) does the cubic equation have a repeated real root, transitioning from no real root to two real roots?
  • This transition is described mathematically by the discrimonant \(\Delta\) changing the sign from negative to positive.

Polynomials for the TU Delft V3 kite at \(f=0\).

Discriminant

The discriminant is positive when
\[ \begin{align*} 4 (1-f)^4 (1+E^2)^3 & > 27 E^2 \left( \frac{\vwsto}{\vw} \right)^4, & & \\ \fragment{0}{{}\left( \frac{\vw}{\vwsto} \right)^4} & \fragment{0}{{} > \frac{27 E^2}{4 (1-f)^4 (1+E^2)^3},} & & \\ \fragment{1}{{}\vw} & \fragment{1}{{} > \sqrt[\Large 4]{\frac{27 E^2}{4(1+E^2)^3}} \frac{\vwsto}{1-f},} & \\ \fragment{2}{{}\vw} & \fragment{2}{{} > \frac{\vwc}{1-f}, \quad \text{with} \quad} & \fragment{2}{{} \vwc } & \fragment{2}{{} = \sqrt[\Large 4]{\frac{27 E^2}{4(1+E^2)^3}}\, \vwsto, \quad \text{and}} \\ & & \fragment{2}{{} \vwsto} & \fragment{2}{{} = \sqrt{\frac{2g}{\rho\CL}\frac{m}{S}}.} \end{align*} \]

But what is the physical meaning of \(\vwc\)?

Cardano’s method

Cubic polynomial

Polynomial for \(f=0\) and physical parameters of the TU Delft V3 kite

Investigate the cubic polynomial for
three different wind speeds,
\(\vw=\) \(0.9\vwc\), \(\vwc\) and \(1.1\vwc\).

By definition \(x = \left( \frac{\va}{\vw} \right)^2 \ge 0\),
\(x<0\) unphysical.

For \(\vw<\vwc\) \((\Delta < 0)\),
⤳ none of the real roots in
physical range.

For \(\vw=\vwc\) \((\Delta = 0)\),
⤳ repeated real root.

For \(\vw>\vwc\) \((\Delta > 0)\).
⤳ two distinct real roots.

Only interested in the lower value.

Full wind speed range

Tangential velocity factor \(\lambda_\mathrm{max} = \lambda(m=0)= E\).


Interpretation

Considering a kite with a static take-off wind speed
\[\vwsto = \sqrt{\frac{2g}{\rho\CL}\frac{m}{S}},\]

the theoretical cut-in wind speed for crosswind operation at \(\beta=\phi=0\) is given by
\[\vwc = \sqrt[\Large 4]{\frac{27 E^2}{4(1+E^2)^3}}\, \vwsto.\]

  • This definition of a lower wind speed limit implies a constant tether length (\(f=0\)).
  • For \(\vw < \vwc\), the kite can not generate a sufficient lift force for sustained flight.
  • For \(\vwc < \vw < \vwsto\), the generated lift force can counteract gravity. In this wind speed regime, the tangential velocity factor \(\lambda\) is a function of the wind speed \(\vw\).
  • For \(\vw > \vwsto\), the generated lift force is too high, driving the kite to a force equilibrium at higher elevation angles \(\beta\), or requiring reeling out (\(f>0\)).

Interpretation

  • By reeling in (\(f<0\)), the flyable wind speed range can be temporarily decreased below the value of \(\vwc\), however, this condition can not be maintained for longer time.
  • This low wind speed regime is particularly important for dynamic launching of kites.
  • While the static take-off wind speed \(\vwsto\) is affected by the lift coefficient \(\CL\), the theoretical cut-in wind speed \(\vwc\) adds the dependence on the lift to drag ratio \(E\).
  • Static take-off wind speed is affected by mass scaling, predominantly via the ratio \(m/S\).
  • Theoretical cut-in wind speed is substantially lower, especially for kites with high aerodynamic performance \(E=L/D\).
  • The analysis is highly idealized and the real cut-in wind speed of the kite will be higher.

Radial flight

Massless radial kite

Reeling factor ⤳ degree of freedom
\[f = \frac{\vk}{\vw}\]

Apparent wind velocity
\[\vva = \vvw - \vvk\]

Decomposition of kite velocity
\[\vvk = \vec{b} + \vec{c}\]
\[\vva = \vvw - \vec{b} - \vec{c}\] \[\vvw = \vva + \vec{c} + \vec{b}\] \[\vw^2 = \left( \va - c \right)^2 + b^2\quad \text{(positive vector magnitudes!)}\] \[\frac{\va}{\vw} = \sqrt{1-\left(\frac{b}{\vw}\right)^2} + \frac{c}{\vw}\]

Displayed is the case for reel in with \(f<0\)







Massless radial kite

Geometric similarity
\[\frac{b}{\vk} = \frac{L}{\Fa} = \frac{1}{\sqrt{1+\frac{1}{E^2}}}\] \[\frac{c}{\vk} = \frac{D}{\Fa} = \frac{1}{\sqrt{1+E^2}}\]

Normalized to the wind speed
\[\frac{b}{\vw} = -\frac{f}{\sqrt{1+\frac{1}{E^2}}}\] \[\frac{c}{\vw} = -\frac{f}{\sqrt{1+E^2}}\]

Minus signs aacount for negative value of \(f\)

Displayed is the case for reel in with \(f<0\)

Massless radial kite

Apparent wind speed
\[\begin{align} \frac{\va}{\vw} & = \sqrt{1 - \frac{f^2}{1 + \frac{1}{E^2}}} - \frac{f}{\sqrt{1 + E^2}}, \\ & = \frac{\sqrt{1 + E^2\left( 1 - f^2 \right)} - f}{\sqrt{1 + E^2}} \end{align}\]

Non-dimensional tether force
\[\frac{\Ft}{qS} = \CL \frac{\left[\sqrt{1+E^2(1-f^2)}-f\right]^2}{E\sqrt{1+E^2}}\]

Power harvesting factor
\[\zeta = \CL \frac{f\left[\sqrt{1+E^2(1-f^2)}-f\right]^2}{E\sqrt{1+E^2}}\]

Displayed is the case for reel in with \(f<0\)

Massless radial kite

Elevation angle ⤳ not a degree of freedom
\[\cos\beta = \frac{\sqrt{1+E^2(1-f^2)}+fE^2}{1+E^2}\]


Homework #1: Derive this expression.


Hint:

  • Formulate two different expressions for the apparent wind speed, one of them containing \(\cos\beta\).
  • Eliminate the apparent wind speed by equating those equations.
  • Solve for \(\cos\beta\).

2022-04-06-resit 2(d)

Massless radial kite

Minimum reeling factor
\[f_{\rm min}=-\sqrt{1+\frac{1}{E^2}}\]


Homework #2:
Derive this expression.


Hint:

  • Use the fact that the radicant in the expression for \(\cos\beta\) has to be greater or equal to zero.

2022-04-06-resit 2(f)

Loyd’s theory - comparison

awesco.eu/awe-explained/

How to use these analytic models?

The wind speed \(\vw\) describes the available energy resource.
The lift-to-drag ratio \(E\) characterizes the specific kite design.

For the ideal crosswind kite prescribe

the tether reeling speed \(\vt\) to calculate the achievable kite crosswind speed \(\vkx\), or
the kite crosswind speed \(\vkx\) to calculate the required tether reeling speed \(\vt\).

For the non-maneuvering kite prescribe

the tether reeling speed \(\vt\) to calculate the achievable elevation angle \(\beta\), or
the elevation angle \(\beta\) to calculate the required tether reeling speed \(\vt\).

References

Anderson, J.D.: Fundamentals of aerodynamics. McGraw Hill LLC (2016)
Cayon, O., Deursen, V. van, Schmehl, R.: Translational dynamics of bridled kites: A reduced-order model in the course reference frame. Wind Energy Science Discussions. 1–38 (2025). https://doi.org/10.5194/wes-2025-205
Cayon, O., Gaunaa, M., Schmehl, R.: Fast aero-structural model of a leading-edge inflatable kite. Energies. 16, 3061 (2023). https://doi.org/10.3390/en16073061
Diehl, M.: Airborne wind energy: Basic concepts and physical foundations. In: Ahrens, U., Diehl, M., and Schmehl, R. (eds.) Airborne wind energy. pp. 3–22. Springer, Berlin Heidelberg (2013)
Friedl, F.: Fault-tolerant design of a pumping kite power flight control system, http://resolver.tudelft.nl/uuid:e704e8aa-2371-437b-8de9-80b3b7067241, (2015)
Joshi, R., Schmehl, R., Kruijff, M.: Power curve modelling and scaling of fixed-wing ground-generation airborne wind energy systems. Wind Energy Science. 9, 2195–2215 (2024). https://doi.org/10.5194/wes-9-2195-2024
Kichenassamy, S.: Continued proportions and tartaglia’s solution of cubic equations. Historia Mathematica. 42, 407–435 (2015). https://doi.org/10.1016/j.hm.2015.03.004
Loyd, M.: Crosswind kite power. Journal of Energy. 4, 106–111 (1980). https://doi.org/10.2514/3.48021
Luchsinger, R.H.: Pumping cycle kite power. In: Ahrens, U., Diehl, M., and Schmehl, R. (eds.) Airborne wind energy. pp. 47–64. Springer, Berlin Heidelberg (2013)
Poland, J.A.W., van Spronsen, J.M., Gaunaa, M., Schmehl, R.: Wind tunnel load measurements of a leading-edge inflatable kite rigid scale model. Wind Energy Science Discussions. 2025, 1–33 (2025). https://doi.org/10.5194/wes-2025-77

Questions?





kitepower.tudelft.nl

// reveal.js plugins