Impossible d'afficher l'image. Votre ordinateur manque peut-être de mémoire pour ouvrir l'image ou l'image est endommagée.

Redémarrez l'ordinateur, puis ouvrez à nouveau le fichier. Si le x rouge est toujours affiché, vous devrez peut-être supprimer l'image avant de la réinsérer.

Narrow Kelvin Waves

Frédéric Moisy, Marc Rabaud

GDR PHENIX, Paris, September 19, 2013

Google images
 Summary

1)  Classical Kelvin wedge for gravity waves

2)  Narrow Kelvin wedge at large Froude
3)  Effect of capillarity
4)  Wave drag and planing hull
5)  Conclusions
 Dispersion relation

Monochromatic plane waves:

=> Wave equation A = A0 expi(kx − ωt)

Hypothesis: irrotational flow (low dissipation) and linearity (small amplitude)

' ! $2 * ρg
2 k ,
Dispersion relation ω = f(k) :
€ )
ω = gk 1+ # & kc =
)( " kc % ,+ γ

ω g * $ k ' 2-
Phase velocity: Vϕ (k) = = ,1+ & ) / Dispersive waves
k k ,+ % k c ( /.

Group velocity:
dω
Vg (k) =
€ dk
 Phase and group velocities in deep water
0,6

Vphase
V ou V (m/s) 0,5 Vgroupe

0,4 Vg
g

0,3
ϕ

Vϕ
0,2

€
0,1
k/k
c

0 €
0 0,5 1 1,5 2 2,5 3 3,5

k/kc
Close to minimum velocity

1,5 mm cylinder moving at 22 cm/s

Existence of Vmin: Thomson verified this point on his yacht with the help of an
eminent guest, Hermann von Helmholtz [O. Darrigol, Worlds of flow].

Kelvin's floating laboratory: the Lalla Rookh

 Lord Kelvin
(William Thomson)
1824-1907

Fig. 25. Kelvin’s ship-waves [Thomson 1887a, plate; perspective view borrowed by Kelvin from
William Froude
R.E. Froude] (1810-1878)
Mach construction or Kelvin wedge?

Mach cone (U > c) Kelvin wedge (any U)

1
Non dispersive media Dispersive media Vg = Vϕ
2

C = Vg = Vϕ g
U sin θ ( k ) = Vϕ =
k
€

C tan θ
sin α (U) = sin θ = tan α =
U 2 + tan 2 θ
-  Crawford, AJP, 1984
-  Lamb, Lighthill or Whitham
 tan θ g
tan α (θ ) = U sin θ ( k ) = Vϕ =
2 + tan 2 θ k

k 25
−1
kg sin α max =
1
tan α ( k ) = 3
k 20
2 −1
kg
α(k) (degrees)
15
g
avec kg = 2
U
10

0 0 1 2
10 10 10
k / kg
 Everything clear?

Photograph by Adrian
Pingstone (Avon Gorge,
Bristol, 2004)

Matlab simulation (F. Moisy)

 Summary

1)  Classical Kelvin wedge for gravity waves

2)  Narrow Kelvin wedge at large Froude
3)  Effect of capillarity
4)  Wave drag and planing hull
5)  Conclusions
 angle (°)
théorie
And yet …
ManipKelvin
60

50

40
angle (°)

30

20

10

0
Les Treilles
0 0,5 1 1,5 2 2,5 3 3,5 (octobre 2001)
Vitesse (m/s)
Jear Walker
(Pour la Science)

Wikipedia
 « Field measurements »

Set of 37 images from Google Earth

close to harbors (resolution 0.5 m)
(a)

Measure ship length L and

wavelength λ

Deduce ship velocity U and Froude

number

(b) U
Fr = α = f (Fr)?
gL
(Color online) Airborne images of ship wakes taken
Wake angles from Google Earth images

19.47

𝐹𝑟=𝑈/√⁠𝑔𝐿 

Fr = 0.15 Fr = 1.1
 Simulation with gaussian
Fr=0.5
Fr pressure field
(Havelock hypothesis 1918)

Fr =1

Fr =2
Origin of narrow wake pattern ? Finite size effect

λg/L=1/2 Largest excited gravity wave:

λg
Fr = 0.28

25
Fr = 0.4 Fr = 0.9 Fr = 2

The actual wake pattern 20

is an convolution of 15
α (deg.)

10

5
by the spectrum of the
disturbance 0 -1 0 1 2
10 10 10 10
k/k
g
 m g
of the pwaves corres
40 cos 1 ( 2/3) ' 35.
Kelvin this wedge of angle
Model (4)
Asympt. (5) -  (Dirac
2 lengths L andpressure
λg, dist
20 Images -  1 dimensionless number Fr
is recovered.
Simul.
α (deg.)

10
8
6 ct
Good
agreement! H
4 At large Froude:

3
cgt
1 gL
0.1 0.2 0.4 0.6 1 2 3 sin(α ) ~ = k
Fr Fr U
c
As in the Mach regime where sin(α ) =
IG. 2. (Color online) Red circles: Log-log plot of the wake U M
ngle ↵Phys.
as a Rev.
function of110,
Letters the214503
Froude(mai
number,
2013) measured from a
 Summary

1)  Classical Kelvin wedge for gravity waves

2)  Narrow Kelvin wedge at large Froude
3)  Effect of capillarity
4)  Wave drag and planing hull
5)  Conclusions
‘Cylindrical duck’ of diameter 3 cm, U = 0.75 m/s (Fr = 1.4)

𝛼≃19°
‘Cylindrical duck’ of diameter 3 cm, U = 2.6 m/s (Fr = 4.8)

𝛼≃7.1°
Cylinder of diameter 1.5 mm, U = 0.6 m/s (Fr = 4.9)

𝛼≃19°
Cylinder of diameter 1.5 mm, U = 2 m/s (Fr = 16)

α ≈ 6°
With capillary waves D ∈ [10,16, 30, 62 ] mm

Bo = 0.59
Bo = 0.94
Bo = 1.8
Bo = 3.6

D
Bo =
(deg.)

λc 1
10

0 1
10 10
-  3 lengths D, λc and λg, Fr = U / (g D)1/2
-  2 dimensionless number, Fr and Bo
With capillary waves

cϕt H

λ
cgt

k
φ
θ α

γ 3 M O
ω ( k ) = gk + k Ut
ρ
ω
U cosθ = cϕ =
k
cg ( k ) U 2 − cϕ2 ( k )
tan α ( k ) =
U 2 − c g ( k ) cϕ ( k )
 35
With capillary waves 1.2 1.4 1.6 1.938 3
30

25
U
= Fr π Bo
cmin

(k) (deg.)
20 Different U/cmin 5

15

γ 10
ω ( k ) = gk + k 3
ρ
5
ω
U sin θ = cϕ =
k 0 0 1 2
10 10 10
k/k
cg ( k ) U 2 − cϕ2 ( k ) g
tan α ( k ) =
U 2 − c g ( k ) cϕ ( k )
g
avec kg = 2 But damping in exp(-2νk2t)
U
With capillary waves D ∈ [10,16, 30, 62 ] mm

Bo = 3.7
Bo = 5.9
Bo = 11
Bo = 23

D
Bo =
λc
α (deg)

1
10
Generalised Froude
number:

* U
Fr =
cg ( D ) 10
0 1
10
U / c (k )
g eff
 Summary

1)  Classical Kelvin wedge for gravity waves

2)  Narrow Kelvin wedge at large Froude
3)  Effect of capillarity
4)  Wave drag and planing hull
5)  Conclusions
Wave-making resistance

-  Known since Froude

-  Existence of a wave drag

(dominant at large velocity)

1
Rw = ρU 2 B 2Cw
2
 boat by an imposed pressure field P (x, y) at the water
face. The resulting surface deformation ⇥(x, y) can then
computed as a Fourier integral (seeWave-making resistance
Eq. 2.17b of Ref. [9],
Eq. 11 of Ref. [1]). From this imposed pressure and cal-
ated wave field, the wave drag is then computed by inte-
ting the product
Havelockof the local pressure
method (1918): by the slope
Imposed of the perturbation P(x,y)
pressure
erface in the direction of the motion:
⇤⇤
⇥
RW = P (x, y) dxdy. (6)
x

On figure Where
4 we have simulated the wave pattern gen-
ted by a moving ZGaussian
Z ⇥ pressure field, g(r) =
dk dk P̂ (kx , ky )
= r /L , where F0 is2 a normalization
2 x y
⇧F0 /L2 ) ⇤(x,
exp y)2⇧ 2 2
exp i[kx x + ky y]
ce, which corresponds here2⌅ 2⌅weight
to the kc2boat
[ (k of+the ) (F k (k =U
0 x
i⇥)2 ]
D). From this simulated surface height, we have computed
wave drag using Eq. 6 for various Froude numbers. The
ults, plottedIninthe
figure
limit5, are in perfect agreement with the
ε→0
ct result found by Benzaquen et al. [1] for a Gaussian
 Wave-making resistance

0.12
A decrease of Simulation (N=2048)
the wave drag 0.1
Benzaquen et al. (2011)

is found.

Cw / (p / ! g b3)2
0.08

0.06

0.04

0.02

0
0 0.5 1 1.5 2 2.5 3
Fr
pressure field: i
Analytical result of Benzaquen, Chevy, ⇤ ⌅2 ⇧ /2
p
D 1 d⇥ d
Raphaël for an imposed gaussian CW =
L3 Fr8 0
⌃ ⇥ ⇥ 4⌥
cos5 ⇥ exp 2⌅Fr cos ⇥ o
pressure field.
(7) n
l
 Wave-making resistance

Wave resistance RW = f(Fr)~Fr6 or 8

Planing hull
Planing Mini-Transat boat
(Fr > 1)
 Conclusions
Fr =
1.1
- Narrow wave wakes exist, and are not explained
by the classical Kelvin argument

-  Capillary regime is under progress.

-  Fr > 1 => 1) Narrow Kelvin wake

2) Planing hull
3) a decrease of the wave drag

Is there a link or just a coincidence?

Thank you!
NASA satellite image (MODIS imager on board the Terra satellite) of a
wave cloud forming off of Amsterdam Island in the far southern Indian
Ocean. Image taken on December 19, 2005.
 References

1.  Thomson W. (Lord Kelvin), 1887, On ship waves,

Trans. Inst. Mech. Eng., 409–433.

2.  J. Lighthill. Waves in fluids. Cambridge University

Press, Cambridge, 1978.

3.  F.S. Crawford, Elementary derivation of the wake

pattern of a boat, Am. J. Phys. 52, 782-785, 1984.

4.  M. Rabaud and F. Moisy, Ship Wakes: Kelvin or

Mach Angle?, Phys. Rev. Letters 110, 214503 (mai
2013).