Droplet ejection using WunenburgerDroplet ejection using acoustic waves: Régis Wunenburger...

High intensity acoustic waves can be used to

o trigger flows,

o deform liquid surfaces,

o control hydrodynamic instabilities,

o mix, atomize liquids,

o transport droplets …

Droplet ejection using acoustic waves: Régis [email protected]

jeudi 4 octobre 2012

mailto:[email protected]:[email protected]

liquid

acoustic transducer

air

Context

Controlled delivery of liquid samples for high-throughput screening

applications

)biological tests, automated formulation(

500 µm

500 µs 1312 µs1375 µs1750 µs




water

air

! Origin of the mist : cavitation ? Faraday-like instability ?

! Big droplet : how big, how many, influence of viscosity ?

! Extension to emulsification )liquid-liquid interface(, encapsulation )3 liquids( …" Experiments

" Comparison with numerical simulation

To learn more : contact Régis Wunenburger - FCIH )[email protected]( 600 fps

Gerris



mailto:[email protected]:[email protected]:[email protected]:[email protected]

Thrusters for nanosatellites

Michel DUDECK, Institut d’Alembert, Université Pierre et Marie Curie 75252 Paris, France, [email protected]

Cubesat standard of nanosatellites for universities proposed by univ. Stanford (1kg, 1W, 10cm)

Inaugural fligh of VEGA on 13 Feb. 2012 from Europe's Spaceport in French Guiana with 9 small satellites from Poland, Italy, Hungary, Spain, Romania, France (Robusta 1)

First French nanosatelliteRobusta 1Univ. Montpellier2

Today:no on-board thruster, strong limitation of missions in space

Different concepts are studied: - ionic liquid ejection by electric field (EPFL Lausanne, Pays-Bas, GB, MIT) - emission of ions after ablation of teflon by magnetic field (Univ. of Surrey, GB) - emission of ions by field effect(Austria) - emission of Xe ion by electrostatic effect (UPMC, UVSQ)

Training: - Physical analysis of the different concepts (from published papers in English) - Analysis of the performances (specific impulse, level of thrust, efficiency, !V…) - State-of-art (projects, first manufactured thruster, results of tests in vacuum chamber, ..) - Advantage/disadvantages

Oct. 3, 2012


http://www.esa.int/esaMI/Launchers_Access_to_Space/ASEKMU0TCNC_1.htmlhttp://www.esa.int/esaMI/Launchers_Access_to_Space/ASEKMU0TCNC_1.htmlhttp://www.esa.int/esaMI/Launchers_Access_to_Space/ASEKMU0TCNC_1.html

Simulation of the flow in recorder Lagrée P-Y [email protected] Fabre B.

expériment ∂’Alembert LAM

- simulation of the instability of the jet- coupling with the cavity - comparisons with a compressible code CAAmeleon

Auvray R.



The Hour Glass

Simulation of granular flows Lagrée P-Y [email protected] Staron Lydie

simulation Gerris µ(I) versus DCM

fine understanding of the structure of the flowsimulation of other flows



pump

measurement point

8

9

7

4

6

2

5

1

3

1 m

0 0.1 0.2 0.3 0.4 0.5 0.6−10

−5

0

5

10

15

20

25

Time(s)

mea

n sp

eed(

cm/s

)

ExperimentalFEFV−1orderFV−2order

Simple artery model!Custom made 70kPa"(Ultrasonic diagnostic system)

Doppler system

Measurement system used

Measurement point

Fig. 22: Measurement system used.

3

2

Am

plitu

de [a.u

.]

Inner pressure

Flow velocity

1

0

Am

plitu

de [a.u

.]

-1

1.41.21.00.80.60.40.20.0

Time [s]

Fig. 23: Observed inner pressure wave and flow velocity.

33

∂Q

∂t+ γ

∂

∂x

Q2

S= −Sρ−1 ∂P

∂x− cf

Q

S

∂S

∂t= −∂Q

∂x

P = P0 + k(R−R0)(1 + ε(R−R0)

R0) + η

∂R

∂t

Simulation of arterial flows

2 kP

a

Incident wave

Reflected waves

83.0 cm

Pressure

2

55.0 cm

83.0 cm

Time

Measurement Simulation

0.5 s

27.5 cm

Fig. 15: Measured and simulated pressure waves (tube D).

Pressure

Time

Measured wave

Simulation (x 1.0)

Simulation (x 2.0)

Simulation (x 3.0)

0.5 s

1 kP

a

Offset level

Fig. 16: Effects of fluid viscous parameters to the pressure waves at the first point.

! "#A

$% &

'( )

BB

C C

Silicone tubesBifurcations

Reflection points Measurement point

Fig. 17: Structure of a simple artery model.

15

Lagrée P-Y [email protected] Fullana J.-M. Wang X.



Simulation of arterial flows Lagrée P-Y [email protected] Fullana J.-M.

where i ¼ 0; 1; 2 refer to the parent and two daughtervessels and Yi " Ai=rci is the characteristic admittanceof the ith vessel where Ai is the cross-sectional area. Thetransmission coefficient is T ¼ 1þ R: Both the reflec-tion and transmission coefficients at the junction dependupon the area and wave speed of each segment. If thereis no reflection, R ¼ 0; the junction is called ‘wellmatched’. We note that this equation is valid for bothforward and backward waves if the parent and daughterbranches are determined relative to the wave rather thanmorphologically. Thus, a single bifurcation will havedifferent reflection coefficients for waves approaching itin the three different vessels. This directional sensitivityof the reflection coefficient means that a junction that iswell matched for forward waves can cause substantialreflections for backward waves (Hardung, 1952).

Calculating the reflection coefficients for the differentbifurcations from the data collected by Westerhof et al.

(1969) and Stergiopulos et al. (1992), we found that thedata resulted in relatively large reflections coefficientsfor several of the junctions in the model. For example,the junction between the thoracic aorta I (segment 18)and the intercostal and the thoracic aorta II (segments26 and 27) have a 50% decrease in area which results ina reflection coefficient RE0:25: These reflections tendedto obscure the effects of the reflections from theperipheral segments and so we decided to adjust theradii of the vessels to ensure that our model system waswell matched for forward travelling waves. This wasdone by holding the parent vessel radius constant andincreasing or decreasing the radii of both daughtervessels by the same fraction until R ¼ 0: Table 1 includesthe adjusted radii and the wave speeds calculated fromthem with the original radii and wave speeds given inbrackets. Similarly, we found that the symmetry in theoriginal data for the length of the arteries in the legsresulted in simultaneous reflections from each of the legswhich resulted in unwanted interactions between re-flected waves and so we somewhat arbitrarily increasedthe length of each artery in the right leg by 1 mm: Theproblem of simultaneous reflections from the arms didnot arise because they are not symmetric.

For the resistances at the end of each terminal artery,the reflection coefficient depend upon the peripheralresistance, Rp (Lighthill, 1978):

RT "dP

DP¼

RP $ rcRP þ rc

:

The values for the terminal reflection coefficients, RT;are included in Table 1.1

2.4. The method of calculation

The method of characteristics outlined above does notmake any assumptions about linearity. Indeed, thesolution of the full, non-linear equations of one-dimensional flow is one of the advantages of themethod. There are two principal non-linearities in thatsolution: the characteristic directions depend upon thevelocity of the blood and the wave speed can be afunction of pressure. Both of these non-linearities can beincorporated into the method, but in this study we havechosen to concentrate our attention on the effects of thecomplex geometry of the arteries and so we will assumethat both of these non-linearities have a negligible effect,i.e. c ¼ constant in each arterial segment and U5c: Weplan to explore the non-linear effects in a subsequentpublication.

With this assumption of linearity, it is possible toapply the concept of the transfer function to the analysis

ARTICLE IN PRESS

Fig. 1. The arterial model used in the calculation. This model wasoriginally introduced by Westerhof et al. (1969) and contains data fordiameter, length, wall thickness and Young’s modulus for the 55largest arteries (after Stergiopulos et al., 1992).

1The use of RP as peripheral resistance and RT as terminal reflectioncoefficient is unfortunate nomenclature, but one that is forced upon usby standard usage.

J.J. Wang, K.H. Parker / Journal of Biomechanics 37 (2004) 457–470 459

Artery 49

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−1

0

1

2

3

4

5x 10−4

Time(s)

Flux

(cm

3 s−1

)

DGFEFVFD

Figure: Flux

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5

3

Time(s)

Pres

sure

DGFEFVFD

Figure: Pressure

Artery 45

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5

3x 10−4

Time(s)

Flux

(cm

3 s−1

)

DGFEFVFD

Figure: Flux

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5

3

Time(s)

Pres

sure

DGFEFVFD

Figure: Pressure

Artery 40

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5x 10−4

Time(s)

Flux

(cm

3 s−1

)

DGFEFVFD

Figure: Flux

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5

Time(s)

Pres

sure

DGFEFVFD

Figure: Pressure

Artery 22

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−2

0

2

4

6

8

10

12

14x 10−4

Time(s)

Flux

(cm

3 s−1

)

DGFEFVFD

Figure: Flux

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5

Time(s)

Pres

sure

DGFEFVFD

Figure: Pressure

Artery 16

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−2

0

2

4

6

8

10

12

14x 10−4

Time(s)

Flux

(cm

3 s−1

)

DGFEFVFD

Figure: Flux

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5

Time(s)

Pres

sure

DGFEFVFD

Figure: Pressure

Flux Pressure

Q(t) P (t)

Vénus of MOdeLIsation

Artery 1

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.02

0

0.02

0.04

0.06

0.08

0.1

Time(s)

Flux

(cm

3 s−1

)

DGFEFVFD

Figure: Flux

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−0.5

0

0.5

1

1.5

2

2.5

Time(s)

Pres

sure

DGFEFVFD

Figure: Pressure

Wang X.

- add viscoelasticity and tension- study the influence of parameters



Maurice Rossi

[email protected]édéric Doumenc

evaporation of polymer


mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]

[email protected]@gmail.com

[email protected]

Gianmarco Pinton

Daniel Fuster & S. Zaleski [email protected]


mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]

Figure à faire avec plus de lumière: fond blanc, avec moins de temps succesifs.Faire une séquance ou l’on voit la paille que je tiens... Indiquer l’intervalle de temps

fig1: le phénomène

g

Time5mm

t=0 t=... t=... t=...

Liqu

id c

olum

n in

free

fall

free fall

Ret

ract

ion

Texte

jerome hoepffner



fig3jerome hoepffner



ça coupeSphère

Notz et basaran. Diagramme de phase

Coupe sous forme de ligament

Coupe en f

orme

ramassée

Est-ce qu’on peut faire une distinction entre les cas qui se coupent en forme de ligament, et les choses qui se coupent en forme bizarre juste lorsque le ligament commence à se ramasser sur lui même (lorsque les deux bout en retractation commencent à se sentir l’un l’autre?)

Sphère

jerome hoepffner



Tim

e

Radius 1

Figure avec fond blancla vorticité et en plus les lignes de courant pour faire voir le venturi?Avec un zoom qui montre le champ de vitesseFigures avec Matlab.

Taylor-Culick speed

t=0

t=...

t=...

t=...

t=...

fig2: le numérique jerome hoepffner



t=[...]

Time

t=[...] t=[...] t=[...]

fig3: la figure des particulesjerome hoepffner



Time

a) Pinching; Oh=???

fig4: la figure du colorant

b) Avoiding the pinching; Oh=???

t=0

jerome hoepffner



Too viscous

Too shortPinch

Ici j’essaye de faire quelque chose dans l’esprit de reprendre le graph de Hutching, mais en le transformant en information pour la localisation en x des evènements: premier pinch et premier et suivants évitements. Chouette: pour le visqueux, ça fait plein d’évitements successifs, qui vont faire de plus en plus de lignes vers le haut.

Sur ce graph, on peut aussi mettre les images de la forme du ligament au moment ou les choses se passent si on trouve la place.

Je crois bien que la forme de ce qui se passe pour lévitement dans le cas semi-infini, va donner des idées claires pour les limites entre le pinch et le no pinch dans le cas.

On pourrai déjà commencer à mettre des points en regardant le film de Gounséti, mais il faudrai metre plus de maillage, et faire les choses plus proprement.

La question laissée de côté: que se passe t-il lorsque les deux parties du ligament se rencontrent: des effets spéciaux qui promeuve le pinch ou le no pinch? Ce sera important si notre critère doit prédire la longueur critique...

2nd Pinch

Avoidanc

e

Breakup

Voici probablement le graphique qui résumera les calculs que nous avons fait en visqueux...

Even in the inviscid case, the thing is too small to make a blob and a neck

Pinch-off delayed by capillary Venturi

event

Pinch

2nd Pinch

Avoidance

There is not even the formation of a neck, thus no pinching and no avoidance

Direct pinch-off , no avoidance

fig5: la figure historiquejerome hoepffner



Timea) Retractation of a cone

Vortex shedding

Self-similar recoil

break-up

Vortex sheddingbreak-up

Time

fig6: la figure de la diversité

b) Ligament in extension

Figure à faire avec du colorant pour avoir des images plus souriantes?

jerome hoepffner


Droplet ejection using WunenburgerDroplet ejection using acoustic waves: Régis Wunenburger...

Documents

Transcript of Droplet ejection using WunenburgerDroplet ejection using acoustic waves: Régis Wunenburger...