• Aucun résultat trouvé

POSTER

N/A
N/A
Protected

Academic year: 2022

Partager "POSTER"

Copied!
1
0
0

Texte intégral

(1)

Lagrangian compressible dynamics in a self-similar incompressible jet

Thomas Basset

1

† , Bianca Viggiano

2

, Thomas Barois

3

, Mathieu Gibert

4

, Nicolas Mordant

5

, Raúl Bayoán Cal

2

, Romain Volk

1

, Mickaël Bourgoin

1

1

Laboratoire de Physique, École Normale Supérieure de Lyon, France.

2

Department of Mechanical and Materials Engineering, Portland State University, USA.

3

Laboratoire Ondes et Matière d’Aquitaine, Université de Bordeaux, France.

4

Institut Néel,

Université Grenoble Alpes, France.

5

Laboratoire des Écoulements Géophysiques et Industriels, Université Grenoble Alpes, France.

A turbulent water jet seeded with tracers

40cm

) A turbulent round free jet...

- round nozzle with a diameter D = 4 mm

- nozzle exit velocity UJ ' 7 m/s

ReD = UJD/ν ' 2.8 × 104 ) ... seeded with tracers...

- neutrally buoyant spherical polystyrene tracers with a diameter dp = 250 µm

- specific seeding only through the nozzle

nozzle seeding ) ... tracked by cameras.

- three high-speed cameras oriented orthogonal to the brown faces (icosahedral geometry) in a back light configuration

- measurement volume spanning 100 mm z 180 mm (jet axis z and z = 0 the nozzle exit position)

- 50 movies of 8000 frames recorded at 6000 fps (inlet valve open some seconds before each recording)

Particle tracking velocimetry

0 200 400 600 800 1000 1200

800

600

400

200

30 40 50 60 70 80

à

Particle detection → Stereoscopic reconstruction [1, 2] → Tracking - Cylindrical coordinates (z, r, θ)

- Tracks convolved with a first-order derivative Gaussian kernel ⇒ velocity

⇒ ' 108 particle positions and velocities

References

[1] Machicoane et al. 2019, Rev. Sci. Instrum. 90 (3), 035112.

[2] Bourgoin & Huisman 2020, Rev. Sci. Instrum. 91 (8), 085105.

[3] Pope 2000, Turbulent Flows, Cambridge University Press.

Mean axial velocity hU (z, r )i

Mean velocity field well-known through Eulerian measurements [3]

- Centerline velocity: U0(z) = hU(z, r = 0)i - Half-width: hU(z, r = r1/2(z))i = 12U0(z)

In the self-similar region (for z & 15D = 60 mm)

U0(z)

UJ = B

(z − z0)/D r1/2(z) = S(z − z0)

⇒ Experimental results: z0 ' 4D, B = 5.3 and S = 0.105

Self-similar radial profile: f(η) = hU(z, r)i/U0(z) with η = r/(zz0)

f(η) = e−Aη2

with A = log(2)/S2

⇒ Same axial velocity field despite the nozzle seeding

Mean radial velocity hV (z, r )i (1/2)

Self-similar radial profile: g(η) = hV (z, r)i/U0(z)

) Continuity equation for the flow velocity field: ∇· U = 0

→b η(ηf (η))0 = (ηg(η))0 [3]

⇒b g(η) = ηf (η) − 1 η

Z η

0 xf (x) dx (1)

Incorrect function (1): flow velocity field U 6= Uϕ tracer velocity field

⇒ Entrained flow not tagged by the tracers due to the nozzle seeding (higher measured radial velocity): influence of entrainment up to the core of the jet

Mean radial velocity hV (z, r )i (2/2)

) Mean tracer density: ϕ(z, r) with ϕ0(z) = ϕ(z, r = 0) Self-similar radial profile: Φ(η) = ϕ(z, r)/ϕ0(z)

φ0(z) ∝ 1 zz0

⇒ Same behavior for Φ(η) and f(η) (except wider Φ(η))

) Continuity equation for the tracer velocity field: ∇·Uϕ) = 0

Φ b gϕ(η) = ηfϕ(η) (2) with fϕ(η) ' f(η) from experiments

Correct function (2): smaller maximum and negative part for η & 0.2 due to some entrained tracers remaining in the tank

⇒ A new model coherent with experimental results

g(η) = gϕ(η) − 1 η

Z η

0 xf(x) dx

| {z }

entrainment term

(3)

A diffusive model

Advection-diffusion equation: ∇·UKT∇ϕ) = 0

with KT the turbulent diffusion coefficient, KcT = KT/(U0r1/2) normalized To be coherent with (3) ⇒

b

KcT(η) = − 1 S

Φ(η) Φ0(η)

1 η

Z η

0 xf(x) dx

⇒ Link KT to the entrainment term and the “compressibility” Φ0(η)/Φ(η) KT linked to the known turbulent viscosity νT through the turbulent Prandtl number σT = νT/KT ' 0.6

⇒ Coherent with scalar transport values [3]

Conclusion

Lagrangian tracer velocity field ∇· Uϕ 6= 0 due to the nozzle seeding but

∇·Uϕ) = 0 and ∇·UKT∇ϕ) = 0 with the new turbulent diffusion coefficient KT linked with entrainment

† Contact: [email protected]

Références

Documents relatifs

Spots due to the cold 1s 3 metastable atoms were recorded when both lasers were off, and the video frames at different time intervals showed the images generated only by

Poursuivre cette histoire longue dans le contexte de l’entreprise, c’est reconnaître que le gouvernement de l’entreprise, son pouvoir exécutif, peut ainsi être repensé

Malheureusement, les déclarations du Président de la République jeudi dernier concernant une éventuelle transformation du CNRS en agence de moyens, suivies de votre

Hocevar, “A reduced complexity decoder architecture via layered decoding of LDPC codes,” in IEEE Workshop on Signal Processing Systems, SIPS 2004, Oct.. Veidenbaum, “Guest

Designing LDPC decoder based on our ASIP architecture model requires a design flow to automatically generate the SIMD matrix, the memory mapping and the PU execution spe-

Comme dans le dispositif des fentes d’Young, deux ondes venant de la même source interfèrent.. Cette lame permet également de recombiner les deux ondes pour

Nous avons pu constater qu’il existe un très large éventail de compétences disponibles dans les laboratoires aquitains, avec des équipes qui ont une reconnaissance nationale

Théorie des nombres : problèmes diophantiens, points rationnels, représentations galoisiennes, géométrie arithmétique, théorie de Hodge p-adique, géométrie des nombres,