Atmospheric CFD modelling for environmental applications at local scale

(1)

Atmospheric CFD modelling

for environmental applications

at local scale

B. Carissimo, S. Lacour, H. Foudhil, L. Musson-Genon, E. Dupont, D. Wendum,

M. Milliez, B. Albriet, E. Demael, L. Laporte,

(2)

Outline

•

Objectives

•

Focus : concentration fluctuations

•

Overview : other activities

(3)

Animation Wind tunnel

Objectives

Develop an integrated modelling tool for studies of the

atmospheric environment at local scale

Dispersion of pollutants and its impact on population and

environment in urban and around industrial areas.

Other impact studies : reactive pollutants, particle

formation and dispersion, noise propagation …

Accidental releases,

Population comfort (wind and turbulence prediction),

(4)

The model: Mercure_Saturne

Developed by CEREA

3-D model adapted to atmospheric flow and dispersion

simulation

Core of the model: CFD model Code_Saturne (EDF) which

can handle complex geometry and complex physics

Unstructured grid, finite volumes Simulations:

•

• Full scale, fine resolution, complex terrainFull scale, fine resolution, complex terrain

•

• Large scale Large scale meteometeo. conditions taken into account. conditions taken into account

•

• kk--

ε

turbulence closure modelturbulence closure model

•

• neutral neutral gazgaz releases, point source, continuous releasesreleases, point source, continuous releases

•

(5)

Concentration time series (wind tunnel measurements) 0 200 400 600 800 1000 Tsite en seconde 0 200 400 600 800 1000 Ci -C BF en ppm 0 200 400 600 800 1000 Tsite en seconde 0 100 200 300 400 Ci -C BF en pp m Ci Concentration moyenne At 1000 m High fluctuations near the source

(6)

Focus : concentration

fluctuations

•

Two approaches

:

•

Hybrid

RANS / PDF

(

PhD Thesis

A.

Radicchi

)

•

Pure RANS

(

(7)

Hybrid approach

RANS/PDF

RANS (k-ε) PDF (SLM¹) u_i , P k ,

f

(PDF Eulérienne de la concentration)

(8)

[ RANS / PDF ]

a

Résolution des EDS pour la trajectoire d'un

échantillon de particules de fluide,

dx

_p,i

u

_p,i

dt

du

_p,i

1

P

x

_i

dt

u

_p,i

u

_i

T

_L

dt

C

0

dW

i u_i , P , x_p,i,u_p,i W_i, T_L

Champs moyens (RANS) Champs instantanés (EDS) Proccessus de Wiener,

Temp characterisitique de la turbulence

(9)

[ RANS / PDF ]

b

Résolution de l'equation pour le melange de la

concentration transportée par chaque particule de fluide,

d

_p p m

dt S

dt

p m Conc. moyenne

Conc. instantanée particule p Temp de mélange

(10)

(11)

Pure RANS approach :

Concentration mean: j j j i j j x c u x C D x x C U t C ∂ ∂ − ∂ ∂ ∂ ∂ = ∂ ∂ + ∂ ∂ ( ' ') ) ( ) ( ρ ρ ) ( ' ' j t t j x C Sc c u ∂ ∂ = − ρ µ µt t Sc turbulent viscosity ε ρ µ_t = C_µ k2 Schmidt number Concentration variance: j j j j j j j i j j x c x c D x C c u x c u x c D x x c U t c ∂ ∂ ∂ ∂ − ∂ ∂ − ∂ ∂ − ∂ ∂ ∂ ∂ = ∂ ∂ + ∂ ∂ ' ' 2 ' ' 2 ) ' ' ( ) ' ( ) ' ' ( 2 2 2 2 ρ ρ ρ ) ' ( ' ' 2 2 j t t j x c Sc c u ∂ ∂ = − ρ µ k c R x c x c D f c j j ε ρ ε ρ _'2 ' ' 2 = = ∂ ∂ ∂ ∂

(12)

The Mock Urban Setting Test

(MUST)

Biltoft (2001).

Yee and Biltoft (2004).

Near full scale experiment

in the U.S. Army Dugway Proving Ground (Utah), conducted for the DTRA (Defense Threat

(13)

Simulations with

Mercure

~4 m ~2 m 0.6 to 1m ~0.3 m ~4 m Mesh: ~800 000 hexahedral elements Dimensions: 240 m x 233 m x 32m

Horizontal grid : lower levels Stretched vertical grid

MUST: 20 selected cases

Detailed study of dynamics and mean concentration with

(14)

Simulations with

Mercure

Root mean square of concentration fluctuations (ppm)

(15)

Simulations with

Mercure

(16)

Simulations with

Mercure

20 simulated cases:

•

• MG = exp (|MG = exp (|lnln Co| Co| –– ||lnln Cp|)Cp|)

•

• VG = exp [|(VG = exp [|(lnln CoCo--lnln Cp) Cp) 22 _|]_|]

•

• FAC2 = fraction of data that FAC2 = fraction of data that satisfy 0.5<Cp/Co<2

satisfy 0.5<Cp/Co<2

Co = MUST observations Co = MUST observations Cp =

Cp = MercureMercure predictions predictions |x| = average over the data set |x| = average over the data set

53.8 % 2.56 0.66 Masts A, B, C, D 50.3 % 2.70 0.51 Tower T 66.4 % 2.22 1.05 All horizontal 60.1 % 2.40 0.82 All 52.6 % 2.61 0.61 All vertical 67.0 % 1.368 1.156 Line4 69.2 % 1.341 0.980 Line3 67.2 % 1.437 1.028 Line2 63.5 % 3.31 0.94 Line1 FAC2 VG MG

Root mean square of concentration fluctuations

(17)

Simulations with

Mercure

10 simulated cases:

FAC2 (neutral) = 46.9 % FAC2 (stable) = 53.3 %

+ 6 %

(18)

Simulations with

Mercure

c’(source) = 1% C(source)

(19)

Cloud modeling : plume + fog

Impact of building and complex terrain on

dispersion

Chemistry of reactive plumes

Aerosols formation (local scale)

Wind energy

Radiative effects

operational ABL prediction (SIRTA)

Traffic and tunnel

}

Following presentations

(20)

Calcul code Mercure iso-concentration

eau liquide Panache Bugey

(21)

Fig. 2 Comparison of the in-situ spectra measured by the aircraft with the simulated ones. a) Horizontal slice at z=1320m above the ground for case B ( -- -- : iso-LWC and —— : iso-N: c). The symbol ∗ shows the maximum of LWC, and the circles o are part of the same arc whose center is ∗. b) shows the measured spectra (black) 660m far from the ∗ and the log-normal curves (blue) obtained from the LWC and Nc values given by the code at the position of the shaded circles (located 660m far from the ∗).

a)

(22)

Stage C. Samba

Évolution de la visibilité

( )0.88 0 7 . 144 ) 02 . 0 ln( l q avec VIS β ρ β = − =

Nouveau paramètre d’étude : la visibilité

Te mps (heure ) A lti tu de ( m ) vis ibilite 1 2 3 4 5 6 7 8 9 0 20 40 60 80 100 120 140 160 180 200 220 10 20 30 40 50 60 70 80 90 Mesures CAILLOU MERCURE Variations spatio-temporelles de la visibilité

3 phases dans l’évolution du brouillard :

La formation, le développement vertical et la dissipation de la couche de brouillard

(23)

IV. Résultats

Spectre des gouttelettes à 2m

Diamètre moyen de l’ordre de 10 µm et 20 µm Nc de l’ordre de quelques centaines de gouttelettes par cm³

Distribution des gouttelettes de nuages (cm-3/µm) à 2m en fonction du diamètre des gouttelettes (µm)

(24)

dispersion

Wind energy

Local ABL prediction (SIRTA)

Traffic and tunnel

(25)

(26)

Résultats pour la situation de vent de Sud (Mercure_Saturne 1.1.3)

(27)

(28)

Overview : other activities

dispersion

Wind energy

Traffic and tunnel

(29)

(30)

(31)

dispersion

Wind energy

Traffic and tunnel

(32)

(33)

dispersion

Wind energy

Traffic and tunnel

(34)

Wind energy : turbulence in complex terrain

and mask effect for large parks (Thesis L. Laporte)