Direct numerical simulations of convective flows Oceanography - Gas - - PowerPoint PPT Presentation

Direct numerical simulations of convective flows

Oceanography - Gas Transfer Sam Fredriksson, University of Gothenburg

Outline of this presentation

• Aim with this

presentation

• Background
• Route forward
• Work so far

Carbon Cycle – IPCC (2007)

How are CO2 fluxes determined?

Typically: 1.CO2 concentration in surface water (pCO2) 2.Flux model: F = k (Caq – Ceq) 3.Piston velocity/gas transfer koefficient (k) calculated from average wind speed at 10 m height.

• D. Bade 2009, Encycl. of Inland Waters

Surfactant influence of flow field and gas transfer

McKenna 2000 Surface vorticity Surface velocity Surface divergence Clean Oley alcohol II k as a funct. of time (surfactant concentration cholesterol)

Gas exchange research strategy

Forming a group of researchers Gas exchange and concentration measurements Wind measurements Turbulence measurements Infrared camera recording Numerical modeling

Route forward numerical simulations

General

• DNS (Direct Numerial Simulations)

with OpenFoam

• DNS with convection
• DNS with surface shear
• DNS with surfactant
• LES (Large Eddy Simulations)
• Waves …

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 urms/Ubar vrms/Ubar wrms/Ubar urms_Lada/Ubar urms_Lada/Ubar wrms_Lada/Ubar

Route forward numerical simulations

General

• DNS with OpenFoam
• DNS with convection
• DNS with surface shear
• DNS with surfactant
• LES/DNS moving surface
• Waves …

Lars Davidsson et al. ”DNS of plane vertical channel with and without buoyancy”, Turbulence, Heat and Mass Transfer, 2003

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0.2 0.4 0.6 0.8 1 1.2 1.4 Umean/Ubar Umean _Lada/Ubar

Route forward numerical simulations

General

• DNS with OpenFoam
• DNS with convection
• DNS with surface shear
• DNS with surfactant
• LES/DNS moving surface
• Waves …

Robert Handler: ”Direct numerical simulations of free convection beneath an air-water interface at low Rayleigh numbers”, Physics of Fluids, 2003

Computational domain

Evaporation

Boussinesq formulation

Momentum and Continuity

• U ∙ U 1
• ν

1 ,

• ∙ U 0

Linear approximation of density Temperature field is governed by a convection-diffusion equation given by

• U ∙ T T S

Temperature field, Raq ≈ 5x106

0.076 m Played 10 times real time

IR-measurements at Bornö

Magnus Gålfalk

≈ 1 m

Lz = 0.122 (Side = 0.17 m) Ra ~ QLz

4

Surfactant - Finite Area Method - fam

Areas for investigation

Analysis

• Mesh convergence – it seems that there is a need for finer meshes

with present formulation compared to Handler et al.

• Size of computational box and mesh resolution
• More thorough analysis and comparison against experiments etc.
• Surfactant formulation, cyclic boundaries for fam
• Find correlation between IR and gas transfer
• Find a good measure/representation of the temperature field (rms,

skewness, flatness, divergence etc.)

• Waves etc.

Thank you