A comparison of wind tunnel measurements and CFD model simulations - PowerPoint PPT Presentation

A comparison of wind tunnel measurements and CFD model simulations for pollutant dispersion at the portal of road tunnel for different wind directions M. Taghavi 1 , B. Carissimo 1 , F. Gourdol 2 , P. Mejean 2 , L. Soulhac 2 , C. Bernagaud 3 , and J.F. Burkhart 3 1 Centre d’Enseignement et de Recherche en Environnement Atmosphérique (CEREA), Paris, France 2 Ecole Central de Lyon, Laboratoire de Mécanique des Fluides et d'Acoustique, Lyon, France 3 Ministère de l'Ecologie, de l'Energie, du Développement Durable et de la Mer, Centre d'Etudes des Tunnels, Pôle Ventilation et Environnement, Lyon, France

Outline : • Context & Objective • Measurements campaign • Model configuration and input • Primary simulation analysis • Uncertainty in tunnel exit conditions • Final simulation evaluation • Conclusion

Context & Objective • Context • tunnel : solution to keep away the increasing traffic from residential areas (especially in urban areas). • tunnel portals problems concerning the air quality may arise due to the concentrated release of polluted air there. • important to have suitable tools to assess the dispersion from roadway tunnels for planning purposes.  Frequent construction of tunnels (increasing)  Obstacles (sometimes natural)  Urban area (traffic, noise, pollution, …) • Objective Code_Saturne CFD model Evaluation

Measurements campaign Schema of tunnel model in wind tunnel

Measurements campaign Description of the tunnel

Measurement campaign : visualization

Measurements campaign Description of the available datasets : A- two different tunnel geometries : canyon and slope B- different angles of wind / tunnel axis : 0, 30, 60, 90, 120, 150, 180 C- measured parameters : velocity and turbulence (hot wire) concentration and variances (FID)

Measurements campaign Different wind / tunnel speed ratio U/V U: tunnel exit speed V: incoming wind reference velocity

Model configuration and input data Mesh caracteristics (Orientation = 90 ° ) Number of cell : X-Y-Z = 68-74-40 (~200000 nodes) Resolution : Axe X= from 8.2 mm on tunnel axis to 190.4 mm on the edge Axe Y= constant at 50mm Axe Z= Variable from 6.4mm at the bottom to 263.3mm at the top

Model configuration and input data Mesh set up 0, 90, 180 (exact WT geometry) 30, 60,120, 150 (approximate WT geometry) (same mesh) Boundary conditions: • Wind, turbulence • Concentration

Model configuration and input data View of tunnel portal (Orientation = 90 ° )

Primary simulation, evaluation Simulation results Configuration 0 ° U/V=1 Wind tunnel results

Primary simulation, evaluation Simulation results Configuration 60 ° U/V=1 Wind tunnel results

Primary simulation, evaluation Simulation results Configuration 120 ° U/V=1 Wind tunnel results

Primary simulation, evaluation Simulation results Configuration 120 ° U/V=2 Wind tunnel results

Primary simulation, evaluation Simulation results Configuration 120 ° U/V=0.25 Wind tunnel results

Detailed horizontal profiles in the canyon (Configuration 0 ° ; U / V=1) Comparaison Modèle et Mesure pour profil vertical a: Comparaison Modèle et Mesure pour profil vertical a: configuration:0, x=1550 et Z=10mm (cas: UsurV=1) configuration:0, x=1000 et Z=10mm (cas: UsurV=1) 0.6 0.4 Cadim (Model) Cadim (Model) Cadim (mesure) Cadim (mesure) 0.3 0.5 0.3 0.4 0.2 Cadim Cadim 0.3 0.2 0.2 0.1 0.1 0.1 Y (mm) Y (mm) 0.0 0.0 -250.0 -200.0 -150.0 -100.0 -50.0 0.0 50.0 100.0 150.0 200.0 250.0 -250.0 -200.0 -150.0 -100.0 -50.0 0.0 50.0 100.0 150.0 200.0 250.0 Comparaison Modèle et Mesure pour profil vertical a: Comparaison Modèle et Mesure pour profil vertical a: configuration:0, x=1550 et Z=50mm (cas: UsurV=1) configuration:0, x=1000 et Z=35mm (cas: UsurV=1) 0.7 0.5 Cadim (Model) Cadim (Model) 0.4 Cadim (mesure) Cadim (mesure) 0.6 0.4 0.5 0.3 0.4 Cadim Cadim 0.3 0.2 0.3 0.2 0.2 0.1 0.1 0.1 Y (mm) Y (mm) 0.0 0.0 -250.0 -200.0 -150.0 -100.0 -50.0 0.0 50.0 100.0 150.0 200.0 250.0 -250.0 -200.0 -150.0 -100.0 -50.0 0.0 50.0 100.0 150.0 200.0 250.0

Uncertainty in tunnel exit conditions Effect of an increase of turbulence at the portal head Comparaison Modèle et Mesure pour profil vertical a: Comparaison Modèle et Mesure pour profil vertical a: configuration:90, x=0 et y=-100mm (cas: UsurV=2) configuration:90, x=0 et y=-500mm (cas: UsurV=2) 140 140 Turb=0.2 Diss=2 Turb=0.2 Diss=2 120 120 Turb=2 Diss=2 Turb=2 Diss=2 Turb=8 Diss=2 100 100 Turb=8 Diss=2 Turb=20 Diss=2 Cadim (mesure) 80 Turb=20 Diss=2 80 Cadim (mesure) 60 60 Z (mm) Z (mm) 40 40 20 20 C adim C adim 0 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 -20 -20 -40 -40 -60 -60 Comparaison Modèle et Mesure pour profil vertical a: Comparaison Modèle et Mesure pour profil vertical a: configuration:90, x=0 et y=-250mm (cas: UsurV=2) configuration:90, x=0 et y=-1000mm (cas: UsurV=2) 140 140 Turb=0.2 Diss=2 Turb=0.2 Diss=2 120 120 Turb=2 Diss=2 Turb=2 Diss=2 Turb=8 Diss=2 100 100 Turb=8 Diss=2 Turb=20 Diss=2 Turb=20 Diss=2 80 Cadim (mesure) 80 Cadim (mesure) 60 60 Z (mm) Z (mm) 40 40 20 20 C adim C adim 0 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 -20 -20 -40 -40 -60 -60

Uncertainty Statistical model evaluation : Tool recommended : BOOT ( Model Validation Kit, http://www.harmo.org/kit ) Scatter Plots

Uncertainty Statistical measures of model/measurements differences : - fractional bias (FB)  - the correlation coefficient (R)  - the normalized mean square error (NMSE)  - fraction of predictions within a factor of two of observations (FAC2)  - the geometric variance (VG)  - the geometric mean bias (MG) 

Sensitivity tests Turbulence and dissipation variation (tunnel head) (Configuration : 90 ° Case : U/V= 2) Inside the canyon: MODEL NMSE CORR FA2 FB OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000, FB=FBfn-FBfp) T02D2 0.12 0.695 0.869 -0.268 (FBfn= 0.007, FBfp= 0.274, FB=FBfn-FBfp) T2D2 0.08 0.821 0.953 -0.195 (FBfn= 0.011, FBfp= 0.205, FB=FBfn-FBfp) T2D30 0.08 0.785 0.935 -0.201 (FBfn= 0.011, FBfp= 0.211, FB=FBfn-FBfp) T8D2 0.05 0.833 0.981 -0.127 (FBfn= 0.020, FBfp= 0.146, FB=FBfn-FBfp) T20D2 0.04 0.832 0.981 -0.073 (FBfn= 0.033 , FBfp= 0.106 , FB=FBfn-FBfp) T20D20 0.04 0.832 0.981 -0.074 (FBfn= 0.033, FBfp= 0.107, FB=FBfn-FBfp) T30D2 0.15 0.760 0.738 0.137 (FBfn= 0.204, FBfp= 0.067, FB=FBfn-FBfp ) Inside and outside the canyon (all): MODEL NMSE CORR FA2 FB OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000, FB=FBfn-FBfp) T02D2 0.22 0.917 0.811 -0.288 (FBfn= 0.017, FBfp= 0.306, FB=FBfn-FBfp) T2D2 0.20 0.904 0.659 -0.229 (FBfn= 0.045, FBfp= 0.274, FB=FBfn-FBfp) T2D30 0.19 0.910 0.672 -0.231 (FBfn= 0.043, FBfp= 0.274, FB=FBfn-FBfp) T8D2 0.21 0.877 0.615 -0.156 (FBfn= 0.082 , FBfp= 0.239 , FB=FBfn-FBfp) T20D2 0.23 0.857 0.568 -0.096 (FBfn= 0.119, FBfp= 0.215, FB=FBfn-FBfp) T20D20 0.23 0.857 0.568 -0.098 (FBfn= 0.119, FBfp= 0.216, FB=FBfn-FBfp) T30D2 0.41 0.780 0.439 0.058 (FBfn= 0.261, FBfp= 0.203, FB=FBfn-FBfp )

Sensitivity tests Turbulence and dissipation variation (tunnel head) (Configuration : 90 ° Case : U/V= 2)

Final simulation, evaluation (90 ° case) Concentration (adim) Velocity U (m/s) -------------------------- Configuration 90° (concentration) ------------------------ ------------------------ Configuration 90° (Vitesse du vent) ------------------------ U/V=0.25 U/V=0.25 MODEL NMSE CORR FA2 FB MODEL NMSE CORR FA2 FB OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) MOD. 0.13 0.953 0.833 -0.099 (FBfn= 0.063, FBfp= 0.162) MOD. 0.15 0.657 0.985 0.269 (FBfn= 0.298, FBfp= 0.029) U/V=0.5 U/V=0.5 MODEL NMSE CORR FA2 FB MODEL NMSE CORR FA2 FB OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) MOD. 0.28 0.913 0.813 -0.263 (FBfn= 0.040, FBfp= 0.302) MOD. 0.05 0.783 0.984 -0.008 (FBfn= 0.084, FBfp= 0.091) U/V=1 U/V=1 MODEL NMSE CORR FA2 FB MODEL NMSE CORR FA2 FB OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) MOD. 0.21 0.909 0.722 -0.145 (FBfn= 0.068, FBfp= 0.213) MOD. 0.03 0.776 0.984 -0.021 (FBfn= 0.050, FBfp= 0.071) U/V=2 U/V=2 MODEL NMSE CORR FA2 FB MODEL NMSE CORR FA2 FB OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) OBS. 0.00 1.000 1.000 0.000 (FBfn= 0.000, FBfp= 0.000) MOD. 0.20 0.899 0.698 -0.158 (FBfn= 0.070, FBfp= 0.228) MOD. 0.03 0.939 0.962 -0.083 (FBfn= 0.025, FBfp= 0.108)