Comparative evaluation of an Comparative evaluation of an Eulerian - - PowerPoint PPT Presentation

Comparative evaluation of an Comparative evaluation of an Eulerian Eulerian CFD and Gaussian plume models for CFD and Gaussian plume models for roadside heavy metal deposition roadside heavy metal deposition roadside heavy metal deposition roadside heavy metal deposition

Mohammad Taghavi1, Bertrand Carissimo1, Karine Sartelet1, Karine Muller2, Irène Korsakissok1

1Centre d’Enseignement et de Recherche sur l’Environnement Atmosphérique (CEREA), Joint Laboratory École des Ponts ParisTech / EDF

R&D, Université Paris-Est, 6-8 Avenue Blaise Pascal, 77420, Champs sur Marne, France

2CETE de Lyon, 46 rue Saint-Théobald, 38081, L'Isle d'Abeau cedex, France

Corresponding author : taghavi@cerea.enpc.fr, Phone +33 164152145, Fax : +33 1 64 15 21 70

• utline
• utline
• Context

Context and objectives and objectives

• Tools and development

Tools and development

• Case study, Model configuration and Input data

Case study, Model configuration and Input data

• Case study, Model configuration and Input data

Case study, Model configuration and Input data

• Inter

Inter-comparison between models, Results comparison between models, Results

• Conclusions

Conclusions

Harmo13, Paris, France Harmo13, Paris, France

• Negative impact on air quality

+ efficient and flexible, important role in world’s economy and human’s life quality.

Context Context and objectives and objectives

Road transport systems: Benefits and problems Benefits and problems

21 toxic substances can mainly be assigned to road traffic. heavy metals are among them (e.g. Pb, Cd, Cu, Sb & Zn). Cadmium ranks 7th on U.S. EPA priority list of hazardous substances. In most European cities emission from road traffic has become the most important source of local air pollution. Harmo13, Paris, France Harmo13, Paris, France

Sector split emissions of primary and secondary fine particulates (Source: EEA)

(EEA member countries; EU-15; EU-27 - EU-15; EFTA-4 and CC-3; efta-4 and CC-3 Synonyms: Iceland, Liechtenstein, Norway, Switzerland)

Vehicular traffic emission contribution at regional scale Vehicular traffic emission contribution at regional scale

Context Context and objectives and objectives

Harmo13, Paris, France Harmo13, Paris, France

Near Near roadside roadside deposition deposition

• non-point source of particle-associated metals
• direct affection (populated area)

Context Context and objectives and objectives

Air pollution at near roadside Air pollution at near roadside

• indirect affection (unpopulated area)

Harmo13, Paris, France Harmo13, Paris, France In France (Source: USIRF) : ~ One million kilometers paved roads ~ Eleven thousand kilometers expressways Controls Policy

Measurements Valuable information about pollution levels Modeling Direct source - receptor relationship, Scenarios study

Context Context and objectives and objectives

Approaches, and appropriates tools Approaches, and appropriates tools Direct source - receptor relationship, Scenarios study Problem at local scale: Complex geometry (flow and turbulence are significantly affected by

the presence of groups of buildings and thus are widely inhomogeneous and unsteady)

Question: Which model is appropriate?

Harmo13, Paris, France Harmo13, Paris, France

Tools and development Tools and development

Mercure_Saturne Mercure_Saturne model model

1 CFD model: Mercure_Saturne 2 Gaussian models: a- ADMS urban V2 (Carruthers et al., 1997) b- Polyphemus_Gaussien (Korsakissok et al., 2008)

• Adapted for atmospheric modeling
• Taking into account complex geometries
• Meshes with any type of cell (tetrahedral, hexahedral, prismaticpyramidal,

polyhedral…)

• Any

type

• f

grid structure (unstructured, block structured, hybrid, conforming or with hanging nodes…)

• Gas and aerosol dispersion

And we added dry and wet deposition scheme

Harmo13, Paris, France Harmo13, Paris, France

Tools and development Tools and development

Dry and wet deposition schemes Dry and wet deposition schemes

• !

" #

• Simplified dry deposition scheme

Simplified wet deposition scheme Harmo13, Paris, France Harmo13, Paris, France

Tools and development Tools and development

Variation of dry deposition rate according to the diameter of the particle

n velocity (cm s-1)

Wind = 10 m/s

Deposition v

this work Roustan (2005)

Harmo13, Paris, France Harmo13, Paris, France

*

Case study, M Case study, Model configuration and Input data

• del configuration and Input data

Measurement campaign, study domain

Harmo13, Paris, France Harmo13, Paris, France

Case study, M Case study, Model configuration and Input data

• del configuration and Input data

Field of modeling, grid (b- structured grid)

Resolution: X: 1m with increasing rate of 1.1 Y: 10 m (constant) Z: 0.6m with increasing rate of 1.1 40,000 calculations nodes

Harmo13, Paris, France Harmo13, Paris, France

Case study, M Case study, Model configuration and Input data

• del configuration and Input data

Emission (Cadmium) Emission (Cadmium)

Modeling software: IMPACT ADEME (V2)

Mass quantity, size distribution, median diameter Tires & brakes wear

Data: Laboratory of Applied Thermodynamics (LAT) and the Transport Research Laboratory (TRL)

Exhaust

Modeling software: IMPACT ADEME (V2) Tires + Brakes (mg/km/j) Exhaust (mg/km/j) Repartition 2,5<PM<10 15,66 16 % 1<PM<2,5 9,27 10 % PM<1 2,82 69,63 74 % Total 27,75 69,63 100%

Result

Harmo13, Paris, France Harmo13, Paris, France

Median diameter for each mode:

PM1 related to the traffic is characterized by 95% particles lower than 0,5 microns (Santamaria et al., 1990; Lonati et al., 2006)

Case study, M Case study, Model configuration and Input data

• del configuration and Input data

Emission Emission Shape and position of emission source

For Mercure_Saturne: Homogenize concentration up to 1.5 m height at the boundary (inflow boundary condition) For Gaussian models: Point sources at 1 and 0.5 m height in the road center for ADMS and Polyphemus_Gaussian respectively

Harmo13, Paris, France Harmo13, Paris, France

Case study, M Case study, Model configuration and Input data

• del configuration and Input data

Background values of Cadmium deposition (Spatial& temporal variation)

regional scale, 2001 (Roustan, 2005) Time variation, 1996-98 (site 4)

Total cadmium deposition rate (gr.km-1.yr-1)

Harmo13, Paris, France Harmo13, Paris, France

Inter Inter-comparison with Gaussian models, Results comparison with Gaussian models, Results

February 3, 1997, weak dispersion (weak wind: 1m/s)

Harmo13, Paris, France Harmo13, Paris, France

Inter Inter-comparison with Gaussian models, Results comparison with Gaussian models, Results

February 20, 1997, normal dispersion (wind: 5.7 m/s)

Harmo13, Paris, France Harmo13, Paris, France

Inter Inter-comparison with Gaussian models, Results comparison with Gaussian models, Results

February 18, 1997, rainy day with normal dispersion (wind: 7.3 m/s)

Harmo13, Paris, France Harmo13, Paris, France Mercure_Saturne ADMS Polyphemus_Gaussian Total deposition

(pg.m-2.s-1)

15.5 12.6 15.0 Wet deposition

(pg.m-2.s-1)

0.39 0.17 0.71

Inter Inter-comparison with Gaussian models, Results comparison with Gaussian models, Results

Total deposition for February 1997 Total deposition for February 1997

Harmo13, Paris, France Harmo13, Paris, France

Inter Inter-comparison with Gaussian models, Results comparison with Gaussian models, Results

Taylor diagrams Taylor diagrams

Correlation: MS=0.97, AD=0.87, PG=0.98 RMS error: RMS error: MS=0.30, AD=0.63, PG=0.21 NSD: MS=0.78, AD=0.46, PG=0.88 Harmo13, Paris, France Harmo13, Paris, France

Conclusions Conclusions

• Size specific parameterization of particle dry deposition velocity and scavenging were

integrated within atmospheric CFD Mercure_Saturne.

• Total cadmium emission from exhaust (combustion) can be considered smaller than PM1
• In the near source area, Gaussian models result are very sensitive to injection height
• For an inert heavy metal like Cadmium, deposition rate is very sensitive to wind speed and

particles size

• Dry deposition is the determining deposition mode for heavy metals near the roads for the
• Dry deposition is the determining deposition mode for heavy metals near the roads for the

traffic emission.

• Our findings indicated cadmium deposition was increased in the immediate vicinity of a large

freeway, and quickly reduced to urban background deposition rates between 200 m downwind

• f the freeway.
• Cadmium was considered as an inert substance, from emission to deposition, same procedure

can be applied for other heavy metals provided by road traffics such as copper, lead and zinc. Harmo13, Paris, France Harmo13, Paris, France

Thank you for your attention Thank you for your attention

Harmo13, Paris, France Harmo13, Paris, France