Detailed numerical modeling of local atmospheric dispersion in an idealized urban area

1
Detailed numerical modeling of local atmospheric
dispersion in an idealized urban area

• M. Milliez, S. Panzarella, B. Carissimo
• CEREA
• Research and Teaching Center for Atmospheric
  Environment
• Chatou/Marne-la-Vallée, FRANCE

2
Outline

• Objectives
• The Mercure model
• Simulation results evaluated with two
  experiments
• Hydraulic simulation of the MUST array
  (R.W.Macdonald, C.E.Ejim, 2002) results and
  comparison
• MUST Mock Urban Setting Test (C.A.Biltoft et
  al., 2001) results and preliminary comparison
• Perspectives

3
Objectives

• Investigate flows and pollution dispersion in an
  urban environment analyses of flow properties
  and concentration fields.
• In order to
• Study the impact on population and environment on
  a local scale
• Describe the average building effects (in terms
  of porosity, drag and turbulence) on flow and
  concentration fields

4
The Mercure model

• Developed by EDF and 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
• Eulerian approach
• Full scale, fine resolution, complex terrain,
  thermal effects
• Large scale meteo. conditions taken into account
• k-? turbulence closure model
• porosity/drag option

5
Hydraulic simulation of the MUST array

• Report Flow and Dispersion Data from a Hydraulic
  Simulation of the MUST array, R.W. Macdonald,
  C.E. Ejim,2002, University of Waterloo, Canada
• Hydraulic flume with an upstream region to
  simulate of a turbulent ABL flow in neutral
  stability conditions.
• Experiments
• 150 scale
• Array of 10x4 obstacles (200mm long, 50mm wide,
  50 mm high)
• Use of heat as a tracer

6
Hydraulic simulation of the MUST array

• continuous release
• several locations upstream and within the array
• 3 different flow direction 0 deg, 30 deg, 45 deg

7
Simulations with the Mercure model

• Mesh 900 000 elements
• Horizontal grid 0.5 m x 0.5m
• Stretched vertical grid 0.5m -gt 1.6 m

• Boundary conditions
• Upstream
• Dowstream gradients 0

8
Simulations with the Mercure model

• 0

• 30

• 45

9
Comparisions
Wind 0 , source upstream

• U

• TKE

10
Comparisions concentration
11
Comparisions concentration
12
The Mock Urban Setting Test

• C.A. Biltoft, et al. Report and data, 2001.
• Near full scale experiment in the U.S. Army
  Dugway Proving Ground (Utah), conducted for the
  DTRA (Defense Thread Reduction Agency )
• Objectives acquire meteo. and dispersion data
  set
• overcome the scaling limitations of laboratories
  simulations.
• Neutral gas releases in a field of containers.
• Array of 10x12 obstacles (12.9 m long, 2.42 m
  wide, 2.54 m high)

13
The Mock Urban Setting Test

• Releases for different meteorological conditions
  in several locations within the array
• 63 continous releases of duration of 15 min
• Data preanalyses and statistics

14
The Mock Urban Setting Test
N
Pneumatic Mast
32-m Tower (digiPIDs at 1-, 2-, 4-, 6-, 8-,10-,
16-m levels)
Line4 h1.6 m
Line5
Line3 h1.6 m
Line2 h1.6 m
Line1 h1.6 m
simulated release point Hs0.15 m
30
simulated wind
15
Simulations with the Mercure model

• Mesh 800 000 hexahedral elements
• Dimensions 240 m x 233 m x 32m

• Horizontal grid lower levels

4 m
0.6 to 1m
2 m
0.3 m

• Stretched vertical grid

4 m
16
Simulations with the Mercure model

• Boundary conditions
• Upstream wind profile in a stable atmosphere
• L100
• Z00.1m
• Wind bearing 28
• u3m/s (z8m)
• Dowstream
• gradients 0
• Top
• symetry and free slip

17
Simulations with the Mercure model
18
Simulations with the Mercure model

• Y86 m

Y106 m
Y146 m
19
Simulations with the Mercure model

• Y86 m

• Y106 m

• Y146 m

20
First comparisons wind speed
Row 5
21
First comparisons TKE
Row 5
22
First comparisons concentration
Row 5
23
First comparisons
24
Conclusions

• Water flume simulations
• Satisfactory results for wind, turbulence and C
  at 0
• Differences in side wall effects at 30, 45 ?
  comparison not conclusive
• MUST simulations
• first results encouraging comparison
• further analyses needed ( inflow profile )

25
Perspectives

• More analyses / comparisons with the MUST
  experiment
• Future thesis work take into account heat
  fluxes and radiative transfers with buildings

26
Acknowledgments

• Rob Macdonald, University of Waterloo (CA)
• water flume experimental data and analysis
• Defense Threat Reduction Agency (USA)
• MUST field experiment database
• B. Carissimo was supported during his sabbatical
  leave by the Comprehensive Atmospheric Modeling
  Program (CAMP) at George Mason University (USA)

27
Thank you
28
Simulation with the Mercure model