Numerical Simulation of Atmospheric Flow within Urban Morphology

1
Numerical Simulation of Atmospheric Flow within
Urban Morphology

• Sang-Mi Lee, D. Zajic, J.-J. Kima, A. Rubio
• and H. J. S. Fernando
•  
• Environmental Fluid Dynamics Program
• Arizona State University
• aSchool of Earth and Environmental Sciences,
• Seoul National University, Korea

2
Outline

• Multiscale Modeling System
• Validation of the CFD using MUST
• Evaluation of the CFD QWIC based on a wind
  tunnel data

3
Multi-Scale Modeling Approaches
Mesoscale Meteorological Model Background flow
field
Synoptic flow
Local circulation
Urban perturbation
Computational Fluid Dynamic Model perturbations
induced by urban morphology
4
Nested Simulation of MM5 CFD
5
The Fifth-generation Penn State/NCAR Mesoscale
Model (MM5)

• Terrain following ? - coordinate
• Non-hydrostatic dynamics
• Four-dimensional data assimilation
• Multiple nest capability
• Physics

6
Governing Equations (in a Boussinesq airflow
system with the Coriolis effect neglected)
(1) Momentum equation
(2) Mass continuity equation
(3) Thermodynamic energy equation
(4) Transport equation for a passive pollutant
7
k-e turbulence closure scheme model
(5) Parameterization of Reynolds stresses
(6) Equations for turbulent kinetic energy and
its dissipation rate
8
Downslope flow 0500 LST
9
Upslope flow 1100 LST
10
Urban database of Oklahoma City
11
Experimental Setup

• Number of Computational Cells
• 140 X 140 X 60 in EW, NS, height, respectively
• Grid distance 10 m X 10 m X 2 m in x, y, z
  direction

12
Westerly Inflow
13
Northwesterly Inflow
14
Part II

• Evaluation of the CFD
• CFD vs. MUST

15
MUST (Mock Urban Setting Test)
Size of array is 176 m by 180 m
Dimensions of buildings are 12.2 x 2.48 x 2.54 m
Frontal packing and plan area density were 0.12
16
First Urban Canyon
17
Computational Domain
110 x 100 x 40 grid elements
18
Validation of CFD model Preliminary test
with MUST
19
Comparison (single building)
20
Part III

• Evaluation of CFD and QWIC
• vs. Wind Tunnel Data

21
QWIC

• QWIC-URB
• Computing wind field based on empirical relations
  around obstacles
• QWIC-PLUME
• Lagrangian particle model

22
QWIC-URB
23
Displacement Zone

• The length of the displacement zone.
• The shape of the displacement zone.

24
Cavity Zone

• The length of the cavity zone.
• The shape of the cavity zone.
• U-velocity in cavity zone.

25
Wake Zone

• The downwind decay parameter.
• The wind in the wake zone.

26
Mass Conservation

• Variational method to enforce mass conservation

27
Lagrangian Particle Model

• X xp U?t ?t(up u)/2
• Y yp V?t ?t(vp v)/2
• Z zp W?t ?t(wp w)/2

28
QWIC-URB (cont.)
29
QWIC-URB (cont.)
30
Numerical Configurations

• Domain 100x100x40m3
• Grid size
• ?x ?y ?z 1m
• Obstacle 10x10x10m3
• Wind direction normal to the face of the
  obstacle.

31
Approaching Flow
32
Wind Tunnel Data (Hosker 1984)
33
CFD model Output
34
QWIC model Output
35
Cavity Zone One cube length behind the center
of an obstacle.
36
Cavity Zone Two cube lengths behind center of
obstacle.
37
Wake Zone Three cube lengths behind center of
obstacle.
38
Summary

• A Urban scale airflow was simulated in a nested
  application of MM5 CFD.
• The CFD tends to over-predict TKE at building top
  level, while predicted mean variables showed
  reasonable agreement.
• Comparing wind tunnel data, CFD over-predicted
  wind speed above a building level, while QWIC
  under-predicted.

39
(No Transcript)