Wind tunnel experiments and CFD simulations show that a recirculation flow region
forms within a street canyon exposed to a perpendicular flow. Both wind tunnel measurements
and CFD predictions show that for aspect ratio
W/H ¼ 0.8 a single vortex forms inside the canyon. In particular, Fig. 6 shows velocity vectors
on the vertical plane in the middle of the street
canyon obtained from wind tunnel experiments
(Fig. 6b) and CFD simulations (Fig. 6c). The
figure shows that the shape and dimension of
the vortex predicted by the CFD simulations is
very similar to that observed by CEDVAL experiments [77] compiled at the University of Hamburg. The ability to capture the right dimension
of the vortex is very important as the vortex
drives the distribution of pollutants within the
street canyon.
As outlined in [79, 80], it should be noted that
typical CFD codes predict dispersion spread that
is smaller than that predicted by the well-validated
integral model ADMS-Urban [81]. From a sensitivity test on the turbulent Schmidt number value
(the standard value is 0.7), the value of 0.4 was
found to be the most appropriate in order to artificially increase plume dispersion, even though
such tuning should be avoided. Figure 6d (top)
shows the dimensionless coordinate y/H versus
the dimensionless concentration K near the floor
(z/H ¼ 0.06) within the street canyon downwind
the source building and dimensionless concentration K versus dimensionless coordinate x/H on the
same vertical plane (bottom). The horizontal
velocity near the floor is negative because the
vortex in the canyon is clockwise. Thus, pollutants are carried toward the leeward side and mixed
0.5
1
0
0.5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
WALL B
Line Sources
Street Canyon
WT
k-e
0.5
1
0
RSM
Z/H
0.5
y/H
Normalized Concentration, c +
1
0
LES
1
0
0.5
1
0
0.5
a
a
b
b
c
d
W = H
0.23 H
0.35 H
H
linesources
−5
−4
8
8
1 6
1 6
2 4
24
32
32
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
WALL A
0.5
1
0
0.5
1
0
1
0
Urban Air Quality: Meteorological Processes,
Fig. 5 (a) Sketch of the street canyon with indication of
the position of line source, grid, and wind tunnel model.
(From [74] CODASC, 2008). (b) Mean normalized
concentrations at wall A (leeward) and wall
B (windward) showing comparison between WT data
(available from CODASC) and numerical results
performed with k-ε, RSM, and LES simulation
Urban Air Quality: Meteorological Processes
183
forms within a street canyon exposed to a perpendicular flow. Both wind tunnel measurements
and CFD predictions show that for aspect ratio
W/H ¼ 0.8 a single vortex forms inside the canyon. In particular, Fig. 6 shows velocity vectors
on the vertical plane in the middle of the street
canyon obtained from wind tunnel experiments
(Fig. 6b) and CFD simulations (Fig. 6c). The
figure shows that the shape and dimension of
the vortex predicted by the CFD simulations is
very similar to that observed by CEDVAL experiments [77] compiled at the University of Hamburg. The ability to capture the right dimension
of the vortex is very important as the vortex
drives the distribution of pollutants within the
street canyon.
As outlined in [79, 80], it should be noted that
typical CFD codes predict dispersion spread that
is smaller than that predicted by the well-validated
integral model ADMS-Urban [81]. From a sensitivity test on the turbulent Schmidt number value
(the standard value is 0.7), the value of 0.4 was
found to be the most appropriate in order to artificially increase plume dispersion, even though
such tuning should be avoided. Figure 6d (top)
shows the dimensionless coordinate y/H versus
the dimensionless concentration K near the floor
(z/H ¼ 0.06) within the street canyon downwind
the source building and dimensionless concentration K versus dimensionless coordinate x/H on the
same vertical plane (bottom). The horizontal
velocity near the floor is negative because the
vortex in the canyon is clockwise. Thus, pollutants are carried toward the leeward side and mixed
0.5
1
0
0.5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
WALL B
Line Sources
Street Canyon
WT
k-e
0.5
1
0
RSM
Z/H
0.5
y/H
Normalized Concentration, c +
1
0
LES
1
0
0.5
1
0
0.5
a
a
b
b
c
d
W = H
0.23 H
0.35 H
H
linesources
−5
−4
8
8
1 6
1 6
2 4
24
32
32
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
−5
−4
−3
−2
−1
0
1
2
3
4
5
WALL A
0.5
1
0
0.5
1
0
1
0
Urban Air Quality: Meteorological Processes,
Fig. 5 (a) Sketch of the street canyon with indication of
the position of line source, grid, and wind tunnel model.
(From [74] CODASC, 2008). (b) Mean normalized
concentrations at wall A (leeward) and wall
B (windward) showing comparison between WT data
(available from CODASC) and numerical results
performed with k-ε, RSM, and LES simulation
Urban Air Quality: Meteorological Processes
183