Isolated Street Canyons
Recent interest has risen in employing LES models
[72, 73] to overcome the shortcomings of RANS in
its inability to capture the unsteady and inherent
fluctuations of the flow field within the street canyon
on which the dispersion of pollutants depends.
RANS, which is the most widely used approach in
industry for the modelling of turbulent flows,
assumes that non-convective transport in a turbulent
flow is governed by stochastic three-dimensional
turbulence possessing a broadband spectrum with
no distinct frequencies and, therefore, models the
entire range of the eddy length scales. This approach
has obvious weaknesses and poses serious uncertainties in flows for which large-scale organized
structures dominate, such as flows around buildings.
In addition, RANS models assume gradient transport, which may not be the case for pollutant
exchange at the roof level of a street canyon. LES,
although computationally more expensive, has an
advantage over RANS in that it explicitly resolves
the majority of the energy carrying large-scale organized structures and the internally or externally
induced periodicity involved, whereas only the universal small-scale eddies are modelled.
As an example, an illustration is provided of the
flow structure and pollutant dispersion within an
urban street canyon of width to height ratio W/
H ¼ 1 using the standard k-ε model, the Reynolds
Stress Model (RSM), and LES, coupled with the
advection-diffusion method for species transport
[74] (Fig. 5). In the case shown, the computational
domain was built by using about one million cells
with smallest dimensions equal to 0.077 H. Numerical results, which include the statistical properties
of pollutant dispersion, for example, the mean concentration distributions, three-dimensional spreads
of the pollutant, etc., are then compared to windtunnel measurements from the online database
[75]. Integrated into the model street, four tracer
gas emitting line sources were used for simulating
the release of traffic exhausts. Mean concentrations
are normalized according to:
c
þ
¼
CU ref H
Q=l
,
where C is the measured/calculated concentration,
H the building height, U ref the flow velocity at
height H in the undisturbed approaching flow, and
Q/l the tracer gas source strength per unit length.
Figure 5 shows normalized concentration contours at wall A (leeward) and wall B (windward)
of the street canyons.
It is observed that amongst the two RANS
models, RSM performed better than standard k-ε,
however, LES proved better than RANS in predicting the concentration distribution because it
was able to capture the unsteady and intermittent
fluctuations of the flow field, and hence, resolve the
transient mixing process within the street canyon.
Regular Building Arrays
Wind tunnel experiments have demonstrated that
street ventilation is reduced in the presence of
upstream buildings. This seems to be because of
the upward displacement of the flow and the consequent perturbed momentum exchange between
the street canyon and the outer region of the flow.
Also, numerical results have indicated that the
surrounding building configuration affects pollutant dispersion in a street canyon, therefore it should
be taken into account in numerical dispersion
modelling (see, for example, [76]). Two examples
of flow and pollutant dispersion distributions for
regular building geometries are shown in Fig. 6,
investigated by means of wind tunnel experiments
and numerical simulations as previously described.
The experiment is set up in a boundary layer
wind tunnel to simulate the case of dispersion of
pollutants from naturally ventilated underground
parking garages [77]. A finite array of idealized
building blocks with 0.1 m by 0.15 m base dimensions and 0.125 m height (three buildings crosswind and seven buildings along wind direction)
are considered, as shown in Fig. 6a. The aspect
ratio of the street canyons resulting from the
building arrangement is W/H ¼ 0.8. Four
ground-level emission sources were mounted
close to the building. RANS type simulations
[78] were performed by employing the standard
k-ε model and the advection-diffusion model. The
computational grid was built by using 500,000
cells, with smallest dimensions equal to 0.005 m.
Flow and dispersion were measured within the
street canyon positioned downwind of the building equipped with the sources.
182
Urban Air Quality: Meteorological Processes
Recent interest has risen in employing LES models
[72, 73] to overcome the shortcomings of RANS in
its inability to capture the unsteady and inherent
fluctuations of the flow field within the street canyon
on which the dispersion of pollutants depends.
RANS, which is the most widely used approach in
industry for the modelling of turbulent flows,
assumes that non-convective transport in a turbulent
flow is governed by stochastic three-dimensional
turbulence possessing a broadband spectrum with
no distinct frequencies and, therefore, models the
entire range of the eddy length scales. This approach
has obvious weaknesses and poses serious uncertainties in flows for which large-scale organized
structures dominate, such as flows around buildings.
In addition, RANS models assume gradient transport, which may not be the case for pollutant
exchange at the roof level of a street canyon. LES,
although computationally more expensive, has an
advantage over RANS in that it explicitly resolves
the majority of the energy carrying large-scale organized structures and the internally or externally
induced periodicity involved, whereas only the universal small-scale eddies are modelled.
As an example, an illustration is provided of the
flow structure and pollutant dispersion within an
urban street canyon of width to height ratio W/
H ¼ 1 using the standard k-ε model, the Reynolds
Stress Model (RSM), and LES, coupled with the
advection-diffusion method for species transport
[74] (Fig. 5). In the case shown, the computational
domain was built by using about one million cells
with smallest dimensions equal to 0.077 H. Numerical results, which include the statistical properties
of pollutant dispersion, for example, the mean concentration distributions, three-dimensional spreads
of the pollutant, etc., are then compared to windtunnel measurements from the online database
[75]. Integrated into the model street, four tracer
gas emitting line sources were used for simulating
the release of traffic exhausts. Mean concentrations
are normalized according to:
c
þ
¼
CU ref H
Q=l
,
where C is the measured/calculated concentration,
H the building height, U ref the flow velocity at
height H in the undisturbed approaching flow, and
Q/l the tracer gas source strength per unit length.
Figure 5 shows normalized concentration contours at wall A (leeward) and wall B (windward)
of the street canyons.
It is observed that amongst the two RANS
models, RSM performed better than standard k-ε,
however, LES proved better than RANS in predicting the concentration distribution because it
was able to capture the unsteady and intermittent
fluctuations of the flow field, and hence, resolve the
transient mixing process within the street canyon.
Regular Building Arrays
Wind tunnel experiments have demonstrated that
street ventilation is reduced in the presence of
upstream buildings. This seems to be because of
the upward displacement of the flow and the consequent perturbed momentum exchange between
the street canyon and the outer region of the flow.
Also, numerical results have indicated that the
surrounding building configuration affects pollutant dispersion in a street canyon, therefore it should
be taken into account in numerical dispersion
modelling (see, for example, [76]). Two examples
of flow and pollutant dispersion distributions for
regular building geometries are shown in Fig. 6,
investigated by means of wind tunnel experiments
and numerical simulations as previously described.
The experiment is set up in a boundary layer
wind tunnel to simulate the case of dispersion of
pollutants from naturally ventilated underground
parking garages [77]. A finite array of idealized
building blocks with 0.1 m by 0.15 m base dimensions and 0.125 m height (three buildings crosswind and seven buildings along wind direction)
are considered, as shown in Fig. 6a. The aspect
ratio of the street canyons resulting from the
building arrangement is W/H ¼ 0.8. Four
ground-level emission sources were mounted
close to the building. RANS type simulations
[78] were performed by employing the standard
k-ε model and the advection-diffusion model. The
computational grid was built by using 500,000
cells, with smallest dimensions equal to 0.005 m.
Flow and dispersion were measured within the
street canyon positioned downwind of the building equipped with the sources.
182
Urban Air Quality: Meteorological Processes