Environmental Engineering Reference
In-Depth Information
was logarithmically extrapolated from the measured data available at three levels
up to 16 m from the South tower. No initial profile was set for the turbulent
kinetic energy. In all the simulations the timestep was 4 × 10
-3
s.
We recall that RAMS, as a meteorological model, was not designed to produce
steady-state conditions. However, we set the boundary conditions so to approach
as close as possible a steady state solution, verifying that a quasi-steady flow was
reached after 4 min of simulation. The dispersion simulation was conducted using
a passive tracer, released after the model reached its quasi-steady condition. The
source in RAMS6-mod corresponded to a grid point at x = −77.46 m, y = 67.47 m
and z = 1.8 m; the emission was continuous.
2. Results and Summary
In the following, some simulations results for MUST field case n. 2681829 are
presented, exemplifying RAMS6-mod performance, and compared with observed
data. We note that the only difference between the simulations presented here is in
the turbulence closure. With respect to horizontal and vertical contour plots of
tracer concentrations (not shown here) we notice that the turbulence closure scheme
has a significant impact on the concentration pattern, both in the distribution and
the shape of the plume. We found out that k-∑ scheme produces larger vertical
spreads, while RNG-k∑ shows a slightly narrower and smoother horizontal distri-
bution of the plume. The deflection of the plume's centerline with respect to the
direction of the upstream wind was reproduced by both closure models in a
similar way.
25
25
25
20
20
20
15
15
15
Z (m)
10
10
10
5
5
5
0
0
0
0
5
speed (ms
-1
)
10
0
1
2
tke (ms
-2
)
3
4
5
0
2
4
Conc (ppm)
Fig. 2.
Wind speed, turbulence and concentration profiles at T tower. Crosses: observations; line
and triangles: k-∑ run; solid line: RNG-k∑ run
