Application of ANSYS-FLUENT for by ps94506


									        Application of ANSYS-FLUENT for meso-scale
                atmospheric flow simulations

                                Gergely Kristóf Ph.D., Norbert Rácz, Miklós Balogh

    Department of Fluid Mechanics, Budapest University of Technology and Economics, Budapest,

CFD simulation packages are known not being useful for modeling atmospheric flows involving
thermal stratification effects such as plume rise, gravity waves, flow around high mountains, or thermal
convections such as see-breeze, heat islands. In order to overcome the underlying mathematical
difficulties in simulation of the processes in atmospheric flows add-on module has been developed for
ANSYS-FLUENT in the form of user defined functions. The new simulation method has been
successfully validated with high accuracy data from laboratory scale thermal convection and gravity
wave experiments. Further development is going on to include moisture transport and its
thermodynamic effects into the simulation system. The applications demonstrate ANSYS-FLUENT's
flexibility in physical modeling and raise hopes that this approach can establish a powerful new
analysis tool for environmental technology and climatology.

atmospheric flow, urban climatology, pollution assessment, gravity waves

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany

1. Introduction

Analyses of flow around built structures as well as the concentration field in the vicinity of pollution
sources are well known application fields of CFD. These investigations usually concentrate on the air
flow within a relatively thin layer, e.g. couple of ten meters, above the ground. Several effects need to
be added to the simulation model when the height of the computational domain needs to be extended.
The most important effects are:
1. thermal stratification;
2. cooling caused by adiabatic expansion related to rising air;
3. compressibility;
4. Coriolis force;
5. moisture transport.
Stable thermal stratification, which is usually present in the atmosphere, tends the flow attacking high
mountain be forced around the mountain, rather allowing flow over the mountain, thus causing a
significant change in important technical characteristics, such as wind power potential or dispersion of
pollutants. Stratification has a strong, usually damping, effect on turbulence as well. Correct shape of
chimney plumes cannot be calculated without taking into account the thermal stratification and the
adiabatic expansion effects. Coriolis force causes significant change in the wind direction, known as
Ekman spiral, even in the first couple of ten meters of the atmospheric boundary layer. Variation of
density due to the changing hydrostatic pressure can also be an important effect above an altitude of a
few (one or two) kilometers.
There are a number of highly developed meso-scale atmospheric simulation codes, such as MM5,
Meso-NH or WRF, including all of the above effects, but the application of these codes is limited when
close field characteristics of geometrical features is of interest. Analyses of contaminant transport in
an urban atmosphere definitly requires CFD based approach because local immission levels depend
very much on fine flow structures such as urban canyon effects, building-specific lofting, and eddies in
the wakes of buildings.
In order to get meaningful output from CFD models some modification to the usual incompressible
models need to be done. To support the practical application of CFD based atmospheric simulation
models it is important to keep in mind that the model should allow large time stepping.
Only very few earlier approaches are known from the literature to the adaptation of different CFD
solvers to meso-scale atmospheric problems. The model implemented by Montavon [1] in CFX solver
is one of the known solutions.
Our aim is to develop a general methodology for including the above effects in CFD solvers and to
implement this scheme in ANSYS-FLUENT for extending the system's capabilities to meso-scale
atmospheric simulations.
In this publication we would like to highlight the potential fields of application for the new method and
to show some validation results.

2. Methodology

Unsteady incompressible with Boussinesq's approximation for the fluid density is used in ANSYS-
FLUENT 6.x solver allowing robust and effective solution. Turbulent transport is modeled by the
Realizable k-ε turbulence model with full buoyancy effects.
Meso-scale effects are treated by a set of user defined functions (UDFs). There are two important
element of our approach:
- a system of transformations effecting the vertical coordinate, vertical component of the velocity
    vector and state variables (pressure, temperature and density);
- user defined volume sources in momentum equations, energy equation and in turbulent transport
    equations for bridging the remaining gap between the governing equations used for solution and
    the correct form of the governing equations including important meso-scale effects.
For mathematical details reader is referred to some of our recent publications [2,3], and a more
detailed explanation on our approach with complete derivation of the source terms and with some
further important validations will be available soon in the scientific literature.

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany

3. Applications
The new approach can be of benefits in several areas of atmospheric simulation where meso-scale to
micro-scale coupled effects can be important such as in:
• Local circulation modeling:
   – Urban heat island convection, ventilation of cities;
   – See breeze;
   – Valley breeze.
• Power generation and pollution control:
   – Assessment of wind power potential, optimization of wind farms;
   – Plumes emitted by cooling towers and chimneys;
   – Dispersion of pollutant in the urban atmosphere.
• Research of meteorological phenomena:
   – Gravity waves;
   – Cloud formation;
   – Flow around high mountain.
• Simulation of disasters:
   – Large scale fires (e.g. in forest fires or town fires);
   – Volcanic plumes.
Some of these applications will be demonstrated in the following sub-chapters.

3.1Urban climatology
Reproducible measured data for atmospheric phenomena is available mostly from laboratory
experiments therefore these measurements can be regarded as good basis of validation. We need to
keep in mind, that the laboratory experiments make approximations as well (eg. compressibility of the
fluid, turbulence and Coriolis force are not accurately treated), therefore full scale validations cannot
be omitted.
High quality experimental data on convection caused by the phenomenon called urban heat island has
been made available by Cenedese, A. and Monti, P. The experiment has been carried out in water
tank with stable thermal stratification enforced by heated and cooled heat exchanger plates attached
to the top and bottom surface of the tank. The heat island convection has been triggered by switching
on a thin circular electric heater, which fixed to the bottom of the tank. PIV measurements have been
used for mapping the velocity field and a set of thermo couple probes has been used for determining
the temperature profile in the symmetry plane.
Since the turbulence was very weak in the experiments large eddy simulation capability of ANSYS-
FLUENT 6 has been employed in the corresponding simulation.
Realistic velocity field (see Fig.1) has been obtained from the 3D LES simulation in the mid-plane.
Height and width of the plume are very close to the experimental observations. Temperature profiles
at different radial positions (Fig.2) as well as profiles of horizontal and vertical velocity components
(Fig.3) are in quantitative agreement with the simulation results.

        Figure 1. Computed flow field in the symmetry plane (time interval of averaging is 300 s).

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany

     2                                                                             2
                                                      x/D - simulation                                                         x/D - experiment

            z/z i
                                                                0                                                                        0
   1.5                                                          0.25             1.5                                                     0,25
     1                                                          0.75               1

   0.5                                                                           0.5
                                                         (T-T a )/dT m                                                            (T-T a )/dT m
     0                                                                             0
         -0.5                 0         0.5              1             1.5          -0.5               0           0.5           1            1.5

  Figure 2. Dimensionless profiles of the temperature perturbation along the axis of the heater plate,
                                 LES results (left) and measured data (right)

     2                                                                            2

                                              x/D                                                                        x/D

                                                    0.5 - experiment                                                           0 - experiment
    1.5                                                                          1.5
                                                    0.5 - LES                                                                  0 - LES

     1                                                                            1

    0.5                                                                          0.5

                                                                    u/U                                                                       w/W
     0                                                                            0
          -0.4         -0.2   0   0.2   0.4    0.6      0.8     1      1.2             -2          0       2   4         6       8       10         12

 Figure 3. Dimensionless profiles of horizontal (left) and vertical (right) velocity components, symbols:
                                    measured data, solid line: LES result
Full scale urban heat island convection has been simulated for the Hungarian city Szeged using on-
site temperature measurements of Unger at al. [5], shown from Fig.4, as boundary condition. Since
only air temperature data was available, instead of specifying surface temperature, a volume heat
source of controlled intensity has been used. Geometry of the bottom surface has been defined on the
basis of natural relief. Characteristic size of the surface grid was 70-80 m within the city and 800-900
m at the rural territory.

    Figure 4. Surface grid with contour plot of elevation (left). Winter time urban heat island intensity
                                    (right): perturbation temperature (in K).

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany

Velocity (Fig.5) and temperature (Fig.6) fields have been analyzed for no wind and for 1.4 m/s wind of
70° angle. The analyses provide valuable information on the distribution of the intensity of country
breeze which plays an important role in the ventilation of the city.

Figure 5. Simulated velocity field for hypothetical case of no geostrophic wind (above) and for 1.4 m/s
                                                 wind (bellow).

  Figure 6. Simulated air temperature above the city for no geostrophic wind (above) and for 1.4 m/s
                                              wind (bellow).

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany

3.2Pollution control
Accurate representation of orography can be very important when assessing the dispersion of
pollutants. We demonstrate this effect on a simulation of a hypotetical plume in the hungarian
mountain Pilis. Natural relief has been downloaded from public database, the geometrical model and
the mesh has been than generated with our purpose developed pre-processor. The advantage of
using ANSYS-FLUENT over special pollution analyses packages, is its potential for detailed
investigation around some important objects by using the unstructured mesh capability of the solver.

       Figure 7. Relief of mountain Pilis used as geometry of the bottom surface in the simulation.

          Figure 8. Shape of a simulated pollution plume for the most frequently occurring wind.

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany

3.3Gravity waves
In atmospheric flow, characterized by stable stratification, internal gravity waves can be induced by
surface features. In steady wind conditions, a steady pattern of gravity waves are usually produced,
which can be used by gliders for ascending to very high altitudes. Physical phenomena responsible for
gravity waves have been taken into account in our modeling approach, therefore gravity wave
measurements can be used for validation of the model.
Experiments in saltwater tank have been carried out by Gyüre B. and Jánosi I.M. [6]. Stable
stratification, with predefined buoyancy frequency N, has been set by setting a gradually decreasing
salt concentration. Perturbations have been generated by towing an obstacle of h height on tens wire
at the bottom of the tank. U denotes the towing speed.

    Figure 9. Contours of computed streamlines and solid lines indicate propagating wave fronts at
                   U/Nh = 1.4 (top) and U/Nh = 0.3 (bottom) non-dimensional flow velocity

         Figure 10. Normalized averaged wave length (left) and amplitude (right) as a function of non-
                                    dimensional horizontal flow velocity

As shown from Fig.9 smooth wave pattern is generated at relatively high towing speed and shorter
waves with possible wave breaking are generated at a lower towing speed. Calculated wave patterns
have been found very similar to those taken from the experiments; wave breaking has been accurately
reproduced. As can be seen in Fig.10, graphically measured wave length and amplitude are in
acceptable agreement with experimental observations.

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany

4. Conclusions

Methodology has been worked out for adding meso-scale atmospheric simulation capabilities to CFD
solvers. The model extensions have been implemented in the form of a UDF package in ANSYS-
FLUENT simulation system. These functions deal with thermal stratification, dry-adiabatic expansion
and compressibility effects as well as with the Corolis force. Inclusion of moisture transport, phase
changes, and surface energy balance are further important step towards the solution of practical
problems of full complexity.
Some validation examples against the results of water tank experiments have been presented. Further
validations against full scale atmospheric observations are subject to ongoing work.
The new method has the potential of extending the applicability of ANSYS-FLUENT system to several
new fields. Unstructured mesh capability of ANSYS CFD products can raise the quality standard of the
simulation of meso-scale to micro-scale coupled atmospheric phenomena in the near future.
Application examples in urban climatology, pollution control and in analyses of gravity waves have
been presented in this publication.

Acknowledgements This work has been supported by the Hungarian Research Fund under
contract number OTKA T049573 and National Research and Development Program under contract
number NKFP 3A/088/2004.

5. References

[1]     Montavon C.: Simulation of atmospheric flow over complex terrain for wind power potential
        assessment, Doctoral thesis, EPFL, Lausanne, 1998
[2]     N.Rácz, G.Kristóf, T.Weidinger, M.Balogh: Simulation of gravity waves and model validation to
        laboratory experiments, CD, Urban Air Quality Conf. Cyprus, 2007
[3]     G.Kristóf, N.Rácz, M.Balogh: Adaptation of pressure based CFD solvers to urban heat island
        convection problems, CD, Urban Air Quality Conf. Cyprus, 2007
[4]     Cenedese, A. and Monti, P.: Interaction between an Urban Heat Island and a Sea-Breeze Flow.
        A Laboratory Study, J. Appl. Meteorol. 42, 1569-1583., 2003
[5]     Unger J.: Intra-urban relationship between surface geometry and urban heat island: review and
        new approach. Climate Research 27, 253-264, 2004
[6]     Gyüre, B. and Jánosi, I.M.: Stratified flow over asymmetric and double bell-shaped obstacles.
        Dynamics of Atmospheres and Oceans 37, 155-170, 2003

ANSYS Conference &
25 CADFEM Users’ Meeting 2007

November 21-23, 2007 Congress Center Dresden, Germany


To top