Microsoft Word - 26jonassen.docx
CCHHEEMMIICCAALL EENNGGIINNEEEERRIINNGG TTRRAANNSSAACCTTIIOONNSS
VOL. 40, 2014
A publication of
The Italian Association
of Chemical Engineering
www.aidic.it/cet
Guest Editor: Renato Del Rosso
Copyright © 2014, AIDIC Servizi S.r.l.,
ISBN 978-88-95608-31-0; ISSN 2283-9216
Implementation of an Integrated Stationary – Non Steady
State Gaussian Modeling System to Simulate Contaminants
Dispersion into the Atmosphere
Massimo Rabonia, Sabrina Copelli*b, Mattia Tigolib, Vincenzo Torrettac,
Giordano Urbinic, Giuseppe Nanoa
aDipartimento di Chimica, Materiali e Ingegneria Chimica “G. Natta” - Politecnico di Milano- via Mancinelli, 7 – 20133 –
Milano – Italy
bDipartimento di Scienza e Alta Tecnologia - Università degli Studi dell’Insubria - via Vico, 46 – 21100 – Varese – Italy
cDipartimento di Biotecnologie e Scienze della Vita - Università degli Studi dell’Insubria - via Dunant, 3 – 21100 –
Varese – Italy
sabrina.copelli@uninsubria.it
Contaminants and odor emissions are some of the major environmental problems that several industrial
sites have to face. Particularly, complaints exhibited by the population living near plants characterized by
both unpleasant odor emissions and hazardous contaminants dispersion (e.g. waste treatment and
disposal plants, wastewater treatment plants, chemical industries, food industries, livestock activities,
rendering plants, tanneries, etc.) are becoming more and more frequent, leading to the establishment of
either legal suits (in the case of existing plants) or constraints with respect to the construction of new
plants. Taking note of the importance of this feature, it is essential to provide industrial sites with easy-to-
use software capable of evaluating the effects on the neighborhood of their atmospheric issuances.
Therefore, the aim of this work has been the implementation of an integrated stationary – non steady state
Gaussian modeling system (based on AERMOD and CALPUFF algorithms) which is able to simulate the
dispersion of both contaminants and odors into an open field choosing automatically the most correct
algorithm to be used. Its robustness and reliability has been tested using both AERMOD and CALPUFF
simulations. The good agreement among the obtained results makes this software particularly suitable for
an easy evaluation of the effects of a generic issuance onto the nearby environment without requiring a
deep knowledge of the dispersion modeling system by the users.
1. Introduction
It is well known that, in the last twenty years, the impact of air pollution on human health and environment
has received an increasing attention from both public organizations and industries (Copelli et al., 2012;
Rada et al., 2014). Particularly, more and more severe regulations concerning the releases of hazardous
air pollutants (HAPs), volatile organic compounds (VOCs) and dusts of very different sizes into the
atmosphere have arisen, leading to a pressing need for complete and reliable simulations of the dispersion
dynamics. Such simulations are often tricky to be carried out by inexperienced users, even when
dedicated software with user-friendly interfaces (e.g., commercial versions of AERMOD and CALPUFF)
are employed, because of a series of features to be taken into account: the presence of a complex domain
on which the diffusion equations need to be integrated (the topography of the area interested by the
dispersion), the modeling of both dry and wet deposition, the effects of gravitational settling, temperature
and wind velocity.
In particular, simulating the behavior of a generic pollutant released to the atmosphere means to determine
its concentration field at any point in space and at any time after the issuance. There are essentially two
DOI: 10.3303/CET1440049
Please cite this article as: Raboni M., Copelli S., Tigoli M., Torretta V., Urbini G., Nano G., 2014, Implementation of an integrated
stationary – non steady state gaussian modeling system to simulate contaminants dispersion into the atmosphere, Chemical Engineering
Transactions, 40, 289-294 DOI: 10.3303/CET1440049
289
ways to simulate the dispersion of a pollutant into the atmosphere: assuming an Eulerian point of view (in
which fixed coordinates in space are employed) or a Lagrangian point of view (in which the coordinates
move with the particles). Nevertheless, all dispersion models use mathematical equations, describing the
atmosphere, dispersion and chemical-physical processes occurring within a plume (continuous emission)
or a puff (discontinuous emission), to calculate contaminants concentrations at various locations.
In this regard, lots of dispersion modeling systems have been proposed. Looking at the methodology on
which all these models are based, it is possible to classify them into four categories (Busini et al., 2012;
Capelli et al., 2013; Holmes and Morawska, 2006; Torti et al., 2013): 1) box models; 2) Gaussian models;
3) Lagrangian models; 4) Computational Fluid Dynamic models.
Focusing on Gaussian models, they are widely used in atmospheric dispersion modeling, in particular for
regulatory purposes, and are often ‘‘nested’’ within Lagrangian and Eulerian models. Such methods are
based on a Gaussian distribution of the plume, in both vertical and horizontal directions, under steady and
unsteady state conditions. Considering a plume (continuous emission), its normal distribution is modified at
huge distances from the source due to the effects of turbulent reflection from the surface of the earth and
at the boundary layer (when the mixing height is low). The width of the plume is determined by σy and σz,
which are defined either by stability classes (Pasquill, 1961) or travel time from the source.
It should be noted that, among all these different methods currently available, the United States
Environmental Protection Agency (US-EPA) recommends CALPUFF model for the evaluation of long-
range transport of pollutants (generally between 50 and 200 km), while in the near field (i.e., within 50 km
from the source), US-EPA identifies as a reference the application of the stationary Gaussian model
AERMOD (US-EPA, 2005).
Particularly, CALPUFF (Shire et al., 2000) is a multi-layer non-steady state puff dispersion model designed
to model the dispersion of gases and particles using space and time varying meteorology basing on
similarity equations, turbulence, emission strengths, transformation and removal. It is able to model point,
line, volume and area sources using an integrated puff formulation incorporating the effects of plume rise,
partial penetration, stack and building effects. Nevertheless, due to some intrinsic limitations, CALPUFF is
not recommended either for performing calculation on timescales shorter than 1 h or for modeling
dispersions that are heavily influenced by turbulence, such as in an urban environment (US-EPA, 2005;
Brode and Anderson, 2008). AERMOD (US-EPA, 2009) is a near field steady state Gaussian plume model
based on planetary boundary layer turbulence structure and scaling concepts, including treatment of both
surface and elevated sources over both simple and complex terrain. In the stable boundary layer the
distribution is assumed to be Gaussian in both the horizontal and vertical directions. AERMOD is able to
model buoyant plumes and incorporates a treatment of lofting, whereby the plume remains near the top of
the boundary layer before mixing with the convective boundary layer (CBL).
Looking at practical features, the use of CALPUFF in the near field is better if there are weather and/or
geographic conditions such as to make inappropriate a simulation with AERMOD (Brode and Anderson,
2008; Dresser and Huizer, 2011). These conditions, which can be defined as “non-stationary”, are very
common if the wind regime involves high frequencies of wind calm and the presence of both land-sea
discontinuities (originated by the presence of coastlines) and complex terrain morphologies that require an
accurate reconstruction of the wind field.
In this work, an integrated stationary – non steady state Gaussian modeling system based on AERMOD
and CALPUFF algorithms has been implemented in order to: 1) automatically switch between AERMOD
and CALPUFF algorithms basing on the pros and cons evidenced above; 2) support inexperienced users
in both pre and post processing.
This software has been then used to simulate the near field (4 km) dispersion into the atmosphere of a
point issuance containing a mixture of different odorous compounds coming from the cleaning air system
of a hypothetical wastewater treatment plant (pumping station and primary settling). The topography taken
into account is particularly complex because it involves the presence of both quite high hills and sea.
Moreover, wind circulation is characterized by strong oscillations during the day (both in terms of
magnitude and direction) thus implying a continuous change in the stability class. In order to perform such
a simulation, the CALPUFF algorithm has been selected and, then, obtained results have been compared
with those ones coming from a real CALPUFF simulation. The good agreement which has been achieved
confirms the reliability of the developed modeling system that can therefore be used for an easy but
reliable evaluation of the dynamics of a generic dispersion process.
2. Overview of the software
DISPSIM is an integrated stationary – non steady state Gaussian modeling system developed in order to
provide an easy evaluation of the effects of a generic issuance of contaminants and odors into an open
290
field. As the Gaussian dispersion models on which DISPSIM is based, that is AERMOD and CALPUFF, it
is able to treat both complex morphologies and terrain conditions (both urban and open-country), dry and
wet deposition, different source types (point, line, volume and area sources), non-steady state emissions
and meteorological conditions, vertical wind share, plume rise, building downwash and, finally, overwater
and coastal interaction effects.
DISPSIM is particularly user-friendly because it automatically selects the most appropriate dispersion
algorithm choosing in between: a near field steady state Gaussian plume model (that is, an “AERMOD-
like” model) and a multi-layer, multi-species non steady state puff dispersion model (that is, a “CALPUFF-
like” model). This aspect is particularly useful because it makes the software very versatile and easy to use
even for inexperienced users who do not posses a deep knowledge of the fluid dynamics influencing the
dispersion of contaminants and odors into the atmosphere (and, therefore, they would not be able to
choose in between different dispersion models implemented into a software). The criteria implemented for
the automatic switch are: a) source emission type (continuous or not); b) presence of wind calm (wind
velocity under 0.2 m s-1); c) presence of complex terrain; d) environment (urban, coastal, …); e) timescale
simulation; f) field extension.
As AERMOD and CALPUFF, DISPSIM is constituted by three different calculation modules: a pre-
processing module, a dispersion module and a post-processing module. The first one, called METSIM, is a
meteorological and geophysical model that develops a hourly wind and temperature three dimensional
field to be used to calculate stability classes employed by the simulation module. It is derived from an
integration of AERMET (the pre-processing module of AERMOD used to compute certain boundary layer
parameters used to estimate profiles of wind, turbulence and temperature) and CALMET (the diagnostic 3-
dimensional meteorological model that pre-processes data for CALPUFF; US-EPA, 2014).METSIM is also
able to retrieve digital terrain elevation data (DTED) from NASA database (NASA, 2000) with a resolution
of about 50x50 m (Beauducel, 2013). Moreover, it is possible to import ArcGIS shapefiles or GeoTIFF
(georeferenced raster maps) representing Corine land cover and area description, respectively, with the
aim of defining urban and open-country terrain conditions (Mathworks, 2010). Sources and specific
receptors location is loaded from an ASCII file, a Google Earth kml file (Google, 2014) or through a
graphical user interface (GUI) which allows picking points on the previously loaded maps.
DISPSIM is the core of the software and it permits to perform the effective calculation of the 3-D
concentration field by implementing both AERMOD and CALPUFF algorithms (US-EPA, 2014).
DISPpost builds different maps showing average contaminants (or odors) concentrations, peak
concentrations, the n-th percentile concentrations (computed using a low-demand memory consuming
algorithm), and threshold limit value overcoming frequencies (and so on) during a pre-determined period of
time (generally, 1-year). Moreover, DISPpost allows to evaluate the time distribution of the overcoming
frequencies at the receptors during the day or the seasons in order to evaluate the critical period for the
nuisance (e.g. nighttime for inhabitants, during the day for schools, summertime for vacationers, etc..).
The major difference with respect to AERMOD and CALPUFF software is that all these modules are not
separated but fully integrated each other.
The parallel computing and the highly structured matrix algorithm capable of processing the different
concentration layers is more effective in the AERMOD-like simulations, in which each time step is
independent from the others. This makes the simulation very short time consuming.
3. Case-study
DISPSIM has been used for assessing the atmospheric issuance of odors coming from a hypothetical air
cleaning plant located in the south of Italy. The site is located near the coast (east) at sea level (Figure 1a).
In the neighborhood, there are: three towns (E, S-SE and S-W), an industrial complex (N), cultivated lands
(Figure 1b) and a beach (highly frequented by the population) (S-E). One of the two towns (S-W) is located
on a hill. The characteristics of the issuance are summarized in Table 1.
The area here considered for the simulation is constituted by a square of 4.0 km of side (near field),
centered on the location of the odors issuance. For the post-processing purposes, the area thus defined
has been unbundled into a Cartesian orthogonal grid consisting of a square mesh of 50 m of side for a
total of 6,561 points per horizontal layer. Moreover, 8 specific receptors, that describe the position of 5
residential buildings, a school and two aggregation points (beach), have been located around the plant
(see Figure 1).
1-year of hourly-averaged meteorological data (lacking data: 0.3%) have been adopted for evaluate the
odor nuisance of the plant.
291
a) b)
Figure 1: Description of the area surrounding the plant. Aerial view showing the source and eight
receptors locations (a); land use overlapped on the technical map (b).
Table 1: Characteristics of the hypothetical issuance treated in the case study.
Description Unit of measurement Value
Source type - Point
Type of efflux - Vertical without deflection
- Continuous
Volumetric flow rate Nm³ h-1 20,000
Outlet temperature °C 30
Source height (above the ground) m 10
Source diameter m 0.75
Odor background concentration OU Nm-3 0
Odor concentration OU Nm-3 4,000
Peak-to-mean value for odor - 2.3
4. Results and discussion
Figure 2a shows the mean odor concentration registered in the area. The eastern zones are the most
exposed because of both its closeness to the source and meteorological conditions (e.g. receptor R2
mean concentration: 0.16 OU Nm-3); on the contrary the hilly urban area (receptors R6, R7 and, especially,
R8) and the southern village (receptor R5) avoid the odor nuisance thanks to their overhead position (only
receptor R6 and R7 are at 0.05 OU Nm-3) and the distance from the source. Therefore it follows that two of
the three towns do not have particular issues.
a) b)
Figure 2: DISPSIM results: mean (a) and maximum (b) odor concentrations (OU Nm-3) in the studied
area.
292
Considering peak concentrations (Figure 2b), the most exposed sites (up to 50 OU Nm-3) are the urban
area near receptor R2 and the western cultivated area located within 1,000 m from the source.
The 98th percentile concentration map (Figure 3a), considered also by some Italian local guidelines
(Regione Lombardia, 2014) for odor nuisance assessment, confirms critical conditions for the nearby
receptors placed in the eastern part of the area (R2, in particular, with 2.3 OU Nm-3). South-eastern
shoreline (the most interested by bathing and tourism; receptors R3 and R4) is affected by 98th percentile
concentrations higher than 1.5 OU Nm-3. Orography confirm to positively influence the odor conditions, as
demonstrated by the receptors R6, R7 and R8 placed on the hill located in the S-W with respect to the
source. The school (receptor R5) has a 98th percentile odor concentration below 1 OU Nm-3.
a) b)
Figure 3: DISPSIM results: 98th percentile (a) concentrations (OU Nm-3) and frequency (as percentage
of the total occurrences) of the overcoming 1 OU Nm-3 detection threshold value (b).
Figure 3b shows the map of the 1 OU Nm-3 overcoming frequency in the studied area during the 1-year
simulation. It can be noticed that some of the south-eastern urban areas are afflicted for more than 175 h
y-1 by odor concentrations greater than 1 OU Nm-3. Considering that 1 OU Nm-3 is the odor detection
threshold statistically perceived by half of the population (DEFRA, 2010), the issue can represent a trouble
for both the inhabitants and the tourism development in the area.
In order to evaluate such situation, DISPpost feature of “time distribution” analysis was applied on
receptors R3 (bathing establishment; Figure 4a) and R5 (school; Figure 4b). Such receptors are located in
typical “discontinuous use” areas. In fact, the bathing establishment is used only in summertime (typically
between May and September) while the school is used in the morning-early afternoon (typically, between 8
am and 4 pm).
a) b)
Figure 4: DISPpost “time distribution” analysis tool: 98th percentile concentrations (OU Nm-3) of: a)
monthly values (excluding night-time simulations) at receptor R3 (bathing establishment) and b) daily
values at receptor R5 (school). The horizontal solid lines show the odor detection threshold.
The monthly 98th percentile concentrations at receptor R3 shows a periodic behavior of odor concentration
which affects negatively the bathing establishment during its activity (odor concentration always above 1.5
OU Nm-3 in summertime, with a peak of 3.8 OU Nm-3 in July): the odor source could be a problem for
tourism development if not adequately treated. For what concern daily distribution at receptor R5,
maximum concentrations are most frequent in the morning (up to 7.4 OU Nm-3 at 5 am) and in the late
afternoon (6.6 OU Nm-3 at 5 pm) as a consequence of the breezes, with negative effects for the students.
293
5. Conclusions
In this work, a nimble - integrated stationary–non steady state Gaussian model to simulate contaminants
dispersion into the atmosphere has been proposed (DISPSIM) and used to simulate the effects of an
odors issuance. The easiness of treating data, such as digital elevation data as well as receptors, source
locations and data time-analysis, combined to the integration of pre-post processing with the dispersion
model helps to reduce computation time (a complete 1-year simulation on a 4.0 x 4.0 km area takes only 5
minutes). Such features elect DISPSIM as a tool for quick air pollution and odor assessment or,
alternatively, a tool for air treatment plant design.
Acknowledgements
The authors wish to acknowledge AirClean Srl (Rho, Milan, Italy) for their technical support.
References
Beauducel F., 2013, ReadHGT script, <http://dds.cr.usgs.gov/srtm/version2_1> accessed: 20.05.2014.
Brode, R.W., Anderson, B., 2008, Technical issues related to CALPUFF near-field applications –
Attachment A, 26th September 2008, US-EPA.
Busini V., Capelli L., Sironi S., Nano G., Rossi A.N., Bonati S., 2012, Comparison of CALPUFF and
AERMOD Models for Odour Dispersion Simulation, Chemical Engineering Transactions, 30, 205-210,
DOI: 10.3303/CET1230035.
Capelli L., Sironi S., Del Rosso R., Guillot J.-M., 2013, Measuring odours in the environment vs. dispersion
modelling: A review, Atmospheric Environment, 79, 731-743, DOI: 10.1016/j.atmosenv.2013.07.029.
Copelli S., Torretta V., Raboni M., Viotti P., Luciano A., Mancini G., Nano G., 2012, Improving
Biotreatment Efficiency of Hot Waste Air Streams: Experimental Upgrade of a Full Plant, Chemical
Engineering Transactions, 30, 49-54, DOI: 10.3303/CET1230009
Google Inc., 2014, Keyhole Markup Language – KML Documentation,
<https://developers.google.com/kml/> accessed 20.05.2014.
Holmes N. S., Morawska L., 2006, A review of dispersion modelling and its application to the dispersion of
particles: An overview of different dispersion models available, Atmospheric Environment, 40, 5902-
5928, DOI: 10.1016/j.atmosenv.2006.06.003.
Department for Environment Food and Rural Affairs (DEFRA), 2010, Odour Guidance for Local
Authorities, DEFRA, London, UK.
Dresser, A.L., Huizer, R.D., 2011, CALPUFF and AERMOD model validation study in the near field:
martins creek revisited, J. Air & Waste Manage. Assoc., 61, 647–659, DOI:10.3155/1047-
3289.61.6.647.
Mathworks Inc., 2014, MatLAB Mapping Toolbox manual.
National Space Agency (NASA), 2000, Shuttle Radar Topography mission,
<http://www2.jpl.nasa.gov/srtm/> accessed 12.02.2014.
Pasquill F., 1961, The estimation of the dispersion of windborne material, Meteorology Magazine, 90, 33-
49.
Rada E.C., Raboni M., Torretta V., Copelli S., Ragazzi M., Caruson P., Istrate I.A., 2014, Removal of
benzene from oil refinery wastewater treatment plant exhausted gases with a multi-stage biofiltration
plant, Revista de Chimie, 65(1), 68-70.
Regione Lombardia, 2014, Guidelines concerning the characterization and permission of gas emissions
into the atmosphere from high-impact activities (In Italian),
<http://www.reti.regione.lombardia.it/cs/Satellite?c=Redazionale_P&childpagename=DG_Reti%2FDeta
il&cid=1213355430857&pagename=DG_RSSWrapper> accessed 20.03.2014.
Shire J.S., Strimaitis D.G., Yamartino R.J., 2000, A user’s guide for the CALPUFF dispersion model
(version 5), Earth Tech Inc., MA, USA.
Torti E., Sibilla S., Raboni M., 2013, An Eulerian-Lagrangian method for the simulation of the oxygen
concentration dissolved by a two-phase turbulent jet system, Computer and Structures, 129, 207-217,
DOI: 10.1016/j.compstruc.2013.05.007
United States - Environmental Protection Agency (US-EPA), 2005. Guideline on Air Quality Models, 40
CFR Part 51, Appendix W. Published in the Federal Register, Vol. 70, No. 216, November 9, 2005.
United States – Environmental Protection Agency (US-EPA), 2009, AERMOD implementation guide (rev
March 19, 2009), US-EPA, NC, USA.
United States – Environmental Protection Agency (US-EPA), 2014, Preferred/Recommended Models,
<http://www.epa.gov/ttn/scram/dispersion_prefrec.htm> accessed 20.03.2014.
294
Controfacciata.pdf
Pagina vuota
Pagina vuota
Pagina vuota