For the Localization of a MSW Incinerator Results of Atmospheric

Results of atmospheric dispersion model for the localization of a MSW incinerator

E. Brizio1, G. Genon2 & M. Poggio2 1Cuneo Province Administration, Italy 2Turin Polytechnic, Italy

Abstract

The present study reports the attempt to localize a MSW incinerator, starting from three proposed sites, on the basis of the atmospheric transport and dispersion modelling. The considered area, around Turin, in North-Western Italy, is characterized by a high percentage of hours of calm wind, so the choice of the atmospheric model is very important in order to have reliable and convincing results. In the present study we used the AERMOD model by the U.S. EPA, that is a hybrid atmospheric dispersion model which can treat the calm hours within a more restrained computational time with respect to Eulerian or Lagrangian models. The investigated issue is generally a very delicate and thorny argument because of the natural opposition of the people against this kind of plant. The first thing to do to treat the localization of a MSW incinerator is to calculate the concentrations at the ground level due to the emitted pollutants and to compare them to the air quality regulations; secondly, one can think about the possibility of building a district heating system or distributing technological vapour service to industrial users, avoiding the respective emission of pollutants. In practice, the concentrations at the ground level when the incinerator is placed in the three proposed sites are very comparable and, for each site, much smaller than the air quality regulations and the measured levels of the area. The best localization is then determined by the possibility of realising a large district heating system. Keywords: incinerator, AERMOD, concentration, district heating, calm hours, MSW, avoided emissions. 1 Introduction

In the present paper we simulated the atmospheric transport and dispersion of pollutants generated by the incineration of the solid waste deriving from the

Development and Application of Computer Techniques to Environmental Studies X G. Latini, G. Passerini & C. A. Brebbia (Eds) © 2004 WIT Press, www.witpress.com, ISBN 1-85312-718-3 148 Development and Application of Computer Techniques to Environmental Studies X mechanical selection of the MSW produced by the metropolitan area of Turin, in North-Western Italy. There are three possible localizations for the plant: Volpiano, 20 km far from Turin, Strada del Francese and Gerbido, respectively in the North-Western and South-Western part of Turin; each of these three sites has different possibilities of realising or increasing the district heating systems.

2 Data at disposal

2.1 Technical features and possible localization

As concerns the plant, we know that the MSW incinerator would be fed by 245.000 t/y of selected MSW (approximately 34 t/h, composed by 38,6% C, 5,1% H, 22,4% O, 1,0% N, 0,1% S, 0,8% Cl, 15,0% ashes and 17,1% H2O, enthalpy of combustion = 3.595 kcal/kg). The thermal power of the plant is 142,2 MW. The emitted flow is approximately 258.138 Nm3/h, with a humidity volume of 15% and a oxigen content near 9,4% (vol/vol for dry smoke). The flue gas temperature has been assumed 140 °C, the stack height is 80 m and the stack diameter is 3 m. The emission limits for the incineration can be referred to the EC regulations, in particular 76/2000/CE; on the other hand, the emission levels, taken from measurements on existing plants, can be often much smaller than the limits, as reported in Table 1.

Table 1: Emission limits and levels for a MSW incinerator.

EMISSION LIMITS EMISSION LEVELS YEARLY MASS POLLUTANT 3 3 (mg/Nm , dry gas, O2 @ 11%) (mg/Nm , dry gas, O2 @ 11%) FLUX (t/y) TSP 10 1 2,2

NO2 200 60 134,4

SO2 50 10 22,4 CO 50 25 55,9 Σ Heavy metals 0,5 0,1 0,2 PAH 0,01 0,002 4 E-3 3 PCDD/DF (ng/m ) 0,1 0,05 1 E-7 Cd+Tl 0,05 0,02 4 E-2 T.O.C. 10 10 22,4 HCl 10 10 22,4 HF 1 1 2,2 Hg 0,05 0,05 0,1 Zn 5 5 11,2

Table 2 reports the different features relating to the electric energy generated by the incinerator, the distributed thermal energy and the avoided emissions from the displaced heating plants for the three sites. It is important to observe that the assumed scenarios would overlap the existing district heating systems which, at present, are supplied by cogeneration plants: as a consequence, the net displaced heating plants are fewer than those corresponding to the total distributed thermal energy whereas the electric energy generated by the cogeneration plants can be

Development and Application of Computer Techniques to Environmental Studies X G. Latini, G. Passerini & C. A. Brebbia (Eds) © 2004 WIT Press, www.witpress.com, ISBN 1-85312-718-3 Development and Application of Computer Techniques to Environmental Studies X 149 larger. We assume that the existing heating plants burn natural gas (77%), gas oil (20,6%) and oil (2,4%). The avoided emissions should be modelled over 6 months per year, 14 h/d; we supposed that the eliminated sources have a flue gas temperature of 180°C, a stack exit gas velocity of 2,5 m/s, a stack height of 20 m and a thermal power of approximately 2 MW each. On the other hand, the incinerator has been assumed to emit continuously over the entire year.

Table 2: Three possible sites for the incinerator.

STRADA DEL VOLPIANO GERBIDO FRANCESE Electric power (GWh/y) 232,6 213,5 205,7 Thermal energy (GWh/y) 103,5 364,7 475,5 Net displaced heating plants (GWh/y) 6,8 127 145,5

NO2: 1,5 NO2: 27,6 NO2: 31,7 SO : 0,7 SO : 10,4 SO : 10,4 Avoided emission (t/y) referable to the 2 2 2 TSP: 0,1 TSP: 1,2 TSP: 1,2 increased district heating system CO: 0,9 CO: 17,8 CO: 20,4 CO2: 1.400 CO2: 25.900 CO2: 29.600 Increment of electric energy generated by existing cogeneration plant 18,8 49,2 75,8 (GWh/y)

2.2 Meteorological data

We have at disposal a meteorological database for 1 year, from June 1st, 1999 to May 31st, 2000, calculated for the entire Province of Turin (Finardi [1]) by means of the diagnostic model MINERVE, which can define the 3-dimensional wind field, and the meteorological pre-processor SURFPRO, which assesses the atmospheric turbulence parameters (the sensible heat flux, the friction velocity, the convective velocity scale, the Monin-Obukhov lenght, the mixing height and the stability classes) on the basis of the land cover and the solar radiation. The models employ surface wind observations from the regional meteorological network as well as from Milano Linate airport radio-soundings and ECMWF (European Centre for medium range Weather Forecast) data. In particular, MINERVE is a mass-consistent model which interpolates the available observations considering the orography of the domain and the surface roughness, giving back, as results, the wind and temperature hourly vertical profiles with an horizontal resolution of 1 km. Figure 1 points out the wind roses for the three sites (Volpiano, Strada del Francese and Gerbido) and for the centre of Turin.

3 The AERMOD model

In order to calculate the concentrations of the pollutants at the ground level referable to the incinerator or to the emissions of the displaced heating plants, we applied the AERMOD model by U.S. EPA [2], which was born to replace the Industrial Source Complex Model (3rd version), up to now used for the environmental planning, the risk assessment and the emission permits in the U.S.A.. The AERMOD model has been designed for the short-range simulations

(distances from sources generally smaller than 50 km), both for simple and complex orography. AERMOD is a steady state plume model; when the boundary layer is stable (SBL), it assumes the concentration distribution to be Gaussian in both the vertical and the horizontal. On the other hand, in the convective boundary layer (CBL) the horizontal distribution is also assumed to be Gaussian but the vertical distribution is described with a bi-Gaussian probability density function (pdf). Moreover, in the CBL, AERMOD treats “plume lofting”, whereby a portion of plume mass, released from a buoyant source, rises to and remains near the top of the boundary layer before becoming mixed into the CBL.

Figure 1: Wind roses: clock-wise, Volpiano (80 m), Strada del Francese (80 m), Gerbido (80 m) and the centre of Turin (30 m).

AERMOD also tracks any plume mass that penetrates into the elevated stable layer and then allows it to re-enter the boundary layer when and if appropriate. For sources in both the CBL and the SBL AERMOD treats the enhancement of lateral dispersion resulting from plume meander. As concerns the treatment of the orography, AERMOD is strongly different from the previous regulatory models (e.g. ISC3) that distinguish among simple, intermediate and complex terrain. Actually, AERMOD has been designed to be more physically realistic: where appropriate, the plume is modelled as either impacting and/or following the terrain by applying the dividing streamline concept; the streamline height depends on the parameter “height scale” which represents the terrain height that dominates the flow in the vicinity of the receptor. One of the major improvements of AERMOD is the characterization of the planetary boundary layer (PBL) through both surface and mixed layer scaling. AERMOD estimates vertical profiles of wind speed, wind direction, vertical and lateral turbulence, temperature and vertical potential temperature gradient based on measurements and extrapolations of those measurements using similarity relationships. Anyway, AERMOD is designed to run with a minimum of observed meteorological parameters, such as a single surface measurement of wind speed, wind direction and ambient temperature. Unlike existing regulatory models, AERMOD accounts for the vertical inhomogeneity of the PBL in its dispersion calculations by averaging the parameters of the actual boundary layer over the portion of the layer that contains plume material between the plume centroid height and the receptor height. Moreover the AERMOD model doesn’t use the Pasquill-Gifford stability classes in order to estimate the dispersion coefficients: on the contrary, these parameters depends directly on the calculated vertical and lateral turbulence. In the end, AERMOD contains improved algorithms for the plume rise, particularly for plumes affected by building wakes (PRIME model), for buoyancy and the urban night time boundary layer. On the basis of the adopted model formulations, unlike existing regulatory models, AERMOD can handle wind speeds smaller than 1 m/s which are usually considered as calm winds; generally, Gaussian model should not be run with low wind speeds as predicted concentrations became unrealistically large in these cases. On the contrary, there is nothing in AERMOD that prevents the use of zero wind speeds; as a matter of fact, while AERMOD is fundamentally a steady-state Gaussian plume model, it contains improved algorithms for dealing with low wind speed conditions. Consequently, AERMOD can produce model estimates for conditions when the wind speed is less than 1 m/s, but still greater than the instrument threshold. The threshold should be the higher of the detection limits for the wind speed and the wind direction measurement instrument. If the wind speed for a given hour is less than the user defined threshold level input to AERMET, that hour is considered calm and no concentration is calculated by AERMOD. The described feature is of fundamental importance when performing long term analysis (seasonal or annual mean concentrations) in areas with high percentage of low wind speed, as in Northern Italy.

4 The atmospheric modelling

The model AERMOD has been applied to domains measuring 40 km * 40 km round the plant, with a spatial resolution of 200 m; in the present study, we adopted the regulatory default options of the model, including the use of stack- tip downwash, the effects of elevated terrain and the calm and missing data processing routines. The urban modelling option, which incorporates the effects of increased mixing height in urban areas under stable atmospheric conditions, has not been applied since the considered domains comprehend the city of Turin (more than 1 million inhabitants) but they are much larger than its urban area. The computational time for the entire domain (40401 points), 8784 hours for each point (actually, we have at disposal the meteorological data for a leap year), is approximately 5 hours for a PC Pentium III, 1.1 GHz, 216 MB RAM. The following Figure 2, 3, 4, 5 and 6 show the isoconcentration maps for the assumed scenarios, determined on the basis of the yearly mean concentration at the ground level. Figure 2, 3 and 4 deal with the incinerator impact in relation to the three proposed localizations whereas Figure 5, 6 and 7 show the avoided concentrations referable to the displaced heating in Settimo Torinese (connected with the site of Volpiano), in Northern Turin (depending on Strada del Francese) and in the centre of Turin (in relation to Gerbido). As concerns the impacts due to the incinerator, we have the largest yearly mean concentrations where the orography is more complex, in particular on the hills which are placed in the Eastern part of the analysed area.

Figure 2: Isoconcentration map for the incinerator placed in Volpiano.

Figure 3: Isoconcentration map for the incinerator placed in Strada del Francese.

Figure 4: Isoconcentration map for the incinerator placed in Gerbido.

Figure 5: Isoconcentration map for the displaced heating in Settimo Torinese.

Figure 6: Isoconcentration map for the displaced heating in Northern Turin.

Figure 7: Isoconcentration map for the displaced heating in the centre of Turin.

As a matter of fact, not all the plume has sufficient kinetic energy to rise to the height of the terrain, and consequently, part of the plume material doesn’t follow the terrain but impacts it. Table 3 reports the yearly mean concentration averaged over the entire domain and the maximum punctual yearly mean concentration due to the incinerator, in relation to the most worrying urban pollutants, i.e. the total suspended powder (TSP) and the nitrogen oxides (NOX); the maximum calculated hourly mean concentrations for the three proposed sites are 0,2 µg/m3 3 for TSP and 12 µg/m for NO2.

Table 3: Calculated yearly mean concentrations at the ground level for the incinerator.

STRADA DEL VOLPIANO GERBIDO FRANCESE Maximum punctual yearly TSP: 0,0081 µg/m3 TSP: 0,0068 µg/m3 TSP: 0,0048 µg/m3 3 3 3 mean concentration NO2: 0,486 µg/m NO2: 0,408 µg/m NO2: 0,288 µg/m Yearly mean concentration TSP: 0,00047 µg/m3 TSP: 0,00038 µg/m3 TSP: 0,00040 µg/m3 3 3 3 averaged over the domain NO2: 0,0282 µg/m NO2: 0,0228 µg/m NO2: 0,0240 µg/m

On the other hand, Table 4 points out the avoided concentrations from the displaced heating plants. In order to compare the avoided emissions with the added ones, the right parameters are the yearly mean concentrations averaged over the domain because the maximum punctual concentrations are usually placed in different points. As one can easily observe, in the case of Northern

Turin and the centre of Turin, the displaced heating produces much larger concentrations at the ground level than the corresponding incinerator (even more than 10 times in the case of TSP), although the total annual emissions are smaller (a half for TSP, CO and SO2, a quarter for NOX); moreover, the maximum concentrations are placed round the sources. This effect can be explained by the physical and technical features of the corresponding sources, which are generally not far from the ground level and with low stack exit gas velocity.

Table 4: Calculated yearly mean concentrations at the ground level for the displaced heating plants.

SETTIMO NORTHERN CENTRE OF

TORINESE TURIN TURIN Maximum punctual yearly TSP: 0,005 µg/m3 TSP: 0,058 µg/m3 TSP: 0,060 µg/m3 3 3 3 mean concentration NO2: 0,081 µg/m NO2: 1,325 µg/m NO2: 1,575 µg/m Yearly mean concentration TSP: 0,00012 µg/m3 TSP: 0,00129 µg/m3 TSP: 0,00171 µg/m3 3 3 3 averaged over the domain NO2: 0,002 µg/m NO2: 0,030 µg/m NO2: 0,045 µg/m

5 Conclusions As for the incinerator emissions, the yearly mean concentrations referable to the three proposed localizations are strongly comparable (the largest difference is 20% between Volpiano, which causes the highest concentrations, and Strada del Francese). Anyway, for each site, the calculated mean concentrations, both yearly and hourly, are much smaller than the air quality limits suggested by the 3 EC regulations (e.g. yearly mean concentration of 40 µg/m for PM10 and NO2). As for the pollutant levels measured in the urban area of Turin, the maximum increments of the yearly mean concentration would be 0,01% for PM10 and 0,6% for NO2; in the same way, with regard to the other pollutants, the highest increment due to the assumed incinerator are very small if compared to existing backgrounds in urban or industrial environments (Rabl et al. [3]). Moreover, at least for PM10, CO, NO2 and SO2, the calculated increments are strongly counterbalanced or exceeded by the avoided concentrations due to the district heating systems. On the basis of the reported data, the worst localization for the incinerator is Volpiano which causes the highest concentrations with no possibilities of large district heating systems; the other localizations are substantially equivalent with a small preference for the site of Gerbido which allows the largest avoided emissions from the displaced heating plants in the centre of Turin. References [1] Finardi, S., Costruzione di una base dati meteorologici sulla Provincia di Torino con l’ausilio di modellistica numerica, ARIANET, 2003. [2] U.S. EPA, Office of Air Quality Planning and Standards, Revised Draft user’s guide for the AMS/EPA regulatory model AERMOD, U.S. EPA, 2002. [3] Rabl, A., Spadaro, J.V., McGavran, P.D., Health risks of air pollution from incinerators: a perspective. Waste Management & Research, ISWA, 1998.