• CO2 excluding short term cycle (biomass burning)
• CO2 short term cycle only (biomass burning)

The study involved African-American and Hispanic children born in the Bronx or Northern Manhattan (New York) between 1998 and 2006. Their mothers, aged from 18 to 35 years, underwent personal air monitoring for PAH exposure during pregnancy. The children were followed up to age 7 years.

After adjustment for child’s sex, age at measurement, ethnicity, and birth weight and maternal receipt of public assistance and pre pregnancy obesity, higher prenatal PAH exposures were significantly associated with higher childhood body size.

Pregnant women exposed to higher concentrations of polycyclic aromatic hydrocarbons were more than twice as likely to have children who were obese by age 7 compared with women with lower levels of exposure.

Poor diets and physical inactivity are the main culprits for obesity, but there is new evidence that air pollution can play a role. Indeed, for many people without the resources to buy healthy food or the time to exercise, prenatal exposure to air pollution make them even more susceptible to obesity.

The increased risk of obesity is another negative effect of exposure to PAHs. Other well known consequences of prenatal exposure to PAHs are a negative effect on childhood IQs and linkage to anxiety, depression and attention problems in young children. Polycyclic aromatic hydrocarbons are also known carcinogens.

It is therefore evident that exposure to PAHs must be reduced. This can be done, for example, by reducing the traffic of diesel vehicles from the city centres.

References

A. Rundle, L. Hoepner, A. Hassoun, S. Oberfield, G. Freyer, D. Holmes, M. Reyes, J. Quinn, D. Camann, F. Perera, R. Whyatt. Association of Childhood Obesity With Maternal Exposure to Ambient Air Polycyclic Aromatic Hydrocarbons During Pregnancy. American Journal of Epidemiology, 2012; DOI: 10.1093/aje/kwr455

Perera et al. Prenatal Airborne Polycyclic Aromatic Hydrocarbon Exposure and Child IQ at Age 5 Years. Pediatrics, August 2009; DOI: 10.1542/peds.2008-3506

Frederica P. Perera, Deliang Tang, Shuang Wang, Julia Vishnevetsky, Bingzhi Zhang, Diurka Diaz, David Camann, Virginia Rauh. Prenatal Polycyclic Aromatic Hydrocarbon (PAH) Exposure and Child Behavior at age 6-7. Environmental Health Perspectives, 2012; DOI: 10.1289/ehp.1104315

The new rule will go into effect in June 2012. Starting from such date to January 2015 natural gas well operators will be allowed to either

• flare the gas that escapes during the drilling process or

• capture it through green completions technology (i.e. systems to reduce methane losses during well completions by capturing what is lost during drilling). The green completions technologies are already widely deployed at wells.

Starting from January 2015, industries must reduce emission completion, the option of solely flare gases will be no more valid.

There are two advantages in green completions technologies:

• they are expected to reduce by 95 percent the emissions of harmful air pollutants from wells, and

• they will enable companies to collect additional natural gas that can be sold.

Among the pollutants emitted from oil and gas wells, there are VOCs (volatile organic compounds), which are responsible for the formation of secondary pollutants (for example ozone), and are dangerous for the human health (benzene for example causes cancer). Moreover the rule will also reduce the emissions of methane, which is a very important greenhouse gas.

Flares are important devices to reduce the emissions of such gases. Enviroware has developed a software which helps to characterise their emissions, estimating the height and diameter of their stack, and evaluating the thermal and acoustic impact.

]]> http://www.enviroware.com/us-epa-air-pollution-standards-for-oil-and-natural-gas-wells/feed/ 0 http://www.enviroware.com/toxflam-online-a-web-tool-for-air-quality-modelling/?utm_source=rss&utm_medium=rss&utm_campaign=toxflam-online-a-web-tool-for-air-quality-modelling http://www.enviroware.com/toxflam-online-a-web-tool-for-air-quality-modelling/#comments Thu, 05 Apr 2012 16:26:02 +0000 admin http://www.enviroware.com/?p=1214 TOXFLAM is a software code based on an analytical solution to the atmospheric dispersion equation for inert substances that may undergo a first-order chemical or physical decay and have a density that is similar to the air density (Bianconi and Tamponi, 1993). This general analytical solution to the dispersion equation describes releases of any duration in a finite mixing layer. The instantaneous release solution (the so-called “Gaussian puff solution”) and the continuous release solution (the so-called “Gaussian plume solution”, as the one used in ISC and AERMOD US-EPA models) are just particular cases of the general one. Moreover, analytical solutions of finite duration having different starting times can be superimposed to describe a time-variable release rate. The model can also give the finite or infinite exposure to the concentrations.

TOXFLAM can simulate the release from a point source (stacks or pool fires assimilated to point sources) and it can compute the plume rise of the pollutant due to mechanical or thermal momentum. TOXFLAM extrapolates the wind speed at the release height according to a vertical profile that depends on land characteristics (throught the roughness length parameter) and the atmospheric stability.

The full mathematical model description can be found in Bianconi and Tamponi (1993) along with sensitivity analysis results. Model validation against measurements can be found in Andretta et al. (1993). You can contact us if you’re interested in model’s applications or use under license.

A simplified version of TOXFLAM is available online on this website, where you can input your data (meteorology, source parameters, domain) and see on Google Earth the results. You can start from one of the three examples of application that are provided and then change the input values to see how this affects the results.

This document introduces a new software to estimate the atmospheric impact of industrial flares. The software allows to design stacks with correct height and diameter in order to minimise risks for workers and nearby structures. It also allows to estimate the thermal impact and acoustic impact of existing or future flares. Flame tilting due to the wind speed can be modelled with a simple methodology and with the Brzustowski and Sommer (B&S) approach. The software also determines the flue gas composition starting from the fuel gas composition. The effective source parameters to be used as input in atmospheric dispersion models are determined with two different methods: US EPA SCREEN3 and TCEQ.

A PDF version of this document is available here.

Software availability

The FLARES software described herein has been developed by Enviroware srl in Visual Basic using the .NET framework. It works under any recent Windows operating system. It can be evaluated at no cost for a limited time period. The permanent license of the software is bound to a single PC, and can be purchased from the web site or directly from the software.

1. Introduction

The majority of chemical plants and refineries have flare systems designed to relieve emergency process upsets that require release of large volumes of gas via flaring or venting operations. Flaring operations are the controlled burning of gases, in open flames and open air, in the course of routine oil and gas production. Venting, on the contrary, is the controlled release of gases into the atmosphere, without combustion.

The gas combustion usually occurs at the top of a flare stack which is placed in a remote site of the plant. Flare systems have been technically described in many papers (e.g. Bader et al. 2011; Hong et al. 2006).

When a flare operates, it generates noise, heat radiation, and emits atmospheric pollutants. If the combustion is efficient, which means to have a good mixing between the fuel gas and air, the emitted gases are mainly water vapour and carbon dioxide. Even if the combustion efficiency may be higher than 90%, other pollutants are generally present, such as carbon monoxide (CO), nitrogen oxides (NOX), sulphur dioxide (SO2), volatile organic compounds (VOC) and particulate matter (PM). VOCs derive by incomplete combustion of the flared gas, or by its conversion to other compounds, such as aldehydes or acids. However, VOC elimination is nearly complete, exceeding the 98%. Concerning smoke formation, it is most probable in streams with high carbon/hydrogen mole ratio (greater than 0.35). Some regions of the world are heavily affected by flares pollution (Obia et al., 2011).

Industries would have all the advantages to avoid flaring, however such operation is necessary under different circumstances as, for example:

• to release high pressure in the plant and avoid catastrophic situations;

• after process upsets, equipment changeover or maintenance;

• to burn vapours collected from the tops of tanks as they are being filled;

• to release unburned process gas from the processing facilities;

• to eliminate the excess gas which cannot be supplied to customers.

Both elevated and ground level flares exist. They can also be classified according to the method used for enhancing mixing between air and fuel. In the steam-assisted flares steam is injected into the combustion zone to promote turbulence and entrain air into the flame. In the air-assisted flares forced air is used to provide the combustion air and mixing; these flares can be used when steam is not available. The non-assisted flares do not have any mechanism for enhancing the air-flame mixing; they are used for gases with a low heat content and a low carbon/hydrogen ratio (such ratio is related to the smoke production). Finally, the pressure-assisted flares use the vent stream pressure to promote mixing between air and flame.

This document describes the structure and some example applications of a software specifically developed to correctly design the stack of subsonic flares, evaluate the heat radiation and noise levels of subsonic or sonic flares, and determine the flue gas composition and the effective stack parameters for atmospheric dispersion applications.

2. Software description

2.1 Theory

The calculations implemented into the software are mainly based on the methodology described in the ANSI/API Standard 521, fifth edition January 2007 (API, 2007), and on the US-EPA technical documentation (EPA, 1995).

The radiation level K (kW/m²) at distance d (m) is calculated, assuming the flame as a single radiant point located at its centre, using the following equation:

where F is the fraction of heat released which is radiated (-), Q is the heat liberated (kW) obtained multiplying the mass flow rate (kg/h) by the heat of combustion (kJ/kg), and τ is the transmissivity (-), which is the fraction of heat radiated transmitted through the atmosphere.

The transmissivity depends on how atmosphere absorbs the radiated heat. Since water vapour is the most important absorber in air, relative humidity will play an important role. In fact, an empirical equation to calculate transmissivity is (Brzuwstoski and Sommer, 1973)

where RH is the relative humidity (%).

A graphical empirical relation allows to relate the heat liberated by the flame to its length. Such graphical relation, and others, have been digitised within the software. Flame tilting due to the wind can be calculated either with a simple method which relates the horizontal and vertical displacement of the flame centre with the ratio between wind speed and gas exit speed, or with the Brzuwstoski and Sommer (1973) approach which depends on the lower explosive limit concentration of the flared gas and on the jet thrust and wind thrust parameters. Once determined the flame length and its tilting, the position of the radiant point is placed half way between the flare tip and the flame tip.

The stack height and diameter can be calculated in two different ways according to the user choice:

• by specifying an allowable radiation at an horizontal distance from the stack base (point B in Figure 1) or

• by specifying an allowable radiation directly under the flame centre (point A in Figure 1).

The distance from the flame centre to the point of interest (A or B) is calculated by inverting equation (1) and assuming a constant unitary transmissivity. If the constrain for determining the stack height is the maximum radiation allowed directly under the flame centre, once calculated the distance d, the stack height is obtained by subtracting the displacement of the flame centre from the stack tip height, which is known. If the constrain is the allowable radiation at a specific horizontal distance from the stack base, the stack height is obtained by applying the Pythagoras’ theorem (the displacement of the flame centre from the stack tip, Δh and Δx, are known after the tilting calculation).

The stack diameter is determined by inverting the equation for the Mach number, which is an input value, and which depends on the mass flow rate, the relative molecular mass of the flare gas, the flowing temperature, the absolute temperature within the flare tip and the compressibility factor.

Once the theorethical stack height and diameter have been calculated with the above procedure, their actual values must be chosen on the market from those just greater than the calculated values.

Equations (1) and (2) are used for determining the heat radiation levels at specific horizontal distances from the stack base and at specific heights above the ground.

The noise levels are calculated as explained in API (2007). Initially the noise level at 30 m is calculated starting from the mass flow (kg/h), the speed of sound in the gas (m/s), and the pressure ratio, which is the ratio between the static pressure upstream from the restriction and the absolute pressure downstream of the restriction. At distances greater than 30 m the noise level decays in a way proportional to the logarithm of the distance.

The flue gas parameters are determined starting from the stream composition expressed in percent of volume, the mass flow rate and the excess of air, and applying stoichiometry. Starting from these data the methodology initially calculates the gas stream parameters (molar weight, normal density, volume and mass flow rate, heat released, and moles of C, H, O, N and S), then the flue gas parameters are determined (molar weight, mass and volume flow rate, normal density, water content, oxygen content in dry flue gas, dry volume flow rate, concentration of pollutants and emission rates of pollutants). The emission rate of SO2 is calculated starting from the sulphur contained within the fuel gas stream. The emission rates of the other pollutants (NOX, CO, VOC and PM) are calculated using the AP42 emission factors. The VOC emission rate can be alternatively calculated indicating a destruction efficiency. For example, if the destruction efficiency is 98%, the 2% of the VOCs within the fuel gas stream will be emitted into the atmosphere.

The effective stack parameters which must be used within the atmospheric dispersion models (e.g. AERMOD, CALPUFF) are calculated using two different methods: the US EPA SCREEN3 (EPA, 1995) method, and the TCEQ (Texas Commission on Environmental Quality) (TCEQ, 2004) method. The main difference between the two methods is that the TCEQ method does not consider an additional plume height for the length of the flame. Both the two methods equate the buoyancy flux calculated by the Briggs (1969) equation with the buoyancy flux calculated starting from the actual buoyancy of the flare combustion gases.

where Q_h is the sensible heat obtained as Q_h = FQ, expressed in cal/s, which is known. Assuming specific values for the stack temperature T (K), the air temperature T_a (K) and the stack exit velocity v (m/s), by inverting the equation, it is possible to obtain a value for the equivalent diameter d_E (m) as

The stack temperature T is often assumed to be 1273 K, while the ambient temperature T_a and the stack exit velocity v depend on the specific location. The exit velocity v must be chosen to avoid the stack tip downwash, therefore v > 1.5 u, where u is the wind speed. If v = 20 m/s and Ta = 293 K, a simplified form of the above equation is

With the TCEQ method the fraction of heat radiated F is calculated by means of the Tan (1967) expression: F = 0.048 M^0.5, where M is the molar weight of the gas stream. On the contrary, with the SCREEN3 EPA method F is usually assumed equal to 0.55.

Using the TCEQ method, the equivalent diameter, the exit speed and the exit temperature are used for calculating the plume rise, and the effective release heigth of the plume is obtained by the sum of the flare stack height and the plume rise height. No additional height for the length of the flame is considered, therefore the user will insert in the atmospheric dispersion model the equivalent diameter, the exit velocity (e.g. 20 m/s), the exit temperature (e.g. 1273 K), and the flare stack height, together with the emission rates of the pollutants.

It is observed that for complete combustion of hydrocarbons the temperature of the flare flame should be 1200 K (Attar et al., 2007), which is also the temperature at which NOX begins to form.

With the EPA SCREEN3 method, an equivalent stack height h_E is calculated by adding to the stack the flame length, assuming a tilt of 45 degrees:

where h_S is the flare stack height and Q is the total heat release (cal/s). When this method is adopted, the user will insert in the atmospheric dispersion model the equivalente diameter, the equivalent stack height, the exit velocity (e.g. 20 m/s), the exit temperature (e.g. 1273 K), together with the emission rates of the pollutants.

2.2 Features

FLARES is a Windows application written in Visual Basic using the .NET framework. A brief list of the main features of the software are:

• Ability to size (i.e. to calculate the minimum values of height and diameter) subsonic industrial flares according to a specified allowable radiation;

• Calculation of thermal radiation and noise levels of subsonic and sonic flares;

• Flame tilting calculated with the Simple approach and with the Brzustowski and Sommer approach;

• Calculation of the flue gas composition starting from the fuel gas composition;

• Calculation of the effective stack parameters needed for dispersion modelling;

• Ability to works with SI (International Standard) and USC (United States Customary) units;

• Load and georeferentiate raster base maps (automatic georeferentiation if the base map file has an associated world file);

• Position the stack over the base map;

• Thermal radiation levels and noise levels are shown at any point moving the mouse over the base map;

• Plots radiation levels and noise levels over the base map;

• Stack can be exported in KML format for Google Earth as a 3D cylinder;

• Thermal radiation levels and noise levels can be exported in KML format for Google Earth.

3. Flue gas composition and atmospheric dispersion

Mass flow rate and stream composition allow to calculate flue gas parameters, among which the heat liberated by the flame and the molar weight, and the flue gas parameters. The net heat release can be used for verifying if the flare will meet the minimum heating value requirements of 40 CFR § 60.18.

Starting from the stream composition indicated in Table 1   the values indicated by Alizadeh-Attar et al. (2007) are used as an example   and using a mass flow rate of 10000 kg/h, an excess of air equal to 250%, a reference oxygen value equal to 3%, and a molar weight of the air equal to 28.85, the output stream composition reported in Table 2 is obtained. Remembering that 1 Btu corresponds to 0.293 Wh, and 1 Nm3 corresponds to 37.326 scf, dividing the total heat released of 578520 kW by the volume flow rate of 47559.2 Nm³/h, and transforming the units, a heating value of 1112.3 Btu/scf is obtained, which compares very well with the value of 1111.5 Btu/scf observed by Attar et al. (2007).

Table 1: Input stream composition (as in Alizadeh-Attar et al., 2007).

Table 2: Output stream composition.

The flue gas parameters calculated for the stream composition reported in Table 1, are reported in Table 3. The SO2 concentration and emission rate are calculated starting from the sulphur content in the fuel assuming that all H2S is converted to SO2, while concentrations and rates of the other pollutants are calculated starting from the AP42 emission factors assuming a light smoking flare (for particulate). The VOC emissions can alternatively be calculated by specifying a destruction efficiency as a software input.

Table 3: Flue gas parameters.

The effective stack parameters needed for calculating the atmospheric dispersion of the pollutants emitted by the flare are calculated using an exit temperature of 1273 K, an exit speed of 20 m/s and an average ambient temperature of 293 K. Moreover, for the SCREEN3 EPA methodology a radiating heat loss equal to 55% is considered, while for the TCEQ methodology such value is calculated by the software as explained by Tan (1967). The calculated effective stack parameters are reported in Table 4. If the SCREEN3 methodology is used, the effective stack height to be specified within the dispersion model is the actual stack height plus the additional geometrical height (which represents the flame length).

Table 4: Effective stack parameters to be used in atmospheric dispersion models.

The two different methodologies for calculating effective diameter and height will reflect on different air quality values when used within atmospheric dispersion models. As an example, the atmospheric concentration values have been calculated using both the methodologies with the TOXFLAM (Bianconi and Tamponi, 1993) analytical dispersion model, assuming a physical stack height of 20 m. The release of a generic pollutant with a rate of 1 g/s lasting for 10 minutes has been assumed. Figure 2 shows the time trend of the instantaneous concentrations estimated with the parameters indicated in Table 4 at a downwind distance of 500 m (wind speed 2 m/s, mixing height 1000 m, air temperature 293 K, stability class B), and 1000 m (wind speed 5 m/s, mixing height 1200 m, air temperature 293 K, stability class D). It is observed that the elevated values are due to the fact that they are instantaneous concentrations, not hourly concentrations. In fact, the hourly concentration calculated for example at 500 m downwind is 27.1 µg/m³with the EPA-SCREEN3 methodology, and 29.1 µg/m³ with the TCEQ methodology. Figure 3 shows the maximum instantaneous concentrations calculated at different downwind distances from the flare with the two methodologies. It is observed that, due to the lower emission height, the TCEQ methodology gives always greater concentrations.

Figure 2a: Instantaneous concentrations calculated at 500 m and with two different meteorological conditions.

Figure 2b: Instantaneous concentrations calculated at 1000 and with two different meteorological conditions.

Figure 3: Maximum instantaneous concentrations calculated at different distances (wind speed 5 m/s, temperature 293 K, mixing height 1200 m, stability class D).

4. Stack sizing

A flare must be located so that it does not present a hazard to surrounding personnel and facilities, moreover for safety reasons a stack is used to elevate it. The height of the stack is determined based on the ground level limitations of thermal radiation intensity. If high surrounding structures are present close to the flare, the limiting thermal radiation may be calculated above the ground at a specific height. In this paragraph the stack diameter and height are calculated by specifying a maximum allowable thermal radiation value at ground at a specific horizontal distance from the stack basis (i.e. point B in Figure 1). The stream parameters calculated in the previous paragraph will be used as input values.

The allowable thermal radiation must be chosen accordingly to the user needs and the plant features. In this example a value of 6.3 kW/m² will be used, which represents the maximum radiant heat intensity in areas where emergency actions lasting up to 30 s can be required by personnel without shielding but with appropriate clothing (which means hard hat, long sleeved shirts, work gloves, long legged pants and work shoes). The decision to consider in the calculation the contribution of the solar radiation depends mostly on the value of the allowable radiation. The maximum values of solar radiation span, approximately, the interval (0.8 – 1.0) kW/m², depending on the position on the Earth and the period of the year. If the allowable radiation is relatively higher as the one indicated above, the inclusion of the solar radiation does not vary significatively the stack height (and then the costs).

On the contrary, for smaller values of the allowable radiation, the solar radiation must be considered because it has a high influence on the stack height. For example, for a continuous flare an allowable radiation of 1.58 kW/m² (which is the maximum radiant heat intensity at any location where personnel with appropriate clothing can be continuously exposed) could be specified, which is almost comparable to the maximum solar radiation.

The flare stack height for different design wind speeds (from 6 m/s to 20 m/s) has been calculated using the stream composition illustrated in the previous paragraph. The input values used are summarised in Table5. The stack height has been calculated both with the simple methodology and with the B&S methodology. The stack diameter does not depend on the wind speed, therefore the two methodologies give always the same result of 0.33 m. On the contrary, the suggested stack height for different design wind speeds and for the two methodologies is illustrated in Figure 4. The suggested stack height is smaller when the B&S method is used.

Once the stack has been sized, it can be exported in Google Earth in the exact place where it is (or will be), as a vertical cylinder with correct height and diameter. An example is available here.

Table 5: Input values used for dimensioning the flare stack.

(*) The air temperature and the lower explosive limit concentration are needed only for the B&S method.

Figure 4: Stack height calculated for different design wind speeds and with the two methodologies (simple and B&S).

It is observed that the B&S method describes the flame tilting, and then the location of the flame centre, better than the simple method. However, the flame tilting with the B&S method must be calculated starting from non-dimensional parameters which contain the wind speed at denominator. This means that the B&S method cannot be used in completely calms conditions. Up to now it is not a big problem of course, because in calm conditions the flame is not tilted. Anyway, the tilting is graphically calculated (by means of curves digitised within FLARES), and the curves are defined only for some values of the adimensional parameters. This means that there will be some conditions which define the minimum and maximum values of the wind speed for the applicability of the B&S method. Defining the parameters A and B as follows

where M_A is the air molar weight, M_G is the gas molar weight, C_L is the lower explosive limit (expressed as a fraction), U_G is the gas exit speed (m/s), D is the flare stack diameter (m), T_A is the air temperature (K) and T_G is the gas temperature (K), the more stringent of the two following conditions must hold:

Considering the values used in this paragraph, where the gas exit speed is 160.8 m/s, the last expression determines the most stringent conditions. The result is that in this case the wind speed must be greater than 0.22 m/s and smaller than 86 m/s; while the maximum wind speed is always respected, the minimum allowable wind speed must be checked. In some cases the minimum wind speed can be close to 1 m/s.

5. Thermal and acoustic impacts

As written in a previous section, beside atmospheric pollutants, when a flare operates, it generates heat radiation and noise.

In the following it will be considered a flare with a stack 22.9 m high, the input gas stream described in Table 1, and a wind speed of 10 m/s. The simple method will be used for the flame tilting. The parameters calculated by FLARES are a power liberated of 578697 kW, a flame length of 52 m, and an exit speed of 171 m/s.

Figure 5 shows the thermal radiation at ground as function of the horizontal distance from the flare stack. The values are shown for a constant unit transmissivity, and for transmissivities calculated for three values of the relative humidity: 40%, 60% and 80%. As shown, the transmissivity decreases increasing the relative humidity and the distance from the emitting point (the flame centre), therefore greater values are calculated for lower relative humidities. The maximum value of the thermal radiation is predicted in all cases for a point at about 15 m from the stack base, it is equal to about 10 kW/m² for the case with unit transmissivity, while it is 1.7 kW/m² less for the case with a relative humidity of 40%. The difference between the maximum thermal radiation predicted for a relative humidity of 40% and the one predicted for 80% is 0.3 kW/m², and it decreases when the distance from the flame centre increases.

Figure 5: Thermal radiation at ground as function of the horizontal distance from the flare stack.

For the same stack, and using some of the input parameters shown in Table 5, the noise level at ground as function of the horizontal distance from the flare stack has been calculated (Figure 6). A pressure ratio of 2 has been used. With these input data FLARES calculates a speed of sound within the flared gas equal to 321.7 m/s, and a noise level at 30 m from the stack tip equal to 98.6 dB. The software keeps such noise level constant for distances smalled than 30 m from the stack tip (it is observed that Figure 6 is plotted against the horizontal distance from the stack base, not against the absolute distance from the stack tip), then it decreases according to the logarithm of the distance.

Maps of thermal radiation and acoustic levels can be exported in Google Earth. An example is available here for the thermal radiation.

6. Conclusions

Flaring is a unavoidable process in the production of oil and gas. A certain quantity of the gas needs to be flared at the production site for reasons of safety. Other times part of the gas produced with the oil is flared for reasons that may be a combination of geography and availability of customers. Flaring is also used in other types of industries, not only oil and gas. For these reasons it is important to have reliable methodologies and software tools which allow to evaluate the environmental impacts of the flaring activity: thermal radiation, noise and air pollution. Moreover, to guarantee the safety of operators and industrial structures and equipments, it is important to design stacks with correct heights and diameters according to the quantity of the flaring gas, its properties and the meteorological conditions (wind speed) of the site. The FLARES software, based on the API 521 Standards, is capable to carry out all these activities. Moreover it calculates the properties of the fuel gas and the flue gas starting from the molar composition of the fuel gas. This document presents some theoretical aspect of the software, describes its features and presents some examples of application.

References

Alizadeh-Attar A., Ghoohestani H.R., Nasr Isfahani I. (2007) Reducing flare emissions from chemical plants and refineries through the application of fuzzy control system. Proceedings of the 6th WSEAS International Conference on Applied Computer Science, Hangzhou, China, April 15 17, 2007.

API Standard 521 Pressure-relieving and Depressuring Systems, FIFTH EDITION, JANUARY 2007.

Bader A., Baukal C.E. Jr and Bussman (2011) Selecting the proper flares systems. Chemical Engineering Progress (AIChE). pp. 45-50. http://www.aiche.org/cep/

Bianconi R. and M. Tamponi (1993) A mathematical model of diffusion from a steady source of short duration in a finite mixing layer. Atmospheric Environment, 27A, 781-792.

Briggs G.A. (1969) Plume Rise, USAEC Critical Review Series.

Brzustowski T.A. and Sommer E.C. Jr. (1973) Predicting Radiant Heating from Flares. Proceedings – Division of Refining, API 1973, Vol. 53, API Washington, D. C., p. 865-893.

EPA (1995) SCREEN3 Model User’s Guide. EPA-454/B-95-004. Pages 60.

Hong J., Baukal C., Schwartz R and Fleifil M. (2006) Industrial scale flare testing. Chemical Engineering Progress (AIChE). pp. 35-39. http://www.aiche.org/cep/

Obia A.E., Okon H.E., Ekum S.A., Eyo-Ita1 E.E., Ekpeni E.A. (2011) The Influence of Gas Flare Particulates and Rainfall on the Corrosion of Galvanized Steel Roofs in the Niger Delta, Nigeria. Journal of Environmental Protection, 2011, 2, 1341-1346.

Tan S. H. (1967) Flare System Design Simplified. Hydrocarbon Processing, Vol. 46, No. 1, 172 176.

TCEQ (2004) http://www.tceq.texas.gov/assets/public/permitting/air/memos/flareparameters.pdf

Crosswinds can also occur when travelling on roads, especially on large bridges and highways, which can be dangerous for motorists because of possible lift force created as well as causing the vehicle to change direction of travel.

Generally, a crosswind is any wind that is blowing perpendicular to a direction.

Each aircraft has a uniquely stated maximum crosswind component derived from flight test experiments. For example a Boeing 727-200 has a maximum crosswind component of 35 knots (17.8 m/s), while a Cessna 172 has a maximum crosswind component of 17 knots (8.7 m/s).

Wind coverage is defined as the percentage of time that crosswinds are below an acceptable velocity. According to the FAA standards (FAA AC 150/5300-13), the minimum wind coverage considering all the observations is 95 percent. This means that for the 95% of the time, the crosswind component must be smaller than the maximum crosswind component of the aircrafts landing in a specific airport.

In designing runway orientation, the most desirable runway is one that has the largest wind coverage and minimum crosswind components.

The WindRose PRO software can be used for analysing a long series of data and calculating, for each possible runway direction, the wind coverage and the crosswind components (maximum, average and median). The software can also be used to evaluate the correct orientation of an existing runway. Moreover, since WindRose PRO allows date/time filtering of the input data, it is possible to evaluate the wind coverage and the crosswind components even for airports which work only in particular seasons (for example during summer) or only during day time. For each direction, the WindRose PRO output contains information about the wind coverage, the maximum crosswind from left and right, the average and the median (i.e. the 50^th percentile) of the absolute value of the crosswind components, the maximum headwind, the maximum tailwind and the average and the median of the absolute value of the headwind components.

As an example, we considered the Cagliari Elmas airport (Sardinia, Italy) whose runway is oriented from South East to North West. We collected the METAR data of the airport for the period 2008-2011, decoded them and prepared for WindRose PRO. More than 98% of the data were valid. The derived wind rose is represented in the following figure (left). It is observed that the prevailing wind direction is aligned with the runway, as it must be. The wind rose of the maximum (red) and average (green) speed shows that the maximum wind speed observed from 2008 to 2011 is 18 m/s, and comes from North West.

The wind coverage has been calculated for a hypothetical aircraft with a maximum crosswind speed of 8 m/s. WindRose PRO has been instructed to test 72 possible runway orientations, starting from 0 degree with steps of 5 degrees. It is observed that orientation N must be intended as the orientation from N degrees to (N+180) degrees. The following figure (left) shows that the maximum wind coverage, which is equal to 100%, is obtained for the two directions 115 degrees and 120 degrees. Only directions from 25 degrees to 50 degrees have a wind coverage smaller than the 95% indicated by the FAA. Finally, an example of numerical results of WindRose PRO is also presented in the figure (right) as a table. For each runway orientation the following parameters are calculated: wind coverage, maximum crosswind from left and right, average and median of the absolute values of crosswinds, maximum headwind and tailwind, average and median of the absolute values of headwind.

For any additional information on WindRose PRO, visit this page or contact us.

Acknowledgements: the foreground image has been taken from the Crosswind landing article on Wikipedia.

A wind rose is a chart which gives a view of how wind speed and wind direction are distributed at a particular location over a specific period of time. It is a very useful representation because a large quantity of data can be summarised in a single plot.

The first step to plot a wind rose with an electronic data sheet is to organise the wind data in a table according to their direction and speed classes. In other words the joint distribution of wind direction and speed must be calculated, as shown for example in the next figure. Each yellow cell contains the number of events observed over a specific time period for a specific combination of wind direction and speed. For example, wind blowing from North (N) with a speed smaller than 1 m/s has been observed 51 times, while wind blowing from North East (NE) with speed between 1 m/s and 2 m/s has been observed 159 times. If available, the user may also specify the average wind speed for each direction, as shown for example in the green cells. The total number of events and the corresponding percentages for each direction and wind speed class are automatically updated.

The example file uses 16 directions and 6 wind speed classes, but their number and contents can be easily modified.

Once the number of observations for each direction and wind speed class has been specified for each yellow cell, three charts are produced: the wind rose, the wind direction distribution and the wind speed distribution. If the average wind speed for each direction is also specified, then a fourth chart is produced representing the rose of the average wind. Examples of these four charts are reported in the following images.

The joint distribution of wind direction and speed must be determined by the user. This task might require long times, particularly for large time series of data. In a wind rose the length of each arm is proportional to the number of events, or the frequency, at which wind was observed from that direction. For a specific direction, the different wind speed frequencies sum up to give the total length of the arm. The wind rose plotted with the Microsoft Excel or Open Office Calc files does have such feature. If you need more professional wind roses and more complex analysis of your data, you might want to evaluate WindRose PRO.

In order to obtain a harmonised approach in air quality modelling over Europe, the Forum for Air Quality Modelling in Europe (Fairmode) was established in 2008 as a joint action of the European Environment Agency and the European Commission’s Joint Research Centre (JRC). The technical reference guide is an output of that joint action

Air quality models are very important tools, since they allow to:

• Assessing the existing air quality situation – for example showing exceedances of EU or national air quality standards, calculating population exposure to pollution and health impacts, and identifying contributions of air pollutants from different sources.

• Air quality forecasting – many national, regional and local authorities have established forecasting systems to warn the public when air pollution episodes are expected.

• Air quality planning – identifying possible measures to reduce emissions, and developing emissions reduction scenarios.

Precisely because of their importance, it is expected that AQ models be comparable, well documented, and validated for their required applications in order to achieve reliable modelling results.

The EEA technical report provides an overview of the use of models with regard to the Directive 2008/50/EC on ambient air quality and cleaner air for Europe. The guide has three key aims:

• provide common technical guidance for the use of air quality (AQ) modelling in relation to the EU’s AQ Directive;

• provide a central reference point for the development of a harmonised approach to modelling;

• promote good practice in AQ modelling.

The technical report is available from the EEA web site.

A wind rose is a chart which gives a view of how wind speed and wind direction are distributed at a particular location over a specific period of time. It is a very useful representation because a large quantity of data can be summarised in a single plot. This type of plot can be used not only to plot wind speed against wind direction, but to plot any variable which depends on wind direction, as for example the average or the maximum wind speed measured over a time period. If a monitoring station measures both wind and concentrations of air pollutants, it is possible to plot the concentration levels of a specific pollutant against wind direction to investigate if higher levels could be related to any specific source.

Wind roses contain important information and are used in different fields as, for example, in air quality studies, in designing energy saving buildings, and in positioning wind turbines.

The simple on line tool available in this web site allows to create wind roses with 16 directions, each one representing an arc 22.5 degrees wide. The first direction is centred on North (i.e. 0 degree), and the last one on NNW (i.e. 337.5 degree). As usual, the wind direction represents the direction from which the wind blows. The colours of the wind speed classes cannot be modified, they are red, yellow, green, cyan, and blue.

Two examples of usage of the tool will be shown in the following.

Example of usage: wind rose

When used for plotting a wind rose, the on line tool requires that the joint frequency distribution of wind direction and speed must be inserted. Therefore the user must analyse the wind data, group them according to the 16 direction, and then according to the wind speed classes adopted. This binning of the wind data is not done automatically by the on line tool, while other software packages (e.g. WindRose PRO) are capable to create the wind rose directly by reading the wind data in many formats.

When you access the page of the on line wind rose tool, some data are preloaded and, to save time, we are going to use such data in this example. Anyway, you can modify such values according to your needs. The first step is to assign a title to the plot we are going to create, we must write it within the Plot title text box, as an example we will simply write Wind rose. We will leave the Smoothing check-box unchecked. On the contrary, all the Draw check-boxes and the Fill check-boxes will be checked, so that the data of all the five columns will be represented and filled. We will change the text of the Labels, which will be used for the legend, inserting for example 0-1 in place of Class 1, then 1-2, 2-3, 3-4 and >4 in place of Class 5.

Click the Proceed button, and you will obtain the following plot. Try to uncheck some of the check-boxes to see how the plot changes.

In a wind rose the length of each arm is proportional to the number of events, or the frequency, at which wind was observed from that direction. For a specific direction, the different wind speed frequencies sum up to give the total length of the arm. The wind rose plotted with the on line tool does not have such feature, since all the frequencies are plotted starting from the centre of the plot, therefore it is more properly a radar plot.

Example of usage: average wind speed

Now suppose you want to plot the average wind speed for each direction. In the Plot title text box write Wind speed. Since there is only one variable to plot, only one Draw check-box must be checked, we will check only the first one. Uncheck the corresponding Fill check-box, so that the curve will be not filled. In the first label insert Average. Then fill in the first column with the following values for the average wind speed 1.3, 1.5, 2.5, 3.3, 2.6, 1.8, 1.7, 1.4, 1.3, 1.7, 2.0, 2.0, 2.3, 1.3, 1.1, 1.0. The values are in m/s and have been calculated using WindRose PRO starting from hourly data for a whole year. The average wind speeds must be inserted from N to NNW, so that 1.3 m/s is the value corresponding to N, and 1.0 is the value corresponding to wind coming from NNW.

Click the Proceed button, and you will obtain the following plot.

This same procedure can be followed for plotting air pollution concentration data if their average, or peak, values are known for each direction.

This article shows how to use the simplified version of the COPERT 4 methodology available on the Enviroware web site. The methodology is based on the contents of the EMEP/CORINAIR Emission Inventory Guidebook – 2007, available on the internet site of the European Environment Agency, more precisely on chapter 7, concerning Road transport. Chapter 7 provides the methodology, emission factors and relevant activity data to calculate:

• the emissions produced by the exhaust systems of road vehicles (SNAP codes 0701 to 0705),

• the non-exhaust emissions such as fuel evaporation from vehicles (SNAP code 0706) and

• the component attrition, which means tyre and brake wear and road abrasion (SNAP codes 0707 and 0708).

The simplified methodology allows to calculate the only exhaust emissions. For many European countries, it gives the bulk emission factors in terms of grams of pollutants emitted per kg of fuel consumed. The emission factors at national level have been obtained applying the detailed COPERT 4 methodology using the activity data derived from TREMOVE. Therefore the detailed COPERT4 methodology has been a-priori applied to obtain the simplified emission factors.

The vehicles categories considered by the simplified COPERT4 methodology are Gasoline Passenger Cars (gPC), Diesel Passenger Cars (dPC), Gasoline Light Duty Vehicles (gLDV), Diesel Light Duty Vehicles (dLDV), Diesel Heavy Duty Vehicles (dHDV), Buses, Mopeds and Motorcycles. The simplified methodology does not deal with LPGs, 2-stroke and gasoline heavy-duty vehicles because of their small contribution to a national inventory.

The simplified methodology allows to calculate the exhaust emissions of carbon monoxide (CO), nitrogen oxides (NOX), non-methane volatile organic compounds (NMVOC), methane (CH4), particulate matter (PM), and carbon dioxide (CO2). All PM emissions refer to PM_2.5, as the coarse fraction (PM_2.5-10) is negligible in vehicle exhaust.

The application of the simplified COPERT 4 methodology must be done keeping in mind that the emission factors

• correspond to a fleet composition estimated for year 2005, therefore their accuracy deteriorates as time distance increases from such year because new technologies appear and the contribution of older technologies decreases;

• correspond to national-wide applications including mixed conditions driving (from urban congestion to free flow highway).

The methodology can be useful for example in simplified emission inventories, where rough estimate of the transport contribution is required. It is observed that the methodology is not suitable to be applied over small areas (e.g. a single town), or for a small time period (e.g. few days), because in such cases it would be even more approximated.

Example of input data

The emission factors are given as function of fuel used by the transport sector, therefore the first step is to obtain information about the total amount of fuel used. Considering for example Italy, for the whole country and for year 2008, such information can be obtained from the internet site of the Italian Oil Union (Unione Petrolifera). In 2008 Italy has consumed, for the road transport sector, 11044 Gg of gasoline, and 25934 Gg of diesel.

Since we want to estimate the emissions in Italy, an assumption that we have to do is that all this fuel has been consumed in Italy, even if a fraction of it has been consumed abroad. Similarly there will be a fraction of fuel sold abroad and consumed in Italy.

Other assumptions are needed to split the fuel consumption among the vehicle classes listed above (gPC, dPC, gLDV, dLDV, dHDV, buses, mopeds and motorcycles). A precise calculation of the consumption split is beyond the scope of this article, however it is worth to say that there are methodologies and software which allow a reliable estimate of the consumption of each vehicle class. The EMITRA software, for example, uses the actual fuel consumption, the number of vehicles and the vehicles fleet as known input data, then calculates the total consumption starting from assumed values of the average speed on different road types and of the average trip length for each vehicle type. If the calculated consumption is equal to the actual consumption, or at least comparable within a degree of acceptability, the fuel split is automatically obtained. Otherwise the procedure is repeated using different values for the average speeds and the average trip lengths until the convergence is reached.

For this example we will simply assume that the fuel is consumed as summarised in the following pie charts, which is, for gasoline, gPC: 93%, gLDV: 4 %, mopeds: 1%, motorcycles: 2%, and for diesel, dPC: 41%, dLDV: 11%, dHDV: 43% and buses: 5%.

Therefore the Gg of fuel consumed by the different vehicle classes is gPC 10270.9; gLDV 441.8; mopeds 110.4; motorcycles 220.9; dPC 10632.9; dLDV 2852.7; dHDV 11151.6 and buses 1296.7.

These numbers are then used as input data for the on line procedure. We need to select Italy among the available countries, then we must insert the fuel consumption for each vehicle class, paying attention to the units because the above numbers are in Gg (i.e. kilotonnes), while the system needs them in Mg (i.e. tonnes). An example of the input mask is shown in figure.

Results

The results of the on line simplified COPERT 4 methodology are given both in numerical terms and graphically. A table (see the figure) gives the total emissions calculated for each pollutant and for each vehicle class. The emission units are automatically decided by the software starting from their values, they can be kg, Mg and Gg. Moreover, six pie charts, one for each pollutant, show the amount of emissions due to each vehicle class.

Using the input data discussed before, the amount of estimated emission due to road transport over the whole Italy are 1910.9 Gg of carbon monoxide, 703.2 Gg of nitrogen oxides, 206.7 Gg of NMVOC, 15.0 Gg of methane, 26.8 Gg of particulate matter and 116.3 Tg of carbon dioxide. As shown by the pie charts below, which are automatically produced by the on line simplified COPERT 4 methodology, the greatest amount of carbon monoxide is emitted by gasoline passenger cars (gPC), which is responsible for the emission of more than 81% of the total. More than 52% of nitrogen dioxides is emitted by diesel heavy duty vehicles, while passenger cars, both gasoline and diesel, are responsible for the emission of about 17% each one. Methane and NMVOC are mostly emitted by gasoline passenger cars (about 60% of the total). Important emissions of particulate matter, which is all PM_2.5, are due to heavy duty vehicles (more than 37% of the total) and to diesel passenger cars (more than 34% of the total). Finally, the greatest emissions of carbon dioxide are due to heavy duty vehicles (30.1%), diesel passenger cars (28.7%) and gasoline passenger cars (27.9%).

In a typical emission inventory, now that we have the total emissions of each pollutant, other steps would follow. Among these steps, four important ones are:

• the spatial disaggregation of the emissions (i.e. how they distribute over the territory);
• the temporal disaggregation of the emissions (i.e. how they distribute over the months, the days of the week and the hours of the day);
• the chemical speciation of the NMVOC (i.e. the determination of the chemical species within this pseudo-species which contains all the volatile organic compounds but methane);
• the NOX speciation into NO and NO2.

Concerning the chemical speciation, the CORINAIR methodology contains the fraction of species (alkanes, cycloalkanes, alkenes, alkines, aldehydes, ketones and aromatics) for each vehicle category and fuel type. Even the NOX speciation, indications are given within the CORINAIR methodology.

The size speciation of particulate matter would be another task in an emission inventory but, as stated above, all the road traffic exhaust emissions of PM refers to PM_2.5. Finally, concerning the PM speciation in elemental and organic carbon, the CORINAIR methodology contains ratios for different vehicle technologies.