Abstract: The cstimation of highway pollution is always a difficult task because the vchicle cmissions depend on several factors like traffic volume, specd, mode of start, fucl type and composition etc. Bccause of the complexity in cstimating traffic cmissions, some computer modcls wcre devclopcd. Presently emissions factor models such as MOBILE, EMFAC,COPERT and dispersion models such as CALINE 4, Urban Air shed Model are available. Models are also nccessary for forecasting and planning purposes. In this paper an attenpt has been made to discuss the fcatures of these computer models, through their input requirements and output.
1.0INTRODUCTION
Highway cmissions are dircctly related to traffic volume, vehicle typc, speed and mode. It is intuitively clear that the more velhicles that occupy a particular space, the greater the cmissions, not as apparent as the effect of vehicle type, and mode. For examplc, one would usually expect a hcavy truck to emit more air pollution (eg. CO) than an automobile. But a heavy truck, using dicsel as fuel may emit less CO han a gasoline powered automobile under the same conditions. The spced also affects emissions in two ways. First, higher speeds demand for higher power, which leads to more fuel consumpion and greater emissions. But with increased speed tlhe vehicle departs an arca more quickly so the emissions, usually reported in granms per mile, are reduccd. Finally the vehicle mode can drastically affect the emissions.
Because of the complexity in estimating the traffic emissions, modelling is required.
There exist two types of modelling processes, emissions quantification and dispersion modelling. The cmission quantification modelling involves the estimation of amount of pollutant load in tonnes. For this purpose, the motor vehicle emission factor models are required. The emission factor models give the emission factor of a particular pollutant in grams per mile. The emission factors so obtained are then multiplied by the vehicle activity to get the total emissions. The dispersion modelling is used to predict the concentration of pollutant. These are useful to predict the ambient air quality.
Models are also nccessary for forccasting and planning purposes. Models are presently being developed to quality combine metcorological forccast models with air pollution dispersion models to enable air forccastS.
* ix AA Tech student, Center for Transportation Engineering (cTe), Department of Civil Engineering. Nat1onal Institute of Technology, Il'arangal. ** Head of Transportation Diviston &Assistant Professor, Center for Transportation Engineering (cTe), Depariment of Cwil Fng1meering, National lnstitute of Techmology, Il'arangal 12 2.0 BASIC EMISSIONESTIMATION METHODOLOGY
Tlhe basic cquation used for estinating motor vchicle cnissions involvcs multiplying activity data by an appropriate cmission factor. It is as given below.
Ep VMT × EFp (1)
Where: Ep = Total enmissions of pollutant p; VMT =Vehicle Miles Travelled; and EFp =Emission factor of pollutant p.
2.1Motor Vehicle Emission Factor Models
The emission factor is defined as the estimated average emission ratc for a given pollutant for a givcn class of vehicles. Instead of simple published cmission factors, however, motor vehicle emission factors are derived from emission factor models. The reason for this is that cmissions from motor vchicles are morc complex and dynamic than most other source types. For example, changes in fucl characteristics, vehicle operating speeds, cmission control technology, ambient temperature, and altitude can all affect cmission factors. In order to account for these and other impacts, an emission factor modcl is normally used that includes the effects of many parameters. Estimates of vehicle emissions are obtaincd by multiplying an estimate of the distance traveled by a given class of vehicles by an appropriate emission factor.
The most advanced models are the U.S. EPA's MOBILE series (the current version is MOBILE 6), PART 5 and the EMFAC model devcloped by CARB. Thesc models use statistical relationships based on thousands of emission tests performed on both new and used vehicles. In addition to standard testing conditions, many of these vehicles have becn tested at other temperatures, with different grades of fuel. and under different driving cycles. Relationships have been devclopcd for vehicles at varying emission control levels, ranging from no control to projections of in-use performance of future low-emission vehicle fleets.
Another enmission factor model, the COPERT model (Andrias and others 1992) applies a methodology developcd by the CORINAIR working group on cmission factors to calculate cmissions from road traffic in the EUcountries.
2.1.IMOBILE 6 Model (http://www.epa.govotaq/models. htm)
MOBILE is an EPA model for cstimating pollution from highway vehicles. The MOBILE model can be Uscd in cvery state of USA except California where the EPA permits the use of EMFAC, The MOBILE modcl consists of an integratcd set of FORTRAN routines that gencrate hydrocarbon (HC), carbon monoxidc (CO), and nitrogcn oxide (NON) cmission factors for gasolinc and dicscl-powered on-road motor vchicles. Hydrocarbon emission factors can be exNpressed either as total hydrocarbons (THC). non methane hydrocarbons (NMHC), volatile organic compounds (VOC). total organic gases (TOG), or non methane organic gases (NMOG).
The first generation of the MOBILE model was created in the mid-1970s The MOBILE model has subsequently undergone several updates and revisions in order to account for changing environmental legislation and teclhnological advances. These updatcd versions also incorporate large amounts of newly collected emissions data in an attempt to more accurately estimate motor vehicle emissions. The most recent version of the MOBILE model is MOBILE 6.
MOBILE6 calculates emission factors for 28 ind1vidual vehicle types in low-and high-altitude regions of the United States. MOBILE6 emission factor estimates depend on various conditions, such as ambient tenmperatures, travel speeds, operating modes, fuel volatility, and mileage accrual rates
MOBILE is used to calculate current and future emission inventories of these emissions at the national and local level. These inventories are used to make decisions about air pollution polic at the local, state and national level. Inventories based on MOBILE are also used to meet the federal Clean Air Act's State Implemcntation Plan (SIP) and transportation conformity rcquirenments, and are sometimes used to meet requirements of the National Environmental Protection Act (NEPA).
The travel activity for MOBILE model is Vehicle Miles of Travel (VMT). The emission factors are multiplied by the VMT to obtain the emission estimates.
Alist of MOBILE6 inpul parameters is provided below. Most of these inputs are optional because the model will supply default values unless altemnate data is provided. Users must provide input data for calendar year, minimum and maximum daily temperature, and fucl volatility.
MOBILE 6 Input Parameters
Calendar ycar Month (January. July) Hourly Temperature Altitude (hugh, low) Weckend/weckday Fucl characteristics (Reid vapor pressure, sulfur content, oxygenate content) Humid1ty and solar load Registration (agc) distribution by vehicle class Annual mileage accumulation byvehicle class Diescl sales fractions by vehicle class and modcl year Avcrage spccddistribution by lhour and roadway DIstributionof vehicle miles travclled by roadway type Enginestarts per day by vehicle class and distribution by hour Eine start soak tinne distribution by hour Tripcnd distribution by hour Average trip lenyth distribution Hot soak duration DIsibution of vehicle miles trIvclled by vchicle class Full, partial, and multiple diurnal distribution by hour Inspection and maintenance (M) program description Anti-tampering inspection program description Stage lI refuelling emissions inspection program description Natural gas vehicle fractions HCspccies output Particle size cut off Emission factors for PM and HAPs
2.1.2 EMFAC (EMission FACtor) Model (http:/www.arb.ca gov msei)
The California Air Resource Board (CARB) has released EMFAC 2002 as thc latest version of the California Mobile Source Emission Inventory (Emission Factors) model for use in the state of California. The model development was coordinatcd with the Califormia Department of Transportation (Caltrans) EMFAC 2002 has a Windows-based interface that allows a user to perform various types of analyses without invoking a command-line program.
The main function of the EMFAC Model is to generate emission factor information for the numerous
vehicle classcs, such as heavy-duty trucks and passenger cars. The EMFAC Model can provide emission estimates for the State as a whole and individually for cach county, air district, or air basin The EMFAC Modcl also provides emissions for cach type of vehicle (cars, trucks, motorcycles, motor homes, etc), In addition, the EMFAC Model can provide emission rates for various conditions (different temperatures, humidities, and speeds) to allow for the evaluation of special scenarios such as day-specific emissions inventorics or assessIment of control mcasure effectiveness. It can produce emission factors for Total Organic Gases, Carbon Monoxide (CO), Oxides of Nitrogen and Particulate matter.
The EMFAC Model is used for a varicty of purposes. First, output from the EMFAC Model is used to create California's annual statewide emissions inventory. Because on-road motor vehicle emissions are a significant part of California's tolal emissions, the output from the EMFAC model provides an important part of tlhe inventory.
The EMFAC Model is also used to evaluate the effectiveness of various motor vehicle control programs, as well as to show lhow California motor vehicle emissions have changed over time and are projected to change in the future.
The travcl activity inputs for EMFAC model are Vehicle Hours of Travel (VHT), Number of trips and parking duration. The emission factors produced by the EMFAC model are used in the CALINE 4 dispersion model.
Input Requirements
Geographical area (state, air basin, district, county)
Speed fractions by scenario ycar, hour and vehicle class Calendar year, Month Model year for which the emission factors are necded Temperaturc, Relative Humidity and speed Vehicle trips per day (total trips per day, by vehicle class, by vehicle and fucl, by vehicle/fuel/hour) Total VMT (total trips per day, by vehicle class, by vehicle and fuel, by vehicle/fuel/hour) Inspection and maintenance (/M) program description RVP for the calendar year and season Hourly temperaturc profile for the scason Hourly RH profile for the season Speed fractions for the calendar ycar ldle times for heavy-duty vehicles
2.1.3 PARTS Emission Factor Model (Mexico Emissions Inventory Program Mauals, May 1996)
The U.S. EPA PARTS emission factor model utilizes FORTRAN routines that are similar to MOBILE to estimate particulate mater (PM) and sulfur oxide (SOx) emission factors for motor vehicles. Although the PARTS nmodel resembles the MOBILE model in several respects, it is at an carlier stage of development because less particulate emission data have been collected. This is mainly the result of ozone precursors (TOG, C0, and NOx) being given higher priority than PM in the U.S. Conscquently, some parameters that affect motor vehicle particulate emissions (e.g., temperature, inspection and maintenance [/M programs, fuel impacts) have not been modeled in PARTS. Also, several assumptions in the model (i.., driving cycles, fuel specifications, emission control systems, engine system deterioration rates) are valid only for the U.S. At the present time. the PART5 model has not been modified for outside use of the U.S.
2.14 COPERT Model (Rentz et al., December 1999)
The COmputer Programme to Calculate Emissions from Road Transport (COPERT) model (Andrias and others 1992) applies a methodology developed by the CORINAIR working group on emission factors to calculate emissions from road trafic in the EU (Eggleston and others 1991). COPERT can be used to estimate vehicle emission factors for carbon monoxide, non-methane hydrocarbons, methane, oxides of nitrogen, total particulate matter, ammonia, and nitrous oxide. Fuel consumption estimates are also provided. Emission factors are estimated for urban, rural, and highwvay driving with an average automobile speed of 25 kilometers per hour, 75 kilometers per hour, and 100 kilometers per hour, respectively. COPERT accounts for cold-star emissions and evaporation losses, and uses an average trip length of 12 kilometers. The latest version in this series is COPERT II.
Input requirements for COPERT II
Fuel specifications Monthly temperatures Fuel reid vapour pressure Cold start parameters for average trip length Monthly canister cflicicncy Flcct Infornation Circulation Information Using his information the modcl calculates thc emission factors or CO, NO,, VOC, and PM for hot stabilized and cold excess starts and cvaporative emissions.
3.0 DISPERSION MODELLING
A Dispersion Model is a tool used to predict the concentration of a contaminant at a
receptor resulting from point, area or volume exhaust sources. Dispersion modcls use mathematical algorithms that simplify atmospheric dispersion and dilution phenomena. Typical model input variables include: contaminant emission rates; stack height, gas temperature; discharge velocity; flow rate; terrain topography; property dimensions; and building elevations. Dispersion modcls are used to prcdict the impact of contaminant emissions at locations surrounding the source.
For vehicular emissions dispersion modelling the following models are uscd:
HIWAY CAL3QHC CALINE4 UAM
The detailed description of the models is given below.
3.1 HIWAY: A Highway Air Pollution Model
The HIWAY (htp:/cesimo. ing ula ve GALA models hiway him) model was developed by U.S EPA for the evaluation of air pollution due to a motorway in flat terrain and for a given traffic condition (assumed to be
cquivalent to a linear pollution source). It is based on the Gaussian plume approximation, which is adapted 10 represent the pollution due to each lane, and corrected to account for low wind conditions.
HIWAY cannot deal with severe obstacles to the air flux (buildings, large tree areas) or with complex terrains
3.2Carbon Monoxide Dispersion Modelling
COemissions are usually generated during peak-hours and are measured in grams per vehicle-mile for use in dispersion models. These models use meteorological, transportation, emission, and other site-specific information topredict concentrations of pollutants downwind fron the modelled source
To nodel the dispersion of CO, models of the Gaussian line-source types are most widely used. If one considers a single isolated point source, such as the smoke stack of a power plant, the plune rises because it is warncr than the surrounding air. As the plume is adverted downwind, it is subjected to atmospheric tusbulence that causes it to diffuse from the source, thereforc, pollutant concentrations decrcase with Incrcasing distance from the centrclinc of the plume. The spreading and wafting of plumes will be influenced by wind speed, direction, and various other dispersion parameters. As the wind speed increases, the distance between the particles within the plume will increase. The net effect is that pollutant concentrations are generally inversely proportional to the wind sped.
The stability and mixing height will also influence the dispersion of the plume. If there is a high degree of atmospheric turbulence, this willtend to spread tlhe plume more rapidly. If the plume has spread vertically so that upper margin of the plume is contained by an inversion, the mixing height is reduced This increases the concentration of the pollutant between the ground and the base of the inversion layer.
The height of the emission source also affects the ground level concentrations. The greater the height of the emission, the further the plume willhave to spread, before significant concentrations are observed at the ground level.
These factors are the principles behind the Gaussian plume models, such as CAL3QHC and CALINE-4, used in mobile source-related analyses. These models calculate how pollutants are dispersed by representing the relationships discussed in the form of mathematical cquations.
CAL3QHC is the EPA-required dispersion model to be used in all areas of USA, cxcept California, which has developed the CALINE-4 model.
3.2.1CALINE 4Model (Paul Benson, P.E. June 1989)
CALINE 4 is a fourth generation line source air quality model developed by the California Department of Transportation that predicts CO impacts ncar roadways. Its main objective is to assist planners to protect public health from adverse effects of excessive CO exposure. The model is based on the Gaussian diffusion equation and employs a mixing zone concept to characterize pollutant dispersion over roadways. For given source strength, meteorology, and site geometry and site characteristics the model can reliably predict (1-hour and 8-hours) pollutant concentrations for receptors located within 150 meters of the roadway. The model can also predict the worst-case scenario (combination of wind speed, direction and stability class),which produces the maximum pollutant concentrations at the pre-identificd receptor points along the highway. CALINE4 comes with a Windows user interface.
Input requirement for CALINE 4
CALINE 4 highway ispersion model requires the following data as input:
Traflic parameters: Traffic volume (hourly and peak), traffic composition (two wheelers, three whcclers, cars, buses, goods vehicle etc.), type of the fuel used by each category of vehicles, fuel quality, average specd of the vehicles. Mctcorological parameters: Wind speed, Wind dircction, stability class, mixing height Emission parameters: Expressed in grams /distance travelled. It is diferent for different catcgories of vehicles and is a function of type of the vehicle, fuel used, average speed of the vehicle and cngine condition etc. Road geometry Road width. med1an width. length and orientation of the road number and length of cach links Type of the terain. Urban or nural. flat or hilly Background concentration of pollutants Receptor location
322 CAL3QHC Model (http www seas upenn edu -ys562 ug cal3qhc html)
CAL3QHC is generally used for modclling cm1ssion at intersections. although it can be used to model frcc-flow cond1uons as well
Input requircmenits for CAL3QHC
To un CAL3OHC, the following inputs are requircd hourly
Wind speed in meters per sccond. Wind angle with respect tothe positive Y-ax1s in degrees. Atmospheric-stability mcasurc. MIxing hcight and width in meters. Reccptor information (e g. number, height, angle of observation, and distance from the road). Roadway characteristics (e g,number of lanes and segment length). Section typc (c.g., at grade. fill, bridge, and depressed). Coordinates of the end points of the link, Signalcycle in usc, Frec-flovw and idle cmission factors (oblaincd from the MOBILE models). Background concentrations of the spec1fic pollutant, and Height of the pollutant source
CAL3QHC works by considerng the ntersection as a series of links on which velncles are in d1ferent modes of opcration The model takes the nput data and calculates average qucue lengths over the specificd time D1fferent enussion factors fronm the MOBILESa model are then appl1ed, on the bas1s of whether the vchicle is in idling (qucued) or n free flow Output is the concentration of the pollutant in parts per mllion Coestinmates are produced for both |- and 8-hr periods during the peak CO scason, generally the WInter.
Orone formation is predicted using photochemical d1spersion models that use mobile source emissions as inputs. The most widely uscd regional dispersion model is the UAM. UAM 1is an urban scale, three dimensional, grid tvpe numerical simulation model. This model incorporates a three-dimensional (3-D) photochenical grid to sinulate the atmosphere. Its purpose is to calculate concentrations of pollutants by simulaing plhy sical and cheical processes in the atmosphere that affect pollutant concentrations The UAM uses atnospheric diffusion or spccies continuityequations that represent a mass balance in which all of the relevant emissions, transport, dilusion. chemical reactions, and removal processes are expressed mathematically.
Input requirement for UAM
For urban applications the model is usually used to simulate a 2- or 3-day ozonc episode. The data
requirement for this is as follows:
Hourly estimates of the height of the mixed layer, which requires day-specific upper-air teinperatures and wind data at various times; A 3-D wind-field for each hour, cover, and the chemical Ambicnt temperature, humidity, atmospheric pressure, solar radiation, cloud species to be simulated, and by carbon-bascd class Hourly gridded emissions for NO, and VOCs. (VOCs must be classified because the UAM enploys carbon-based chenmical kinctic nechanisms.)
Output from UAM
Some typical output of the UAM include the following
Average conccntrations by hour and grid square for all specics and Instantaneous concentrations for cach species by grid square at the beginning of the average period
4.0 SCOPE OF THE MODELS AND CONCLUSIONS
All the models described above are designed for U.S. fleet only except the COPERT Model. The Models include he defaults, which are designed to represent the "national average" input data values. Users who desire a more precise estimate of local emissions can substitute information that more specifically reflects
local conditions. Use of local input data will be particularly common when the local emission inventory is to be constructed from separate estimates of roadways, geographic areas, or imes of day, in which fleet or trafficconditions vary considerably.
For instance, the Environment Canada has developed a Canadian version of the MOBILE5, named as MOBILE 5.C Anderson, Kanaroglou et al, TRR 1520) to adjust the differences between the Unite States and Canada. Similarly, In order to account for possibly different vehicle fleets and driving behavior in Mexico, the MOBILE model has been modified for the Mexico City, Monterrey, and Ciudad Juárez metropolitan areas (Mexico emissions Inventory Program Manuals). A technology cquivalence matrix for cxhaust and evaporative emission factors were developed. It can be seen that a 1994 Mexican light-duty gasoline vchicle (LDGV) would be cquivalent 1o a 1988 U.S. LDGV. In some instances, a certain Mexican model ycar might be equivalent to one U.S. model ycar for cxhaust control technology and another for Cvaporative control technology.
Beforc applying these modcls tothe Indian traffic conditions, there is nccd for adjustment of the modcl to rcflect thc Indian conditions. Otherwise, the results will produce uncertain results. REFERENCES "Users Guide to MOBILE 6 Mobile Source October 2002. Assessment and Standards Divis1onEmisston Factor Model EPA420-R-02-028 US. Environmental Protection Agency 2 Chariton Koundis., Leonidas Ntziachristos and ZisSis Samaras. "COPERT II., toCalculate Emissions from Road Transport" User Manual (Version 2 1). Computer Program 3 November 2000 Dana L Coe Douglas S Eisinger Jeffrey D. Prouty."User's Guide for CL4 A User-Friendly Interface for the CALINE 4 Model for Transportation Project Impact Assessments", User s Guide 4 STI-997480-1814-UG Fieber, J,B Austin and JHeiken Character1stics of MOBILE and EMFAC Models Maldonado, H "Methodology to Calculate Emisson Factors for On-Road Motor lehiclex Technical report California Air Resources Board 6 0 Rentz S Nunge. U Karl T Holtmann. T. Zundel, Final Report, "Feasibil1n Study on the Development of a DesIgn for an Emission Projection Afodel, December 1999 Paul Benson, PECALINE4 - ADIspersion Model for Predicting Air Pollution Concentrations near Highways" ReVised November 1986, Revised June 1989 Rad1an International, Mexico Emissions Inventory Program Manuals. "OLUME II Motor V'ehicle Inventory Development yfinal 9. USDepartment of Transportation, TMIP. 'work shop on Transportation Ar Oual1h Analys1s' Module 3. Travel and Emissions Model Interactions 10 William P. Anderson. Pavlos S Kanaroglou. Enc J Miller, and Ronald N Buliung. "Smulat1ng Automob1ile Emissions in an Integrated Urban Model' Transportation Rescarch Record I520 | Website References http //wvwwepa gov/otag/models htm http //www arb ca gov/msei/msen htm http //www tmip.fhwa dot gov http //www vergina eng.auth gr/mech/lat/copert http //www kincctrics com/ibrary/sp%20-%20air-modellngfinal pdf http://www ces1mo ing ula ve/GAIA/models/hiway htm http //www ess co at/GAlA/modcls/msqldb/modcls/hiway/toc htm http //www cpa gov/asmdnerl/urban html http //wwwccl rutgers cdu/-ss/thes1s/thesi-node 56 html http //www epa gov/ttn/conference/erl0/mobile/wolfpresentat1on pdf http //www scas. upenn cdu/-sys562/ug/cal3qhc html http //www strc org/cal3qhc html
