Measuring and Modeling Vehicle Emissions: Methodology and Applications

Transcription

1 Measuring and Modeling Vehicle Emissions: Methodology and Applications Nagui M. Rouphail, PhD Director, ITRE and Professor of Civil Engineering North Carolina State University Raleigh, NC, USA Distinguished Lecturer Seminar Series Center for Multimodal Solutions for Congestion Mitigation 1 University of Florida December 3, 2009

2 Acknowledgments The research program is a team, not individual effort Dr. H. C. Frey, Professor of Environmental Engineering, NC State University Dr. Haibo Zhai, CMU, formerly Post-Doc at NC State Dr. Alper Unal, World Resources Institute, Post-Doc Numerous graduate and Undergraduate students 2

3 Presentation Outline Motivation Description of Research Program and PEMS Application domains I. Direct assessment of traffic signal control II. Modal emission models III. Development of regional emissions inventories Summary and conclusions 3

4 Motivation Transportation: 28% of U.S. energy consumption Daily, highway vehicles consume: 8.9 million barrels gasoline & 2.7 million barrels diesel fuel Annually and nationally, highway vehicles contribute: Carbon Monoxide (CO) 60% 60 million tons Nitrogen Oxides (NO x ) 36% 7.4 million tons Volatile Organic Compounds (VOC) 27% 4.5 million tons Outlook More competition for limited energy resources Higher market penetration of Alternate Fuel Vehicles (AFV) Interest on how roadway infrastructure design and operations may effect energy consumption and emissions 4

5 Growth Percent 200% 150% 100% 50% Contribution To Greenhouse gases Motivation Growth in U.S. Highway Passenger Vehicle Miles Traveled, Real Personal Income, and Population (Percent Change VMT Growth Trends since 1970) Continue Real Personal Income Population Passenger VMT 0% Source: Calculated from U.S. Department of Transportation, Federal Highway Administration, Highway Statistics Perecentage of On-Road Vehicle Emissions in NEI 80% 60% 40% 20% 0% Year CO NOx VOC Contribution To National Emissions Inventory Despite Recent Dip 5

6 Thousand Vehicle Sold Total Flex-fuel Electric hybrid Diesel Gas Technology Fuel cell and electric Motivation Project MP of Alternative Vehicles From Annual Energy Outlook 2007, U.S. Department of Energy, alternative technologies are projected to exceed 27% of light duty vehicle sales by 2030 in the national range. Number of Vehicles 150, ,000 50, Number of E85 Fueled Vehicles in U.S. 6

7 Motivation Addressing national need to assess energy and emission consequences at the local and corridor levels, e.g. Comparing different controls (signals vs. roundabouts) Comparing different road designs (i.e., grade, curvature effects) Comparing different vehicles and fuels Comparing different routes and facility classes (freeways vs. arterials) Identifying emission hotspots 7

8 Why not MOBILE / MOVES? MOVES (Motor Vehicle Emissions Simulator) still under development; no official emission rate estimator released more likely in 2010 U.S. EPA s MOBILE6 is currently used to address a wide variety of air pollution modeling needs.. but limited ability to represent real world driving cycles and micro scale influences at high temporal and spatial resolution, such as acceleration and deceleration, road grade. one cycle used to represent driving on local roadways and freeway ramps. CO 2 emission rates not adjusted by speed and facility type. limited capability to model advanced technologies and alternative fuels. Lacking link based emissions estimators to couple with transportation models for high resolution emission inventory estimates. To estimate link level average emission rates for all vehicle classes To evaluate the sensitivity of average emission rates to average link based speed and facility type. 8

9 Research Program at NC State Founded upon extensive, real world data collection of vehicle dynamics, energy use and emissions (EUE) Focused on assessing variability in EUE due to: Vehicular factors (fuel, engine size, condition, etc.) Infrastructure effects (grades, signals, curves, etc.) Driver effects (level of agressiveness) Ambient conditions (temp., humidity, etc.) Data are also used to drive EUE model development at the micro (sec by sec) and macro (mean speed) scales Fully compatible with the proposed MOVES binning method Uses in vehicle instrumentation using Portable Emissions Measurement Systems (PEMS) 9

10 What is PEMS?? A Portable Emissions Measurement System that: Scans engine data and vehicle dynamics Reports Vehicle Emissions: via Gas analyzers for NO, HC, CO, CO 2 and opacity (Particulate Matter) Integrates vehicle location and micro movement: (GPS) Collects supplemental data e.g., ambient weather conditions, intersection position, etc.. Can be set up in vehicle in about 15 minutes; connections fully reversible 10

11 PEMS Details Engine Scanner OBD II Adapter System Interface Main Unit GPS Sampling probe 11 Frey, H.C., A. Unal, N.M. Rouphail, and J.D. Colyar, On Road Measurement of Vehicle Tailpipe Emissions Using a Portable Instrument, Journal of Air & Waste Management Assoc., 53(8): (August 2003).

12 PEMS Limitations Special retrofit may be needed for certain older vehicle models Relies on energy and emission sampling from one or few vehicles, but is not representative of the fleet Does not characterize local traffic information (average speed, density, flows) Other options Remote sensing Dynamometers 12

13 Emissions episodes account for the majority of trip emissions, Sample Micro scale Data PEMS Speed(mph) enters queue Morrisville Pkwy. crosses intersection Weston Pkwy. Aviation Pkwy. Airport Blvd Elapsed Time (minutes) Not necessarily Proportional to VMT 13

14 Application I Direct Assessment of Traffic Signal Timing Improvements Objective: Empirical Verification of Signal Timing Effects Impacts on Emissions Measurements on over 20 different vehicles 1,250 one way on road trips 4,000 miles traveled 160 vehicle hours 10 drivers Four corridors: focus on Walnut Street coordination effects Unal, A., N.M. Rouphail, and H.C. Frey, Effect of Arterial Signalization and Level of Service on Measured Vehicle Emissions, Transportation Research Record, Journal of the Transportation Research Board, No. 1842, pp (2003). 14

15 Application I Summary of Walnut St. results Speed(mph) Cary Towne Blvd. Maynard Rd. Mall Access Sturdivant Dr. Nottingham Dr. Buck Jones Rd. Crossroads Blvd. Dillard Dr. 20 U.S. 64/1 EB Ramp 0 0 enters queue 1 crosses signalized intersection 2 3 Elapsed Time (minutes) 4 5 BEFORE 1 CO (g/sec) Speed(mph) Elapsed Time (minutes) Cary Towne Blvd. Maynard Rd. Sturdivant Dr. Buck Jones Rd. U.S. 64/1 EB Ramp Dillard Dr. 20 Mall Access Nottingham Dr. Crossroads Blvd enters queue crosses signalized intersection Elapsed Time (minutes) CO (g/sec) Elapsed Time (minutes) Delays 40 % Stops 30 % CO 12 % NO 8 % HC 12 % 15 AFTER Controlled for vehicle, driver, arterial, direction and time of day (24 runs)

16 Application II Modal Emission Models Modes distinguish between power requirements and the resulting emission rates generated Can be defined at many levels Acceleration / Deceleration / Cruise / Idle (trip based) Vehicle Specific Power (MOVES paradigm Second by second) consistent with micro simulation output Speed Range (Link based emissions) consistent with travel demand model outputs 16

17 Application II Modal Emission Models 4 Mode model on Walnut Street 100 Percent of Driving Modes North South Cumulative Northbound Distance (mi) Idle Accel. Decel. Cruise 17

18 4 Mode Modal Emissions Model 100 NO (mg/sec) HC (mg/sec) CO (mg/sec) CO 2 (g/sec) Average Modal Emissions CO 2 Emissions (g/s) Idling 2 - Acceleration 3 - Deceleration 4 - Cruise Idle Cavalier Acceleration 2.2 L Caravan Deceleration 3.3 L Tahoe 5.3 Cruise L 18 Coelho, M., T. Farias and N. Rouphail (2006), Effect of Roundabout Operations on Pollutant Emissions, in Transportation Research Part D, Transport and Environment, Vol. 11 (5), pp

19 Application II Micro Scale Modal Emission Models Models of emission rates for LDGV (grams/sec) Vehicle Specific Power (VSP): Uses micro vehicle data accounting for the effects of instantaneous Speed ; Acceleration ; Grade ; Aerodynamics VSP = v 1 3 [ 1.1a (sin(tan ( grade))) ] + ρv VSP binned into 14 modes (changes in MOVES) Applicable to the estimation of emissions when actual or synthetic speed profiles are available: e.g., aasidra ; CORSIM, VISSIM or other micro models 19

20 Micro Scale Modal Emission Model using VSP VSP Mode Definition VSP Mode Definition 1 VSP < VSP < VSP < VSP < VSP < VSP < VSP < VSP < VSP < VSP < VSP < VSP < 33 EF m CO (g/s) E = VSP < VSP 39 N M n= 1 m= 1 U mn EF m VSP BIN (m)

21 Application II: Link Level Modal Emission Models Motivation: Application to macro supply and demand models Justification: Within narrow average speed bands, VSP link trip based distribution fairly stable Speed (km/h) Average Speed: km/h 9 Runs Distance (km) Percentage of Time (%) Average Speed: km/h 9 Runs VSP Mode Zhai, H., H.C. Frey, and N.M. Rouphail, A Vehicle Specific Power Approach to Speed and Facility Specific Emissions Estimates for Diesel Transit Buses, Environ. Science and Technology, 2008, 42 (21), pp

22 Average NO Emission Rate (mg/s) Average CO Emission Rate (mg/s) Application II: Link/ OD Level Modal Emission Models Sample: 588 runs covering 6 routes, 2TOD, 3 Primary Veh. (~ 195 hrs) 2005 Chevrolet Cavalier 2.2 L 2005 Dodge Caravan 3.3 L 2005 Chevrolet Tahoe 5.3 L Average Vehicle Speed (km/h) Chevrolet Cavalier 2.2 L 2005 Dodge Caravan 3.3 L 2005 Chevrolet Tahoe 5.3 L Average Vehicle Speed (km/h) Average HC Emission Rate (mg/s) Average Fuel Consymption Rate (g/s) Average Vehicle Speed (km/h) Chevrolet Cavalier 2.2 L Dodge Caravan 3.3 L 2005 Chevrolet Tahoe 5.3 L Chevrolet Cavalier 2.2 L 2005 Dodge Caravan 3.3 L 2005 Chevrolet Tahoe 5.3 L Average Vehicle Speed (km/h)

23 Effect of speed on LDGV Link Emission Rates Higher emission rates per unit time at higher speeds, BUT less travel time... so effect on total emissions varies CO 2 (g/s) CO (mg/s) >50 Speed (km/h) >50 Speed (km/h) NO (mg/s) HC (mg/s) >50 Speed (km/h) >50 Speed (km/h) Frey, H.C., K. Zhang, and N.M. Rouphail, Fuel Use and Emissions Comparisons for Alternative Routes, Time of Day, Road Grade, and Vehicles Based on In Use Measurements, Environmental Science and Technology, 42(7): (April 2008). 23

24 Application III : Regional Emissions Inventory Estimation An emissions inventory is a database that lists, by source, the amount of air pollutants discharged into the atmosphere of a community during a given time period. Emission inventory data are used to: calculate permit fees, perform toxic air pollutant modeling evaluations, conduct modeling for federally required criteria pollutant attainment and maintenance plans, provide appropriate data to the public, monitor progress and trends and (Source: EPA) 24

25 Application III : Regional Emissions Inventory Estimation: Highway Vehicles -Meteorology - Facility type - Average cycle speed - Vehicle class & age - Vehicle fuel & technology - I/M program, standards -Year Speed- and facility- specific link emission rates for a given technology Regional Travel demand model Basic Emission Rates Technology correction factors HEV, Diesel, CNG..(if necessary) Speed correction factors for a given technology Link-based vehicle volume/travel time Emission factor for a given technology 25 On-Road Mobile Source Emission Inventory

26 EF Generalized Link Emissions ( ) BER TCF SCF Y, C, f, v C, T T, f, V = α Y, T, f, V EF BER = Generalized Emission Factor /sec = Basic Emission Rate fixed Standard Driving Cycle, Vehicle, Fuel Alpha= Cycle Adjustment Factor TCF = Technology Correction Factor SCF = Speed Correction Factor T = Technology Y = Year F = Fuel V = Speed 26

27 Datasets for BER, SCF & TCF Parameter Vehicle Fuel & Technology Source Basic Emission Rates Speed Correction Factors Fuel Economy Technology Correction Factors Travel Patterns LDGV, LDDV, HDDT, HDDB LDGV, HDDT HDDB LDDV LDGV MOBILE6 NCSU PEMS EPA PEMS Portugal PEMS EPA LDDV, HEV, CNG Cars Fuel Economy Guide: EPA & DOE E85, HEV, CNG Cars EPA Cert. Tests B20 trucks, CNG Buses Charlotte/Triangle Region Models Literature* ITRE, NCSU 27

28 Basic emission rates (BERs) BERs (CO, NO x and HC) derived from MOBILE6 under average speeds of 19.6 mph (LA4 cycle) for light duty vehicles and 20.0 mph for heavy duty vehicles For light duty gasoline and diesel vehicles, basic CO 2 emission rates are estimated based on fuel carbon content & fuel economy For heavy duty diesel vehicles, CO 2 emission rates are derived from MOBILE6 28

29 Speed correction factors (SCF s) Ratio between a link s average emission rate for a speed and a rates for reference speed range (e.g km/h) Estimated using Vehicle Specific Power based on realworld speed profiles NO x (mg/sec) Speed (km/h) Speed Correction Factor HC CO NOx CO Speed (km/h) 29

30 Technology correction factors (TCFs) Account for emissions differences of alternative relative to conventional fuel and vehicle technologies HC, CO and NO x, for E85, HEV and CNG from EPA s annual certification tests For CO 2, for HEV and CNG from fuel economy comparisons, and for E85 from fuel combustion theoretical analysis (Zhai et al., 2007) For alternative heavy duty vehicles, TCFs are estimated from the literature 30

31 Emission Factors by Speed and Vehicle Technology on Arterials, CY LDGV E85 CNG LDDV HEV 120 CO 2 (g/s) CO (mg/s) NO x (mg/s) Speed (km/h) Speed (km/h) HC (mg/s) Speed (km/h) Speed (km/h) Frey, H.C., H. Zhai, and N.M. Rouphail (2009), Regional On Road Vehicle Running Emissions Modeling and Evaluation for Conventional and Alternative Vehicle Technologies, Environmental Science and Technology, 43(21):

32 Triangle Regional Model Network AM Peak Baseline Scenario Durham Chapel Hill Raleigh 32

33 Triangle Regional Model Vehicle Activity Item Summary 4 hour Morning Peak Vehicle Class b Freeway / Ramp Facility Type Arterial Local Collect. Total Link Average Speed (km/h) Number of Links in Model 1,580 3,990 9,510 15,080 Cars 3,589 3,775 3,185 10,548 Vehicle Kilometers Traveled (10 3 VKT) Trucks 1,687 1, ,744 Buses Percentage of Total VKT (%) Sum 5,322 4,858 4,208 14,388 Car Trucks Buses Sum

34 Regional Emissions Inventory TE = CT ( vol ) EF ct t ct ct ct Where: ct = combination of vehicle class and technology; EF ct = link-based emission factor for vehicle class (ct) (g/sec); t ct = average link travel time of vehicle class (ct) (second); vol i, ct = travel volume of vehicle on link for vehicle class (ct) (vehicles/hr) TE = total emissions for a single link (g/hr). 34 Source: Frey et al 2008

35 Scenarios Analyzed Baseline Scenario: Considers only conventional vehicle technologies for 2005 base year Alternative Scenario: considers conventional technologies and market penetration of alternative fleet technologies for the 2005 base year. Future Scenario: considers conventional and alternative vehicle technologies, emission control standards and programs for the calendar year

36 Scenarios Vehicle Technologies Vehicle Class Car Truck Fuel & Tech. Fleet Penetration of Each Vehicle Class (%) Present (2005) Future Scenario (2030) Baseline Alternative Future LDGV E HEV LDDV CNG EV & Fuel Cell HDDT B20 Trucks Bus HDDB CNG Bus

37 Relative Emission Changes During Morning Peak Hour Scenarios and Pollutants Comparison HC CO NO x CO 2 Alternative vs. Baseline No VMT Growth vs. Baseline VMT Growth vs. Baseline Full Useful Life EPA Emissions Standards by Tier in grams/mile Tier Phase Dates Weight Based? PCAR NOX PCAR CO PCAR PM LLDT NOX LLDT CO LLDT PM I YES II (Bin 6) NO

38 38 Network Emission Characterization Vehicle Type Pollutant HC CO Baseline Scenario 10.6% 0.6% Future Scenario 0.9% 5.5% 41.1% 31.6% 58.0% 62.9% 4.0% 0.2% Vehicle Miles Traveled Distribution 26.0% 0.7% 73.3% 5.6% 88.8% 17.4% 3.5% 95.8% 27.5% NO x 77.0% 2.2% 41.2% 69.0% 2.3% 36.2% Car Truck Bus 38 CO % 61.4%

39 Transportation Network Emissions Characterization by Facility Type Pollutant HC 28.8% Baseline Scenario 38.9% Vehicle Miles Traveled Distribution 29.3% 37.0% 32.3% 25.4% CO 44.5% 33.8% 30.1% 22.8% NO x 47.9% Freeway+ Ramp CO % 26.0% 43.7% % Arterial Local and Collector %

40 Sensitivity of Emissions Reductions to Market Penetration (η) of Advanced Vehicle Technologies Alternative vs. Baseline Scenario 5% 0% 1- η Conventional η Advanced 22% 36% E85 HEV LDDV CNG EV & Fuel Cell 37% Market Penetration 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% Perecentage of Emission Reduction -20% -40% -60% -80% -100% HC CO NOx CO2 40

41 Conclusions PEMS enable the use of direct emission measurements for assessing impact of infrastructure design and operations PEMS enables development of a variety of modeling platforms from micro to macro scale Emissions are highest under the acceleration mode, followed by cruise, deceleration and idle Signal coordination shows very strong promise as an emission reduction tool with about 15 20% average reduction across pollutants 41

42 Conclusions Emission rates (per sec.) increase monotonically with link or route average speed Emission rates (per km) decrease monotonically with link or route average speed.. Total emissions effect may vary.. Road grades, particularly those above 3% result in significant emission rate increases The largest fraction of emissions occurs on freeways including ramps in the morning peak hours, followed by arterials. 42

43 Conclusions Passenger cars are the major emission sources for HC and CO, whereas trucks are the major emission sources for NO x and CO 2 in the morning peak. Market penetration of alternative vehicle technologies may decrease emissions. Emissions reductions increase when market penetration increases. In the short term, HC and CO emissions have larger reduction potentials than NO x and CO 2. NO x may also have significant emissions reductions in the long term. Replacement of Tier I with Tier II vehicles has the strongest potential for emission reductions. Benefits also vary depending on VMT changes 43

44 Measuring and Modeling Vehicle Emissions: Methodology and Applications Nagui M. Rouphail Thank you for your attention!!! 44