December 04

1

On-Road Emissions in Asia Measured by Remote Sensing

Niranjan VescioPresentation at BAQ 2004

Agra, IndiaDecember 8, 2004

2

Presentation Outline

1. Introduction to Vehicle Remote Sensing

a. RSD Technology

b. RSD Measurements

2. Asia Remote Sensing Data India Data

3. Motorcycle Measurement Challenges

a. Improvement Plans

4. Concluding Remarks

3

Contributors Ms. K. Easter, Mr. K. Balakrishnam

US Asia Environmental Partnership

Dr. Balraj Bahnot, Director Automotive Research Association of India PB 832 Pune, India

Mr. Nitin Manawat, Managing Director Environmental Systems Products-India Powai, Mumbai, India [nitin.manawat@etest.com]

Dr. Donald H. Stedman, Gary A. Bishop University of Denver, Department of Chemistry and Biochemistry, Denver, CO, USA [dstedman@du.edu]

Dr. Peter McClintock Applied Analysis Tiburon, CA, USA [peter.mcclintock@pacbell.net]

India Study Sponsor

Data Collection

Data Analysis

December 04

4

Introduction to Remote Sensing

5

Measurement Principle

Spectroscopy (Light Absorption)

IR

Opacity

UV

6

Opacity (Smoke/PM) Measurement

Source:Presentation by David Kittleson, University of Minnesota, Department of Mechanical Engineering, et al “Chemical & Physical Characteristics of Diesel

Aerosol,” presented at the 12th Annual CRC Conference, April 15-17, 2002.

OldNew

0.550.23

Traditional OpacimeterNew Smoke Channel

1. Diesel EmissionsPM Mass Shift

Old to New1500nm to 250nm

2. Measurement λOpacimeter to RSD SF

550nm to 230nm

0.25 1.5

7

Remote Sensing Technology

www.rsdaccuscan.com

8

The On-Road Advantage

On-road emissions cause poor air quality.

Remote sensing measures on-road emissions.

9

1. RSDs Measure Ratios…

___CHX + ___N2 +___O2 ___CO2 + ___CO + ___CHX + NO2

2. RSDs Calculate Concentrations1. Simple Explanation - http://www.rsd-remotesensing.com/user_info.asp under “RSD for Non-Technicals”

2. Math Derivation - http://www.feat.biochem.du.edu/whatsafeat.html under “standard combustion equation.”

RSD MeasurementsRatios to Concentrations

Q

Q’

Q’’

10

Fuel Specific Mass EmissionsCO, HC, NO per kg of Fuel Burned

4.71)'2.23(1

28

QQ

Q

kgFUEL

gCO

4.71)'2.23(1

'2.244

QQ

Q

kgFUEL

gHC

4.71)'2.23(1

''30

QQ

Q

kgFUEL

gNO

2CO

COQ

2CO

HC'Q

2CO

NO'' Q

Where,

Smoke Factor of 1 = 10gmSmoke/kg of Fuel

11

On-Road Emissions

On-road emissions are not certification emissions.

Nevertheless new vehicles measured on-road usually read well below their certification values.

On-road emissions are not the same as readings from scheduled emission tests.

On average the two correlate very well

1. Second vs. Second ( a good analyzer)

2. Results vs. Result ( can screen vehicles)

3. Average vs. Average ( can characterize fleet emissions)

12

Correlation Test SetupJapan Petroleum Energy Center – February 2003

RSD

HORIBADirect analyzer

blower

Engine Cooling fan To bench analyzer

A/F sensor

to HORIBAMEXA9100EGR

Suction of blower

Mirror

AccuScan4000

Exhaust pipe Si tube

13

Correlation Test Method & ResultsMitsubishi Lancer

Several different drive modes were utilized: LA4 drive mode Steady state drive modes (60,40,20kph)

Horiba bench analyzers were compared to RSD: Comparisons between gas ratios and their

mean values. CO/CO2, HC/CO2, NO/CO2

14

On-Road Emissions

On-road emissions are not certification emissions.

Nevertheless new vehicles measured on-road usually read well below their certification values.

On-road emissions are not the same as readings from scheduled emission tests.

On average the two correlate very well Second vs. Second ( a good analyzer)

Result vs. Result ( can screen vehicles)

Average vs. Average ( can characterize fleet emissions)

15

RSD vs. I/M 240 - HC

IM240 HC g/mi vs. RSD HC ppm propane(Each point is one model year 1992-2002, N=83,737)

y = 0.0032x - 0.2559

R2 = 0.9794

0.0

0.1

0.1

0.2

0.2

0.3

0.3

0.4

80 100 120 140 160 180 200

RSD HC ppm propane

IM24

0 g

/mi

16

On-Road Emissions

On-road emissions are not certification emissions.

Nevertheless new vehicles measured on-road usually read well below their certification values.

On-road emissions are not the same as readings from scheduled emission tests.

On average the two correlate very well Second vs. Second ( a good analyzer)

Results vs. Result ( can screen vehicles)

Average vs. Average ( can characterize fleet emissions)

17

Exhaust Emissions Variability“Key: Knowing when dealing with it”

All emissions tests show low variability on new vehicles.

All emissions tests show high variability when testing older, and especially older broken, vehicles. (Bishop et al, 1996; www.feat.biochem.du.edu)

Even Federal Test Procedure (FTP) testing performed 21 times shows this high variability.

18

Gra

ms/

mil

e

AQIRP data

19

Station-Based Test Variability”Unlikely to Identify Variability”

Scheduled emission tests tend to be performed only once for vehicles which pass.

Scheduled emission tests are only repeated for vehicles which fail the first time.

Therefore, vehicle emissions variability less likely to be identified through conventional station-based testing.

20

Remote Sensing Test Variability“More Likely to Identify Variability”

Remote sensing readings can include as many as 5000 vehicles per day.

Variability of readings taken day after day on the same vehicle is readily apparent.

Sources of Variability (descending order of magnitude):

1. Vehicle (site selection, cutpoints)

2. Ambient Conditions (track & account for changes)

3. Instrument Set-up (audits, self-diagnostics)

4. Instrument Itself (building a better analyzer)

21

Gross Emitter ID Study“On-Road Screening”

22

BAR Pullover Study Method

Upstream Dual RSD Screen:

Downstream Roadside ASM:

Dual RSDs

Vehicles exceeding RSD Cutpoints

23

RSD Gross Emitter ID Performance

On-road gross emitters pulled over by a policeman have more than an 82-88% chance of failing a California emissions test based on one RSD exceedance.

>92% chance based on 2 RSD readings above cutpoints.

[“REMOTE SENSING DEVICE HIGH EMITTER IDENTIFICATION WITH CONFIRMATORY ROADSIDE INSPECTION”; California BAR Final Report 2001 – 06 August 30, 2001, copy available on www.feat.biochem.du.edu.]

December 04

24

Asia Remote Sensing Data

25

Locations & Year Measurements Mean gCO/kg

Mean gHC/kg

Mean gNO/kg

Bangkok THA, 1993 5,260 264 220 *

Hong Kong, 1993 5,891 115 20 *

Katmandu NPL, 1993 11,227 362 189 *

Kuala Lumpur MAL, 1995 9,478 209 22 31

Seoul KOR, 1993 3,104 100 14.7 *

Taipei TWN, 1993 12,062 180 23.4 *

Singapore 1995

2004

1,681 148

327

52410

Tokyo JPN, 1995 3,881 67 7.6 *

Melbourne AUS, 1992 5,260 149 6.8 24

New Delhi, India, 2004 10,800 106 61 20

Auckland NZ, 2004 34,400 89 13 10

Average USA 1989-92 34,000 113 26 *

Average USA 2003 63,000 39 3.5 4.9

Everywhere a few vehicle produce the majority of the motor vehicle emissions.

Asian on-road

emissions show large variabilitySingapore

appears comparable to the US

Asia RSD Data

26

Auto emissions by decile. New Delhi, 2004

0

1000

2000

3000

4000

5000

6000

1 2 3 4 5 6 7 8 9 10

Decile

pp

m (

CO

/10)

COHCNO

Average is dominated by a small fraction of high emitters

Majority of vehicles have negligible emissions

27

India Data Collection Summary

4 Days each in Pune and Delhi

2 Pune Sites, 3 Delhi Sites

~10,000 valid records each in Pune and Delhi

Speed/Acceleration measurements were not used to to filter measurements (i.e. Vehicle specific power).

28

Pune RSD Sites

29

Delhi RSD Sites

30

One Pune Site

31

One Delhi Site

32

CO Emissions (g/kg fuel)

CO Emissions g/kg of fuel

-

50

100

150

200

250

300

350

400

Two

wh

ee

lm

oto

rcy

cle

Two

wh

ee

lm

oto

rs

coo

ter

Th

ree

wh

ee

ler

Co

mm

erc

ial

veh

icle

s

Fo

ur

wh

eel

veh

icle

s

Bu

s

Pune

Delhi

-

50

100

150

200

250

300

Two

wh

ee

lm

oto

rcy

cle

Two

wh

ee

lm

oto

rs

coo

ter

Th

ree

wh

ee

ler

Co

mm

erc

ial

veh

icle

s

Fo

ur

wh

eel

veh

icle

s

Bu

s

Pune

Delhi

33

HC Emissions (g/kg fuel)

HC Emissions g/kg of fuel

-

50

100

150

200

250

300

Two

wh

ee

lm

oto

rcy

cle

Two

wh

ee

lm

oto

rs

co

ote

r

Th

ree

wh

ee

ler

Co

mm

erc

ial

ve

hic

les

Fo

ur

wh

ee

lv

eh

icle

s

Bu

s

Pune

Delhi

34

NO Emissions (g/kg fuel)

NO Emissions g/kg of fuel

-

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

Two

wh

eel

mo

torc

ycle

Two

wh

eel

mo

tor

sco

ote

r

Th

ree

wh

eele

r

Co

mm

erci

alve

hic

les

Fo

ur

wh

eel

veh

icle

s

Bu

s

Pune

Delhi

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

Two

wh

eel

mo

torc

ycle

Two

wh

eel

mo

tor

sco

ote

r

Th

ree

wh

eele

r

Co

mm

erci

alve

hic

les

Fo

ur

wh

eel

veh

icle

s

Bu

s

Pune

Delhi

35

Smoke Emissions (g/kg fuel)

Smoke Emissions g/kg of fuel

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

Two

wh

ee

lm

oto

rcy

cle

Two

wh

ee

lm

oto

rs

co

ote

r

Th

ree

wh

ee

ler

Co

mm

erc

ial

ve

hic

les

Fo

ur

wh

ee

lv

eh

icle

s

Bu

s

Pune

Delhi

36

Pollutant Contribution (% of Total)

0%

10%

20%

30%

40%

50%

60%

Two wheelmotorcycle

Two wheelmotor

scooter

Threewheeler

Heavy dieselvehicles

Four wheelvehicles

Per

cen

t o

f To

tal

CO

HC

NO

Smoke

Fleet

HC

Smoke

NOx

NOx

37

NO Mean Emissions by Decile

Bus

Comm

ercia

l veh

icle

Four w

heel v

ehicl

e

Three w

heeler

Two whee

l mot

orcy

cle

Two whee

l mot

or sc

ooter

12

34

56

78

910

0

1000

2000

3000

4000

5000

6000

7000

8000

NO ppm

Decile

RSD NO Mean Emissons of Each Decile

38

Smoke Mean Emissions by Decile

Two whee

l mot

or sc

ooter

Two whee

l mot

orcy

cle

Comm

ercia

l veh

icle

Three w

heeler

Four w

heel v

ehicl

e

Bus 12 3 4 5 6 7 8 9 10

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

Smoke Factor

Decile

RSD Smoke Mean Emissions of Each Decile

39

RSD Smoke - Singapore

Diesel vehicle smoke increases steadily with

age

Petrol vehicles have less smoke but smoke

also increases with age

40

RSD Diesel Smoke - Singapore

Average Smoke

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

1980

& o

lder

1981

-198

5

1986

-199

0

1991

-199

5

1996

-200

0

2001

& n

ewer

Sm

oke

HGV - D

BUSES - D

LGV - D

TAXIS - D

41

RSD Patrol Smoke - Singapore

Average Smoke

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

1980

& o

lder

1981

-198

5

1986

-199

0

1991

-199

5

1996

-200

0

2001

& n

ewer

Sm

oke

MOTORCYCLE - P

BUSES - P

LGV - P

CARS - P

42

India Data Collection Efficiency

Pune Delhi

TypeVehicles

ObservedEmissions Measured

% Emissions Measured

Vehicles Observed

Emissions Measured

% Emissions Measured

Approx. Delhi Registrations

2WM 7,126 2,095 29% 2,705 293 11% 1,600,000 2WS 4,704 1,326 28% 1,987 249 13% 1,000,000 3WV 3,695 2,065 56% 1,602 529 33% 73,772 4WD 975 759 78% 1,349 973 72% 122,625 4WV 1,689 1,558 92% 8,750 6,846 78% 1,233,064 Bus - CNG 1,565 1,318 84% 24,678 Other 1608 773 3502

19,797 8,576 21,460 10,208 4,054,139

December 04

43

RSD Motorcycle Measurement

44

2-Wheeler Measurement Challenges

1. Exhaust Volume Small samples

2. Exhaust Location Variable height and directions

3. Driving Mode Many Coasts and Decelerations

4. Dual Vehicle Passes Contaminated or Blocked

Measurement

45

RSD4000 System

Remote sensingbeam

Exhaust

IR Reference

CO2 Channel

Trigger

Acquired CO2 Channel

TRUCK MODE DATA ACQUISITION

High Pipe Low Pipe

RSD3000

46

3-mirror option 5-mirror option

Motorcycle RSD Concept

Loaded Mode Measurement Dedicated Lane Multiple Beam Heights

2.5m

47

Concluding Remarks…

REMOTE SENSING – RSD4000 introduces UV smoke channel. RSDs offer the advantage of measuring on-road emissions. Emissions variability recognized in multiple RSD measurements. RSDs proven & widely used for screening and characterization.

INDIA REMOTE SENSING DATA - CO uniform across types and proportional to fleet #s. HC higher in rich burning 2/3-wheelers.

Lower in Delhi 3-Ws (credit CNG program) Low in Delhi’s CNG buses.

NO highest among lean burning Comm. Diesels and Petrol 4-Ws Highest rates from CNG buses, but should assess contribution.

Smoke highest in rich burners (2/3-Ws), lowest in lean burning high NOx emitters. Lower in Delhi’s 3-Ws and CNG Buses Rises with age and higher in Diesels.

MOTORCYCLE MEASUREMENT - 2-W capture efficiency must be improved.

Loaded mode measurement in a dedicated lane with more versatile beam deployment.

48

Web Sites

University of Denver: www.feat.biochem.du.edu DU reports,

publications and downloadable data. www.sign.du.edu SMART SIGN 24/7 RSD

operation and live web camera.

Environmental Systems Products www.esp-global.com – general ESP Product Info www.rsdaccuscan.com – history, legislation,

reports, documents, technology, products