India Telephone Seminar - ESP Global

1

Environmental Systems Products Holdings Inc.

Presented to ARAI on March 18th/19th, 2004Dr. Donald Stedman, Niranjan Vescio, Gary Full

2

Agenda

Instrumentation – Gary FullRSD Gas Calculations – Dr. StedmanSummary Statistics – Dr. Stedman2-Wheeler Improvement – Gary FullOther Discussion - All

3

RSD4000 Technology Device Measurement Overview

Presented by Gary Full

4

Optical Path

RCVR

SRC

SDMTMM

Direction of vehicle travel

Outgoing beam

Return beam

5

Optical Beam

IR filament source and UV light source are combined and positioned so as to appear at the focus of the source mirror. Result is a collimated (nearly parallel rays) exit light beam.Transfer mirror offsets the beam and returns it to the receiver optical section of SDM.Receiver mirror collects and focuses the received light.UV light is split out via a dichroic mirror and distributed to aspectrometer via fiber optic cable for NO and SF measurements.IR light is sent to a rotating (12,000 RPM) spinning mirror scanner assembly which distributes light to individual IR detectors for CO, CO2, REF, and HC detection. Scanner produced 2400 light pulses/second at each detector.Thermoelectrically cooled detectors (~-5 degree C) produce stable, highly sensitive, low noise electrical signal outputs.

6

Optical Beam Path

7

Signal Conditioning/Firmware

Each detector output is AC-coupled via a preamplifier.Discrete gain selection (0.5, 1, 2, 4, or 8).Q=10 band-pass filter produces sine waves at carrier frequency of 2400 Hz. Gas information is represented by the amplitude modulation of the carrier.Sine waves are then rectified.A synchronous integrator provides an average over 24 sine wave periods to produce a result that is sampled by a 16-bit A/D converter every 10-milliseconds (100 samples/second).IR gas channel voltages (CO,CO2, and HC) are ratioed to the REF channel to provide a signal proportional to gas transmittance (ratio of emitted light to received light).

8

Gas Measurement Processing

Curves of gas transmittance versus gas amount (number of molecules/unit beam cross-section) are stored in the SDM.Measured transmittances are converted to gas amounts.Gas amount measurements (ambient gas readings) at front of the vehicle are recorded for CO, CO2, and HC.50-samples of information are recorded at the rear of the vehicle.Amount of gas attributable to the vehicle is found by subtracting the ambient value from each of the 50 samples.

9

0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.8550

0

50

100

150

200

250

300

350P2 CO2 (%-cm), std-day

Normalized ratio (CO2/REF)

Am

ount

of g

as %

-cm

350

50−

estCO2plot

.9.2 ratioCO2plot kr CO2⋅

3b-1

CO2 Response Curve

10

Plume Decay Graphs

11

Gas Measurement Processing continued

Aggregate gas ratios (a single value for the possible 50 point sets) are calculated for CO/CO2, HC/CO2, and NO/CO2. The aggregate values are least squares linear regression fits of thegas value array with respect to the CO2 array. E.g. CO array versus CO2 array. For the aggregate slope calculations we reject data points immediately before and after beam blocks/unblocks, and where the sum of the gas amounts are below a threshold.We consider a gas measurement “valid” if there are at least 5-points available in the slope calculation – “invalid” otherwise.

12

Source(s)

Detectors/filters

CO2

CO

HC

NOx

REF

Using calibration curves(results of coefficienting)determine the numberof molecules in the pathfor each of the gas species.

nCO2

nCO

nHC

nNOx

Determine gas ratios

nCO/nCO2

nHC/nCO2

nNOx/nCO2

THE BASIC INSTRUMENT

Software Software Gas Ratios

Results Reporting

nCO/nCO2

nHC/nCO2

nNOx/nCO2

%CO

%CO2

%HC

%NOx

COMBUSTION EQUATION

%CO2 = 1/ [ C1 * nCO / nCO2 + C2 +C3 * nHC / nCO2 + C4 * nNOx / nCO2]

%CO = %CO2 * nCO / nCO2

%HC = %CO2 * nHC / nCO2

%NOx = %CO2 * nNOx / nCO2

ASSUMPTIONS that impact determination of C1, C2, C3, and C41. The measured gases are the results of combustion and only combustion gases.2. The fuel hydrogen to carbon ratio is known.3. The oxidizer is air.

13

Gas Measurement Processing continued

We also report “stoichiometric” (meaning the concentration that would be measured by a tail pipe probe if there were no excess air in the exhaust beyond what is needed for combustion) gas concentrations. These gas concentrations are the result of combustion chemistry where the basic assumptions are that the aggregate carbon to hydrogen content of the fuel is known and that air is the oxidizer in the combustion process.

14

Other RSD Subsystems/Options

Speed/accelerationVehicle/license plate pictureWeather measurementGPS for time base as well as positionWireless communications links- 300 meters- Cellular links

PC storage and transfer media“Smart” Signs / motorist displays

15

RSD Calculations

Presented by Dr. Donald Stedman

16

On-road Results from Denver

Sept 2003

Slides prepared by Donald H. Stedman, March 18, 2004 for ESPwww.feat.biochem.du.edu

17

IR Signal versus time

0

1

2

3

4

5

6

7

8

9

0 100 200 300 400 500 600

Time ms

IR s

igna

l RefCOCO2HC

IR Signal vs Time

18

Gas readings versus time

-1

0

1

2

3

4

5

6

7

8

0 100 200 300 400 500 600

Time ms

% in

8 c

m.

COCO2HC

Gas Readings vs Time

19

CO versus CO2

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

0.1

0 1 2 3 4 5 6 7 8

CO2

CO

CO vs CO2

20

HC versus CO2

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

0.1

0 1 2 3 4 5 6 7 8

% HC in 8 cm

%C

O2

in 8

cm

HC

HC vs CO2

21

Figure 2b: IR Signal versus time

0

1

2

3

4

5

6

7

8

0 50 100 150 200 250 300 350 400 450 500Time ms

Sig

nal V

RefCOCO2HC

IR Signal vs Time

22

Figure 3b: Gas readings versus time

-1

0

1

2

3

4

5

6

7

0 50 100 150 200 250 300 350 400 450 500

Time ms

% in

8 c

m.

COCO2HC

Gas Readings vs Time

23

Figure 4b: CO versus CO2

y = 0.7999x - 3.7888R2 = 0.9976

0

0.2

0.4

0.6

0.8

1

0 1 2 3 4 5 6 7

CO2

CO

CO vs CO2

24

Proof that the on-road readings are correct

Excellent Correlation by MY to IM240, published and repeated in Chicago and Phoenix and other years.

25

Denver 1999 CO

y = 0.6785x + 14.003R2 = 0.9736

0

50

100

150

200

250

300

0 50 100 150 200 250 300

Average CO from IM (g/kg fuel)

Ave

rage

CO

from

RSD

(g/k

g fu

el)

82

88

95

26

Denver 1999 HC

y = 0.4108x + 0.8305R2 = 0.9783

02468

101214161820

0 5 10 15 20

Average HC from IM (g/kg fuel)

Ave

rage

HC

from

RSD

(g/k

g fu

el)

82

8895

27

Denver 1999 NO

y = 1.0848x + 1.7888R2 = 0.9783

02468

101214161820

0 5 10 15 20

Average NOx from IM (g/kg fuel)

Ave

rage

NO

from

RSD

(g/k

g fu

el) 82

88

96

28

Proof from Roadside Pullovers

Using a single remote sensor to pull over on-road high emitters, BAR found 83%-88% ASM failure rates on CO or HC or NO depending on the remote sensing cut point chosen.

At a 10,000 vehicle per day site, a very high cut point (top 1%)will fail 100 vehicles per day!

California Bureau of Automotive Repair, “Remote Sensing Device high emitter identification with confirmatory roadside inspection” Final Report 2001-06.

29

A few broken vehicles cause most of the emissions.

RSD is ideal for emissions inventory and to identify broken vehicles

30

Repeatability

Low emitters with good air/fuel ratio control have very repeatable emissions regardless of the test method.High emitters with poor air/fuel ratio control have very erratic emissions regardless of the test method."Motor Vehicle Emissions Variability", G.A. Bishop, D.H. Stedman, L. Ashbaugh, J. Air Waste Manage. Assoc., 46:667-675, 1996.

31

Gra

ms/

mile

AQIRP data21 FTP tests on each vehicle.

32

Hit rates

Exact validity criteria for University of Denver data see CRC reports on our web site www.feat.biochem.du.eduHit-rate depends upon the exhaust plume magnitude (vehicle load, speed and engine size and measurement site) and software conservatism. With good sites we routinely observe a hit rate >96% on automobiles.With the folded optical beam system we observe >95% hit rate for 95 cc motorcycles.

33

Data analysis from India

Presented by Dr. Donald Stedman

34

Gas Flag

18189103867803

9.29%92%168913115584WV

5.36%78%9752167594WD

20.31%56%3695163020653WV

25.86%28%4704337813262WS

39.18%29%7126503120952WM

Fleet%V %TotalxVCType

35

CO Contribution by Vehicle Type/km

0%5%

10%15%20%25%30%35%40%45%

Two whee

l motorcycle

Two wheel m

otor scooter

Three wheeler

Heavy diesel vehicles

Four wheel v

ehiclesPe

rcen

tage

of T

otal

CO

COfleet%

CO

36

HC Contribution by Vehicle Type/km

0%5%

10%15%20%25%30%35%40%45%

Two wheel m

otorcycle

Two wheel m

otor scooter

Three wheeler

Heavy diesel vehicles

Four wheel v

ehiclesPe

rcen

tage

of T

otal

HC

HCfleet %

HC

37

NOx Contribution by Vehicle Type/km

0%10%20%30%40%50%60%

Two wheel m

otorcycle

Two wheel m

otor scooter

Three wheeler

Heavy diesel vehicles

Four wheel v

ehiclesPe

rcen

tage

of T

otal

NO

x

NOFleet%

NOx

38

Smoke Contribution by Vehicle Type/km

0%5%

10%15%20%25%30%35%40%45%50%

Two wheel m

otorcycle

Two wheel m

otor scooter

Three wheeler

Heavy diesel vehicles

Four wheel v

ehiclesPerc

enta

ge o

f Tot

al S

mok

e

Smoke (<25)Fleet%

Smoke

39

CO Emissions/kg of fuel

050

100150200250300350400

Two whe

el moto

rcycle

Two whe

el moto

r sco

oter

Three

whe

eler

Heavy d

iesel

vehic

les

Four w

heel v

ehicl

es

CO Emissions

40

HC Emissions/kg of fuel

050

100150200250300

Two whe

el moto

rcycle

Two whe

el moto

r sco

oter

Three w

heele

r

Heavy d

iesel

vehic

les

Four w

heel v

ehicl

es

HC Emissions

41

NO emissions/kg of fuel

6.59

0.16 0.51

19.52

11.68

0.00

5.00

10.00

15.00

20.00

25.00

Two whe

el moto

rcycle

Two whe

el moto

r sco

oter

Three w

heele

r

Heavy d

iesel

vehic

lesFou

r whe

el veh

icles

NO Emissions

42

Smoke gm/kg of fuel (assumes all black)

0.005.00

10.0015.0020.0025.0030.0035.00

Two whe

el moto

rcycle

Two whe

el moto

r sco

oter

Three w

heele

r

Heavy d

iesel

vehic

lesFou

r whe

el veh

icles

Smoke

43

Improvements Necessary for Improved 2-wheeler measurement

Presented by Gary Full

44

1. Traffic Diversion

2-wheeler and 3-wheeler traffic must be diverted to a single file lane in which measurement equipment can be deployed.

45

2. Equipment Modification

Present technology equipment must be operable for measurements for path lengths of 3 to 5 meters.Present equipment is calibrated for measurement path lengths of 5 to 10 meters.

46

3. Loading Mechanism

Suggest that a portable ramp be deployed at the measurement station. 0.3 meter rise over a 3-meter length. Similar ramp down.

47

4. Multiple Analyzers

Two analyzers may be required deployed at two different heights to assure plume intersection with measurement beams.

48

Other Discussion

CutpointsCorrelations