Using mobile measurements to update on-road transportation emissions inventories

Shaojun Zhang, Assistant ProfessorSchool of Environment, Tsinghua University

2019 International Emissions Inventory ConferenceDallas, TexasJul 31, 2019

2019 International Emissions Inventory Conference

Shaojun Zhang, Hui Wang, Xinyu Liang and Ye Wu School of Environment, Tsinghua University

K. Max Zhang Sibley School of Mechanical and Aerospace Engineering, Cornell University

Parikshit DeshmukhJacobs Technology Inc.

Richard Baldauf, James Faircloth, and Richard Snow U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

We acknowledge the support from Ministry of Science and Technology (MOST) of China’s international collaborative project with U.S. EPA.

Authors and Acknowledgements

Outline

BackgroundMethods

Results and discussion

Summary

3

Official source apportionment of ambient PM2.5 pollution

Other relevant concerns : high fraction of nitrate aerosol (e.g., over 30% in Beijing), summertime O3 pollution, and exceedance of NO2 concentrations in urban areas

Vehicle emissions and air pollution problems in China

45%

Beijing (81 µg/m3, 2015) Beijing (58 µg/m3, 2017)

29%

Shanghai (53 μg/m3)

Vehicle/Mobile sources Coal combustionDust IndustrialResidential, argicultural and others

4

Data source: Ministry of Ecological Environment of China; Note: Cross-boundary transport not included

HDDV emissions control in China

Controlling air pollutant emissions from HDDVs are one of the prioritized tasks in the new “Clean Air Action Plan (2018-2010)”

Generally, the European regulatory systems have been used in China China III and earlier stages: typically no modern aftertreatments China IV and China V (implemented nationwide since 2015):

SCR as mainstream aftertreatment for NOX control (2.0 g/kWh for China V) Almost no DPF adoption

China VI: recently implemented since Jul 2019 in a few cities/regions PEMS regulations: conformity factors of 1.5 for NOX and 2.0 for PN (China VIb) OBD requirements: capability of remote on-board sensor-based monitoring

5

Technology innovation to enhance in-use surveillance

PEMS1~2 vehs per

day

Dynamometer1~2 vehs per

day

Regulated methods Real-world efficient methodsNon-intrusive

Plume chasing50~80 vehs per day2~5 mins per veh

Remote sensing1000+ vehs per dayShort snap per veh

High accuracySmall sample size

Low accuracyLarge sample size

Applicable to key fleetsNOx sensor required

Remote OBD monitoring

OBM data loggerData collection and

transmission (to cloud)

Accuracy

Fleet coverage6

Joint workshop at Tsinghua University, Beijing

Technical communication at EPA ORD NRMRL in RTP, NC

• Sponsored by Ministry of Science and Technology of China (MOST)-USEPA collaborative agreement on technology innovation for environmental protection

• A focus on developing cost-effective road emission measurement methods • Collaborative project on motor vehicle plume chasing and joint research workshops

Tsinghua-NRMRL collaborative project

7

Outline

BackgroundMethods

Results and discussion

Summary

8

Research framework

• Gridded emission input

• WRF/2D-VBS CMAQ

• Impact from updated EFs on ambient NO2, O3 and PM2.5

• Updating key EFs for HDVs

• Updating fleet-level emission inventories

3. Emission inventory

update2. Plume chasing

1. Method evaluation

4. Air quality simulations

• Concurrent PEMS-chasing tests

• Optimize EF calculation methods

• Large-scale field tests

• EF database based on plume chasing

9

Mobile chasing platform

10

Deep cycle battery set

PM2.5

CO2

NOX /NO

Sample inlet

BC

Camcorder

GPS

PN1

Computer

10

Licor 820

Eco physicsnCLD66

Magee AE33(880 nm)

TSI DustTrak 8530

TSI CPC3007

Pollutant analyzers (Resolution: 1Hz)

Concurrent PEMS-chasing tests

Experimental routes in Beijing.

• 12 heavy-duty diesel trucks recruited for concurrent PEMS-chasing tests to evaluate plume

chasing for NOX emission factors.

11

Local roads (20~40 kph)

Express roads (50~80 kph)

PEMS: SEMTECH-EcoStar

Test fleet: 4 China III (no SCR), 4 China IV (SCR) and 4 China V (SCR)

Testing procedure and EF calculations• Measure road background and exhaust plumes intended for targeted vehicles

12

• NOX EF (g NOx per kg fuel) calculations• Approach A: linear regression without BG subtraction (slope indicates fuel based NOx emissions)

• Approach B: moving average ratio with BG subtraction

Mobile chasing for HDV emissions Nearly 3000 vehicles measured by using mobile chasing since 2017

JJJ region: more than 1200 vehicles

PRD region: more than 1000 vehicles

Sichuan: more than 200 vehicles

Fen-Wei Plain & Central China: ~500 vehicles

13

Vehicle emission models in China

International models (e.g., MOBILE/MOVES; COPERT) have been used to estimate vehicle emissions in China since 1990s.

Local emission measurements are fundamental to develop emission models and reduce uncertainty in policy decisions.

− EMBEV (Tsinghua): initially sponsored by the Beijing EPB, and based on thousands of LDV dynamometer tests and hundreds of LDV/HDV PEMS tests

− China’s National Emission Inventory Guidebook (Tsinghua and VECC): developed for the MEP in 2015, largely based on the EMBEV

− EMBEV-Link: city-scale link-level emission inventory models based on real-world traffic datasets

The Guidebook for road sector will be updated soon, as supported by China MOST

14

EMBEV: Zhang et al., Atmos. Environ., 2014; EMBEV-China: Wu et al., Sci. Total Environ., 2017; EMBEV-Link: Yang et al., Atmos. Chem. Phys., 2019

EMBEV emission factors: heavy-duty diesel vehicles

Major data sources: PEMS tests for pre-China IV HDVs (and China IV urban buses) Mode-specific emission rates available (i.e., similar to MOVES) Few tests for China IV and China V trucks, and their emission factors are largely estimated

based on the reduction trends in European fleets (i.e., 30% ↓ from China III to China IV)

Estimated emission factors (GVW>12 t) under a typical driving cycle (40 km h-1)

1.46

0.55 0.440.22 0.12 0.02

0.0

0.5

1.0

1.5

2.0

Emis

sion

fact

or (g

/km

)

PM

3.31

0.730.42 0.21 0.10 0.10

0

2

4

Emis

sion

fact

or (g

/km

)

HC15.21

8.446.71 7.19

5.04 4.28

0

4

8

12

16

Emis

sion

fact

or (g

/km

)

NOx

Wu et al., 2017, STOTEN 15

Gridded emission input

Monitor siteRoad

NOXemission

16

• Real-world traffic datasets aided allocation• Total fleet emissions by vehicle category• Temporal and spatial allocations, based on specific road lengths (considered the effects from vehicle

volumes, urban/rural, city, speed)

~900 traffic sites (mostly express roads)

Air quality simulation systems

WRFv3.3 configuration− Planetary boundary layer: Mellor-Yamada-Janjic TKE− Land surface layer: Noah− Microphysics: WSM3− Cumulus scheme: G-D− Long/short wave radiation: RRTM

CMAQv5.0.1 enhanced by 2D-volatility basis set (VBS) module

− Improve the simulation performance of SOA− Gas-phase chemistry scheme: SAPRC99− Aerosol module: AERO6

• WRF-CMAQ model system− Three-layer nested simulation (36km-12km-4km)− Focus region: Beijing-Tianjin-Hebei (JJJ)− Duration: January, May, August and November in 2020

17

36 km×36 km

4 km×4 km

Outline

BackgroundMethods

Results and discussion

Summary

18

Comparison of NOX emission factors between mobile chasing and PEMS

Vehicle-specific relative errors within ±20%

Fleet-average errors: -3.8% (local roads by approach A) and -0.3% (freeways by approach B)

Note: The solid markers indicate measurements on local roads, while the hollow markers indicate measurements on freeways. 19

0 20 40 60 80 100 120 1400

20

40

60

80

100

120

140

V9V10V11V12

V5V6 V7V8

V2 V3 V4

Cha

sing

EFs

(g N

OX /

kg-fu

el)

PEMS EFs (g NOX /kg-fuel)

China Ⅲ China Ⅳ China Ⅴ V1

Local roads_Approach AFreeways_Approach B

0 20 40 60 80 100 120 1400

20

40

60

80

100

120

140

China III China IV China V V1 V9

V10V11V12

V5V6V7V8

V2 V3 V4

Cha

sing

EFs

(g N

OX /k

g-fu

el)

PEMS EFs (g NOX /kg-fuel)

Local roads_Approach AFreeways_Approach A

a.

0 20 40 60 80 100 120 1400

20

40

60

80

100

120

140

V9V10V11V12

V5V6 V7V8

V2 V3 V4

Cha

sing

EFs

(g N

OX /

kg-fu

el)

PEMS EFs (g NOX /kg-fuel)

China III China IV China V

b.

V1

Local roads_Approach BFreeways_Approach B

Concurrent PEMS-Chasing comparison(12 HDVs, 245 chasing tests)

0

20

40

60

80

100

120

140

China VChina IV

Local roads Freeways

NO

x EF

s(g/

kg_f

uel)

Veh. IDV1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12

China III

NOX emission factors on local roadsand freeways for PEMS results

Mobile chasing for HDV emissions——NOX

NOX emissions: No significant improvement in China IV diesel trucks (claimed to have adopted SCR) compared with China III trucks; limited improvements for China V diesel trucks.

− Failure to refill urea tanks and tampering with the SCR device are suspected NG HDVs: Significantly higher NOX emissions (lean-burn engines, no NOx aftertreatment) HDDVs registered in various provinces: Beijing not significantly improved compared to other provinces

China III China IV China V NG0

40

80

120

160

200

Diesel

25%-75% 5%-95% Median

NO

X E

Fs

（ g/kg_fuel）

Mean

GD BJ LN SD SX HN SHX SC HB TJ Others0

20

40

60

80

100

120

NO

X E

Fs

（ g/kg_fuel）

25%-75% 5%-95% Median Mean

Note: GD:Guangdong; BJ:Beijing; LN:Liaoning; SD:Shandong; SX:Shanxi; HN:Henan; ; SHX:Shaanxi; SC:Sichaun; HB:Hebei；TJ: Tianjin

Impact of emission standard and fuel Impact of registration place——HDDV

20

BC emissions: decrease with emission standards; high emitters exist for China III to China V

HDDV in various provinces: related to different vehicle emission control policies

Mobile chasing for HDV emissions——black carbon (BC)

China III China IV China V NG0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0 25%-75% 5%-95% Median

BC

EFs

（ g/kg_fuel）

Mean

Diesel

GD BJ LN SD SX HN SHX SC HB TJ Others0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5 25%-75% 5%-95% Median

BC

EFs

（ g/kg_fuel）

Mean

Note: GD:Guangdong; BJ:Beijing; LN:Liaoning; SD:Shandong; SX:Shanxi; HN:Henan; ; SHX:Shaanxi; SC:Sichaun; HB:Hebei；TJ: Tianjin

Impact of emission standard and fuel Impact of registration place——HDDV

21

A revisit of NOx and PM2.5 emissions for HDVs

22

0

2

4

6

8

10

2010 2012 2014 2016 2018 2020

Emis

sion

s (m

illion

tons

)

NOX

by EMBEVby Chasing results

Total emission trends

0.0

0.5

1.0

1.5

2.0

Emis

sion

fact

or (g

/km

)

PM2.5

EMBEV updated by Chasing results

0

5

10

15

20

Emis

sion

fact

or (g

/km

)

NOX

Emission factor

0

100

200

300

400

500

2010 2012 2014 2016 2018 2020Emis

sion

s (th

ousa

nd to

ns)

PM2.5

by EMBEVby Chasing results

* Average GVW: 20 t, Driving speed: 40 km/h, BC/PM2.5=0.7

High emitter20%

10%

8%

---- Medium ---- Mean ---- EMBEV

NOX PM2.5

↑ baseline

↓ updated

Air quality impacts: nitrogen dioxide

Monthly average NO2 concentrations are estimated to increase 2-4 μg m-3 (12%-15%) in Beijing-Tianjin-Hebei region when the simulations apply the chasing NOX emission factors.

2020 baseline NO2 concentration

Jan May

Aug Nov

NO2 concentration change

Jan May

Aug Nov

23

Air quality impacts: ozone

Summer: O3 concentrations increased due to increased NOX emissions Winter: O3 concentrations decreased by ~2 ppbv due to strong VOC-limited conditions.

2020 baseline 8-h maximum O3 concentration

Jan May

Aug Nov

O3 concentration change

Jan May

Aug Nov

24

Air quality impacts: PM2.5 and key components

Summer: PM2.5 is predicted to increase 1.2-2.3 μg m-3 (4%-5%) due to increased nitrate. Winter: decreased atmospheric oxidability (i.e., O3) leads to PM2.5 reduction from baseline

2020 baseline PM2.5 concentration

Jan May

Aug Nov

PM2.5 concentration change

-2.0 -1.0 0.0 1.0 2.0

-1.0 1.0 3.0

Jan May

Aug Nov

Jan

Aug

ECPOMSOMNH4NO3SO4PM2.5

ECPOMSOMNH4NO3SO4PM2.5

Changes by aerosol component

25

Summary

On-road evaluations were conducted for plume chasing tests by comparing with PEMS measurements.

Chasing tests indicate that NOx emissions have not been significantly improved with the emission standard for HDVs in China.

BC emissions for HDVs have been reduced with the emission standard; however, high emitters were still found among the fleet.

With updated NOx emission factors, which will not decline before 2020, we estimated increased ambient concentrations for annual NO2 and PM2.5 , and summer O3 in the Jing-Jin-Ji region.

26

Thank you for your attention !Contact. E. zhsjun@Tsinghua.edu.cn

Acknowledgements to ongoing and former project sponsors: Ministry of Science and Technology (MOST) of China's International Science and Technology Cooperation Program (2016YFE0106300; i.e., the MOST-EPA collaboration program), and the National Key Research and Development Program of China (2017YFC0212100).

Disclaimer: The views expressed in this presentation are those of the author and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency