Ceci est une version HTML d'une pièce jointe de la demande d'accès à l'information 'Additional test vehicle identities'.

EMISSIONS OF A LIGHT DUTY 
MULTIFUEL VEHICLE UNDER URBAN 
DRIVING CONDITIONS
Provenza Alessio,
Bonnel Pierre, Carriero Massimo, Perujo Adolfo, Weiss Martin
European Commission, Directorate General Joint Research Centre (JRC), Institute for Energy and 
Transport, Sustainable Transport Unit
Via Enrico Fermi, 2749 - 21027 Ispra (VA) - Italy

In collaboration with
Dondi Davide, De Santoli Livio, Fraticelli Fabio
CITERA Research Center, Rome, Italy
http://www.jrc.ec.europa.eu/
CRC Real-World Emissions Workshop – San Diego – March 25-28 2012 

General background: legislation
• Directive 2009/30/EC, (fuel quality) target: reduce GHG 
emissions per unit of energy at least by 6% until 2020.
• Regulation 79/2009 target: ensure the proper functioning 
of hydrogen-powered motor vehicles by specifying 
harmonized safety requirements.
• No existing specific methods to type-approve vehicles 
fueled by H2-CNG blends.

General background
Assessment of transport technologies using real-world 
emissions measurements
CO2 regulations of light (EC/443/2009) and heavy-duty 
vehicles (on-going development of EU HDV CO2 testing 
procedures: Model based, real-world PEMS measurements 
could be used as validation test methods)
Objective of the present research: development of data 
evaluation methods for real-world CO2 emissions, to 
establish  links with vehicle/engine characteristics and/or 
operating conditions (speed, road grade, etc...)

In-use emissions testing
Recent Legislative Developments:
Publication of the PEMS based In-Service Conformity (ISC) 
provisions for the future EURO VI standards, (also 
applicable to EURO V engines)
European PEMS Pilot Program for Non Road Mobile 
Machinery (NRMM) engines
PEMS candidate method to check and to limit the Real 
Driving Emissions (RDE) of Light Duty Vehicles from Euro 6 
standards onwards (2014)

Objectives of the study
To develop methods to make use of the in-use emissions 
data collected with PEMS.
To study  the exhaust emissions as function of the CNG-
hydrogen in real-world driving conditions.
To relate on-road data with test cell data
To asses the greenhouse gasses emission reduction 
obtained by means of hydrogen blends.



Data used for the study
Real-world measurements from a EURO 4 light-duty vehicle
5 different fuels used: Gasoline, CNG, CNG+10%H2, 

CNG+20%H2, CNG+30%H2.
Vehicle tested on an urban route 17Km long (two tests for 

each fuel)


Vehicle specifications
BRAND
FIAT
MODEL
Panda
YEAR
2007
DISPLACEMENT
1242 cm3
Gasoline
FUEL
CNG
Hydrogen-CNG Blends
Max Power
44/38 kW
(gasoline/CNG)
After treatment
3-way catalyst



PEMS installation



PEMS installation

Results: Trip averaged emissions
300
0.25
250
0.20
200
) 0.15
)
Gasoline
Gasoline
m
m
K
K
/
CNG
CNG
150
/
(g
CN(g
CNG+10%H2
x G+10%H2
2
CO
CNNOG+20%H2
0.10
CNG+20%H2
CNG+30%H2
CNG+30%H2
100
0.05
50
0
0.00
Test 1
Test 2
Test 1
Test 2

Trip averaged emissions
Trip averaged emissions strongly influenced by traffic conditions 
Difficult to have direct comparison with standard test cycle due to: 
Road grade, idling, different average vehicle speed.

30
1800
1600
25
1400
)
1200
h 20
/
)
Ga(ssoline
Gasoline
 (Km
  1000
d
e

CNG
CNG
15
CNidlingG+10%H2
CNG+10%H2
 spe

e
a 800
g
CNtG+20%H2
CNG+20%H2
a
r

To
e
CNG+30%H2
CNG+20%H2
Av 10
600
400
5
200
0
0
Test 1
Test 2
Test 1
Test 2

Averaging window approach
Moving averaging window – Distance based (4 and 11 km, 
using a time increment equal to the data sampling frequency)
Data binning according to the parameters governing the vehicle 
dynamics and therefore the fuel consumption and the 
emissions
Average speed
Average road grade
Average relative positive acceleration (RPA)
Exception: Aerodynamic effects assumed to remain constant within a speed 

range (effect of front wind had to be neglected)
Statistics conducted in the different bins
Comparison with ECE (urban part of NEDC)



Results
• Averaging reference distance = NEDC length, 11.007 km.
• Average road grade between -0.25% and 0.25%.
• Distribution of CO2 emissions (g/km) as function of  the average 
window speed. 
Test 1
Test 2

Results
• Averaging reference distance = ECE 15 length, 4.052 km.
• Average road grade between -0.25% and 0.25%.
• Distribution of CO2 emissions (g/km) as function of  the average 
window speed.
300
300
250
250
200
)
200
)
m
Ga
m soline
Gasoline
/K
/K
(g
CNG
(g
CNG
2
2
CNG+10%H2
CNG+10%H2
CO
CO
150
CNG+2
1500%H2
CNG+20%H2
CNG+30%H2
CNG+30%H2
100
100
Test 1
Test 2
50
50
0
10
20
30
40
50
0
10
20
30
40
50
60
70
Average Window Speed (Km/h)
Window average speed (Km/h)

Results

Averaging reference distance = ECE 15 length, 4.052 km.

Average road grade between -0.25% and 0.25%, max idle 33%.

Distribution of CO2 emissions (g/km) as function of  the window 
relative positive acceleration in the speed bin 15-25 Km/h.
300
R² = 0.7872
250
R² = 0.8745
R² = 0.9746
200
)
m

R² = 0.9782
Gasoline
/K
(g

CNG
2
R² = 0.781
CNG+10%H2
CO 150
CNG+20%H2
CNG+30%H2
100
50
0
5
10
15
20
25
Window  Relative Positive Acceleration (m/s2)

Results

Averaging reference distance = ECE 15 length, 4.052 km.

Average road grade between -0.25% and 0.25%.

Speed bin 15-25 Km/h.

Comparison between dyno emissions and on-road emissions
300
250
200
)
/Km
ECE
150
(g 2
TEST 1
CO
TEST 2
100
50
0
Gasoline
CNG
CNG+10%H2
CNG+20%H2
CNG+30%H2

Results

Averaging reference distance = ECE 15 length, 4.052 km.

Average road grade between -0.25% and 0.25%.

Speed bin 15-25 Km/h.

Comparison between dyno emissions and on-road emissions
0.3
0.25
0.2
)
m
/K

ECE
0.15
(g x
TEST 1
NO
TEST 2
0.1
0.05
0
Gasoline
CNG
CNG+10%H2
CNG+20%H2
CNG+30%H2

Conclusions: methodology
Indicators proposed for a systematic data binning method
RPA was found to be a good indicator to explain the 
variability in the CO2 emissions at constant road grade and 
speed
The averaging window method provides an efficient way to 
link emissions and average operating characteristics

Conclusions: results
Use of hydrogen blends effectively reduce on-road CO2
emissions
NOx emission generally increase as the hydrogen content in 

the blend increases
Largest spread of CO2 results in bins corresponding to the 

largest RPA spread.


Special thanks to:
S. Alessandrini and F. Forni, 

PEMS technicians
A. Ceci, Driver