Influence of Fuel Ethanol Content on PN and PM from Direct Injection Gasoline Engines
Dr. Piotr Bielaczyc, Joseph Woodburn, Dr. Andrzej Szczotka
BOSMAL Automotive Research and Development Institute Ltd
Concern over particulate matter emissions from automotive engines has been expressed for decades [1]. This has lead to the introduction of both mass and number-based emissions limits for Diesel vehicles in the European Union; direct-injection petrol engines have now also had a particle number limit set, and it is likely that all automotive internal combustion engines will be subject to particle mass and number limits in the future. These limits have necessitated (or will necessitate) the implementation of measures aimed at reducing particle emissions, through changes to the calibration, fuel injectors, etc; and by implementing filtration devices (Diesel particulate filters, Gasoline particulate filters). The direct injection concept exhibits better fuel consumption and this engine type will eventually come to dominate park-ignited automotive engines. Ambitious targets for the implementation of biofuels in the European Union are slowly changing the composition of both petrol and Diesel fuels used in the EU, with impacts on particle emissions. Thus, investigations into the effect of ethanol blends on direct injection engines are of great importance for the future. A series of experiments was performed to determine the impact of fuel ethanol content and ambient temperature on particle number and mass emissions from a European gasoline vehicle featuring a direct injection engine. Testing was performed according to the EU legislative test procedure, over the New European Driving Cycle, using a chassis dynamometer, foil-backed TX40 filters for quantification of particle mass and a PMP-compliant particle counting system for quantification of particle number emissions. Tests were performed using ethanol blends E5 (i.e. 5% ethanol by volume), E10, E25 and E50 at both temperatures currently specified for SI vehicles in EU legislation: +24 °C and -7 °C. Figures 1a and 1b present particle mean number and mass emissions results obtained for the various petrol-ethanol blends.
0.0
2.5
5.0
7.5
10.0
E5 E10 E25 E50
PM
em
issi
on
[mg
/km
]
Fuel blend
PM, 24 deg C
PM, -7 deg C
Nodata
1E+10
1E+11
1E+12
1E+13
E5 E10 E25 E50
PN
em
issi
on
[#/k
m]
Fuel blend
PN, 24 deg C
PN, -7 deg C
Nodata
Figure 1 – a) particle mass (PM) emissions obtained over the NEDC at +24 °C and -7 °C for various ethanol blends
Figure 1 –b) particle number (PN) emissions obtained over the NEDC at +24 °C and -7 °C for various ethanol blends
Usage of an oxygenated fuel is generally thought to improve the combustion process and reduce particulate emissions. However, despite the increased oxygen content of the fuel, higher ethanol blends showed decreases in particle number and mass emissions which ranged from small to non-existent. The E10 and E50 blends performed well at the lower test temperature regarding their mass emissions, but showed no real advantage in terms of number emissions. Regarding comparison to the legislative limits, which are mandated for a 20-30 °C ambient temperature, the mass limit was fulfilled for all fuel blends; the phase-in particle number
limit was just met for all blends, but the long-term limit of 6.00E11 #/km was exceeded for every fuel blend.
In line with previous examinations (e.g. [2,3]), the semi-log space created by co-plotting particle mass emissions (linear scale) and particle number emissions (log10 scale) was employed. While the two metrics are not directly comparable, as they measure somewhat different aspects, graphical and numerical explorations of the relationship between particle mass and number can be informative for a variety of reasons, both theoretical and practical [1]. Figure 2 presents particle number emissions plotted as a function of particle mass emissions, alongside the future legislative limits.
y = 9E+11x + 4E+11R² = 0.9204
1.00E+11
1.00E+12
1.00E+13
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
PN
em
issi
on
[mg
/km
]
PM emission [mg/km]
SIDI limit SIDI limit - phase-in Data Trend (ethanol blends)
Figure 2 – a semi-log plot of particle number and mass results from one direct injection vehicle tested over the NEDC, with ethanol blends E5 to E50. Raw results (from both phases of the cycle) are presented in comparison to
the EU legislative limits
The trendline’s y intercept value was found to of a similar order of magnitude to the gradient, and thus the trendline was not forced through the origin. The observed PN/PM gradient is reasonably close to previously reported correlations for Diesel vehicles (with DPFs) [e.g. 2], but the large y-intercept value implies the presence of large numbers of particles of almost zero mass, in contrast to correlations observed for Diesel engines (with a DPF). Possible reasons include the physicochemical parameters of the two fuels – and the impact this has on the relative sizes of the volatile and non-volatile particle fractions, the presence or absence of a particulate filter, injection pressure, combustion temperature, etc. Fuel ethanol content was observed to correlate only weakly with PN/PM ratio (Figure 3), as previously reported in another study [3]. Further testing would be necessary to definitely determine the existence of a correlation between fuel ethanol content and this metric. In order to determine the nature of the response of PM and PN emissions to the ethanol content of the fuel blend, it would be necessary to conduct further testing with a greater number of ethanol blends of low and moderate ethanol content. Additionally, in order to further understand the effect of ambient temperature, tests could be conducted at temperatures between the two mandated EU test temperatures.
1.00E+11
1.00E+12
1.00E+13
0 10 20 30 40 50
PN
/PM
ra
tio
[#/m
g]
Blend ethanol content [Vol. %]
UDC EUDC
Figure 3 – The ratio of PN to PM results obtained at 24°C for various ethanol blends
Considering the concurrent trends of increasing interest in direct injection petrol engines and increasing concern over particle number and surface area, research in this area remains an intriguing necessity. Since usage of both ethanol and direct injection engines will increase in the future, the subject addressed in this research is of considerable interest regarding future emissions scenarios. Likewise, further investigations into the relative merits of each particulate matter metric (mass, number, size distribution, effective surface area) and their reliability and relevance to real world emissions are paramount.
References:
[1] Kittlelson, D.B. Engines and nanoparticles: a review. Journal of Aerosol Science, 29, 5-6, 575-588, 1998. DOI: 10.1016/S0021-8502(97)10037-4.
[2] Bielaczyc, P., et al. Correlation between particle mass and number for Euro 5 compression ignition vehicles, Book of Abstracts, The 15th ETH-Conference on Combustion Generated Nanoparticles, Zurich, Switzerland, 26-29 June 2011.
[3] Maricq, M.M., et al. The Impact of Ethanol Fuel Blends on PM Emissions from a Light-Duty GDI Vehicle, Aerosol Science and Technology, 46, 5, 2012. DOI: 10.1080/02786826.2011.648780.
16th ETH Conference on Combustion Generated Nanoparticles24th – 27th June 2012, Zurich, Switzerland
Investigations on fuel ethanol content and bi t t t ti l b ambient temperature on particle number
and mass emissions from vehicles featuring direct injection gasoline engines
Piotr Bielaczyc, Andrzej Szczotka, Joseph WoodburnBOSMAL Automotive R&D Institute Ltd, Poland
P. Bielaczyc 25th June 2012 16th ETH-Conference 1
,
Presentation contentsPresentation contents
Research context & Experimental aims
Introduction
Experimental Details
Results & Discussion
Conclusions
P. Bielaczyc 25th June 2012 16th ETH-Conference 2
Research context & Experimental aimsResearch context & Experimental aims The physicochemical characteristics of standard petrol and ethanol
differ widely. The addition of even small quanitities of ethanol to
petrol has been shown to affect gaseous exhaust emissions. Other
studies have reported effects on PM and PN emissions.
This study seeks to obtain further information on PM (mass) and
PN (number) emissions with fuels of varying ethanol content.
Interest in direct injection engines is increasing and this engine type Interest in direct injection engines is increasing and this engine type
will be used much more widely in the future. Interest is using
greater quantities of ethanol in gasoline is also increasing The greater quantities of ethanol in gasoline is also increasing. The
research presented here is therefore of great importance for future
t ti i i
P. Bielaczyc 25th June 2012 16th ETH-Conference 3
automotive emissions.
Introduction:
Concern over particulate matter emissions from
Introduction:
Concern over particulate matter emissions from automotive engines has been expressed for decades
This has lead to the introduction of both mass and number-based emissions limits for Diesel vehicles in the European Union; direct-injection petrol engines have now also had a particle number limit set: 6.00x1012
#/km (for the year 2014); 6 00x1011 #/km for the year#/km (for the year 2014); 6.00x1011 #/km for the year2017
Interest is gowing in using increased quantities of ethanol in petrol fuel blends for automotive applications
P. Bielaczyc 25th June 2012 16th ETH-Conference 4
p pp
Background: ethanol as an automotive fuel:
Using ethanol as a vehicular fuel is not a new ideag
Primary alcohol, formula C2H5OH Suitable for use in SI engines, usually blendedg y
h l0% ethanol 100 % gasoline
100% ethanol0% gasoline
E85E5 E10 …etc…
P. Bielaczyc 25th June 2012 16th ETH-Conference 5
Background: petrol-ethanol blends worldwide
Sweden: E85E85
EU:
USA: E E
E5…E10 India: E5, E10
Brazil: E18-25,
E10, E15Australia:
E105,E85, E100
P. Bielaczyc 25th June 2012 16th ETH-Conference 6
Background: ethanol as an automotive fuel:
Interest has been growing in ethanol Interest has been growing in ethanol
Potentially sustainable, renewable biofuel, depending on feedstock
The EU’s Fuel Quality Directive (2009/30/EC) enabled more
widespread use of ethanol in petrol in the EU (blends up to E10)
Currently the most widely-used biofuel
All ‘standard’ gasoline sold in the EU is currently E5g y 5
US EPA approved E15 for use in recent model year cars
P. Bielaczyc 25th June 2012 16th ETH-Conference 7
Direct Injection Spark Ignition (DISI) engineect ject o Spa g t o ( S ) e g e
The major advantages of a j gDISI engine are increased fuel efficiency and high power outputp p
Emissions levels can also be more accurately be more accurately controlled with the DISI system
No throttling losses in some DISI engines, when compared to a compared to a conventional fuel-injected engine
P. Bielaczyc 25th June 2012 16th ETH-Conference 8
Direct Injection Gasoline engineDirect Injection Gasoline engine Engine speed is controlled by the
engine control unit/engine engine control unit/engine management system (EMS), which regulates fuel injection function and ignition timing
High fuel efficiency is an advantage; emissions of particulate matter are a disadvantagedisadvantage
The literature suggests that DISI PM emissions are higher than SI MPI PM emissions are higher than SI MPI PM emissions and DISI PN emissions are much much higher than SI MPI PN emissions
DISI engines will come to have the largest market share for SI engines in the future
P. Bielaczyc 25th June 2012 16th ETH-Conference 9
the future
0,20Vehicle type A - CI - IDI
Background: Example emissions
0,16
ypVehicle type B - CI - IDIVehicle type C - CI - IDIVehicle type D - CI - DIVehicle type E - CI - DI
C
D0,12
n [
g/k
m]
ypVehicle type F - SI - DIVehicle type G - SI - MPIVehicle type H - SI - MPI
A
AA
B
B
B
CC
C
DDD
EE
F
0,08
M e
mis
sio
n
AA
BB
C DD
EE
E
FF
0,04
PM
FF
F
GGGGGHHHHH
0,00
195 s 195 s 390 s 400 s UDC+EUDC195 s 195 s 390 s 400 s UDC+EUDC
PM emissions for different vehicles with SI and CI engineSource: Bielaczyc P., Szczotka A., Pajdowski P. EAEC Conference 2001
P. Bielaczyc 25th June 2012 16th ETH-Conference 10
Pajdowski P. EAEC Conference 2001
Background: Filter overviewCI - IDI SI - DI
SI - MPICI - DI
Filters with particulate matters from NEDC cycle for different types of vehicles
CI - IDI SI - DI SI - MPI SI - CNG
Filters with particulate matters from first 195 s of NEDC cycle for different types of vehiclesSource: Bielaczyc P., Szczotka A., P jd ki P EAEC C f
P. Bielaczyc 25th June 2012 16th ETH-Conference 11
Pajdowski P. EAEC Conference 2001
Experimental Details
P. Bielaczyc 25th June 2012 16th ETH-Conference 12
Experimental Details : test facility Schematic diagram of BOSMAL’s emissions measurement setup, as used to
measure emissions from the test vehicles
Experimental Details : test facility
P. Bielaczyc 25th June 2012 16th ETH-Conference 13
Experimental Details : test facility BOSMAL’s climate-controlled test chamber
(-35 °C to +60 °C) housing a chassisdynamometer and windspeed fan Emissions analysers
Dilution tunnel and particle countery p
Testing carried out in accordance with EU & PMP test requirements.
Test temperatures: +24 °C and -7 °C Driving cycle: NEDC
Three tests performed on each fuel (meanvalues plotted in graphs)
Blends E5, E10 and E50 tested at -7 °C (twot t bl d l l tt d h )
P. Bielaczyc 25th June 2012 16th ETH-Conference 14
Driving cycle: NEDC tests blend; mean values plotted on graphs)
Experimental Details: VehicleExperimental Details: VehicleVehicle type Passenger car
E i i S d d E Emission Standard Euro 5Engine capacity [dm3] Approx. 1.4
Max. Power [kw] 67Max. Torque [Nm] 200
Aftertreatment system Close Coupled TWC
λ λ = 1
(stoichiometric)
P. Bielaczyc 25th June 2012 16th ETH-Conference 15
Results & Discussion
P. Bielaczyc 25th June 2012 16th ETH-Conference 16
Results: PM & PN during the NEDCResults: PM & PN during the NEDC
NEDC
1,00E+13PN
3
3,5PM
2
2,5
3
1,00E+12
1
1,5
2
PN
1,00E+11
E 5 E 10 E 25 E 50 E 5 EOT
0
0,5PM
/k /kE 5 E 10 E 25 E 50 E 5 EOT
• Limited effect for PN (considering test-to-test variability); stronger effect for PM
#/km mg/km
P. Bielaczyc 25th June 2012 16th ETH-Conference 17
• Mass and number emissions lowest from the E50 blend and highest from the E10 blend
Results: PM & PN during the UDCResults: PM & PN during the UDC
UDCPN PM
1,00E+13
PN
5
6
PM
1 00E+12 3
4
1,00E 12
1
2
3
PN
1,00E+11
E 5 E 10 E 25 E 50 E 5 EOT
0
1PM
/k5 0 5 50 5 O
• Limited effect for PN (considering test-to-test variability); stronger effect for PM
#/km mg/km
P. Bielaczyc 25th June 2012 16th ETH-Conference 18
• Mass and number emissions lowest from the E50 blend and highest from the E10 blend
Results: PM & PN during the EUDCResults: PM & PN during the EUDCEUDCPN PM
1,00E+13
1,61,82
1,00E+12
11,21,4,
1,00E+11
0 20,40,60,8
PN
1,00E+10
E 5 E 10 E 25 E 50 E 5 EOT
00,2PM
#/km mg/km
• Wide variation between blends for PM; relatively little for PN (considering test-to-test variability)• Mass and number emissions highest from the E10 blend• Low emissions from the E50 blend despite a substantially increased fuel flow rate
#/km mg/km
P. Bielaczyc 25th June 2012 16th ETH-Conference 19
• Low emissions from the E50 blend, despite a substantially increased fuel flow rate
Results: PN/PM ratios for the UDC and EUDC3,00E+12
UDC EUDC
2,50E+12
mg]
1 50E 12
2,00E+12
M ra
tio [#
/
1,00E+12
1,50E+12
PN/PM
5,00E+11
,
0 10 20 30 40 50Blend ethanol content [Vol. %]
• No obvious correlation observed between ethanol content and the PN/PM ratio• Very low repeatability
P. Bielaczyc 25th June 2012 16th ETH-Conference 20
• Very low repeatability
Results: PM & PN during the NEDC at 7 CResults: PM & PN during the NEDC at -7 CNEDC
1,00E+13PN
12,00PM
8,00
10,00
1,00E+12
4,00
6,00
PN
1,00E+11 0,00
2,00PM
E 5 E 10 E 50
• Wide variation between blends for PM; very low emissions for the E50 blend
mg/km#/km
P. Bielaczyc 25th June 2012 16th ETH-Conference 21
y
Results: PM & PN during the UDC at 7 CResults: PM & PN during the UDC at -7 C
UDC
1 00E 13PN
25 00PM
1,00E+13
20,00
25,00
1,00E+1210 00
15,00
5,00
10,00PN
PM
1,00E+11
E 5 E 10 E 50
0,00
mg/km#/km
P. Bielaczyc 25th June 2012 16th ETH-Conference 22
Results: PM & PN during the EUDC at 7 CResults: PM & PN during the EUDC at -7 C
EUDC
1 00E 13PN
25 00PM
1,00E+13
20,00
25,00
1,00E+1210 00
15,00
5,00
10,00UDC PN
UDC PM
1,00E+11
E 5 E 10 E 50
0,00
mg/km#/km
P. Bielaczyc 25th June 2012 16th ETH-Conference 23
Results: Temperature influence for E51401.00E+12
E5 (24 d C)
1201.00E+11
E5 (24 deg C)
E5 (-7 deg C)
Vehicle speed
100
1.00E+10
/h]
#/cm
3 ]
60
80
1.00E+09
icle
sp
eed
[km
on
cen
trat
ion
[#
40
1.00E+08
Veh
Par
ticl
e c o
201.00E+07
01.00E+060 200 400 600 800 1000 1200
Time [s]
• Significantly higher emission at -7 °C for idling and transientsDiff ll b di f d 650 d
P. Bielaczyc 25th June 2012 16th ETH-Conference 24
• Differences all but disappear after around 650 seconds
1401.00E+13E10 (24 deg C)
Results: Temperature influence for E10
1201.00E+12
E10 (24 deg C)
E10 (-7 deg C)
Vehicle speed
1001.00E+11
m/h
]
[#/c
m3 ]
60
80
1.00E+09
1.00E+10
hic
le s
pee
d [
km
con
cen
trat
ion
[
401.00E+08
Veh
Par
ticl
e c
201.00E+07
01.00E+060 200 400 600 800 1000 1200
Time [s]
• Significantly higher emission at -7 °C for idling and transientsDiff ll b di f d 650 d
P. Bielaczyc 25th June 2012 16th ETH-Conference 25
• Differences all but disappear after around 650 seconds
1401.00E+12E50 (24 deg C)
Results: Temperature influence for E50
1201.00E+11
E50 (24 deg C)
E50 (-7 deg C)
Vehicle speed
100
1.00E+10
m/h
]
[#/c
m3 ]
60
80
1.00E+09
hic
le s
pee
d [
km
con
cen
trat
ion
[
40
1.00E+08
Veh
Par
ticl
e c
201.00E+07
01.00E+060 200 400 600 800 1000 1200
Time [s]
• Significantly higher emission at -7 °C for idling and transients up to around 400 seconds• Surprisingly emission at -7 °C in fact lower for some portions of the cycle (550-900 seconds)
P. Bielaczyc 25th June 2012 16th ETH-Conference 26
• Surprisingly, emission at -7 C in fact lower for some portions of the cycle (550-900 seconds)
Conclusions
P. Bielaczyc 25th June 2012 16th ETH-Conference 27
Conclusions (1/2)Conclusions (1/2)
At 24 C PN emissions were high but were (just) under At 24 C, PN emissions were high, but were (just) underthe upcoming limit of 6.00E12 #/km for all test fuels.PM emissions were all well under the DISI limit.
Emissions of PN and PM increased substantially at -7 C,but the difference varied with the ethanol content of thefuel.
Possible reasons include the physicochemicalt f th t f l d th i t thi hparameters of the two fuels – and the impact this has on
the relative sizes of the volatile and non-volatile particlefractions catalyst light-off time and the oxidation offractions, catalyst light off time and the oxidation ofparticles in the TWC.
P. Bielaczyc 25th June 2012 16th ETH-Conference 28
Conclusions (2/2)Conclusions (2/2)
While considerable variation was observed, fuelethanol content was observed to correlate onlyethanol content was observed to correlate onlyweakly with the PN/PM ratio. The results suggestthat large numbers of particles of negligible massthat large numbers of particles of negligible masswere emitted.
Despite the increased oxygen content of the fuel, Despite the increased oxygen content of the fuel,higher ethanol blends showed limited to non-existent decreases in particle number and masse ste t dec eases pa t c e u be a d assemissions, although the E50 blend did appear toperform well.
P. Bielaczyc 25th June 2012 16th ETH-Conference 29
p
Thank you for your attention
Any questions?
Thank you for your attention
Any questions?
Contact: [email protected]
P. Bielaczyc 25th June 2012 16th ETH-Conference 30