RPT 14 06
RPT 14 06
RPT 14 06
6/14
Impact of FAME on
the performance of
three Euro 4 light-
duty diesel vehicles
Part 1: Fuel
consumption and
regulated emissions
report no. 6/14
Impact of FAME on
the performance of
three Euro 4 light-
duty diesel vehicles
Part 1: Fuel
consumption and
regulated emissions
Prepared for the Concawe Fuels and Emissions Management Group by its Special
Task Force FE/STF-25:
L. Jansen (Chair)
J. Ariztegui
J-F. Bernard
M.D. Cárdenas Almena
R. Clark
N.G. Elliott
P. Gomez-Acebo
C. Higham
B. Kelecom
L. Pellegrini
P. Scott
I
report no. 6/14
ABSTRACT
By 2020, EU legislation will require that 10% of the total transport fuel energy demand is met by
the use of renewable energy, primarily by blending bio-components. Although many types of
blending components for diesel fuels are being considered to achieve this requirement, Fatty Acid
Methyl Esters (FAME) are the most likely to be used in significant volumes over the coming decade.
FAME products have been used in Europe for many years, both as blends and as neat fuels, in
certain niche markets.
One unanswered question concerning FAME/diesel fuel blends is the effect of FAME on fuel
consumption. Since FAME has a slightly lower energy content compared to hydrocarbon-only
fuels, a higher volumetric fuel consumption is expected unless the vehicle is able to compensate
in some way for the energy loss associated with the bio-component in diesel fuel.
To answer this question, Concawe completed a vehicle study in which four diesel fuel blends with
FAME (as Rapeseed Methyl Ester (RME)) were tested in three Euro 4 light-duty passenger cars,
each equipped with different after-treatment technologies. The FAME contents of these fuels
varied from 0% to 50% v/v in order to accentuate the effect of FAME on the energy content of the
blended diesel fuels. The programme was statistically designed to give a robust and repeatable
testing schedule so that fuel consumption and tailpipe emissions data could be reliably collected
over regulatory and transient driving cycles. The vehicle study was conducted for Concawe by the
Laboratory for Applied Thermodynamics of the Aristotle University of Thessaloniki, Greece.
Fuel consumption data for all three vehicles over all driving cycles show that the volumetric fuel
consumption increases in direct proportion with increasing FAME content and the decreasing
volumetric lower heating value (energy content) of the FAME/diesel fuel blends. There was no
detectable change in the energy efficiency of the vehicles on different fuel blends and they were
not able to compensate for the lower energy content of the FAME/diesel blends through improved
performance.
Increasing the FAME content also reduced the PM but increased the NOx, HC, and CO emissions.
The overall impact of FAME on tailpipe emissions was small when compared to the variations in
emissions seen for different driving cycles and for different vehicles over the same driving cycle.
No significant difference in emissions performance was observed for the two types of Diesel
Particulate Filter (DPF) aftertreatment systems that were tested in these vehicles.
It is expected that these results will be of importance to those interested in the impact of FAME in
diesel fuel on Well-to-Wheels fuel consumption and on tailpipe emissions from modern light-duty
passenger cars.
KEYWORDS
Exhaust emissions, diesel fuel, engine technology, vehicles, fuel quality, NEDC, Euro 4, fuel
consumption, particulate mass (PM), NOx, FAME, RME
NOTE:
Considerable efforts have been made to assure the accuracy and reliability of the information
contained in this publication. However, neither Concawe nor any company participating in
Concawe can accept liability for any loss, damage or injury whatsoever resulting from the use of
this information.
This report does not necessarily represent the views of any company participating in Concawe.
II
report no. 6/14
CONTENTS Page
1. INTRODUCTION 1
1.1. IMPACT OF FAME ON FUEL CONSUMPTION: PUBLISHED
LITERATURE 6
1.2. IMPACT OF FAME ON REGULATED EMISSIONS:
PUBLISHED LITERATURE 6
2. EXPERIMENTAL 8
2.1. DIESEL VEHICLES 8
2.2. BIODIESEL TEST FUELS 10
2.3. LUBRICANT SELECTION 12
3. METHODOLOGY 13
3.1. TEST PROTOCOL 13
3.2. STATISTICAL ANALYSIS 15
5. CONCLUSIONS 36
6. GLOSSARY 39
7. ACKNOWLEDGEMENTS 42
8. REFERENCES 43
APPENDIX 1 IMPACT OF FAME ON FC: PUBLISHED LITERATURE 46
APPENDIX 3 ANALYTICAL DATA FOR RME AND DIESEL TEST FUELS 58
APPENDIX 4 TEST PROTOCOL 60
APPENDIX 5 STATISTICAL DATA ANALYSIS 66
APPENDIX 6 EMISSIONS FROM DIESEL VEHICLES 68
APPENDIX 7 MEANS AND STANDARD DEVIATIONS 71
APPENDIX 8 FUEL CONSUMPTION CALCULATIONS 78
APPENDIX 9 FC RESULTS OVER OTHER TEST CONDITIONS 80
APPENDIX 10 RPM DATA OVER NEDC FOR ALL VEHICLES 85
APPENDIX 11 ADDITIONAL MEASUREMENTS 86
III
report no. 6/14
SUMMARY
This vehicle study was designed by Concawe to investigate the effect of FAME on
fuel consumption and tailpipe emissions from three Euro 4 light duty passenger cars.
All vehicles had common rail turbocharged engines and were equipped with different
exhaust aftertreatment technologies. Two of the three vehicles had Diesel Particulate
Filters (DPF) that were regenerated using different strategies. All test work was
carried out at the Laboratory of Applied Thermodynamics (LAT) of the Aristotle
University of Thessaloniki, Greece.
Four fuels were evaluated in this study. One base diesel (B0 complying with the
EN 590 specification [3]) was blended with commercially sourced Rapeseed Methyl
Ester (RME) (complying with the EN 14214 specification [4]) to give three diesel/RME
blends at 10%, 30% and 50% v/v RME. These concentrations were selected in order
to anticipate future increases in biofuel blending levels for use in compatible vehicles
and to accentuate the impact of RME on vehicle performance and emissions.
The fuel consumption data for all three vehicles over all driving cycles show that the
volumetric fuel consumption increases in direct proportion to the increasing FAME
concentration and the decreasing volumetric lower heating value (energy content) of
the FAME/diesel fuel blends. There was no apparent change in the energy efficiency
of the vehicles on different fuel blends and the vehicles were not able to compensate
for the lower energy content of the FAME/diesel blends through improved
performance.
IV
report no. 6/14
Increasing the FAME content also reduced the particulate mass (PM) emissions and
the number of solid particles but increased the NOx, HC, and CO emissions. The
overall impact of FAME on emissions was small when compared to the variations in
tailpipe emissions seen for different driving cycles and for different vehicles over the
same driving cycle. No significant difference in emissions performance was observed
for the two types of Diesel Particulate Filter (DPF) aftertreatment systems evaluated
in this study.
V
report no. 6/14
1. INTRODUCTION
The Renewable Energy Directive (RED, 2009/28/EC) [2] will require 10% renewable
energy in transport fuels by 2020 within the European Union while the Fuel Quality
Directive (FQD, 2009/30/EC [28]) will require reductions in GHG emissions from
transport fuels. Changes to the European gasoline and diesel fuel specifications have
already been made to enable higher blending of bio-components into market gasoline
and diesel fuels in order to enable these requirements.
In the near term, biofuel obligations for diesel fuels will most likely be fulfilled with
FAME. In the longer term, bio-components from biomass and from other sources may
become more readily available and will undoubtedly be required in order to achieve
the EU ambitions. Although this is likely, the current test programme focussed on
FAME bio-components only and alternatives to FAME, such as Hydrogenated
Vegetable Oils (HVO) and Biomass-to-Liquids (BTL) blend components, may be
considered for future studies.
FAME can be used as a neat fuel (B100) but is most widely used in low concentration
blends with fossil diesel fuel. Currently, the EN 590 specification [3] allows up to 7%
v/v FAME meeting the EN 14214 specification [4] to be blended into diesel fuel. Many
different FAME types are commonly used throughout Europe and the rest of the world.
Rapeseed Methyl Ester (RME) is the most frequently used, however, and a single
batch of RME was chosen for this study (see Section 2.2).
FAME use is an everyday practice in most EU Member States and its share in diesel
fuel is rising year by year. B100 is allowed in some national legislation and some
adapted vehicles, including those in captive fleets, can operate with high percentages
of FAME, typically up to 30% v/v FAME (B30). As a result, any potential impact of
biodiesel blends on vehicle emissions is already affecting vehicle fleet pollutant
emission levels. These impacts could gain greater significance in view of the stricter
emission standards (Euro 5) that are now in place [19].
Of special interest is the potential impact that higher FAME contents may have on
vehicle fuel consumption and on regulated and unregulated emissions. Previous
vehicle test programmes [5,6,7,8] have studied the effect of bio-components on both
fuel consumption and regulated emissions but many of the more systematic studies
have focused on heavy duty engines that were not yet equipped with exhaust
aftertreatment systems. The most comprehensive assessment of heavy-duty vehicle
emissions performance was published by the US EPA [9] which showed that NOx
emissions increase and PM, CO, and HC emissions decrease as a function of
increasing FAME content. The published literature does not adequately address the
impact of FAME on modern light-duty diesel vehicle performance, however, and there
is an overall gap in understanding on performance issues related to light-duty diesel
vehicles that are important to the European marketplace.
The fuel consumption (FC) of light-duty vehicles operating on biodiesel blends is also
an important question as attention increasingly focuses on the GHG reductions that
can be achieved from these fuel blends. In most Well-to-Wheels (WTW) studies [10],
the vehicle's efficiency is assumed not to change when the engine runs on an
oxygenated fuel, that is, the same megajoules (MJ) of fuel are needed to complete a
prescribed driving cycle for both hydrocarbon-only and oxygenated diesel fuels. This
means, however, that a slightly higher volumetric FC is expected for oxygenated fuels
because the volumetric energy content of these fuels is somewhat lower than that of
hydrocarbon-only fuels. This effect is easily observed in gasoline vehicles fuelled with
1
report no. 6/14
ethanol blends and will be more evident in diesel vehicles if the concentration of FAME
in diesel were to increase.
For this reason, it is important to know whether modern vehicles are capable of
recovering a portion of this volumetric penalty through better engine efficiency when
running on oxygenated fuels. The published literature is not entirely clear on this point,
in part because the energy content of FAME/diesel blends is only slightly lower than
that of hydrocarbon-only diesel fuels. Detecting small differences in volumetric FC is
difficult experimentally.
Accordingly, Concawe designed and carried out this statistically robust test
programme to answer this question by generating new FC data from European diesel
passenger cars. At the same time, this study provided an opportunity to re-evaluate
the impact of oxygenate blends on vehicle exhaust emissions. Some results from this
study have been published elsewhere [25].
The objective of this test programme then was to generate definitive FC data and an
emissions database on diesel fuels that contain varying concentrations of RME. This
work was conducted on three diesel vehicles where our evaluation of recent literature
indicates that there is a lack of reliable or consistent data.
Three vehicles complying with the Euro 4 emissions standard [18] were selected for
this study. These vehicles were equipped with different exhaust aftertreatment
systems and were chosen as representatives of the European diesel passenger car
fleet. All vehicles used Exhaust Gas Recirculation (EGR) for NOx reduction and some
form of Diesel Oxidation Catalyst (DOC). In addition, Vehicles 1 and 3 were equipped
with Diesel Particulate Filters (DPF) that were regenerated using two different
strategies. Details on the vehicles and their aftertreatment systems can be found in
Section 2.1.
Four fuels were tested in this study. One base diesel fuel (Fuel B0 meeting the EN 590
specification) was blended with a single batch of commercial RME (meeting the
EN 14214 specification) to give blends containing 10% v/v (B10), 30% v/v (B30) and
50% v/v (B50) RME. These RME concentrations were selected in order to anticipate
future increases in biofuel blending levels for use in compatible vehicles and to
accentuate the impact of RME on vehicle performance and emissions.
Details on the test fuels are given in Section 2.2. RME was chosen for this study
because it is oxidatively more stable and is manufactured from a common feedstock
used for FAME production in the EU-27 as shown in Figure 1 [11]. The type of
feedstock used to manufacture the FAME is not expected to play a large part in the
fuel consumption and exhaust emissions results studied here. A summary of
published literature related to this question is provided in Appendix 2.
2
report no. 6/14
EU-27 (2008)
6,000
5,000
FAME Feedstock
Other Feedstocks
4,000
(tonnes)
Palm Oil
3,000 Rapeseed Oil
2,000 Soybean Oil
1,000
Sunflower Oil
All vehicle/fuel combinations were tested five times on the B0 fuel and four times each
on the three RME-containing fuels using a statistical design (see Section 3.2). A
range of test cycles and steady state conditions were evaluated:
New European Driving Cycle (NEDC), including the Urban Driving Cycle (UDC,
also called the ECE-15) and the Extra Urban Driving Cycle (EUDC);
ARTEMIS Urban and Road driving cycles [12]; and
120 and 50 km/h steady state conditions.
These driving cycles allowed any differences between cold and hot start tests to be
identified along with any changes as a result of transient versus steady state
operation.
At each test condition, all regulated tailpipe emissions were measured (NOx, HC, CO,
PM) along with CO2 and FC data and these measurements are included in this report.
For some test conditions, additional measurements were made of modal NO and NOx
emissions and also of total particle number (PN). Additionally, particle size
distributions were collected under steady-state driving conditions and these
measurements were carried out in accordance with PMP 1 protocols. The results from
both the NO and PN measurements are included in a second report [1] as well as the
analysis of PM composition and carbonyl emissions. Over the course of the vehicle
study, the oxidation stability of the RME/diesel blends was monitored and the results
are included in Appendix 11.
Vehicles selected for this study were equipped with DPF aftertreatment systems to
reduce particle emissions. Among the different types of DPF technologies in common
use, the fuel-borne catalyst technology has a reasonable market share on the short-
to medium-term but is not considered to be the favourite DPF option for the longer-
term. In-cylinder post injection is expected to be the dominant DPF technology for the
longer-term (Figure 2). Diesel Particulate NOx Reduction (DPNR) technology is not
widely used today and it has been noted that regeneration of the DPNR system occurs
more frequently compared to standard DPF technology increasing overall fuel
3
report no. 6/14
consumption. In this study, the DPF regeneration events were monitored to ensure
that they did not occur during FC and emissions measurements.
Figure 2 Expected EU market share for different DPF regeneration strategies in 2012
for light-duty passenger cars (Source: Ricardo (2008))
Especially with the in-cylinder post-injection strategy, higher FAME levels in diesel
fuel can increase the higher-boiling fuel components (including FAME) that appear in
the lubricant. This is because DPF regeneration is achieved by periodically injecting
extra fuel into the cylinder at the end of the combustion cycle. This can lead to
changes in the lubricant performance and, in worse cases, damage to the oil pump
and engine due to higher total lubricant volumes.
Among the three main types of DPF regeneration strategies, the in-cylinder post-
injection strategy is the most commonly used for light-duty diesel vehicles today
(Figure 2). This approach (used in Vehicle 1) and, to a lesser extent, the fuel-borne
catalyst approach (used in Vehicle 3) are expected to be sensitive to this lubricant
problem. The third DPF regeneration strategy, in-exhaust injection, is not expected to
result in fuel accumulation in the lubricant because the fuel injector is mounted in the
aftertreatment system, typically in front of the DPF. This approach is similar to non-
DPF aftertreatment systems in terms of the impact of FAME in fuel on the lubricant
system.
Diesel fuels having higher FAME contents have also been reported to increase the
build-up of ash on the DPF leading to higher backpressure across the filter and the
need for more frequent DPF regeneration cycles. There is some evidence that FAME
can also reduce the balance point temperature for passive soot combustion. The
different energy contents of FAME and FAME/diesel blends could also be an issue
for engine calibration settings. There are interesting questions and an additional
Concawe study has been conducted to investigate some of these effects.
4
report no. 6/14
In this study, Vehicle 2 was not equipped with a DPF and was included in order to
give a clear indication of how the engine responded to fuel without complications from
exhaust system regeneration. This also allowed read-across of the results to other
Concawe studies that had been previously done on non-DPF vehicles [13,14,15,16].
FAME has about 14% (on a mass basis) less energy content compared to fossil diesel
fuel. This lower energy content is expected to increase vehicle fuel consumption on a
volumetric basis proportional to the FAME concentration. Well-to-Wheel (WTW)
analyses [10] have shown the potential benefits in energy content and GHG
emissions for various FAME products when used as neat fuels (Figure 3). It is
important to know from a Tank-to-Wheels (TTW) and WTW perspective whether the
efficiency of modern engines changes when running on FAME-containing fuels.
Figure 3 WTW expended fossil energy and GHG emissions savings for 100% bio-
diesel pathways compared to conventional diesel (2020+ DICI vehicle) [10]
5
report no. 6/14
Before beginning the vehicle study, a literature review was completed in order to
evaluate the previously reported impacts of FAME on fuel consumption (see
Appendix 1). Of the 99 papers examined in this review, 45 papers were selected as
promising candidates for information on the impact of FAME-containing fuels on FC.
Although these papers were studied in detail, it became apparent that none of the
experiments reported in these papers were designed to look specifically at changes
in engine efficiency. A best attempt has been made, however, to analyse them from
this viewpoint but the robustness of the experimental design has led to the published
efficiency data being of questionable value.
The key findings from the published literature are summarized here. First, there is
little to no discrimination in FC effects for small changes in FAME content, e.g. from
B0 to B10. Where papers have claimed an engine efficiency benefit when running on
oxygenated fuels, the results are generally associated with a lack of discrimination
between fuels. Analysing the data from these papers “second-hand” has required
calculations of energy consumption or engine efficiency and these have had to rely
on the Lower Heating Values (LHV) reported in the papers. Some of these LHVs are
of questionable accuracy and represent another potential source of error in an
analysis of published results.
The same published literature was evaluated to identify the impact of FAME in diesel
fuel on regulated emissions (see Appendix 3). The majority of the published data
6
report no. 6/14
refers to work carried out on heavy duty engines and covers an extensive range of
fuel types, FAME concentrations, engines, and test protocols. In 2002, the US
Environmental Protection Agency (EPA) published a comprehensive statistical
analysis of all available data [9] from which the following graph has been extracted.
The effects shown in Figure 4 represent the most widely reported view of biodiesel
effects on regulated emissions. However, it should be remembered that these results
are from a collection of published studies that predominantly focused on heavy duty
engines (and primarily on US market engines) that were not equipped with exhaust
aftertreatment and tested only over hot start test cycles. It may not be reasonable to
assume that these results will be representative of modern European light duty
vehicles that are equipped with a variety of aftertreatment technologies and are
certified over a cold start test cycle. There are considerably fewer publications related
to modern light-duty diesel vehicles and the results that have been reported are
generally less consistent than those from the heavy duty tests. For this reason, it was
important to include both regulated and unregulated exhaust emissions in this study.
7
report no. 6/14
2. EXPERIMENTAL
Three light-duty diesel passenger cars complying with the Euro 4 emission regulations
[17,18] were selected for this test programme. The choice of vehicles was made in
order to cover the most prominent engine and exhaust aftertreatment technologies
currently present in the European fleet. For this reason, all three vehicles were
equipped with direct injection, high injection pressure common rail and turbocharged
engines. All vehicles used EGR for NOx control and a DOC for reducing CO and HC
emissions. Vehicles 1 and 3 were also equipped with a DPF for controlling PM
emissions. Previous results have shown that DPFs can be effective at reducing PN
emissions, namely those ultrafine particles that have aerodynamic mobility diameters
less than about 0.1µm.
Two different types of DPF were chosen for this study based on their regeneration
strategy. Vehicle 1 was equipped with a catalysed DPF (CDPF) while Vehicle 3 used
a fuel-borne additive to regenerate the DPF. As shown in Figure 2, these two DPF
regeneration technologies are likely to be the most widely used options for DPF
regeneration in the European light-duty diesel fleet. The vehicles and their
characteristics are shown in Table 1.
8
report no. 6/14
The average regulated emissions from these vehicles over the NEDC are shown in
Figure 5 for all five tests on the hydrocarbon-only diesel fuel (B0). Compared to the
Euro 4 and 5 emissions limits shown by the horizontal bars, all three vehicles
complied with the Euro 4 limits, although the NOx emissions for Vehicle 3 and the PM
emissions for Vehicle 2 were borderline. The DPF-equipped vehicles reduced the PM
emissions to levels that were substantially lower than that required even by the
upcoming Euro 5 emissions regulation [19].
2 Vehicle 3 is the ‘golden vehicle’ that was previously used as the round-robin test vehicle in the European
Particle Measurement Programme (PMP) [20,21]
9
report no. 6/14
0.5 Vehicle 1
Vehicle 2
Regulated Emissions (g/km)
0.4 Vehicle 3
0.3
Euro 4 NEDC Limits
Euro 5 NEDC Limits
0.2
0.1
0.0
HC CO NOx PM
Although some European countries, mainly Germany and Austria have marketed
B100 fuels (that is, 100% FAME), low level blends of FAME in diesel fuel are more
commonly used across the European market. Provided that the FAME blend stock
meets the EN 14214 standard [4], up to 7% v/v FAME can be blended into
conventional diesel fuels according to the European EN 590 standard [3]. CEN is also
working to produce a specification for a fuel with FAME content in diesel fuel up to
10% v/v FAME (B10) for compatible vehicles. For fleet applications, there is also
some use of mid-range blends, typically 20-30% v/v FAME in diesel fuel.
In principle, FAME, either neat or in blends with diesel fuel, can be used in normal
diesel engines. The cetane number of FAME is typically above 50 and the product
distils within the normal diesel distillation range, although on the high end of this
range. The properties of FAME are sufficiently different from conventional diesel,
however, that some impacts on combustion can be expected.
Diesel fuel containing up to 7% v/v FAME and meeting the EN 590 standard can be
used in all conventional diesel vehicles, according to warranty information provided
by the auto manufacturers. At higher FAME levels in diesel fuel, FAME has been
reported to affect engine and aftertreatment performance and some materials, such
as fuel system seals and gaskets. Auto manufacturers have expressed some
concerns about whether the composition and oxidative stability of FAME, as specified
in EN 14214, are adequate to avoid these problems and this is the subject of on-going
work within CEN.
10
report no. 6/14
Four fuels were specially blended for this study: a reference diesel fuel (B0) and 3
blends of this diesel fuel with FAME. The base fuel complied with the European
EN 590 standard and had a maximum 10 mg/kg sulphur content.
The FAME was a European Rapeseed Methyl Ester (RME) complying with the
European EN 14214 standard and contained 1,000ppm of a commercial antioxidant
additive (butylated hydroxy toluene (BHT)) to ensure oxidation stability. The Rancimat
oxidation stability was monitored throughout the study (see Appendix 11).
The RME was blended into the EN 590 diesel fuel at three different concentrations,
resulting in the following four test fuels:
The properties of these four test fuels are presented in Appendix 3. Table 2 lists the
fuel parameters that are most important for the fuel consumption evaluation.
As shown in Table 2, increasing the FAME content from B0 to B50 resulted in a small
(~7%) reduction in LHV. For this reason, the test protocol was carefully designed with
sufficient statistical power in order to control experimental variability and to
differentiate between small differences in FC and regulated emissions.
3 HiTEC® 4678
4 Average of two or more measurements conducted at Concawe Member Companies
11
report no. 6/14
A single lubricant was used in all three vehicles in order to avoid the influence of
different lubricants on engine performance and emissions. The selection was made
after consulting with the manufacturers of all three test vehicles. The lubricant was an
SAE grade 0W-30 for diesel engines (ACEA class B3-B5) with low sulphur,
phosphorus and ash contents.
12
report no. 6/14
3. METHODOLOGY
From the analysis of the published literature and the small change in LHV for the
RME/diesel blends, it was apparent that repeatable, reliable and consistent
measurements would be required in order to interpret the impact of RME
concentration on fuel consumption and emissions. For this reason, a robust and
statistical experimental design was required.
Before starting a test on a given vehicle, the lubricant was changed and the vehicle
was driven for at least 1000km in order to condition the lubricant. During this mileage
accumulation, information was collected on the DPF-equipped Vehicles 1 and 3 to
identify the indicators for the onset of DPF regeneration. This ensured that DPF
regenerations would be observed if they occurred during the daily measurement
protocol.
From this information, the mileage interval at which the DPF regeneration would occur
could be determined and steps taken to avoid regeneration during testing. An
additional step was included in the vehicle conditioning to assess how close the
vehicle was to regeneration and, if required, a ‘forced’ regeneration was carried out
to ensure that regeneration did not occur during the next day’s testing. More
information on DPF regeneration is given in Section 4.4.1.
Figure A4-1 in Appendix 4 shows the complete daily procedure for vehicle
preparation and testing which consisted of the following steps:
Pre-condition the vehicle on the designated test fuel, including a controlled DPF
regeneration, if required;
Cold soak the vehicle overnight;
Complete fuel consumption and emissions measurements in the following
sequence:
NEDC on the cold vehicle consisting of the UDC (also called the ECE-15)
and the EUDC,
ARTEMIS Urban and Road transient cycles [12],
120km/h steady-state with some measurements also collected at a
50km/h steady-state condition.
Strict adherence to this protocol ensured that all vehicle preparation, fuel changes,
and fuel consumption and emissions measurements were carried out in a repeatable
manner for each vehicle/fuel combination.
The RME/diesel blends (B10, B30, and B50) were tested four times in each vehicle
while the base fuel (B0) was tested five times in order to improve estimates of baseline
performance and engine drift with time. A typical test order is shown below:
Test 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Fuel B0 B10 B50 B30 B0 B10 B30 B50 B0 B50 B30 B10 B0 B10 B50 B30 B0
13
report no. 6/14
In this example, the three RME/diesel blends are tested in four randomized blocks
and the five B0 tests are positioned between adjacent blocks and at the start and end
of the testing.
The block design approach reduces the risk that fuel effects will be confounded by
potential extraneous sources of variability while the absence of back-to-back
measurements ensures that the results are truly independent. Different randomized
orders were used for each vehicle (see Appendix 4). Seventeen days of testing were
required to complete testing on one vehicle and a single driver was used for all vehicle
tests.
A repeat test criterion was included in the test protocol. If the ratio of the highest to
the lowest FC measurements over the cold NEDC for a particular vehicle and fuel
combination exceeded 1.05, then a repeat test was to be conducted on that vehicle
and fuel as soon as possible or at the end of the test programme. This criterion was
based on historical repeatability data from various sources. In this study, the variation
in FC results did not exceed this ratio and no repeat tests were required. Similar
criteria were not set for the other tailpipe emissions and cycles.
An example is shown in Figure 6 for FC results for Vehicle 3 over the NEDC. The
left-hand chart shows the FC results in the actual daily testing order while the right-
hand chart shows the same results rearranged according to the fuel tested. The
results of this type were statistically analyzed as described in Section 3.2.
Figure 6 FC results for Vehicle 3 over the NEDC in actual testing order and in fuel
order
Vehicle 3: NEDC Results in Testing Order Vehicle 3: NEDC Results in Fuel Order
6.8 6.8
Fuel Consumption
Fuel Consumption
6.7 6.7
(l/100km)
(l/100km)
6.6 6.6
6.5 6.5
6.4 6.4
6.3 6.3
B0
B30
B10
B50
B0
B50
B30
B10
B10
B30
B50
B50
B30
B10
B30
B30
B30
B50
B50
B50
B50
B0
B0
B0
B0
B0
B0
B0
B0
B10
B10
B10
B10
B30
From work carried out by PMP, it is known that the extent of particle fill on the DPF
can affect results with measured PN decreasing as the DPF fills [20,21]. However,
there was little scope to take this into account within the test protocol without severely
increasing the length of the study. From the analysis of the PN emissions [1], no
significant discontinuities were observed between PN measurements made before
and after a DPF regeneration event.
14
report no. 6/14
Particle number (PN), both solid (carbonaceous) and total particle counts, for all
cycles and the particle size distribution for the steady state tests [1];
Analysis of PM composition from selected test conditions. The composition
analysis included the percentage of Soluble Organic Fraction (SOF) by Gas
Chromatography (GC); sulphate and nitrate anions by Ion Chromatography (IC);
and elemental carbon (EC) by difference. The fuel- and lube-derived fractions of
the SOF were also estimated by GC. The results of these measurements are
included in [1];
Analysis of carbonyl (aldehyde and ketone) unregulated emissions from selected
cycles [22].
Table 3 summarises all of the measurements that were completed over each test
condition.
The total number of solid particles was measured according to the PMP protocol. A
schematic and a full description of the equipment and protocol are given in [20,21].
This report focuses on the results obtained from both FC and regulated emissions
measurements, including the analysis of NO/NOx ratio as a tool for understanding
catalyst and vehicle effects. The composition of filtered PM samples and unregulated
emissions, including PN, PN size distribution, and carbonyl emissions, are
summarized in [1].
This section describes the statistical methods used to analyse the data from this
programme. These are similar to those used in earlier Concawe studies [13,14,15,16].
More details can be found in Appendix 5.
15
report no. 6/14
Each FC and emissions measurement (CO2, HC, CO, NOx, NO, NO/NOx, and PM)
was examined separately on a vehicle-by-vehicle and cycle-by-cycle basis.
In the EPEFE gasoline project [23] and in the other emission and particulate studies
cited above, the variability in emissions and particulate measurements has typically
been found to follow a lognormal distribution with the degree of scatter increasing as
the emission level increases.
The standard deviation vs. mean plots in Appendix 7 show that HC, CO, NOx, NO,
and PM measurements behaved in this way in the present study in those cycles where
there were noticeable differences in emissions between fuels (note these plots were
plotted before outliers had been removed or the results had been trend corrected).
This assumption is harder to verify rigorously for CO 2 and FC because the
measurements differ little from fuel to fuel in absolute terms on any particular vehicle.
Nevertheless, all subsequent statistical analyses are based on the assumption of
lognormality because the physical mechanisms suggest that this is the most plausible
model for emissions (and therefore FC) data.5
The data were examined for outliers by inspecting studentized residuals (residuals
divided by their Standard Errors (SE)). Less than 1% of the measured emissions
values were rejected (see Appendix 5 for details).
Consistent with previous Concawe studies, arithmetic means have been used to
summarize all fuel consumption, gaseous emission and PM measurements, despite
the lognormality in the data. This is because logarithmic (i.e. geometric means)
underestimate total emissions to the atmosphere (see Appendix 5). Therefore in the
various plots in subsequent sections, all measurements are plotted on the original
g/km or l/100km scale. SEs and error bars were computed using weighted analysis
of (co)variance techniques (see Appendix 5).
In the bar charts presented in Section 4, the error bars show the mean value ±1.4 x
SE of the mean value. The factor 1.4 in this equation was selected for consistency
with EPEFE [23] and with previous Concawe studies [13,13,14,16]. Emissions from
two fuels will not be significantly different from one another at P<5% 6 unless there is
a sizeable gap between their error bars.
Statistically significant time trends (at P<1% 7) were found in 228 of the 114 data sets
(3 vehicles × 7 measurements × up to 6 cycles9) for HC, CO, NOx, NO, PM, CO2, and
FC. These sets of data points can be adjusted to what they might have been if all of
the tests had been conducted halfway through the test programme using statistical
“analysis of covariance” techniques. Trends were corrected on the natural and not the
logarithmic scale using an appropriately weighted analysis (see Appendix 5).
Both the uncorrected and corrected means are shown in Appendix 6. It can be seen
that trend correction generally had little effect on mean emissions because a
statistically robust randomised block design was used to set the test order (see
16
report no. 6/14
Section 3.1). Nevertheless, the correction reduced SEs and error bars helped to
improve the discrimination between different fuels. For this reason, corrected means
are shown for these variables in Figures 8 to 15.
17
report no. 6/14
Appendix 10 shows the actual engine speed record from each vehicle’s Engine
Control Unit (ECU) regarding calculated engine load (as provided by the On-Board
Diagnostics (OBD)). As expected, for every vehicle, the engine speed following the
predefined cycle profile and gearshift strategy did not vary with different fuels.
However, a clear increase in the average calculated engine load (as reported by the
OBD) is observed with different fuels for all cycles (Figure 7). In this figure, the B0
fuel is shown in blue (bottom) while the B50 fuel is shown in red (top). The engine
management system (EMS) calculates the load on the basis of the fuel volume
injected into the cylinder.
Figure 7 Calculated engine load from the ECU for Vehicle 2 over the NEDC
The reduction in the fuel’s energy content with increasing RME resulted in an increase
of the fuel volume injected into the cylinder. This is interpreted by the ECU as an
increase in the engine load (by means of the ECU-calculated engine load). However,
the power output (actual load) does not change.
Table 4 summarizes the engine load data recorded from the vehicle’s ECU. The B10
fuel caused an increase of about 1% in the ECU-calculated engine load reflecting the
lower volumetric energy content of this B10 blend. The corresponding reductions were
about 3% for B30 and about 5% for B50. The reductions for the B10 and B30 fuels
are considered to be small and most probably do not affect the engine control
strategy. It is possible, however, that the increase in the ECU-calculated engine load
with B50 is large enough to affect the EMS resulting in slightly different operating
18
report no. 6/14
The primary objective of this study was to determine whether modern vehicles
improve their efficiency when running on FAME/diesel fuels. If this were to occur, the
volumetric FC should be lower than would be predicted from the volumetric Lower
Heating Value (VLHV) of the RME/diesel fuel blend. Since the change in VLHV is not
large, higher RME contents and a rigorous test protocol were used in this study as
described previously.
For planning purposes, however, the vehicle’s efficiency or energy consumption (in
MJ/100km) is more important so that CO2 emissions can be evaluated over the entire
WTW fuel production and use chain. Because GHG emissions are linked to energy
consumption, CO2 emissions from the vehicle also depend on the composition of the
fuel, especially the hydrogen to carbon (H/C) ratio of the fuel. Energy consumption is
generally more informative for comparing the WTW performance of different fuels in
vehicles over a prescribed driving cycle.
In this study, tailpipe emissions measurements were made for CO2, CO and HC10.
These measurements, together with the H/C of the test fuel, allowed the FC to be
calculated by carbon balance (see Appendix 8 for details). Because the RME
contains oxygen, its energy content per litre (or per kilogram) is lower than the energy
content of a litre (or kilogram) of fossil diesel fuel. Therefore, for the same vehicle
efficiency, the volume of fuel needed to complete a given driving cycle will be slightly
higher for an RME-containing fuel than it will be for a fossil diesel fuel.
To calculate the energy consumption or vehicle efficiency, the LHV of the fuel blend
must also be known by direct measurement. Because all of the fuel and vehicle
measurements have some degree of experimental uncertainty, a carefully designed
study must be completed with multiple repeat tests in order to unambiguously identify
10 In comparison to the CO2 emissions, the CO and HC emissions are very small.
19
report no. 6/14
Although CO2 emissions are important and can be measured directly, they cannot be
directly compared with energy consumption if the fuel compositions are different. As
was the case in this study, however, the variation in H/C ratio between different fuels
was not very large so the correlation between CO2 emissions and energy
consumption is quite close.
Figures 8 and 9 show the CO2 emissions and FC, respectively, for all vehicles and
all driving cycles. In all of the bar charts shown in this report, the error bars represent
±1.4 x SE of the mean value for the indicated fuel and vehicle (see Section 3.2 for an
explanation). Emissions from two different fuels will not be significantly different from
one another at P<5% unless there is a sizeable gap between their error bars.
Figure 8 Average CO2 emissions for all vehicles and driving cycles
350
B0
300 B10
B30
250
CO2 Emissions (g/km)
B50
200
150
100
50
0
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
20
report no. 6/14
12 B0
B10
Fuel Consumption (l/100km)
10 B30
B50
8
0
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
UDC EUDC NEDC Artem Artem 120kph
cold Urban Road
As shown in Figures 8 and 9, both the CO2 emissions and the calculated volumetric
FC increase with increasing RME content as expected from the slightly lower energy
content of the RME/diesel blends. The same trends are observed for all vehicles and
cycles. Repeat measurements of CO2 emissions on the same vehicle, fuel, and
driving cycle varied by up to ± 3% about their mean which was within the range of
expected variation. It is very clear, however, that the changes in CO 2 emissions and
volumetric FC are much larger between different driving cycles and vehicles than they
are from changes in the fuel composition.
Focusing on the fuel variations only, however, the CO2 emissions generally increased
as the RME content increased. The CO2 emissions with the B50 fuel were on the
order of 1 to 3% higher than for the B0 fuel in most cases. All vehicles responded in
a similar way for CO2 emissions and volumetric FC with increasing RME content.
Over the NEDC, the vehicle is driven according to a prescribed cycle of speed versus
time (see Figure A4-1 in Appendix 4). For fuels having different LHVs, this means
that different masses of fuel will be consumed over the regulatory NEDC and
converted to CO2 exhaust emissions through combustion.
As the RME content increases, the mass of fuel consumed increases because the
LHV of the RME (and therefore the LHV of the RME/diesel blend) decreases. Much
of this additional mass, however, comes from the oxygen in the RME which does not
affect CO2 emissions directly.
On the other hand, CO2 emissions are affected by the fraction of fuel energy that
comes from carbon compared to hydrogen, that is, the mass H/C ratio of the fuel. For
the fuels used in this study, the mass H/C ratio varied very little so we would not
expect to see any differences in CO2 emissions between fuels if the energy efficiency
of the vehicle did not change when running on different fuels.
21
report no. 6/14
This analysis was completed for the NEDC and the results are shown in Figure 10.
The left-hand charts show the measured FC (in l/100km) versus the VLHV (in MJ/l)
of the test fuel for all three vehicles over the NEDC. The error bars show 95%
confidence limits for the average FC values plotted and the solid line is a best fit
through the data points. The correlation between the FC and VLHV is obviously very
good.
In the right-hand figures, the percent change in FC is plotted versus the percent
change in [1/VLHV] using the B0 results for each vehicle as the basis. The error bars
again represent 95% confidence limits and the solid line is a best fit through the data
points. The slope of this best fit line is also reported in each figure. The dotted lines
are the 95% confidence limits around the best fit line and the dashed line is a one-to-
one correlation line.
Vehicle 1 Vehicle 1
7.9 9
Fuel Consumption
7.8
% Change in FC
7
7.7 B50
7.6 B30 5 y = 1.545x
B50
7.5 B30
3
7.4 B10
B0 1 B10
7.3
B0
7.2 -1
0 1 2 3 4
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4
% Change in (1/VLHV)
Volumetric Lower Heating Value
Vehicle 2 Vehicle 2
6.6 9
Fuel Consumption
% Change in FC
6.5 7
6.4 B50
5 y = 1.406x B50
6.3 B30
B10 3
6.2 B30
B0 B10
6.1 1
B0
6.0 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Volumetric Lower Heating Value % Change in (1/VLHV)
Vehicle 3 Vehicle 3
6.8 9
Fuel Consumption
B50
% Change in FC
7
6.7
B30 5
6.6 y = 0.917x
3 B50
B10 B0
6.5 B30
1
B0 B10
6.4 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Volumetric Lower Heating Value % Change in (1/VLHV)
22
report no. 6/14
The best fit lines lie above the one-to-one correlation lines for Vehicles 1 and 2,
suggesting that the engine efficiency over the NEDC is actually lower on the higher
RME fuels. However, these correlation lines are not outside the 95% confidence
limits. This means that the change in FC can be simply explained by the loss in energy
content of the RME-containing fuels, within the experimental uncertainty inherent in
the measurements.
Within the statistical precision, the data show that the volumetric FC increases in
direct proportion to the decrease in the VLHV of the RME/diesel fuel blends over the
NEDC. There is no evidence that the use of RME as a blending component provides
an engine efficiency benefit and the vehicles are not able to compensate for the lower
energy content of the FAME/diesel blends through better engine efficiency on the
oxygenated fuels. Although a recent study [27] suggests that future engines could be
further engineered to adapt to differences in fuel composition, this is clearly not the
case with the vehicles tested in this study.
Similar trends in FC vs. VLHV were found for the same vehicles and fuels over the
ARTEMIS Urban cycle (Figure 11). Comparable data for the UDC and EUDC portions
of the NEDC and for the 120 km/h steady-state condition are shown in Appendix 9.
23
report no. 6/14
Figure 11 Change in FC as a function of the fuel’s VLHV over the ARTEMIS Urban
cycle
Vehicle 1 Vehicle 1
12.2 9
12.1
Fuel Consumption
12.0 B50 7
% Chnage in FC
B30 y = 1.7067x
11.9
5 B50
11.8 B10 B30
11.7
3 B10
11.6
11.5 B0 1
11.4 B0
11.3 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Vehicle 2 Vehicle 2
10.0 9
9.9
Fuel Consumption
7
% Chnage in FC
9.8
B50
9.7 5
B30 y = 1.3686x
9.6 B50
9.5 3 B30
B10
9.4 B0 1 B10
9.3 B0
9.2 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
% Change in (1/VLHV)
Volumetric Lower Heating Value
Vehicle 3 Vehicle 3
10.7 9
Fuel Consumption
10.6
7
% Chnage in FC
10.5 B50
10.4 5
B30
10.3 B0 3
10.2 B10 y = 0.4868x B50
1
10.1
B0 B30
10.0 -1 B10
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
24
report no. 6/14
This section presents the results from the regulated emissions measurements. In
each case, the absolute average values recorded over each cycle are provided with
error bars that represent 1.4 x the standard error of the measurements.
Figure 12 shows the average NOx emissions measured for each vehicle over all
driving cycles and demonstrates the effect of driving cycle, vehicle, and fuel on NOx
emissions.
Figure 12 Average NOx emissions for all vehicles and driving cycles
2.0
1.8 B0
1.6
B10
B30
1.4
NOx Emissions (g/km)
B50
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
The Euro 5 and 6 emissions regulations may present some engine calibration and
aftertreatment challenges, especially with respect to controlling NOx emissions under
all operating conditions. NOx emissions are the sum of NO and NO2 in the exhaust
gas. Although diesel engines emit mainly NO, the aftertreatment system, especially
the DOC, increases the oxidation of the engine-out NO to NO2. The NO emissions
and NO/NOx ratio are discussed in Section 4.4.
From the NOx emissions results shown in Figure 12, it is observed that RME has a
consistent effect on NOx emissions over all cycles and in all three vehicles. The NOx
emissions increased linearly with increasing RME content in the fuel blend. For most
vehicles and cycles, the increases were in the 1-10% range but increases up to 20%
were observed in some casces for the B50 fuel.
Over the NEDC, all three vehicles responded in a similar way with the NOx emissions
increasing as the RME content of the fuel increased. Over the NEDC sub-cycles,
however, Vehicles 1 and 3 responded quite differently, with Vehicle 1 giving higher
25
report no. 6/14
NOx emissions over the UDC and Vehicle 3 giving higher NOx emissions over the
EUDC.
These differences were much larger than the effects found with changing RME
content. In Vehicle 2, however, the NOx emissions over the UDC were essentially
unchanged with increasing RME content but did increase slightly over the EUDC with
increasing RME content. Since the differences found in the sub-cycles gave similar
results for all vehicles across the NEDC, it is possible that the sub-cycle differences
are associated with different EMS strategies in all three vehicles.
Over the ARTEMIS cycles, all vehicles gave broadly similar trends with respect to
NOx emissions. Significant differences in NOx emission levels were observed
between legislated and non-legislated cycles; in some cases, emissions over the
ARTEMIS cycles were almost 10 times higher than over the NEDC.
Overall, the effect of increasing RME content on NOx emissions from the light-duty
diesel vehicles in this study is consistent with that reported previously for heavy-duty
vehicles [9]. Literature relating to the effect in light duty applications is less consistent,
although generally it is observed that NOx emissions increase with increases in FAME
content. However, it has also been reported that a reduction in NOx emissions was
observed as the FAME content of the diesel fuel increased [24].
4.3.2. PM Emissions
Figure 13 shows the Particulate Mass (PM) emissions for all vehicles and for three
driving cycles. PM emissions were not measured over the NEDC sub-cycles or over
the 120kph steady-state condition.
Figure 13 Average PM emissions for all vehicles and applicable driving cycles
0.06
B0
0.05 B10
B30
B50
PM Emissions (g/km)
0.04
0.03
0.02
0.01
0
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
26
report no. 6/14
A single filter was used to collect the PM over each complete cycle, including the
NEDC. From Figure 13, it can be seen that the absolute PM emissions from the DPF-
equipped Vehicles 1 and 3 were very low and in some cases close to the measured
background. No statistically significant differences were observed from these vehicles
with changes in the RME content.
For Vehicle 2, the PM emissions were more than 10 times higher than those from the
DPF-equipped vehicles and decreased with increasing RME content. Over the NEDC,
reductions of up to 50% were observed with the B50 fuel. Increasing the RME content
was somewhat more beneficial over the NEDC than over the ARTEMIS cycles.
27
report no. 6/14
The carbon monoxide (CO) and hydrocarbon (HC) emissions from each vehicle are
shown in Figures 14 and 15, respectively.
B0
0.4
B10
B30
CO Emissions (g/km)
0.3
B50
0.2
0.1
0.0
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
UDC EUDC NEDC Artem Artem 120kph
cold Urban Road
0.06 B0
B10
0.05
B30
HC Emissions (g/km)
0.04 B50
0.03
0.02
0.01
0.00
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
As shown in these figures, the absolute emissions levels for both HC and CO were
very low for all vehicles and considerably lower than the Euro 4 regulatory limits. For
all hot start cycles, the absolute values for HC were very close to background levels
and no statistically significant differences were seen between fuels.
28
report no. 6/14
The highest concentrations of both CO and HC emissions were seen over the cold
start UDC, with lower CO and HC emissions observed over the EUDC after the DOC
had fully warmed up. Thus, over the NEDC, the CO and HC emissions are primarily
due to emissions occurring over the UDC. No significant differences in the HC
measurements were observed over the EUDC. For all vehicles, RME significantly
increased the CO emissions over the cold start UDC while emissions were low over
the EUDC with no clear patterns. For a few driving cycles, the increase in CO and HC
emissions was statistically significant between the B0 and B50 fuels.
Over the ARTEMIS Urban and Road cycles, Vehicle 3 showed higher CO emissions
than Vehicles 1 and 2. A significant reduction in CO emissions was observed with
increasing RME content in this case and a smaller effect was observed for HC
emission. For this vehicle, RME appeared to have a positive effect on both CO and
HC emissions.
As previously mentioned, the DPF-equipped vehicles were tested before starting the
test programme in order to understand their DPF regeneration frequency. This was
important to ensure that no DPF regeneration events occurred during a single day’s
testing. The DPF-equipped vehicles were constantly monitored for possible
regeneration events that could occur during the measurements and, when they were
found to be close in mileage to the next regeneration event, they were ‘forced’ into a
regeneration event following the completion of the day’s testing. Thus, all
measurements taken over the complete programme were unaffected by DPF
regeneration.
In the test sequence followed, the DPF in Vehicle 1 regenerated three times while the
DPF in Vehicle 3 regenerated twice. Typical evidence of a regeneration event
compared to standard operation is provided in Figures 16 and 17.
29
report no. 6/14
Figure 16 Catalyst temperatures for a DPF regeneration event (23-2 series) for
Vehicle 1 compared to standard (average) operation over 3 driving cycles
Catalyst IN 23-2
1000
Catalyst OUT 23-2
900 Catalyst IN Average
Catalyst OUT Average
800
Temperature ( C)
700
o
600
500
400
300
200
100
NEDC Ar. Urban Ar. Road
0
0 500 1000 1500 2000 2500 3000 3500
Time (sec)
Figure 17 Exhaust temperature during a DPF regeneration event (Texhaust 3-6) compared
to standard operation (Texhaust) for Vehicle 3
700
500
Temperature (oC)
400
300
200
100
0
0 500 1000 1500 2000 2500 3000
Time (sec)
From previous PMP work, it is known that the extent of particle fill on the DPF can
affect results with measured particle numbers decreasing as the DPF fills [20].
However, within the test protocol there was little scope to take this into account without
severely extending the length of the study.
30
report no. 6/14
4.4.2.1. NO Emissions
During the study, NO emissions were measured separately from the legislated bag
emissions for NOx. In a vehicle without a DOC, the majority of the NOx emissions are
emitted as NO. In the presence of a DOC, however, a larger fraction of the NO will be
oxidised to NO2 and emitted at the tailpipe. These NO2 emissions are of increasing
concern for local air quality and human health.
The measurement of NO and also the investigation of the NO/NOx ratio allowed the
effects due to the different aftertreatment technologies to be measured and compared
when operating on different RME/diesel blends.
Figure 18 shows the NO emissions measured from each vehicle over all driving
cycles. NO emissions data were not collected at the steady-state conditions.
0.5
B0
B10
0.4 B30
B50
NO emissions (g/km)
0.3
0.2
0.1
0
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
From this figure, it can be seen that the effect of increasing RME concentration on
NO emissions follows similar trends to those already presented in Section 4.3.1 for
NOx emissions. Over the UDC, the NO emissions were similar for Vehicles 1 and 3
while they were much higher for Vehicle 3 over the EUDC. Consequently, the NO
emissions for Vehicle 3 were also higher over the NEDC. Compared to the NEDC and
its sub-cycles, the NO emissions were much higher over the ARTEMIS cycles and
more sensitive to changes in the RME concentration.
31
report no. 6/14
on local air quality and on human health. In the framework of these measurements, it
was attempted to investigate whether there was any impact of RME content on the
NO/NOx ratio. Figure 19 shows the normalized NO/NOx ratio as measured over the
legislated cycles. The NO and NO/NOx ratio data are tabulated in Appendix 6. The
NO/NOx ratio is not shown for the 120 kph steady-state because the NO emissions
were not measured at this condition.
Figure 19 NO/NOx emissions ratio for all vehicles and driving cycles
0.6
B0
B10
0.5
B30
B50
0.4
NO/NOx
0.3
0.2
0.1
0
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
Veh. 1
Veh. 2
Veh. 3
As shown in Figure 19, the NO/NOx ratio depends strongly on the vehicle and driving
cycle but there is very little effect of RME concentration. Over the EUDC (hot portion),
all vehicles show a trend to lower NO/NOx ratio with increasing biodiesel
concentration. This trend is most pronounced for Vehicle 3 with the NO/NOx ratio over
the EUDC being approximately 10% lower for the three biodiesel blends compared to
the reference fuel. As a result, the NO/NOx ratio is about 8% lower over the NEDC
for this vehicle indicating a shift towards slightly more NO2 when biodiesel is present.
Analysis of the NO from the regulated bag emissions procedure is not the best way
to measure the NO/NOx ratio, however, because the elapsed time between bag
sampling and analysis can change the NO concentrations in a way that is not
repeatable or controllable. For this reason, the average instantaneous NO/NOx ratio
was also measured using a fast response NOx analyser in addition to the bag
emissions measurements. The results of these measurements are shown in Figure
20.
32
report no. 6/14
Figure 20 Average instantaneous NO/NOx ratio over NEDC as recorded using the fast
NOx analyzer for B0 and B50 fuels
1.2
.
1 B0: 0.43
Diesel B50: 0.46
0.8
0.6
0.4
0.2
0
0 200 400 600 800 1000
time (s)
Vehicle 1
1
Average NO/NOx ratio (concentrations) .
Average NO/NOx :
0.9
B0: 0.57
0.8 B50: 0.58
0.7
0.6
0.5
0.4
0.3
0.2
B50
0.1 Diesel
0
0 200 400 600 800 1000
time (s)
Vehicle 2
1.6
.
Average NO/NOx:
Average NO/NOx ratio (concentrations)
1.4
B0: 0.72
1.2 B50: 0.66
0.8
0.6
0.4 B50
0.2 Diesel
0
0 200 400 600 800 1000
time (s)
Vehicle 3
33
report no. 6/14
Generally, the results of the analysis made with the fast NOx analyser are consistent
with those from the bag samples. The instantaneous measurements indicate that the
use of B50 does not affect the NO/NOx ratio in the case of Vehicles 1 and 2. However,
the use of B50 on Vehicle 3 led to an approximately 8.5% lower NO/NOx ratio over
the NEDC compared to the B0 fuel. The differentiation between the B0 and the B50
NO/NOx ratio begins at the start of the EUDC (800 seconds) and onwards.
One explanation for this observation may be related to the exhaust aftertreatment
systems of the 3 vehicles. RME, due to its oxygen content, tends to enhance the
oxidation of certain pollutants such as CO, HC, PM, and NO. However, the oxidation
catalysts on the first two vehicles masked these oxidation effects in the case of CO
and HC (Section 4.3.3). The difference in exhaust gas oxidation in Vehicle 3 may be
due to a smaller DOC, a different type of catalyst, or different optimisation to the other
vehicles. In the CO and HC case, biodiesel use led to reductions in these pollutants.
It is therefore expected that increasing FAME use in this vehicle would have the same
impact on NO, increasing its conversion to NO2 and lowering the NO/NOx ratio, as is
observed. Although the difference is not large, more study on NO/NOx engine-out
emissions with FAME/diesel blends would be appropriate in order to better
understand the effects of engines, fuel, and aftertreatment system.
It is possible that the presence of the RME interacts with the DOC in some way, either
by affecting the light-off period (which would be especially noticeable over the UDC)
or by affecting the efficiency of the catalyst as it warms up. Interactions between fuel
species and catalyst operation have been seen before. From the EPEFE programme
[23], it was shown that the removal of 2-3 ring PAH from the fuel resulted in a reduction
in catalyst efficiency. From this observation, it was surmised that higher molecular
weight hydrocarbons are oxidised more easily compared to lower molecular weight
molecules. While the effect seen here suggests the opposite with respect to higher
molecular weight hydrocarbons, the presence of an oxygenated fuel may also play a
part. These interactions are not clearly understood and more information on the
interaction between fuel components and catalyst performance is needed in order to
fully interpret the results.
34
report no. 6/14
Figure 21 Normalized HC vs. normalized CO emissions over the UDC cold start cycle
2
Normalised HC emissions
1.5
0.5
0.5 1 1.5 2
Normalised CO emissions
35
report no. 6/14
5. CONCLUSIONS
The primary objective of this study was to determine whether modern vehicles
improve their efficiency when running on RME/diesel fuel blends. If this were to occur,
the volumetric FC should be lower than predicted from the lower energy content (LHV)
of the RME/diesel fuel blend.
This study shows that the volumetric FC is inversely proportional to the volumetric
energy content of the RME/diesel fuel blends in all vehicles and over the NEDC and
ARTEMIS Urban cycles. Comparable data shown in Appendix 9 show that the same
conclusion applies to the UDC, EUDC, ARTEMIS Road, and the 120 km/h steady-
state condition. In other words, the vehicles were not able to compensate for the
energy contents of RME/diesel blends through better engine efficiency on the
oxygenated fuels.
As anticipated, the CO2 emissions and volumetric FC increase with increasing RME
content. The same trends are observed for all vehicles and cycles, where vehicle and
cycle differences are much larger than fuel effects. Focusing on the fuel effects only,
however, the CO2 emissions generally increased linearly with the RME content, the
increase for the B50 fuel being between 1-3% in most cases. These conclusions are
based on measurements in which DPF regenerations were specifically avoided and
the results could be different under ‘real world’ conditions. Figure 22 shows the
average percentage change from baseline in CO 2 emissions and volumetric FC for
the RME containing fuels evaluated in this study.
Fuel Consumption
4.0
CO2
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
B10 B30 B50
36
report no. 6/14
Regulated Emissions
RME appears to have a consistent effect on NOx emissions over all cycles in all
three vehicles with NOx emissions increasing linearly with increasing RME content in
the fuel blend. The NOx emissions increase up to 20% with B50 in some cases with
most differences being in the 1-10% range depending on the fuel blend and vehicle.
37
report no. 6/14
Figure 23 Average percentage change in regulated emissions for all vehicles over the
NEDC
-20
-40
-60
B10 B30 B50
The NO/NOx ratio depends strongly on the vehicle and driving cycle but there is very
little effect of RME concentration. A small trend to lower NO/NOx ratio with increasing
biodiesel concentration is most pronounced in Vehicle 3 with the NO/NOx ratio over
the EUDC being about 10% lower for the three biodiesel blends compared to the B0
fuel. For this vehicle, a small shift towards more NO2 is observed when biodiesel is
employed which is believed to be due to its DOC performance. Observations made
from the bag emissions analyses were confirmed with an on-line analyser that
measured instantaneous NOx.
In most cases, CO and HC differences with RME are of about the same order in
relative terms. Thus it is concluded that RME has a generally uniform effect on tailpipe
CO and HC, again possibly implying a global impact on DOC performance.
38
report no. 6/14
6. GLOSSARY
CO Carbon Monoxide
DF Dilution Factor
FC Fuel Consumption
39
report no. 6/14
HC Hydrocarbon
HD Heavy-duty
LD Light-duty
NO Nitric Oxide
PN Particle Number
40
report no. 6/14
TC Turbo Charged
41
report no. 6/14
7. ACKNOWLEDGEMENTS
All those at LAT who worked enthusiastically for this project and particularly to:
Mr. Theodoros Tzamkiozis, Mr. George Vagiatas, Mr. Christos Samaras, Mr.
Argirios Tzilvelis, Mr. Thanassis Papazacharias and Mrs. Kousoulidou for her
remarks on the analysis of the results;
Mr. Evangelos Georgiadis for his valuable remarks;
The Association for Emissions Control by Catalyst (AECC), Brussels, Belgium,
for providing Vehicle 3 (the ‘golden vehicle' [8]) used in this study;
Mr. Samaras and Mr. Gouzelis of Hellenic Petroleum SA for helping with the fuel
storage;
Professor Stamoulis Stournas, National Technical University of Athens, Athens,
Greece, for analytical measurements on the biodiesel fuels;
Afton Chemical Company for supplying the HiTEC® 4678 diesel performance
additive package.
While this project was being completed, we were saddened to hear that our
collaborator, Professor Stam Stournas at the National Technical University of Athens,
had passed away. He was a valued contributor to our project, as well as to European
research on fuels and vehicles and his many contributions will be missed.
42
report no. 6/14
8. REFERENCES
3. CEN (2009) Automotive fuels – diesel – requirements and test methods. EN 590.
Brussels: Comité Européen de Normalisation
4. CEN (2009) Automotive fuels – fatty acid methyl esters (FAME) for diesel engines –
requirements and test methods. EN 14214. Brussels: Comité Européen de
Normalisation
6. Agarwal, A.K. (2007) Biofuels (alcohols and biodiesel) applications as fuels for
internal combustion engines. Progress in Energy and Combustion Science 33, 3,
233–271
7. Graboski, M.S. and McCormick, R.L. (1998) Combustion of fat and vegetable oil
derived fuels in diesel engines. Progress in Energy & Combustion Science 24, 2, 125-
164
10. JEC (2014) Well-To-Wheels analysis of future automotive fuels and powertrains in
the European context. Version 4a, Report EUR 26236 EN. Brussels : JRC, EUCAR,
CONCAWE
11. Wood Mackenzie (2009) Food and Fuel: The Outlook for Biofuels in 2020
12. de Haan, P. and Keller, M. (2001) Real-world driving cycles for emission
measurements: ARTEMIS and Swiss cycles. BUWAL-Bericht SRU Nr. 255. Bern:
Swiss Agency for Environment, Forests and Landscape
13. CONCAWE (2002) Evaluation of diesel fuel cetane and aromatics effects on
emissions from euro-3 engines. Report No. 4/02. Brussels: CONCAWE
14. CONCAWE (2005) Fuel effects on the characteristics of particle emissions from
advanced engines and vehicles. Report No. 1/05. Brussels: CONCAWE
15. CONCAWE (2005) Fuel effects on emissions from advanced diesel engines and
vehicles. Report No 2/05. Brussels: CONCAWE
43
report no. 6/14
16. CONCAWE (2009) Comparison of particle emissions from advanced vehicles using
DG TREN and PMP measurement protocols. Report No. 2/09. Brussels: CONCAWE
17. EU (1998) Directive 98/69/EC of the European Parliament and of the Council of
13 October 1998 relating to measures to be taken against air pollution by emissions
from motor vehicles and amending Council Directive 70/220/EEC. Official Journal of
the European Communities No. L350, 28.12.1998
18. EU (2000) Directive 1999/96/EC of the European Parliament and of the Council of 13
December 1999 on the approximation of the laws of the Member States relating to
measures to be taken against the emission of gaseous and particulate pollutants from
compression ignition engines for use in vehicles, and the emission of gaseous
pollutants from positive ignition engines fuelled with natural gas or liquefied petroleum
gas for use in vehicles and amending Council Directive 88/77/EEC. Official Journal of
the European Communities No. L44, 16.02.2000
19. EU (2007) Regulation (EC) No 715/2007 of the European Parliament and of the
Council of 20 June 2007 on type approval of motor vehicles with respect to emissions
from light passenger and commercial vehicles (Euro 5 and Euro 6) and on access to
vehicle repair and maintenance information. Official Journal of the European
Communities No. L171, 29.06.2007
20. Andersson, J. et al (2006) Update on the PMP Phase 3 Light Duty inter-laboratory
correlation exercise. 10th ETH Conference, August 2006. Zurich: Swiss Federal
Institute of Technology
22. US EPA (1999) Compendium of methods for the determination of toxic organic
compounds in ambient air. Determination of formaldehyde in ambient air using
adsorbent cartridge followed by High Pressure Liquid Chromatography (HPLC).
Compendium Method TO-11A. EPA/625/R-96/010b. Cincinnati, OH: Center for
Environmental Research Information
23. EPEFE (1995) European programme on emissions, fuels and engine technologies.
EPEFE Report on behalf of ACEA and EUROPIA
25. Rose, K.D. et al (2010) Impact of Biodiesel Blends on Fuel Consumption and
Emissions in Euro 4 Compliant Vehicles. SAE Paper No. 2010-01-1484. Warrendale,
PA: Society of Automotive Engineers
26. Barnett, V. and Lewis, T. (1994) Outliers in Statistical Data, 3rd edition. New York: J.
Wiley & Sons
27. Ricardo (2011) Biodiesel: improved fuel consumption, reduced emissions. Ricardo
Quarterly Review Q4 2011, 16-19
44
report no. 6/14
28. EU (2009) Directive 2009/30/EC of the European Parliament and of the Council of 23
April 2009 amending Directive 98/70/EC as regards the specification of petrol, diesel
and gas-oil and introducing a mechanism to monitor and reduce greenhouse gas
emissions and amending Council Directive 1999/32/EC as regards the specification
of fuel used by inland waterway vessels and repealing Directive 93/12/EEC. Official
Journal of the European Communities No. L140, 05.06.2009
45
report no. 6/14
The effects of FAME on pollutant emissions (see Appendix 2), FC, and engine operation have
been reported in many studies. As one would expect, however, these studies have been based on
different engines, test protocols, and FAME types and concentrations. For this reason, it is difficult
to obtain from the already published literature a clear assessment of the impact of biodiesel fuels
on vehicle fuel consumption.
In an extensive review, Lapuerta et al [A1-1] concluded with a summary table covering the impacts
of biodiesel fuels on gaseous pollutants, power output and fuel consumption (see Appendix 2,
Table A2-1). This table presents the percentage of scientific studies reporting either increases,
decreases or no changes in various vehicle emissions and operating characteristics. The results
are qualitative because the different test conditions and protocols used in the studies did not permit
a quantitative comparison. In contrast to the EPA study results [9] shown in Figure 4, this table
shows that in some cases, such as for NOx emissions and engine efficiency, no straightforward
conclusion can be drawn.
Kousoulidou et al [A1-2] also performed an extensive literature review for the European
Environment Agency (EEA) on the impact of biodiesel fuels on pollutant emissions and FC. From
this review, FC was reported not to change for low concentration FAME blends and increased with
the reduction in the fuel’s energy content for higher FAME blends, consistent with the results from
our study.
In a study by Fontaras et al [A1-3], five different biodiesel blends were tested in order to examine
their impact on emissions and FC of a common-rail passenger car. Limited effects were observed
on CO2 emissions, with only two of the five biodiesel blends (containing PME and SUME) providing
statistically significant differences. However, these effects were rather small and in opposite
directions, and no global conclusion on the effect of biodiesel on tailpipe CO 2 emissions could be
drawn.
In addition, a recent study by Martini et al [A1-4] on two passenger cars did not lead to consistent
conclusions. Tests with B30 and B100 in a Euro 3 common-rail passenger car over the NEDC,
confirmed that the impact on emissions was greater with higher biodiesel content. In the same
study, however, the FC did not seem to be affected by FAME at low concentrations while the
increase in volumetric FC was limited to 3% for neat FAME (B100).
In addition to the reviews cited above, Concawe has performed its own review evaluating the
potential impact of FAME on fuel consumption. Of the 99 papers examined in this review, 22 papers
provided reliable information. When these papers were studied in detail, it was apparent that the
experiments were not designed to look specifically at changes in engine efficiency. A best attempt
has been made to analyse these papers for engine efficiency trends and, where appropriate, apply
calculations to the published data. Unfortunately, the lack of robust experimental design has led to
the resulting efficiency data being of questionable quality.
Our assessment of key papers from the published literature is summarised below.
Discriminating differences in FC measurements was not very good, particularly for small
changes in FAME content, e.g. between B0 to B10 [A1-14,A1-15,A1-16,A1-17].
46
report no. 6/14
Where studies claimed an engine efficiency benefit with increasing FAME content, the
results were generally associated with a lack of statistical discrimination between fuels
[A1-18,A1-19].
Analysing the data “second-hand” from these papers required calculations of energy
consumption or engine efficiency, but these relied on the LHVs that were reported in the
papers. Some of these LHVs were of questionable accuracy and represent another
potential source of error in the analysis [A1-20,A1-21,A1-22,A1-23].
Some papers reported that lower energy efficiency when using biodiesel could be
recovered by optimising the engine operation [A1-24,A1-25].
The overall conclusion from our literature review on FC effects was that no improvement in engine
efficiency has been observed from the use of FAME and FAME blends compared to fossil diesel
fuel that compensates for the reduction in energy content of the FAME/diesel blend. There is
certainly no evidence to suggest an improvement in engine efficiency while there is some evidence
to suggest that a reduction in engine efficiency is observed with FAME and with FAME/diesel
blends. In this respect, our assessment is consistent with that of Kousoulidou et al [A1-2] where
no improvements in engine efficiency were reported and the key controlling factor for the fuel
consumption of FAME blends was the change in the fuel’s LHV.
47
report no. 6/14
5. Horn, U. et al (2007) Neat Biodiesel Fuel Engine Tests and Preliminary Modelling.
SAE Paper No. 2007-01-0616. Warrendale, PA: Society of Automotive Engineers
Using a single cylinder research engine, rather large error bars (up to 10%) were
reported. Fossil diesel fuel was shown to have up to 31% lower FC compared to
biodiesel fuels. Both brake thermal efficiency graphs and calculations presented in
this paper showed that B100 had 5-14% higher energy consumption compared to
fossil diesel.
11 The information shown in italics is Concawe’s own assessment of key findings from the cited paper.
48
report no. 6/14
12. Sinha, S. et al (2005) Performance Evaluation of a Biodiesel (Rice Bran Oil Methyl
ester) Fuelled Transport Diesel Engine. SAE Paper No. 2005-01-1730. Warrendale,
PA: Society of Automotive Engineers
Unexpected trends and the lack of error bars raise some questions about these
results. The paper reported small improvements in thermal efficiency (1.5%-3%) at
full load. The paper also reported that the BSEC for all biodiesel blends was lower
than for fossil diesel and that B20 and B10 showed lower BSEC at all loads.
13. Corgard, D.D. et al (2001) Effects of alternative fuels and intake port geometry on
HSDI diesel engine performance and emissions. SAE Paper No. 2001-01-0647.
Warrendale, PA: Society of Automotive Engineers
This paper used a small bore DI diesel engine. The main interest of the study was air
flow and swirl effects and not biodiesel fuels. There is a mention in the text that no
change in BSFC was observed for 0%, 15% and 30% biodiesel blends, but this may
reflect the repeatability of the experimental method rather than any actual engine
efficiency effects.
14. Haas, M.J. et al (2001) Engine performance of biodiesel fuel prepared from soybean
soapstock: A high quality renewable fuel produced from a waste feedstock. Energy &
Fuels, 15, 5, 1207-1212
Using an HD engine, this paper mentioned that B100 had a worse BSFC than
reference certification diesel, but that the BSFC for B20 was the same as the
reference diesel. This may reflect the discrimination of the BSFC method, rather than
a real effect. There were no specific references to changes in engine efficiency.
49
report no. 6/14
(e.g. injection timing and EGR) on performance. Plots suggest that the thermal
efficiency was worse with increasing FAME content, but by only a very small amount.
The magnitude of errors was not reported making definitive statements difficult.
20. Grimaldi, C.N. et al (2001) Performance and emissions of common rail DI diesel
engine using fossil and different bio-derived fuels. SAE Paper No. 2001-01-2017.
Warrendale, PA: Society of Automotive Engineers
This paper reported consistent expectations for FAME giving rise to lower smoke
emissions and poorer fuel consumption. There was no evidence of better engine
efficiency with biodiesel fuels. The paper commented that the FC results and torque
were completely in line with the higher density and lower energy contents (LHV) of
the biodiesel blends.
21. Li, H. et al (2007) Study of Emission and Combustion Characteristics of RME B100
Biodiesel from a Heavy duty DI Diesel Engine. SAE Paper No. 2007-01-0074.
Warrendale, PA: Society of Automotive Engineers
Using an HD engine, this paper focused on emissions. The greater fuel consumption
observed for FAME (12%-15%) was attributed to its lower energy content (LHV) No
reference was made to engine efficiency and LHV values for the fuels tested were not
reported, thus no additional calculations could be applied to this paper’s results.
22. Kawano, D. et al (2006) Application of Biodiesel Fuel to Modern Diesel Engine. SAE
Paper No. 2006-01-0233. Warrendale, PA: Society of Automotive Engineers
50
report no. 6/14
Using a 4L, 4-cylinder engine, this paper did not mention efficiency measurements on
biodiesel fuels. Some BSFC data were presented but it was not possible to make any
definitive statements on energy efficiency without having LHV data on the test fuels.
24. Last, R.J. et al (1995) Emissions and performance characteristics of a 4-stroke, direct-
injected diesel engine fuelled with blends of biodiesel and low sulfur diesel fuel. SAE
Paper No. 950054. Warrendale, PA: Society of Automotive Engineers
This paper focused on an HD DI engine equipped with hydraulically actuated
electronically controlled (HEUI) unit injectors, a 2-way catalyst, and EGR. The paper
reported that the mass fuel consumption with biodiesel was worse than with fossil
diesel. When the engine was optimised for a specific fuel, then the mass FC was
similar to that found on fossil diesel.
25. Suryawanski, J.G. et al (2005) Effect of Injection Timing Retard on Emissions and
performance of a Pongamia Oil Methyl Ester Fuelled CI Engine. SAE Paper No. 2005-
01-3677. Warrendale, PA: Society of Automotive Engineers
Using a KIRLOSTAR TV1 4-stroke single cylinder engine, the brake thermal efficiency
was found to be similar for all blends of pongamia oil methyl ester compared to fossil
diesel at standard injection timing. With retarded injection timing (4°) and at lower
BMEP, the biodiesel showed lower energy consumption but this difference
disappeared as the BMEP increased.
26. Horn, U. et al (2007) Detailed Heat Release Analyses with Regards to Combustion of
RME and Oxygenated Fuels in an HSDI Diesel Engine. SAE Paper No. 2007-01-
0627. Warrendale, PA: Society of Automotive Engineers
This paper focused on the effect of EGR on key performance parameters. No fuel
consumption data were reported. Results were acquired from detailed heat release
analysis on a high-speed DI (HSDI) diesel engine. The energy conversion efficiency
for the heat release in the cylinder was calculated from the maximum heat release
and the supplied energy. This efficiency was slightly higher for European diesel fuel
compared to RME.
27. Bittle, J.A. et al (2010) Biodiesel Effects on Influencing Parameters of Brake Fuel
Conversion Efficiency in a Medium-duty Diesel Engine. J Eng Gas Turbines & Power,
132, 12, 122801-122810
Changes in fuel consumption were investigated at nine different operating conditions
for fossil diesel and B100 derived from palm oil. No changes in brake fuel conversion
efficiency were observed except at high loads conditions.
51
report no. 6/14
Compared to FC effects, many more studies have been reported on the impact of FAME on
regulated and unregulated emissions. For this reason, this Appendix addresses only a subset of
recent publications and does not claim to be a comprehensive literature survey. Some examples
are provided in each instance as representative of many more that address essentially the same
topic.
The effect of FAME type and concentration on regulated emissions has been widely studied. The
majority of publications are related to emissions from HD engines and have covered a wide range
of different test procedures and protocols. A range of biodiesel types have been tested, including
both animal and vegetable based components, both in their unreacted and also in their esterified
state. A range of concentrations have been examined up to and including 100% biodiesel on a
variety of engines (both commercial and research) and including a wide range of test conditions.
Because of this diversity in tests and results, it is difficult to develop a consensus from reading the
literature alone.
In 2002, however, the US EPA [A2-1] completed an analysis involving more than 800 sets of
emissions data. These data were collected from a range of studies which included all of the
variables mentioned above and a detailed statistical analysis was completed to summarise the
effects of biodiesel fuels on regulated emissions. Figure A2-1 is reproduced from this study and is
frequently cited to show the impact of increasing biodiesel content on emissions.
Figure A2-1 Average emission impacts of biodiesel for heavy-duty highway engines [A2-1]
Since this EPA study was reported, this graph has represented the most widely held view on the
effects of biodiesel on regulated emissions. It should be emphasized, however, that the study was
carried out only on heavy duty engines (primarily US engines) and on a variety of fuels and test
procedures (including hot start tests) and did not include engines equipped with after-treatment
52
report no. 6/14
There are also European publications that relate to heavy duty testing which predate this study.
Some gave results in line with those reported by the EPA [A2-1,A2-3,A2-4] while others [A2-5,A2-
6] reported differences, normally with respect to NOx emissions, where a reduction in NOx with
increasing FAME was reported.
Since 2002, there have been many more publications on heavy duty engine results, again including
most of the variables mentioned above. Short of carrying out another EPA-type statistical analysis,
it is difficult to compare the magnitude of the results published. However, it is possible to look at
the results ‘in the round’ and establish whether they agree with the directional trends already
reported in the EPA study.
From our evaluation, many publications were found that agree completely with the EPA trends [A2-
7,A2-8,A2-9,A2-10,A2-11] while others show some variation, again, most commonly with respect
to NOx [A2-12,A2-13]. One paper [A2-14] addressed the use of a catalytic converter on a heavy
duty engine and reported that its use resulted in a corresponding reduction in NOx as the biodiesel
concentration increased.
However, results from heavy duty engines, and especially engines typically used in the US market,
cannot necessarily be extrapolated to the European passenger car fleet in which common rail
engines dominate, exhaust aftertreatment systems such as oxidation catalysts and DPFs have
become standard equipment and the diesel fuel used as basic blendstock has a significantly higher
cetane number than in the US. Furthermore, the European certification test emphasises cold
engine starting conditions where the presence of biodiesel can have a different impact compared
to hot start conditions.
There are considerably fewer publications related to the effect of biodiesel in light duty applications.
Again, the available papers cover a wide range of variables with respect to FAME, concentration
of blends, engine conditions etc. One early paper [A2-4] presented limited results using neat RME
and concluded that all regulated emissions (including NOx) were reduced when FAME was used.
Another relatively early paper [A2-15] showed reductions in HC and CO and increases in NOx, but
no effect on PM.
In a recent and quite extensive review, Lapuerta et al [A2-24] summarised these results in the
following table regarding the impacts of biodiesel on gaseous pollutants, power output, and FC
(Table A2-1). The table shows the percentage of scientific studies reporting either increases,
decreases or no changes in various emissions and operation characteristics. The results are
qualitative because the different conditions and protocols used in the studies do not make a direct
comparison possible. In contrast to the EPA study, this table indicates that there are general trends
but, in certain cases such as for NOx and efficiency, no straightforward conclusion can be drawn.
53
report no. 6/14
Brake-specific fuel 98 2 - -
consumption
Thermal efficiency 8 80 4 8
NOx emissions 85 10 5 -
PM emissions 3 2 95 -
THC emissions 1 3 95 1
CO emissions 2 7 90 1
a
Many references included in this category have reported both increases and decreases depending on engine load
conditions, engine type, engine operation temperature, etc.
LAT/AUTh has also performed an extensive literature review [A2-23] for the EEA on the impact of
biodiesel on pollutant emissions and fuel consumption. The results confirmed that the use of
biodiesel results in higher NOx and lower PM emissions. The size of these effects is related to the
biodiesel concentration, the vehicle operating conditions, and the engine technology. The effect of
biodiesel on CO and HC is generally reported to be beneficial. However, only limited studies were
conducted on modern vehicles equipped with diesel oxidation catalysts or other exhaust
aftertreatment technologies, configurations which might change this picture. Because of this, the
picture regarding CO and HC might be significantly different for modern diesel passenger cars.
In a recent LAT study [A2-26], five different biodiesel blends were tested in order to examine their
impact on the emissions and consumption of a common-rail passenger car. Small effects were
observed on CO2 emissions, with only two (PME, SUME) of the five blends providing statistically
significant differences. However, these differences were rather limited and in opposite directions
and there was no global conclusion on the effect of biodiesel on tailpipe CO2 emissions. The effect
of biodiesel on HC and CO emissions was more prominent over the cold-start driving cycles where
the absolute HC and CO emission levels were higher. Over these cycles, B10 fuels resulted in
~25% higher HC and CO emissions than B0 diesel fuel. The ratios of HC and CO emissions over
cold-start and hot driving cycles were different for the B0 and B10 fuels. These results indicated
that DOC performance was different when biodiesel was used. The effect of different biodiesel
blends on NOx emissions was variable, ranging from –7% to +11% on average, depending on the
type of biodiesel feedstock.
In addition, a recent study from the Joint Research Centre [A2-26] on two passenger cars did not
lead to consistent conclusions. NEDC tests on a Euro 3 common-rail equipped car using B30 and
B100 fuels confirmed that the regulated emissions were higher with increasing biodiesel content.
The largest effect was observed when neat biodiesel (B100) was used, suggesting that the different
properties of the fuels resulted in a non-optimized engine operation, leading to significant increases
in certain pollutants, such as CO, HC and NOx. However, fuel consumption did not seem to be
affected by the presence of biodiesel at low concentrations while the increase in fuel consumption
was limited to 3% with neat biodiesel (B100). The same fuels, when used on a unit-injector
54
report no. 6/14
equipped Euro 3 car, led to completely different observations concerning the impact of biodiesel
on modern passenger cars. Therefore, the effect of biodiesel blends on NOx emissions from
passenger cars is not straightforward and appears to depend on feedstock, vehicle technology and
operating conditions.
55
report no. 6/14
REFERENCES TO APPENDIX A2
4. Staat et al (1995). The effects of rapeseed oil methyl ester on diesel engine
performance, exhaust emissions and long term behavior. SAE Paper No. 950053.
Warrendale, PA: Society of Automotive Engineers
6. Haas, M.J. et al (2001) Engine performance of biodiesel fuel prepared from soybean
soapstock: A high quality renewable fuel produced from a waste feedstock. Energy &
Fuels, 15, 5, 1207-1212
10. Krahl et al (2005). The influence of fuel design on the exhaust gas emissions and
health effects. SAE Paper No. 2005-01-3772. Warrendale, PA: Society of Automotive
Engineers
12. Sharp et al (2005). Heavy duty diesel engine emission tests using special biodiesel
fuel. SAE Paper No. 2005-01-3671. Warrendale, PA: Society of Automotive
Engineers
56
report no. 6/14
14. Peterson et al (1996). Emissions testing with blends of esters of rapeseed oil with and
without a catalytic converter. SAE Paper No. 961114. Warrendale, PA: Society of
Automotive Engineers
15. Schramm et al (1999). Emissions from a diesel vehicle operated on alternative fuels
in Copenhagen. SAE Paper No. 1999-01-3603. Warrendale, PA: Society of
Automotive Engineers
17. Grimaldi et al (2002). Common rail HSDI diesel engine combustion and emissions
with fossil/bio-derived fuel blends. SAE Paper No. 2002-01-0865. Warrendale, PA:
Society of Automotive Engineers
18. McGill et al (2003). Emission performance of selected biodiesel fuels. SAE Paper
No. 2003-01-1866. Warrendale, PA: Society of Automotive Engineers
19. Kawano et al (2006). Application of biodiesel fuel to modern diesel engine. SAE Paper
No. 2006-01-0233. Warrendale, PA: Society of Automotive Engineers
20. Lance et al (2004). Emissions performance of pure vegetable oil in two European light
duty vehicles. SAE Paper No. 2004-01-1881. Warrendale, PA: Society of Automotive
Engineers
24. Lapuerta, M. et al (2008) Effect of biodiesel fuels on diesel engine emissions. Science
Direct: Progress in Energy and Combustion Science 34, 198-223
27. Martini, G. et al (2007) Effect of biodiesel fuels on pollutant emissions from Euro 3 LD
diesel vehicles (1). Report EUR 22745 EN. Ispra : DG JRC - IES
57
report no. 6/14
Concawe Data13
Carbon content % wt ASTM D5291 76.98
Hydrogen content % wt ASTM D5291 12.00
Oxygen content % wt In-house method 10.3
H/C Molar Ratio Calculated 1.86
Gross heating value MJ/kg ASTM D240/IP 12 39.73
Net heating value MJ/kg ASTM D240/IP 12 37.18
58
report no. 6/14
Test Fuels
Fuel Property Units Test Method B0 B10 B30 B50
Data from Fuel Blender14
Cetane Number EN ISO 5165 53.2 53.6 53.8 53.9
Cetane Index (4 variable) ASTM D4737 60.3 59.1 57.6 56.1
Density at 15°C kg/m3 EN ISO 12185 823.1 829.1 841.0 853.0
Initial Boiling Point °C EN ISO 3405 202.8 204.9 206.1 214.3
Distillation T50 °C EN ISO 3405 271.2 277.2 293.6 312.0
Distillation T95 °C EN ISO 3405 316.5 330.5 339.3 344.3
Final Boiling Point °C EN ISO 3405 326.5 337.3 344.2 352.1
Flash point °C EN ISO 2719 80 76 91 97
CFPP °C EN116 -22 -21 -25 -21
Cloud Point °C EN ISO 3015 NM -18 -14 -11
Viscosity at 40°C mm2/s EN ISO 3104 2.661 2.785 3.081 3.410
Total Aromatics % m/m EN 12916 22.3 18.3 15.0 10.6
Mono-aromatics % m/m EN 12916 20.8 17.1 14.1 9.9
Di-aromatics % m/m EN 12916 1.4 1.2 0.9 0.7
Tri-aromatics % m/m EN 12916 0.1 0.1 <0.1 <0.1
Polycyclic Aromatic
% m/m EN 12916 1.2 1.2 0.9 0.6
Hydrocarbons
FAME Content % vol EN 14078 <0.1 10.7 30.6 50.9
Sulphur mg/kg EN ISO 20846 <3 <3 <3 <3
Copper Corrosion
EN ISO 2160 1A 1A 1A 1A
3h @ 50°C
Oxidation Stability g/m3 EN ISO 12205 4 0.2 0.1 0.1
Oxidation Stability hrs EN 15751 NM 64.6 31.3 21.5
Water Content mg/kg EN ISO 12937 22 37 138 138
Ash Content % m/m EN ISO 6245 0.002 <0.001 <0.001 <0.001
HFRR Micron EN ISO 12156 227 156 171 168
59
report no. 6/14
Vehicle fuelling
As described below, no back-to-back measurements with the same fuel were conducted.
Therefore, in order to facilitate fuel changes between measurements, minimize cross fuel
contamination and ensure optimal washing of the vehicle fuelling system, it was decided to use an
external fuel tank for these tests. Guidelines were received from the manufacturers in order to
ensure proper engine fuelling and operation. When required, an external low pressure pump was
employed for vehicle fuelling.
For simulating vehicle resistances on the chassis dynamometer, data regarding coast down times
and reference mass were provided by the manufacturers. Particularly for Vehicle C, different
reference masses were employed for fuel consumption and for gaseous pollutant tests. Since the
scope of the study was primarily to test the effect of biodiesel on fuel consumption, the reference
mass corresponding to the fuel consumption test was used for all tests.
For addressing the needs of the study, a protocol was selected which combined the European
NEDC certification cycle and other test cycles that simulate “real world” driving. The objective was
to obtain data regarding the effect of RME over the NEDC, which is one of the most referenced
cycles, but also understand how RME might impact more “real world” operation, particularly over
urban and semi-urban conditions. The “real world” driving cycles were developed in the framework
of the ARTEMIS project [12] and are considered representative of city (URBAN) and rural (ROAD)
driving conditions in Europe. In addition, two steady state modes were used (at 50 and 120 km/h)
for measuring particle size distributions and particle number emissions. The speed versus time
profiles for the NEDC and ARTEMIS cycles are presented in Figure A4-1. The NEDC profile also
shows the UDC and EUDC portions.
The daily measurement protocol started with the NEDC, which is a cold-start driving cycle. The
NEDC consists of an urban part (UDC) where the engine starts from room temperature and an
extra-urban part (EUDC) for testing the car at higher than urban speeds. The NEDC was followed
by 2 x EUDC for vehicle conditioning purposes before completing the two measurements
associated with the ARTEMIS Urban and Road cycles. Each vehicle testing day ended with tests
at constant speed (50 and 120 km/h). Each constant speed test consisted of 5-10 minutes of
conditioning and approximately 15 minutes of measurement.
60
report no. 6/14
After the measurements were completed, fuel change followed by vehicle conditioning took place.
The tank was emptied and washed with the new fuel and the vehicle was then operated for a short
time on the new fuel. During the first 15 seconds of operation with the new fuel, the returning fuel
was collected and discarded in order to avoid contamination caused by any of the previous fuel
remaining in the fuel lines. The vehicle was left to idle for 5 minutes and then the conditioning
process was initiated. A detailed flowchart of the measurement and conditioning processes is
provided in Figure A4-2.
61
report no. 6/14
Testing sequence
No back-to-back measurements were performed with the same fuel. Each fuel was tested once
and then another fuel was used. The fuel testing order was different for each vehicle. The test
sequences performed are presented in Table A4-1.
Table A4-1 Testing sequence for the three vehicles (with dates and vehicle mileage)
As shown in this table, the testing sequences were divided into sub-blocks of four fuels. Each block
constituted of all four fuels examined and always began with a B0 fuel. The rest of the fuels were
tested in random order. This scheme was decided after analyzing the repeatability of previous
measurements which defined the total number of repetitions necessary. As mentioned, each
RME/diesel blend was tested 4 times on each vehicle and the reference B0 fuel five times.
During the measurement and the conditioning periods, data were retrieved from the engine ECU
on various important operating parameters. The data were recorded using a standard OBD-II tool.
Among the parameters monitored and recorded were: engine RPM, engine load, acceleration
pedal position (if available), EGR rate (if available), air flow, exhaust gas temperature (if available)
and possible engine fault codes.
Monitoring and recording these data helped evaluate the repeatability of each vehicle’s engine
operation and additionally helped to identify possible DPF regenerations. In the first case, the
engine RPM and load were used to compare how the vehicle was driven over each driving cycle
and in addition for identifying possible changes in EMS caused by the lower energy content of the
test fuel. In the second case, exhaust gas temperature data upstream of the DPF were used to
identify possible regenerations that might occur during the measurements of Vehicle 1. Such
regenerations would distort the fuel consumption and emissions results of the particular test.
62
report no. 6/14
Emissions sampling
During the study, different emissions were measured and various sampling techniques were
followed. A brief summary of the pollutants/parameters sampled is provided in Table 3 and in
Section 2.
The distinction between NEDC and its sub-cycles is made for emissions that are calculated
both over NEDC as a whole and for the individual sub-cycles.
Over the 120 km/h mode indicative regulated pollutant emissions will be presented based
on calculations made using modal data. However this was not a standardized
sampling/emissions calculation activity.
Fast NOx and high volume PM sampling was conducted only for B0 and B50 fuels. One
sample of PM was collected for all 3 cycles in this case.
More detailed information is given in the following paragraphs.
Emission measurements over NEDC and ARTEMIS were conducted following the European
regulations (Directive 70/220/EEC and amendments). The exhaust gas was primarily diluted and
conditioned by means of Constant Volume Sampling (CVS). A 6 m long corrugated stainless steel
tube transferred the exhaust from the tailpipe to the CVS tunnel inlet. The tube was insulated to
minimize heat losses and particle thermophoresis and was clamped onto the vehicle exhaust pipe
with a metal-to-metal connection to avoid exposing the hot exhaust gas to any synthetic material
connectors. A flow rate of approximately 500 Nm 3/h was maintained in the CVS tunnel by a positive
displacement pump. The dilution air was filtered through a HEPA class H13/EN1822 filter at the
inlet of the dilution tunnel. Proportional diluted exhaust samples were collected in bags for gaseous
pollutants measurements.
Gaseous pollutants were measured with laboratory analyzers as foreseen by European legislation
(chemiluminescence for NOx, flame ionization detector for HC and non dispersive infrared for CO
and CO2). Fuel consumption was derived by means of the exhaust-to-fuel carbon balance, taking
into account the oxygen content of fuels (see Appendix 8). PM samples were collected on 47mm
PTFE-coated glass fibre filters (Pallflex TX40HI20-WW) following the PMP specifications as
presented below.
In order to obtain an indication of the test fuels’ effect over motorway driving conditions, modal
data from the 120 km/h steady state mode were employed for calculating CO2, CO, HC and NOx
emissions and fuel consumption. For these calculations the instantaneous signals of the
aforementioned analyzers were used.
Test order
The programme was conducted using the statistically designed test order specified in Table A4-2
below.
Fuels B10, B30 and B50 will be tested four times in each vehicle, while the base fuel B0 will be
tested five times in order to obtain improved estimates of baseline performance and engine drift.
The three FAME blends B10, B30 and B50 are tested in four randomized blocks, and the base fuel
tests are positioned between adjacent blocks and at the start and end of testing. Different
randomized orders are used for each vehicle.
63
report no. 6/14
PM sampling
Particulate Matter (PM) sampling was performed following the specifications of the PMP protocol.
A separate filter was used for each of the three driving cycles (NEDC, ARTEMIS Urban and Road)
for measuring PM emissions.
In addition to the vehicle tests, six additional blank tests were performed for determining
background levels for both PM and carbonyl compounds (see below). During these 6 tests, the
sampling procedure for PM and carbonyl compounds was repeated identically as for vehicle
measurements but the CVS was disconnected from the vehicle tailpipe.
After weighing and calculating PM emissions, the filters were packed in order to be used for
determining the soluble organic fraction (SOF) of PM, anions and elemental carbon (EC) by
difference. Both PM filters and cartridges were stored according to Concawe’s recommendations
and were sent for analysis periodically as the measurements progressed. It is noted here that all
PM filters (including the ones from ARTEMIS Road which are not scheduled for analysis) were
stored in case some of them need to be used in the future.
The SOF analysis and measurements of aldehyde and ketone emissions were completed by the
Environmental Pollution Control Laboratory (EPCL) of the Department of Chemistry at AUTh.
Filter Preparation
Pallflex TX40 Fluorocarbon coated glass fibre filters were used. The filter batch was always
recorded. The filter diameter was 47 mm.
The particle sample filters (both blank and loaded) were conditioned (as regards temperature and
humidity) in a clean room, under controlled temperature (22±3oC) and humidity conditions
(45±8%), according to PMP regulation. The filters were placed on a grounded aluminium plate
during their conditioning period. Moreover, they were placed under a perforated aluminium cover
in order to be protected from dust and be in contact with the environment at the same time.
Three reference filters were kept in the clean room and were weighed at the same time as the
blank and loaded filters, in-line with the PMP regulation. Each sample filter (blank and loaded) was
weighed more than once during its conditioning period. The conditioning period is set to 2- 80 h by
the PMP procedure. However, because of the subsequent non-regulated pollutant analyses the
64
report no. 6/14
loaded filters were normally kept 24-48 h in the clean room and then immediately stored, in order
to ensure no change of the PM composition.
Microgram balance
The analytical balance used was Mettler-Toledo UMX2 with 0.1 μg resolution. The balance was
grounded by its placement on an anti-static plate and the particulate filters on a grounded
aluminium mat to avoid development of static charge. A reference weight was weighed during the
testing period together with the reference filters. The balance precision (standard deviation) for the
reference weight was 0.9 μg during the whole measuring period.
Filter storage
After their final weighing, the filters were packed in order to be first stored and then sent for PM
speciation analyses. The filter paper was folded in half with the side containing the particulate
deposit on the inside. The folded filter was wrapped in aluminium foil and the foil-wrapped filter
was placed in a suitably-sized self-sealing plastic bag. The plastic bag was stored in the chilled
area of a refrigerator.
Each bag was separately labelled. The bag labelling included the filter code number which
identifies a unique test and the loaded filter conditioning time. The date of the final weighing, the
weight of the filter, the batch number of the filter, and the initials of the person preparing the filter
sample bag were also recorded.
65
report no. 6/14
This Appendix provides additional information on the statistical analysis methods discussed in
Section 3.2.
The data were examined for possible outliers and trends by examining studentized residuals
(residuals divided by their standard errors) in analysis of (co)variance models fitted to the
measured emissions, or fuel consumption, for a particular vehicle and cycle on both the natural
and log-transformed scale. In this study, we fitted a one-way ANOVA model to each vehicle × cycle
× emission combination with emission or ln(emission) as the response variable and fuel as the
classifying factor. Trends were sought by treating test order as a covariate.
The studentized residuals were compared against the upper 5% and 1% points tabulated in [26].
Suspicious results were queried with the originating laboratory and were not rejected unless there
were sound engineering reasons to believe that something untoward had happened in that
particular test. In the event, only a few results were rejected, as follows:
Vehicle 1:
CO results in NEDC/UDC/EUDC on 6 March 2009 on B50 (abnormally high)
CO result at 120 km/h on 19 March 2009 on B10 (abnormally high)
HC results in Artemis Road & Urban cycles on 27 March 2009 on B50 (abnormally high)
Vehicle 2:
NOx and NO results in NEDC/UDC/EUDC on 9 April 2009 on B50 (abnormally high); the
corresponding NO/NOx results were also rejected
PM result in NEDC on 28 April 2009 on B50 (abnormally low)
In addition, the following results became outliers after the trend correction had been applied and
so were removed before calculating trend corrected means:
Vehicle 1:
CO2 and FC results in Artemis Urban cycle on 31 March 2009 on B0 (abnormally high)
In Appendix 6 and the bar charts in Section 4, arithmetic means are used for fuel consumption,
gaseous emissions, the NO/NOx ratio, and PM despite the lognormality assumed in the data.
Geometric (i.e. logarithmic) means give excellent comparisons between fuels on a percentage
basis but have the disadvantage of underestimating total emissions to the atmosphere. Arithmetic
means give better estimates of total emissions to the atmosphere but can be inflated unduly by
isolated high results.
Each vehicle cycle emission measurement data set was analysed separately. The arithmetic
mean emissions and fuel consumption, and their standard errors, were estimated for the various
fuels from a weighted analysis of variance or covariance in which each measurement was assigned
a weight equal to
weight = 1 / (mean emission for that fuel and vehicle)2
to take account of the lognormality in the data (see [23], Annex 05).
66
report no. 6/14
In the bar charts presented in Section 4, the error bars show the
mean value ±1.4 x standard error of mean
The 1.4 factor was chosen for consistency with both the EPEFE [23] and recent Concawe reports
[13,14,15,16]. The original rationale was that when two fuels were significantly different from one
another at P<5%, their error bars would not overlap; this factor also gave 84% confidence that the
true mean lay within the limits shown.
Error bars based on a 1.4 factor err on the side of being slightly too narrow for determining
significant differences in the present programme as fewer tests were carried out. Such an
interpretation would require error bars based on a factor between 1.52 and 1.56, depending on the
exact number of valid tests and whether or not a trend correction has been applied. Therefore
there needs to be a sizeable gap between the error bars for two fuels for their means to be
considered significantly different from one another at P<5%.
67
report no. 6/14
The arithmetic mean emissions and fuel consumption from each vehicle, fuel, and driving cycle
are summarized in this Appendix. Both the uncorrected and corrected means are shown wherever
a trend correction has been applied. See Section 3.2 of the report for more details.
68
report no. 6/14
69
report no. 6/14
70
report no. 6/14
Note: These plots show all data before trend correction or the deletion of outliers.
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
71
report no. 6/14
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
72
report no. 6/14
73
report no. 6/14
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
74
report no. 6/14
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
Vehicle 2 Vehicle 2
Vehicle 1 Vehicle 1
Vehicle 3 Vehicle 3
75
report no. 6/14
Vehicle 2 Vehicle 1
Vehicle 3
76
report no. 6/14
Vehicle 2
Vehicle 1
Vehicle 3
Vehicle 2
Vehicle 1
Vehicle 3
Vehicle 2
Vehicle 1
Vehicle 3
77
report no. 6/14
According to the regulated procedure fuel consumption is calculated based on the carbon balance
between the exhaust gas and the fuel consumed. The procedure assumes a constant
hydrogen/carbon ratio in the fuel for which a fixed value is provided for diesel (CH 1.86 or C/H =
0.155). The procedure does not foresee yet values for biodiesel and its blends with diesel. In order
to accurately calculate fuel consumption and correct for the biodiesel hydrogen-carbon ratios and
oxygen content the fundamental equation (below) for carbon balance fuel consumption has been
applied.
Where:
FCl/100km = calculated fuel consumption in l/100 km
CWFFuel = carbon weight (mass) fraction of the fuel
CWFExh = carbon weight (mass) fraction of the exhaust hydrocarbons
0.429 = carbon weight (mass) fraction of CO
0.273 = carbon weight (mass) fraction of CO2
HC = HC emissions for the test (g/km)
CO = CO emissions for the test (g/km)
CO2 = CO2 emissions for the test (g/km)
SG = density of the fuel (kg/l)
The Dilution Factors (DF) used for deriving pollutant emissions in the case of the four test fuels
and were calculated from the C/H/O ratios. Although these corrections led to minor changes in fuel
consumption (≤1%), they were considered important because the primary scope of the study was
to investigate the effect of biodiesel on vehicle energy efficiency. In addition the differentiations
expected between the test fuels were of similar order (0.5-5%) and therefore the correction was
important. Diesel fuel consumption and emissions were calculated solely on the legislated basis
using only the actual fuel density of B0 and not the predefined value.
With regard to Euro 5+ emissions legislation for diesel passenger cars and corresponding
development trends in exhaust gas measurement technology, such as advanced particulate
counter devices, the DF plays an important role for the measurement of regulated emissions over
78
report no. 6/14
the NEDC. Because the biofuel-containing diesel fuels (e.g., B10) have a carbon-weight fraction
that is different from B0, the DF must be calculated based on the biodiesel’s actual C/H/O ratio in
order to ensure constant dilution with fresh air in the CVS system while measuring tailpipe
regulated emissions. The DF is calculated in the automatisation system (e.g. GEM device) of the
chassis dynamometer from individual emissions bag concentrations and C/H/O lab fuel analysis
parameters by using the formulae from Regulation No. 83 Revision 3 - Amendment 3 (22 July
2009) as shown below:
where:
And
X = 100 * Cx / [Cx+(Hy/2)+(3,76*(Cx+(Hy/4)-Oz/2))]
where
79
report no. 6/14
In the following figures, the left-hand charts show the measured FC (in l/100km) versus the VLHV
(in MJ/l) of the four test fuels for each of the three vehicles over one particular cycle. The error bars
represent 95% confidence limits, calculated as ± 2 SE based on the repeatability of multiple FC
measurements. The solid line is a best fit through the data points and the dotted lines show 95%
confidence limits for the true regression line. The dashed line is the one-to-one correlation line.
In the right-hand figures, the percent change in FC is plotted versus the percent change in [1/VLHV]
using the B0 results for each vehicle as the basis. The error bars again represent 95% confidence
limits and the solid line is a best fit through the data points. The slope of this line is also shown.
The dotted lines are the 95% confidence limits around the best fit line and the dashed line is a one-
to-one correlation line.
Vehicle 1 Vehicle 1
11.6 9
11.4
Fuel Consumption
7
B50
% Chnage in FC
y = 1.925x
11.2 B30 B50
5
B30
11.0
B10 3
10.8 B10
B0
10.6 1
B0
10.4 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Vehicle 2 Vehicle 2
8.9 9
8.8
Fuel Consumption
7
% Chnage in FC
8.7 B50
8.6 5
y = 1.3552x B50
8.5 B30
3
B10 B30
8.4
B0 1 B10
8.3
B0
8.2 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Vehicle 3 Vehicle 3
8.8 9
8.7
Fuel Consumption
B50 7
% Chnage in FC
8.6
5
8.5 B30
B0 3 B50
8.4 y = 0.7212x
B10 1
8.3 B30
B0
8.2 -1 B10
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
80
report no. 6/14
Vehicle 1 Vehicle 1
5.7 9
Fuel Consumption
% chnage in FC
5.6
B50
B30 5
5.5 y = 1.0916x
3 B50
B30
5.4 B0 1
B10
B0 B10
5.3 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Vehicle 2 Vehicle 2
5.2 9
5.1 B50 7
Fuel Consumption
% chnage in FC
7
5.7
% chnage in FC
B50 5
5.6
B30 y = 1.0749x B50
3
B30
5.5 B10 1
B0 B10
B0
5.4 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
% change in (1/ VLHV)
Volumetric Lower Heating Value
81
report no. 6/14
Vehicle 1 Vehicle 1
5.5 9
5.4
Fuel Consumption
% chnage in FC
B50
5.3 5
y = 1.2138x B50
B30
5.2 B10 3
B0 B10 B30
5.1 1
B0
5.0 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Volumetric Lower Heating Value % change in (1/VLHV)
Vehicle 2 Vehicle 2
5.6 9
5.5 7
Fuel Consumption
B50
% chnage in FC
5.4 5
B30 y = 1.2248x B50
5.3 B10 3
B0
5.2 1
B30
B10
B0
5.1 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Volumetric Lower Heating Value % change in (1/VLHV)
Vehicle 3 Vehicle 3
6.2 9
Fuel Consumption
6.1 B50 7
% chnage in FC
5
6.0 B30
y = 0.9239x B50
3
B10 B0
5.9 B30
1
B0 B10
5.8 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
82
report no. 6/14
Vehicle 1 Vehicle 1
12.2 9
12.1
B50
Fuel Consumption
12.0 7
% Chnage in FC
B30 y = 1.7067x
11.9
5 B50
11.8 B10 B30
11.7
3 B10
11.6
11.5 B0 1
11.4 B0
11.3 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Vehicle 2 Vehicle 2
10.0 9
9.9
Fuel Consumption
7
% Chnage in FC
9.8
B50
9.7 5
B30 y = 1.3686x
9.6 B50
9.5 3 B30
B10
9.4 B0 B10
1
9.3 B0
9.2 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
% Change in (1/VLHV)
Volumetric Lower Heating Value
Vehicle 3 Vehicle 3
10.7 9
10.6
7
Fuel Consumption
B50
% Chnage in FC
10.5
10.4 5
B30
10.3 B0 3
10.2 B10 y = 0.4868x B50
1
10.1 B30
B0
10.0 -1 B10
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
83
report no. 6/14
Vehicle 1 Vehicle 1
5.9 9
5.8 7
Fuel Consumption
% chnage in FC
5.7
B50
5
5.6 B30 y = 1.1367x B50
3
5.5 B0
B10 1
B30
5.4
B0
5.3 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Volumetric Lower Heating Value % change in (1/VLHV)
Vehicle 2 Vehicle 2
5.8 9
5.7 7
Fuel Consumption
% chnage in FC
5.5 B0 3
B10 B30 B50
5.4 1 y = 0.7096x
B0
5.3 -1 B10
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Volumetric Lower Heating Value % change in (1/VLHV)
Vehicle 3 Vehicle 3
6.6 9
6.5 7
Fuel Consumption
% chnage in FC
6.4 B50 5
6.3 B30
y = 0.9736x B50
3
6.2 B10 B0 B30
6.1 1
B0 B10
6.0 -1
34.0 34.2 34.4 34.6 34.8 35.0 35.2 35.4 0 1 2 3 4
Volumetric Lower Heating Value % change in (1/VLHV)
84
report no. 6/14
Vehicle 1
Vehicle 2
Vehicle 3
85
report no. 6/14
At the beginning and the end of each vehicle’s test sequence samples of the test blends were
collected. These samples were tested for oxidation stability in order to monitor possible oxidation
of the fuels while the testing progressed. Samples for the modified Rancimat tests were taken from
the barrels opened the following days:
EN15751
Sample Sampling date Viscosity Rancimat Time Acid Number
cSt @ 40C hours mg KOH/g
B0 28-Feb-09 2.665
70.0
60.0
50.0
40.0
86
30.0
report no. 6/14
70.0
60.0
50.0
40.0
30.0
Viscosity
20.0
Rancimat Time
10.0 Acid Number
0.0
30-Jun-09
12-Mar-09
9-Apr-09
7-Apr-09
8-Apr-09
9-Jun-09
2-Jul-09
4-Mar-09
4-Mar-09
25-May-09
27-May-09
26-May-09
87
Concawe
Boulevard du Souverain 165
B-1160 Brussels
Belgium
Tel: +32-2-566 91 60
Fax: +32-2-566 91 81
e-mail: [email protected]
website: http://www.concawe.org