You are on page 1of 14


The advantages of applying Compressed Natural Gas (CNG)
as a fuel for internal combustion engines are well known. In
addition to a signifcant operating cost savings due to a lower
fuel price relative to diesel, there is an opportunity to reduce
the engine's emissions. With CNG combustion, some
emissions, such as Particulate Matter (PM) and Carbon
Dioxide (CO2), are inherently reduced relative to diesel fueled
engines due to the nature of the combustion and the molecular
makeup of the fuel. However, it is important to consider the
impact on all emissions, including Total Hydrocarbons (THC)
and Carbon Monoxide (CO), which can increase with the use
of CNG. Nitrogen Oxides (NOx) emission is often reported to
decrease with the use of CNG, but the ability to realize this
beneft is signifcantly impacted by the control strategy and
calibration applied.
FEV has investigated the emissions and performance impact
of operating a heavy-duty diesel engine with CNG in a dual fuel
mode. The CNG was introduced via injectors mounted to an
inlet pipe located upstream of the intake manifold. The
fumigation approach included a mixer to improve the
distribution of gas prior to delivery to the cylinder. The initial
investigations sought to determine how the performance of a
heavy-duty diesel engine would be affected by the introduction
of CNG. For this effort there was no change to the base engine
calibration, and the ability to maximize substitution of diesel
with CNG was investigated. It was observed that the ability to
maximize substitution of diesel with CNG across the operating
map was limited by extremely high THC levels, combustion
instability and limitations in peak cylinder pressure and exhaust
gas temperature.
With the application of a simplifed engine calibration with a
single diesel injection and Exhaust Gas Recirculation (EGR),
timing adjustments allowed higher CNG substitution levels in
several areas of the operating map. A further increase in gas
substitution along with higher fuel conversion effciency,
improved combustion stability and even lower emissions could
be achieved through Reactivity Controlled Compression
Ignition (RCCI) combustion. This approach required a unique
injection strategy along with a careful balance of EGR rates
and boost pressure. Under this combustion regime it was
possible to observe a simultaneous reduction of NOx and PM
emissions, approaching engine-out emission levels that could
avoid, or signifcantly minimize, aftertreatment of these
With the desire to quickly apply CNG systems to existing diesel
engine architecture in an effort to reap the beneft in fuel cost
savings, manufacturers and system developers must be careful
to understand the full impact on the engine's performance and
emissions. Tests conducted as part of this investigation have
revealed that an un-optimized approach to CNG introduction
can lead to extreme THC emissions that mostly consist of
Methane (CH4). In addition, the maximum gas substitution
level is signifcantly limited in most regions of the engine
operating map. Thus, the ability to specifcally tune the
calibration for operation with CNG is essential to achieving the
maximum beneft in fuel cost savings and emission control.
The use of CNG in heavy-duty applications has seen a sharp
growth in the last couple of years as reported in studies
conducted by the Energy Information Administration (EIA) [1].
EIA studies project CNG consumption in heavy-duty
applications to increase at 11.4% per year through 2040. The
growth is mainly seen in the off-road market, but other sectors
such as on-highway, marine and locomotive are also
embracing a switch to CNG at a rapid rate.
There are two main technology approaches through which
CNG can be applied for heavy-duty engines. A common
approach for existing diesel engines is to operate in a dual fuel
regime where diesel and CNG are combusted simultaneously.
Investigation of Diesel and CNG Combustion in a
Dual Fuel Regime and as an Enabler to Achieve RCCI
Published 04/01/2014
Mufaddel Dahodwala, Satyum Joshi, Erik W. Koehler, and Michael Franke
FEV Inc.
CITATION: Dahodwala, M., Joshi, S., Koehler, E., and Franke, M., "Investigation of Diesel and CNG Combustion in a Dual
Fuel Regime and as an Enabler to Achieve RCCI Combustion," SAE Technical Paper 2014-01-1308, 2014,
Copyright © 2014 SAE International
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
A second approach considers a dedicated CNG combustion
system that requires the addition of an ignition source. The
main technological challenges that apply to this type of engine
conversion are studied in detail by Ribas [2].
As an additional consideration, past studies have shown that
due to throttling losses and lower compression ratio, effciency
of a Spark Ignited (SI) CNG engine is lower compared to a dual
fuel diesel/CNG engine, particularly at part loads [3]. Brake
Mean Effective Pressure (BMEP) capability is also reduced
with an SI CNG engine, leading to lower torque at low speeds
when compared to a dual fuel diesel/CNG engine. An SI CNG
engine does however have an advantage on the exhaust gas
aftertreatment, as stoichiometric operation allows the
application of a conventional and comparatively simple 3-way
In the case of dual fuel diesel/CNG engines, two types of
technologies are currently considered depending on the
method used for CNG induction. In the frst type, both the
diesel fuel and natural gas are directly injected into the
combustion chamber using either two separate injectors or a
special injector with a dual-concentric needle design [4]. In the
second type, natural gas is either fumigated at a single point
into the intake path of the engine before or after the
turbocharger compressor and is premixed with air and EGR or
injected at multiple points in the intake port of the engine.
Intake fumigation at single point is the most widely used
method of CNG induction for on-highway applications, due to
its simplicity, and is the approach applied for this work.
Much research has been conducted to understand the
combustion behavior of dual fuel diesel/CNG engines [4, 5, 6,
7, 8, 9, 10]. As reported by most researchers, the major
diffculty with dual fuel operation is the challenge of providing
high levels of CNG substitution, especially at low and medium
loads. Attempts to apply high substitution levels lead to lower
engine effciency and higher concentrations of CO and
unburned hydrocarbon emissions. At higher loads, the
improvement in CNG utilization leads to improvements in both
engine performance and exhaust emissions. To date a large
portion of the research has been focused on either single
cylinder research engines or light-duty production engines.
Relatively few studies look at engine performance and
corresponding emission impact of dual fuel concepts applied to
medium-duty and heavy-duty production engines. One study
on a heavy-duty engine was conducted by Guzman et. al. [4],
wherein the authors recalibrated the base diesel engine to
operate in dual fuel mode to quantify the CO2 and PM
emissions during the transient portion of the Heavy-Duty
Federal Test Procedure (HD FTP) and World Harmonized
Transient Cycle (WHTC). Their work found that, with a properly
calibrated dual fuel engine signifcant reductions in PM and
CO2 emissions could be achieved in transient operation when
compared to a baseline diesel engine.
In recent years, the heavy-duty market has seen the strongest
growth in the application of dual fuel technologies, with refuse
trucks and buses leading the way. In terms of on-highway
applications, most dual fuel engines currently in use are End
Of Life (EOL) retrofts that do not need to comply with stringent
emission standards. To date, few companies have certifed
dual fuel engines to US2010 emission standards. One of those
that has certifed to US2010 has stated a fuel cost savings of
up to 30% and a maximum CNG substitution of 70%.
One motivation behind this study was to understand the
limitations imposed by a US2010-compliant on-highway diesel
engine in terms of maximum CNG substitution. Additional
complexity to this approach results from the management of
cylinder-to-cylinder variations, which in some cases makes the
most promising strategy developed on a single cylinder engine
not feasible for a production engine without applying signifcant
design changes. One of the main requirements for employing a
dual fuel approach is to have the fexibility of operating in
diesel-only mode. Design changes to specifcally allow better
operation with CNG will not be attractive to the end user if they
compromise the performance of the engine in diesel-only
The frst part of the study aims at understanding engine
behavior with introduction of CNG without changes to the base
calibration, as is typically done by retroft companies. To
accomplish this, 13 load points were selected, and a maximum
CNG substitution sweep was conducted at each of these
points. The second part of the study investigated the impact of
diesel calibration changes on the allowable maximum CNG
substitution and thus the achievable fuel cost savings. The
calibration optimization was completed at all 13 load points.
The fnal part of this study is focused on exploring advanced
combustion concepts with CNG and seeks to maximize
substitution and lower emissions.
Experimental Setup
A production heavy-duty diesel engine was used for this
investigation. The 2010 inline 6-cylinder engine has a
displacement of 13 liters, is rated at 425 HP and is On-Board
Diagnostics (OBD) - II compliant. It is equipped with a high
pressure common rail fuel injection system, cooled high
pressure EGR, and a compression ratio of 16.5:1. CNG was
introduced into the intake using eight CNG injectors located
downstream of the charge air cooler. A mixer was installed
downstream of the CNG introduction location to support equal
distribution of the CNG in the intake manifold. Figure 1 depicts
the test cell setup applied for this investigation. A Rapid
Controller Prototyping (RCP) system was used for controlling
the amount of CNG induction. The production engine was
equipped with a Diesel Oxidation Catalyst (DOC), Diesel
Particulate Filter (DPF) and Selective Catalytic Reduction
(SCR) catalyst, however for this investigation only a DOC was
installed and the backpressure was adjusted via a back
pressure valve to simulate the absent aftertreatment
components. The aftertreatment module in the Engine Control
Unit (ECU) was deactivated to prevent it from affecting the
base engine performance.
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
Figure 1. Test Cell Schematic
The engine was coupled to a 560 kW Alternating Current (AC)
dynamometer. CNG and diesel fuel fow measurements were
accomplished using Micro Motion fow meters while the air fow
was measured using an ABB air fow meter. Engine-out and
tailpipe gaseous emissions were measured with a dual channel
Horiba MEXA 7500 DEGR emission bench. The intake
manifold was instrumented for CO2 concentration
measurement, which was used for calculation of the EGR rate.
PM emission was determined through the Motor Industry
Research Association (MIRA) calculation, which utilized data
provided by a smoke meter drawing sample upstream of the
DOC. The engine was instrumented with Kistler 6061B
in-cylinder pressure transducers to allow cylinder pressure
measurements on all six cylinders.
Test Matrix
Based on operating points from the Ramped Modal Cycle
(RMC), HD FTP and highway cycles for this engine, 13 specifc
test points were selected for this study. The points were
selected to provide a good balance between certifcation cycles
and real-world operation. The program test points are overlaid
upon the cycle operating points in Figure 2. For the baseline
substitution study, the CNG substitution was increased in 10%
increments until the engine became unstable (Indicated Mean
Effective Pressure (IMEP) Coeffcient of Variation (COV) less
than 10%) or a limit of peak pressure or maximum exhaust
temperature was reached.
Figure 2. Test points selected based on the operating points on the
RMC, HD FTP and highway cycles
For optimization studies, the effect of main injection timing and
CNG substitution was studied at each of the selected operating
points. Figure 3 shows an example test matrix for one
particular engine speed. The main injection timing was swept
at different substitution levels across the three load points. At
high loads the timing could not be advanced because of peak
pressure limitations, and it was found that higher substitution
could only be achieved by retarding the main timing. RCCI
combustion was evaluated at the low load points. The test
matrix for RCCI combustion at low loads was an extension of
the test matrix shown in Figure 3. The timings were advanced
to 50-80 deg Crank Angle (CA) before Top Dead Center
(bTDC) at substitution levels of 60-80%. The CNG substitution
was calculated on an energy basis. The details of the
calculation are outlined in the Appendix.
Figure 3. Test matrix applied for optimization at a fixed speed
Result and Discussion
Baseline Maximum Substitution
A review was conducted to understand the methods employed
by retroft manufacturers to substitute CNG for diesel without
changes to the base calibration. Two popular approaches are
to manipulate the boost signal and to bleed off part of the fuel
supplied to the injectors back to the fuel tank to reduce the
injected diesel fuel for a particular load. These approaches are
not directly possible when considering an engine equipped with
OBD. Without OBD calibration changes, such approaches
would cause a low boost or injector fault and result in a de-rate
of the engine. Thus, to avoid this infuence for the current work,
the reduction in diesel quantity was obtained by reducing the
pedal demand and maintaining load by introducing CNG into
the system. With this approach, the control parameter set
points, including the injection events, EGR rates and rail
pressures, were not consistent for different substitution levels
within the same load point. These variables were controlled by
the ECU based on the commanded pedal (fueling) and the
engine speed.
In an effort to maintain this paper within a reasonable length
the results presented without calibration changes will be limited
to 1500 rpm and three different load points; 6 bar, 14 bar and
24 bar. These points will be analyzed to understand the effect
of CNG substitution on engine performance and emissions.
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
Figure 4 shows the Apparent Heat Release Rate (AHRR) and
cylinder pressure traces for 6 bar BMEP at three different
substitution levels. The emission impact and corresponding
control variables for different substitution levels at this load
point are provided in Figure 5. From these fgures, deterioration
in combustion effciency can be seen with an increase in CNG
substitution, possibly due to the reduction in injected diesel fuel
leading to lower combustion chamber temperatures. This led to
signifcant increase in hydrocarbon emissions, Brake Specifc
Fuel Consumption (BSFC) (calculated on a diesel equivalence
fuel fow rate) and brake specifc cost, which is defned as the
net fuel cost per kilowatt hour of energy (Appendix). An
increase in NOx emission was observed up to 40% CNG
substitution, due to reduction in EGR, but then dropped as the
combustion effciency deteriorated. The EGR reduction can be
attributed to the richer lambda values observed with increasing
CNG substitution as the fresh air charge is displaced with
CNG. Increased PM emissions at higher substitution were
possibly due to the increased soluble organic fraction at higher
loads and a richer lambda value.
Figure 4. AHRR and in-cylinder pressure for 6 bar BMEP load at 1500
rpm at three different substitution levels
A reduced peak in the heat release trace and a late burning of
CNG, also observed by Maxey et al. [5], was clearly visible with
an increase in substitution. Observations regarding the
combustion behavior are in agreement with those of Maxey et
al.; wherein the reduction in diesel fuel quantity reduces the
amount of diesel available to ignite the less reactive CNG [5].
Figures 6 outlines the AHRR and cylinder pressure trace for 14
bar BMEP, while Figure 7 shows the emission impact with
increasing CNG substitution at the same load. Similar trends
for NOx, PM and THC emissions were observed for 14 bar
BMEP as seen at 6 bar BMEP, however, here the combustion
effciency remained above 90% leading to a net beneft in
brake specifc cost. At this load, a maximum substitution of
98% could be reached where CNG was ignited by the diesel
pilot injection and an SI type heat release trace was observed.
Figure 5. Emissions and performance parameters at 6 bar BMEP,
plotted against CNG substitution
A NOx emission increase was closely coupled to a reduced
EGR rate and an earlier heat release. In general, the
combustion COV, as also observed by Sun et al. [6], was
affected by CNG substitution percentage and the EGR quantity.
The cylinder pressure rise rate and the peak cylinder pressure
increased signifcantly at 24 bar BMEP, shown in Figure 8, due
to the simultaneous combustion of diesel and CNG. This
limited CNG substitution to 36% at this load point. Combustion
effciency remained above 98.8% due to the higher combustion
temperatures. The NOx, PM and THC emissions trends
presented in Figure 9 for 24 bar BMEP are similar to those
observed at 6 bar and 14 bar BMEP.
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
Figure 6. AHRR and in-cylinder pressure for 14 bar BMEP load at 1500
Figure 7. Emissions and performance variables at 14 bar BMEP,
plotted against CNG substitution
Figure 8. AHRR and in-cylinder pressure for 24 bar BMEP load at 1500
Figure 9. Emissions and performance variables at 24 bar BMEP,
plotted against CNG substitution
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
Figure 10 shows the practical CNG substitution map when
access to the calibration is not available and thus no calibration
changes are applied. This map is determined by maintaining
the engine-out emissions within a level that would allow the
engine to meet US2010 emission legislation when conventional
diesel aftertreatment is applied. The average NOx emission for
the baseline diesel calibration was approximately 2.0 g/bhp-hr,
therefore a NOx limit of 2.0 g/bhp-hr, along with a NMHC limit
of 0.238 g/bhp-hr, was applied as limiting criteria.
From Figure 10 we can see that the average substitution would
be limited to less than 10% if the base calibration is carried
over and no changes are applied. There are three major
reasons for the limited substitution shown in Figure 10. First, at
low loads the amount of unburned THC is very high. Second,
at mid loads the NOx emissions tend to increase due to a
reduction in EGR caused by running richer lambda values.
Lastly, the peak pressures increase with increase in CNG
substitution at higher loads, leading to the necessity to observe
mechanical limits. Most dual fuel retrofts on the market today
are applied to EOL vehicles that do not require emission
certifcation. These applications can therefore accept increases
in emission levels and apply higher CNG substitution levels.
Figure 10. Practical CNG substitution map without base engine
calibration changes and meeting US2010 emission standards
Calibration Optimization
As shown in the previous section, only very low CNG
substitution levels can be achieved when changes to the base
diesel calibration are not considered. Therefore, to increase the
CNG substitution levels, a modifcation of the base calibration is
required. The base calibration used three injections along with
EGR to reach the desired NOx, PM and Non-Methane
Hydrocarbons (NMHC) targets when operating on diesel fuel. To
understand the effect of each injection event, a series of tests
were conducted at 1500 rpm and 6 bar BMEP and multiple
substitution levels. Three injection strategies were explored:
main injection only, main + pilot injection and main + post
injection. In order to eliminate the effect of EGR, the EGR rate
was reduced and maintained at a near-zero level throughout
modifcation of the base calibration. A summary of the injection
strategy employed for each case is provided in Table 1.
Figure 11 shows the impact of each of these strategies on the
NOx and THC emissions at various CNG substitution levels.
The NOx emissions decrease with an increase in CNG
substitution levels for all three strategies, but none of the
strategies show a signifcant advantage in their ability to control
THC emissions. Therefore, to simplify the diesel injection
control, a main injection only strategy was applied for the
calibration optimization.
Table 1. Overview of the three injection strategies employed at 1500
rpm, 6 bar BMEP and different CNG substitution
Figure 11. Comparison of NOx and THC emission for multiple CNG
substitution levels at 1500 rpm and 6 bar BMEP
It was also realized that EGR was required to control the NOx
emissions, as shown in Figure 12.
A study was then conducted with multiple CNG substitution
levels and diesel injection timings for the main injection only
case. At each speed and load point the EGR rate and injection
pressure were held constant. The boost pressure was
controlled by the ECU based on the existing calibration. The
diesel injection timing advance was restricted to 20 deg bTDC,
above which the possibility of spraying the diesel fuel directly
on the cylinder liner was high. The CNG substitution was
limited to 80%; above this level the combustion was unstable
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
and resulted in very high THC emissions. The optimization
study was conducted at all 13 points as outlined in the Test
Matrix section. To describe the impact of injection timing and
CNG substitution on emission and performance, the following
text focuses on the results obtained at 1500 rpm and three
different load points.
Figure 12. Comparison of NOx emissions for multiple CNG substitution
levels at 1500 rpm and 6 bar BMEP
Figure 13 shows the effect of diesel injection timing and CNG
substitution levels on the NOx, PM and THC emissions at 1500
rpm and 6 bar BMEP. A decrease in NOx emissions and an
increase in THC emissions can be observed as CNG
substitution levels are increased at constant injection timing.
Figure 13. Injection timing study for different CNG substitution levels at
1500 rpm, 6 bar BMEP with 35% EGR
Figure 14 compares the AHRR and pressure traces for three
different substitution levels at a fxed injection timing of 11 deg
bTDC. As shown in Figure 14, the initial peak of the AHRR
trace reduces as CNG substitution is increased, while the
longer burn durations lead to lower combustion temperatures
and lower NOx emissions. However, at a fxed substitution
level the NOx emissions increase as the injection timing is
advanced, but the THC emissions decrease.
Figure 14. AHRR and cylinder pressure trace for 1500 rpm, 6 bar
BMEP at 11 deg bTDC with three different substitution levels
Comparing the AHRR traces in Figure 15, it can be observed
that, as the injection timing is advanced, the AHRR traces
move toward Top Dead Center (TDC). This leads to higher
in-cylinder temperatures causing higher NOx emission
formation but at the same time allowing more time for the CNG
to burn at higher temperatures and reducing the THC
emissions. Although the best NOx emission results are
obtained between 60% and 80% substitution levels and
approximately 12 deg bTDC injection timing, at this point, the
unburned THC emissions are very high. Thus, a compromise
must be identifed between the maximum substitution and the
allowable NOx and THC emissions. The PM emissions, on the
other hand, are very low across the complete optimization
range, with a minimum achieved at 20 deg bTDC between 20%
and 60% CNG substitution levels. Unfortunately, the NOX
emissions in this range are very high.
Figure 16 shows the NOx, PM and THC emissions at 1500 rpm
and 13 bar BMEP. The NOx emission results are similar in
terms of a decreasing trend with increase in CNG substitution
levels at the same injection timing and an increase in NOx
emissions with injection timing advance at a constant
substitution level. Again, as in the previous case, the regions
where the NOx emissions are lowest have very high THC
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
emissions and vice-versa. Thus, at mid loads as well, the
amount of CNG substitution is dependent upon the trade-off
between NOx and THC emissions.
Figure 15. AHRR and cylinder pressure trace for 1500 rpm, 6 bar
BMEP at 60% CNG substitution with three different injection timings
Figure 16. Injection timing study for different CNG substitution levels at
1500 rpm, 14 bar BMEP with 38% EGR
Figure 17 shows the impact of injection timing at three different
substitution levels for 24 bar BMEP. Contrary to the approach
at the lower loads, the injection timing was retarded at high
load to allow for higher CNG substitution levels. By retarding
the injection timing, the mechanical limits of maximum cylinder
pressure were avoided. As the injection timing was retarded
with increasing CNG substitution levels, the NOx and PM
emissions were decreasing with a familiar increase in THC
emissions. Comparing the AHRR traces in Figure 18, it is
observed that as the CNG substitution is increased the peak of
the AHRR trace moves away from TDC, matching the earlier
fnding of a longer burn duration with less reactive CNG leading
to reduced NOx formation. At a fxed substitution level, a
retarded timing reduced the NOx emission as expected, but the
THC emissions increased. The ability to further retard the
injection timing is limited beyond a certain crank angle since
the longer CNG burn duration leads to high THC emissions.
Figure 17. Injection timing study for different CNG substitution levels at
1500 rpm, 24 bar BMEP with 28% EGR
Based on the main timing and CNG substitution studies
conducted at the 13 speed and load points, a map was
generated showing the maximum CNG substitution that is
possible when allowing changes to the base calibration. The
map is again determined by observing the limits imposed by
the US2010 emission legislation.
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
As shown in Figure 19 the average substitution achieved is
approximately 50%, with higher substitution achieved at the
higher load points where the retarded injection timing strategy
was employed to control the pressure rise rates. With this
substitution level, there was a penalty in the CH4 emissions,
but these emissions are not regulated by the US2010 emission
legislation, and therefore, no limit was imposed on the CH4
levels while generating the optimized map. At low loads the
substitution is mainly limited by the unburned hydrocarbons
due to a limited diesel quantity available for igniting the
premixed CNG charge.
One possible approach to overcoming these limitations is to
change from conventional diesel combustion to reactivity
controlled combustion. RCCI is a dual fuel strategy in which a
high reactivity fuel (diesel) is injected early enough to allow
complete mixing with a low reactivity fuel (CNG) to create a
diverse reactivity map within the chamber, allowing for
controlled and complete auto ignition of the premixed charge.
The RCCI combustion strategy not only allows higher gas
substitution levels but also has the potential to simultaneously
reduce NOx and PM emissions.
Figure 18. AHRR and cylinder pressure trace for 1500 rpm, 24 bar
BMEP with three different substitution levels
Figure 19. Optimized CNG substitution map with calibration changes
and meeting US2010 emission standards
Investigation of RCCI Combustion
Figure 20. AHRR and cylinder pressure traces for three different main
injection timings at 60% substitution (6 bar BMEP at 1500rpm)
The existence of very high hydrocarbon emissions at low loads
impeded higher substitution levels during the optimization
effort. As a step toward further optimization, it was considered
that the higher fuel reactivity difference between CNG and
diesel could potentially enable RCCI combustion, despite a
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
relatively high compression ratio. To explore this concept, the
main injection timing for diesel was advanced well beyond
conventional diesel injection timings at a low load point.
Figures 20 and 21 show the AHRR and the corresponding
emissions at four different main injection timings and a 60%
CNG substitution level. Up to a main injection timing of 32 deg
bTDC, the peak pressures and temperatures continue to
increase, leading to high levels of NOx formation. THC, CO
and smoke were reduced due to an increased peak
combustion temperature.
Figure 21. Emissions and performance variables at 1500 rpm and 6
bar BMEP, plotted against main injection timing
However, with further advancement in timing, beyond 32 deg
bTDC, the AHRR trace moved closer to TDC. This resulted in
lower peak combustion temperatures and pressures causing a
sharp reduction in NOx emissions. The heat release trace is
wider with a lower peak and clearly visible two-stage heat
release that includes a cool fame region. The heat release
profle closely resembled that of RCCI combustion [11], and it
was concluded that the spatial fuel reactivity gradient led to the
controlled heat release [11].
Although the reactivity gradient governed the heat release rate,
EGR was necessary to control the combustion phasing for this
high compression ratio engine. At 1500 rpm and 6.0 bar BMEP,
the maximum possible EGR level was 37%. This EGR level
was held constant for each injection timing investigated at this
operating point. The NOx emissions at the most advanced
injection timing of 65.2 bTDC was below the US2010 emission
legislation, indicating the potential of this concept in eliminating
NOx aftertreatment. A summary of the operating condition and
observed emissions at the injection timing of 65.2 bTDC has
been provided in Table 2.
Table 2. Operating conditions and performance for RCCI at 6 bar
BMEP and 1500 rpm
It could be shown that with further advanced timings of up to
80 deg bTDC the combustion phasing did not exhibit further
change. However, an increase in CO emissions was observed
with further timing advancement, possibly due to the
impingement of liquid fuel on the cylinder walls.
Figure 22 shows an optimized CNG substitution map when
RCCI combustion was included at the low load points. The
average CNG substitution was improved from approximately
50% to above 65% for the complete engine map.
Figure 23 shows the cost savings based on the optimized CNG
substitution map outlined in Figure 22. A maximum cost
savings of up to 40% can be achieved with advanced
combustion concepts tested at the lower speed and load
As outlined in Table 3, operation in dual fuel mode (that makes
use of RCCI at low loads) allows, on average, a 43% reduction
in NOx emissions, a 68% reduction in PM emissions and a
22% reduction in fuel costs.
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
Additional work is ongoing to identify the factors that contribute
toward achieving RCCI combustion at different speed and load
points. Results from these investigations will be further
explored and discussed in a future publication.
Figure 22. Optimized CNG substitution map with calibration changes,
including RCCI combustion at low loads and meeting US2010 emission
Figure 23. Cost savings in percentage from baseline diesel operation
Table 3. Average cost savings and emissions reduction in percent,
obtained from the optimized CNG substitution map that includes RCCI
combustion at low loads
1. Assuming no access to the base engine calibration, a
baseline maximum substitution map with practical emissions
was created. It was found that the cost and emission
benefts of running on CNG were limited at lower loads due
to high hydrocarbon emissions and at higher loads due to
peak cylinder pressures and turbine inlet temperatures.
2. In an optimization step, strategies were developed
primarily involving diesel injection timing change and CNG
substitution level. Advancing the diesel injection timing at
lower loads led to a reduction in THC emissions. However,
at higher loads, slightly retarded injection timing helped
in achieving higher CNG substitution. The optimized
substitution map achieved an average of 49% substitution
across the test point map with 15% reduction in NOx and
43% reduction in PM emissions. At higher engine speeds,
it appears the reduced time for CNG/air mixing results in a
varying cylinder-to-cylinder distribution of CNG, and thus
deteriorated combustion stability.
3. RCCI combustion could be achieved at low loads,
which enabled even higher CNG substitution and lower
emissions. A maximum of 50% net indicated thermal
effciency was observed at 6 bar BMEP load point along
with 75% reduction in both NOx and PM emission. The
potential benefts of RCCI combustion were limited due
to the un-optimized combustion chamber design and high
compression ratio.
Future Work
This work shows that an RCCI combustion strategy for burning
CNG and diesel fuel is most promising in terms of NOx
emission control and fuel consumption reduction. However,
challenges remain that must be resolved to allow extended use
of RCCI combustion at higher loads. Studies are currently
ongoing to explore the factors that affect the ability to achieve
RCCI combustion at different speed and load points. Along with
calibration optimization, changes in engine design, including
piston bowl design, compression ratio and nozzle
confguration, are being considered. Port fuel injection of CNG
in an effort to reduce the cylinder-to-cylinder variation observed
with RCCI is also being investigated. Solutions enabling RCCI
combustion and maximizing CNG substitution in dual fuel
mode must also respect the option for diesel-only operation
when CNG is not available.
1. US Energy Information Administration, “Natural Gas,” http://, Aug 2013
2. Ribas X., “Heavy Duty liquefied Natural Gas engine
developments to meet future emissions requirements,
methodology and real application,” FISITA paper
F2010F013, 2010.
3. Ouellette, P., “Cummins Westport Spark-Ignited (SI)
and High Pressure Direct Injection (HPDI) Natural Gas
Engines,” presented at NGVTF 2003, USA, January 28-29,
4. Barroso Guzman, P., Ribas, X., García, J.Sr., and Pita,
M.Sr., “PM and CO2 Reduction in a Dual-fuel Heavy-
duty Diesel Engine during the Freeway Part of Transient
Worldwide Emission Tests,” SAE Technical Paper 2013-01-
2759, 2013.
5. Maxey, C., Kalaskar, V., Kang, D., and Boehman, A.,
“Impact of Supplemental Natural Gas on Engine Efficiency,
Performance, and Emissions,” SAE Technical Paper 2013-
01-0847, 2013, doi:10.4271/2013-01-0847.
6. Sun, L., liu, Y., Zhou, L., and Zeng, K., “Experimental
Investigation of Cycle-by-Cycle Variations in a Natural Gas/
Diesel Dual Fuel Engine with EGR,” SAE Technical Paper
2013-01-0853, 2013, doi:10.4271/2013-01-0853.
7. Aroonsrisopon, T., Salad, M., Wirojsakunchai, E.,
Wannatong, K. et al., “Injection Strategies for Operational
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
Improvement of Diesel Dual Fuel Engines under Low Load
Conditions,” SAE Technical Paper 2009-01-1855, 2009,
8. Kowalewicz A., and Woloszyn R., “Comparison of
performance end emissions of turbocharged CI engine
fuelled either with diesel fuel or CNG and diesel fuel,”
Combustion Engines, PTNSS-2011-SC-117, 2011.
9. Papagiannakis, R., Hountalas, D., and Kotsiopoulos, P.,
“Experimental and Theoretical Analysis of the Combustion
and Pollutants Formation Mechanisms in Dual Fuel DI
Diesel Engines,” SAE Technical Paper 2005-01-1726,
2005, doi:10.4271/2005-01-1726.
10. Papagiannakis, R., Hountalas, D., Rakopoulos, C.,
and Rakopoulos, D., “Combustion and Performance
Characteristics of a DI Diesel Engine Operating from Low to
High Natural Gas Supplement Ratios at Various Operating
Conditions,” SAE Technical Paper 2008-01-1392, 2008,
11. Nieman, D., Dempsey, A., and Reitz, R., “Heavy-Duty
RCCI Operation Using Natural Gas and Diesel,” SAE Int. J.
Engines 5(2):270-285, 2012, doi:10.4271/2012-01-0379.
Contact Information
Koehler, Erik
Manager, Performance and Emissions
Commercial Engines; FEV, Inc.
Dahodwala, Mufaddel
Senior Engineer, Performance and Emissions
Commercial Engines; FEV, Inc.
The authors would like to thank FEV, Inc. management for
encouraging this research effort and providing the resources
necessary to accomplish its goals. The authors would also like
to thank those who took the time to review this paper and
provided valuable feedback.
AC - Alternating Current
AHRR - Apparent Heat Release Rate
BMEP - Brake Mean Effective Pressure
BSC - Brake Specifc Cost
BSFC - Brake Specifc Fuel Consumption
bTDC - Before Top Dead Center
CH4 - Methane
CNG - Compressed Natural Gas
CO - Carbon Monoxide
CO2 - Carbon Dioxide
COV - Coeffcient of Variance
DOC - Diesel Oxidation Catalyst
DPF - Diesel Particulate Filter
ECU - Engine Control Unit
EGR - Exhaust Gas Recirculation
EIA - Energy Information Administration
EOL - End Of Life
FTP - Federal Test Procedure
GGE - Gasoline Gallon Equivalent
HD - Heavy-Duty
IMEP - Indicated Mean Effective Pressure
MIRA - Motor Industry Research Association
NMHC - Non-Methane Hydrocarbon
NOx - Nitrogen Oxides
OBD - On Board Diagnostics
PM - Particulate Matter
RCCI - Reactivity Controlled Compression Ignition
RCP - Rapid Controller Prototyping
RMC - Ramped Modal Cycle
SCR - Selective Catalytic Reduction
SI - Spark Ignited
TDC - Top Dead Center
THC - Total Hydrocarbons
WHTC - World Harmonized Transient Cycle
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
= Mass fow rate of CNG
= Lower heating value of CNG
= Mass fow rate of diesel
Brake specifc cost is a parameter that accounts for the amount of cost savings that one can get through CNG substitution. The average
national cost of CNG (2.09$/GGE) and Diesel (3.84$/Gallon) in May 2013 were used for this calculation.
= Average National Cost of CNG in $/Kg (1 GGE CNG = 2.567 Kg CNG)
= Average National Cost of Diesel in $/Kg (1 US Gallon Diesel = 3.149 Kg Diesel)
P = Engine Power (W)
Optimum CNG substitution at different operating points was calculated using the composite desirability index for multiple response
optimization. The target, maximum values and weight factors were selected for NOx, PM, NMHC, CO, BSC and ringing intensity
according to regulation limits.
Desirability can be calculated for individual responses by the following formula:
The composite desirability can then be expressed as:
Y = Response (NOx, PM, BSC…etc)
w = Weighing factor
n = Number of Responses
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM
CD = Composite Desirability Index
d = Desirability
The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE’s peer review process under the supervision of the session
organizer. The process requires a minimum of three (3) reviews by industry experts.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical,
photocopying, recording, or otherwise, without the prior written permission of SAE International.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the
ISSN 0148-7191
Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM