Progress in Energy and Combustion Science

Progress in Energy and Combustion Science 90 (2022) 100995
Contents lists available at ScienceDirect
Progress in Energy and Combustion Science

journal homepage: www.elsevier.com/locate/pecs
Synthesis gas as a fuel for internal combustion engines in transportation

Amin Paykani a *, Hamed Chehrmonavari a, Athanasios Tsolakis b, Terry Alger c,
William F. Northrop d, Rolf D. Reitz e
a
School of Physics, Engineering and Computer Science, University of Hertfordshire, Hatfield, AL10 9AB, UK
b
Department of Mechanical Engineering, School of Engineering, University of Birmingham, Birmingham B15 2TT, UK
c
Automotive Propulsion Systems, Powertrain Engineering Division, Southwest Research Institute, TX, USA
d
Department of Mechanical Engineering, University of Minnesota, 111 Church St. SE Minneapolis, MN, 55419, USA
e
Engine Research Center, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53705, USA
A R T I C L E I N F O A B S T R A C T
Keywords: The adverse environmental impact of fossil fuel combustion in engines has motivated research towards using
IC engine alternative low-carbon fuels. In recent years, there has been an increased interest in studying the combustion of
Syngas fuel mixtures consisting mainly of hydrogen and carbon monoxide, referred to as syngas, which can be
Fuel reforming
considered as a promising fuel toward cleaner combustion technologies for power generation. This paper pro-
Combustion
Efficiency
vides an extensive review of syngas production and application in internal combustion (IC) engines as the pri-
Emissions mary or secondary fuel. First, a brief overview of syngas as a fuel is presented, introducing the various methods
for its production, focusing on its historical use and summarizing the merits and drawbacks of using syngas as a
fuel. Then its physicochemical properties relevant to IC engines are reviewed, highlighting studies on the
fundamental combustion characteristics, such as ignition delay time and laminar and turbulent flame speeds. The
main body of the paper is devoted to reviewing the effect of syngas utilization on performance and emissions
characteristics of spark ignition (SI), compression ignition (CI), homogeneous charge compression ignition
(HCCI), and advanced dual-fuel engines such as reactivity-controlled compression ignition (RCCI) engines.
Finally, various on-board fuel reforming techniques for syngas production and use in vehicles are reviewed as a
potential route towards further increases in efficiency and decreases in emissions of IC engines. These are then
related to the research reported on the behavior of syngas and its blends in IC engines. It was found that the
selection of the syngas production method, choice of the base fuel for reforming, its physicochemical properties,
combustion strategy, and engine combustion system and operating conditions play critical roles in dictating the
potential advantages of syngas use in IC engines. The discussion of the present review paper provides valuable
insights for future research on syngas as a possible fuel for IC engines for transport.
minimize greenhouse gas (GHG) emissions on a well-to-wheel basis to

meet worldwide emissions legislation [4]. The next generation of
1. Introduction cleaner fuels is governed by several important factors, including the
availability of the feedstock, production efficiency and environmental
Worldwide, approximately 80% of useful energy, including high- impact, distribution and storage, safety issues, integration and
quality heat, propulsion work, and electricity, is produced by compatibility with engines, vehicle, and transportation systems – costs,
combustion-driven processes using fossil fuels. The main detriments of regional variations, standardization and public acceptance [5,6]. In this
the application of fossil fuels in IC engines are their contribution to regard, natural gas [7,8], hydrogen [9] and ammonia [10,11] and
global warming and environmental pollution, and therefore govern- optimized engines [12] have recently drawn much attention from en-
ments are encouraging the use of zero/low-carbon fuels in power and gine researchers to meet the future emission standards.
propulsion systems for achieving energy security and meeting emissions Hydrogen (H2) has been considered as a viable energy carrier for fuel
targets [1,2]. In the coming decades, both IC engine efficiency and fuel cells and future IC engines. The wide flammability range of hydrogen
technology must improve and contribute to achieving carbon-neutral makes it suitable for engine operation over a wide range of air-fuel
transportation [3]. Engine technology must also use fuels that
* Corresponding author:
E-mail address: a.paykani@herts.ac.uk (A. Paykani).
https://doi.org/10.1016/j.pecs.2022.100995
Received 30 June 2020; Received in revised form 3 February 2022; Accepted 6 February 2022
Available online 18 February 2022
0360-1285/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
A. Paykani et al. Progress in Energy and Combustion Science 90 (2022) 100995
Nomenclature HRR Heat release rate

HTHR High-temperature heat release
Greek HWFET Highway Fuel Economy Test
λ Air–fuel equivalence ratio ICE Internal combustion engine
ϕ Fuel–air equivalence ratio IDT Ignition delay time
γ Ratio of specific heats IGCC Integrated gasification combined cycle
θ Crankshaft angle IMEP Indicated mean effective pressure
η Efficiency ITE Indicated thermal efficiency
LFS Laminar flame speed
Abbreviations LD Light duty
AFR Air fuel ratio LFG Landfill gas
AHRR Apparent heat release rate LHV Lower heating value
AKI Anti-knock index LPG Liquefied petroleum gas
ATDC After top dead center MAP Manifold absolute pressure
ATR Autothermal reforming MBT Maximum brake torque
BMEP Brake mean effective pressure MES Methane enriched syngas
BSFC Brake specific fuel consumption MFB Mass fraction burned
BTDC Before top dead center ML Machine learning
BTE Brake thermal efficiency MN Methane number
CA Crank angle NIMEP Net indicated mean effective pressure
CDC Conventional diesel combustion NOx Nitrogen oxides
CFD Computational fluid dynamics NTC Negative temperature coefficient
CFR Cooperative fuel research NVO Negative valve overlap
CHP Combined heat and power PG Producer gas
CI Compression ignition POX Partial oxidation
CN Cetane number PPM Parts per million
CNG Compressed natural gas PRF Primary reference fuel
CO Carbon monoxide PRIEMER Premixed mixture ignition in the end-gas region
COV Coefficient of variation RCCI Reactivity controlled compression ignition
CR Compression ratio R-EGR Reformed exhaust gas recirculation
CRI Common rail injector RG Reformer gas
D-EGR Dedicated exhaust gas recirculation RMG Reformed methanol gas
DFBG Dual fluidized bed gasifier RON Research octane number
DF Dual fuel SAC Superadiabatic combustion
DI Direct-injection SAE Society of Automotive Engineers
DISI Direct injection spark ignition SEC Specific energy consumption
DME Dimethyl ether SI Spark ignition
DNS Direct numerical simulation SMR Steam methane reforming
EGR Exhaust gas recirculation SOC Start of combustion
EPA Environmental protection agency Syngas Synthesis gas
ERC Engine research center SZM Single zone model
FTP Federal test procedure SwRI Southwest Research Institute
FT Fischer –Tropsch TCR Thermochemical recuperation
GA Genetic algorithm TFR Thermochemical fuel reforming
GDI Gasoline direct injector TFS Turbulent flame speed
GHG Greenhouse gas emissions THC Total hydrocarbons
GIE Gross indicated efficiency TIDR Total inert dilution ratio
GT Gas turbine TWC Three-way catalyst
GTL Gas-to-liquid UHC Unburned hydrocarbons
HCCI Homogenous charge compression ignition WGS Water–gas shift
HD Heavy-duty WHR Waste heat recovery
HE Hydrous ethanol WOT Wide-open throttle
HOME Honge oil and its methyl ester
mixtures, and the lean mixture operation can increase the fuel economy In contrast, hydrogen’s high auto-ignition temperature (858 K) ne-
of the hydrogen-fueled engines. Moreover, its high diffusivity and flame cessitates the use of a spark plug or a supplementary low auto-ignition
speed result in a faster uniform fuel/air mixture and an improved temperature fuel. On the other hand, its low ignition energy increases
combustion inside the cylinder [13,14]. The carbon-free structure of the the propensity of phenomena like pre-ignition, backfire and knock. NOx
hydrogen also makes it an excellent fuel for the clean and efficient emissions can also be increased due to the higher combustion temper-
operation of IC engines. Fig. 1 shows the energy density of various fuels atures associated with hydrogen combustion in IC engines [15].
based on lower heating values (LHV) compared to hydrogen. As can be Hydrogen can be produced in various ways to be stored in a pres-
seen, hydrogen has nearly three times the energy content of gasoline on surized tank on-board in a vehicle aiming at improving both combustion
a mass basis. and aftertreatment processes [17]. The comparison of hydrogen usage in
2
stored liquid fuels on vehicles can mitigate many of the hydrogen stor-
age and combustion challenges [27–31], and this will be extensively
discussed in the upcoming sections.
1.1. What is syngas?
Syngas (an abbreviation for synthesis gas) can be derived by gasifi-

cation processes from natural gas, heavy oil, biomass, and coal (carbon-
containing fuels) for stationary application or by chemical processes
from liquid and gaseous fuels for vehicular application, in which a
complex and long-chain molecule converts to simpler molecules
including H2 (hydrogen), CO (carbon monoxide), and CH4 (methane),
CO2 (carbon dioxide), H2O (water vapor), and N2 (nitrogen). Based on
the process adopted and the type of feedstock used, end-products as a
form of syngas will have varying compositions and heating values, thus
different terms such as “town gas”, “wood gas”, “water gas”, “producer
gas”, “reformer gas”, “power gas”, and etc. have been used in the litera-
ture. A possible range for the volumetric composition of syngas is shown
in Table 2, indicating that it is comprised mainly of H2 and CO. The
preferred composition generally depends upon the purpose of syngas
usage, feedstock, and production cost. For example, if the goal is
hydrogen production by a biomass gasification process, much more cost
and effort are needed for adequate purification than the reaction
Fig. 1. Gravimetric and volumetric density of hydrogen and other fuels [16]. reforming process.
Syngas can also be produced through fuel reforming with steam and/
vehicles through on-board hydrogen production, hydrogen storage, and or air using compact systems that can be located on-board a vehicle [27,
stationary power generation systems is summarized in Table 1. The low 33,34]. External on-board and in-cylinder fuel reforming techniques
ignition energy and volumetric energy density make storage of hydrogen have been identified as possible pathways to produce syngas on-board in
on-board a challenging task [18,19]. Hydrogen as a low-density fuel IC engines to improve efficiency and emissions [33]. The details of
needs to be compressed at high-pressure tanks (350–700 bar) to allow an on-board fuel reforming and syngas production will be covered in sec-
adequate driving range of more than 500 km with safety and cost to be tion 4.
the main challenging issues. Difficulties involved in the storage and
transport of hydrogen still present barriers to the commercial use of 1.2. Syngas applications
hydrogen as a secondary or a main fuel for IC engines [19,20]. However,
hydrogen is currently viewed as the most promising clean energy carrier Syngas can be used as an intermediate to create other attractive clean
of the future, and the hydrogen infrastructure is now rapidly expanding fuels such as ammonia (NH3), dimethyl ether (DME), and methanol
in many counties, both in the EU and outside. This has sparked research (CH3OH) [35–39]. Another use of syngas is as a primary chemical
interest in the energy conversion technologies, particularly fuel cells building block in petrochemical and refining processes [40]. Syngas can
that are capable of using hydrogen as the fuel [21]. be used in many different applications, including:
While the merits of operating IC engines with varying concentrations
of hydrogen have been well documented [25], the key challenges • Power generation in existing coal power plants [22]
remain unresolved, including practical approaches to hydrogen pro- • Combined heat and power (CHP) plants [41]
duction without sacrificing system efficiency and/or sufficiently • Integrated gasification combined cycles (IGCC) [42]
addressing its adverse combustion properties [26]. Syngas is considered • Fuel cells [43–46]
an intermediate step in the transition from carbon-based fuels to • Production of transportation fuels from gas-to-liquid (GTL) processes
H2-based fuels, since it is composed mainly of hydrogen (H2) and carbon [47]
monoxide (CO). In IC engines, using syngas produced from on-board • Gas turbines [48–50]
• Directly as the primary or secondary fuel in IC engines [51]
Table 1
Comparison of hydrogen usage in vehicles through on-board hydrogen pro-
duction, hydrogen storage, and stationary power generation systems [22–24].
Transportation (on-board Transportation Stationary power
in vehicles) (storage in vehicles) generation Table 2
A possible range for syngas composition on a volume percent basis [29,32].
• Ability to operate on H2 • Ability to operate on • Ability to operate on
partly 100% H2 100% H2 Species Min Max
• More complicated design • Simple design • Simple design
• Less space required • More space required • No space problem H2 8.6 61.9
• Few safety concerns • Serious safety • No safety problem CO 22.3 55.4
• Challenge with the catalyst concerns • Hydrogen can be CH4 0.0 8.2
activation under a wide • Challenge with produced or can be CO2 1.6 30.0
range of temperature and hydrogen production, used from the N2+Ar 0.2 49.3
reactants changes store and distribution storage system H2O - 39.8
• Moderate cost • Highest cost • Lowest cost H2/CO 0.33 2.36*
• Fast response during • Fast response during • Aesthetic issues are LHV (MJ/m3) 5.02 12.57
transient operation is a transient operation not important *
The broad experimental studies of syngas-fed IC engines reviewed in sec-
challenge
tions 3 and 4 have been performed with simulated syngas whose H2/CO ratio
• Aesthetic issues
can go beyond this value, i.e., H2/CO: 3/1 (by vol.).
3
1.3. Syngas production methods power generation. The biomass used consisted of bones and meat resi-
dues sampled directly from the industrial line, characterized by high
Solid fuel (coal, biomass, wastes) feedstocks are sources for syngas water content, about 42% in mass, and potential health risks. They re-
production [52,53] by traditional gasification methods [54,55]. Also, ported that syngas with a composition of 19.1%CO, 17%H2, and 1.6%
liquid and gaseous feedstocks can be converted to syngas by CH4 was produced.
non-gasification methods [56] that comprise one or a chain of reforming New gasification concepts such as dual fluidized bed gasifiers (DFBG)
processes using the catalysts; such processes are expected to be viable [76], plasma gasification [77–80], and supercritical water gasification
in-situ methods of syngas production for automotive engines. In this [81,82], have demonstrated the potential to improve syngas quality
section, only a brief review of different syngas production methods is from biomass gasification. However, they are not of significant interest
presented since several comprehensive review papers [57,58] explain at present, and are more applicable for stationary power generation.
them in detail and emphasize the IC engine operation with gasification Moreover, wood and solid fuels seem to be not effective for transport
fuel [59,60] and non-gasification fuel [33]. applications due to their low energy density. A further analysis on this
topic is presented in section 3.
1.3.1. Gasification processes
Biomass gasification converts a solid fuel into syngas which can be 1.3.2. Non-gasification processes
burnt in stationary gas turbines and IC engines. Gasification is a ther- Non-gasification processes include catalytic fuel reforming, where
mochemical process that converts a solid hydrocarbon feedstock like the parent fuel reacts with steam and/or oxygen in a heterogeneous
coal, petroleum coke, refinery residuals, biomass, and municipal solid process to produce syngas [27,83]; and non-catalytic processes like
waste, into syngas without consuming a large part of its heating value plasma-assisted reforming [84]. The three main global reactions in fuel
[58]. In gasification, the carbon-containing feedstock reacts with a reforming are partial oxidation (POX), steam reforming (SR), and
gasifying agent like oxygen, steam, and air, which breaks down the auto-thermal reforming (ATR) [47,85], which will be explained in the
mixture into syngas, as illustrated in Fig. 2 [54,57,61,62]. The type and following.
design of the gasifier, gasification temperature, type and flow rates of When a mixture of sub-stoichiometric oxygen and feedstock reacts to
the feedstock and oxidizing agents, and type and amount of catalysts produce syngas, this self-sustaining chemical reaction is called “partial
[63], are the main parameters in influencing the gasification process and oxidation (POX),” It is an exothermic reaction. The complete combustion
syngas production [64–70]. Oxygen as the gasifying agent produces reaction forming CO2 is restricted due to insufficient O2 in the reactants
syngas called “medium syngas,” and if air is used, then it is called “pro- [86,87]. In endothermic steam reforming (SR), a high-temperature
ducer gas” [71]; the following sub-section will clarify the terms used in mixture of fuel and steam is sent through a bed of catalyst. The heat
the rest of the paper. Using gasification methods, in addition to H2 and required to drive the endothermic reaction is provided externally (e.g.,
CO, other typical products including H2O, CO, CH4, unwanted tars, electric heaters or combustion of fuel). The syngas stream from the
small amounts of NH3, and H2S are also produced. The relative amount steam reforming can then flow to a water gas shift (WGS) reactor, where
of each species depends on the feedstock and the gasification process additional steam converts CO to CO2 and generates additional H2
[57]. Modeling approaches for biomass gasification have been exten- [88–91].
sively reviewed in [72]. When both the POX and SR reforming reactions are combined in a
For syngas production, a typical solid fuel gasification system single reactor, it is called autothermal reforming (ATR). Industrially, the
coupled to spark ignition and dual-fuel diesel engines is illustrated in ATR reactor is implemented in a refractory-lined pressure vessel with a
Fig. 3 [73]. The system includes a gasifier, a gas clean-up and cooling combustion chamber, burner, and catalyst bed. In this reactor, a sub-
system, a gas mixer, a starting blower, and the engine. The engine draws stoichiometric environment oxidizes fuel, natural gas in most cases,
air from the gasifier during the intake stroke, through the cleaning with O2. In the ATR reaction, POX provides heat for later SR reactions,
system, and from the gas mixer where air is combined with the syngas thus controlling the exit temperature of the reactor. Endothermic dry
[74]. For example, Marculescu et al. [80] used food processing industry reforming (DR) where the CO2 is reacting with hydrocarbons feedstock
waste for energy conversion, using gasification and an IC engine for to create H2 and CO can also be employed [92–95]. The advantages and
Fig. 2. Schematics of biomass gasification process [66].
4
Fig. 3. a) Syngas production system for an SI engine (Reprinted from [73] with permission of Elsevier); b) diagram of the entrained-flow gasifier for power gen-
eration on dual-fuel diesel engine (Reprinted from [75] with permission of Elsevier).
5
disadvantages of these processes and the corresponding chemical re- 1956 by Szezich [108]. Later, further interest in biomass-based fuel
actions are summarized in Table 3 as main reforming reactions and in application in engines was sparked by the oil crisis in the mid-1970s
Table 4 as side reactions that may occur. These fuel reforming processes [109]. Simultaneously, in response to the oil price shocks, US
have been proposed for on-board hydrogen-rich gas production in Congress passed the first national standards for tailpipe-out emissions to
vehicular applications, discussed in detail in section 4. Note that the goal increase miles per gallon (mpg) of passenger vehicles within ten years
is high hydrogen content; therefore, understanding the nature of these [110], which was the onset of adopting further stringent regulations for
reactions aids us in deciding which reactions should participate in oil-based fuels. As already mentioned, by that time, syngas use was
reforming processes within a designed reactor and which ones to avoid. mainly limited to IGCC for stationary power generation [40,111,112],
and as an intermediate to produce other biofuels [113,114].
1.3.3. Syngas as an IC engine fuel On the other hand, fuel reforming has been used for producing
Syngas as an IC engine fuel can be utilized in power generation and syngas industrially for more than 70 years [115], but the concept of
transportation, produced mainly by gasification and reforming pro- compact on-board reforming for engine applications dates back to the
cesses, respectively. As mentioned above, syngas names would be 1970s [116]. Investigations carried out in 1973–1975 in the US with a
different based on composition and LHV, while the terminology we use Chevrolet car with an engine equipped with a syngas generator
throughout the current work can be seen in Fig. 4, taken from IC engine- demonstrated a decrease in petrol consumption by 26% when driving
relevant published works. It is worth noting that the figure does not according to the Federal Drive Cycle CVS-3 [117]. On-board fuel
encompass all viable routes for syngas production. reforming was first practically applied to a carbureted gasoline engine
As seen in the first row of Fig. 4, producer gas (PG) or low heating through a project named “Boston Reformed Fuel Car” by Newkirk and
value (LHV) fuel [59] is a gasification product with low LHV due to Abel [118], which had an on-board non-catalytic system operated at
using air as the gasifying agent and, therefore, the presence of high N2 higher than T = 1100 ◦ C for the steam reforming reactions of gasoline.
content in products. With steam or oxygen or a combination of both Martin [119] at the University of Arizona demonstrated a better concept
utilized as the gasifying agent, the gasification product can be upgraded of this project using a catalyst to reduce the temperature of the steam
to fuel with a medium LHV. In general, the gasification process suffers reforming of gasoline to around T = 620 ◦ C. In the meantime, Houseman
from low thermal efficiencies (~ < 50%) due to the extra step of and Cerini [120] from the Jet Propulsion Laboratory (JPL) evaluated
vaporizing the moisture contained in biomass and then from tar in the on-board hydrogen generation by adding partial oxidation reforming to
products [101]. Noting that in addition to gas constituents, tar and steam reforming along with the catalyst to perform the whole reforming
residue co-exist in the end-product of gasification, but their amounts process for both hydrocarbon and alcohol feedstocks fueled SI engine. A
reduce as high-temperature steam reforming is employed, as seen in series of works from Lindström and Sjöström [121–123] at the Royal
Fig. 5. Thus, gasification fuel presumably creates concerns of NOx Institute of Technology (Sweden) was seemingly the first investigation
emissions (due to probably high N2 content) and deposit formation in on directly using the recirculated exhaust gases as the primary heat
the combustion chamber, which should be considered when this fuel is supplier for driving the endothermic reactions of the steam reforming
being used as an IC engine fuel. Several IC engine related studies, process and as species that can participate in the reforming process. At
particularly with SI engines, with gasification fuels have been investi- the University of Birmingham, Jones [124,125] evaluated the feasibility
gated so far, which will be discussed in section 3. of this concept as exhaust-gas reforming in more detail by incorporating
In the second row of Fig. 5, reformate or reformer gas (RG) is a the exhaust gases and using the entire fuel feedstock in the reforming
reforming reaction product, which often has high H2 and CO contents process. For more reading on this topic, there are two notable reviews in
with negligible N2 and high heating value. RG can be economically the literature [126,127], and additional discussions are provided in
produced with in-situ methods, which will be discussed partially in section 4.
section 3 (sub-sections of 3.2.2, and 3.3) and comprehensively in sec- The development of syngas-fueled IC engines was not commercial-
tion 4 (on-board fuel reforming technologies). ized because of the low lifespan of the catalysts and tightening of NOx
The overall purpose of using syngas as a primary or secondary IC emission standards. However, research on syngas as a potential fuel for
engine fuel is to benefit from renewable energy sources and clean power IC engines in transportation has gained much interest in recent decades
conversion technologies without scarifying efficiency, power derating, with advances in catalysts and the potential for on-board generation
and pollutant emission formation of engines fueled with syngas [102, through waste heat recovery (WHR). An average of 30% substitution of
103]. However, its advantages are dependent on the physicochemical fossil fuels by syngas has been proposed, based on extrapolating from the
properties of syngas (i.e., the amount of CO and H2) (see section 2), the current EU Directives [128,129]. For further information, several review
type of combustion engine (see section 3), and the fuel reforming papers [27,51,59] and a book chapter [29] have been published on the
technique used for on-board syngas production (see section 4). historical and trends of syngas for power generation.
1.4. Historical use of syngas as an engine fuel

1.5. Scope and structure of the current review
The first application of syngas in vehicles was reported in the 1920s
when German engineer Georges Imbert developed a wood gas generator There are many studies in the literature on the use of syngas in
for automobile use [104]. The gases were cleaned and dried, and then different types of IC engines, and it has been shown that syngas can be
fed into the vehicle’s combustion engine. Later, during World War II, used as a renewable and alternative low-carbon fuel for both spark
shortages of petroleum fuels led to further development of wood gas ignition (SI) and compression ignition (CI) engines [130–133]. How-
vehicles [105,106]. By 1945 syngas was used for trucks, buses, agri- ever, there are still numerous subjects such as on-board syngas pro-
culture, energy generation, and industrial machinery. After the end of duction and application in advanced combustion strategies that require
World War II, the application of syngas was changed to integrated further investigation and review. This work is focused on determining
gasification combined cycles (IGCC) for stationary power generation. syngas’ potential as a renewable IC engine fuel in transportation. To do
This could be due to less interest in syngas usage in IC engines related to this:
the sharp drop in global oil price and intensified ambitions towards high
power engines by increasing the anti-knock index (AKI) of gasoline fuel. • Section 2 presents the fundamental physicochemical characteristics
The 1950s and 1960s are recalled as the “Power Wars” age [107]. The of syngas, which ultimately help us perceive the combustion
potential of syngas production and application in a low compression behavior of syngas as an IC engine fuel under engine-relevant con-
ratio SI engine fueled with a syngas-methane mixture was studied in ditions. Due to the composition variability of syngas, various H2/CO
6
Table 3
Advantages and disadvantages of the three main fuel reforming technologies [43,96-98].
Reforming Chemical formulaEnthalpy of Characteristics Advantages Disadvantage
process reaction (kJ mole−1) 1
Steam y Endothermic • Highest H2 generation (H2/CO ~ Highly endothermic

Cx Hy + xH2 O→xCO + x + H2 •
Reforming 2 3:1 by vol.) • Low rate of hydrogen production
(SR) ΔhR = +12591 • Need significant quantity of water
• Complex system regarding working in high
temperature
• Careful thermal management (S/C ratio
control)
• Less fuel flexibility (only certain fuel)
Partial x y Exothermic • Simple system (absence of external Lowest H2 generation
Cx Hy + O2 →xCO + H2 •
Oxidation 2 2 heat and water) and compactness • Lower heating value of reformate than base fuel
1
Reforming ΔhR = −676 • Rapid start • Catalyst’s deactivation issues by Sulphur and
(POX) • Fast response to temperature and coke depositions
reactant changes • Non-catalytic partial oxidation needs costly
• Not have thermal management materials because it operates at very high
issue temperatures
• High fuel-type flexibility
Autothermal z Combination of exothermic • Simple System (lack of water and • H2 and CO selectivity is small, because the
Cx Hy + zH2 O + x− O2 →
Reforming 2 and endothermic external heat requirements) device uses two different catalyst types
y
(ATR) xCO2 + z + H2 (thermally neutral) • Compact • During load changes and start-up, careful con-
2
• Quick to start trol system is required to balance SR and POX
1
With assuming that fuel is n-octane, and reactants and products are both at 273.15 K and 1 bar.
Table 4
Other side reforming reactions that may take place [99,100].
Reforming process Reaction Enthalpy of reaction(kJ mole−1) Notes
Dry reforming y ΔhR = +15881 Lower H2 yield than SR (H2/CO ~ 1:1 by vol.).
Cx Hy + xCO2 →2xCO + H2
(DR) 2 Likely to happen at high temperature and low pressure.
Combustion y y ΔhR = −51161 If enough O2 is available.

Cx Hy + x + O2 →xCO2 + H2 O
(complete oxidation) 4 2
Water-gas shift (WGS) CO + H2 O

↼
CO2 + H2 ΔhR = −411 A route to CO purification and thus increased H2 yield at low temperatures.
⇁
Methanation 1 ↼
CO + 3H2 CH4 + H2 O ΔhR = −2061 H2 will be consumed to increase CH4 yield.
⇁
1
Methanation 2 ↼
CO2 + 4H2 CH4 + 2H2 O ΔhR = −165
⇁
Boudouard ↼
2CO2 CO2 + C ΔhR = −1721 Undesired reaction in catalytic reforming
⇁
Thermal Decomposition C8H18 → CH4 + C7H14 Endothermic May occur on the catalyst and increase CH4 yield.
Hydrogenolysis C8H18 + H2 → CH4 + C7H16 Exothermic May occur on the catalyst and increase CH4 yield.
1
Assuming that the fuel is n-octane, and reactants and products are both at 273.15 K and 1 bar.
Fig. 4. Syngas terminology used in the current review paper.
7
• The transition from syngas as an IC engine fuel to syngas as a

transportation fuel requires producing and carrying the syngas on-
board a vehicle, such that both fuel economy and engine efficiency
can increase. We discuss all proposed on-board fuel reforming
technologies compatible with IC engine in section 4 to achieve these
goals, especially paying attention to waste heat recovery (WHR)
from engine exhaust.
• The final section represents the summary and demonstrates sugges-
tions to make a feasible future for syngas as a transportation fuel.
2. Fundamentals
2.1. Physicochemical properties of syngas relevant to engines
Table 5 lists the physicochemical properties of typical syngas mix-

tures compared to hydrogen, methane, biogas, and conventional fossil
fuels. It is apparent that syngas has the highest laminar flame speed and
widest flammability limits amongst all fuels after hydrogen, making it
beneficial for application as a secondary fuel in IC engines, especially at
low engine loads. It means that a wide range of stable operation (from
rich to lean) could be achieved with syngas fuel, especially promoting
the ultra-lean combustion in spark-ignition engines. The volumetric
energy content is the main parameter determining the fuel injection
system design and required storage tank for vehicular applications
[134]. In the case of syngas use in IC engines, the fuel injection system
size should be larger, or the injection duration should be longer.
Although there is no available data for the minimum ignition energy
of syngas mixtures, it can be concluded that the high hydrogen content
in syngas could facilitate cold starts and guarantee rapid ignition [135],
considering the unwanted ignitions, specifically pre-ignition and back-
fire (see sub-section 3.1.1). A scarcity in data has also continued for the
Fig. 5. Yield of product for biomass gasification with thermal decomposition
and steam reforming (Reprinted from [101] with permissions of Elsevier). quenching distance, and a lower value means a more complete com-
bustion [136], but with a higher heat transfer losses.
Detailed combustion characteristics of syngas mixtures for gas tur-
mixtures are considered in the first place, then the effects of diluents
bine engine-relevant conditions have already been reviewed by Lee et al.
are discussed.
[137] and Jithin et al. [138]. In the following section, characteristics of
• To better understand the effects of syngas on combustion, emissions,
the ignition delay times, laminar and turbulent flame speeds of syngas
and performance of IC engines with conventional and advanced
mixtures under engine-relevant conditions are reviewed and discussed
combustion strategies, section 3 reviews the syngas application in SI,
with a focus on those aspects which have not been addressed carefully in
HCCI and dual-fuel engines. In SI engine section, the power gener-
previous reviews.
ation applications of the syngas-fueled engines are also included.
Table 5
Physicochemical properties of a typical syngas mixtures compared to other fuels [59,139-143].
Properties Syn11 Syn21* Syn31 Syn41þ Syn51þ Biogas Hydrogen Carbon Methane Gasoline Diesel
* monoxide
Density [kg/m 3] 2 0.54 0.67 0.68 1.05 1.04 1.11 0.0824 1.145 0.656 719.7 832
Molecular weight [kg/kmol] 2 13.91 15 15.2 23.2 34.4 2.0 28 16.04 103 200
Stoichiometric air/fuel ratio [kg/ 5.3 4.58 7.23 1.4 2.07 5.67 34.2 2.5 17.2 14.7 14.7
kg]
Flammability limits [vol.% in air] 24-60 6.06- 5.8- 7-21.6 13.4-58 7.5-14 4-75 12.5-74 5-15 1.4-7.6 0.6-
74.2 41.4 7.5
Flammability limit [ϕ] 0.2-7.2 - - - - - 0.1-7.5 0.3-6.8 0.4-1.6 0.7-4.3 1.0-
6.5
Autoignition temperature [K] 980 873-923 873- 898 898 923 858 882 813 550 589
923
3
Minimum ignition energy [mJ] - - - - - - 0.02 - 0.28 0.24 -
Laminar flame speed [m/s] 3 1.0 1.8 - 0.5 0.5 0.25 1.8-2.8 0.4 0.38 0.37- -
0.43
Adiabatic flame temperature [K] 3 2584 2385 2400 - 2200 2145 2390 2214 2214 2580 -
Quenching distance [mm] 3 - - - - - - 0.64 1.6 2.1 2.84 -
Lower heating value [MJ/kg] 15.7 17.54 24.4 5 7.47 17.0 119.7 10.1 50.0 43.4 42.6
Volumetric energy content [MJ/ 8.47 11.75 16.59 5.25 7.84 18.87 9.86 11.56 32.8 31235 35443
m3]
1
Syn1:57/43:H2/CO, Syn2:50/50:H2/CO, Syn3:40/40/20:H2/CO/CH4, Syn4: 22.6/24.3/2.2/9.3/41.2:H2/CO/CH4/CO2/N2, Syn5: 19.6/29.6/5.27/5.41/40.56:
H2/CO/ CH4/CO2/N2 and Biogas: 55.6/42.3/2.1: CH4/CO2/N2 (by volume).
2
at NTP: normal temperature (T=298.15 K) and pressure (p=1 bar). 3 at ϕ=1 (stoichiometry).
*
A typical reformate, + A typical producer gas
8
+ OH reaction. Those works were completely reviewed in [157,158].

Recently, in order to justify the observed discrepancies, Mansfield
and Wooldridge [159] indicated two distinct ignition behaviors attrib-
uted to the strong (or spontaneous ignition which is referred to as
spatially homogeneous ignition) and mild ignition (or deflagration
ignition which is characterized by the onset of localized reaction sites)
for a syngas, with a molar ratio of H2/CO = 0.7, mixed with air (O2/inert
gas molar ratio = 1/3.76) at ϕ = 0.1 and 0.5, T = 870-1150 K, and p =
3-15.4 bar through high-speed imaging in RCM experiments. Two
high-speed imaging frames of that work are depicted in Fig. 7, in which
multiple localized flame-like structures can be seen on the right side. The
authors [159] indicated that the ignition behavior is weakly affected by
the molar ratio of H2/CO, and, by comparison to other work [160], is not
strongly dependent on the combustion device used, albeit strongly
dependent on the initial thermodynamic state (i.e., pressure, tempera-
ture, and equivalence ratio). Also, they stated that the mild ignition
behavior is not avoidable by reducing the equivalence ratio towards lean
condition; however, the available chemical kinetics mechanisms used in
the zero-dimensional (0-D) homogeneous reactor modelling would be
capable of predicting the IDTs quite accurately.
Fig. 6. Discrepancies of ignition delay times between five homogenous reactor As mentioned earlier, a review by Lee [137] evaluated IDTs of syngas
models’ predictions and three experimental datasets for syngas (H2/CO: 50/50, mixtures at different H2:CO ratios with different pressures ranging from
by vol.)/air mixture at ϕ = 0.5 and p = 20.2 bar (Reprinted from [156] with 1.6 to 49 atm, by collecting data from Krejci et al. [161,162], Mathieu
permission of Elsevier). et al. [163], Vasu et al. [164], Keromnes et al. [165], Gersen et al. [166],
and Mittal et al. [152], as shown in Fig. 8. Relying on this review and
2.2. Ignition delay time (IDT) related works, the effects of various parameters on IDTs of syngas
mixtures are summarized below:
The ignition delay times, IDTs, of syngas mixtures were measured in Temperature and equivalence ratio (ϕ) effects:
earlier studies mostly using a shock-tube apparatus confined to low While increasing the temperature shortens the IDT substantially, the
pressure (up to 2.2 bar) and low-to-high temperature (up to 2850 K) IDT of the various compositions of H2/CO syngas at lean equivalence
conditions, which are less applicable to typical gas turbines and IC en- ratios (ϕ) has a decreasing trend, but only at higher pressures, e.g., 49
gines operating conditions [144–150]. The emergence of advanced in- atm, as seen in Fig. 8(h). Conversely, at lower pressures, e.g., 7.9 atm,
tegrated system technologies like syngas-fired IGCC as an alternative to the IDTs can have an increasing trend by reducing the ϕ, which is the
conventional coal-fired power plants, together with reported discrep- case of Fig. 8(b).
ancies between measurements and reaction model predictions of IDTs of Pressure effect:
syngas mixtures (see Fig. 6), has provoked researchers to reproduce The influence of pressure on IDT of 50:50 H2:CO at ϕ = 1.6, and at p
experimental measurements and refine the reaction kinetics for syngas = 8, 12, and 32 atm are shown in Fig. 9 (left). The complex behavior of
oxidation at higher pressures and temperatures [151,152]. Thus, some IDT can be seen with altering pressure at different temperature regions.
later studies were conducted using rapid compression machines (RCM) At high temperatures (>1250 K), higher pressures significantly shorten
[152–154] and shock-tubes [155,156] at higher pressures, suggesting the IDT, while at intermediate temperatures (1110 K<T<1250 K),
that the observed discrepancies between measured and calculated IDTs increasing pressure from 1.6 to 32 atm exhibits a mild decreasing and
at these pressures are due to not considering the overall activation en- even increasing decreasing trend for IDT, which implies that a compe-
ergy correctly, HO2/H2O2 chemistry dominance in the tition effect may exist. Actually, this was attributed to the pressure
chain-propagation regime, and particularly the role of CO + HO2 = CO2 dependence of H2 ignition [165], which is the competition between the
Fig. 7. High-speed imaging of homogenous (left) and inhomogeneous (right) ignition behaviors of syngas, with a molar ratio of H2/CO = 0.7, mixed with air (O2/
inert gas molar ratio of 1/3.76) illustrating uniform and non-uniform chemiluminescence (Reprinted from [159] with permission of Elsevier).
9
Fig. 8. Ignition delay times of different H2:CO mixtures as a function of temperature at different pressures (Reprinted from [137] with permission of Elsevier).
chain branching reaction (H + O2 → O + OH) that dominates in the concentration (i.e., CO < 60%), adding CO to H2 did not lead to a sig-
high-temperature region, and the pressure-dependent propagation re- nificant increase in IDT, as shown by Gersen et al. [166] (see Fig. 8 (a)).
action (H + O2(+M) → HO2(+M)) that dominates in the Further discussions regarding the CO’s inhibiting effect have been re-
low-temperature region. Moreover, the latter reaction is more favorable ported in [152,166].
at high pressures and is also pressure-dependent, inhibiting the chain CH4 effect:
branching and resulting in longer IDT at elevated pressures, as seen in Under engine-relevant conditions, it has been reported [167–169]
Fig. 9 (right). that the addition of H2 has ignition-promoting effect in CH4 mixtures,
CO effect: with substantially reduced IDT by replacing CH4 with a 50% mole
Though the H2-chemistry is dominant at low CO content, the CO has fraction of H2, particularly at higher temperatures. For the purpose of
an inhibiting effect on the IDTs of H2/CO fuel due to the decrease in IDT measurement in CH4/H2/CO/oxidizer mixtures, Gersen et al. [166]
activation energy. Lower activation energy can be detected in the high- carried out experiments and calculations at ϕ = 0.5 and 1, p = 20-80 bar,
temperature region, followed by a very steep increase at around T = and T = 900-1100 K conditions using an RCM and the SENKIN code
1000 K. The inhibition effect is much more prominent as pressure in- [170]. With the addition of CH4 (> 50% by mole fraction) to the
creases from ~15 to 30 bar; that is, the CO oxidation is retarded by H2/CO-oxidizer mixture, CO, with low mole fraction (≤ 30%), had no
increasing pressure. Beyond p=30 bar, the incremental variation in the inhibiting effect on either IDTs of CH4/H2/CO or CH4/CO, while CH4
inhibiting effect would be trivial. These mentioned trends can be seen in had this effect, particularly at high temperatures, as illustrated in Fig. 10
the labeled pictures of (a), (c), (e), and (g) in Fig. 8. For high H2 (a). In contrast to CO, CH4 with an even low mole fraction (~ 6%) in the
10
Fig. 9. Effect of pressure on the ignition delay times for 50:50 H2:CO at ϕ = 0.5 (left) and for a mixture of 0.7 H2 + O2 + 3.76 Ar (right) (Reprinted from [137,165]
with permission of Elsevier).
Fig. 10. Effect of CH4 on ignition delay times of syngas (Reprinted from [166,171] with permission of Elsevier).
(a) ϕ = 0.1, and p ~ 40 atm [166]
(b) ϕ = 0.1, and p ~ 15 atm [171]
fuel showed an ignition-inhibiting effect due to OH consumption by CH4 the data set:
+ OH = CH3 + H2O, as can be seen from Fig. 10 (b); noting that the
12500
ignition stage is depicted as a two-step ignition process, i.e., τign1 and τ τign = 3.7 × 10−6 P−0.5 ϕ−0.4 χ O2 −5.4 exp (1)
RT
ign2, due to the presence of two separate regions of fast pressure increase
on the recorded pressure-time history plot. This behavior has not been In this correlation, τign denotes the ignition delay time [ms], p pres-
explicitly reported before [171] for syngas fuel mixtures. The experi- sure [bar], T temperature [K], φ the equivalence ratio, and χ O2 the ox-
mental works measuring IDTs of syngas mixtures under engine-relevant ygen mole fraction. The experimental data were also in good agreement
conditions are summarized in Table 6. with model predictions based on the detailed mechanism proposed by
There have been several correlations proposed in the literature for Davis [151].
IDT calculation, which can provide a means to estimate syngas ignition As a final note, it is also important to study ignition delay times of
properties in a simplified form, where detailed kinetic models are syngas blended with other gases such as CO2, and H2O, since syngas is
computationally expensive. Walton et al. [154] performed ignition not suitable as a sole fuel for some type of IC engines (as will be dis-
studies of simulated syngas mixtures of H2, CO, O2, N2, and CO2 in an cussed in section 3). The effects of diluents like H2O [163,173], CO2
RCM. The IDT data spanned pressures between p = 7.1-26.7 bar, tem- (negligible) [163], impurities (e.g., trimethylsilanol) [171] and NH3
peratures from T = 855-1051 K, equivalence ratios from ϕ = 0.1-1.0, (negligible) [163] on IDT of syngas mixtures have been studied in only a
oxygen mole fractions from 15% to 20% and H2:CO ratios from 0.25 to few studies. It was suggested that H2O has contradictory impacts on the
4.0 (molar basis). Regression analysis yielded the following best-fit to IDTs under various pressures, increasing IDT at high pressures as
11
Table 6 Table 7
Available experimental datasets for ignition delay times of syngas mixtures. Measured MN for various test gases [182].
Year Author(s) Facility Mixture (by p T (K) ϕ (-) Test gas H2 CO CH4 CO2 N2 MN
vol.) (bar)
Reformer gas 44.5 2.30 38.1 13.0 2.10 62.4
2007 Mittal et al. RCM* H2/CO/O2/ 15- 950- 0.36- Coal gas 22.3 63.1 - 1.30 13.3 30.0
[152,153] N2/Ar 50 1100 1.60 Wood gas or PG 39.2 23.5 8.50 26.4 2.40 61.4
2007 Walton et al. RCM H2/CO/O2/ 7.1- 855- 0.1- Digester gas - - 60.8 37.8 1.50 139.1
[154] N2/CO2 26.4 1051 1.0 Landfill gas - - 60.5 39.5 - 139.5
2007 Peterson et al. ST* H2/CO/O2/ ~ 20 943- 0.5
[156] N2/CO2 1148
2012 Gersen et al. RCM CH4/H2/ 20- 900- 0.5 & comparing the tested fuels to reference blends, Malenshek and Olsen
[166] CO/O2/N2/ 80 1100 1
[182] conducted a direct comparison by developing a test procedure
Ar
2012 Krejci et al. ST H2/CO/O2/ 1.5- 960- 0.5
with a computer-controlled gas blending system, minimizing the effects
[161] Ar 32 2000 of other uncontrolled variables besides CR, such as ambient temperature
2013 Kéromnès et al. RCM & H2/CO/O2/ 1-70 914- 0.1- and atmospheric pressure, and producing maximum knock by sweeping
[165] ST N2/Ar 2220 4.0 the air-fuel ratio (rather than ϕ=1 for indirect comparison). Some results
2014 Mansfield and RCM H2/CO/O2/ 3-15 870- 0.1-
for a wide range of syngas composition are presented in Table 7.
Wooldridge N2 1150 0.5
[159] Experimental data, such as that represented in Table 7, were used in
2015 Mansfield and RCM CH4/H2/ 5& 1010- 0.1 the literature to compare with the programs that estimate the MN of
Wooldridge CO/O2/N2 15 1110 natural gas as a function of its composition. It was found that estimates
[171] + TMS**
can deviate considerably with respect to experimental data when syngas
2015 Ouyang et al. ST CH4/H2/ 5-10- 1100- 0.5-
[172] CO/O2/N2 15 1700 1.0-
is considered as fuel [183]. These deviations are mainly caused by the
/Ar 2.0 various and high concentrations of carbon monoxide, and especially
* carbon dioxide (as a knock suppressor) and hydrogen (as a knock
RCM: Rapid compression machine, ST: Shock tube.
** propagator) in syngas. The other reason, for example, for early version
TMS: Trimethylsilanol which is added to syngas as impurity, 10-100 ppm.
of AVL program was that carbon monoxide effects are not included
[177]. Diaz et al. [177] carried out a comprehensive analysis of the
HO2/H2O2 reactions become more profound, thereby increasing the
auto-ignition tendency of binary mixtures of hydrogen and carbon
reactivity of the mixture [173]. Although substantial progress has been
monoxide diluted with carbon dioxide. They proposed a statistically
obtained in our awareness of IDT for H2/CO syngas, there are still few
validated second-order regression model to predict MN. Results revealed
studies to achieve a general description for a representative H2/CO/di-
that increasing carbon dioxide concentration in the mixtures strongly
luents/(possible) impurities syngas under engine-relevant conditions.
increases the MN of the mixtures. A great fraction of H2 and CO intensely
reduces the MN, whereas the CO2 or N2 raise knock resistance and thus
2.3. Methane number (MN) increase the MN [176]. Lechner et al. [184] showed that the classical
methane number or the propane knock index1 fail for the syngas com-
In the IC engine community, to isolate the knock propensity of the positions in practice. Based on a review of published fuel data, 31 gas
fuel from the specific engine design and condition, a practical measure mixtures were developed using a D-optimal statistical design consisting
was required to indicate the ability of a fuel to resist knock [174]. Due to of H2, CO, CH4, CO2, and higher hydrocarbons up to C3. Additionally,
this, octane number is defined, and some octane rating methods like the pure methane, pure propane, pure hydrogen and a biogas mixture
motor octane number (MON) have been established by ASTM for liquid consisting of 60 vol% CH4 and 40 vol% CO2 were examined as reference
fuels based on primary reference fuel blends, in which iso-octane-like fuels. Practically all syngas blends were more prone to knock than
fuel is assigned with ON (or MON) of 100 and n-heptane-like fuel with methane or biogas. Admixtures of higher hydrocarbons were found to
ON (or MON) of 0; which means fuels with high MON resist substantially increase the knock propensity. This indicates the promi-
auto-igniting more than those with low MON. Extending this concept to nent role of higher hydrocarbons with regard to the knock propensity of
gaseous fuels leads to the so-called methane number (MN), taken from fuel gases. Lean equivalence ratios, exhaust gas recirculation, and the
the work done for the Austrian AVL Company through 1964-1969 [175]. addition of water vapor were effective measures to mitigate the risk of
The system rating of MN employs a reference fuel blend of methane and knock.
hydrogen, by which methane and hydrogen are given to MN of 100 and Under the MN determination, hydrogen should have the lowest
zero, respectively. CH4-CO2 mixtures were also considered as reference resistance to knocking, which somehow is not true since some work has
mixtures for MN beyond 100, defining 100 plus the volume of existing presented a research octane number (RON) of around 130 for hydrogen.
CO2 for MON; for example, 20%CO2 + 80%CH4 (by vol.) results in MON Also, hydrogen and methane as reference fuels in the MN rating system
of 120 while 20%H2 + 80%CH4 (by vol.) leads to MON of 80. Therefore, have very different characteristics in flame speeds and ignition delays, in
it is expected that an engine fueled with a syngas fuel having a higher contrast to those in the ON rating system, which have similar flame
CO2 fraction leads to a rise in the upper limit of compression ratio (CR) speeds [185]. It was seen that the MN could be correlated to the adia-
without knock issue. More generally, once the MN number through this batic flame temperature and laminar flame speed, a high MN fuel having
method is determined, regulating both spark timing and CR can be a low adiabatic flame temperature and long flame initiation period
employed to avoid knock, applicable to an engine fueling with either [176]. However, compared to H2, propane (C3H8) has shown more
100% syngas or syngas blended with other gaseous fuels. To determine knock tendency, which may stem from its lower auto-ignition
the MN, some researchers [176–179] have conducted their own pro- temperature.
cedure and compared it to developed MN programs, such as the AVL To summarize, there is still no comprehensive consensus to deter-
Methane program V3.20 [180] and the MWM MN program [181], which mine the knock propensity of gaseous, namely syngas fuels. The
are partly looked at in the following. currently potential errors in the MN method, such as various operating
Generally, the procedure for specifying the MN of tested fuel is
basically a comparative basis in which unknown fuel is indirectly
compared to a map of reference blends at the same CR via a curve;
however, this subject is out of scope for the current review and can be 1
The propane knock index is one of MN alternatives to index knock pro-
found elsewhere [182,183]. As an alternative approach for indirectly pensity of gaseous fuels.
12
Fig. 11. Published measured data of laminar flame speeds for the model syngas mixture of H2/CO: 5/95 with oxidizer (by vol.) at atmospheric (symbols), 5 and 10
atm (lines) (adopted and extended from [194] with permission of Elsevier, other references [165,188,189,192,193,195,196]).
engine conditions and heavy hydrocarbon presence in syngas, lead to For the high-pressure operating condition, Sun et al. [196] were one
keep investigating this matter. of the first groups that conducted the LFS measurement of H2/CO mix-
tures at pressures up to p=40 atm. They substituted nitrogen (N2) with
helium (He) to increase the Lewis number (Le) of the mixture, mini-
2.4. Laminar flame speed (LFS)
mizing flame-front instability3 due to wrinkling and cell formation over
the flame surface. It is also noteworthy that if the ignition energy sup-
Laminar flame speed is a crucial parameter, which results from the
plied by the spark plug of a SI engine is not enough to drive the flame
net effects of the diffusivity, exothermicity, and reactivity of the mixture
front to a minimum radius at which stretch effects are decreased, the
[186], and has a strong dependency on pressure, temperature, mixture
propagating flame of a mixture with Le>1 can extinguish [199]. Refer-
composition, and equivalence ratio [187]. The LFS of syngas-air mix-
ring back to Fig. 11 and now to Fig. 13 (left) for pressures higher than
tures at typical temperatures and pressures (NTP) conditions were
p=10 atm, it is obvious that increasing initial pressure substantially
extensively reported in [188–194], and Fig. 11 shows an example of
reduces LFS, which is attributed to hydrodynamic instabilities due to
published data for LFS of a syngas mixture (H2/CO: 5/95 by vol.) at
decrease in flame thickness as seen in Fig. 13 (right). Some studies are
atmospheric and high-pressure conditions. As can be seen, a relatively
available concerning the LFS of syngas mixtures at higher pressure [162,
decent consistency in measured LFS values can be observed among all
165,194-196, 200,201]; most of them used helium rather than nitrogen.
data, with a maximum variance of 10 cm/s peak-to-peak, while the
High-H2 syngas, the target fuel for IC engines, is more susceptible to
difference was seen to be larger as pressure increases.
flame front instabilities, taken from LFS of pure H2-oxidizer mixtures
Moreover, the discrepancies among measured LFS data are as high as
[19]. Both thermal-diffusive and hydrodynamic instabilities are
40 cm/s at fuel-rich conditions (ϕ > 2) by increasing H2 concentration
involved in amplifying the flame front’s destabilization by affecting the
by just 10%. Fig. 12 shows that the maximum LFS value increases
Lewis number and flame thickness, respectively. So, as compared with
approximately from 65 in Fig. 11 to 130 cm/s by substituting 20% of CO
CO, H2 has a more dominant effect on flame stabilities [202]. To mea-
with H2 in the syngas mixture. This can be attributed to the wide range
sure LFS of syngas with H2/CO: 85:15% (by vol.) at elevated pressures
of data sets used for extrapolation2 (to eliminate stretch effects from
up to p = 10 bar and lean conditions (ϕ = 0.5-0.6), which is relevant to
data), the applied stretch extrapolation model, the chosen flame radius
IC engines with premixed charge, Goswami et al. [203] conducted ex-
range, and the inhibiting effect of carbonyl compounds coming from the
periments using the heat-flux method and then compared the resulting
steel containers utilized to store the CO content of syngas. More infor-
LFSs with those obtained from the available mechanisms [165,204-206]
mation on fundamentals and technical possible sources of observed
by kinetic modeling. They observed significant differences between
discrepancies and uncertainties are described in [187] and [197] for
experiment and modeling, attributed to the uncertainty in key reactions,
each distinct fuel-air mixture, and in [198] for spherically expanding
and suggested that further studies are requisite to modify their rate
syngas flames.
2 3
To correctly determine the LFS, particularly in spherical propagating flame Two common intrinsic flame instabilities have been pointed out within the
at elevated pressures, stretch effects due to flame curvature and flame un- present work, the thermo-diffusivity instability that is characterized by Lewis
steadiness should be subtracted from experimental data to avoid scattering the number and the hydrodynamic (as known as the Darrieus-Landau) instability,
data. Stretch effects can be eliminated by using non- and linear extrapolation which is due to density jump across the flame-front, that is characterized by the
equations. flame thickness and thermal expansion ratio.
13
Fig. 12. Published measured data of laminar flame speeds for the model syngas mixture of H2/CO: 25/75 with oxidizer (by vol.) at atmospheric (adopted from [194]
with permission of Elsevier).
Fig. 13. Effect of pressure increase on laminar flame speeds of syngas diluted in helium (left) and flame front growth of syngas diluted in air with flame radius of 60
mm (right), for H2:CO = 25:75 (by vol.) and at room temperature; p/pi: pressure ratio, σ: thermal expansion ratio, andδf: flame thickness. (Reprinted from [137] with
permission of Elsevier).
constants. Later, Zhang et al. [207] studied the LFS of the same lean
high-H2 syngas of the previous study [203] at elevated pressures but
using the spherical flame method. Fig. 14 depicts that the measured LFS
is seen to diminish monotonically as pressure increases, due to the fact
that SL ~ p(n/2−1) [197] and the overall order of reaction n is normally
less than 2, due to the third body, inhibiting the H + O2 + M → HO2 + M
reaction. As can be seen from Fig. 14, while mechanisms have better
predicted the results of the latter study compared to that of the former
study, the reason is unclear and needs additional study. In addition, the
preferential diffusivity of high-H2 syngas must be noted here due to the
high mass diffusivity of H2 [19], as it plays a vital role in the future of
flame front structure [208]. Interestingly, at elevated pressures and lean
conditions but lower H2 content (35:65 H2:CO mixtures), syngas laminar
flames were determined [209] to be extremely unstable, resulting in
cellular structures in the spreading flame front surface, where SL ~p−0.15
having a fairly small reduction in pressure compared to lean methane
flames where SL ~p−0.50. As a result, it is challenging to measure the LFS
of syngas-air mixtures free of flame instability effects without diluents. It
is almost impossible to obtain accurate LFSs before the onset of flame
instabilities at lean and elevated pressure conditions [210].
Based on the above discussions and Refs. [211,212], it is worth
Fig. 14. Laminar flame speed of 85%H 2–15%CO/12.5%O 2–87.5%He mixture noting that lean high-H2 syngas (H2/CO) flames at elevated pressures
as a function of pressure (Reprinted from [207] with permission of Elsevier). tend to wrinkle and exhibit cellularity even at the early stage of kernel
development. Such unstable cellular flames can be self-accelerated by
14
gradually increasing their frontal surface area, and, thus, this should be The LFSs of H2O diluted syngas-air mixtures have been explored by
noted as a reason for knocking combustion in SI engines. Both intrinsic some researchers at atmospheric conditions using both stagnation [224]
flame instabilities and the acceleration characteristic of the flame front, and spherical [193] flame methods, consistently concluding that H2O
at decreased critical (minimum) flame radius, result in more difficulties addition will non- and monotonically reduce the LFS of CO- and H2-rich
in accurately measuring the LFS. Accordingly, Xie et al. [211] proposed syngas, respectively. For higher pressures up to p=10 bar, Santner et al.
an updated method, based on Bradley et al. [213], to obtain the LFS at [225] found a monotonic drop in mass burning rate with increasing H2O
various flame radii without the acceleration effect. Identification and concentration for an equimolar H2/CO syngas. Besides, they indicated
characterization of self-acceleration of syngas flame are still in their that the inhibitory effect of H2O addition would be increased with
infancy and an ongoing topic for research. increasing pressure. In a research on CH4 blended with H2/CO, Zhou
As H2/CO mixtures do not represent real syngas and always refer to et al. [226] conducted measurements and numerical simulations at
bottled gas or simulated gas, the effect of dilutions with CO2, N2, H2O, at Tin=303 K, ϕ = 0.6–1.5 and p = 0.1–0.5 MPa with a wide range of
least on LFS of H2/CO mixtures must be investigated by scrutinizing the bio-syngas (H2/CO/CH4) compositions. They experimentally deter-
thermal, chemical, and transport impacts of each of them. It has been mined and measured LFS of premixed bio-syngas/air flames under
separately reported in [195,200,202,214,215] that CO2 decreases the various compositions of fuel mixtures, as depicted in Fig. 16, in which
flame temperature, thereby reducing the rate of H2/CO oxidation and αBasis, as a base for comparison, is set to H2/CO/CH4 = 40/40/20 (% by
thus the LFS, which means the chemical effect of CO2 is dominating. Han vol.) with constant mole ratios among species; and αH2-60 equals
et al. [200] measured the LFS of H2/CO (3/1, 1/1, and 1/3 by vol.) with H2/CO/CH4 = 60/26.6/13.3. It was observed that the Li mechanism
CO2 dilution for elevated pressures p=1, 4, and 10 bar and temperatures [204] is well in line with the measured LFS of premixed bio-syngas
T=298, 375, and 450 K using the outward spherical flame propagation flames, particularly under fuel-lean conditions. There is a slight differ-
method, and indicated that increasing CO2 dilution linearly decreases ence between the expected findings and the experimental ones for
the LFS and directly affects the elementary reaction rates. A compre-
hensive study by Vu [202] revealed that CO2- and N2-diluted syngas-air
flames are more sensitive to stretch effects than a He-diluted one
because of the reduced Markstein number with increasing diluent con-
centrations. Also, from Fig. 15, it can be seen that the effective Lewis
number (defined as a combination of fuel and oxidizer Lewis numbers) is
reduced in both cases of CO2 and N2 dilution and will be more prone to
thermo-diffusive instability at lean conditions; according to such trend,
Parthap et al. [214] showed the minimum permissible quantity of the
equivalence ratio to achieve flame stability would be shifted from 0.6 to
0.8 for increasing CO2 concertation from 0 to 30%, respectively, for a
H2/CO: 1/1 (by vol.) mixture. A recent study [216] showed that for a
diluted H2/CO mixture, the flame would be more stabilized by
increasing ϕ when the H2 fraction is more than 50%. Also, it was shown
that the flame instability of high-H2 syngas (as H2/CO mixture) scarcely
varies with increasing CO2 or N2 dilution fraction compared to a
high-CO syngas whose flame becomes more stable with rising dilution.
While decreasing LFS with both CO2 and N2 addition has been
proven [217–222], syngas-air mixtures have higher LFS with N2 dilu-
tion, which has a thermally dominant impact [222]. Therefore,
increasing CO2 dilution has a more substantial inhibiting effect on LFS
due to its direct engagement in the chemical reactions through CO + OH
↔ CO2 + H [214,223], and hence it is necessary to evaluate it with more Fig. 16. Laminar flame speed of H2/CO/CH4/air mixtures at various compo-
caution in predicting syngas containing CO2 by a reaction model. sitions (Reprinted from [226] with permission of Elsevier).
Fig. 15. Effective Lewis number at rich/lean conditions and room temperature with the (Reprinted from [202] with permission of Elsevier).
15
fuel-rich conditions. results of such studies are vastly dependent on the capability of the
It can also be seen in Fig. 16 that the peak value of LFS moves to- developed reaction mechanisms in predicting the LFS under desired
wards fuel-rich conditions with increasing H2 content in the mixture due simulation conditions.
to the very high diffusivity of hydrogen [227], which has a severe, Finally, measured LFS data can establish correlations [245], which is
increasing trend with ϕ. Conversely, the addition of CH4 to the mixture a simpler and more convenient approach than detailed simulations with
leads to shifting the peak value to the lean side with a quite large kinetic mechanisms, for easy use in estimating the LFS of syngas-based
reduction in the quantity of LFS due to the inherent slower LFS of CH4. mixtures during engine combustion simulations [191]. However, those
Also, with an increase in CO fraction in the fuel, the LFS is not much are only applicable to regions far away from stretch and ignition energy
influenced. The thermal and chemical kinetic analysis shows that the effects. An example of the kind of LFS correlation is given below, which
addition of CO has much more impact on the adiabatic flame tempera- is a function of equivalence ratio, temperature, and pressure [194]:
ture but plays a minor role in its chemical kinetic impact as opposed to c d
( ) T p
the addition of H2 [228]. The influence of H2 addition to LFS is mainly Su = Su0 1 + a(ϕ − 1) + b(ϕ − 1)2 (2)
due to its chemical effect at lean and stoichiometric conditions. How- T0 p0
ever, with the addition of around 75% CO, the rise in laminar flamma-
where Su0 is the reference LFS at certain point (ϕ = 1, T0 = 298.15 K, p0
bility is attributed to the high adiabatic flame temperature (thermal
=1 bar), ϕ the equivalence ratio, T the temperature in K and p the
effect) [229].
pressure of the mixture in bar. The fitting coefficients, a, b, c, and d were
The effects of inert gases like N2, CO2, and H2O, and CH4 on the
obtained for different hydrogen concentrations using a nonlinear, least-
laminar flame speeds of syngas, in the form of H2/CO, were briefly
squares method. This correlation is valid only for non-cellular laminar
reviewed here and are extensively discussed in [137,230]. The LFS of
flames within the equivalence ratio range of 0.6 < ϕ < 5. The pressure
syngas in the form of producer gas (PG) have also been addressed in
and temperature ranges are a function of the equivalence ratio and
[231–235], which is the case of 100% syngas fueled IC engines in sta-
hydrogen fraction.
tionary applications. Monteiro et al. [234] investigated the impact of
Semi-empirical correlations are also available to predict the LFS of
various compositions of PG generated through fluidized bed, updraft,
H2/CO/air flames with N2 and CO2 dilution [246]. Although such cor-
and downdraft gasification on the LFS, using the spherical flame
relations are preferred to comprehensive chemical kinetic models in
method. Fig. 17 illustrates that the LFS of two studied PG over a wide
terms of computational cost, they have been restricted to specific con-
range of ϕ have a similar trend to those of syngas with high CH4 content,
ditions, beyond which the predictions are inaccurate. This is because
although the maximum values are at stoichiometric conditions. Tippa
many earlier correlations were based on a small collection of experi-
et al. [236,237], concluded that PG combustion with higher H2, CO, and
mental data, mostly carried out over a narrow range of engine operating
N2 fractions and comparatively lesser amounts of CH4 and CO2 are stable
conditions [73,191,196,238,245,247,248]. In addition, no flame speed
to preferential diffusion effects and predictable for minor changes in
correlation is available that allows the calculation of LFS for blends of
composition.
syngas and gaseous fuels. For gaseous blends, blending has a non-linear
Other researchers have tried to determine the LFS of syngas blended
effect on the LFS, resulting in complicated mathematical correlation
with hydrocarbon fuels [238–240] and with exhaust gas recirculation
equations to capture this behavior [249]. Therefore, the use of reaction
(EGR) [241,242], which is the primary case in syngas fueled engines
kinetics calculations is preferred for the combustion study of
through reforming fuel on-board a vehicle, of which detailed discussions
syngas-fueled engines. With this in mind, data-driven machine learning
can be found in sections 3 and 4. For instance, Kwon and Min [238]
(ML) algorithms [250], as a more recent solution, can be trained to
calculated the LFS of syngas/iso-octane-air mixtures by a 1-D model
overcome this issue based on a large dataset having high accuracy to
under conditions relevant to SI engines. The EGR effect was also
anticipate the LFS of syngas with large fluctuations in their composition
included using different fractions of burnt gas up to 0.5 in volume. The
at various conditions. This approach also allows more researchers to
LFS of the studied mixtures were observed to be proportional to the
participate in developing syngas-based combustion systems without
initial temperature, LHV, the fraction of syngas in the fuel mixture and
requiring the in-depth knowledge that only qualified experts have.
H2 fraction in the syngas, and inversely proportional to the initial
In summary, instabilities induced by high H2-containing syngas at
pressure, deviation of equivalence ratio from the stoichiometric condi-
elevated pressures make LFS measurements with reasonable accuracy
tion, and the fraction of burnt gas. A few studies have also considered the
impossible. In general, with an increase in pressure and concentration of
combustion characteristics of syngas mixtures in premixed SI engines
H2, the LFSs for premixed H2/CO-oxidizer mixtures reduce and increase,
[73,243,244]. However, the extent of accuracy and reliability in the
respectively. Diluting syngas with N2, CO2, or H2O decreases the LFS.
The maximum flame speed occurs at fuel-rich conditions, while the
maximum flame temperature occurs near the stoichiometric condition.
Also, the LFS has shown a strong dependency on pressure, inlet tem-
perature, and equivalence ratio. It was shown that extensive modifica-
tions in kinetic mechanisms should be taken into account. There is a
broad experimental data collection available in the literature for the
combustion of syngas with various concentrations of H2:CO. Nonethe-
less, research on syngas mixed with other gases such as CO2 and H2O is
still limited, and a few experimental data are available for the LFS of
syngas mixed with CH4 at elevated pressures [158,195,239,251]. New
measurements with thorough uncertainty analysis on the LFS of syngas
(H2/CO + diluents) solely and blended with other fuels (like gasoline,
diesel, etc.), which are of great interest in vehicular applications at
elevated pressures and engine-relevant conditions, will help us to un-
derstand syngas combustion behavior properly.
Fig. 17. Laminar flame speeds comparison between two producer gas gener- 2.5. Turbulent flame speed (TFS)
ated by up- and downdraft gasification with H2/CO [189], H2/CO/N2 [217,
231] mixtures (Reprinted form [234] with permission of Elsevier). Like the LFS, the turbulent flame speed (TFS), ST, is of interest and is
16
defined when the surface area of thin laminar flames interacting with intensities used in the measurements. It is believed that these fuel effects
the turbulent flow is increased. Unlike LFS that exclusively depends on are due to the reactant mixture’s stretch sensitivity, which has a strong
mixture properties, TFS is also a function of turbulent flow, enclosed by effect on the positively strained leading points of the turbulent flame
the IC engine geometry. So, measuring TFS is sensitive to the experi- brush for non-zero Markstein length fuels [265].
mental rig used, and it was confirmed [252,253] that it is an The second topic was devoted to extracting ST correlations for ease of
experiment-dependent variable. Turbulent spherical flames propagating use in combustion modeling. Regarding syngas-air mixtures at higher
inside the IC engine consume the unburned mixture when the pressure than atmospheric, some of them are presented in Table 8.
flame-front enters to reaction zones, thereby determining the heat However, such correlations are obviously incapable of capturing all
release rate. From an engine viewpoint, TFS is dependent on sensitivities of the TFS. For example, Venkateswaran et al. [267] used ST
trapped-mass conditions, residual turbulence at trapping, and the mean databases for negative Markstein length fuel blends, i.e., mixtures with
piston speed, while processes such as auto-ignition, combustion, and small Lewis number like H2-rich syngas- or H2-air mixtures, and high-
heat conduction are also affected [254]. Generally, Lipatnikov and lighted that it is challenging to distinguish the effects of time-scale,
Chomiak [253] demonstrated that flame stretch and instability could length-scale, and ReT on these correlations. Thus, future works are
significantly affect TFS, depending on the conditions and mixture needed with novel correlations in a more unified form.
properties. As the final point of this section, the detailed experimental study on
Because of high stretch sensitivity due to the high mass diffusivity of chemistry-turbulence interaction and instabilities of the flame front for
H2 content [255], the turbulent flame properties of model syngas simpler fuels like H2 particularly, in turbulent flame combustion is next
(H2/CO) mixtures are of great importance and interest. A few helpful to impossible at this instant, while it is possible with the aid of the direct
experimental measurements on the H2/CO-air mixtures have been car- numerical simulation (DNS) [268]. An exciting subject addressed with
ried out under elevated pressures, such as those by stabilized DNS is the effect of preferential diffusivity on turbulent spherical syngas
Bunsen-type flames [256–259] and centrally-ignited propagating flames flames on the lean side of stoichiometric. This is due to the well-known
[209,260]. The findings of several flame configurations suggested that behavior of lean hydrogen-air mixtures, which is a non-unity Lewis
the TFS of lean hydrogen flames are around a factor of two faster than number fuel and differs from heavy hydrocarbon fuels. In an example
those of hydrocarbon flames, although their magnitudes vary signifi- study of three-dimensional DNS with detailed chemistry for lean 70:30
cantly for various flame geometries [29]. Regarding measuring the TFS H2/CO syngas at p = 4 bar, it was found [269] that thermo-diffusive
of syngas-air mixtures in high-pressure environments, the two most instabilities have negligible impact on cellularity development of
discussed topics in the literature are: flame structures under high turbulence levels, though the flame accel-
eration and species diffusive flux are still influenced by the preferential
1- The impact of pressure increase, fuel composition, flame-front sen- diffusion. The same authors [270] showed how the interaction between
sitives, and ReT on TFS. equivalence ratio variation and instabilities could affect the heat release
2- Extracting ST correlations from experimental data for a particular rate. Performing two-dimensional DNS for a turbulent expanding 50/50
fuel-air mixture at specific conditions and generalizing them. H2/CO syngas flame at ϕ = 0.6 and atmospheric conditions, Bhide and
Sreedhara [271] found heat release rate enhancement and a rise in flame
Herein, the turbulent Reynolds number is defined as ReT = (u′ l0)v, speed for even a quiescent medium due to high mass diffusivity of
where u′ , l0 and ν are the turbulent intensity or root-mean-square (r.m.s) hydrogen. Under greater turbulence levels, a further increase in HRR
turbulent fluctuating velocity, the integral length scale, and the kine- and a shift in the peak location of HRR towards low-temperature zones
matic viscosity of the reactants, respectively. In the first topic, the in- were also seen. The shifting trend was ascribed to atomic hydrogen
fluence of pressure, two scenarios have been reported to date. The first species, and it disappeared at high pressure and temperature conditions.
states that the TFS increases with increasing pressure due to the The differential diffusion effect (related to the different values of the
improvement of flame instabilities, at constant u′ andl0, regardless of the species Le) was also observed to diminish when raising the CO fraction
fact that ReT also increases due to the inverse proportion of ν to pressure to 70%.
elevation. However, the second proved that TFS decreases, similar to In conclusion, while noting that high-pressure turbulent combustion
LFS in a negative exponential behavior, with increasing pressure as the
ReT is held constant [260,261].
As instances of the first scenario, Liu et al. [209] studied TFS for Table 8
35:65 H2:CO-air mixtures at ϕ = 0.5 and 0.7 over an initial pressure Some correlations based on measured data for the turbulent flame speed (ST) of
range of p = 1–10 bar. They found that ST increases with increasing syngas-air mixtures at elevated pressures.
pressure (ST~p0.15 for ϕ = 0.7 and ST~p0.11 for ϕ = 0.5) at a fixed r.m.s Author(s) Correlation for ST Constants Eq.
turbulent fluctuating velocity (u′ ≈ 1.4 m/s), often attributed to cellular
Daniele ST P0= 1 bar, T0 = (3)
instability promotion due to thinner flame thickness (hydrodynamic et al. = 1K
SL
instability) [253,262]. At higher pressure up to 20 bar, Daniele et al. (2011) u 0.63 l0 −0.37 P
′ 0.63
T −0.63 a = 337.5
[257] mentioned that the ST increase is because of diffusive-thermal [257] B a
SL [ δ L ′ ] P0 T0
instability co-acting with hydrodynamic impact. Additionally, it was Wang et al. ST P u n For H2/CO: 35/ (4)
∝a
(2013) SL P0 SL 65
demonstrated that increasing u′ /SL is much more effective in increasing
[263] B a = 3.8, n ≈ 0.42
ST/SL than increasing pressure, particularly in the weak turbulence Shy et al. ST For lean H2/CO: (5)
= aDab
regime. Besides the pressure influence, the increase of ST/SL with (2015) u′ 35/65 with Le ≈
increasing hydrogen fraction in syngas (H2/CO) mixtures was observed [260] S 0.76: a=0.12,
b=0.5
using OH-PLIF measurements in a work by Wang et al. [263], which
Sun and Zhu ST 0 = (1.325.XH2 + 0.916).u + 0.597.
′
- (6)
resulted in the promotion of the intensity of flame front wrinkles due to (2020) e2.XH2
amplified preferential diffusive-thermal instability. In addition to results [266] S
consistent with prior studies, Venkateswaran et al. [264] displayed
Note(s): the Damköhler number is defined asDa = (l0/u′ )(SL/δL), where δL is the
measured quantities of ST for 90:10 H2:CO that are around three times
laminar flame thickness which is approximated byδL ≈ α/SL, and α is the thermal
larger than for CH4 blends, with the same SL,u′ , and operating condi- diffusivity of unburned mixture.
tions. An important conclusion from the work of the same authors [255] XH2 : Hydrogen fraction in H2/CO syngas.
was that fuel effects on ST are not simply a low turbulence intensity B
: Bunsen-type flame method,
phenomenon – it clearly persists over the entire range of turbulence S
: Spherical-type flame method.
17
cannot be directly inferred from atmospheric turbulent combustion re- pyrolysis/reforming process in a small-scale naturally aspirated (NA)
sults [209], there is no consensus among published findings at higher single-cylinder SI engine with variable compression ratio from 8.47 and
pressures than atmospheric, nor even a general description for turbulent 11.9 and an installed mixer instead of the original carburetor. The gas
premixed syngas flames. There are also very few studies presenting the produced from gasification was hydrogen-rich, while that from the
turbulent combustion of spherical (model) syngas flames compared to two-step pyrolysis was methane-rich. They found that both gases pro-
Bunsen-type flames. As in Daniele et al. [257] study under gas vided similar thermal efficiency compared to CNG. In addition, the
turbine-relevant conditions, there is a need to study turbulent spherical hydrogen-rich gas was observed to show a wider stable engine operation
syngas flames with boundary conditions matched with those in an IC range with the λ up to 2.0, and NOx and HC emissions were quite low for
engine, also considering the dilution effect. it compared to both CNG and the methane-rich gas. Bika [275] inves-
tigated the combustion characteristics and knock limits of a
3. Syngas application in IC engines syngas-fueled SI engine with a variable compression ratio
(CR=6:1-10:1). The fuels investigated were pure H2, H2/CO: 75/25, and
3.1. Syngas in SI engines H2/CO: 50/50 (% by vol.). He reported that increase in the CO fraction
increased the knock limit of syngas and advanced the ignition timing of
Research on syngas application in SI engines has been focused on MBT but did not affect the indicated efficiency up to H2/CO: 50/50. In
improving the fuel consumption (or fuel conversion efficiency) and addition, the maximum heat HRR was observed with pure hydrogen.
engine-out emissions. All employed strategies that result in a more However, the compression ratio increase affected the peak pressure
efficient SI engine4 can be divided into two categories, along with the more than by the CO percentage increase. A maximum ITE of 32% was
associated engine design modifications: reported with H2/CO: 50/50 at ϕ=0.6 and CR=10:1. The general
conclusion was that the extended combustion duration with a higher CO
1- Improve stoichiometric operation with a combination of downsizing, percentage provides more knock resistance at a higher compression
boosting, and exhaust gas recirculation (EGR) strategies. ratio.
2- Move towards lean or ultra-lean operation through the intake charge As mentioned earlier, the power output of SI engines is usually de-
dilution and advanced direct injection strategies. rated during operation with 100% syngas. For instance, Mustafi et al.
[272] studied a syngas fuel named “Powergas” that contained 44%H2,
The lean combustion concept in gasoline SI engines has allowed 52% CO, and 4% N2 produced from the Aqua-fuel process5, to evaluate
higher brake thermal efficiency due to reducing pumping and heat the performance and emissions with Powergas (15.3 MJ/kg LHV) in a
transfer losses, while keeping NOx emissions low because of lower peak variable CR Ricardo E6 single-cylinder SI engine and also compared it
combustion temperatures. However, ultra-lean combustion requires with those with gasoline at lean (ϕ = 0.81) and CNG at stoichiometric
careful consideration in order to operate the engine beyond the lean conditions (ϕ = 1.0). The Powergas indicated no superior performance
limit or dilution limit. While a number of methods, including intake relative to CNG and gasoline as the main fuel in the SI engine. They
design improvement, advanced ignition systems, and EGR stratification, observed more than 20% decline in brake torque output and more than
could address this, reformate containing hydrogen-rich gas can serve as 100% increase in BSFC compared to gasoline and CNG at 1500 rpm and
an extender of lean and flammability limits of the engine. Additionally, CR = 8.1:1, as seen in Fig. 18. Although the engine-out THC and CO
since the choice of fuel directly affects the engine’s lean limit, alterna- emissions with syngas fueling were lower than those with gasoline and
tive fuels with reformate can also be used in SI engines. Also, synergies CNG, both CO2 and NOx emissions were higher. Similar power losses
can emerge from syngas production via on-board fuel reforming and have also been reported with syngas combustion in SI engines [272,274,
thermochemical energy recovery techniques. However, the focus of this 276-280], especially at lower CR and closer to stoichiometric conditions
section is to run SI engines with syngas as the primary fuel, which can and operating away from the MBT timing. Munoz et al. [276] also found
only improve engine’s efficiency by extending its lean limit. around 50% reduction in power output of a carbureted SI engine with
producer gas compared to gasoline, at CR=8.1:1.
3.1.1. 100% Syngas in carbureted and port fuel injection (PFI) SI engines Shah et al. [277] experimentally studied the performance of a
The majority of studies on 100% syngas-fueled SI engine have used single-cylinder SI engine with syngas produced from biomass gasifica-
carburetor or port fuel injection (PFI) technologies in power generation tion. The syngas was composed of 16.2–24.2% CO, 13–19.4% H2,
for two main reasons. First, the SI engine can integrate thermo- and bio- 1.2–6.4% CH4, 9.3–13.8% CO2, and the balance N2 with an LHV of about
chemical processes in which feedstock can be supplied from viable, eco- 5.79 MJ/Nm3. Based on the results obtained, the overall engine effi-
friendly energy resources, e.g., biomass. Second, since the LHV of syngas ciency from the combustion of both syngas and gasoline fuels was found
is lower than that of typical fuels such as gasoline and natural gas, the to be similar at the same maximum power output, albeit higher
engine output power and torque could drop by 20-40% with syngas maximum power output was achieved by gasoline operation. CO was
operation, as reported in several published works [59,272,273]. This observed to be 30–96% lower with syngas than gasoline at each oper-
drawback makes using 100% syngas in SI engines via carburetor or PFI ating point due to the higher carbon content of gasoline fuel. In addition,
technologies unsuitable for vehicular applications. However, it is a good the HC emission of syngas was negligible at less than 40 ppm in all
choice for stationary applications where some limited operating points operating conditions due to the presence of very less HC (1.2–6.4%) in
can be optimized by MBT timing. Also, the engine-out emissions can the syngas used in this study. However, the main reason for HC reduc-
vary by changing the ways of syngas production or adding some other tion could be the increased H2 content in the fuel mixture, which is
species to syngas. A brief review of 100% syngas fueled SI engines for responsible for promoting fast conversion of HCs [273]. It was also re-
stationary and transportation applications is presented below. ported that NOx emissions in syngas operation were lower, about
Syngas composition variability can affect the level of extending the 54–94%, compared to gasoline operation. This was due to the lower
lean limit of engine combustion. Ando et al. [274] evaluated the per- cylinder temperatures as a result of the lower LHV of syngas. The
formance of low-BTU gases produced from gasification and a two-step operating fuel/air equivalence ratio (ϕ) in this study was not directly
declared while appearing that the operation of a designed
gasoline-based engine is stoichiometric for gasoline and is lean for
4
Here we can consider two ways of reduced pumping losses (also referred to
as either throttling or gas exchange losses) at low to part loads and/or
5
decreased friction losses at high speeds while maintaining emissions low Aqua-fuel technology transforms clean by-products like glycerin into
enough. renewable fuels for standard generators.
18
Fig. 18. Brake torque and BSFC versus engine speed at WOT and MBT timing and CR = 8:1 (Reprinted from [272] with permission of Elsevier).
Fig. 19. Brake thermal efficiency contour maps for specified fuels at ϕ = 1, 0.85, and 0.7 and a fixed ignition timing of 15◦ BTDC and CR = 10.7:1 (Reprinted from
[281] with permission of Elsevier).
19
syngas. It is complicated to run a 100% syngas-fueled SI engine close to high methane fraction of the syngas fuel accounts for the increased
stoichiometric conditions (i.e., ϕ > 0.75) because the retardation of levels of HC emissions. The general increasing CO2 trend of syngas cases
ignition timing after the TDC is required due to hydrogen presence in the with respect to gasoline and methane cases seen was due to the CO2
syngas composition (irrespective of unstable combustion possibility due content in syngas. Nevertheless, a fixed injection timing (15◦ BTDC) of
to high hydrogen content in syngas). With this in mind, the following gasoline used as a baseline for all cases in this study may result in a
two studies are presented for fixed and varying ignition timings. deviation from a realistic comparison.
In a study intended for vehicular applications, Arroyo et al. [281] In the case of varying ignition timing to achieve MBT, Ran et al.
conducted an experimental study to compare the operation of a NA SI [282] conducted experiments to differentiate the performance and
engine fueled with two different syngas cases and two typical gasoline emissions of a NA PFI single-cylinder CFR engine operating on E10 (90%
and methane fuels, at ϕ = 1, 0.85, and 0.7 in full and medium-loads over gasoline + 10% ethanol vol.), CNG (95% CH4 vol.), ethanol (95%
a wide range of engine speeds. Two syngas fuels produced from catalytic ethanol + 5% water vol.), and syngas (60% H2 + 40% CO vol.) at an
decomposition of biogas with molar composition of 23% H2 and 23% engine speed of 1200 rpm and CR = 8:1. The ϕ was in the range of
CO, 26% CH4, and 28% CO2, with the LHV=14.095 MJ/kg as syngas #1, 0.74-0.99, 0.63-0.98, 0.76-0.97, and 0.3-0.78 for fuels in the above-
and 40% H2, 39% CO, 11% CH4, 10% CO2, with LHV=16.525 MJ/kg as mentioned order. The high heat release rates and spark timing restricted
syngas #2, were tested. A considerable loss in BTE was observed in the upper side of ϕ for syngas (see Fig. 20 (b)). Among tested fuels,
gaseous fueling cases at stoichiometric conditions; however, a slight syngas indicated the lowest net IMEP due to having the lowest volu-
increase in power occurred at ϕ=0.85, as shown in Fig. 19 (gasoline did metric efficiency and LHV. Generally, the volumetric efficiency of syn-
not achieve stable combustion at ϕ=0.7). As the authors reported, the gas and CNG due to their gaseous nature and lack of evaporative cooling
reason for higher NOx emissions was the increased H2 and the decreased effects is lower than for liquid fuels. Also, the stoichiometric air/fuel
diluent contents, e.g., CO2, in syngas composition. The established ratio of syngas is lower than CNG; therefore, extra mass fuel is injected
comparison with syngas #1, syngas #2, and methane showed that the for syngas fueling, resulting in more degradation of volumetric
Fig. 20. Volumetric efficiency (a), gross heat release rate (b), net indicated efficiency (c), and combustion efficiency(d) for specified fuels at MBT timings and CR =
8:1 (Reprinted from [282] with permission of Elsevier).
20
efficiency, as can be seen in Fig. 20 (a). Moreover, the lowest CA0-10%

(flame development period) and CA10-90% (rapid burn period) belonged
to syngas because of the high laminar flame speed and high diffusivity of
the H2 content in syngas. However, the smaller quenching distance of
H2, which implies the flame could travel nearer to the cylinder wall,
leading to higher heat loss, reducing the net indicated efficiency (Fig. 20
(c)) at closer stoichiometric conditions.
Another work [283] tried to draw a comparison between syngas, and
biogas, biogas-syngas blend fueling cases in a NA SI engine at various
MBT timings and at a fixed engine speed of 1500 rpm, using experiments
and numerical simulations. The syngas used had a molar composition of
17% H2, 15% CO, 4% CH4, 15% CO2, and the balance N2 (HHV= 5.65
MJ/nm3) and biogas had 65% CH4, 35% CO2 (HHV=25.83 MJ/kg). The
corresponding ITE, NOx emissions, higher temperature, and maximum
pressure derivative for both syngas and biogas fueling are shown in
Fig. 21. As seen, all observed variables have a relatively distinct sensi-
tivity to where ignition timing occurs, revealing that the previous work Fig. 22. In-cylinder pressure versus crank angle at different CRs (Reprinted
lacks a reasonable basis for comparison. Although the syngas fueled SI from [284] with permission of Elsevier).
engine due to having high H2, low CO2, and presence of N2 contents
produces more NOx and has the highest temperature regions over the
entire ignition timing sweep than the biogas one, lower NOx formation
belongs to syngas, 3 g/kWh, rather than biogas, 7 g/kWh, at maximum
MBT operation, where the maximum ITE happens (26◦ and 50◦ BTDC for
syngas and biogas, respectively). Additionally, syngas fueling not only
indicates 1.5% more ITE compared to biogas at maximum MBT, but also
tends to maintain high ITE over the whole range of ignition timing in
contrast to biogas, indicating the superior combustion characteristics of
H2 relative to CH4.
3.1.2. 100% syngas in CI engines retrofitted to SI mode

To mitigate the power loss with syngas use in SI engines, Sridhar
et al. [284,285] experimentally studied syngas combustion in a
multi-cylinder SI engine modified from a diesel engine at varying CRs
(11:5:1 to 17:1). Fig. 22 shows in-cylinder pressure traces at different
CRs, which did not show any trace of knock for all the studied loads.
Performance of the engine at higher CRs was smooth and proved that Fig. 23. Power generation efficiency of the syngas, diesel and natural gas en-
syngas use in the SI mode at higher CR is technically feasible. Homdoung gines at different loads (Reprinted from [289] with permission of Elsevier).
et al. [286,287] also reported that for CRs as high as 17:1, the engine
fueled with producer gas (PG, see section 1) could operate smoothly density recorded in the diesel engine (1.5-12%) was much higher than
without knock risk. This was attributed to the faster burn rate due to the that in the PG case (0-2%). Zero-dimensional (0-D) modeling was also
presence of H2 in the PG. Although the maximum BTE for the PG engine employed to compare the brake power results obtained under a
was lower than the original diesel engine by about 3.5%, and the smoke wide-open throttle (WOT) condition, and a good agreement was noticed
Fig. 21. The effects of ignition timing on other variables for syngas (left) and biogas (right) fueling, at CR= 12.9 and 1500 rpm (Reprinted from [283] with
permission of Elsevier).
21
with experimental results [288]. Following the work of Hagos [292], Fiore et al. [297] numerically
Raman and Ram [289] studied the efficiency of a syngas-fueled en- investigated the effect of piston shapes and injection specifications on
gine at different loads. Fig. 23 illustrates the comparison of syngas, the performance and emissions of a syngas-fueled DI SI engine. The
diesel, and natural gas engines in power generation efficiency at parametric study included three different piston bowl profiles and
different load conditions. The maximum power generation efficiency of half-angles of injection of 30◦ , 45◦ , 52.5◦ and 60◦ . The SOI timing was
the diesel, natural gas, and syngas engine is 28%, 24%, and 21%, also varied from 90◦ BTDC to 130◦ BTDC. They found that the Omega
respectively. Again, a reduction in torque and power de-rating for the shaped profile can provide best performance, despite of maximum NOx
syngas engine was observed, but at a lower amount. The syngas engine emissions compared to the high clearance and low clearance combustion
operates at 21% power generation efficiency at the full load condition, cups. The results were very sensitive to the injection angle and SOI,
which is about 20% and 10% less than that of the diesel and natural gas which affected the fuel distribution.
engines. Particulate matter (PM) emission:
It can be mentioned that increasing the engine compression ratio Compared to PFI engines, particulate matter (PM) emissions lead to a
through use of CI engines retrofitted to the SI mode is an efficient way to known big challenge in DISI engines, especially for ultrafine particles,
increase brake power. However, an average of 16% power loss can be which harmfully affect human health [298]. The sources of PM forma-
observed in the gas mode compared to diesel operation at a comparable tion in DISI stem from mixture inhomogeneity because of a lack of time
CR. Recently, Park et al. [290] demonstrated that high compression for fuel atomization and spray impingement, which depends on many
ratios and intake boosting could improve syngas (30% H2, 25% CO, 45% factors, including operating engine conditions, air-fuel equivalence
CO2) fueled SI engine power and efficiency. Engine operation with a ratio, injection timing, spark timing, and fuel-specific properties [299].
high compression ratio of 17.1:1 was possible in the lean combustion While fueling a DISI engine with hydrocarbon-free gaseous fuels is
mode owing to the low combustion temperature. The gross ITE in the seemingly expected to lead to lower particle emission, Thawko et al.
lean combustion mode was 18.4% higher than that in the stoichiometric [300] surprisingly showed a 170% increase in total particle concentra-
combustion mode, mainly because of a significant reduction in the heat tion of reformate (H2/CO2:75/25 %by mole) fueling compared to gas-
transfer loss. However, the gross indicated power in the lean combustion oline at the same and fixed engine speed and load. The most likely
mode was 25.6% lower than that in the stoichiometric combustion reason for the high measured particle is attributed to the interaction
mode. In summary, stoichiometric operation using syngas was appro- between lubricating oil and high-pressure reformate jet, as the authors
priate for a high-power engine generator, whereas lean operation using claimed. Due to its lower density, the required time for syngas fuel in-
syngas is more suitable for a high-efficiency engine generator. jection at high pressure is much longer than that for the liquid fuel.
Besides, the small flame quenching distance of hydrogen content in
3.1.3. 100% Syngas in direct injection (DI) SI engines syngas may result in the syngas flame traveling closer to the lubricated
Hagos et al. [291–293] were the first researchers to utilize syngas cylinder walls. Thus, the chance of lubricant participating in the com-
(50%H2/50%CO) as the primary fuel in a direct-injection spark-ignition bustion rises. In terms of particle size distribution, it was indicated that
(DISI) engine, and then compared its combustion, performance, and the nucleation mode (< 50 nm) is primarily responsible for the
emissions characteristics to those with CNG operation (as the baseline). discrepancy in total particle number concentration between reformate
It should be noted that besides overcoming the issues related to the and gasoline fueled DISI engines, as can be seen in Fig. 25 (right).
volumetric efficiency of gaseous fuels, the backfire phenomenon is also In contrast, Bogarra et al. [301] indicated that the addition of
being addressed by the DI ignition system. As well-known, imple- reformate (H2/CO/CO2/H2O: 5/2.7/11/9.0 %vol.) plus EGR (called
mentation of gaseous injectors in comparison with liquid ones has its R-EGR, see Section 4.1) as supplementary fuel to gasoline in a gasoline
own difficulty in terms of injector geometry and engine-related char- direct injection (GDI) engine reduced the engine PM emissions even
acteristics such as the interaction between piston design6 and the mixing more than the addition of EGR only, as shown in Fig. 26. It was found
process with late injection, injector width pulse, etc. An improvement to that the reformate combustion can significantly decrease the PM emis-
the previous work [272], as the authors stated, was a slight decrease in sions, however, the reduction was dependent on the PM nature. Refor-
brake torque from ranging 20-30% to 14-19% at the same speed of 1500 mate combustion was also found to remove soot cores more efficiently
rpm, due to eliminating issues attributed to volumetric efficiency pen- than the volatile PM. Moreover, fuel reforming technology did not show
alty of delivering gaseous fuels at the intake. Despite the improvement, significant detrimental influence on the TWC operation for the studied
Hagos et al. [294] came across a limitation on the injection duration at conditions.
late injection timings and higher BSFC with syngas fueling than CNG
fueling. Thus, the engine operation was limited to near stoichiometric 3.1.4. Syngas addition in SI engines
conditions, leading to a restriction in maximum brake power and a As mentioned earlier, a possible use of syngas in SI engines is as an
request for a bigger fuel tank. These all were attributed to the low LHV of additive to enrich or improve the performance of another fuel such as
syngas, which is lower than a fifth of LHV of CNG (see Table 5). To methane or a methane-rich biofuel (natural gas or landfill gas); and
overcome this, in a series of works Hagos et al. [294–296] investigated gasoline. Since syngas has a lower heating value than these two fuels, the
enriching the syngas with methane (a higher LHV fuel) in a DISI engine. improvement obtained by adding it is not really on the engine power but
The new fuel was named methane enriched syngas (MES), consisting of rather on the efficiency, combustion, and pollutant emissions. Con-
H2/CO/CH4: 40/40/20 (% by vol.). It was found that the MES shortens cerning knock studies, Gerty and Heywood [302] experimentally
the injection duration of syngas and extends the air/fuel stoichiometric investigated the performance of a knock-limited SI engine by consid-
ratio (λ) compared to CNG and syngas at the same engine speed. ering the effects of relative air/fuel ratio, inlet boost pressure,
Furthermore, the MES improved the maximum BTE and the BSFC by compression ratio, and reformate added to the tested fuel-air mixtures.
30.2% and 21.3%, with a little effect on the CO, NOx, and THC emissions The tested fuels were: Unleaded test gasoline (UTG96, representing as
[294] (see Fig. 24). gasoline or indolene fuel with octane number of 96), primary reference
fuel (PRF, a mixture of iso-octane and n-heptane), and toluene reference
fuel (TRF, a mixture of toluene and n-heptane). A bottled mixture, as
6 reformate, of H2/CO/N2: 25/26/49 (% by vol.) was utilized to simulate
Even by eliminating the physical processes associated with the spray of
liquid fuel (droplet breakup to evaporation), a sufficient time for the charge the output of a POX fuel reformer (named Plasmatron, see subsection
mixing process at late injection is still needed because the shape of the piston 4.1.1). The mixture was considered the replacement for a portion of the
head affects the charge distribution, which may result in an uneven air/fuel main fuel (gasoline), which then mixes with the gasoline in the intake
ratio in the chamber. manifold. Fig. 27 (a) shows the combustion retard, which was defined as
22
Fig. 24. Comparison of injection duration at maximum IMEP (left) and maximum brake torque vs. engine speed (right) for syngas, MES, and CNG at CR=14:1
(Reprinted from [294] with permission of Elsevier).
Fig. 25. Total PN vs. time (left) and total PN vs. particle diameter (right) comparisons between gasoline and methanol steam reforming (MSR) products as the fuel of
a DISI engine at IMEP of 5 bar and engine speed= 2800 rpm (Reprinted from [300] with permission of Elsevier).
was most effective when added to PRFs that are entirely composed of
alkane hydrocarbons. The combustion retard needed to avoid knocking
decreased by about 2◦ CA per 3% reformate fraction. When applied to
TRF fuels composed of more than 20% n-heptane with octane numbers
of 95 or lower, reformate addition was only marginally less effective
than for PRF fuels. As the content of alkanes declined and the aromatic
content increased, reformate addition seemed to be less successful for
TRFs to the point where there was no value in reformate addition to pure
toluene.
Using the same reformate and at low CR (9.8:1), a direct comparison
between pure hydrogen and reformate additions showed that both of
them could extend both peak efficiency and the lean limit of indolene
Fig. 26. Total PN vs. particle size comparison among no EGR, EGR, and REGR,
as well as with after and before using the three way catalyst (TWC) operation combustion linearly [303,304]. On the contrary, at high CR (13.4:1), the
(Reprinted from [301] with permission of Elsevier). reformate addition was no longer had a linear relationship, while pure
hydrogen still indicated an obvious, proportional/linear extension of
the combustion phasing disparity at the current state and the optimal peak efficiency [305], as can be seen in Fig. 28 (left). The main reason
one (i.e., MBT spark timing), as a function of reformate fraction. A for this behavior was due to the proportional extension of the location of
reduction in combustion retard can be seen with increasing reformate the Δθ10 − 90%=30◦ CA. However, no distinguishable curves were seen
fraction. The reduction with the addition of reformate for each fuel was with increase of reformate, as shown in Fig. 28 (right).
normalized, as shown in Fig. 27(b). As can be seen, reformate addition Smith and Bartley [306] evaluated stoichiometric operation of nat-
ural gas + EGR and natural gas with syngas + EGR in a single-cylinder
23
Fig. 27. a) Combustion retard to just avoid knock with increased reformate fraction for fuels with PRFs and TRFs; CR = 11.6:1, λ = 1.0, 1500 rpm, MAP for 40%
boost (NIMEPMBT = 14.7 bar, 1.4 times un-boosted NIMEPMBT). b) Decrease of combustion retard with increased reformed fuel fraction at 1500 rpm, PRFs at λ =
1.0 (40% boost) and λ = 1.3, TRFs at λ = 1.0 (40% boost) and λ=1.3, and UTGs at λ = 1.3. Data is from all three compression ratios [302].
Fig. 28. Effect of air dilution on net indicated engine efficiency and CoV NIMEP (left) and on 10-90% burn duration (right) for various amounts of reformate
(produced by Plasmatron) and pure hydrogen; at MBT, 1500 RPM, CR = 13.4:1, and NIMEP = 3.5 bar [305].
Caterpillar engine converted to SI operation. A 44% extension of EGR 3.2. Syngas in HCCI engines
tolerance on a mass basis and a 77% reduction in raw NOx emissions for
using syngas with natural gas were detected compared to the natural gas Research on syngas utilization in advanced combustion strategies
case. Cha et al. [307] experimentally investigated the operating char- has significantly increased due to its clean-burning properties and
acteristic of an SI engine for power generation upon mixing CH4 and possible production from several paths. The homogeneous charge
simulated syngas (H2 and CO) in a diesel engine that was modified to run compression ignition (HCCI) engine is a low-temperature combustion
in the SI mode with a CR=13:1. Fig. 29 shows the fuel conversion effi- (LTC) concept which combines the best features of SI and CI engines.
ciency versus NOx emissions for various fuel mixtures, where NOx The HCCI engine has shown the potential for higher efficiency and ultra-
emissions were elevated as fuel conversion efficiency increased. How- low NOx and soot emissions, however it has some drawbacks in terms of
ever, this increase was insignificant for syngas compared to H2 or CO limited operating range, combustion control and excessive unburned
when added to CH4-air mixtures at a ratio of 5% of the overall LHV. products [319,320]. In this part, syngas-fueled HCCI combustion is
Since on-board fuel reforming methods can produce syngas as the pri- analyzed in detail.
mary fuel, this will be discussed in section 4.
A summary of the main research works published on syngas utili- 3.2.1. 100% syngas fueling
zation in SI engine is presented in Table 9. Future study of 100% syngas- According to the finding of previously published papers, HCCI
fueled SI engines can be followed in designing an appropriate injection combustion with syngas has not shown superior performance over other
system for gaseous fuel like syngas in DISI engines and applying fuels. It faces a similar barrier, such as the limited operating window and
advanced spark systems like multiple injections. a need for regulating the intake temperature to control the combustion
event. For instance, Stenlåås et al. [321] focused on HCCI combustion of
100% syngas, which was reformed methanol gas (RMG), a mixture of
hydrogen and carbon monoxide in the volume ratio 2:1, and then
24
combustion, Maurya et al. [327] used a stochastic reactor model (SRM)

with a detailed chemical kinetic mechanism (32 species and 173 re-
actions) to numerically investigate syngas fueled HCCI combustion
validated against previous experiments [323]. Within the SRM, a
probability density function (PDF) was used to define the local (vari-
able) and global (constant) physical parameters. For various engine
operating conditions, simulations were performed by changing intake
temperature, engine load, and speed at a constant compression ratio of
19:1, and the obtained operating window is shown in Fig. 30. The in-
clined hatched line represents operating conditions with a combustion
efficiency of less than 80%. The area with PPRR > 5 MPa/ms is repre-
sented by the crossed hatched region, and the contour lines in the plot
represent IMEP. Operating conditions with combustion efficiency higher
than 80% and PPRR < 5 MPa/ms are considered as the operating area of
the HCCI engine. It can be seen that increasing the intake temperature
(or Tivc, specified at intake valve closing) can help the operating range in
the lean-side of stoichiometry, but at the expense of lowered IMEP due to
decreased volumetric efficiency. At higher engine speed, suitable oper-
ating range has adversely shifted to a more bounded region that is highly
sensitive to intake temperature and excess air (λ or air-fuel ratio). In this
Fig. 29. NOx emissions versus fuel conversion efficiency for various compo- case, the chance of incomplete or vigorous combustion is severe. The
nents added to CH4-air mixtures at different spark timings (Reprinted and maximum combustion efficiency of 98% is reported in the study.
modified from [307] with permission of Elsevier).
While the abovementioned works have dealt with 100% syngas in
the form of simulated reformate gas (with only H2 and CO), there are a
compared it to pure hydrogen fueling [322] in SI operating mode. They number of published works with producer gas generated from a biomass
found that a higher temperature for RMG fueling than pure hydrogen gasification process. Przybyla et al. [328] compared SI and HCCI fueled
was required to achieve the same combustion duration with HCCI mode. with low calorific producer gas under the same load; they reported that
Both hydrogen- and RMG-fueled HCCI did not provide sufficient exhaust the syngas-fueled HCCI engine had extremely high CO and a lower
energy for the on-board reforming process by waste heat recovery. In indicated efficiency than SI, equaling a more than 10% loss in the base
summary, the operating mode based on 100% reformate was more fuel’s chemical energy. It is worth mentioned that the use of high intake
favorable for SI mode operation. temperature to control the combustion in syngas-fueled HCCI engines
Several years later, Bika et al. [323] measured the HCCI performance has been reported in previous works. Bhaduri et al. [329] employed high
on 100% syngas fueling at two equivalence ratios, 0.26 and 0.30, and an temperature strategy as a positive advantage over a syngas-fueled SI
engine speed of 1800 rpm, with a diesel engine converted to HCCI mode engine to introduce a novel concept of unprocessed syngas-fueled HCCI,
by adding PFI injection and intake air heating systems. Three compo- which has tar contents and moisture. Because the intake temperature
sitions were considered: H2=100%, H2/CO=3/1, and H2/CO=1/1 by was set to around T=250◦ C, which is higher than the dew point of the
volume. By increasing intake temperature, it was seen that the tars, issues related to tar condensation were eradicated. According to the
in-cylinder pressure, temperature, and heat release started to increase, authors, moisture in syngas tends to delay combustion, while tars’ effect
and finally led to advancing the combustion phasing, which had com- was negligible. Nevertheless, such an intake temperature is not accept-
bustion efficiency in the range of ηc = 83%-88%. Also, by altering the able for vehicular applications, but the readers can refer to their later
equivalence ratio towards leaner regions and increasing CO in syngas, works [330–332] for in-depth information on this concept. We believe
the intake temperature was increased to maintain the best IMEP that a brief discussion of syngas impurities can also be useful here.
conditions.
With the aim of improving HCCI combustion and reduction of 3.2.2. Syngas addition
pollution emissions based on the syngas-air mixture, a few modeling Reformate as ignition controller:
studies [324,325] have been carried out to explore the effect of various The operating range of kinetically-controlled HCCI combustion is
species addition. Starik et al. [325] assessed the potential of the addition narrow due to lack of any direct control on ignition timing; this can lead
of a highly reactive species such as ozone or singlet delta oxygen (SDO) to extremely high pressure rise rates corresponding to nearly constant
into an HCCI fueled with syngas (50%H2/50%CO by volume). They volume heat release of a homogeneous air-fuel mixture in the combus-
noticed that the presence of 0.38% of O3 mole fraction (=1% of SDO) in tion chamber at high loads (for low octane fuels), and poor combustion
total oxygen could diminish the dependency on high intake temperature and misfiring at low loads (for high octane fuels). To mitigate this,
by accelerating and intensifying the ignition and combustion event, various techniques such as intake charge preheating [333], EGR [334,
especially under the low load regime (φ = 0.2); this resulted in a 7-14% 335], boosting [336,337], using two fuels with different reactivity (i.e.,
improvement of engine power output and up to a 15% reduction of both one has a high octane and the other has a high cetane number) [338,
CO and NO emissions due to lack of excessO2. 339] have been proposed which proved to have a profound effect on
For better understanding of auto-ignition and combustion develop- control of ignition event. For more details about these techniques, one
ment in HCCI based on 100% syngas, Yamasaki and Kaneko [326] can refer to [340,341].
conducted an experimental and modeling study using mock syngas with Syngas or reformate plays the role of the ignition-timing controller
varying compositions, ranging from 20% to 40% for H2, 7-27% for CO, by either retardation (if it is the primary fuel with high octane number,
20-40% for N2, 10-30% for CO2, and a 3% CH4 by volume. An e.g., syngas/diesel fueling case) or advancement (if used as the sec-
auto-ignition temperature of T=1100 K, almost similar to hydrocarbon ondary fuel with higher combustibility characteristics, e.g., natural gas/
fuels, was observed for syngas, governed by in-cylinder gas temperature syngas) of the main combustion, which leads to widening the load
and not by the fuel composition tested. It was also found that H2 has a operating range of the HCCI engine. In the former case, syngas has two
longer conversion time than CO during the combustion process, and widespread effects on fuels, showing two stages of heat release such as
combustion duration can be roughly determined by the volumetric ratio diesel, n-heptane, and dimethyl ether (DME). These result in prolonged
of H2/ CO2. For detailed analysis of the syngas combustion in HCCI ignition delay times, and widened combustion duration (in other words,
25
Table 9
Summary of syngas utilization studies in SI engines.
Author(s) Objective(s) Syngas Experimental info Simulation Finding(s)
composition
(% vol.)
100% Syngas in carbureted, port fuel injection (PFI), direct injection (DI) SI engines
Mustafi et al. A comparison study H2/CO/N2: 44/52/ A 4-stroke, 1-cyl, SI Engine simulation Compared to gasoline and CNG:
[272] between ‘Powergas’, 4.0 engine, CRs= 8:1-11:1, program ISIS 23-30% ↓ in brake torque, ↑NOx, ↑CO2 while HC
(2006) S gasoline, and natural gas LHV = 15.3 MJ/kg rpm= 1000-2000, WOT, and CO were negligible.
fueled SI engine ITs = MBT timings
Conte and Effect of RG addition on H2/CO/N2: 21/24/ A 4-stroke, 2-cyl, PFI SI — When the stability limit was met, combustion took
Boulouchos flame speed and flame 55 engine, CR= 8.7:1, place in a very similar way for different
[308] front propagation in LHV = 5.3 MJ/kg rpm= 2000, BMEP= 2 gasoline–RG blends.
(2006) premixed, homogeneous bar, ITs = MBT timings
charge gasoline engines
Papagiannakis Comparison between H2/CO/CH4/CO2/ A 4-stroke, 20-cyl, TC, SI Two-zone ↑syngas fuel concentration with ↑engine load:
et al. [309] conventional natural gas N2: 19/29/6/8/38 engine, CR= 11:1, 1500 thermodynamic ↑peak cylinder pressure and an improvement of
(2007) (NG) and syngas fueled SI LHV = 7.45 MJ/kg rpm model the engine efficiency. Compared to NG: ↑CO, ↑NOx
engine emissions
Rakopoulos and Development of a zero- // // A zero-dimensional, More reliable NO emissions prediction with
Michos [310] dimensional, multi-zone multi-zone, multiple burned zones than single burned zone.
(2008) combustion model for thermodynamic Calculated NO were kept low at all engine loads (<
predicting the performance combustion 200 ppm) with lean combustion of syngas.
and NO emissions of a SI model
engine fueled with syngas
Shah et al. [277] Evaluation of syngas fueled H2/CO/CH4/CO2/ A 4-stroke, 1-cyl, NA, SI — Lower power output (by about 1000 W), a
(2010) S SI engine-alternator and N2: engine, capacity= 5.5 significant ↑ in CO2 (33-167%) but lower NOx (54-
compared it with gasoline 13-19.4/16.2-24.2/ kW 84%) and CO (30-96%) compared to gasoline
operation 1.2-6.4/9.3-13.8/ operation
balance
LHV = 5.79 MJ/kg
Arroyo et al. Efficiency and emissions of Syn1: H2/CO/CH4/ A 4-stroke, 2-cyl, NA, SI — Both syngas cases have ↑ CO2 trend at all tested ϕ,
[281] a SI engine fueled with CO2: 23/23/26/28 engine, CR= 10.7:1, and a ↓ BTE among others.
(2013) V syngas, CH4, and gasoline Syn2: H2/CO/CH4/ rpm= 2000-4500, Syn2 and CH4 had the highest NOx and lowest CO,
CO2: 40/39/11/10 loads= 50% and 100%, respectively.
ϕ= 0.7-0.85-1, IT=
15◦ BTDC
Arroyo et al. Combustion behavior of SI Syn1: // // — In all speeds and ϕ: CoV of IMEP:
[311] engine with syngas, CH4, Syn2: // biogas>CH4>syn1>gasoline>syn2.
(2014) V biogas, and gasoline Biogas: CH4/CO2: Maximum HHR and cylinder Pressure, highest
60/40 NOx, and lowest HC belonged to syn2.
Arroyo et al. Effects of IT and // //, various ITs= 11◦ -59◦ — ↑H2, ↑CO, and ↓CO2 contents: ↑advancement of IT
[312] supercharging BTDC, intake boost= 1 retardation. Right ϕ and IT result in higher
(2015) V to 1.3 bar efficiencies than the gasoline. ↑boost pressure:
↓CO2, ↓CO, ↑NOx, ~HC.
Hagos et al. First attempt at study on a H2/CO: 50/50 A 4-stroke, 1-cyl, DISI — Compared to CNG operation: ↓brake torque (14.2-
[292] direct-injected (DI) syngas LHV = 17.54 MJ/kg engine, CR= 14:1, 19.6%) and ↓BTE (16.7%), and ↑BSFC. At higher
(2014) fueled SI engine and rpm=1500-2400, ϕ= loads: ↑NOx and ~CO. THC was negligible.
compare it with CNG 0.67-1, a fixed SOI of
(baseline) 180◦ BTDC
Hagos et al. Methane-enriched syngas Syngas: // // — At all tested BMEP: BTE at 1500 rpm:
[294] (MES) in DISI engine MES: H2/CO/CH4: CNG>MES>syngas. BTE at 2100 rpm:
(2015) 40/40/20 LHV= MES>CNG>syngas. BSFC at both rpms:
24.4 MJ/kg syngas>MES>CNG. Little effect on CO, THC and
NOx.
Hagos et al. Study of effect of injection H2/CO: 50/50 //, SOIs= 120◦ and 180◦ — SOI= 120◦ compared to 180◦ BTDC: Better BTE,
[293] timing on a syngas fueled BTDC power, and BMEP at all engine speeds. With
(2016) DISI engine ↑speed, BTE started to ↓. THC, NOx, and CO all
worsened.
Hagos et al. Effect of SOI timing on a H2/CO: 50/50 // — Based on improved performance and emissions:
[295] DISI engine with 20% CH4 CH4 augmentation: the engine well-performed with SOI=180 but with
(2017) augmented syngas 20% (volume/ SOI=120 just at lower speeds.
volume)
Hagos et al. Engine speed and A/F ratio // A 4-stroke, 1-cyl, DISI — With ↑λ: ↑CoV of IMEP in all speeds, ↑flame
[296] (λ) effects on the CH4 engine, CR= 14:1, development and ↑flame propagation durations.
(2019) augmented H2-rich syngas rpm=1500-2400, ϕ= All the durations were more sensitive to λ at lowest
combustion in DISI engine 0.3-0.8, SOI= 180◦ speed as compared to the other speeds.
BTDC, WOT
Kan et al. [283] Effects of IT, H2 and CH4 H2/CO/CH4/CO2/ A 4-stroke, 4-cyl, SI KIVA4-CHEMKIN Under WOT and MBT operation:
(2018) contents on SI engine O2/N2: 17/15/4/ engine, CR= 12.9:1, ITE (%): syngas (39) > biogas (37.5), NOx (g/
fueled with syngas and 15/0.14/48.86 rpm= 1500, ϕ=0.8, IT= kWh): syngas (3.3) < biogas (7.2)
biogas Biogas: CH4/CO2: variable At advanced ITs: ↑H2 (11-20% by vol.): was less
65/35 sensitive to NOx change, than ↑CH4 (55-88% by
vol.).
(continued on next page)
26
Table 9 (continued )
Author(s) Objective(s) Syngas Experimental info Simulation Finding(s)
composition
(% vol.)
Ran et al. [313] The potential of lean SI H2/CO: 60/40 A 4-stroke, 1-cyl, SI With ↑CR: ↑engine efficiency and improved lean
(2020) combustion with E10, LHV = 20.4 MJ/kg engine, CRs= 9:1-11:1, burn capability for all tested fuels. The leanest
CNG, ethanol, and syngas. rpm= 1200, intake misfire limit (0.4<ϕ<0.7) was achieved with
pressure= 0.75 bar, ϕ= syngas with low NOx, low CO, and no THC.
0.3-1, various ITs
100% syngas in CI engines retrofitted to SI mode
Sridhar et al. Use of producer gas (PG) in H2/CO/CH4/CO2/ A 4-stroke, 3-cyl, NA, CI — Smooth combustion at CR=17:1. To achieve MBT,
[284] engines at high CRs H2O/N2: 19/19/2/ engine, CRs= 11.5:1- a retardation in spark timings is needed. Compared
(2001) 12/2/46 17:1, 1500 rpm, various to diesel mode, at equal CR, the maximum power
LHV = 4.5 MJ/kg ITs loss was 16% in SI mode with PG.
Sridhar et al. Develop a PG engine // A 4-stroke, 3-cyl, NA, CI — 20 – 30% loss in power compared to diesel
[131] integrated with a biomass engine, CRs= 11.5:1- (baseline) case
(2005) gasifier, with a new 17:1 At CR =17:1 and maximum MBT: both NOx and
designed gas carburetor. A 4-stroke, 12-cyl, TC, CI CO were minimum.
engine, CR= 12:1
A 4-stroke, 6-cyl, NA, SI
engine, CR= 10:1
Gamino and Simulation of syngas H2/CO/CH4/ CO2/ — 2-D thermodynamic A predictive model with capability to use in
Aguillon [314] combustion with a multi- N2: 6/25/5/11/53 code and commercial parametric studies of syngas combustion.
(2010) spark ignition system CFD package,
PHOENICS
Ulfvik et al. Engine performance, H2/CO/CH4/CO2/ A 4-stroke, 1-cyl, CR= — Compared to NG: ↓NOx, ↓THC, and ↓maximum
[315] efficiency and emissions of N2: 30/20/1/15/34 12.6:1, rpm= 1050, IMEP, ↑CO, wider the lambda range, with PG
(2011) SI engine with PG and ITs= variable under comparable conditions.
natural gas
Raman and Ram Comparison of PG engine H2/CO/CH4/CO2/ A 4-stroke, 6-cyl, NA CI — At all loads:
[289] with diesel and natural gas H2O/N2: 23/21/ engine, CR= 12:1, 1500 Overall efficiency: Diesel>NG >PG.
(2013) (NG) engines at various 0.9/9.0/2.0/46.1 rpm, a fixed IT of 28◦ Specific fuel consumption: PG>NG>PG.
loads BTDC At full load the discrepancies were decreased.
Shivapuji and The PG impact on overall H2/CO/CH4/CO2/ A 4-stroke, 6-cyl CI zero-dimensional TC imbalance led to the highest power de-rating.
Dasappa [316] engine performance with H2O/N2: 19/19/2/ engine, CR= 16.5 in CI model The selection and optimization of TCs were
(2014) turbocharger (TC) 12/2/46 mode and 10.5 in SI described as the technique for recovering the loss
mode, 1500 rpm of non-thermodynamic power.
Shivapuji and Analysis of thermo- // – Effect of H2 A 4-stroke, 2-cyl, CHEMKIN ↑H2 part from 12.8% to 19.4% by vol.: ↑BTE
Dasappa [317] physical properties and ratio SI research engine, CR= ↑H2 part up to 37.2% by vol.:
(2015) experimental analysis of 11:1, 1500 rpm ↑engine cooling load, ↓BTE, due to increasing the
the proportion of fuel combustion stage
hydrogen on syngas fueled
by SI engine
Cha et al. [307] Study of a SI engine by H2/CO: 50/50 A 4-stroke, 4-cyl, TC, CI — ↑Syngas mixing ratio (5%, 10%, and 15%):
(2015) adding syngas into CH4, engine, CR= 13:1, Combustion was improved, ↑fuel conversion
with a mixing ratio various ITs efficiency, ↑NOx, and the rate of ↑ was non-linear.
A/F and spark timings should be calibrated.
Homdoung et al. Effects of variable CRs, H2/CO/CH4/CO2/ A 4-stroke, 1-cyl, CI —
[286] engine speeds and loads on O2/N2: 8.5/30.5/ engine, CRs= 9.1-17:1, The modified engine worked well at higher CRs
(2015) the performance of a small 0.35/4.8/6.3/49.55 rpm=1000-2000, with PG than with gasoline. Up to 23.9%, BTE was
engine operated on PG 8.5%H2; 30.5% CO; load=20%-100% obtained.
0.35% CH4; 4.8% Rather to the original diesel:
CO2; 6.3%O2 and ↓BTE (11.3%), ↓engine’s smoke density, ↑CO, with
49.55%N2 similar HC emissions.
Kim and Kim Feasibility assessment of H2/CO/CO2: 70/ A 4-stroke, 1-cyl, CI 1-D multicylinder
[318] hydrogen-rich syngas 15/15 engine, CRs= 15:1, engine model in GT- The vehicle with the syngas engine could travel for
spark-ignition engine for rpm=1800 SUITE 152.6 km without refilling. The syngas bus
heavy-duty long-distance exhibited lower fuel consumption than the
vehicle application equivalent CNG bus by 15%.
Denotations: ↑: increase, ↓: decrease, ~: nearly constant or insignificant, //: similar to above.

IMEP: indicated mean effective pressure, CR: compression ratio, ITE: indicated thermal efficiency, BTE: brake thermal efficiency, CoV IMEP: coefficient of variation of
IMEP, TC: turbocharge, NA: naturally aspirated, IT: ignition timing, SOI: start of ignition, ϕ: the fuel/air equivalence ratio, WOT: wide open throttle, and MBT:
maximum brake torque.
smoother combustion). This can be justified according to the following: 2- Chemical effect: OH generated from the first HR stage may be
consumed by hydrogen (i.e., H2 + OH = H2O + H), suppressing
1- Dilution (or concentration) effect: O2-deficiency happening by chemical kinetics of the mixture before the combustion.
replacing O2 with H2 leads to weakening the low-temperature HR
(LTHR or partial oxidation) stage of the fuel, which is a preparation In the latter case, the strong thermal effect of syngas on high octane
step of combusting the in-cylinder mixture, resulting in lowering the fuels like natural gas may result in rising the specific heat capacity of the
hydroxyl (OH) concentration, and thus retarding the high- mixture and increasing the compression temperature of in-cylinder
temperature HR (HTHR or combustion). It was shown that the charge eventually, depending on intake temperature. To highlight the
magnitude of LTHR has a direct correlation with advancing or thermal effect, valving strategies may be helpful without needing intake
retarding the HTHR [342]. preheating. Eng et al. [343] reopened the exhaust valve during the
27
Fig. 30. Operating range for syngas fueled HCCI engine under various engine speeds at a constant compression ratio [327].
intake stroke to obtain more hot gases in the cylinder, i.e., exhaust They stated that this concept with HCCI combustion can result in
re-breathing. The combustion chamber of an HCCI-mode engine fueled high thermal efficiency over a wide operable window of equivalence
with iso-octane and partial oxidation gas (POX gas) added to the intake ratio and high overall system efficiency, which is comparable to the fuel
experiences a high quantity of internal residuals and thermal stratifi- cell system. A brief summary of their work is given as follows:
cation; this causes the ignition timing to advance. Another technique is
negative valve overlap (NVO), in which the residuals are trapped and 1- Using an optimal compression ratio along with a suitable DME/MRG
recompressed, as described in section 4. ratio (with increasing amount of MRG) resulted in a higher thermal
Dimethyl ether (DME) has some more interesting combustion char- efficiency for HCCI combustion, and a higher overall thermal effi-
acteristics than diesel (i.e., soot-free combustion and lower NOx emis- ciency can be achieved by using waste heat recovery (WHR) [346].
sions), but still faces challenges with regard to its physical properties (i. 2- Higher compression ratio led to advanced ignition timing, but with
e., low viscosity, which leads to higher leakage and in turn engine higher heat transfer (cooling) loss [346] due to the small quenching
durability concerns) [344]. Similar to syngas, DME can benefit from distance and fast burning velocity of hydrogen content in MRG
on-board fuel reforming in vehicles via a methanol-steam reforming (syngas) [353].
process [345]. With these mentioned potentials for DME, and emerging 3- While both H2 and CO have influence on retarding the second stage
dual-fueling strategies for achieving controllable HCCI combustion, HR, the role of H2 is much more intense due to having a larger effect
Shudo et al. [346–351] conducted a number of experiments supported on first stage HR by consuming the OH radical [347]. Also, com-
by chemical kinetics analysis, to run a single-cylinder SI engine con- bustion of hydrogen during LTHR is controlled by H + O2 → OH + O
verted to the HCCI mode fueled with DME and MRG (methanol reformed and less reactive HO2 [349]. CO2 also caused a slight retardation
gas or syngas). Both fuels were produced from endothermic methanol [348].
steam reforming using engine exhaust gases, and then followed by 4- Despite larger amounts of H2 being useful for ignition control effects,
methane dehydration for DME generation; with a single fuel tank H2 has a lower effect on WHR.
required for liquid methanol, as seen in Fig. 31 (left). While using 5- Highest overall thermal efficiency was expected from thermal
simulated reformate comprising 67% H2 and 33% CO (by volume) in the decomposition [350] and highest output would be from steam
experiments, they found that ignition of the second heat release (main reforming, which resulted in MRG with higherH2 content.
combustion) and engine load could be controlled by altering the pro-
portion of the MRG/DME ratio; by increasing the ratio, the overall A collection of works [354–356], which have studied use of a cata-
system efficiency reached 49% and a 14% rise in engine efficiency over lytic reforming reactor in the exhaust stream of a natural gas HCCI en-
the baseline was obtained (see Fig. 31 (right)). This controllable HCCI gine, have shown that hydrogen-rich gas could be a substitute for intake
concept was achieved due to delayed ignition and increased temperature preheating for the same emitted NOx, greatly benefiting low engine
before the main combustion, corresponding to the hydrogen content of loads. These works have been expanded to form on-board fuel reforming
the MRG. Similarly, they proposed this concept by storing a tank of DME strategies, which are called reformed exhaust gas recirculation (R-EGR),
and partial oxidation of it to produce DME reformed gas (DRG) [352]. as described in the next section. Those works concluded that employing
Fig. 31. Schematic of proposed concept (left) heat balance of this concept (right) (Reprinted from [349] with permission of Elsevier).
28
on-board fuel reforming products to extend the lower load-limit, along Later in 2008, the same researchers [361] examined the effects of
with using the supercharging and high internal EGR technologies for varying simulated RG compositions on HCCI combustion timing with
increasing the upper load-limit could be a solution for extending the natural gas, iso-octane, and n-heptane. The RG compositions chosen had
entire operating window of HCCI engines fueled with natural gas or H2/CO ratios of 3/1 (H2/CO=75/25 %by volume) and 1/1
gasoline [357]. (H2/CO=50/50 %by vol.). The first and latter are a typical output of
At the University of Alberta, a series of research works have been natural gas steam reforming and gasoline partial oxidation in an
experimentally conducted to capture the effect of reformer gas (RG) on-board fuel reformer, respectively. Fig. 33 reveals the impact of two
blending on the base fuels, including high-octane primary reference selected RG compositions on ITE and SOC for selected fuels. For
fuels (PRF) [358], low-octane PRF [359], natural gas [360,361], iso-octane, both compositions retarded the combustion timing similarly
n-heptane [361,362], and iso-octane [361] in a variable compression (i.e., +1.5 CAD per each 10% of RG); which implies that iso-octane may
ratio, single-cylinder CFR engine operated in the HCCI mode. The effect behave more sensitively to the thermal effect than the chemical effect of
of varying RG blend fraction (mRG +m mRG
base fuel
∗ 100) was studied, while the H2 content in RG. Also, at the same IMEP, the RG 75/25 moved the
intake temperature, air-fuel ratio, and EGR were kept constant. For mixtures towards the richer region compared to the RG 50/50, which
low-octane PRF, such as PRF0 (i.e., 100% n-heptane) and PRF20 (i.e., was operated at the leaner condition. Fig. 33(a) shows the ITE with the
20% iso-octane and 80% n-heptane), it was observed, in conformity with expanded rich-limit of RG75/25 decreased, presumably due to increased
the above discussion, that increasing values of RG replacement leads to a heat transfer losses. For natural gas, both compositions advanced the
retarded and prolonged combustion event even after TDC [359]. As a combustion timing with no obvious discrepancy by increasing the spe-
reported case, 20% of RG replacement delayed the start of combustion cific heat ratio and compression temperature of the mixture. The sup-
(SOC) by about 14 CAD (3 ms) and lengthened the combustion duration porting modeling studies showed that the thermal effect is dominant,
by 50%, and increased the indicated power and thermal efficiencies by and the chemical effect arising from the various compositions used has
17% and 12%, respectively. This remarks that combustion can be an insignificant effect. Also, ITE with RG75/50 is increased due to H2
controlled by syngas without using other strategies like preheating the properties in extending the lean limit of the mixture (see Fig. 33(c)).
intake charge, adjusting dilution level or EGR, and by using valving For n-heptane, both compositions retarded the combustion timing by
strategies, which do have slow response or penalty in engine power. suppressing the LTHR stage (see Fig. 33(e)), in which H2 indicates a
They also applied a similar procedure to high-octane PRF [358], PRF80 stronger effect than CO (see Fig. 34). Indicated efficiency of the RG 75/
and PRF100, except by increasing CR and intake temperature (from 25 fueling decreased after reaching an RG fraction of 10%, staying away
100◦ C to 140◦ C). Fig. 32 displays representative in-cylinder pressure from optimal combustion phasing of n-heptane, with more retardation
traces and their corresponding heat release. As can be seen, the peak (see Fig. 33(b)).
in-cylinder pressure and pressure rise rate were reduced by increasing Thus, it is evident that RG with emphasis on the H2 content acts as an
the RG, with peak pressure shifting from approximately TDC to LTHR inhibitor (chemical effect) due to its participation in low tem-
approximately 10◦ ATDC. RG substitution effectively smoothed com- perature oxidation for fuels having two-stage heat releases like n-hep-
bustion by retarding the combustion phasing towards a more efficient tane, PRFs, and DME. In this regard, Gou et al. experimentally [364] and
timing for HCCI combustion, and by widening the combustion duration. numerically [365] studied the thermal, chemical, and dilution effects of
Replacing RG did not impact NOx emissions but increased CO and HC hydrogen enrichment on HCCI fueled with n-heptane. Recently, based
marginally, therefore, with a slight decrease in combustion efficiency. on numerical validation against experimental data acquired from the
The indicated fuel conversion efficiency was slightly increased, because work of Hosseini and Checkel [358,359], multi-zone modeling studies
the combustion phasing was better than the baseline case. It was found with detailed chemical kinetics and with artificial inert species have
that 40% of RG blended in a PRF80 fuel (by mass) has an octane number been carried out to assist in the fundamental understanding of the
as close as PRF100, but with totally different combustion characteristics. chemical, dilution, and thermal effects of RG blending on the base fuels
Thus, the calculated octane number is not a good indicator of the during the HCCI combustion [363,366-370]. For n-heptane [370] and
characteristics of HCCI combustion. low-PRF fuels [368], the dominance of the chemical effect among other
Fig. 32. Effect of reformer gas on in-cylinder pressure traces and heat release rates with PRF80 at CR = 14.4:1 and EGR = 33% [363].
29
Fig. 33. Effect of RG composition on ITE and start of combustion of HCCI combustion of iso-octane (a & d), n-heptane (b & e), and natural gas (c & e) [361].
Fig. 34. Heat release of HCCI combustion of n-heptane for: (a) RG 75/25 and (b) RG 50/50 [361].
effects on retarding the SOC was confirmed, mostly due to H2 presence Table 10 summarizes the main syngas works on HCCI engines pub-
in low-temperature oxidation reactions during the negative coefficient lished in the literature. Syngas as a primary fuel of an HCCI engine has
temperature or NTC (period between LTHR and HTHR). For high PRF serious obstacles such as its lower combustion efficiency than 90%, high
fuels, the dilution effect increases by increasing the PRF number [368]. CO emissions, and excessive dependency on intake preheating and high
30
Table 10
Summary of syngas utilization studies in HCCI engines.
Author(s) Experiments/Simulation Main fuel Fuel addition (% Objective(s) Finding(s)
by volume)
Syngas addition (Bold ones showed retarding effect on SOC), others showed advancing effect on SOC)
Shudo and 1-cyl SI converted to HCCI, CR= Dimethyl-ether A model gas for Using onboard generated DME Combustion ignition timings were
Ono [346] 8.3, 9.7, 12.4, 15.8, rpm= 1000, (DME) methanol reformed and MRG by waste heat recovery controlled by DME/MRG ratio.
(2002) Tintake= at room temp. gas (MRG): to control HCCI combustion With ↑ CR: HRR advances and cooling loss
H2/CO: 67/33 timing increases
Peucheret and A modified 1-cyl SI, Natural gas 10% of Reformate A solution to: Proposed an onboard catalytic exhaust gas
Wyszynski CR= 12.0-14.5, Tintake=140◦ C- (NG) was added into High auto-ignition of NG fueling fuel reforming strategy for NG HCCI engine.
[354] 230 C
◦
intake. needs a high compression ratio Stable combustion
(2004) H2/CO/CH4/CO2/ (CR) and high intake Lower intake temperature requirement
O2/N2: 10/0.3/0.4/ temperatures at low load ↓ NOx ↑ HC ↑ CO.
9.0/0.8/79.5
Hosseini and 1-cyl CFR, CR= 16.0, 17.0, 18.5, CNG Reformer gas (RG) A solution to: By ↑ RG replacement fraction:
Checkel rpm= 470, 700, Tintake= 140◦ C was introduced to Pure NG-fueled HCCI with knock Expanding the operating window to very
[360] the intake charge severity, high NOx and low lean range:
(2006) (0-60% replacement performance ↓ Knock
with the main fuel ↓ NOx ↑ CO ↑ Combustion eff.
by mass)
H2/CO = 3/1
Hosseini and 1-cyl CFR, CR= 14.4, 16.0, PRF80 (20% n- RG A solution to SI at part load: By ↑ RG replacement fraction:
Checkel rpm= 700, Tintake=140◦ C heptane, 80% (0-40% replacement RG effects on high octane fueling Expanding and shifting the engine operating
[358] iso-octane) by mass) in an engine with dual SI/HCCI window towards ultra-lean:
(2007) PRF100 RG1: H2/CO = 3/1 modes (improving HCCI ↓ Maximum cylinder pressure
RG2: H2/CO = 1/1 operating at part load) ∼ NOx ↑ HC ↑ CO ↓ Combustion efficiency.
Hosseini and 1-cyl CFR PRF0 RG A solution to high emissions of By ↑ RG replacement fraction:
Checkel CR= 9.5, 11.5, rpm= 700, PRF20 (0-40% replacement diesel engines: Expanding the engine operating window on
[359] Tintake= 100◦ C by mass) RG effects on low octane fueling the rich side:
(2007) H2/CO = 3/1 in an effective HCCI ↓ Peak cylinder pressure and PRR
↑ IMEP ↓ CoV IMEP ↑ Engine power
∼ NOx ↑ HC ↑ CO.
Hosseini and 1-cyl CFR, CR= 14.4, rpm= 800, iso-octane H2/CO = 3/1 Thermal and chemical effects of At high initial temperature:
Checkel Tintake=100◦ C H2/CO = 1/1 RG blending Thermal effect dominates (advancing the
[361] combustion phasing). At moderate
(2008) temperature: Chemical effect dominates
(retarding the combustion phasing).
Overall: insignificant effect on combustion
Hosseini at al. 1-cyl CFR n-heptane RG blending A solution to: With RG blend fraction of 10%:
[371] CR= 11.8, rpm= 800, Tintake= (0-40% replacement Knock propensity at high load 4.4 CA retardation in SOC,
○
(2009) 110◦ C by mass) ↑ H2content was more effective in retarding,

RG1: H2/CO = 3/1 ↑IMEP by 0.25 bar, ↑ITE by 2.8%
RG2: H2/CO = 1/1
Voshtani et al. Single-zone and multi-zone NG RG1: H2/CO = 3/1 RG enrichment effects on a HCCI
[372] thermodynamic-kinetic models, RG2: H2/CO = 1/1 combustion engine operating Both chemical and thermal effect advance
(2014) based on [360] with NG. the SOC, but dilution retards that.
chemical > dilution > thermal
H2 and CO in RG at low and high amount of
RG have strong responsibility of altering
SOC.
Zheng et al. Zero-dimensional (0D) and DME DME reformed gas Effect of REGR on HCCI fueled REGR (EGR + DRG): can delay ignition
[373] three-dimensional (3D) (DRG) produced by with DME time, allowing main combustion closer to
(2014) combustion models REGR strategy: TDC, minimizing negative compression
H2/CO: 66.7/33.3 work.
Compared to EGR-only: ↓HC, ↓CO.
Neshat et al. Multi-zone model PRFs (0-30% Effect of RG blending on HCCI By RG addition: ↓ H abstraction reactions
[368] and semi detailed chemical- replacement) combustion of PRFs rate ↑ decomposition fuel time ↓ LTHRR
(2016) kinetic mechanism H2/CO = 1/1 peak.
↑ PRF number: ↑ dilution ↓ chemical effect.
↓ NOx ↑ HC ↑ CO
Reyhanian Multi-zone model NG H2/CO = 3/1 Understanding of various effects
and Hosseini with detailed chemical-kinetic iso-octane of RG addition, with composition Thermal effect:
[370] mechanism normal- of RG in HCCI combustion SOC was advanced for all fuels.
(2018) + heptane Chemical effect:
an artificial inert species method SOC was advanced with NG and iso-octane
and retarded with n-heptane, mainly due to
H2.
Dilution effect:
SOC was retarded for all fuels.
2D CFD simulation iso-octane H2/CO = 2/1 Combustion and emission Ignition timing and combustion duration
Kozlov et al. At ϕ= 0.2, 0.4 characteristics of HCCI operating were retarded and decreased, respectively,
[374] on iso-octane/syngas especially at ϕ= 0.2.
(2018)
(continued on next page)
31
Table 10 (continued )
Author(s) Experiments/Simulation Main fuel Fuel addition (% Objective(s) Finding(s)
by volume)
↑ NOx at ϕ= 0.4
↓ HC ↓ CO at ϕ= 0.2
100% syngas fueling (%vol)
Achilles et al. Modified, 1-cyl truck/bus PG: — Comparing producer gas (PG) PG as compared to NG fueling:
[375] engine, CR= 22.5:1, 1050 rpm, H2/CO/CH4/ and natural gas (NG) in HCCI Wider ϕ range, faster combustion, more
(2011) Tintake=80-180 ◦ C for NG, CO2/N2: 30/20/ engine smooth and stable, ↓maximum IMEP,
Tintake=75-145 ◦ C for PG 1.0/15/34 ↓intake temperature requirement, ↓THC,
and ↑CO emissions.
Bika et al. 1-cyl engine, CR=21.2:1, 1800 RG 1: H2/CO: — Studying HCCI engine with two ↑CO fraction in RG: ↑auto-ignition
[323] rpm, Tintake=60-90◦ C 75/25 different syngas compositions at temperature, ↑intake temperature
(2012) RG 2: H2/CO: ϕ=0.26 and 0.3 requirement, ↓combustion efficiency (83-
50/50 88%). At low load and high intake temp.:
↑↑heat transfer to wall.
Przybyla et al. SI engine with CR=8.2, HCCI H2/CO/CH4/ — Fueling of SI (ϕ=1) and HCCI ITE: SI>HCCI by about 3.5%,
[328] engine with CR=20, CO2/N2: 10.1/ (ϕ=0.5) engines with low Energy loss in exhaust based on chemical
(2016) Tintake=132-153◦ C 24.9/2/5.2/ heating value PG energy of the fuel:
57.8 HCCI (10%) > SI (2%). The SI efficiency can
be easier than HCCI optimized.
Bhaduri et al. Air cooled mono-cylinder HCCI H2:CO ratio — Proposing a novel tar-tolerant Moisture in syngas: water’s damping effect
[329] engine CR= 12, rpm=1500, from 30:70 to HCCI engine in response to on chemical reaction rate, delay in
(2015) Tintake=230-270◦ C 55:45% various syngas compositions combustion (positive effect). Tars in syngas:
(with water and tars) ↑LHV of syngas, but their effect on the
combustion phasing were negligible.
Bhaduri et al. Air cooled mono-cylinder HCCI H2/CO/CH4/ — Testing the stability of the engine Over a 24 h test: high ITEs of 33%-39%, a
[330] engine CR= 12.2, rpm=1500, CO2/N2: in response to the naturally low IMEP of about 2.5 bar, and a relatively
(2016) Tintake=250◦ C 15-25/15-25/1- occurring random variations in high NOx (due to ammonia in the syngas)
3/8-15/balance the syngas composition were observed.
Bhaduri et al. Air cooled mono-cylinder HCCI H2/CO/CH4/ — EGR effect on a tar tolerant HCCI ↑EGR from 0% to 25%:
[376] engine CR= 12.2, rpm=1500, CO2/N2: engine ↑IMEP from 2.5 bar to 3.3 bar, via damping
(2017) Tintake=250◦ C 18/22/1.0/11/ maximum pressure rise and allowing higher
48 ϕ.
Starik et al. Zero-dimensional single-zone H2/CO: 50/50 — Improvement of combustion in a Ignition can be accelerated before TDC, thus
[377] thermochemical model syngas fueled HCCI engine by ↓intake temperature. At the same IT, 1% of
(2017) adding ozone or excited oxygen SDO in total oxygen may ↑ the power by 7-
molecules 14% compared to the base case with higher
intake temperature.
Jamsran et al. 1-cyl 126TI-II series, Doosan H2:CO:CO2: — Investigate the combustion and ↑ intake temperature and boosting: Low-
[378] Infracore, CR= 17.1, rpm= 30/2545 emissions of an HCCI engine calorific syngas (15% H2, 15% CO, and 70%
(2021) 1800, Tintake=140◦ C-140◦ C- 30/15/55 with various syngas CO2 3.34 MJ/Nm3) can be auto-ignited
230◦ C 15/25/60 compositions
15/15/70
Denotations: ↑: increase, ↓: decrease, ~: nearly constant or insignificant, //: similar to above.

IMEP: indicated mean effective pressure, CR: compression ratio, ITE: indicated thermal efficiency, BTE: brake thermal efficiency, CoV IMEP: coefficient of variation of
IMEP, TC: turbocharge, NA: naturally aspirated, IT: ignition timing, SOI: start of ignition, ϕ: the fuel/air equivalence ratio, WOT: wide open throttle, and MBT:
maximum brake torque.
compression ratios, hindering it from progress in vehicular engines. In conventional diesel combustion (CDC). Similar to compressed natural
contrast, syngas as a fuel additive can support the HCCI combustion of gas (CNG) and liquefied petroleum gas (LPG) [382], syngas is also a
other fuels to push its operating window to a broader range with good candidate for the premixed fuel in dual fuel mode owing to its
attainable high thermal efficiency comparable to that of SI combustion. variable composition, flexible feedstocks, and possibility of on-board
Briefly, syngas addition can be regarded as a means of extending the production on a vehicle [383,384]. In this part, syngas utilization in
lean- or the rich-side of the operating equivalence ratio of HCCI com- dual-fuel combustion is analyzed and reviewed in detail.
bustion fueled with low- and high- octane fuels, respectively.
3.3.1. Diesel/syngas engine performance and emissions
Several studies have been focused on using syngas in dual-fuel con-
3.3. Syngas in dual-fuel engines cepts, where diesel is directly injected into the cylinder as a pilot fuel to
initiate the ignition, while syngas is introduced into the intake port as a
Stable combustion of a premixed syngas-air mixture is problematic in premixed fuel. Syngas application in diesel engines as a secondary en-
CI engines due to the low reactivity and high auto-ignition temperature ergy carrier has been considered to be a promising alternative method
(> 500 ◦ C) of syngas [59]. Also, more stringent diesel engine emissions mainly for emissions reduction. In the early stages of the research, ef-
regulations have directed research towards further development of forts were dedicated to clarifying and distinguishing dual-fuel combus-
advanced combustion engines. Meanwhile, dual-fuel operation tion from CDC, which led to the recognition of two-stage combustion for
[379–381] with a premixed fuel and pilot-injected diesel has shown dual fueling at higher engine loads. Specifically, Garnier et al. [385]
interesting advantages in terms of efficiency and emissions over
32
established a predictive model to determine the ignition delay and injection strategy has not received much attention. It is also worth
combustion of a syngas/diesel CI engine, based on an empirical model noting that higher efficiency with H2/diesel dual operation than
and three Wiebe functions (for pilot-fuel, gaseous-fuel, and diffusive diesel-only operation has been obtained when H2 is injected directly
combustions) calibrated with experimental data. After that, some ex- within the cylinder [393,394]. Antunes et al. [393] indicated fuel effi-
periments were conducted to characterize syngas/diesel dual-fuel ciencies of 34% for H2/diesel and 43% for H2-only compared to 28% for
combustion in terms of engine performance and emissions by varying the diesel-only case.
syngas compositions. Sahoo et al. [386] evaluated the dual-fuel concept At Okayama University, a group has conducted research on dual-fuel
using syngas with a H2/CO ratio = 1/1 by volume in a single-cylinder operation [395–402] with different producer gas generated by biomass
diesel engine at a constant 1500 rpm and injection timing of 23◦ gasification (BMG) and coke-oven gasification (COG) in a supercharged,
BTDC by altering the load from 0 to 100%. Compared to the diesel-only single-cylinder diesel engine. They reported performance and emissions
mode, a lower BTE with a peak reduction of 21% at 80% load, increased characteristics of a dual fuel syngas/diesel engine considering the effect
exhaust gas temperature (50-100◦ C), increased CO and HC emissions for of pilot injection. Their key findings are summarized below:
all engine loads were observed; however, both NOx emissions and peak
cylinder pressure were decreased. Also, by varying H2/CO’s volumetric • The effect of injection pressure and amount of pilot fuel (diesel):
ratio during various loads [387], the maximum HC, CO, and CO2 Smoke emissions were increased with increasing amount of injected
emissions and minimum BTE occurred with 100% CO; however, the diesel, while decreasing with increasing injection pressure (i.e., 2
opposite behavior was observed with 100%H2 [388]. Moreover, from mg/cycle with 400 bar has the same measured smoke with 10 mg/
the second law efficiency point of view, the same authors [389] cycle with 800 bar) [395].
concluded that a syngas/diesel CI engine’s theoretical performance fa- • The effect of pilot-injection strategy with timings ranging from 4 to
vors higher load (beyond 40% load) over lower loads, where the exergy 12 BTDC, injection pressures of 40, 60 and 80 bar, pilot quantities of
○
destruction increases due to the heat transfer losses, see Fig. 35 (right). 2 to 10 mg/cycle: 80 MPa with 3 mg/cycle was optimal for best
Bika et al. [390] experimentally assessed the cycle efficiency of a performance [396].
single-cylinder syngas/diesel engine for four diesel substitutions of 0% • The lowest energy content PG used showed the lowest level of NOx
(diesel-only baseline case), 10%, 20%, and 40%, at two net IMEPs of 2 emissions and smoke-free operation with smooth combustion [398].
and 4 bar, see Fig. 35 (left). Bottled gases of H2 and CO with varying • The effect of varying hydrogen content in PGs from 13.7% to 20% by
proportions (H2/CO: 100/0, 75/25, 25/75, and 0/100 by %volume) volume: High hydrogen content in syngas led to a wide operating
were used to simulate syngas. Compared to the diesel baseline case, they
displayed lowered efficiencies at both loads (10-25% for 2 bar and
5-17% for 4 bar IMEP), presumably attributed to the unburned syngas
discharged in the exhaust stream. Moreover, NOx emissions were
increased for all syngas proportions at 4 bar IMEP.
So far, syngas/diesel dual fuel operation findings have shown a
common trend of efficiency degradation particularly at low- and part-
engine loads with the decrease in combustion efficiency being due to
increased H2 and CO. Wagemakers and Leermakers [382] reviewed the
comparison of dual fueling of diesel with various gaseous fuels like CNG,
LPG, syngas, and hydrogen. They reported that all gaseous fuels could
reduce soot emissions except for syngas. As a consequence of the dual
fuel mode for diesel substitution, methane demonstrated better perfor-
mance than syngas. The drawbacks generally reported from engines
operating on gaseous fuels/diesel at part loads are excessive HC and CO
emissions, and high fuel consumption, resulting in lower combustion
efficiency and low thermal efficiency [391]. Also, a recent review on
natural gas (NG)/diesel engines reported the same trend with up to a
2.1% fall in engine power and low BTE at low-to-part engine loads,
while a maximum rise of 3% at high loads would be attainable compared
to the diesel engine [392]. The only reported solution to the stated
Fig. 36. Premixed mixture ignition in the end-gas region (PREMIER) com-
problems was to increase the H2 content of the syngas, while the pilot bustion concept (Reprinted from [401] with permission of Elsevier).
Fig. 35. Comparison of cycle efficiency vs. diesel substitution at 4 bar IMEP (left) [390] and exergy efficiency vs. engine load (right) [389] for various H2/CO
volumetric ratio (Reprinted with permission of Elsevier).
33
range (i.e., equivalence ratio) and high thermal efficiency; however, would augment more stability in the flame of the H2+PG/ diesel case,
increased NOx emissions [397]. presumably attributed to the high stability of H2-air flames while the PG
flames tend to become unstable. Nevertheless, the increase in H2 from a
Following this work, Azimov et al. [403] suggested a novel com- certain amount (here, 40% of H2 in PG/H2 mixture) resulted conversely
bustion strategy with inspiration from the knock event in the end-gas in the destabilization of the flame, which may be attributed to a decrease
regions of SI engines. The natural gas engine with a homogenous in the Markstein length [412]. At higher loads, much oxygen was
air-fuel mixture could be controlled; when natural gas combusts with the consumed by PG/H2 while hydrocarbon fuel received not enough oxy-
flame front supported concurrently by pilot-injected diesel fuel, which gen; thus, a reduction in BTE would be sensible.
has already been ignited. They called it premixed mixture ignition in the Fig. 38 depicts the variation in emissions at 80% load conditions for
end-gas region (PREMIER) combustion; also, shifting from diffusion the three cases, with PG/H2 ratio of 60:40. Because of the low pilot
combustion to flame front combustion for syngas/diesel dual-mode was quantity, HC emission is high with gaseous fuel replacement at higher
reported in [404]. With the proposed combustion strategy illustrated in engine loads. As noted above at higher loads, larger gaseous fuel sub-
Fig. 36, the effect of H2 and CO2 content in syngas on the performance stitutions, which replace some of the air during the intake stroke, can
and emission of a syngas/diesel engine was investigated under lean lead to increased utilization of oxygen and the reduction of oxygen
operating conditions [401]. PREMIER combustion was observed pri- available for further combustion of the pilot diesel fuel, thus raising the
marily when the amount of pilot fuel used was minimal. Additionally, HC level. Higher CO emissions implies that incomplete combustion and
they [401] stated that an increase in syngas’s hydrogen content short- burning of lubrication oil take place, together with the lower heating
ened the main combustion period and caused the mean combustion value of PG, CO-containing PG (lower mean effective pressure), lower
temperature, IMEP, and efficiency to increase. They also found that adiabatic flame temperature, and lower over-mixing. For the PG case,
diesel could not be substituted completely, or syngas could not be used the drop in NOx may be due to the lower adiabatic flame temperature of
as a single fuel in the diesel engine. PG and the absence of nitrogen in the PG, as well as a more uniform
Some other researchers [396,405-409] have explored the feasibility temperature distribution achieved with the gaseous fuel-air mixture,
of the dual-fuel concept with syngas generated from various sources by allowing the number of high-temperature regions around the diesel
either gasification like producer gas or on-board reforming processes flame to decrease.
like reformer gas. Dohle et al. [409–411] performed a series of experi- The summary of relevant literature that has been reviewed so far
mental studies to investigate the combustion characteristics of a indicates:
dual-fuel diesel engine fueled by three cases of H2, PG, which was
generated from rice husk, and PG in combination with H2 (PG + H2) as a 1- Feasibility of syngas/diesel among gaseous fuels/diesel as a good
secondary fuel. Fig. 37 (left) shows the thermal efficiency variations choice for dual-fuel operation: according to all cited work here, while
with five different combinations of gaseous fuel replacement at 13%, it has no beneficial aspects in terms of engine performance, and even
40%, and 80 % engine loads. Again, up to 30% load, a decrease of BTE also emissions, it performs well at low and part loads.
was found due to the poor combustion characteristic of low calorific 2- The influence of syngas compositions confirmed that increasing the
producer gas even in combination with H2 at low loads. While by H2 content of the syngas has considerable merit in combustion
increasing the total mass quantity of fuel through increasing the engine improvement.
load, the combustion of the dual fuel mode improved due to the higher 3- Modifying the pilot injection strategy can be a promising way of
flame velocity and diffusivity of gaseous fuels compared to diesel. As extending the prospects of the dual-fuel mode with syngas.
compared to H2/diesel and PG/diesel cases (see Fig. 37 (right)), H2
Fig. 37. Brake thermal efficiency for PG+H2/diesel mixtures (left) and for H2/diesel and PG/diesel (right) at different loads (Reprinted from [409] with permission
of Elsevier).
34
Fig. 38. UHC, CO and NOx emissions at 80% load for different Diesel+PG+H2 mixtures (Reprinted from [409] with permission of Elsevier).
4- This concept’s flexibility with a broad range of synthetic gases compression stroke. With respect to emissions, the introduction of syn-
derived from varying sources has led to more attention in research. gas increases NOx emissions due to the higher temperature induced by
the higher adiabatic flame temperatures of carbon monoxide and
The pilot injection strategy of Roy [396], is regarded as a means of hydrogen. It also decreases soot emissions, due to the higher tempera-
controlling the in-cylinder reactivity and combustion phasing. By ture of combustion resulting from the hydrogen content of the syngas,
introducing two fuels having different reactivity (i.e., low and high which allows soot oxidation. It is also confirmed in a recent study by
reactivity) with a pair of separate injections for the high-reactivity fuel Zhong et al. [425] that hydrogen in syngas promotes NOx emission but
within the cylinder control can be achieved. This concept has been suppresses soot emission. Even though HC and CO emissions at medium
presented by Kokjohn et al. [413], who named it reactivity-controlled load operation are relatively low, syngas addition slightly reduces HC
compression ignition (RCCI). In terms of commercialization, RCCI emissions and increases CO emissions due to the syngas CO content, as
[380,381,414,415] is under more investigation for single-fuel operation, depicted in Fig. 40. In another work, Rahnama et al. [418] numerically
where high reactivity like diesel as the parent fuel, can be converted into studied the effects of reformer gas composition (3% syngas enrichment
a low-reactivity fuel by suitable in-situ on-board fuel reforming [34, of intake air) on the combustion and emission characteristics of a natural
416-424]. gas/diesel RCCI engine at a low load condition. They observed that with
Rahnama et al. [34] investigated the effects of reformer gas, increasing CO fraction in the syngas, peak pressure, ringing intensity,
hydrogen, and nitrogen on the combustion characteristics of a and pressure rise rate increased significantly. Using a mixture with a
heavy-duty RCCI engine fueled with natural gas/diesel. The impact of higher CO content, a shorter ignition delay and combustion time and
the introduction of syngas (up to 5%vol) at medium load engine oper- advanced CA50 were achieved. These were interesting results but was
ation was also investigated. As shown in Fig. 39, by adding syngas, peak attributed to the lower cylinder temperatures as a result of using lower
pressure, combustion, and gross indicated efficiency values significantly intake temperatures with high-H2 syngas.
increase, but after a certain point, the gross efficiency drops due to Chuahy and Kokjohn [416] were the first to experimentally examine
improper phasing in combustion and heat release during the the effect of varying syngas composition on a single-fuel RCCI fueled
Fig. 39. Syngas addition effects on cylinder pressure and HRR, gross indicated efficiency and ringing intensity at 9 bar IMEP for a natural gas/diesel RCCI engine
(Reprinted from [34] with permission of Elsevier).
35

Progress in Energy and Combustion Science

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Progress in Energy and Combustion Science

Uploaded by

Copyright:

Available Formats

Progress in Energy and Combustion Science 90 (2022) 100995

Contents lists available at ScienceDirect

Progress in Energy and Combustion Science

Synthesis gas as a fuel for internal combustion engines in transportation

minimize greenhouse gas (GHG) emissions on a well-to-wheel basis to

Nomenclature HRR Heat release rate

1.1. What is syngas?

Syngas (an abbreviation for synthesis gas) can be derived by gasiﬁ-

Fig. 2. Schematics of biomass gasiﬁcation process [66].

1.4. Historical use of syngas as an engine fuel

Steam y Endothermic • Highest H2 generation (H2/CO ~ Highly endothermic

Combustion y y ΔhR = −51161 If enough O2 is available.

Water-gas shift (WGS) CO + H2 O

Fig. 4. Syngas terminology used in the current review paper.

• The transition from syngas as an IC engine fuel to syngas as a

2.1. Physicochemical properties of syngas relevant to engines

Table 5 lists the physicochemical properties of typical syngas mix-

+ OH reaction. Those works were completely reviewed in [157,158].

efﬁciency, as can be seen in Fig. 20 (a). Moreover, the lowest CA0-10%

3.1.2. 100% syngas in CI engines retroﬁtted to SI mode

combustion, Maurya et al. [327] used a stochastic reactor model (SRM)

(continued on next page)

Denotations: ↑: increase, ↓: decrease, ~: nearly constant or insigniﬁcant, //: similar to above.

(2009) 110◦ C by mass) ↑ H2content was more effective in retarding,

Denotations: ↑: increase, ↓: decrease, ~: nearly constant or insigniﬁcant, //: similar to above.

You might also like