1. Introduction
Limiting global warming requires rapid, far-reaching actions in land, energy, industry, buildings, transport and cities [
1]. There is already strong growth of alternative energy sources. However, sustainable liquid fuels will be needed for decades to satisfy the world’s growing energy demands [
2]. Furthermore, internal combustion engines are set to continue as prime energy producers because there is an established worldwide infrastructure for liquid fuel distribution [
3] and the engines offer fuel efficiency, strength and durability [
4]. Combustion systems need to be able to operate with a variety of renewable and sustainable fuels and yet also meet increasingly stringent emission legislation. Expanding fuel choice calls for a clearer focus diversification, quality, and usability of any new fuel [
5]. Alternative fuels can reduce diesel engine emissions and several beneficial results have already been obtained in practice [
6]. However, availability of alternative fuels must evolve to ensure energy security and sustainability. Moreover, their higher price presents a challenge when making new fuel choices [
6]. According to Sirviö [
6], an alternative fuel must fulfil quality assurance and standard requirements to secure consumer acceptance. In addition, after-treatment applications support a conventional and alternative target for fuels to reduce exhaust emissions.
The distinguishing features of diesel combustion are spontaneous ignition (resulting from the high air temperature and pressure in the cylinder when injection commences) and the fact that fuel-air mixing controls the burning rate [
7]. Diesel exhaust emissions depend greatly on fuel and lubricating oil properties, engine parameters and exhaust after-treatment technology. Driving and environmental conditions also influence emissions [
8]. Cylinder gases’ temperature, pressure, density and composition as well as injection timing and injector type, have a direct effect on combustion and emission formation for a given fuel [
9]. In-cylinder pressure has a fundamental role in combustion characteristic analyses and it is needed to control combustion-related parameters. Primarily, the peak cylinder pressure depends on the burned fuel fraction during the premixed combustion phase. A large amount of fuel burned in premixed combustion corresponds to a high peak cylinder pressure. [
10]
The delay period between the start of injection and start of combustion must be kept short, because injection timing is used to control combustion timing. A short delay is also needed to hold the maximum cylinder gas pressure below the maximum the engine can tolerate. [
7] Therefore, the main ignition characteristic of a diesel fuel—its cetane number (CN) must be above a certain value.
This study focused on how the use of various alternative fuels affects combustion, especially in-cylinder combustion. A baseline fuel and six alternative liquid fuels were investigated in a high-speed compression-ignition (CI) engine to understand their combustion properties. The fuels were light fuel oil (LFO, baseline), marine gas oil (MGO), kerosene, rapeseed methyl ester (RME), renewable diesel (HVO), renewable wood-based naphtha and its blend with LFO.
One potential substitute for conventional fuels is mineral origin alternative oils, such as recycled waste oils. Recycling waste lubricant oils (WLO) into diesel-like fuels is an environmentally friendly solution because waste oils are classified as a hazardous waste [
11,
12]. In this study, the first alternative was marine gas oil (MGO), produced from waste lubricating oils. MGO is commonly known as a shipping fuel, and currently it is largely used when a vessel is inside an emission control area (ECA) or within EU ports. Sulphur emission regulations are becoming increasingly stringent and, in the case of older ships, the use of MGO is suggested to be more beneficial than retrofitting scrubbers [
13]. Recently, Wang and Ni [
14] and Gabiña et al. [
4] have also investigated diesel-like fuel (DLF) manufactured from WLO.
Kerosene is primarily used in the aviation sector. However, the U.S. Army and the North Atlantic Treaty Organization (NATO) have a single military fuel policy—mandating the use of kerosene-based jet fuels in ground vehicles equipped with CI engines, thereby simplifying supply chain logistics [
15]. Kerosene has been studied for its pollutants reduction potential. Amara et al. [
16] presents studies where kerosene’s higher volatility lowered NO
X emissions compared to diesel in a CI engine. Hissa et al. [
17] also investigated neat kerosene in a combustion research unit (CRU) and concluded that other fuel may be needed for starting and stopping the engine, or different injection nozzles used. Kerosene’s low lubricity may also cause a malfunction in the injection system. Nevertheless, kerosene was considered an interesting alternative for future special CI engine applications.
First generation biodiesels, known as FAME or fatty acid methyl esters, have been studied for a long time. Despite some problems, they still arouse great interest [
6]. In Europe, the maximum FAME content in conventional diesel fuel is 7 V-% according to EN590:2013. However, in the United States, the renewable share is in some cases above 20 V-%, and even their own standard EN 16709 for B20 and B30 fuels. [
6] Biodiesel fuels’ different types of properties have a strong relation with their fatty acid composition. Poor cold properties and poor storage stability are the main problems associated with the use of FAME fuels [
6]. The review article of Shahabuddin et al. [
18] reported that the combustion characteristics of a bio-fuelled engine are slightly different from the engine running with petroleum diesel. This was one of the main reasons also to include rapeseed methyl ester (RME) in the present study.
Hydrotreated vegetable oils (HVO) are one of the easiest ways to increase the bio-component in diesel fuel. Moreover, they are highly recommended by vehicle and engine manufacturers [
19]. The production process of hydrotreated fuels differs from FAME and this neat renewable diesel can be used in diesel engines without blending. HVO is sulphur and aromatic-free fuel which can be produced from various wastes such as animal fats, residues of forest industry and wastes from the food industry. HVO offers some advantages compared to FAME (e.g., RME), including a higher cetane number (CN), better storage stability and fewer problems with cold operability and deposits [
8]. HVO’s properties correspond to those of traditional fossil fuels but a significant reduction of up to 80% in greenhouse gaseous emissions is achieved [
5]. Furthermore, crude tall oil-based (CTO) fuels do not compete with the food chain and there is no direct land-use change [
5]. A Finnish forestry company manufactured the CTO renewable diesel used in the study.
Renewable, wood-based naphtha is another interesting and novel fuel. This study’s naphtha was made from CTO extracted in the pulp production process. This colourless, sulphur-free, paraffinic product is chemically pure hydrocarbon and can be used as a bio-component in fossil gasoline [
20]. Conventionally, naphtha is produced from crude oil and is suitable for use in compression-ignition (CI) engines [
14,
21,
22]. However, raw naphtha has a low cetane number (CN) which may cause prolonged ignition delay (ID) and hence may retard the start of combustion [
23]. Hissa et al. [
17] observed that ignition of neat wood-based naphtha had a retarded start of combustion and prolonged ID when measured in a CRU. This may limit use of neat naphtha as fuel in diesel engines. Neat naphtha also needs other fuel (e.g., LFO) for starting and stopping the engine. However, there is very little research into wood-based naphtha, prompting the inclusion of this fuel in the current study as an interesting renewable option, especially for blending.
Various liquid fuel options are needed in the near future for flexible power generation, marine and heavy-duty applications. Consequently, this study was carried out to evaluate the effects of alternative fuels’ properties in an engine use and to promote the development of fuel processes and standard to meet engine requirements. The focus was on how various properties of the alternative fuels affects combustion, especially in-cylinder combustion. The cylinder pressures, heat release rates (HRR), mass fractions burned (MFB) and combustion duration were determined at an intermediate speed at three loads. The experimental engine was a high-speed common-rail diesel engine intended for non-road applications. All measurements were performed under steady operation conditions without engine modifications. The study continued with the emission analysis in the article by Ovaska et al. [
24].
2. Materials and Methods
The engine experiments were conducted by the University of Vaasa (UV) at the Internal Combustion Engine (ICE) laboratory of the Technobothnia Research Centre in Vaasa, Finland.
2.1. Fuels
The studied fuels were selected with the aim of increasing the choice of fuel alternatives for non-road compression-ignition engines. The baseline fuel was a commercial low-sulphur light fuel oil (LFO) from Teboil, Finland. Jet A-1 aviation fuel (100% kerosene) was produced by Neste, Finland. Marine gas oil (MGO) produced from waste lubricant oils, was a product of STR Tecoil, Finland. Hydrotreated vegetable oil (HVO), also known as CTO-based renewable diesel, was produced by UPM, Finland. Rapeseed methyl ester (RME) was a product of Analytik-Service Gesellschaft mbH, Germany. Wood-based naphtha was made from CTO, extracted in the pulp production process, and was produced by UPM, Finland. Renewable naphtha was also blended with LFO (20 V-% naphtha and 80 V-% LFO), also known as naphtha20. Chemically, all the fuels contained several hydrocarbon compounds, therefore simple chemical formulas could not be given.
Table 1 summarises the specifications of the studied fuels.
Cetane number (CN) is an indicator of the ignition quality of diesel fuel [
7]. A high CN indicates good ignitability in a diesel engine. As shown in
Table 1, the CN of the fuels ranged from 34 (neat naphtha) to 68 (recycled MGO). The CN of kerosene, naphtha20, LFO, RME and HVO were 41, 51, 52, 53 and 65, respectively.
There is an indirect relationship between density and other fuel parameters such as CN, viscosity, volatility and distillation characteristics [
25]. It has also been shown that fuel density can have a direct effect on the progress of fuel pressure in the injection system and a consequent effect on the dynamic start of fuel injection [
25]. Neat naphtha had the lowest density (722 kg m
−3), markedly lower than that of the baseline LFO (827 kg m
−3). Lower density may result in lower engine power and also affects volumetric fuel consumption. The highest density was measured for RME (883 kg m
−3).
A higher mass-basis lower heating value (LHV) may result in a higher heat input to the engine, i.e., higher cylinder pressures and increased power output. The LHV of most of the studied fuels was 43–44 MJ kg−1. The one exception was RME, with a LHV of 37 to 38 MJ kg−1.
Legislation has driven down fuel’s sulphur content. A lower, new sulphur limit (5,000 mg kg
−1) will be implemented in the maritime sector on 1 January 2020 [
26]. In Finland, non-road diesel engine fuels must comply with EN 590:2013 standard, setting a maximum limit for sulphur content of 10 mg kg
−1 [
27]. Regarding the studied fuels, kerosene (1000 mg kg
−1) and MGO (<100 mg kg
−1) had sulphur exceeding EN 590:2013’s limit.
Diesel fuel injection systems are lubricated mainly by the fuel itself. However, when sulphur level of a fuel is reduced, the process also destroys some of the fuel’s natural lubricants [
28]. To avoid wear in the fuel injection system, a minor portion of lubrication additive (1:4000) was added to kerosene to avoid possible malfunctions in engine use. This amount of lubrication additive should not have any effect on the autoignition properties of the fuel. Also, MGO had lower lubricity than LFO. This might be shown as injection system failures over the time. In addition, longer engine measurements are required to examine the effect of the lubricity.
Safety issues are also critical when introducing new fuels. Fuel suppliers have to specify the fuel properties and confirm compliance with industry standards. These properties include flashpoint, combustibility, stability, compatibility, viscosity, lubricity, etc. Every one of these properties, if not properly addressed, can affect equipment performance and/or reliability and above all, affect safety of personnel or the engine’s safe operation [
31]. In the present study, the low flash points of neat naphtha (20 °C) and kerosene (38 °C), and their low viscosity (naphtha 0.50 mm
2 s
−1 and kerosene 0.94 mm
2 s
−1) necessitated additional safety procedures when handling the fuels and making measurements in the engine. For the studied fuels, the compatibility of the engine, fuel system and auxiliaries was ensured. Additionally, the engine operators used respirator masks and gloves during the measurements.
2.2. Engine
The test engine, an AGCO Power 44 AWI, was a high-speed four-cylinder diesel engine for non-road applications. It was turbocharged, intercooled and had Bosch common-rail fuel-injection. It was loaded by means of a Horiba eddy-current dynamometer WT300. The engine’s specification is in
Table 2.
2.3. Analytical Instruments
LabVIEW system-design software was used to collect the sensor data from the engine. The recorded variables were engine speed and torque, cylinder pressure and injection timing, duration and quantity. A WinEEM4 program provided by the engine manufacturer, AGCO Power, controlled fuel injection according to load-speed requests. The basic settings of WinEEM4 were the same for all fuels. A schematic of the setup of the test bench is in
Figure 1.
A piezoelectric Kistler 6125C pressure sensor was used to measure in-cylinder pressure. The sensor was mounted on the head of the fourth cylinder. A charge amplifier filtered and amplified the signal, which was then transmitted to a Kistler KIBOX combustion analyser. The crankshaft position was recorded by a crank-angle encoder (Kistler 2614B1), which can output a crank-angle signal with a resolution of 0.1 °CA by means of an optical sensor. The cylinder-pressure data was averaged over 100 consecutive cycles to smooth irregular combustion. The averaged data was used to calculate HRR.
The HRR and MFB were calculated via AVL Concerto’s data-processing platform, using the Thermodynamics2 macro. The macro used a calculation resolution of 0.2 °CA. The start of the calculation was set at −30 °CA. The data were filtered with the DigitalFilter macro and a frequency of 2500 Hz. For the HRR results, the average values of in-cylinder pressure were calculated first. Thereafter, the macro was used to calculate HRR values. Finally, the HRR curve was filtered. In contrast, for the MFB results, pressure values were first filtered and then the macro was used. The average values of 100 cycles were not used for MFB results, establishing the standard deviations.
2.4. Experimental Matrix and Measurement Procedure
All measurements were performed under steady operation conditions without engine modifications. The measurements were conducted based on ISO 8178-4 standard, known as non-road steady cycle (NRSC). Additionally, the 25% (3.2 bar) load point was measured at an intermediate speed. With low-viscosity kerosene and neat naphtha, the default engine control parameters made running the engine possible only at an intermediate speed. The added 25% load point gave more information during the experiment, because no engine parameter optimization was applied. Multistage injection (pilot, main and post injections) was used throughout the study. The operating conditions chosen were an engine speed of 1500 min
−1 and three different loads—Brake-mean effective pressures (BMEP) of 3.2 bar, 6.4 bar and 9.7 bar. The loads are in
Table 3.
2.5. Injection Strategy
Pilot injection shortens the ID of a fuel by increasing in-cylinder temperatures for main injections. Post injections, in turn, are used to reduce particulate and soot emissions, primarily at lighter loads and lower engine speeds [
7].
For all the fuels at all loads, the pilot injection was constant. It started at 6.8 ± 0.1 °CA BTDC and ended at 4.1 °CA BTDC. The main injection started at exactly 0.9 °CA BTDC. The duration of the main injection varied not only with the load, but also according to the fuel’s heating value. The main injection was always longest with RME (5.2 to 8.2 °CA depending on the load), due to its lower LHV. The main injection durations for the other fuels were very similar (4.8−7.7 °CA according to engine load). The start of post injection was linked to the end of main injection, and was thus delayed according to the fuel’s LHV. The duration of post injection varied only slightly in the range of 3.3 °CA and 3.6 °CA and was independent of the load. The pilot and main injections are the most important factors when considering ignition delay, that is, HRR and MFB values.
The engine was started with LFO in all fuel tests. Then, fuel was changed to test fuel and the engine was warmed-up and the load was applied. The intake-air temperature was adjusted to 100 °C downstream of the charge-air cooler to support auto-ignition of the fuels at each load. The temperature was controlled manually by regulating the flow of cooling water to the heat exchanger. The valve setting was kept constant. Therefore, the charge temperature changed with the load. All measurements were taken only after the engine had stabilized, as determined by stability of the temperature of coolant water, intake air and exhaust.
4. Conclusions
The present study was carried out to evaluate the effects of the properties of alternative fuels in engine use, especially in-cylinder combustion, and to promote the development of fuel processes and standards to meet engine requirements. The results are useful to understand current trends in the fuel market and the impact that alternative fuels have in CI engine combustion.
Baseline fuel and six alternative liquid fuels were investigated in a high-speed diesel engine for non-road applications. The fuels were light fuel oil (LFO, baseline), recycled, waste lubricating oil (WLO), origin marine gas oil (MGO), kerosene, rapeseed methyl ester (RME), renewable diesel from crude tall oil (HVO), renewable wood-based naphtha and its blend with LFO (naphtha20). All measurements were performed under steady operation conditions without engine modifications. Multistage injections (pilot, main and post injections) were used throughout and the engine-speed of 1500 min−1 was maintained while conducting studies at three different engine-loads with brake mean effective pressures (BMEP) of 3.2 bar, 6.4 bar and 9.7 bar.
The investigated combustion parameters were very similar with all studied fuels. WLO-based MGO showed good combustion performance due to its high CN and short ID. The highest peak of cylinder pressure was measured for MGO, but the overall differences in cylinder pressures between fuels were minor. MGO’s high CN and short ID meant the combustion started slightly earlier, but it also ended slightly later, giving a longer CD at the low and medium-load points. MGO’s high viscosity may increase the combustion duration by hindering mixture formation. However, at the highest load, MGO’s CD decreased relative to the other studied fuels. Similarly, MGO’s combustion performance relative to the other fuels improved when the engine load was increased. However, high viscosity and its high sulphur-content are limiting factors for the use of MGO.
Naphtha’s low CN increased the ID, but neat naphtha burned rapidly due to its low viscosity and density and the high volatility of its lighter compounds, which improved early combustion. Compared to LFO, naphtha’s CD shortened as engine load decreased. Naphtha was level with MGO and kerosene in showing the highest cylinder pressures at all loads. Naphtha’s HRR curve was slightly ahead of the other fuels, maybe because its low flash point and low viscosity promoted good air/fuel mixing. Neat naphtha’s prolonged ID detracts from its combustion performance and limits its use as a drop-in fuel in a diesel engine. Moreover, neat naphtha needs other fuel (e.g. LFO) for starting and stopping the engine.
Naphtha20 was a blend of LFO and CTO-based naphtha—it showed overall combustion performance comparable to LFO’s. That was also the case for CTO-based diesel, HVO. Both naphtha20 and HVO may be used in a diesel engine without any modifications.
Kerosene’s light fractions not only increased its HRRs and cylinder pressures, but also decreased combustion duration compared to LFO. Nevertheless, kerosene’s high sulphur-content limits its use in CI engines.
MGO with RME were favourable in terms of CO and HC emissions while the lowest NOX emissions were recorded with HVO. Smoke emission was negligible for all fuels. However, the emission results of neat naphtha or kerosene were not included.
All investigated alternative fuels can be used in a non-road engine without any modifications, except neat renewable naphtha. Despite the improved premixed combustion due to low viscosity and high-volatile, light compounds of naphtha, the ignition delay was prolonged and the engine load was limited. Moreover, naphtha had poor auto-ignition properties and required LFO for starting and stopping the engine. However, as a blend with LFO, renewable naphtha is suitable for CI engines if safety issues associated with its low flash point are solved.