Experimental study of a kerosene fuelled internal combustion engine

Experimental study of a kerosene fuelled internal combustion engine Tomás Formosinho Sanchez Instituto Superior Técnico, Technical University of Lisbon, Av. Rovisco Pais, 149-1 Lisboa, Portugal; Email: tomasformozinho@gmail.com Abstract Mistral Engines, a Swiss Aeronautical Company wishes to be able to have operating Wankel engines on kerosene fuel because of its market availability. Kerosene is a fuel with lower cetane number than diesel fuel, thus it should give a longer ignition delay. This makes it viable for lower emissions since the longer ignition delay means longer time for the fuel to mix with the in-cylinder gas prior to combustion onset. With the cooperation of the Industrial Energy Systems Laboratory of the Swiss Federal Institute of Technology in Lausanne, a study on the combustion of kerosene in an internal combustion engine (ICE) is here presented. This study is performed in a mono cylinder ICE with variable Compression Ratio (CR) and ariable alve Timing (T). Different strategies, such as Homogeneous Charge Compression Ignition (HCCI), Homogeneous Charge Spark Ignition (HC), Controlled Auto Ignition (CAI) and a combustion pre-chamber concept are considered. 1. Introduction The transportation sector, including aviation, an essential part of our modern society, represents the largest part of the petroleum based fuels consumption. Its importance has been growing continuously during the last decades. The aviation sector normally uses kerosene based fuels in jet engines. The application of this type of fuel in internal combustion engines (Wankel engine and reciprocating piston engines) is a big challenge as a result of their lower fuel tolerance than jet engines. Nevertheless, Mistral Company wants to adapt its Wankel engines to that fuel since the market availability is bigger than the used 1LL avgas gasoline. Therefore, a study on the kerosene combustion in a single cylinder engine with variable CR and T is undertaken. The engine is naturally aspirated and the kerosene is direct injected into the combustion chamber. To understand the behaviour of kerosene combustion in an ICE several parameters were investigated. The influence of the fuel temperature, the timing of the pre-injection and the richness of the mixture. Furthermore, and CI were investigated and HCCI was attempted with a new concept of combustion pre-chamber. Controlled Auto Ignition (CAI) and HCCI combustion are radically different from the conventional combustion in a gasoline engine and CI diffusion combustion in a diesel engine. The combination of diluted and premixed fuel and air mixture with multiple ignition sites throughout the combustion chamber eliminates the high combustion temperature zones and prevents the production of soot particles, hence producing ultra low NO X and particulate emissions. The use of lean, or more often diluted, air-fuel mixture with recycled burned gases permits unthrottled operation of a CAI/HCCI 1

gasoline engine, thus yielding higher engine efficiency and better fuel economy than combustion. Therefore, CAI/HCCI combustion represents for the first time a combustion technology that can simultaneously reduce both and NO X particulate emissions from a diesel engine and has the capability of achieving simultaneous reduction in fuel consumption and NO X emissions from a gasoline engine. Based on these promises, an alternative solution is intended to auto ignite the mixture in several points as in HCCI combustion. An auto ignition pre-chamber jet ignition is here studied and presented. A glow plug inside the pre-chamber should ignite virtually simultaneously an important air-fuel mixture that will form jets flowing out of the nozzles igniting, in several sites, the remaining of the air-fuel in the main chamber. The jet ignition will develop significantly faster than the spark ignition type flame front propagation. 2. Combustion pre-chamber Occasionally, in CI engines, fuel is injected into a pre-chamber where a Glow Plug helps to ignite the mixture on cold starts. However, a new concept was developed where fuel is injected into the main combustion chamber. A pre-chamber is located where the spark plug was. The air-fuel mixture is forced into the prechamber through six nozzles at high velocity during the compression stroke. Fuel is injected into a very turbulent flow which leads to desired high mixing rates. The rich mixture with very high swirl in the pre-chamber will ignite readily and combust very quickly, finally igniting the main chamber gases.[1] Flame jets are intended to burn quickly the whole mixture in a sort of HCCI combustion. Bosch glow plug Combustion pre-chamber a) Cylinder head and pre-chamber b) Combustion pre-chamber concept Figure 1 Cylinder head and combustion pre-chamber images Figure 1 a) shows the images of the pre-chamber designed for this engine. The Bosch glow plug (yellow) selected was the GPM92 from a Bosch catalogue. The pre-chamber concept is shown in Figure 1 b). The nozzle orifices can be observed as well. Given that the glow plug reaches 1ºC of temperature, the pre-chamber was made of heat resistant austenitic stainless steel X15CrNiSi25-2 (Böhler H525). 2

Consequently, the erosion process of the nozzle orifices is also slowed. The real pictures of the pre-chamber are shown in Figure 2. a) Bosch glow plug and pre-chamber b) side view of pre-chamber Figure 2 Real photos of pre-chamber and glow plug 3. Lean combustion mode and HCCI The use HCCI combustion in internal combustion engines is of interest because it has the potential to produce low NOx and Particle Mater (PM) emissions while providing diesel-like efficiency. In HCCI combustion, a premixed charge of fuel and air auto-ignites at multiple points in the cylinder near TDC, resulting in rapid combustion with very little flame propagation. In order to prevent excessive knocking during HCCI combustion, it must take place in a dilute environment, resulting from either operating fuel lean or providing high levels of either internal or external Exhaust Gas Recirculation (EGR). Operating the engine in a dilute environment can substantially reduce the pumping losses, thus providing the main efficiency advantage compared to engines.[2] Lean combustion burning is a solution that allows relatively high compression ratios combined with significantly NOX emissions without after-treatment. HCCI is a form of internal combustion in which well mixed fuel and oxidizer are compressed to the point of auto-ignition. It combines characteristics from engines (HC: homogeneous charge spark ignition) and CI (SCCI: stratified charge compression ignition) engines. The pressure and temperature of the mixture are raised by compression until the entire mixture reaches the point of auto ignition.[1] 4. Results and Discussion 4.1. Pre-Injection timing variation Several measures were taken in order to study the effects of the position/timing variation of the pre-injection on the combustion process in the engine. This test consists of varying the timing of the pre-injection and keeping all the other parameters constant. Therefore, earlier injections should give a longer mixing period of time conducting to more homogeneous mixtures. If properly chosen, sufficient time is available between the end of injection and the start of ignition, assuring a relatively good homogeneous mixture, which will result in fully premixed combustion. The use of direct injection compared to port injection should be an advantage since the fuel is injected during the compression stroke, the gas temperature and density are higher than at intake conditions, enhancing the vaporization process and thus reducing the time to prepare the mixture and/or avoiding the need to heat up the intake air. 3

The test was performed at 15rpm and the duration of injection was.4ms for the pre-injection and.85ms for the post-injection. Thus the fuel quantity injected was 7.66x1-5 kg per cycle. Although the fuel injected per cycle was constant, the lambda sensor revealed values for λ within [.96; 1.3]. The pre-injection timings were the following: -36º, -63º,-9º,-117º, and -144ºCA BTDC. The postinjection timing was kept at a constant -23ºCA BTDC. There was no spark ignition. Therefore the ignition mode was auto-ignition. As a result, homogeneous compression ignition combustion should be achieved without the adaptation of fuel supply in the intake port. Direct injection was made at 95 bars of pressure with a CR of 13. 6 35 Pressure[bar] 5 4 3 2 1-1 -5 5 1 15 2 θ=-36º θ=-63º θ=-9º θ=-117º θ=-144º Instantaneous Heat Release [J] 3 25 2 15 1 5-1 -5 5 1 15 2-5 θ=-36º θ=-63º θ=-9º θ=-117º θ=-144º a) Pressure vs crank angle b) IHR vs crank angle Figure 3 Pressure and IHR as function of crank angle for pre-injection timing variation Figure 3 a) shows the evolution of the pressure as function of crank angle. There is a clear tendency to shift the pressure curve at the vicinity of TDC as function of the timing of injection. The later the pre-injection, the closer the curve is to TDC. The same can be observed in Figure 3 b), the Instantaneous Heat Release (IHR) shifts the same way as the pressure curve. Furthermore, it is noticeable that too late pre-injections have oscillating behaviour at the end of High Temperature Heat Release (HTHR). At pre-injection timing of θ = 36º big oscillations of IHR are visible, especially between 3 to 5ºCA (at the end of HTHR). These oscillations can be a sign of fuel stratification, small clouds of rich mixture that could auto-ignite locally or knock (as some authors suggest [3]) and consequently some peaks on the IHR are found. This is visible when Figure 3 a) and Figure 4 b) are observed. They both show high oscillations between 3 to 5ºCA with a pre-injection timing of θ = 36º. As a matter of fact, during the experiments, the later the pre-injection, the heavier the sound of the engine was perceptible. Concerning the Cumulative Heat Release (CHR), Figure 4 a) shows the specific performance of each of the tests carried out. It can be observed that too late pre-injection ( θ = 36º ) noticeably affects the CHR, shifting the curve to the vicinity of the TDC and delaying the combustion process. There seems to be an ideal pre-injection timing at θ = 9º where the cumulative heat release is delivered more rapidly than the other pre-injection timings. 4

1 18 9 16 Cumulative Heat Release [%] 8 7 6 5 4 3 2 θ=-36º θ=-63º θ=-9º θ=-117º θ=-144º Cylinder Gas Temperature [K] 14 12 1 8 6 4 θ=-36º θ=-63º θ=-9º θ=-117º θ=-144º 1 2-2 -1 1 2 3 4 5 6-2 -1 1 2 3 4 5 6 a) CHR vs crank angle b) Cylinder gas temperature vs crank angle Figure 4 CHR and Cylinder gas temperature as function of crank angle for pre-injection timing variation Figure 4 b) shows the cylinder gas temperature profile of each pre-injection timing strategy. There is a clear tendency to shift up the gas temperature as the pre-injection delay increases. Thus, preinjection strategies closer to TDC appear to have gas temperatures higher than earlier pre-injection. Consequently, later pre-injections have hotter exhaust temperatures as well. It is also possible to observe temperature oscillations in the later pre-injection. As said before, this can occur due to fuel stratification that can locally ignite after the main combustion. 3.5 3 2.5 2 1.5 1.5-16 -14-12 -1-8 -6-4 -2 Q5%[CAD] 6 58 56 54 52 5 48 46 44 42 4-16 -14-12 -1-8 -6-4 -2 Combustion Duration [CAD] Pre-Injection [degrees before TDC] Pre-Injection [degrees before TDC] a) Q5% vs pre-injection timing b) Combustion duration vs pre-injection timing Figure 5 Q5% and Combustion duration as function of crank angle for pre-injection timing variation In Figure 5 a) the curve shifting observed in Figure 3 is emphasised. In fact, it shows the position of 5% of the CHR as function of the various pre-injection strategies. Again, a tendency can be observed. As the pre-injection comes later, the 5% of CHR approximates the TDC. Figure 5 b) shows the combustion duration in CAD as function of the various pre-injection strategies. It can be noticed that the combustion duration increases with the increase of the preinjection delay. Moreover, an optimal strategy seems to occur, where the combustion duration is the fastest. That strategy is, when the pre-injection is done at θ = 9º. 4.2. Excess of air ratio λ variation with pre-chamber, and CI comparison The test was performed at 15rpm and the timing of the injections was -126ºCA for the preinjection and -26ºCA for the post-injection. As a result, a very homogeneous mixture is intended to be achieved. There was no spark ignition; therefore, there is no direct control on start of combustion. The Bosch Glow Plug shown in Figure 2 a) was constantly on. Hence, HCCI should be achieved with the help of jet flames that leave the pre-chamber. The duration of injection was.4ms for the pre-injection 5

and [1.9; 1.63; 1.6;.71;.58;.52;.48] ms for the post-injection. Thus the fuel quantity injected was [14.9, 12.44; 8.95; 6.8; 6.1; 5.64; 5.39] x1-5 kg per cycle. The lambda sensor revealed the following values: λ = [.65;.7;.85; 1.; 1.15; 1.3; 1.45]. Direct injection was made at 95 bars of pressure with a CR of 13. 5 5 Pressure[bar] 4 3 2 λ=.65 λ=.7 λ=.85 λ=1. λ=1.15 λ=1.3 Pressure[bar] 4 3 2 λ=.7 λ=.85 λ=1. λ=1.15 λ=1.3 1 1-6 -4-2 2 4 6-6 -4-2 2 4 6 a) Pressure vs crank angle (Pre-Chamber) b) Pressure vs crank angle () Figure 6 Pressure (Pre-chamber and ) as function of crank angle for various λ Figure 6 compares the cylinder pressure for both ignition methods. As mentioned, the prechamber tremendously affects the start of combustion. There is a clear propensity that the prechamber delays the start of combustion (see also Figure 8 a)). The maximum pressure peak of the pre-chamber always appears for higher crank angle degrees than in tests. It is also noticeable that the peak of pressure in the pre-chamber tests is always lower than the corresponding ones. Since the start of combustion is delayed, the piston is already in the expansion stroke on its way to BDC; therefore lower peaks of pressure during combustion are expected. If the pressure curves are shifted to the right and have lower peaks of pressure in the pre-chamber tests, thus all the other curves (cylinder gas temperature, IHR and CHR) should behave similarly as well. In fact delaying the start of combustion has an important effect on the IHR curve as well; this can be shown in Figure 7. Instantaneous Heat Release [J] 3 25 2 15 1 5-5 5 1 15 2 25-5 λ=.65 λ=.7 λ=.85 λ=1. λ=1.15 λ=1.3 Cumulative Heat Release [%] 1 9 8 7 6 5 4 3 2 1-2 -1 1 2 3 4 5 6 λ=.65 λ=.7 λ=.85 λ=1. λ=1.15 λ=1.3 a) IHR vs crank angle (Pre-chamber) b) CHR vs crank angle (Pre-chamber) Figure 7 IHR and CHR (Pre-chamber) as function of crank angle for various λ Figure 8 compares 5% of CHR (Q5%) and the combustion duration between the pre-chamber, and CI tests. As said before, Figure 8 a) shows that there is a clear propensity of delaying the start of combustion when the pre-chamber is used. 6

In Figure 8 b) it is clear that the pre-chamber severely affects the combustion duration. When used it can burn the fuel faster then flame propagation on mode (1ºCA faster, maximum). Thus, the flame jets flowing out of the pre-chamber should burn better and quicker than the mixture in the cylinder. If the combustion duration is reduced, its efficiency should be increased as predicted in HCCI theory. 1 6 Q5%[CAD] 8 6 CI with Pre-Chamber 4 CI 2.6.7.8.9 1. 1.1 1.2 1.3 1.4 1.5-2 -4 Excess of air ratio λ Combustion Duration [CAD] 5 4 CI with Pre-Chamber 3 CI 2 1.6.7.8.9 1. 1.1 1.2 1.3 1.4 1.5 Excess of air ratio λ a) Q5% vs λ b) Combustion duration vs λ Figure 8 Q5% and Combustion duration (Pre-chamber, and CI) as function of λ Figure 9 a) proves that combustion efficiency is in fact higher when the pre-chamber is used. Combustion Efficiency/Combustion Efficiency(Pre-chamber) 1.2 1..98.96.94.92 CI with Pre-Chamber CI Thermodynamic Efficiency [%].54.53.52.51.5.49.48 CI with Pre-Chamber CI.9.6.7.8.9 1. 1.1 1.2 1.3 1.4 1.5 Excess of air ratio λ.47.6.7.8.9 1. 1.1 1.2 1.3 1.4 1.5 Excess of air ratio λ a) Combustion Efficiency vs λ b) Thermodynamic Efficiency vs λ Figure 9 Combustion and Thermodynamic efficiencies (Pre-chamber, and CI) as function of λ Although the combustion efficiency is higher to the pre-chamber tests, the thermodynamic efficiency (see Figure 9 b)) presents the inverse results. This is due to the definition of the thermodynamic efficiency. In fact, the IMEP gross is severely affected by the big delay of start of combustion. In Figure 1 it is possible to observe that the CHR (QMEP) is higher, for leaner and richer mixture, in the pre-chamber tests than in mode. Comparing with CI mode, pre-chamber present always higher QMEP values. This also helps to conclude that with pre-chamber the fuel burns better, delivering more heat, thus more work is produced and consequently more power. 7

12 QMEP [bar] 11 1 9 8 CI with Pre-Chamber CI 7 6.6.7.8.9 1. 1.1 1.2 1.3 1.4 1.5 Excess of air ratio λ Figure 1 QMEP (Pre-chamber, and CI) as function of λ The combustion and thermodynamics efficiencies should give an idea how kerosene behaves inside the combustion chamber in different conditions. Some conclusions are now outlined. 5. Conclusions To understand the behaviour of kerosene combustion in an ICE several parameters were investigated: influence of fuel temperature, timing of the pre-injection and richness of the mixture. Furthermore, and CI were investigated and HCCI was attempted with a new concept of combustion pre-chamber. The experiments allow to drawn the following conclusions. 1. Start of combustion is sensitive to excess of air ratio in and CI mode, presenting a large delay when the mixture is very rich (λ<.7). 2. Engine start is difficult on CI mode when engine is cold. 3. Better efficiencies are obtained for leaner mixtures, presenting a maximum when λ=1.15. 4. Too late pre-injections lead to IHR instability at the end of HTHR. Fuel stratification may take place and auto-ignition of small clouds of rich mixture happens. 5. Curves shifting are similar with variation of λ for both and pre-chamber mode. 6. The pre-chamber leads to larger ignition delay. In fact, the ignition is not controlled as in mode. 7. The use of pre-chamber leads to overall better efficiencies (combustion and indicated), thus better engine performance. 8. The use pre-chamber and HCCI approach leads to improved combustion stability. 9. The long ignition delay, which can partly lead to low soot emissions, could be beneficial for expanding the HCCI operation area. In combination with the market availability of kerosene, it has the possibility to be a suitable HCCI fuel. 6. Perspectives In this preliminary study, the pre-chamber results seem to be encouraging since overall better efficiencies were achieved when compared with mode. However, the ignition delay seems to be very high. For this reason, even better results are expected if the ignition delay could be properly controlled. 8

A comparison between CI and CI with pre-chamber should complete the final conclusions. It should also help to conclude if there is a positive influence in the use of the combustion pre-chamber when compared with normal CI and mode. Therefore, further tests should be carried out with and without pre-chamber using CI mode. EGR can be used to increase the mixture dilution and augment its temperature so that the ignition delay can be reduced. Thus, a Controlled Auto Ignition (CAI) of the HCCI mixture should be achieved. Air assist DI combustion systems can be used for better fuel vaporization. This methodology can also reduce the poor start problems once it enhances homogenization. Pollutants measurement should be executed once the aviation market will have its own restrictions as in automobile market. As the present work intends to study the kerosene combustion behaviour inside of an ICE for aviation applications, the air admission pressure and its temperature should be fitted to the altitude conditions. Since Mistral Company intends to use Wankel engines, the results presented in this paper should not be conclusive. Tests on a Wankel engine must be performed so that the study can be validated on the application field. Formulae Instantaneous heat release (IHR) dq 1 d dp = γσ p( θ) + ( θ) dθ γ 1 dθ dθ Indicated Mean Effective Pressure (IMEP) IMEP p d p d 36 net = = = 18 36 Wi p d IMEP 18 gross = Fuel Mean Effective Pressure (FuelMEP) The fuel mean effective pressure is defined as: where σ m FuelMEP = f LH Heat Release Mean Effective Pressure (QMEP) Q QMEP= η th = η comb = m [kg] is the mass of fuel supplied per f cycle, LH [J/kg] is the lower heating value for the fuel and [m 3 ] is the displacement of the engine. Thermodynamic Efficiency IMEP gross QMEP Combustion Efficiency QMEP FuelMEP 9 Symbols and acronyms After bottom dead ABDC centre ATDC After top dead centre Before bottom dead BBDC centre BDC Bottom dead centre Break mean effective BMEP pressure [bar] BTDC Before top dead centre CA or CAD Craft shaft angle [º] CAI Controlled auto igniton CI Compression ignition CR Compression ratio DI Direct Injection ECU Engine unit control Exhaust gas EGR recirculation EC Exhaust valve closure EO Exhaust valve opening Fuel mean effective FuelMEP pressure [bar] Homogeneous Charge HCCI Compression Ignition High temperature heat HTHR release Instantaneous heat IHR release [J] Indicated mean effective IMEP pressure [bar] IC Intake valve closure

IO LH N NO X P Q QMEP RPM TDC UA T γ γ σ Intake valve opening Low heat value [MJ/kg] Engine speed [RPM, RPS, Hz] Oxides of nitrogen Cylinder pressure [bar] Cylinder energy release [J,kJ] Heat release mean effective pressure Rotation per minute [rpm] Spark ignition Top dead centre Unmanned Aeronautical ehicles Total cylinder volume [cm 3 ] ariable valve timing Ratio between specific heats, adiabatic polytrophic index Evaluated polytrophic index η comb η Combustion efficiency gas Gas exchange efficiency Thermodynamic η th efficiency θ Crank shaft angle [º] Air-to-fuel ratio or λ excess of air ratio References [1] MEIER, M : Etude expérimentale de l auto-inflammation d une préchambre de moteur monocylindre, EPFL Master Thesis 26. [2] SZYBIST, J, & BUNTING, B: Chemistry Impacts in Gasoline HCCI, Oak Ridge National Laboratory, 26. [3] ZHAO, H: HCCI and CAI engines for the automotive industry, Woodhead Publishing Limited, Cambridge, 27. [4] TÄSCHLER, C: Caractérisation expérimentale de l inflammation d un spray de kérosène, dans les conditions d un moteur, EPFL Master Thesis 28. 1