The Effect of Inlet Valve Timing and Engine Speed on Dual Fuel NG-Diesel Combustion in a Large Bore Engine.
The bivalent engine operation with natural gas and diesel fuel moved again in the focus of engine research. The main factor behind this development is the fuel flexibility to reduce the operating costs by choosing the fuel with the lowest available price. Furthermore, the gas operation offers the shipping sector the possibility to fulfil the IMO Tier 3 emission regulations without exhaust gas aftertreatment . This engine concept usually ignites the natural gas by a small amount of the diesel fuel. A combustion process using simultaneously two entirely different fuels is called dual fuel combustion. Thereby, the term dual fuel should not be confused with an engine operation switching between two different fuels. These are better described as bi-fuel or bivalent engines . Furthermore, the term bivalent dual fuel engine is not a contradiction. It describes an engine concept that can switch between two combustion processes with different main fuels and that, in addition, realizes one operation mode in a dual fuel combustion process.
Based on the capability of the injection system to provide a certain energetic amount of pilot fuel, the dual fuel combustion can be divided into three different concepts :
* the pilot fuel ignition concept,
* the substitution concept and
* the fully flexible dual fuel concept.
In a pilot fuel ignition concept, the injection system can inject a diesel mass with an energetic amount of less than 1% but cannot realize the full load operation with the diesel fuel. Usually, series production diesel engines retrofitted with a natural gas injection systems result in a substitution concept. The full load operation with diesel fuel is thereby possible but a diesel amount much greater than 1% has to be used to ignite the natural gas in dual fuel operation. A fully flexible dual fuel concept can operate with an arbitrary energetic content in dual fuel operation and no restriction exists for the full load diesel operation .
In this paper, the dual fuel combustion of a premixed natural gas/air charge with pilot ignition is studied. Therefore, only engine concepts with port fuel injected natural gas will further be explained below. A comprehensive overview and illustration of dual fuel engine concepts is given in . In high speed large bore engines, wide range injectors are currently in development. With wide range injectors, only one injection system is needed for full load diesel operation and dual fuel combustion with an energetic pilot fuel amount of less than 1%. The current challenges in the development of these systems are to ensure a stable injection quantity for the pilot fuel at a sufficient fuel atomization and to avoid nozzle coking due to the reduced diesel mass at full load . Therefore, current pilot injection systems are the benchmark for these systems. In marine propulsion systems, the large bore medium speed engines are usually equipped with two separate liquid fuel systems. The main injection system is used for full load operation and can also be operated with heavy fuel oil. The second injection system is the pilot fuel injection system operated with diesel fuel. The pilot injector is typically based on high speed heavy duty applications. Thus, the pilot diesel mass corresponds to the high load operation of these injection systems. Furthermore, the pilot injection system is active in all operation modes (cf. ) and, as a consequence, nozzle coking is usually not an issue concerning these pilot injection systems.
With a low energetic amount of pilot fuel, the natural gas/diesel dual fuel combustion process starts to show a positive dwell between the end of injection and the ignition of the diesel fuel [6, 7, 8, 9]. This behavior is typical for Reactivity Controlled Compression Ignition (RCCI) combustion. In RCCI combustion a high reactivity fuel ignites the low reactivity fuel during the compression stroke. Furthermore, reactivity gradient remains in the cylinder charge between both fuels as the high reactivity fuel is injected later in the compression stroke. Due to the positive dwell, the ignition is mainly kinetically driven . The term RCCI for such combustion processes is relatively new compared to the pilot ignited duel fuel combustion process. It was developed during the investigation of HCCI and its extension to higher loads by the use of two fuels with different reactivity. A good summary of the development of these low temperature combustion (LTC) strategies can be found in .
In conventional diesel combustion, the mixture preparation depends mainly on the atomization process. With a positive dwell, the mixture preparation starts depending on the processes during and after the end of injection (EoI). Besides the increased time to homogenize the pilot fuel, the ramp-down during the EoI enhances the entrainment resulting in lean mixtures near the nozzle and fuel-rich regions at the spray tip. Furthermore, this axial equivalence ratio distribution is in contrast to the steady-state period of a conventional diesel injection where the fuel rich regions are located near the injector nozzle [12, 13]. This phenomenon, called entrainment wave, has to be considered in pilot ignited dual fuel combustion to understand the ignition process.
A good overview ofthe current findings and topics of natural gas/diesel dual fuel combustion in heavy duty engines can be found in  and the related papers [15, 16, 17, 18, 19, 20]. These investigations presented the different combustion modes of dual fuel operation, the influence of in-cylinder motion on combustion and emissions, and the main challenges for full load operation caused by pre-ignitions and nozzle coking. In [7, 8, 9, 10, 21], the combustion process in high speed large bore dual fuel engines was investigated in detail. These studies revealed that a low energetic content, less than 1.5% of pilot fuel, was necessary to achieve low N[O.sub.x] emissions. Moreover, a reversal point could be observed in large bore engines. At the reversal point, a further advancing of the injection timing led to a retarded combustion phasing and, as a consequence, the one-to-one pairing between SoE and MFB50% at constant AFER, as normally seen in spark ignited engines, was no longer valid. With the appearance of the reversal point, early injection timing was no longer limited by engine knock. Instead, the combustion stability restricted the engine operation. In , different large bore dual fuel engine concepts were examined and a variable valve train was identified as a necessary feature for a competitive bivalent dual fuel engine. Up to now, no detailed study is currently known on the effects of different valve timings in case of dual fuel combustion in large bore engines. Therefore, the influence of different valve timings on pilot diesel/natural gas dual fuel combustion is investigated in detail throughout this paper. Furthermore, the influence of the engine speed is studied to identify the challenges for mobile applications and the increased need for transient engine operation.
Single-Cylinder Gas Engine, Test Bench, and Data Acquisition
The single-cylinder engine used in this investigation is based on a series production large bore gas engine. The displaced volume, stroke and bore were retained in the single cylinder engine (SCE) and the connecting rod dimensions and crankshaft were reworked to resist 30 MPa cylinder peak pressure. A production piston and cylinder head were used throughout all experiments. Therefore, only the water sleeve in the cylinder head had to be replaced to fit the injector. This resulted in a combustion chamber with piston bowl in combination with a flat cylinder head and a diesel injector in central position. The detailed engine and injector specifications are listed in Table 1. To reduce the mass forces the SCE has been equipped with a mass balance of the first and second order. The valve train consists of two overhead camshafts that control the valves by two cam followers. Different valve timings are realized through replacement of the corresponding camshaft. Two different inlet timings were investigated in the experiments. These are shown in Figure 1, in blue the MaxCC timing designed for maximum cylinder charge after inlet valve closing (IVC), and in red the Miller timing with the aim to reduce the charge temperature during the compression stroke. The exhaust valve timing (black) remained unchanged during all investigations.
A scheme of the test bench is shown in Figure 2. The fuel gas was obtained from the low-pressure natural gas grid. To raise the supply pressure, the natural gas was compressed and stored at 280 bar in a high-pressure tank. Thus, fluctuations of the gas composition were avoided during engine operation. Fuel gas mass flow was measured at high pressure with a Coriolis mass flow meter Rheonik RHE015 before the pressure was reduced to 16 bar supply pressure for the test bench. Thereafter a dome pressure regulator was installed to keep the pressure on the injection system constant at 9 bar independent of the natural gas mass flow. Twelve Bosch CNG injectors type NGI-2-S metered the natural gas mass flow and a venturi nozzle mixed it into the fresh air charge which was delivered by screw-type compressor. An AERZEN ZC11.3 lobed impeller flow meter measured the flow rate and a CS Instruments FA 300-2 Ex hygrometer recorded the humidity contained in the charged air. The installed flow heater allowed an air temperature up to 80 [degrees]C in the experimental investigations. A conventional common rail system was used for delivering a diesel fuel pressure up to 1600 bar.
Furthermore, an adjustable flap was installed in the exhaust gas system in order to simulate the back pressure of a turbocharger turbine. The back pressure was calculated according to the first turbocharger equation . The necessary values for the calculation were measured and averaged during the engine operation. The turbocharger efficiency was assumed with 75% during all experiments. The exhaust gas was analyzed by an Ansyco GASMET CX4000 FTIR and the oxygen concentration was measured by an M&C type PMA paramagnetic oxygen analyzer. The measured quantities were the ideal volumetric concentration of C[O.sub.2], CO, HCHO, THC (in [C.sub.3][H.sub.8] equivalent), C[H.sub.4] and N[O.sub.x] which were converted to mg/[mn.sup.3] at 5% [O.sub.2] in dry exhaust gas according to the regulations applied to stationary gas engines in Germany . An ABB NGC 8206 monitored the natural gas composition. As a result of both measurement techniques, the air-fuel equivalence ratio was calculated according to the method proposed by Brettschneider . Thereby, the influence of pilot diesel in the lambda calculation was neglected because of the small energetic content of two percent. The coolant and oil temperature of the engine were independently controlled to investigate their influence on engine operation. The cylinder head temperature was measured with four type K thermocouples between the valves. Furthermore, the intake, exhaust and cylinder liner temperatures were investigated. To analyze the working cycle the cylinder pressure was indicated using a Kistler 6061B piezoelectric pressure transducer and type 5011 charge amplifier. A Kistler piezoresistive pressure transducer type 4045A10 and 4075A10 and the corresponding charge amplifier type 4603 recorded the intake respective to the exhaust pressure. The recording rate of the pressure traces was set to 0.1[degrees]CA. The control and data acquisition system were developed based on the National Instruments PXI and cRIO platform. NI LabVIEW was used to develop the software.
Test Procedure and Measurement Data Evaluation
The dual fuel combustion process was analyzed varying the SoE at different AFER. In principle, to set the AFER three possibilities are available:
* [[??].sub.FuelGas] = f ([lambda]), [[??].sub.Air] = const,
* [[??].sub.Air] = f ([lambda]) [[??].sub.FuelGas] = const and
* [[??].sub.Air] = f ([lambda], SoE), [[??].sub.FuelGas] = f ([lambda], SoE) with the constraint IMEP = const.
A constant air mass flow leads only to minor changes in the cylinder charge density compared to an unchanged fuel gas mass flow. However, a constant air mass flow results in a changing maximum brake torque (MBT) with the variation of the AFER. Indeed, in both cases, the IMEP changes for different injection timings and combustion phasings. The third operating strategy is to adapt both mass flows to obtain a constant IMEP at the desired AFER. This entails a more extensive engine control strategy and a different cylinder charge density for every operating point. Pilot Diesel ignition is significantly controlled by the density and temperature in the combustion chamber. The in-cylinder charge temperature during compression has remained almost constant in the investigated load range using the same valve timing. However, experiments not presented in this paper revealed that density is a major influencing factor on the ignition behavior. For this reason, the first strategy with constant air mass flow was chosen for the analysis of the combustion process.
The strategy to obtain the engine operating map as a function of SoE and AFER is illustrated below. At first the AFER was gradually increased to determine the lean misfire limit. In this study, the misfire limit is defined as a standard deviation of the IMEP above three percent. Occasional misfire but also a continuous decrease in engine temperature and lapse of the combustion were the consequence exceeding this threshold. As a next step, the influence of SoE was measured at the lean misfire limit. Thereafter, the AFER was decreased in steps to investigate the impact of SoE on the engine behavior. Depending on the engine operating conditions different operating limits were investigated for the SoE. The limits observed for early SoE were either engine knocking or misfire and late SoE were limited by excessive exhaust gas temperatures (600 [degrees]C) or high THC emissions (above 3500 mg/[mn.sup.3]). To avoid injector nozzle coking a cylinder head temperature of 280 [degrees]C between the exhaust valves was set as a threshold value for rich AFER. For this reason, no pre-ignitions with richer AFER were observed in the experiments.
A number of 149 working cycles were recorded at every operating point. Subsequently, these were analyzed with a three pressure analysis (TPA) using GTPower v7.3.0 . In the TPA the gas exchange was calculated using intake and exhaust gas pressure traces. This result was then used together with the cylinder pressure trace to calculate the burn rate, efficiency analysis and related quantities. The measured natural gas composition and air humidity were applied in the modeling of the heating value and internal energy. Afterwards, to obtain a better understanding of the engine operating performance
* the indicated efficiency,
* the mass fraction burnt 50% (MFB50%),
* the mass fraction burnt 2% (MFB2%),
* the ignition delay, defined as the duration between SoE and MFB2%,
* the nitrogen oxides and
* THC emissions
were processed and visualized in engine operating maps or characteristic lines as a function of SoE and AFER using National Instruments Diadem. Efficiency analysis and apparent heat release rate (AHRR) were applied to compare operating points. The AHRR was calculated according to :
[d[Q.sub.H/d[phi]] = [[kappa]/[kappa]-1]p [dV/d[phi]]+[1/[kappa]+1]V[dp/d[phi]] Eq. (1)
Therefore, the cylinder pressure was averaged over the number of working cycles and the isentropic exponent was set to a constant value of [kappa] = 1.32. This method was chosen to obtain a validity check for the calculated burn rate with GTPower v7.3.Q because the AHRR was calculated using a numerical integration scheme. This was considered necessary due to the recorded pressure oscillations caused by the phenomena called ringing . The efficiency analysis in GTPower calculated the efficiency losses caused by
* unburned fuel ([[eta].sub.L,unburned]),
* combustion duration ([[eta].sub.L,Comb]),
* heat transfer ([[eta].sub.L,WH]) and
* gas exchange ([[eta].sub.L,GE]).
To calculate these losses, in a first step, the theoretical efficiency [[eta].sub.theo] was determined in GTPower. For this purpose, the thermodynamic cycle was calculated assuming a perfect gas exchange and an instantaneous combustion of all fuel contained in the combustion chamber. The compression and expansion losses were added as positive work. Furthermore, the heat transfer was neglected during the calculation of the theoretical efficiency. To obtain the losses caused by unburned fuel, the thermodynamic cycle was again calculated but the amount of burnt fuel was reduced according to the unburned components measured in the exhaust gas. The difference between both working cycles corresponded to efficiency loss caused by unburned fuel. In a next step, besides the unburned fuel, the heat release rate determined in the TPA was applied to the calculation of the thermodynamic cycle. Again, the difference between the last two working cycles was computed resulting in the losses caused by combustion duration. The heat transfer losses were computed in the same way. The gas exchange losses were obtained using the following formula:
[[eta].sub.L,GE] = [[eta].sub.theo] - [[eta].sub.L,unburned] - [[eta].sub.L,Comb]-[[eta].sub.L,WH]-[[eta].sub.i]. Eq. (2)
Thereby, the indicated efficiency, needed to determine [[eta].sub.L,GE], was calculated based on the cylinder pressure trace. A detailed description of this methodology is given in .
Results and Discussion
The natural gas composition remained constant during the measurement of one engine map. Between these, the gas chromatograph recorded slight fluctuations in the gas composition leading to a variation of the methane number between 98 and 99. During all investigations the composition consisted of about 96 mole-% methane, 2.5 mole-% ethane and 1.5 mole-% propane and longer-chained hydrocarbons. The pilot diesel conformed to Diesel EN 590 and had a cetane number of 54 (ACZ, DIN EN 15195:2015-02). During all experiments the diesel fuel was taken from the same tank. The injected volume was 10.4 [mm.sup.3] at an injection pressure of 750 bar corresponding to an energetic content of two percent in the investigated load range. To obtain the energizing profile, the injection rate was measured using an injection analyzer which operated corresponding to the Bosch method .
Three different operating conditions were investigated. First, the effect of Miller valve timing on the engine operation was investigated. During this engine map, the intake temperature was set to 60 [degrees]C due to the low combustion chamber temperatures and the resulting difficulties in the engine start behavior. Air mass flow was adjusted to a value of 500 kg/h at 1500 rpm. Second, MaxCC valve timing was analyzed to see the difference in ignition and combustion as a function of compression temperature. The intake temperature was lowered to 50 [degrees]C to reduce engine knocking. Under these circumstances a difference of nearly 100 [degrees]C in compression end temperature was present between these inlet valve timings as shown in Figure 3. The air mass flow and engine speed remained unchanged to ensure a similar cylinder charge density accepting a reduction in the IMEP. Third, the Miller valve timing with a reduced speed of 1000 rpm was investigated to understand the impact of the chemical time scale on engine operation. The intake temperature remained at 60 [degrees]C and the cylinder charge density was kept constant with an air mass flow of 333 kg/h. During all investigations, the intake and back pressure were adjusted to obtain the desired air mass flow and to simulate a turbocharger with an efficiency of 75 percent. The operating conditions are summarized in Table 2.
Miller Valve Timing
The basic engine behavior is observed using an engine map that shows MFB50% as a function of AFER and SoE as plotted in Figure 4. The AFER operating range was small with a width of about [DELTA][lambda] = 0.1 and values ranging from [lambda] = 1.53 to 1.63. As mentioned above, lean AFER were limited due to misfire and richer AFER overheated and coked the injector nozzles. Exhaust gas temperature limited late injection timings and early SoE resulted in misfire. Moreover, the one-to-one pairing between the start of energizing (SoE) and the mass fraction burnt 50% at constant air-fuel equivalence ratio and was no longer valid. In other words, advancing the injection timing at a constant AFER did not advance the MFB50% anymore. MFB50% as a function of SoE featured a reversal point at which advancing the SoE retarded the MFB50% (cf. Figure 5(a)). Furthermore, the N[O.sub.x] TA Luft curve corresponding to a N[O.sub.x] concentration of 500 mg/[mn.sup.3] in the exhaust gas at 5% oxygen concentration is shown in the engine map. Crossing this curve to higher values of AFER decreased the N[O.sub.x] emissions below 500 mg/[mn.sup.3] and vice versa. The minimum MFB50% of 12.3[degrees]CA ATDC was achieved for the richest AFER in the engine map. At this point also the maximum indicated efficiency (45.67%) was obtained considering the engine operating limits but ignoring emission legislation. In compliance with the German regulations, the maximum indicated efficiency of43.04% was found for earliest SoE on the N[O.sub.x] TA Luft curve.
A different visualization of the engine behavior is given in Figure 5 showing the different engine parameters as a function of SoE. This corresponds to a cut in the respective engine map at the specified AFER. MFB50% and THC emissions as function of SoE are plotted for [lambda] =1.56, 1.58 and 1.6 in Figure 5(a). The reversal point in the MFB50% curve was observed for every AFER and shifted towards later SoE for leaner gas mixtures. Furthermore, a leaner AFER retarded the combustion phasing for a given injection timing. The THC emissions level rose with a decrease in natural gas mass flow. Moreover, a reduction of THC emissions was measured while retarding the SoE. This was due to rising exhaust gas temperatures and improved conditions for the post-oxidation process. The trend for the N[O.sub.x] emissions and indicated efficiency is shown in Figure 5(b). The minimum of MFB50% corresponded, as expected, to the maximum of the indicated efficiency and the N[O.sub.x] emissions. The slight differences observed were a consequence of the interpolation process between the measured points. Besides, a richer AFER led to a higher level in the N[O.sub.x] emissions and the indicated efficiency as already described above. Furthermore, at an identical N[O.sub.x] emission value and AFER the earlier SoE resulted in a higher indicated efficiency. As a consequence, a N[O.sub.x] limited combustion process needs a SoE in advance of the reversal point to obtain the highest possible indicated efficiency. The characteristic lines of MFB2% and ignition delay are shown in Figure 5(c). The trend of MFB2% was similar to MFB50%. A reversal point was observed and a leaner AFER is retarding the MFB2%. The ignition delay increased with leaner mixtures but also with advancing the SoE. This is in contrast to a spark ignited combustion process which only shows a slight dependence of the ignition delay by the ignition timing. As the injection pressure remained unchanged, the hydraulic delay has only been an offset to the definition used in this paper and is not causing this behavior. In fact, the slight changes in MFB2% of about four degrees compared to about 15 degrees in SoE resulted in the observed characteristic of the ignition delay.
To investigate the ignition and combustion process in detail, Figure 6 shows the AHRR and the injection rate for selected operating points of the engine map at an AFER of [lambda] = 1.59. The first apparent heat release is enlarged to see the ignition in detail. The operating conditions and resulting quantities of the combustion process are listed in Table 3. Furthermore, the efficiency analysis is presented for each operating point. The highest AHRR was observed at a SoE of 695[degrees] CA with the first heat release starting directly after the EoI. An earlier SoE led to a later combustion phasing and smaller AHRR but the start of combustion (SoC) remained almost identical. Only the injection was advanced and the time between EoI and ignition rose. This is the reason for the ignition delay behavior described above and plotted in Figure 5(c). As expected, the SoE shift to 704[degrees] CA resulted in a later SoC and delayed combustion with the lowest peak AHRR. Interestingly, in this case the first heat release was observed during the diesel injection.
A closer examination of the efficiency analysis (Table 3) showed only minor differences in the losses caused by unburned fuel gas and gas exchange that was due to a similar level of THC emissions and only small differences in charge and back pressure. The high pressure ratio even led to a positive indicated work during the gas exchange. Also the theoretical efficiencies were on comparable level. As a consequence, only the combustion phasing has determined the indicated efficiencies resulting in the highest indicated efficiency of 43.96% for the SoE 696[degrees] CA case.
Further explanation concerning the ignition and combustion is possible taking into consideration the temperature during the compression. Due to the lower combustion chamber temperatures with Miller valve timing (cf. Figure 3), a certain degree crank angle and compression temperature near TDC was necessary for the ignition of the pilot diesel. This caused an increase in the ignition delay with advancing the SoE and consequently a longer duration between EoI and SoC. As a consequence this led to better mixing and homogenizing of the pilot diesel with the surrounding natural gas/air mixture. Furthermore, the volume covered with diesel fuel increased and the diesel fuel concentration was continuously decreasing, resulting in a slower combustion without changing the SoC and also in a lower local combustion temperature and therefore lower N[O.sub.x] emissions. Exactly this behavior was observed, shifting the SoE from 695 to 689[degrees] CA. With the SoE retarded to 704[degrees] CA, the ignition temperature was reached before a sufficient diesel mixture preparation was present. As a result, the SoC was retarded less than SoE was and overlapped with the injection process. Moreover, the delayed combustion process decreased the in cylinder peak temperatures and thus the N[O.sub.x] emissions but also the indicated efficiency.
MaxCC Valve Timing
The MFB50% engine map using MaxCC valve timing is shown in Figure 7. The AFER operating limits remained the same compared to Miller valve timing but the operating range was moved towards leaner AFER ranging from [lambda] = 1.88 to 2.08 and the width was increased to [DELTA][lambda] = 0.2. A different behavior was observed for a variation of the SoE. At rich AFER advancing the SoE always resulted in earlier MFB50% and the engine operation was limited due to the onset of knocking. With the AFER leaner than [lambda] = 2.00 a reversal point appeared and engine misfire got the new operation limit at early SoE. Furthermore, THC emissions started limiting late injection timings over the whole AFER range. The leaner AFER reduced the combustion chamber temperatures and thus led to incomplete burning and low exhaust gas temperatures. The global minimum MFB50% was 7[degrees] CA ATDC measured at an AFER of [lambda] = 2.00. In contrast to Miller valve timing, the global maximum indicated efficiency of 45.65% neglecting emission regulations was not located at this operating point but remained at the local minimum MFB50% at the rich AFER border ([lambda] = 1.88). The maximum indicated efficiency of 45.11% with N[O.sub.x] emissions below 500 mg/[mn.sup.3] was achieved at the reversal point with an AFER of [lambda] = 2.04 and SoE of 685[degrees] CA. This was also different compared to Miller valve timing where the maximum efficiency regarding the emission regulations was on the N[O.sub.x] TA Luft curve and at earliest SoE.
To gain a better understanding of the combustion process, the characteristic curves at the lean and rich operating limits are examined in detail. The curves for MFB50% and THC emission are shown in Figure 8(a). In contrast to Miller valve timing, the SoE had to be retarded with richer AFER to obtain a safe engine operation. Indeed, the curves plotted with an AFER of [lambda] = 1.91 and [lambda] = 2.08 have no common SoE anymore. As described above, the reversal point was observed at lean operating conditions while it disappeared at richer AFER. Due to the leaner AFER, the level of THC emissions was in general higher compared to Miller valve timing. Furthermore, a retarded MFB50% resulted in higher THC emissions because the lower exhaust gas temperatures inhibited the post-oxidation process. In other words, the completeness of the combustion was now determining the THC emissions.
The curve progression of the N[O.sub.x] emission and indicated efficiency (cf. Figure 8(b)) corresponds to the respective MFB50% trend but the maximum indicated efficiency has been located at the minimum MFB50% at the rich AFER border ([lambda] = 1.91) and not at the global minimum of MFB50%. However, the lean AFER achieved higher indicated efficiencies compared to [lambda] = 1.91 under the constraint not to exceed a certain N[O.sub.x] emission level. This was caused by the rapid decrease of the N[O.sub.x] emission advancing the SoE before the reversal point. As shown in Figure 8(c), the ignition delay and MFB2% at [lambda] = 2.08 exhibited the same behavior as observed with Miller valve timing. Only the timing of MFB2% was significantly advanced to -11[degrees] CA ATDC while the values for the ignition delay remained in the same range. However, the combustion process changed drastically with rich AFER. MFB2% started severely changing with different injection timings whereas the ignition delay remained almost constant. Only a slight increase appeared with advancing the SoE. Note that this behavior corresponds to a combustion process with spark ignition. Furthermore, the combustion process was investigated at the AFER operating limits using selected operating points.
Figure 9 shows the AHRR and injection rate of the SoE variation at [lambda] = 1.91. The first difference between the inlet valve timings was the intersection of the AHRR and related injection rate with MaxCC camshaft at [lambda] = 1.91, which means that SoC was always before EoI. This intersection had almost the same duration and only a slight increase was observed with later SoE. Second, the SoC was advanced with earlier SoE whereas with Miller valve timing the ignition remained unchanged with a further advancing of the SoE. Both observations corresponded to the curve progression shown in Figure 8(c) where the ignition delay decreased and MFB2% was retarded with later SoE. Further notice that a later SoE also decreased the peak AHRR and delayed the combustion.
The corresponding operating conditions and resulting quantities of the combustion process are listed in Table 4. Due to the longer inlet valve opening, the charging pressure was significantly reduced to keep an air mass flow of 500 kg/h. Compared to Miller valve timing, the gas exchange work started to achieve negative values with the losses rising to [[eta].sub.L,GE] = 0.48% ([lambda] = 1.91). In contrast, the theoretical efficiency increased with MaxCC camshaft due to the higher AFER but remained at the same level within the SoE variation. Retarding the SoE with MaxCC led to a later combustion phasing and as consequence the THC emission increased. Thus, the losses caused by combustion and unburned fuel rose with later SoE. This is different from Miller valve timing where only the combustion losses determined the indicated efficiency. Notice that the wall heat losses were not further mentioned, as they behave opposite to the combustion losses. Comparing the level of these losses between the different valve timings, MaxCC led to greater wall heat losses and lower combustion losses. This is due to the higher temperature level and the accelerated combustion (MFB50%). Furthermore, the higher AFER resulted in a higher level of THC emissions and thus the losses caused by unburned fuel were greater in the case of MaxCC valve timing. A more detailed comparison will be given later in this section. To summarize so far, the highest indicated efficiency of45.65% was achieved with MaxCC at [lambda] = 1.91 and the earliest SoE of 694[degrees] CA.
The combustion process changed significantly with the AFER increased to [lambda] = 2.08, as shown in Figure 10. The intersection between the AHRR and ignition delay disappeared, and thus the heat release became similar to the combustion process with Miller valve timing again. Notice that, the MFB50% was almost identical in the investigated operating points to visualize the behavior of the AHRR at the reversal point. To further describe the curve progression of the AHRR, the AHRR is divided into three different sections. The first characteristic point is the first apparent heat release called SoC as before (A). This point is not identical with the MFB2%, as this conversion rate is reached significantly later. The second section is the heat release until the turning point (B) named first AHRR (A-B) which, for example, is present in the blue curve at about -13[degrees] CA ATDC. This point is not present in every AHRR, especially at earlier SoE, and then the definition should rather be understood as an aid to dividing the AHRR. Finally, the remaining third section is then named main AHRR. In all SoE cases in Figure 10, the SoC was not significantly changing at an AFER of [lambda] = 2.08. The first AHRR started earlier at the MFB50% reversal point (SoE = 684[degrees] CA) compared to the other energizing timings. Retarding the SoE led to a later first AHRR with a higher slope until the turning point. Afterwards, the gradient of the AHRR decreased falling below the MFB50% and SoE 681[degrees] CA case which showed no first AHRR. The SoE 681[degrees] CA operating point had the largest incline and thus reached the maximum AHRR earlier than the SoE 687[degrees] CA case and almost at the same time as the MFB50% case (SoE = 684[degrees] CA), although the first AHRR was later. Due to the delayed combustion during the expansion the SoE 681[degrees] CA had no advantage in the combustion timing compared to a SoE of 687[degrees] CA.
The combustion process with Miller valve timing exhibited the same behavior. In Figure 6, the detail only did not enlarge this part of the oxidation because the focus was set to highlight the SoC.
The operating conditions and efficiency analysis are shown in Table 5. Due to the leaner AFER of [lambda] = 2.08, the exhaust gas temperature decreased. This resulted in an increase of the exhaust back pressure and the gas exchange losses. Furthermore, the theoretical efficiency was increasing compared to the AFER [lambda] = 1.91 operating points. As expected, the MFB50% ([lambda] = 2.08) case resulted in the highest indicated efficiency because of the lower combustion and unburned fuel losses. The combustion and wall heat losses between SoE 681 and 687[degrees] CA showed no significant difference, the slight disadvantage in combustion losses for SoE 681[degrees] CA can be interpreted as a consequence of the delayed combustion. The higher indicated efficiency of the SoE 687[degrees] CA case was mainly caused by its higher theoretical efficiency.
As described above, the penetrated volume and the diesel concentration are important parameters that determine the combustion process. Moreover, the spray tip penetration and the covered volume are not only determined by the injector geometry and injection pressure but also by the cylinder charge motion. With an AFER of [lambda] = 2.08 the ignition delay increased and as a result a swirl motion will significantly influence and improve the fuel mixing process prior combustion, and therefore will not only affect the post-oxidation process. Furthermore, the swirl will bend the spray and therefore reduce the spray tip penetration in radial direction. Notice that the cylinder head used in the investigations was equipped with a swirl port. Due to that, the cylinder charge motion is stronger with a later IVC and thus a better mixing is expected.
With the MFB2% achieved later than the SoC observed in the AHRR in Figure 10, the SoC seems to be the first-stage ignition of the diesel fuel. As a consequence, the first AHRR then corresponds to the second-stage ignition and combustion of the natural gas/diesel/air-mixture and the resulting main combustion is the oxidation of the remaining natural gas/air-mixture. It appears that a minimum MFB50% combustion with a fixed amount of pilot diesel is then a compromise between a large volume of fresh gas penetrated with high reactivity fuel and a certain local diesel concentration remaining for a fast ignition and first AHRR. At the reversal point (SoE 684[degrees] CA), the optimum amount of volume is penetrated with pilot diesel to ignite early and release an appropriate combustion energy to raise the pressure and temperature in the combustion chamber which results in a fast oxidation of the remaining natural gas/air mixture. Retarding the SoE leads to a smaller covered volume with pilot fuel but with a higher local diesel concentration. As a consequence, the first AHRR is retarded and less energy is released, but it has a faster burning rate. Due to the later and lower heat released, a lower cylinder charge temperature results in a more delayed main AHRR. With advancing the SoE to 681[degrees] CA, the covered volume is increased resulting in a lower pilot fuel concentration which allows the largest incline in the AHRR without a significant first AHRR and turning point. The lower THC emission compared to SoE 687[degrees] CA are very likely a result of the slightly delayed combustion and higher penetrated volume with high reactivity fuel. This theory explains the in-cylinder processes taking place around the reversal point.
A next step is to understand the boundary conditions for this engine behavior. At an AFER of [lambda] = 1.91 (Table 4 and Figure 9) advancing the SoE caused an earlier MFB50% and the intersection between the injection and combustion remained almost constant. Therefore, the ignition delay remained constant leading to the conclusion that the SoC under these circumstances is mixing controlled. With MaxCC valve timing and the later SoE at [lambda] = 1.91, the in-cylinder temperature was always high enough to ignite the pilot diesel. A further advancing of the SoE to increase the ignition delay was limited due to engine knocking. The leaner natural gas/air-mixture ([lambda] = 2.08, Table 5 and Figure 10) resulted in slower combustion, lower combustion temperature and finally increased the knocking resistance. Thus, the SoE had to be advanced to achieve the optimum combustion phasing where the temperature again limited the ignition of the pilot diesel. To sum up, to obtain a reversal point the in-cylinder temperature has to limit ignition of the pilot fuel. Therefore, the ignitability of the pilot fuel and engine knock is the limitation and every measure lowering the ignitability and increasing the knock resistance should enhance this behavior. In this study, low in-cylinder temperatures (Miller) and lean mixtures (MaxCC) were used to create a reversal point. Furthermore, for example a lower compression ratio, a less reactive pilot fuel, etc. should also be applicable. As expected, in  a reduction of the diesel mass also proved to be successful. Further notice that an interaction between the spray and the piston or cylinder liner as a reason for the reversal point was eliminated in additional studies using spray tip penetration measurements and CFD simulations. Moreover, the simulations revealed that the swirl is the main reason for avoiding the interaction between the spray tip and the piston. For the sake of brevity these studies are not presented in this paper.
To compare operating points in different operating ranges such as Miller and MaxCC valve timing, a suitable criterion is needed. A typical goal in engine development is to achieve maximum efficiency at a desired engine load and a given emission limit. Thus, the German emission regulation TA Luft is applied in this study and the operating points with the highest indicated efficiencies and N[O.sub.x] emissions below 500 mg/[mn.sup.3] are used to compare the different valve timings. Notice that the IMEP is not kept constant in this paper as the ignition is the main focus of this study. Recall that therefore the cylinder charge density has to remain constant while changing the AFER. Nevertheless, important information can be gained despite this disadvantage. Due to the high CoV at early SoE and operation below 500 mg/[mn.sup.3], a safety distance to the threshold value was needed to comply with the emission regulations at any point of time without further control technology. This is the reason that the N[O.sub.x] emissions are significantly lower than 500 mg/[mn.sup.3] at the compared operating points in Figure 11 and Table 6.
With the engine operation below 500 mg/[mn.sup.3] at maximum indicated efficiency, no intersection between the injection rate and the AHRR is observed in Figure 11 for both valve timings. However, Miller valve timing resulted in a longer ignition delay and more retarded combustion compared to MaxCC. Nevertheless, the combustion process with Miller camshaft had a higher gradient and peak AHRR.
The corresponding efficiency analysis is shown in Table 6. As reported above, the earlier MFB50% with MaxCC valve timing led to lower combustion losses and therefore to higher wall heat losses compared to early IVC. Furthermore, the higher AFER with MaxCC camshaft resulted in a higher theoretical efficiency compared to the Miller case. In return, the Miller valve timing led to higher combustion and exhaust gas temperatures. Hence, the MaxCC valve timing had higher THC emissions and as a consequence higher losses caused by unburned fuel. Moreover, the ratio between the exhaust gas back pressure and the charging pressure was higher in the MaxCC case. This resulted in a gas exchange process (cf. Figure 12) with no positive work in contrast to the engine operation with Miller camshaft. In sum, the indicated efficiency of 45.11% with MaxCC was 2.06% higher compared to the Miller valve timing with an efficiency of43.04% under the constraint of N[O.sub.x] emission below 500 mg/[mn.sup.3].
The main disadvantage of the Miller valve timing was the late combustion phasing and the resulting combustion losses. Nevertheless, an earlier combustion phasing with N[O.sub.x] emissions below the threshold was not possible. The N[O.sub.x] emissions in the pilot ignited dual fuel combustion process consists of the N[O.sub.x] emissions formed during the ignition and combustion of the diesel/natural gas/air-mixture and the subsequent flame propagation in the natural gas/air-mixture. Due to the higher adiabatic flame temperature with Miller valve timing at [lambda] = 1.59, the combustion phasing needs to be set after TDC in the expansion stroke to keep the combustion temperature and N[O.sub.x] formation rate low. Furthermore, an early injection before the reversal point is needed to homogenize the diesel fuel. This leads to a lower diesel concentration and therefore a lower local temperature and N[O.sub.x] emissions. In contrast, the lower adiabatic temperature of the fresh gas with [lambda] = 2.04 allows a combustion phasing early enough to compensate the losses caused by the unburned fuel. Nevertheless, the conclusion that MaxCC valve timing is the optimal solution for a dual fuel combustion system is not correct. Indeed, an optimal valve timing results in a compression end temperature and consequently in an AFER that leads to an optimum between the combustion phasing and the THC emissions. In the present case, the optimum valve timing would be in between the Miller and MaxCC camshaft. But also, different measures such as increasing the compression ratio with Miller valve timing must be considered when seeking the optimum, as the engine operation is affected by the effective compression ratio. Furthermore, it is expected that the optimal effective compression ratio depends on the engine load and therefore has to be adapted according to the engine operating conditions.
Miller Valve Timing 1000 rpm
Figure 13 shows the MFB50% engine map using a reduced engine speed of 1000 rpm. The operating limits remained qualitatively identical compared to the higher engine speed with Miller valve timing. Thus, exhaust gas temperature and engine misfire were still limiting the SoE. Furthermore, a reversal point is observed throughout the engine map. With the lean misfire limit shifted to [lambda] = 1.75, the AFER operating range was extended to a width of [DELTA][lambda] = 0.19 which was almost equal to the MaxCC valve timing. The N[O.sub.x] TA Luft curve was shifted towards leaner AFER compared to Miller valve timing with 1500 rpm and thereby reduced the operating range with N[O.sub.x] emissions below 500 mg/[mn.sup.3]. Besides, the minimum MFB50% at the rich AFER limit was decreased from 12.3 (Miller, 1500 rpm) to 5.5[degrees] CA ATDC (Miller, 1000 rpm) with a maximum indicated efficiency of 46.72%. Notice that the maximum indicated efficiency in compliance with the emission regulation was also found for earliest SoE at the N[O.sub.x] TA Luft curve. In this case, the indicated efficiency was increased to 44.58% compared to 43.04% with Miller valve timing at 1500 rpm.
The characteristic curves of the MFB50% and the THC emissions can be seen in Figure 14(a). The curve progression of the MFB50% had no significant difference compared to the higher engine speed. At equal AFER and valve timing, the reduced engine speed only showed an earlier combustion phasing and reversal point. Furthermore, the lower engine speed led to a smaller operating range at excessive AFER. For this operating condition, also a different curve progression in THC emission was observed. Like MaxCC valve timing, the THC emissions were increased with retarding the MFB50%. This behavior was not observed until an AFER of [lambda] = 1.66 where retarding the MFB50% still reduced the THC emissions. As a consequence, the combustion temperature that inhibited the post oxidation process was located between [lambda] = 1.66 and 1.74. This transition was also observed by Konigsson  but at a lower AFER of about [lambda] = 1.6. Based on further investigations that are not presented in this paper, it is expected that this turning point in engine operation not only depends on the fuel properties but also on the engine speed or the time scale of the combustion process.
Figure 14(b) shows the characteristic curves of the N[O.sub.x] emissions and the indicated efficiency. The curve progression was as expected, and the only difference compared to higher engine speed was the higher level of the indicated efficiency and the higher N[O.sub.x] emissions at equal AFER. The rapid increase in N[O.sub.x] emissions with lower engine speed showed the sensitivity of the N[O.sub.x] formation rate of the present time scale. The MFB2% and ignition delay curves are plotted in Figure 14(c). With lower engine speed, the MFB2% was significantly advanced at an equal AFER compared to Miller valve timing at 1500 rpm. As the time scale increases with the lower engine speed, this behavior is expected. In contrast, the ignition delay remained almost identical which was caused by the change in the SoE operating range at 1000 rpm.
Figure 15 shows the injection rate and AHRR of the SoE variation at [lambda] = 1.72 and 1000 rpm. The corresponding operating conditions and efficiency analysis are listed in Table 7. Notice that the crank-angle-based injection duration and hydraulic delay decreased while the maximum crank-angle-based injection rate increased compared to the cases with higher engine speed (cf. Figure 16). This was a result of the changed time-scale and not caused by a different injection pressure or injected volume. The trends in AHRR and its relation to the injection rate remained qualitatively the same compared to Miller valve timing at 1500 rpm. Nevertheless, small differences were observed in the AHRR. At an engine speed of 1000 rpm, the maximum of the AHRR was higher in the early SoE case (691[degrees] CA) compared to the MFB50% case (697[degrees] CA). Furthermore, the position of maximum AHRR showed no significant difference between SoE 697 and 703[degrees] CA case.
Like MaxCC valve timing, the indicated efficiency at [lambda] = 1.72 (Table 7) depended on the combustion losses and the losses caused by unburned fuel due to the inhibited post-oxidation process. A more detailed explanation of the efficiency analysis will be given below with the comparison of the different engine speeds.
The basis of comparison remained the same as described earlier with Miller and MaxCC valve timing. The operating points with maximum indicated efficiency and N[O.sub.x] emission below 500 mg/[mn.sup.3] were chosen from each engine map. The difference in IMEP was 0.9 bar between the operating points as shown in Table 8. The lower power output at 1000 rpm was due to the higher AFER needed to fulfil the N[O.sub.x] threshold. Recall that the ignition is the main focus of the study and thus this difference has to be accepted.
Figure 16 illustrates the AHRR and injection rate for both engine speeds with Miller valve timing. The first difference observed was the injection rate due to the different time scale. Second, the SoC and combustion phasing was advanced with the lower engine speed. On the contrary, the maximum of the AHRR showed no significant difference between both engine speeds.
The efficiency analysis in Table 8 reveals a more positive gas exchange work with lower engine speed. Both operating points showed no big difference in the exhaust gas back pressure. The charging pressure was about 1 bar higher in the case of 1500 rpm. Thus, the pressure ratio would predict an advantage for 1500 rpm and is not able to explain the situation without further checking the gas exchange. Therefore, Figure 17 shows the cylinder pressure during the gas exchange for both operating points. After EVO, the cylinder pressure decreased in both cases, but a faster decline to a lower pressure level was observed with the lower engine speed. Furthermore, the exhaust gases were pushed out at a lower cylinder pressure level at 1000 rpm. This effect was a result of the longer time-based opening period of the exhaust valve with lower engine speed which led to a reduced effective flow resistance. Finally, this resulted in a lower pressure-volume work by the piston and thus fewer losses in this period of the scavenging at 1000 rpm. In the first period of the intake stroke, a higher in-cylinder pressure was achieved at 1500 rpm. Afterwards, the pressure declined to the same pressure level as 1000 rpm due to the higher flow resistance while closing the inlet valve. Thus, the higher engine speed led to a greater (positive) pressure-volume work during the intake stroke. But in sum, this cannot compensate the losses during the exhaust stroke at 1500 rpm and thus the lower engine speed resulted in a better gas exchange as is also seen by comparing the areas of the gas exchange. Furthermore, the leaner AFER at 1000 rpm raised the theoretical efficiency compared to the higher engine speed but also increased the losses by unburned fuel. Moreover, as expected, fewer combustion losses at 1000 rpm were achieved due to the earlier combustion phasing at higher AFER. To sum up, at 1000 rpm, a significantly better gas exchange and combustion in combination with a higher theoretical efficiency compensated the higher losses due to unburned fuel and wall heat transfer and resulted in an indicated efficiency of 44.58% which was 1.54% higher compared to the 1500 rpm engine operation.
Summary and Conclusion
The effects of reduced engine speed and compression end temperature on the ignition and combustion process of a pilot ignited large bore gas engine were investigated in detail. The cylinder charge density was kept constant during all investigations. Therefore, the air mass flow was only adjusted to the engine speed. During an engine map, the AFER was modified by regulating the natural gas mass flow. The injected pilot diesel volume of 10.4 [mm.sup.3], corresponding to approx. 2% energetic content, remained unchanged in all experiments. The key observations for the different valve timings and engine speeds are summarized below.
Miller valve timing, 1500 rpm:
* The engine operation showed a small AFER operating range from [lambda] = 1.53 to 1.63.
* Late SoE were limited due to excessive exhaust gas temperature and early SoE resulted in engine misfire.
* The one-to-one pairing between SoE and MFB50% at constant AFER was no longer observed with Miller valve timing. The MFB50% curve exhibited a reversal point at which further advancing of the SoE resulted in a retarded MFB50%
* The ignition delay increased significantly, while the MFB2% changed only slightly, with advancing the SoE at constant AFER.
MaxCC valve timing, 1500 rpm:
* A notable difference in the engine operation was observed, as the AFER operating range was shifted to leaner values ranging from [lambda] = 1.88 to 2.08.
* Due to the decreased combustion temperature, the post oxidation process was inhibited and thus a retarded MFB50% resulted in increased THC emission. As a consequence, late SoE were limited by excessive THC emissions.
* Like Miller valve timing, a reversal point was observed for lean AFER until [lambda] = 2.0 with MaxCC and thus early SoE were limited due to engine misfire. With richer AFER, engine knock started to limit further advancing of the SoE and the reversal point disappeared in the MFB50% curves.
* This behavior resulted in a change of the curve progression in the MFB2% and ignition delay curves. With AFER leaner than [lambda] = 2.0 the curve progression remained qualitatively the same compared to Miller valve timing, only the MFB2% values were significantly advanced. At richer AFER, the ignition delay stayed almost constant and therefore the EU2% was significantly advanced with earlier SoE, which is according to the behavior of a spark ignited (gas) engine.
Miller valve timing, 1000 rpm:
* With the reduced engine speed and Miller valve timing, the engine operation map and operating limits remained quantitatively the same. Only the AFER operating range was significantly increased from [lambda] = 1.56 to 1.75.
* Due to the AFER operating range, a transition in the behavior of the THC emissions was observed as the post oxidation process was inhibited with AFER leaner than [lambda] = 1.66.
Based on these results the necessary conditions for the appearance of a reversal point were identified. To obtain a reversal point, it is necessary to advance the injection before the ignition temperature of pilot fuel is achieved in the combustion chamber. As a result, the ignition delay increases and the mixture preparation start depending on the processes after the injection and not only on the spray break up. These processes are the increased entrainment with the ramp down of the injection rate, the cylinder charge motion and the time for mixing between the pilot diesel and the natural gas/air-mixture . To advance the injection before the ignition temperature two measures, and their combination, are possible. First, the reduction of the ignitability of the pilot fuel and second an increased knock resistance of the main fuel. In the present study a reduced cylinder charge temperature with Miller valve timing was used to increase the ignition delay of the pilot fuel. In the case of MaxCC valve timing the knock resistance was increased due to leaner AFER. But also other measures, like e.g. different fuel properties or compression ratios, are expected to control this behavior.
This research was funded by the Bavarian Research Foundation (BFS) as part of the project "Effizienzsteigerung von DF Motoren bei Mitteldrucken >24bar" (AZ-1082-13).
AFER - Air-Fuel Equivalence Ratio
AHRR - Apparent Heat Release Rate
ATDC - After Top Dead Center
BTC - Bottom Dead Center
CA - Crank Angle
CH4 - Methane
CO - Carbon Monoxide
CO2 - Carbon Dioxide
CoV - Coefficient of Variance
EoI - End of Injection
EVO - Exhaust Valve Opening
FTIR - Fourier Transform Infrared Spectroscopy
HCCI - Homogeneous Charge Compression Ignition
HCHO - Formaldehyde
IMEP - Indicated Mean Effective Pressure
IVC - Inlet Valve Closing
IVO - Inlet Valve Opening
LTC - Low Temperature Combustion
MBT - Maximum Brake Torque
MFB - Mass Fraction Burned
NGC - Natural Gas Chromatograph
N[O.sub.x] - Oxides of Nitrogen (NO and N[O.sub.2])
SoC - Start of Combustion
SoE - Start of Energizing
SCE - Single Cylinder Engine
TDC - Top Dead Center
THC - Total HydroCarbons
TPA - Three Pressure Analysis
[1.] Nylund, I. and Ott, M., "Development of a Dual Fuel Technology for Slow-Speed Engines," CIMAC Congress 2013, Shanghai, 2013.
[2.] Karim, G.A., Dual Fuel Diesel Engines, (Boca Raton, CRC Press, 2015), ISBN-13: 978-1-4987-0309-3.
[3.] Redtenbacher, C., Kiesling, C., Wimmer, A. et al., "Dual Fuel Brennverfahren - Ein zukunftsweisendes Konzept vom PKW- bis zum Gro[ss]motorenbereich?," 37. Internationales Wiener Motorensymposium, Wien, 2016.
[4.] Senghaas, C., "Simlified L'Orange Fuel Injection System for Dual Fuel Applications," CIMAC Congress 2016, Helsinki, 2016.
[5.] Banck, A., "Dual Fuel Engine Optimized for Marine Applications," CIMAC Congress 2016, Helsinki, 2016.
[6.] Redtenbacher, C., Kiesling, C., Malin, M. et al., "Potential and Limitations of Dual Fuel Operation of High Speed Large Engines," Proceedings of the ASME 2016 Internal Combustion Engine Fall Technical Conference, ICEF2016-9359, Greenville, SC, 2016.
[7.] Kiesling, C., Redtenbacher, C., Wimmer, A. et al., "Detailed Assessment of an Advanced Wide Range Diesel Injector for Dual Fuel Operation of Large Engines," CIMAC Congress 2016, Helsinki, 2016.
[8.] Redtenbacher, C., Kiesling, C., Malin, M. et al., "Requirements for Diesel Pilot Injection of Diesel-Gas Dual Fuel Engines to Achieve the Highest Efficiency with the Lowest Emissions," 10th Dessau Gas Engine Conference, Dessau, 2017.
[9.] Henke, B., Buchholz, B., Schleef, K. et al., "Pilot Injection Strategies for Medium Speed Dual Fuel Engines," 10th Dessau Gas Engine Conference, Dessau, 2017.
[10.] Kokjohn, S.L., Musculus, M.P.B., and Reitz, R.D., "Evaluating Temperature and Fuel Stratification for Heat-Release Rate Control in a Reactivity-Controlled Compression-Ignition Engine Using Optical Diagnostics and Chemical Kinetics Modeling," Combustion and Flame 162(2015):2729-2742, 2015, doi:10.1016/j.combustflame.2015.04.009.
[11.] Nieman, D., Dempsey, A., and Reitz, R., "Heavy-Duty RCCI Operation Using Natural Gas and Diesel," SAE Int. J. Engines 5(2):270-285, 2012, doi:10.4271/2012-01-0379.
[12.] Musculus, M.P.B., Miles, P.C., and Pickett, L.M., "Conceptual Models for Partially Premixed Low-Temperature Diesel Combustion," Progress in Energy and Combustion Science 39(2013):246-283, 2013, doi:10.1016/i.pecs.2012.09.0001.
[13.] Musculus, M. and Kattke, K., "Entrainment Waves in Diesel Jets," SAE Int. J. Engines 2(1):1170-1193, 2009, doi:10.427l/2009-01-1355.
[14.] Konigsson, F., "On the Combustion in the CNG-Diesel Dual Fuel Engine," Ph.D. thesis, Department of Machine Design, Royal Institute of Technology, Stockholm, 2014.
[15.] Konigsson, F., Stalhammar, P., and Angstrom, H., "Combustion Modes in a Diesel-CNG Dual Fuel Engine," SAE Technical Paper 2011-01-1962, 2011, doi:10.4271/2011-01-1962.
[16.] Konigsson, F., Stalhammar, P., and Angstrom, H., "Characterization and Potential of Dual Fuel Combustion in a Modern Diesel Engine," SAE Technical Paper 2011-01-2223, 2011, doi:10.4271/2011-01-2223.
[17.] Konigsson, F., Stalhammar, P., and Angstrom, H., "Controlling the Injector Tip Temperature in a Diesel Dual Fuel Engine," SAE Technical Paper 2012-01-0826, 2012, doi:10.4271/2012-01-0826.
[18.] Konigsson, F., Kuyper, J., Stalhammar, P., and Angstrom, H., "The Influence of Crevices on Hydrocarbon Emissions from a Diesel-Methane Dual Fuel Engine," SAE Int. J. Engines 6(2):751-765, 2013, doi:10.4271/2013-01-0848.
[19.] Konigsson, F., Dembinski, H., and Angstrom, H., "The Influence of In-Cylinder Flows on Emissions and Heat Transfer from Methane-Diesel Dual Fuel Combustion," SAE Int. J. Engines 6(4):1877-1887, 2013, doi:10.4271/2013-01-2509.
[20.] Konigsson, F., Risberg, P., and Angstrom, H., "Nozzle Coking in CNG-Diesel Dual Fuel Engines," SAE Technical Paper 2014-01-2700, 2014, doi:10.4271/20l4-01-2700.
[21.] Kammerdiener, T., Schlick, H., and Schonbacher, M., "Concept Investigations for a Dual Fuel Engine Based on Experimental Studies on a High Speed Single Cylinder Engine," 9th Dessau Gas Engine Conference, Dessau, 2015.
[22.] Lange, H., "Investigation of Alternative Dual Fuel Engine Concepts," CIMAC Congress 2016, Helsinki, 2016.
[23.] Nguyen-Schafer, H., Rotordynamics of Automotive Turbochargers, (Berlin, Springer, 2015), ISBN-13:978-3319176437.
[24.] Bundesministerium fur Umwelt, Naturschutz und Reaktorsicherheit, "Erste Allgemeine Verwaltungsvorschrift zum Bundes-Immisionsschutzgesetz (Technische Anleitung zur Reinhaltung der Luft - TA Luft)," 2002.
[25.] Brettschneider, J., "Berechnung des Luftverhaltnisses [lambda] on Luft-Kraftstoff-Gemischen und des Einflusses von Me[??]fehlern auf [lambda]," Bosch Technische Berichte 6(4):177-186, 1979.
[26.] Gamma Technologies, "GT Suite. Engine Performance Manual," Version 7.3, 2012.
[27.] Bargende, M., "Ein Gleichungsansatz zur Berechnung der instationaren Wandwarmeverluste im Hochdruckteil von Ottomotoren," Dissertation, Technische Universitat Darmstadt, Darmstadt, 1990.
[28.] Kirsten, M., Pirker, G., Redtenbacher, C., Wimmer, A. et al., "Advanced Knock Detection for Diesel/Natural Gas Engine Operation," SAE Int. J. Engines 9(3):1571-1583, 2016, doi:10.4271/2016-01-0785.
[29.] Bosch, W., "Der Einspritzgesetz-Indikator, ein neues Me[beta]gerat zur direkten Bestimmung des Einspritzgesetzes von Einzeleinspritzungen," Motorentechnische Zeitschrift 25(7):268-282, 1964.
Stefan Weber, Technical University of Munich
Richard Stegmann, Maximilian Prager, and Georg Wachtmeister, Technical University of Munich
Received: 28 Nov 2017
Revised: 04 Feb 2018
Accepted: 20 Feb 2018
e-Available: 18 Apr 2018
Natural Gas, Gas Engines, Dual Fuel Combustion, RCCI, Engine Speed, Inlet Valve
Timing, Miller Valve Timing
Weber, S., Stegmann, R., Prager, M., and Wachtmeister, G., "The Effect of Inlet Valve Timing and Engine Speed on Dual Fuel NG-Diesel Combustion in a Large Bore Engine," SAE Int. J. Engines 11(2):2018, doi:10.4271/03-11-02-0015.
TABLE 1 Single cylinder engine and injector specifications. Engine specifications Displaced volume 4.77 [dm.sup.3] Stroke 210 mm Bore 170 mm Connecting rod length 480 mm Compression ratio 12.7:1 Number of inlet valves 2 Number of exhaust valves 2 Injector specifications Injector type Bosch CRI2.16 Cup (tip) type Mini-sac Number of holes & 3, equally-spaced arrangement Spray pattern included angle 160[degrees] Nozzle orifice treatment None (square-edged) Nozzle orifice diameter 0.100 mm, cylindrical Static flow rate (without 77 ml (@ 100 bar & 30 s) needle) TABLE 2 Operating conditions used in the different engine maps. Engine Map Operating Conditions Miller, MaxCC, Miller, 1500 rpm 1500 rpm 1000 rpm AFER 1.53-1.63 1.88-2.08 1.56-1.75 Air mass flow 500 kg/h 500 kg/h 333 kg/h Intake 60 [degrees]C 50 [degrees]C 60 [degrees]C temperature Turbocharger 75[degrees]% 75% 75% efficiency Rail pressure 750 bar 750 bar 750 bar Injected 10.4 [mm.sup.3] 10.4 [mm.sup.3] 10.4 [mm.sup.3] volume TABLE 3 Operating conditions and efficiency analysis for the cases with SoE timings 689, 695, and 704[degrees] CA. The boundary conditions were Miller valve timing, 60 [degrees]C intake temperature, 1500 rpm, 500 kg/h air mass flow, 750 bar rail pressure and 10.4 [mm.sup.3] injected diesel fuel. Operating conditions SoE SoE SoE 689[degrees] 695[degrees] 704[degrees] CA CA CA SoE in [degrees]CA 689 695 704 Rail Pressure in bar 750 750 750 Injected volume in 10.4 10.4 10.4 [mm.sup.3] AFER 1.59 1.59 1.59 Charging pressure in 5.40 5.39 5.37 bar Back pressure in bar 2.70 2.72 2.64 MFB50% in [degrees]CA ATDC 19.8 17.8 24.1 IMEP in bar 18.2 18.7 17.7 N[O.sub.x] in mg 457 604 308 /[mn.sup.3] THC in mg/[mn.sup.3] 1082 1107 1014 Efficiency analysis [[eta].sub.theo] 52.79% 53.14% 52.99% [[eta].sub.L,unburned] 1.51% 1.55% 1.42% [[eta].sub.L,Comb] 6.30% 5.51% 7.95% [[eta].sub.L,WH] 2.36% 2.50% 2.12% [[eta].sub.L,GE] -0.42% -0.38% -0.45% [[eta].sub.I] 43.04% 43.96% 41.95% TABLE 4 Operating conditions and efficiency analysis for the cases with SoE timings 694, 698, and 702[degrees] CA. The boundary conditions were MaxCC valve timing, 50 [degrees]C intake temperature, 1500 rpm, 500 kg/h air mass flow, 750 bar rail pressure and 10.4 [mm.sup.3] injected diesel fuel. Operating conditions SoE SoE SoE 694[degrees] 698[degrees] 702[degrees] CA CA CA SoE in [degrees]CA 694 698 702 Rail pressure 750 750 750 in bar Injected volume 10.4 10.4 10.4 in [mm.sup.3] AFER 1.90 1.91 1.90 Charging 2.72 2.71 2.70 pressure in bar Back pressure 1.84 1.83 1.79 in bar MFB50% in 8.5 11.8 16.6 [degrees]CA ATDC IMEP in bar 16.1 15.9 15.3 N[O.sub.x] in mg/ 818 523 279 [mn.sup.3] THC in mg/ 1799 1998 2722 [mn.sup.3] Efficiency analysis [[eta].sub.theo] 54.51% 54.72% 54.81% [[eta].sub.L,unburned] 2.21% 2.46% 3.33% [[eta].sub.L,Comb] 2.33% 3.20% 4.63% [[eta].sub.L,WH] 3.87% 3.51% 3.08% [[eta].sub.L,GE] 0.45% 0.48% 0.45% [[eta].sub.I] 45.65% 45.07% 43.32% TABLE 5 Operating conditions and efficiency analysis for the cases with SoE timings 681, 684, and 687[degrees] CA. The boundary conditions were MaxCC valve timing, 50 [degrees]C intake temperature, 1500 rpm, 500 kg/h air mass flow, 750 bar rail pressure and 10.4 [mm.sup.3] injected diesel fuel. Operating conditions SoE SoE SoE 681[degrees] 684[degrees] 687[degrees] CA CA CA SoE in [degrees]CA 681 684 687 Rail pressure 750 750 750 in bar Injected 10.4 10.4 10.4 volume in [mm.sup.3] AFER 2.08 2.08 2.08 Charging 2.70 2.70 2.70 pressure in bar Back pressure 1.90 1.90 1.89 in bar MFB50% in 8.8 7.7 8.8 [degrees]CA ATDC IMEP in bar 14.1 14.3 14.1 N[O.sub.x] in mg/ 268 339 317 [mn.sup.3] THC in mg/ 3745 3499 4005 [mn.sup.3] Efficiency analysis [[eta].sub.theo] 55.26%% 55.36%% 55.64%% [[eta].sub.L,unburned] 4.50%% 4.20%% 4.73%% [[eta].sub.L,Comb] 2.46%% 2.26%% 2.44%% [[eta].sub.L,WH] 3.77%% 3.91%% 3.78%% [[eta].sub.L,GE] 0.71%% 0.70%% 0.69%% [[eta].sub.I] 43.82% 44.29% 44.00% TABLE 6 Operating conditions and efficiency analysis for the cases comparing inlet valve timing MaxCC and Miller. The boundary conditions were 1500 rpm, 500 kg/h air mass flow, 750 bar rail pressure and 10.4[mm.sup.3] injected diesel fuel for both cases. The intake temperatures were 50 [degrees]C for MaxCC and 60 [degrees]C in case of Miller. The N[O.sub.x] emissions are below 500 mg/[mn.sup.3] to meet the TA Luft  regulations. Operating conditions MaxCC Miller SoE in [degrees]CA 685 689 Rail Pressure in bar 750 750 Injected Volume in 10.4 10.4 [mm.sup.3] AFER 2.04 1.59 Charging Pressure in 2.70 5.40 bar Back Pressure in bar 1.89 2.70 MFB50% in [degrees]CA 7.2 19.8 ATDC IMEP in bar 14.8 18.2 N[O.sub.x] in mg/[mn.sup.3] 462 457 THC in mg/[mn.sup.3] 2614 1082 Efficiency analysis [[eta].sub.theo] 55.11% 52.79% [[eta].sub.L,unburned] 3.17% 1.51% [[eta].sub.L,Comb] 2.20% 6.30% [[eta].sub.L,WH] 4.02% 2.36% [[eta].sub.L,GE] 0.61% -0.42% [[eta].sub.I] 45.11% 43.04% TABLE 7 Operating conditions and efficiency analysis for the cases with SoE timings 691, 697, and 702[degrees] CA. The boundary conditions were Miller valve timing, 60 [degrees]C intake temperature, 1000 rpm, 333 kg/h air mass flow, 750 bar rail pressure and 10.4 [mm.sup.3] injected diesel fuel. Operating conditions SoE SoE SoE 691[degrees] 697[degrees] 703[degrees] CA CA CA SoE in [degrees]CA 691 697 703 Rail pressure 750 750 750 in bar Injected 10.4 10.4 10.4 volume in [mm.sup.3] AFER 1.72 1.72 1.72 Charging 4.36 4.37 4.35 pressure in bar Back pressure 2.49 2.54 2.47 in bar MFB50% in 16.0 13.4 18.5 [degrees]CA ATDC IMEP in bar 17.3 17.6 16.9 N[O.sub.x] in mg/ 337 757 383 [mn.sup.3] THC in mg/ 2218 2061 2337 [mn.sup.3] Efficiency analysis [[eta].sub.theo] 53.87%% 53.86%% 53.61%% [[eta].sub.L,unburned] 2.64%% 2.49%% 2.78%% [[eta].sub.L,Comb] 4.73%% 3.93%% 5.65%% [[eta].sub.L,WH] 3.32%% 3.63%% 3.10%% [[eta].sub.L,GE] -1.40%% -1.34%% -1.40%% [[eta].sub.I] 44.58% 45.15% 43.48% TABLE 8 Operating conditions and efficiency analysis for the cases comparing the engine speed. The boundary conditions were Miller valve timing, 60 [degrees]C intake temperature, 750 bar rail pressure and 10.4 [mm.sup.3] injected diesel fuel for both cases. To obtain equal cylinder charge density the air mass flow was reduced to 333 kg/h for the 1000 rpm case The N[O.sub.x] emissions are below 500 mg/[mn.sup.3] to meet the TA Luft  regulations. Operating conditions 1000 rpm 1500 rpm SoE in [degrees]CA 703 689 Rail pressure in bar 750 750 Injected volume in 10.4 10.4 [mm.sup.3] AFER 1.72 1.59 Charging pressure in 4.36 5.40 bar Back pressure in bar 2.49 2.70 MFB50% in [degrees]CA 16.0 19.8 ATDC IMEP in bar 17.3 18.2 N[O.sub.x] in mg/[mn.sup.3] 337 457 THC in mg/[mn.sup.3] 2218 1082 Efficiency analysis [[eta].sub.theo] 53.87%% 52.79%% [[eta].sub.L,unburned] 2.64%% 1.51% [[eta].sub.L,Comb] 4.73%% 6.30%% [[eta].sub.L,WH] 3.32%% 2.36%% [[eta].sub.L,GE] -1.40%% -0.42%% [[eta].sub.I] 44.58% 43.04%
|Printer friendly Cite/link Email Feedback|
|Author:||Weber, Stefan; Stegmann, Richard; Prager, Maximilian; Wachtmeister, Georg|
|Publication:||SAE International Journal of Engines|
|Article Type:||Technical report|
|Date:||Apr 1, 2018|
|Previous Article:||Elasto-Hydrodynamic Bearing Model in Powertrain Multi-Body Simulation.|
|Next Article:||A Novel Approach towards Stable and Low Emission Stratified Lean Combustion Employing Two Solenoid Multi-Hole Direct Injectors.|