Experimental Study of Air and EGR Dilution in a Pre-Chamber Spark-Ignited Engine Fueled by Methane

Abstract

Improving the efficiency of spark-ignited (SI) engines while simultaneously reducing emissions remains a critical challenge in meeting global energy demands and increasingly stringent environmental regulations. Lean burn combustion is a proven strategy for increasing efficiency in SI engines. However, the air dilution level is limited by the mixture’s ignition ability and poor combustion efficiency and stability. A promising method to extend the dilution limit and ensure stable combustion is the implementation of an active pre-chamber combustion system. The pre-chamber spark-ignited (PCSI) engine facilitates stable and rapid combustion of very lean mixtures in the main chamber by utilizing high ignition energy from multiple flame jets penetrating from the pre-chamber (PC) to the main chamber (MC). Together with the increase in efficiency by dilution of the mixture, nitrogen oxide (NO_X) emissions are lowered. However, at peak efficiencies, the NO_X emissions are still too high and require aftertreatment. The use of exhaust gas recirculation (EGR) as a dilutant might enable simple aftertreatment by using a three-way catalyst. This study experimentally investigates the use of EGR as a dilution method in a PCSI engine fueled by methane and analyzes the benefits and drawbacks compared to the use of air as a dilution method. The experimental results are categorized into three sets: measurements at wide open throttle (WOT) conditions, at a constant engine load of indicated mean effective pressure (IMEP) of 5 bar, and at IMEP = 7 bar, all at a fixed engine speed of 1600 rpm. The experimental results were further enhanced with numerical 1D/0D simulations to obtain parameters such as the residual combustion products and excess air ratio in the pre-chamber, which could not be directly measured during the experimental testing. The findings indicate that air dilution achieves higher indicated efficiency than EGR, at all operating conditions. However, EGR shows an increasing trend in indicated efficiency with the increase in EGR rates but is limited due to misfires. In both dilution approaches, at peak efficiencies, aftertreatment is required for exhaust gases because they are above the legal limit, but a significant decrease in NO_X emissions can be observed.

1. Introduction

Despite the growing trend of electrification of the transport sector, internal combustion engines will still be widely used in heavy machinery and heavy-duty transport. The main drawbacks of conventional internal combustion engines are high greenhouse gas emissions, e.g., carbon dioxide (CO₂), and pollutant emissions, including nitrogen oxide (NO_X), total hydrocarbons (THC), and particles, e.g., soot, with NO_X and soot emissions being particularly challenging for diesel-fueled engines [1]. These emissions must be reduced to meet the demands of increasingly stringent legislative requirements for internal combustion engines. Strategies such as combustion of lean mixtures [2], catalytic converters, exhaust gas recirculation [3], and the use of gaseous fuels [4] can effectively mitigate greenhouse gas and pollutant emissions.

Gaseous fuel that is widely used and considered an effective alternative fuel option is natural gas, whose main constituent is methane (CH₄) [5]. Natural gas is suitable for application in spark ignited (SI) engines because of its high resistance to excessive knock, compared to gasoline or propane (C₃H₈). High knock resistance means engines can operate with a high compression ratio and, as a result, achieve higher efficiency. The main advantage of methane in terms of exhaust emissions in comparison to liquid fuels is that it consists of only one carbon atom in a molecule. Therefore, it is less prone to agglomerating and forming particles and produces less carbon dioxide (CO₂) [6,7]. However, methane has higher ignition temperatures and slower combustion rates; hence, requires the implementation of alternative combustion approaches to improve ignitability and enhance flame propagation speed, such as pre-chamber spark ignition (PCSI) [6].

Pre-chamber spark ignition (PCSI) combustion simultaneously improves combustion efficiency, lowers emissions, enhances mixture ignitability, and shortens the combustion duration in spark-ignition engines, which makes it suitable for use with natural gas, as it effectively mitigates the fuel’s inherent limitations [8]. The active pre-chamber, i.e., the one with the additional pre-chamber injector, can operate with very diluted mixtures. Lean limit in the main chamber depends on the pre-chamber volume [9] and pre-chamber injection start timing and duration [10]. It has been concluded in [11] that by using lean mixtures (global excess air ratio higher than 1.6) in gasoline operated engines, indicated efficiency can be increased by 5% while decreasing nitrogen oxide (NO_X) emissions by 93% compared to stoichiometric mixtures, with the majority of the emissions being formed in the pre-chamber. However, when using liquid fuels, despite their high energy density, the increase in the pre-chamber fuel injection duration and pressure leads to the increase in particle number (PN) emission [12]. Since it has been shown that pre-chamber initiated combustion can be operated with various fuels [13] and it offers the potential to mitigate challenges of natural gas combustion, the use of natural gas in PCSI is of research interest.

It has been proven that PCSI combustion with methane ensures stable and fast combustion for lean conditions in the main chamber [14,15]. The performance of the PCSI engine operated with natural gas has been experimentally studied in [16,17]. While using a passive pre-chamber, the one without auxiliary fuel injection, does increase engine efficiency, the dilution of the engine operation is limited [16]. When operating with methane, a passive pre-chamber achieves a lean limit of global excess air ratio (measured in the exhaust) λ_global = 1.7, while an active pre-chamber extends the lean limit to λ_global = 2.9 [17]. Some studies on PCSI combustion with natural gas show that large flexibility in the initiation of charge ignition is possible with the appropriate amount of fuel in both the pre-chamber and the main chamber [18].

However, one of the bigger challenges with lean mixture operation in SI engines is the inefficiency of three-way catalysts regarding NO_X reduction caused by non-stoichiometric operation. Previous results obtained with gasoline operation [9] showed that peak efficiency is obtained at moderately lean mixtures where NO_X emissions are still high, while optimization of operating parameters with the aim of reducing NO_X emissions below legal limits resulted in an efficiency penalty. Therefore, to operate the engine at peak efficiency, one must employ NO_X aftertreatment systems. In lean conditions, NO_X can be lowered by lean NO_X traps or SCR catalytic converters. Both systems result in higher costs compared to a simple three-way catalyst, as the systems are more complex and they introduce additional operating costs, e.g., urea in SCR [19]. On the other hand, if the dilutant of the mixture is not air but exhaust gas recirculation (EGR), the usage of a three-way catalyst [20] might be enabled. Several studies with natural gas operated engines with passive pre-chambers and EGR have been conducted [21], and it has been concluded that combustion residuals in the pre-chamber cause too slow combustion, which leads to increased cycle-to-cycle variations and a tendency to misfire [22].

Numerical simulations conducted by the authors confirmed that an active pre-chamber engine fueled with methane using EGR dilution achieves higher indicated efficiency at operating conditions with EGR dilution compared to air dilution for the same maximum brake torque. It demonstrates the significant potential of the EGR dilution approach [19]. Furthermore, EGR reduces NO_X emissions up to 50%, while it slightly reduces CO₂ emissions and increases CO emissions due to lower combustion temperatures [23]. Since EGR slows down the combustion in [24], an active pre-chamber was used to extend EGR dilutant tolerance from 20% to over 30% when compared to conventional SI engine operated with gasoline. It has been shown that spark timing in TJI combustion obtained by using a pre-chamber is much less sensitive to increases in EGR than in SI combustion mode, due to the shorter ignition delay time, which at the same time results in higher combustion stability [8]. Furthermore, in [25], the use of an active pre-chamber increased the indicated efficiency and EGR tolerance rate compared to operation with a passive pre-chamber.

Previous studies have demonstrated the feasibility of PCSI combustion with methane combined with EGR. However, there is a lack of direct comparison between performance results obtained when dilution of mixture is made by air or EGR, where EGR operation is constrained by the air to fuel ratio being stoichiometric. This study aims to bridge that gap and investigates the influence of two different dilution approaches (air and EGR) on combustion, performance, and exhaust emissions in a PCSI engine fueled by methane. By directly comparing these two dilution approaches, it is possible to clarify the trade-offs between engine efficiency and emissions for each approach and suggest which dilution approach would be more favorable. The study is performed experimentally with a carefully designed experimental plan so that the results can be effectively compared. The design of experiments and expected operating range of the engine are based on results from the previous numerical study performed by the authors [20], and similar trends have now been confirmed experimentally.

2. Methods

This study presents an analysis of experimental results of a PCSI engine fueled by methane, operating with two different dilution strategies: dilution by air and dilution by exhaust gases. To complement the experimental findings and provide insights into the aspects that are not directly measurable, a numerical framework was employed [26]. The following sections detail the experimental setup, numerical modeling approach, and measurement methodologies.

2.1. Experimental Setup

The experimental measurements were performed at the Laboratory of IC Engines and Motor Vehicles of the Faculty of Mechanical Engineering and Naval Architecture at the University of Zagreb. The schematic layout of the experimental setup is shown in Figure 1.

The center of the setup is a highly modified air-cooled single cylinder Hatz 1D81Z engine that can operate in different combustion modes and is currently equipped with a modular active pre-chamber system. The compression ratio of the engine can be varied by using cylinder support rings and head gaskets of different thicknesses, which changes the distance between the piston top and cylinder head. Since the measured thickness in the unloaded state varies from the thickness when the gasket is mounted, the calculated uncertainty of the compression ratio obtained in this way at a compression ratio equal to 16 is 0.15. During the entire experimental measurement, the engine was not dismantled, and the cylinder head gasket was kept inside the whole time. Also, before each day of measurement, motorized pressure traces at predefined boundary conditions were taken to check the whole system, including the compression ratio. Therefore, even though there is some uncertainty regarding the exact value of the compression ratio, it was kept constant during the entire experiment. Specifications of the research engine are given in

Table 1. Specifications of the research engine.

Hatz 1D81Z
Engine type	4 stroke, air cooled
Displacement	667 cm3
Stroke	85 mm
Bore	100 mm
Connecting rod length	127 mm
Compression ratio	16
Number of valves	2
Inlet Valve Opens/Closes	36 °CA bTDC/60 °CA aBDC
Exhaust Valve Opens/Closes	54 °CA bBDC/21 °CA aTDC

.

The engine is fueled by compressed methane (CH₄), which is delivered with high pressure to the PC injector, and the pressure is adjusted by a two-stage pressure regulator at the methane tank. The port fuel injector operates at low pressure (an overpressure of 4 bar) that is maintained by a second inline two-stage pressure regulator. The total methane mass flow is measured by a Proline Promass A100 Coriolis flow meter (Endress+Hauser, Reinach, Switzerland) whose specifications are shown in

Table 2. Specifications of Proline Promass A100 Coriolis flow meter.

Parameter	Value
Measuring range	0 to 450 kg/h
Max. measurement error	Mass flow (liquid): ±0.1% Volume flow (liquid): ±0.1% Mass flow (gas): ±0.5% Density (liquid): ±0.0005 g/cm3
Medium temperature range	−50 to +205 °C

.

The active pre-chamber system consists of a spark plug and a pre-chamber injector and the modular system design that enables the change in pre-chamber geometry by the change in pre-chamber tip. The geometry parameters of the pre-chamber used in these experiments are given in

Table 3. Pre-chamber geometry.

Parameter	Value
Pre-chamber Volume (VPC)	1911 mm3
Throat diameter (dPC)	5 mm
Orifice diameter (dorf)	1.15 mm
Number of orifices (n)	6
Volume ratio $\left({\displaystyle \frac{V_{PC}}{V_{clr}}}\right)$	4.3%
Orifice area to volume ratio $\left({\displaystyle \frac{d_{orf}^2·\pi }{4}}·n\right)/V_{PC}$	0.033 cm−1

.

During experiments, the cylinder and the pre-chamber pressure, intake and exhaust pressure and temperatures, fuel and air mass flows, and emissions of CO, THC, NO_X, and soot were measured. THC emissions were measured with the Environnement Graphite 52M heated flame ionization detector (HFID) analyzer (Envea, Reims, France), CO and CO₂ were measured with the Environnement MIR 2M non-dispersive infrared (NDIR) analyzer (Envea, Reims, France), while NO_X emissions were measured by the ECM NOx 5210t device (ECM, Los Altos, CA, USA). The measurement is conducted by sampling low speed data of each operating point at a frequency of 1 Hz over a duration of 60 s, and then averaging the results. For high-speed data (crank angle resolved data, e.g., cylinder pressure), the pressure profiles are sampled with a frequency of 0.5 °CA during the whole cycle and 0.1 °CA during combustion over a period of cycles and then cycle averaged. The method was previously analyzed [27], and measurement uncertainty of 0.48% points for indicated efficiency is calculated using standardized methods. Detailed description of the experimental equipment can be found in [28]. Additionally, soot content was measured by the AVL Smoke Meter (AVL, Graz, Austria), and its technical data is given in

Table 4. Technical data of AVL Smoke Meter.

Measurement principle	Filter paper blackening
Measured value output	FSN (filter smoke number) mg/m3 (soot concentration)
Measurement range	0 to 10 FSN
Detection limit	0.002 FSN or 0.02 mg/m3

.

The EGR percentage is determined by measuring the CO₂ concentration at the engine intake and exhaust. The ratio of the recirculated CO₂ volume fraction at the intake to the CO₂ volume fraction at the exhaust corresponds to the total fraction of recirculated exhaust gas. The formula for calculating the EGR fraction is given in the following expression:

$$\mathrm{E}\mathrm{G}\mathrm{R}\mathrm{ }\left(\mathrm{\%}\right)={\displaystyle \frac{{\left[{\mathrm{C}\mathrm{O}}_2\right]}_{\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{k}\mathrm{e}}-{\left[{\mathrm{C}\mathrm{O}}_2\right]}_{{\mathrm{C}\mathrm{O}}_2\_\mathrm{a}\mathrm{m}\mathrm{b}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t}}}{{\left[{\mathrm{C}\mathrm{O}}_2\right]}_{\mathrm{e}\mathrm{x}\mathrm{h}\mathrm{a}\mathrm{u}\mathrm{s}\mathrm{t}}-{\left[{\mathrm{C}\mathrm{O}}_2\right]}_{{\mathrm{C}\mathrm{O}}_2\_\mathrm{a}\mathrm{m}\mathrm{b}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t}}}}·100$$

2.2. Numerical Framework

Figure 2 shows the concept of the numerical framework that was used in this study and developed by the authors [26].

The numerical framework consists of a reduced 0D/1D model made in AVL Boost^TM v2013.2, extensively validated based on experimental results and 3D-CFD validations in previous studies [20]. The boundary conditions that are imposed on the numerical simulations are the intake pressure profile and intake temperature, exhaust pressure and temperature, cylinder head temperature along with the operating parameters such as spark timing, fuel flow, etc.

The numerical simulations in this analysis were used to provide better insight into certain results that cannot be measured directly. In this case the reduced 0D/1D model made in AVL Boost^TM was employed to provide details of mixture formation in the pre-chamber, e.g., the values of excess air ratio and residual gases in the pre-chamber before and during combustion. Each operating point was calibrated individually by adjusting model parameters (such as fuel flow, heat transfer, and temperatures, etc.) to match the experimental main chamber and pre-chamber pressure profiles. An example of the calibrated main- and pre-chamber pressure profiles for an experimental operating point is shown in Figure 3.

2.3. Design of Experiments

The experimental engine’s compression ratio (CR) is set at CR = 16.0, based on findings from the previous study [20], although the experimental procedure and conditions applied here differ substantially from those used in that work. All measurements were conducted at an engine speed of 1600 rpm, allowing for comparison with previous experiments conducted on the same setup. The intake air temperature was set to 33 °C by using the intake air heater.

The experiments are divided into 3 sets with the aim of comparing the performance results obtained by two different mixture dilution approaches: dilution by air and dilution by EGR.

The first set of experiments was conducted at wide open throttle (WOT) conditions. Port fuel injection was performed with the start of injection at 190 °CA bTDC for all cases, while the pre-chamber injection was performed with an overpressure of 50 bar, and the start of injection was mostly kept at 120 °CA bTDC, but was adjusted to earlier phases for higher levels of EGR to improve pre-chamber scavenging and achieve stable combustion. For the air dilution approach, an excess air ratio sweep was performed from λ_global = 1.0 to λ_global = 1.8, and for each level of λ_global, a spark sweep was performed. λ_global is used to denote the average excess air ratio of the main chamber and the pre-chamber mixture measured at the exhaust pipe. The pre-chamber injection duration was adjusted to achieve stable combustion. For the EGR dilution approach, the excess air ratio was kept at λ_global = 1.0 while the EGR level was gradually increased from EGR = 0% to EGR = 19.9%. For each level of EGR, a spark sweep was performed. The pre-chamber injection duration and injection start were adjusted to improve PC scavenging and achieve stable combustion. To ensure optimal EGR flow to the intake, the exhaust pressure was increased by partially closing the valve at the exhaust pipe.

The second set of experiments was conducted at a constant engine load of IMEP = 5 bar. The engine load of IMEP = 5 bar was selected based on the results from the first set of experiments, which enables a wide range of operating points. The engine load was kept constant by throttling the intake air. Port fuel injection was performed with a start of injection at 190 °CA bTDC for all cases, while the pre-chamber injection was performed with an overpressure of 50 bar, and the start of injection was mostly kept at 120 °CA bTDC. However, it was moved to earlier phases for higher levels of EGR to improve pre-chamber scavenging and achieve stable combustion. For the air dilution approach, an excess air ratio sweep was performed from λ_global = 1.0 to λ_global = 1.7, and for each level of λ_global, a spark sweep was performed. The pre-chamber injection duration was adjusted to achieve stable combustion. For the EGR dilution approach, the excess air ratio was kept at λ_global = 1.0 while the EGR level was gradually increased from EGR = 0% to EGR = 18.9%. For each level of EGR, a spark sweep was performed. The pre-chamber injection duration and injection start were adjusted to improve PC scavenging and achieve stable combustion. In this case the exhaust backpressure valve was fully open because good EGR flow was already achieved with the lower intake pressure caused by throttling.

The third set of experiments was conducted at a constant engine load of IMEP = 7 bar. The engine load was kept constant by boosting the intake pressure for the air dilution approach and by throttling the intake air for the EGR dilution approach. In the air dilution approach, boosting was applied in order to achieve very lean mixtures, whereas in the EGR dilution approach, throttling was necessary to enable high EGR rates, i.e., a large amount of recirculated exhaust gas with a reduced intake air fraction. Again, port fuel injection was performed with a start of injection of 190 °CA bTDC for all cases, while the pre-chamber injection was performed with a pressure of 50 bar, and the start of injection was mostly kept at 120 °CA bTDC but was moved to earlier phases for higher levels of EGR to improve pre-chamber scavenging and achieve stable combustion. The start of injection both into the port and pre-chamber was set based on conclusions from the literature, as a compromise between providing sufficient time for mixture preparation in the pre-chamber and preventing injected fuel from escaping into the main chamber. PC DOI, on the other hand, is determined by the required fuel mass to be injected into the pre-chamber, since the injector is calibrated and its characteristic provides the relation between injection duration and injected fuel mass. For the air dilution approach, an excess air ratio sweep was performed from λ_global = 1.0 to λ_global = 1.7, and for each level of λ_global, a spark sweep was performed. The pre-chamber injection duration was adjusted to achieve stable combustion. For the EGR dilution approach, the excess air ratio was kept at λ_global = 1.0 while the EGR level was gradually increased from EGR = 0% to EGR = 20.1%. For each level of EGR, a spark sweep was performed. The pre-chamber injection duration and injection start were adjusted to improve PC scavenging and achieve stable combustion. In this set the exhaust backpressure valve was fully open because good EGR flow was already achieved with the lower intake pressure caused by throttling.

3. Results

The Section 3 is split into three subsections. The first subsection is an overview of the operating points and the analysis of the first set of experiments with wide open throttle. The second subsection is an overview of the operating points and the analysis of the second set of experiments conducted at the same engine load of IMEP = 5 bar. The third and last subsection is an overview of the operating points and the analysis of the third set of experiments conducted at the same engine load of IMEP = 7 bar.

3.1. First Set of Results—WOT Conditions

The operating points from the first set of experiments for the air dilution approach with the highest indicated efficiencies are given in

Table 5. First set of experimental operating points—air dilution.

Operating Point	IMEP [bar]	λglobal [-]	ST [°CA bTDC]	PC DOI [ms]	Intake Pressure [bar]
OP1	7.91	0.99	−6	0.40	1.03
OP2	7.13	1.20	−4	0.55	1.03
OP3	6.83	1.29	−2	0.70	1.03
OP4	6.48	1.39	2	0.45	1.03
OP5	6.10	1.49	2	0.50	1.02
OP6	5.63	1.59	6	0.29	1.02
OP7	4.96	1.69	6	0.50	1.03
OP8	4.09	1.79	8	0.80	1.02

, where the indicated efficiency is defined as the ratio of the indicated work to the fuel’s energy input.

Although, the lean limit is not at λ_global = 1.8, the mixture was not further diluted because of a significant drop in combustion efficiency.

The operating points from the first set of experiments for the EGR dilution approach with the highest indicated efficiencies are given in

Table 6. First set of experimental operating points—EGR dilution.

Operating Point	IMEP [bar]	λglobal [-]	ST [°CA bTDC]	PC DOI [ms]	EGR [%]	Intake Pressure [bar]
OP9	8.12	1.00	−6	0.30	0.0	1.03
OP10	8.01	1.00	−2	0.30	4.4	1.04
OP11	7.73	0.99	2	0.30	9.9	1.05
OP12	7.48	1.00	6	0.30	15.2	1.05
OP13	7.09	1.00	16	0.30	19.9	1.05

.

EGR level was increased to the highest possible level, which still enabled operation without misfires.

Figure 4 shows the comparison of performance results of MBT operating points for air and EGR dilution. The obtained results of indicated efficiencies, combustion efficiencies, CoV (IMEP), CA50, and IMEP are plotted against the excess air ratio and EGR level, respectively.

Using the air dilution approach, the highest indicated efficiency of 39.5% is achieved at λ_global = 1.4. In contrast, the EGR approach reaches its peak indicated efficiency of 39.4% at the maximum EGR level of 19.9%. A clear trend of indicated efficiency increase with the increase in EGR can be observed. Results of the combustion efficiency show that the EGR approach consistently maintains high combustion efficiency (above 95%) across all operating points. In contrast, the air dilution approach experiences a significant drop in combustion efficiency from λ_global = 1.7 and results in combustion efficiency of 77.0% at λ_global = 1.8. This is the reason why the mixture was not diluted further, as the operation with combustion efficiency below 77% does not provide useful operating conditions. Both approaches exhibit high combustion stability (CoV (IMEP) below 2%). The EGR approach shows a constant increase in indicated efficiency with the increase in EGR and high combustion stability, suggesting that the EGR level should be further increased to potentially enhance the indicated efficiency. However, further increase in EGR was not feasible due to sudden misfires occurring at the next EGR level. The misfire issue will be further elaborated in a later subsection.

With the air dilution approach for the peak indicated efficiency conditions (λ_global = 1.4), the CA50 values are in the expected optimal range (CA50 = 10° aTDC). For the close to stoichiometric conditions, due to extreme knock tendency, the combustion is delayed and is far from the optimal range, while for the lean mixtures, the combustion phase is advanced due to slow and weak combustion. In the EGR dilution approach, the points with high levels of EGR had CA50 values close to the optimal range because EGR lowers the knock tendency and enables the operation with advanced spark timings. An increase in EGR rate results in lower exhaust temperatures (from 629 °C at EGR = 0% to 513 °C at EGR = 19.9%) and in combustion temperatures, which then increases indicated efficiency. Similarly, in the air dilution approach, exhaust temperature decreases due to the increase in the excess air ratio; however, indicated efficiency declines with air dilution higher than λ_global = 1.4 because of the incomplete combustion and deviation from the optimal CA50 values. Incomplete combustion leads to unstable combustion, which is observed by an increase in CoV (IMEP), particularly as the excess air ratio becomes larger than 1.6 (λ_global > 1.6). It should be noted that the pre-chamber duration of injection (PC DOI) was adjusted to ensure stable combustion in the air dilution approach. In the EGR dilution approach, the PC DOI was kept constant because the mixture in the main chamber was predominantly stoichiometric, which primarily influenced the occurrence of misfires, and modifying the PC DOI had little effect. To reduce misfires, pre-chamber purging should be applied.

When comparing indicated efficiencies of the air dilution and EGR dilution approaches at operating points with the same IMEP of 7.1 bar, the EGR approach shows 3.4% points higher indicated efficiency, achieving 39.4% at an EGR rate of 19.9%, compared to 36% at an excess air ratio of λ_global = 1.2. It has to be noted that for IMEP = 7.1, the excess air ratio in the air dilution case is not optimal, so the difference between air and EGR dilution might be caused by non-optimal operation. Therefore, the following two sets of experiments were performed. From the first set of experiments, it could be concluded that the operation with EGR dilution while maintaining λ_global = 1.0, is possible and that in this setup the highest EGR level was around 20% when misfires started to occur.

Table 7. Excess air ration and combustion products concentration at ST in the pre-chamber—WOT (numerical results).

Operating Point	λPC @ST [-]	PC Comb. Products@ST [%]	Operating Point	λPC @ST [-]	PC Comb. Products@ST [%]
OP1	0.68	5.8	OP9	0.72	5.6
OP2	0.69	5.6	OP10	0.81	10.5
OP3	0.70	5.2	OP11	0.82	14.3
OP4	0.84	5.4	OP12	0.79	18.0
OP5	0.82	5.4	OP13	0.73	21.4
OP6	1.11	5.4
OP7	0.87	5.7
OP8	0.79	5.9

contains the values of the pre-chamber excess air ratio and combustion products concentration at spark timing, as obtained by the 0D/1D simulations.

The PC mixture is rich for all operating points (except a slightly lean mixture at OP6) to ensure reliable ignition and stable combustion. Even for the stoichiometric cases, a small amount of fuel had to be injected into the PC to improve scavenging and ensure stable combustion. The PC combustion products concentration for the air dilution approach is at standard and acceptable levels, while in the EGR dilution approach, the increase in EGR causes a gradual increase in the combustion products concentration in the PC, which then leads to misfires and limits further increase in EGR.

Figure 5 shows the comparison of exhaust emission results of MBT operating points for air and EGR dilution. The obtained results of THC, CO, and NO_X emissions are plotted against the excess air ratio and EGR level, respectively.

With the air dilution approach, THC and CO emissions increase with the increase in the excess air ratio due to the drop in combustion efficiency, with CO emissions having their first significant drop when moving from a stoichiometric mixture (λ_global = 1.0) to lean conditions. NO_X emissions are decreasing with the increase in the excess air ratio due to lower combustion temperatures. With the EGR dilution approach, THC emissions increase with the increase in EGR, while the CO emissions with some EGR are higher than the operating point with EGR = 0% and are at similar levels as in air dilution with stoichiometric conditions. By comparing the operating points with the highest indicated efficiencies of both approaches, the THC and NO_X emissions are at similar levels, but the EGR approach has higher CO emissions.

Figure 6 shows the soot emissions measured with the AVL Smoke Meter for the air and EGR dilution approaches. Both approaches have very low soot emissions, shown as filter smoke number (FSN), and are at the lower limit of the AVL Smoke Meter measuring range. Soot emissions are negligible for both dilution approaches because of the use of methane as a fuel. Since the soot emissions are very low and are such in all following experiments, they will not be shown for the two following sets of results.

On first note, the comparison of results presented here with the results in the authors’ previous numerical study [20], might result in the conclusion that there is a difference in results and conclusions. However, the difference in the results of the numerical study arises primarily from differences in the experimental and simulation conditions. In the prior study, the total fuel mass was kept constant (with variations of less than 4%), and the pre-chamber excess air ratio was fixed at λ_PC = 1.0. The excess air ratio in EGR cases was not set to stoichiometric conditions but changed as the EGR was increased. On the other hand, in the experiments described in this study for the EGR dilution approach, the excess air ratio was kept constant and stoichiometric to enable the use of the three-way catalyst. Also, the λ_PC was adjusted for each operating point to ensure stable combustion. As a result, the pre-chamber excess air ratio in all experimental operating points was below 1.0, meaning more fuel was delivered to the pre-chamber, which directly affects indicated efficiency.

However, the experimental results from the experimental study are comparable to the simulation results from the numerical study [20] for one specific case of CR = 16, λ_global = 1.2 at EGR = 0% and EGR = 20%, where λ_global = 1.0 was achieved in the EGR dilution case. For this simulated operating point, with 20% EGR, the indicated efficiency exceeded 38%, while for λ_global = 1.2 at EGR = 0%, it was slightly above 36% (a 2 percentage points difference). In comparison, the experimentally obtained indicated efficiency was 36.1% at EGR = 0% and 39.4% at EGR = 20% (a 3.3 percentage point difference) under WOT conditions and IMEP = 7 bar, which is in reasonable agreement with the simulation results.

It should also be noted that the previous simulations were performed using a 0D/1D approach, which cannot predict misfires, allowing the model to explore higher EGR rates than are achievable experimentally. These factors collectively explain the observed differences between the numerical predictions and the experimental results.

3.2. Second Set of Results—IMEP = 5 Bar

The operating points from the second set of experiments for the air dilution approach with the highest indicated efficiencies are given in

Table 8. Second set of operating points—air dilution (experimental results).

Operating Point	IMEP [bar]	λglobal [-]	ST [°CA bTDC]	PC DOI [ms]	Intake Pressure [bar]
OP14	5.05	1.00	4	0.40	0.69
OP15	5.11	1.10	4	0.40	0.74
OP16	5.02	1.20	5	0.40	0.77
OP17	5.14	1.30	3	0.60	0.81
OP18	5.08	1.40	4	0.65	0.83
OP19	5.00	1.50	6	0.65	0.86
OP20	5.15	1.60	6	0.65	0.93
OP21	5.05	1.66	8	0.65	0.96

.

The operating points from the second set of experiments for the EGR dilution approach with the highest indicated efficiencies are given in

Table 9. Second set of operating points—EGR dilution (experimental results).

Operating Point	IMEP [bar]	λglobal [-]	ST [°CA bTDC]	PC DOI [ms]	EGR [%]	Intake Pressure [bar]
OP22	4.98	1.00	2	0.30	0.0	0.72
OP23	5.06	1.01	4	0.30	5.9	0.76
OP24	5.04	1.01	4	0.30	10.5	0.77
OP25	4.99	1.00	12	0.30	15.3	0.8
OP26	5.13	1.02	16	0.15	18.9	0.83

.

Figure 7 shows the comparison of performance results of MBT operating points of air and EGR dilution at the same engine load of IMEP = 5 bar. The obtained results of indicated efficiencies, combustion efficiencies, CoV (IMEP), and CA50 are plotted against the excess air ratio and EGR level, respectively. One can note that the base operating point with λ_global = 1.0 and EGR = 0%, which is basically the same on left and right sides of the results, shows a slight difference. The difference is because the results of air and EGR dilution were obtained on different days and the atmospheric conditions changed. Therefore, in the discussion, the conclusions regarding indicated efficiencies are made from observations of relative rather than absolute values, with relative value being the percentage change compared to base operating point of that day.

The air dilution approach has a higher increase in indicated efficiency than the EGR approach. The highest indicated efficiency in the air dilution approach is achieved at λ_global = 1.6 and equals 38%, which is an increase of 13.7%, while in the EGR dilution approach, the highest indicated efficiency is achieved at the highest possible level of EGR (18.9%) and equals 36%, which is an increase of 10.1%. Combustion efficiency is high (above 95%) for all operating points in the EGR dilution approach, while the air dilution approach has a significant drop in combustion efficiency above λ_global = 1.6 (below 90%). Both approaches have high combustion stability through all operating points. Similarly to operation with WOT, the EGR approach shows high combustion stability at the highest EGR level, which would suggest the EGR level could further be increased and potentially increase the indicated efficiency. However, this was not possible because of sudden misfires, which occur in the same manner as in WOT conditions.

To show an example of a misfire at high levels of EGR, in Figure 8 the IMEP of 300 consecutive cycles is plotted for one operating point with a high EGR level. As can be seen, the cyclic variation in IMEP is low, which corresponds to the results of previously shown CoV (IMEP), but a sudden misfire occurs because of the high level of combustion products in the PC, which results in a misfire in the PC and then a lack of combustion in the main chamber.

The occurrence of misfire depends on high EGR ratios and cannot be corrected by changing the PC DOI. Figure 9 shows two cases with the same EGR percentage but different PC DOI values (0.8 and 0.4 ms). The variation in PC DOI did not lead to a stable operation without misfires.

In

Table 10. Excess air ration and combustion products concentration at ST in the pre-chamber—IMEP = 5 bar (numerical results).

Operating Point	λPC @ST [-]	PC Comb. Products@ST [%]	Operating Point	λPC @ST [-]	PC Comb. Products@ST [%]
P14	0.63	8.0	OP22	0.77	8.0
OP15	0.71	7.8	OP23	0.77	12.7
OP16	0.76	7.5	OP24	0.76	15.8
OP17	0.72	6.6	OP25	0.70	19.5
OP18	0.74	6.4	OP26	0.69	21.2
OP19	0.78	6.3
OP20	0.82	6.1
OP21	0.83	5.9

, the values of the pre-chamber excess air ratio and combustion products concentration at spark timing obtained from the 0D/1D simulations are given.

Similarly to the WOT conditions, the PC mixture is rich for all operating points with intake throttling. In the air dilution approach, the PC combustion products concentration is at acceptable levels, while the increase in EGR causes a gradual increase in the PC combustion products concentration. One can note that a similar value of the PC combustion products concentration at the borderline case is obtained although the overall EGR level is not the same.

Figure 10 shows the comparison of exhaust emission results of MBT operating points of air and EGR dilution at the same engine load of IMEP = 5 bar. The obtained results of THC, CO, and NO_X emission are, again, shown as absolute, but discussed as relative values to the reference operating points (λ_global = 1.0 and EGR = 0%) to compensate for differences in environmental conditions.

In the air dilution approach, as expected, THC and CO emissions increase with the increase in excess air ratio due to the decrease in combustion efficiency, with a significant decrease in CO emissions when combustion is moved from stoichiometric conditions. The NO_X emissions are decreasing with the increase in excess air ratio due to lower combustion temperatures, having a decrease of 94.6% at peak efficiency and being close to the limit of 400 mg/kWh. With the EGR dilution approach, THC emissions have a slight increase with the increase in EGR, while the CO emissions are generally decreasing. The air dilution approach causes a much bigger increase in THC compared to EGR dilution approach due to the more significant decrease in combustion efficiency. The NO_X emissions are decreasing with the increase in EGR due to lower combustion temperatures. However, a decrease of NO_X emission is lower at peak efficiency in EGR dilution compared to air dilution approach.

According to given results, only operating points with air dilution at λ ≥ 1.6 are below permitted NO_X limitations of the Euro VI (2013) [29] stage for heavy-duty compression–ignition (CI) diesel engines under steady-state test conditions, while EGR dilution does not achieve NO_X emissions below that limit. At IMEP = 5 bar, it can be concluded that air dilution brings greater advantage in indicated efficiency and raw NO_X emissions at excess air ratio λ = 1.6 compared to the EGR dilution method at peak EGR level. Significantly higher THC emissions for the air dilution method could be reduced by using an oxidation catalyst. On the other hand, the NO_X, CO, and THC emissions in EGR dilution could be reduced by the application of a three-way catalyst. There is a potential in the EGR dilution approach for an even higher increase in indicated efficiency if some form of purging of the PC were to be performed.

3.3. Third Set of Results—IMEP = 7 Bar

The operating points from the third set of experiments for the air dilution approach with the highest indicated efficiencies are given in

Table 11. Third set of operating points—air dilution (experimental results).

Operating Point	IMEP [bar]	λglobal [-]	ST [°CA bTDC]	PC DOI [ms]	Intake Pressure [bar]
OP27	7.04	1.01	−6	0.30	0.94
OP28	7.01	1.09	−4	0.30	0.97
OP29	7.04	1.20	−4	0.40	1.06
OP30	6.92	1.31	−2	0.40	1.08
OP31	7.00	1.40	0	0.40	1.11
OP32	7.03	1.51	0	0.40	1.17
OP33	7.07	1.61	2	0.40	1.22
OP34	6.96	1.70	4	0.40	1.26

.

The operating points from the third set of experiments for the EGR dilution approach with the highest indicated efficiencies are given in

Table 12. Third set of operating points—EGR dilution (experimental results).

Operating Point	IMEP [bar]	λglobal [-]	ST [°CA bTDC]	PC DOI [ms]	EGR [%]	Intake Pressure [bar]
OP35	7.10	1.00	−4	0.30	0.0	0.92
OP36	7.04	1.01	0	0.30	6.3	0.93
OP37	6.99	1.00	4	0.30	10.1	0.95
OP38	7.01	1.00	10	0.20	15.2	0.96
OP39	7.00	1.00	20	0.20	20.1	0.99

.

Figure 11 shows the comparison of performance results of MBT operating points of air and EGR dilution at the same engine load of IMEP = 7 bar. The obtained results of indicated efficiencies, combustion efficiencies, CoV (IMEP), and CA50 are plotted against the excess air ratio and EGR level, respectively. The values of indicated efficiencies are, again, shown as absolute, but discussed in relative terms with respect to the reference operating points (λ_global = 1.0 and EGR = 0%).

Again, the air dilution approach has a higher increase in indicated efficiency than the EGR approach. The highest indicated efficiency in the air dilution approach is achieved at λ_global = 1.6 and equals 38.1%, which is an increase of 11.5%, while in EGR dilution approach, the highest indicated efficiency is achieved at the highest possible level of EGR (20.1%) and equals 38.3%, which is an increase of 9.6%. By looking at the combustion efficiency, the EGR approach has again high combustion efficiency (above 95%) for all operating points, while the air dilution approach has a drop in combustion efficiency with the increase in λ; however, at λ_global = 1.7, it is still above 90%. Both approaches have high combustion stability. Similarly to previous operations, the EGR approach shows stable combustion by CoV IMEP measure, but further increases in EGR are limited due to sudden misfires, which occur because of high combustion product concentrations in the PC.

In

Table 13. Excess air ration and combustion products concentration at ST in the pre-chamber—IMEP = 7 bar (numerical results).

Operating Point	λPC @ST [-]	PC Comb. Products@ST [%]	Operating Point	λPC @ST [-]	PC Comb. Products@ST [%]
OP14	0.82	6.3	OP22	0.77	6.8
OP15	0.96	6.4	OP23	0.76	11.8
OP16	0.82	7.0	OP24	0.74	14.9
OP17	0.86	7.2	OP25	0.71	17.7
OP18	0.92	7.2	OP26	0.63	20.8
OP19	0.96	7.1
OP20	1.00	7.0
OP21	1.01	7.1

, the values of the pre-chamber excess air ratio and combustion products concentration at spark timing obtained from the 0D/1D simulations are given.

Like in the first two sets of results, the PC mixture is rich for all operating points. The PC combustion products concentration for the air dilution approach is, again, at acceptable levels, while the increase in EGR causes a gradual increase in the PC combustion products concentration. While at IMEP = 5 bar, all operating points in the air dilution approach are richer than λ_PC@ST < 0.83; at IMEP = 7 bar, all measured points operate at λ_PC@ST > 0.82. This difference arises from the delayed spark timing and slightly shorter duration of injection in the pre-chamber, resulting in a leaner mixture in the pre-chamber at spark timing. CA50 is delayed in both the EGR and air dilution approaches compared to combustion at IMEP = 5 bar, due to the delayed spark timing used to manage the fast combustion rate and high knock tendency.

For the reference operating points at IMEP = 7 bar, the combustion is more rapid, and a greater knock tendency is obtained compared to IMEP = 5 bar. Therefore, the low air or EGR dilution cases have delayed spark timing and delayed CA50, which is far from optimal. The increase in dilution enables advancement of CA50 and obtaining CA50 closer to optimal; however, in the air dilution case at peak efficiency, the CA50 is still slightly delayed.

Figure 12 shows the comparison of exhaust emission results of MBT operating points of air and EGR dilution at the same engine load of IMEP = 7 bar. The obtained results of THC, CO, and NO_X emissions are, again, shown in absolute values, but discussed in relative terms with respect to the reference operating points.

In the air dilution approach, in contrast to IMEP = 5 bar, set the THC emissions first slightly decrease with the increase in excess air ratio, and then with the excess air ratio higher than λ_global = 1.4, the THC emissions increase due to the decrease in combustion efficiency. The CO emissions are lower than the emission at the reference operating point for the whole excess air ratio sweep. The NO_X emissions are decreasing with the increase in excess air ratio starting with λ_global = 1.3 due to lower combustion temperatures. With the EGR dilution approach, THC emissions increase with the increase in EGR, while CO emissions do not have a clear trend. The NO_X emissions are decreasing with the increase in EGR, as was expected. By comparing the operating points with the highest indicated efficiencies of both approaches, the CO emissions are much lower for the air dilution approach than for the EGR approach; however, the EGR approach has 1.4 times lower THC emissions, and the air dilution approach has a slightly greater decrease in NO_X emissions of 0.9% points, compared to the EGR approach.

Although both air and EGR dilution achieve substantial reductions in NO_X emissions, neither approach meets the permitted NO_X limit according to the permitted limit of the Euro VI (2013) [29] stage for heavy-duty compression-ignition (CI) diesel engines under steady-state test conditions. However, in the EGR dilution approach, the fact that the operation is stoichiometric enables the effective usage of the three-way catalyst to reduce NO_X emissions while maintaining high IMEP and ensuring high indicated efficiency.

4. Conclusions

An experimental investigation, enhanced with numerical analysis, of the air and EGR dilution methods in a PCSI engine fueled by methane is performed to investigate the pros and cons of each dilution approach. Also, the comparison of overall engine performance and emissions at the same operating conditions is made. The following conclusions can be drawn from this study:

• Overall, by comparing the two dilution approaches, the highest indicated efficiencies are achieved with the air dilution approach, at all operating conditions, at WOT conditions and at the constant engine loads of IMEP = 5 bar and IMEP = 7 bar. However, EGR shows an increasing trend in indicated efficiency with higher EGR rates. Misfires currently limit the assessment of the full potential of the EGR dilution principle. EGR could potentially provide higher efficiency if the misfire issue were resolved, for example, through pre-chamber purging.
• In air dilution approach the increase in indicated efficiency by dilution is at WOT conditions 16% (at λ_global = 1.4), at the engine load of IMEP = 5 bar the increase is 13.7% (at λ_global = 1.6), and at the engine load of IMEP = 7 bar the increase is 11.5% (at λ_global = 1.6).
• In EGR dilution approach, there is a clear trend of the increase in indicated efficiency with the increase in EGR with peak indicated efficiency increased by, 9.1% at WOT conditions (EGR = 19.9%), at the engine load of IMEP = 5 bar the increase is 10.1% (EGR = 18.9%), and at the engine load of IMEP = 7 bar the increase is 9.6% (EGR = 20.1%).
• The decrease in combustion efficiency at higher excess air ratios is the main cause of the decrease in indicated efficiency for higher dilution levels in air dilution approach. On the other hand, sudden misfires prevented further increase in EGR level to improve efficiency, despite stable combustion in terms of CoV IMEP. Misfires are caused by lack of ignition in pre-chamber most probably caused by high rates of combustion products in the PC at spark timing. A possible solution for that would be the addition of air purging to the PC to improve scavenging.
• The EGR dilution has a stronger suppression of knock compared to air dilution, which manifests in lower EGR dilution being able to run at more advanced CA50s at higher load cases (IMEP = 7 bar).
• In both dilution approaches, the increase in dilution results in increased THC emission; however, at peak indicated efficiency, the increase in THC emission of the air dilution approach is much higher. Both operations require aftertreatment, as the raw emissions are well above legal limits.
• In air dilution, the CO emission is lower at diluted mixture compared to stoichiometric operation; however, at peak indicated efficiency operation, aftertreatment is required. In the EGR dilution approach, the CO emission trends are not clear, but aftertreatment is required as well. Total CO emission values are similar for both dilution approaches.
• The NO_X reduction for the operating points with peak efficiencies and different dilution approaches is similar for IMEP = 7 bar, while the reduction is much higher for the air dilution than the EGR dilution approach at peak efficiency for IMEP = 5 bar. In the air dilution approach at peak efficiency for IMEP = 5 bar, the NO_X levels are close to legal limits; however, at IMEP = 7 bar, the emissions are well above the legal limit and would require aftertreatment as well as in the EGR dilution approach.

By comparing the air and EGR dilution approaches the peak indicated efficiency in air dilution is higher by 3.6% at lower load and 1.9% at higher load but requires a more expensive and more complex aftertreatment system. In the air dilution approach, peak efficiency was achieved, and there is no more potential for increase, while the EGR dilution approach has potential for even higher indicated efficiency, but requires some form of pre-chamber air purging that would lower the amount of combustion products in the pre-chamber and stop misfires.