Assessment of Charge Dilution Strategies to Reduce Fuel Consumption in Natural Gas-Fuelled Heavy-Duty Spark Ignition Engines

Abstract

The need to decarbonize the road transport sector is driving the evaluation of alternative solutions. From a long-term perspective, biomethane and e-methane are particularly attractive as green energy carriers and a part of the solutions for the sustainable freight on-road transport, as they offer significant CO₂-equivalent emissions savings in a net Well-to-Wheel assessment. However, to make methane-fuelled spark ignition (SI) heavy-duty (HD) engines competitive in the market, their efficiency must be comparable to the top-performing diesel applications that dominate the sector. To this end, dilution techniques such as exhaust gas recirculation (EGR) or lean air–fuel mixtures represent promising solutions. Within limits specific to the engine’s tolerance to the used strategy, charge dilution can improve thermal efficiency impact on the pumping and wall heat loss, and the heat capacity ratio (γ). However, their potential has never been explored in the case of methane SI HD engines characterized by a semi diesel-like combustion system architecture. This work presents an experimental study to characterize the energy and pollutant emission performance of a state-of-the-art SI HD gas single-cylinder engine (SCE) operating with EGR or with lean conditions. The engine type is representative of most HD powertrains used for long-haul purposes. The designed test plan is representative of the majority of on-road operating conditions providing an overview of the impact of the two dilution methods on the overall engine performance. The results highlight that both techniques are effective for achieving significant fuel savings, with lean combustion being more tolerable and yielding higher efficiency improvements (10% peak vs. 5% with EGR).

1. Introduction

Nowadays, the promotion of green energy and technologies is fundamental to decrease the greenhouse gas (GHG) emissions caused by the road transport sector. A mix of solutions will probably be the answer in the long-term scenario. In fact, the scientific community is focusing on various alternatives to traditional fuels for feeding internal combustion engines, with particular attention on hydrogen, both in its pure form [1] and in dual-fuel applications [2]. Biofuels have strong and unexplored potential to play a key role in this context of uncertainty. Second-generation bio-fuels use biomass as a primary source (in the EU, mainly municipal and agricultural wastes and special crops on infertile soils), without impacting the food chain. Biomass captures CO₂ throughout its lifetime (storing a “bio-credit”), and by burning its by-products, just a part of it is re-emitted in the air, leading to significant benefits in the net GHG emissions balance [3]. The environmental impact of each specific bio-fuel depends on the biomass used as source as well as on the processes for its conversion, and the advantages can be seen in a Well-to-Wheel (WtW) analysis (from energy source supply to on-road emissions). Compressed and liquified bio-methane (CBM and LBM) stands out among them; its WtW emission factor is about 90% lower than the traditional fuels. LBM can be considered practically equivalent from a chemical point of view to high-purity liquified natural gas (LNG) [4,5] that has the highest hydrogen-to-carbon (H/C) ratio among hydrocarbons used on the road. It is made mainly from methane (CH₄) and its combustion leads to lower CO₂eq production for the same amount of energy produced (lower emission factor compared to gasoline and diesel) [6], underlining that LBM is a valid alternative in the mid/long-term scenario to reduce both Tank-to-Wheel and WtW emissions. NG is specifically suitable for HD SI engine configurations for long-haul applications since it has a high motor octane number (MON) [7], leading to a lower risk of knock (abnormal early combustion of the end-gas) compared to petroleum fuels and to the possibility of using higher compression ratios to enhance efficiency [8]. Moreover, stoichiometric SI engines allow for the use of the three-way catalyst (TWC) to abate the noxious gaseous emissions, including unburned CH₄, that is particularly impactful as its global warming potential (GWP) is 21 times higher than CO₂ over a 100-year period [9].

The freight transport sector, in which long-haul vehicles represent a relevant fraction of the transport demand, is a so-called hard to abate sector in terms of GHG emission reduction targets [10], leading to evaluate SI NG engines now, and bio-methane and e-methane in the medium–long terms, as alternatives to the diesel engines. The main challenge for stoichiometric SI HD engines is their lower efficiency compared to the modern diesel engines [9]: fuel economy is critical for long-haul HD applications. Several techniques can be evaluated to improve HD NG engines’ efficiency, preserving performance and durability. This category includes the dilution strategies, namely exhaust gas recirculation (EGR) and lean combustion, which are particularly interesting because they are relatively easier to implement than other techniques (e.g., water injection, variable compression ratio, waste heat recovery [11,12,13]) and can achieve important benefits.

EGR consists of deflecting a fraction of the exhaust gas to the intake manifold to dilute the fresh air entering the cylinder. A typical system configuration is the high-pressure (HP) EGR. The exhaust gases are picked up before the turbine, and they are cooled and introduced in the intake system after the intercooler (in the low-pressure configuration, the sample point is after the turbine and the re-introduction point is upstream the compressor). A dedicated valve handles the amount of recirculating gas. In Figure 1, a scheme of the HP EGR is reported.

To quantify the EGR quantity re-introduced at the intake, the EGR ratio is defined as pointed out by the following equation:

$$\mathrm{EGR}\mathrm{ratio}={\displaystyle \frac{{\mathrm{m}}_{\mathrm{EGR}}}{{\mathrm{m}}_{\mathrm{EGR}}+{\mathrm{m}}_{\mathrm{fresh}\mathrm{air}}}}\times 100$$

(1)

In SI engines, the use of an EGR system can enhance the thermal efficiency due to the reduced combustion temperature and lower pumping losses in part load operations [14]. Inert exhaust gases do not actively participate in combustion, but they absorb the released heat, which lowers the overall combustion temperature. This results in reduced heat losses to the cylinder walls and exhaust temperatures, ultimately enhancing efficiency. Additionally, since EGR dilution decreases the available oxygen, more intake pressure is required to maintain the same volume of fresh air. This increases efficiency by reducing pumping losses at low-to-medium loads [15]. The use of EGR dilution also allows for the use of TWCs, as combustion can occur under stoichiometric conditions, contributing to significant reductions in harmful gaseous emissions and helping meet EURO standards [16,17]. Furthermore, as the EGR ratio rises, tailpipe NO_x emissions decrease [15,18] due to the reduction in combustion temperature which also makes EGR competitive as a knock control strategy [8]. Several studies have identified an optimal EGR ratio for each engine operating point that maximizes thermal efficiency in SI engines. Additionally, Durrett [19] suggests that the heat capacity ratio (γ) in stoichiometric EGR applications is higher compared to non-diluted operations. However, the potential for improvement is generally smaller than that achieved through air/fuel lean-burn combustion [19].

Lean air/fuel combustion consists in introducing “excess” of fresh air in the cylinder with reference to the stoichiometric quantity needed. If the engine is running, it pulls more air than needed for a stoichiometric combustion. Air dilution decreases the combustion temperature, reducing the wall heat loss [20]. Moreover, within a certain range, lean operations can enhance combustion efficiency by facilitating fuel combustion in an air-rich environment, which helps minimize the formation of unburned hydrocarbons. Lean operations applied to SI NG engines lead to higher volumetric efficiency and indicated thermal efficiency compared to the stoichiometric case, along with a rise in peak pressure and a reduction in exhaust temperature [21]. Despite the lower combustion temperature, the oxidizing environment promotes the formation of nitrogen oxides (NO_x) until specific lean lambda values (typically around λ = 1.1–1.2) are reached, after which NO_x levels begin to decrease [21,22,23]. As the TWC is not effective in lean operations [20,24], the latter aspect is not negligible and prompts the evaluation of specific after-treatment system (ATS) concepts [25].

However, charge dilution slows down the combustion and the flame front speed decreases: an engine can tolerate a certain amount of dilution (identified by the generic term of “dilution limit”), after which it misfires and instabilities appear, with high cycle-to-cycle variation and unburnt hydrocarbon emissions [8,15,26].

This paper presents experimental work conducted on a modern HD SI single-cylinder engine (SCE) fuelled by NG, installed at CNR-STEMS. This study was carried out in seven different engine operating points to evaluate the engine’s tolerance to EGR and lambda lean dilution, as well as to characterize fuel-saving advantages and combustion characteristics. Studies conducted with such a systematic approach in investigating the effects of dilution with fresh air and EGR on a new design HD EURO 6 engine, representative of a category whose decarbonization is complex and necessary, are not present in the literature. The findings offer valuable insights for a comprehensive assessment of dilution strategies, highlighting their potential for enabling the use of gas engines, which is also relevant for future hydrogen applications. The results suggest that both dilution strategies show interesting potential in increasing the indicated efficiency. Lean combustion shows better results in this regard (with peaks of 10%), but the NO_x values recorded for non-extreme dilutions are a serious issue. On the other hand, EGR shows more moderate fuel consumption reductions, but it succeeds in reducing the NO_x produced. For both strategies, once the dilution limit is reached, small additional percentages trigger uncontrollable instability due to frequent misfires.

2. Methodology

The experimental campaign was conducted on an SCE test bench designed and developed in-house, featuring an automatic control system implemented on a real-time programmable controller (CompactRIO by National Instruments, Austin, TX, USA) to provide maximum flexibility in controlling and monitoring combustion, and the parameters involved. The heart of the test bench is a four-stroke single-cylinder port fuel injection (PFI) HD SI engine, based on a reference multi cylinder engine at the state of the art. In

Table 1. SCE main characteristics.

Engine Type	Single Cylinder
Displacement	2147 cm3
Bore	135 mm
Stroke	150 mm
Valves per cylinder	4
Compression ratio	11.9
Combustion chamber type	Bowl-in piston and flat head
Ignition	Inductive discharge + spark plug
Injection	Port Fuel at 8 bar

, engine characteristics are reported.

In Figure 2 is reported a scheme of the intake, exhaust, and EGR lines with the equipped sensors, and Figure 3 shows a picture of the SCE.

The 2.1 L displacement of this specific reference engine is highly representative of the most used powertrains in the HD sector [27]: every relevant result achieved during the research activities could be significant for the decarbonization of a large portion of the sector.

The engine is coupled with an asynchronous dynamometer specific for HD engines. The acquisition and control of the various sensors and engine control parameters has been made through the mentioned homemade developed software. The indicated quantities, such as in-cylinder pressure and rate of heat release (RoHR) traces, the indicated mean effective pressure (IMEP), and the mass burned fractions (MBFs), derive from a Kistler Group (Winterthur, Switzerland) high-speed piezoelectric pressure transducer (model 6052C), and are collected through the X-Ion by AVL (Graz, Austria). The emissions are measured at the exhaust line by a Horiba (Kyoto, Japan) MEXA 7100DEGR gas analyser. In

Table 2. Measurement ranges and accuracy of Horiba MEXA 7100DEGR.

Emission Species	Range	Linearity/Accuracy
CO	0–10%	<1%
CO2	0–20%	1%
O2	0–25%	0.5%
NOx	0–5000 ppm	<1%
THC	0–10,000 ppm	0.5%

, its characteristics are reported.

The fuel used is high-purity natural gas (made by 99.5% of methane), stored in pressurized tanks with certified composition. The EGR ratio evaluation is made with two different strategies. Alongside the usual UEGO lambda sensor (LSU 4.9 type) by Bosch (Stuttgart, Germany) mounted at the tailpipe to enable the automatic fuel dosing, a second UEGO sensor at the intake manifold has been installed. This is mandatory to estimate the EGR ratio without external systems, to keep the injected fuel automatically controlled even in the EGR campaign. In fact, the oxygen volumetric concentrations derived by the UEGO sensors signals can be used for a first estimation of the EGR ratio according to the following formula:

$$\mathrm{EGR}\mathrm{ratio}={\displaystyle \frac{{(\mathrm{O}}_{2\mathrm{env}}-{\mathrm{O}}_{2\mathrm{int}})}{({\mathrm{O}}_{2\mathrm{env}}-{\mathrm{O}}_{2\mathrm{exh}})}}\times 100$$

(2)

The O_2env. is the concentration of oxygen in the environment, O_2int and O_2exh are the ones in the intake manifold and in the exhaust line.

The Horiba MEXA is also used for an online evaluation of the EGR ratio. In fact, the utilized model makes it possible to sample the CO₂ both in the tailpipe and in the intake manifold and uses a formula similar to the previous oxygen concentration-based one to calculate the EGR ratio. The Horiba estimation is expected to be more reliable, and in the subsequent Section 3, the results will always be related to the latter.

The required lambda value is set in the control software, and the actual exhaust gas lambda is controlled in a closed loop by adjusting the injection width, based on feedback from the UEGO sensor. Furthermore, Horiba MEXA calculates the exhaust gas title through the carbon balance equation and can be considered more reliable than UEGO sensors in CH₄ applications as their calibration curve is based on chemically different fuels. The official calibration curve implemented in the control software has undergone an initial refinement, but it is not yet optimized for lambda values above 1.2. For this reason, the lean campaign results will be shown as function of Horiba MEXA lambda values.

In

Table 3. Tested engine points and methodology.

RPM	Load [%]	Testing Methodology
1100	30	- λ = 1.00, 1.05, 1.10, 1.20, 1.30 …. lean limit - EGR ratio = 0%, 5%, 10% …. EGR limit - IMEP and MBF50 fixed values Threshold to identify dilution limit: - IMEP’s Coefficient Of Variation (COVIMEP) = 4%; - Indicated Specific Fuel Consumption (ISFC) trend reversal point
1100	50
1200	40
1200	70
1400	30
1400	50

, the tested engine points and testing methodology are reported, in terms of engine speed per load percentage.

The six points move from low speed and load to medium/high ones since they are the most representative of real driving operations following the homologation World Harmonized Transient Cycle: almost 50% of the engine points reproduced along it are in the range of 1000–1400 RPM within the 70% load. The MBF50s and backpressures are the ones utilized by the reference multicylinder engine in its map. The dilution range extends to the 4% COV_IMEP limit and/or ISFC’s positive trend reversal point, at 5% intervals for EGR and 0.05–0.1 lambda value for lean combustion; the last useful value may be within the step interval. The lambda values refer to the setpoint used in the experiments for the automatic control loop based on the UEGO sensor signal; the actual lambdas are slightly higher (starting from λ = 1.20) due to the reasons stated above. The IMEP and MBF50 were compensated by adjusting the intake pressure and spark timing.

The tests were carried out in steady-state conditions. After an initial engine warm-up procedure, data acquisition begins following a sufficient stabilization period at each specific point, during which temperatures, pressures, and flows reach steady state. A dedicated software section monitors the stationarity of the key parameters. The indicated curves and quantities represent the average values over 200 engine cycles; the test bench data, control parameters, and emissions are mean values collected over a 60 s acquisition period.

3. Results and Discussion

This section presents the main results of the lean combustion and EGR campaigns separately. For clarity and brevity, the indicated curves (in-cylinder pressure, RoHR, and HR) that will be shown later will only refer to the 1200 RPM × 40% load point (also referred to as “cruise”), identified as the most representative of the engine’s on-road operating conditions.

3.1. Lean Combustion Campaign Results

Figure 4 shows the in-cylinder pressure and RoHR curves for the cruise point as a function of lambda dilution. The lambda values shown in the graph legends represent the desired values. For clarity reasons, lambda steps of 0.2 are reported up to the most diluted point: this way, the differences are more easily noticeable.

As dilution progresses, more intake pressure is needed to maintain the IMEP, and therefore the peak firing pressure (PFP) also increases. The combustion becomes longer moving towards leaner lambda values, as demonstrated by the RoHR profiles, which show lower peaks and becomes more flattened. The most diluted point shows a slightly earlier combustion barycentre visually compared to the others, indicating that the correct phasing of combustion was beginning to become overly complex.

In Figure 5, a comparison among the ISFC of the non-diluted running conditions is reported. The values are expressed as percentage variation with respect to the “cruise” point.

In the tested range of engine speeds, load drives the consumption: the higher the load, the higher the efficiency. As a matter of fact, the 1200 × 70% shows the best performance. In the following, to emphasize the trends observed in dilution, the scaling will be based on the undiluted condition for each engine point.

In Figure 6, the ISFC trends as a function of lambda values derived by the Horiba analyser (referred to as “Lambda MEXA [-]”) are reported. The curves are grouped for engine speed, and the values are normalized on the specific fuel consumption value of the non-diluted condition of each specific point.

The trends are decreasing until a certain dilution value, beyond which the consumption tends to rise again. The re-increasing behaviour is not observed for 1200 × 40% and 1400 × 50% because combustion exceeds the stability limits before reaching the reversal. Lean combustion leads to significant efficiency enhancements, particularly at lower loads (thanks to the additional contribution of pumping loss’ reduction), with peak fuel savings reaching 9–10%. At higher loads, the benefits stand around a remarkable 7–8%.

In Figure 7, the COV_IMEP trends are reported.

The graphs show that conditions at lower loads are more easily subject to an increase in COV_IMEP, even at the lower lambda values. The trend is more stable at higher loads. This is likely due to greater turbulence and higher temperatures in the combustion chamber (more reactive environment), which also explains the lower COV_IMEP at higher load points in stoichiometric conditions. A higher amount of excess fresh air is needed to sufficiently slow down the flame front to the point where combustion stability worsens compared to the stoichiometric point. However, for all conditions, a stability limit (known as “lean limit”) is reached: even small additional increases in dilution lead to uncontrollable combustion instability due to the occurrence of several very late combustion and misfire cycles. The combustion starting in an overly diluted environment is weaker as it is not supported by a reactive environment and the flame front is excessively slowed down, affecting its repeatability and efficiency; a misfire is the extreme situation in which the activation energy provided by the spark is not sufficient to guarantee the development and/or propagation of the kernel. This behaviour was recorded for 1100 × 30%, 1100 × 50%, 1200 × 40%, and 1400 × 30%, but is also true for the other ones as experimentally verified during the tests. However, the most diluted points that manage to avoid the sudden increase in COV_IMEP are those with the highest efficiency. Apparently, for this engine type, the “dilution limit” maximizes performance, indicating the potential progress that could be made by introducing technologies to increase the latter.

In Figure 8, the CO₂, NO_x, CO, and unburned CH₄ emissions are shown. The emissions are normalized on the value provided in grams on kWh by the non-diluted condition of each specific point.

NO_x emissions increase significantly with small dilutions, peaking around λ = 1.20 (with a 50–60% increase), as the excess oxygen promotes nitrogen oxidation until the drop in combustion temperatures counterbalances it, and the trend reverses. This is a critical issue for lean combustion, as the high increase in NO_x emissions cannot be abated by a TWC. The carbon oxides collapse as soon as the combustion moves to the lean range, and CO₂ remains almost constant while the unburned CH₄ shows a slight decrease and then rise again for the extreme values of the dilution ranges. These results suggest that combustion efficiency improves with dilution, until it starts to worsen as the lean limit is approached. This can be demonstrated by applying the carbon balance equation [20] to characterise the combustion efficiency, whose results are shown in Figure 9.

Lean combustion has a significant impact on combustion efficiency, with a peak around λ = 1.20, after which decreases. As expected, the higher the load, the higher the combustion efficiency in the whole dilution range.

Figure 10a shows how the energy potentially obtainable from the ideal combustion of the injected fuel (considered as 100% for each individual point) is distributed between the losses (non-ideal combustion, heat losses, losses due to residual exhaust gas energy) and the indicated work obtained for the cruise point as λ progresses. The total potential energy was calculated by multiplying the mass of the fuel by its calorific value. The indicated work was calculated from IMEP and displacement. The residual energy of the exhaust gases is calculated using the following formula:

$${\mathrm{Energy}}_{\mathrm{exh}}={\mathrm{c}}_{\mathrm{p}}\times {\mathrm{mass}}_{\mathrm{exh}}\times ∆\mathrm{T}$$

(3)

The mass trapped in the cylinder (air and fuel) is multiplied by the temperature difference recorded between the section immediately after the engine outlet (before the exhaust tank) and the section immediately before the intake valves (downstream of the intake tank). This procedure makes it possible to calculate how much of the energy obtained from combustion is still present in the exhaust gases, after completing the thermodynamic cycle, neglecting the blow-by flow. Since the accurate estimation of the thermal loss is complex at the current state of the test bench, its contribution was calculated as a complement to the total potential energy, taking into account the calculation made for the other losses. In Figure 10b, the percentage variations of the exhaust gas temperature and total mass trapped in the combustion chamber relative to the stoichiometric point are displayed.

The indicated work obtained tends to be progressively increase in percentage until λ = 1.33, where it plateaus, confirming the trend of ISFC shown in Figure 6. The losses due to non-ideal combustion reflect the trends shown in the previous Figure 9. Heat losses, on the other hand, increase their contribution up to lambda 1.1. This behaviour can be explained by the fact that a higher percentage of fuel is burned completely (combustion efficiency significantly higher than that at the stoichiometric point), generating higher peak temperatures not compensated by the excess air even though the flame front is slower, as depicted in Figure 4. The peak temperatures are significantly reduced for higher dilutions where heat losses strongly reduce their impact (from 30% at lambda 1.1 to 24% at lambda 1.47).

As dilution increases, the contribution from residual exhaust gas energy steadily rises. Figure 10b provides insights on the two main quantities affecting this parameter. As the combustion becomes leaner, the exhaust temperatures decrease, reaching a reduction of up to 10% at lambda 1.47. However, this is more than counterbalanced by the increase in exhaust mass, which rises by over 30%. The combination of these two effects results in a higher share of energy lost in the exhaust gases. The analysis shows that the overall increase in indicated efficiency in lean combustion is strongly dependent on the reduction in heat losses. The other important parameter is certainly the γ, but its contribution cannot be exposed with this analysis and will be the subject of future studies. These results represent an upper limit of the gain, because more boost means greater compressor losses, and therefore the efficiency increase in a multi cylinder engine should be slightly lower.

3.2. EGR Combustion Campaign Results

In Figure 11, the mean in-cylinder pressure cycles and RoHR traces are shown for the cruise point as the EGR ratio increases. EGR steps of 10% are reported for clarity reasons.

The characteristics of in-cylinder pressure and RoHR as EGR dilution increases are similar to those observed in lean combustion: as the EGR ratio increases, the duration lengthens, and more advanced ignition timing is needed to match the barycentre, as well as higher intake pressure to match the IMEP. The difference lies in the source of the higher pressure compared to the non-diluted point. In this case, it is not the fresh air that increases, but rather the recirculated EGR mass that is added to the fresh air. Also in this case, the most diluted point was performed without maintaining the MBF50, which is a symptom of the high combustion instability.

In Figure 12, the ISFC trends recorded for the tested points are reported, following the same engine speed regrouping and normalization criteria of the lean campaign. The EGR ratio is referred to as “EGR [%]” in the graphs’ x-axis.

As the EGR ratio increases, the ISFC decreases, with a peak reduction of 5% at the 1200 × 40% point at EGR ratio of about 15%. At the same engine speed, ISFC reduction is more pronounced at lower loads, while the benefit decreases at higher loads. This is because the system operates under boosting conditions (with intake pressures higher than atmospheric pressure), and there is no reduction in the pumping cycle as a constant differential pressure between exhaust and intake needs to be maintained. As load increases, combustion with EGR becomes more complex to stabilize, as seen at 1200 RPM. In fact, the 1200 × 70% shows no advantage in ISFC, even with an EGR ratio of 5%; instead, fuel consumption increases significantly as the ratio moves towards 10%, showing an anticipated reversal point in the trend compared to the other conditions. These features highlight that EGR can furnish valuable percentages of fuel saving but loses effectiveness at higher loads.

In Figure 13, the COV_IMEP trends are shown.

COV_IMEP remains almost constant to the non-diluted point, with a slight increase followed by a strong instability due to late combustions and misfires, similar to lean combustion. The cause is consistent with the previously discussed explanation: the inert components (mainly CO₂ and H₂O) in the combustion chamber reduce the combustion temperature, slowing down the combustion process. The EGR ratio at which the combustion becomes uncontrollable is diverse among the points. Following the ISFC pattern, the COV_IMEP trend of the highest load underlines that this technique might not be particularly suitable for higher loads. Even in this case, the EGR ratio closest to the sudden combustion instability (“EGR limit”) is the one with the lowest fuel consumption.

In Figure 14, the normalized values (according to the same criteria followed in the lean lambda campaign) of gaseous emissions are provided.

EGR proves to be highly effective in reducing NO_x emissions with a peak of −60% at 1100 × 50% at the highest EGR ratio tested within the COV_IMEP threshold. The CO₂ slightly increases, just like unburned CH₄, while CO does not show any repeatable trend; no systematic trend is observed in combustion efficiency, except for the higher loads’ advantage over lower ones.

As presented in the lean combustion results section, Figure 15a quantifies the percentage contribution of each specific loss type to the total potential energy obtainable from the ideal combustion of the injected fuel, along with the indicated work acquired, as dilution progresses for the 1200 × 40% engine point. In this case, the gas mass used for calculating the residual energy in the exhaust gas considers the air, fuel, and EGR mass. This evaluation considers the engine as a control volume in which these three masses compose the whole trapped mass that undergoes the thermodynamic cycle. Moreover, the EGR flow sent back to the intake is cooled, losing most of the energy still available at the exhaust port, and equalizes its temperature with that of the fuel and air at the intake port. It would not be correct to exclude the EGR mass from the calculation even though it is recirculating, as its energy at the intake differs from that at the exhaust, and the residual energy would be underestimated. Figure 15b shows the trends for exhaust mass and temperature.

The collected indicated work increases its share on the total potentially obtainable energy as EGR dilution progresses, excluding the EGR 18% point which is beyond the EGR limit (as COV_IMEP exceeds the threshold, and the positive trend in specific fuel consumption reverses). Compared to the lean lambda campaign, the energy lost due to non-ideal combustion remains almost constant, highlighting a combustion efficiency that is virtually unchanged. As expected, heat losses tend to decrease with dilution because an increasing portion of the in-cylinder charge is almost inert to combustion, lowering the temperatures. The residual energy of the exhaust gases tends to slightly increase its impact, due to a higher percentage increase in the exhaust mass compared to the decrease in exhaust temperatures.

In HD gas engines, EGR can be particularly useful for reducing fuel consumption at low loads, where the efficiency gain percentage is significant and results in a strong reduction of engine-out NO_x. The possibility of exploiting a traditional ATS systems is not to be underestimated. Lean combustion shows better results in terms of fuel consumption, even at higher loads. However, a turbocharger capable of handling sufficiently high pressures would be required. The option of reducing the lambda level as load increases to ensure proper use of the turbocharger unit could be feasible, although this would lead to higher NO_x emissions. Diesel-derived NO_x reduction technologies, when adapted for use with methane, could help achieve high efficiencies and low emissions in the whole engine map.

4. Conclusions

In this study, the potential of enhancing modern methane-fuelled HD engines’ efficiency through charge dilution techniques (lean combustion and EGR) is explored. This aspect is particularly important in the mid-to-long-term scenario because biomethane and e-methane-powered engines could, alongside green H₂, be a very low environmental impact solution and replace diesel engines; however, they currently face efficiency challenges. The dilution techniques can be evaluated on SI NG engines as alternative combustion methods, with relatively low implementation issues, to reduce fuel consumption. The experiments were conducted on a NG 2.1 Liter SCE with bowl-in piston and flat cylinder head (diesel-derived) to embrace the most used unitary displacement engine in the long-haul HD vehicles [27]. Six engine points were characterized in lean and EGR dilution, to verify the engine’s tolerance to both methods, maintaining IMEP and MBF50 of the non-diluted condition. Each point was pushed towards the “lean/EGR limit” identified by a 4% COV_IMEP threshold. Both techniques demonstrated their capability to reduce fuel consumption, with different pro and cons.

The gradual slowing of the flame front as dilution progresses, which can be due to a less energetic kernel as well as a less reactive environment, results in specific limits for each engine point, where slightly higher percentages cause a sudden and significant increase in COV_IMEP towards unacceptable values due to the occurrence of many misfire and late combustion cycles. This phenomenon is experimentally characterized by the presence of these dilution “breaking points”, after which the combustion becomes suddenly uncontrollable.

Both techniques were capable of reducing the indicated specific fuel consumption. Lean combustion prevails, with greater reductions: the peak fuel saving resembles a 5% for EGR and 10% for lean combustion. Lower loads perform better due to the reduction in pumping losses. For every tested point, the lowest ISFC value is obtained at the “dilution limit”, immediately preceding the onset of the strong and sudden combustion instability. These aspects suggest that increasing the dilution limit within combustion stability tolerability could further improve the results obtained.

The best performance achieved in lean combustion still has a significant drawback due to higher NO_x emissions, which can only be reduced with very high dilution levels that may not necessarily correspond to replicable combustion stability conditions. Evaluating innovative technologies to expand the lean limit could be crucial to avoid the use of dedicated post-treatment systems. Additionally, a combined dilution approach using both techniques could ideally mitigate the increase in NO_x emissions, achieving intermediate fuel consumption gains while maintaining lower emissions. However, the impact on the flame front and combustion stability would need further study. Heat losses increase as the working point moves from the stoichiometric condition to a lambda of about 1.1, due to an increase in peak temperatures resulting from the improvement in combustion efficiency. Moving beyond, the heat losses are reduced. On the other hand, EGR tends to reduce heat losses’ percentage impact even at lower dilution levels. Both strategies tend to increase the residual energy in the exhaust gases, particularly in lean combustion, due to the higher gas mass variation compared to the decrease in temperatures. This aspect can be positive for improving the efficiency of the turbo group but suggests that evaluating earlier combustion timings as dilution increases might be a path to improve performance, while clearly being cautious of the consequences on other parameters (e.g., combustion temperatures and associated heat losses). The activity presented lays the foundation for evaluating the main benefits and limitations of operating under diluted conditions in HD SI gas engines. From this solid starting point, various innovative technologies aimed at improving performance can be studied.