Impact of Hydrogen-Enriched CNG on Combustion Phasing, Fuel Efficiency, and Emissions in a Multi-Cylinder Spark-Ignition Engine under Varying Loads

2.1 Engine specifications

The experimental investigation was conducted using a 6.5-liter, naturally aspirated, water-cooled, four-stroke, multi-cylinder SI engine. The engine features a total displacement of 6.5 liters. It is equipped with an electronically adjusting mixer system, conventional spark plug ignition and is designed to operate on gaseous fuels. The compression ratio was maintained at 17.1:1 to ensure compatibility with both pure CNG and HCNG blends without inducing knocking. A high-energy ignition coil and an advanced ignition timing controller were employed to provide precise spark timing, particularly critical for higher hydrogen content blends due to their lower ignition energy and faster combustion characteristics. The engine specifications are given in Table 1.

Table 1. Engine specifications

The engine was mounted on a heavy-duty testbed and coupled with an eddy current dynamometer to allow for precise control of speed and load conditions during the tests. The engine's performance and combustion metrics were monitored in real-time using a centralized control system integrated with advanced instrumentation.

2.2 Fuel blends

Four fuel configurations were evaluated:

• CNG (0% hydrogen) – Serving as the baseline fuel, composed primarily of methane (~95% CH₄), with small amounts of Ethane, Nitrogen, and Carbon Dioxide. The detailed CNG composition is given in Table 2.
• 18HCNG – A blend containing 18% hydrogen by volume and 82% CNG.
• 25HCNG – A mid-range blend with 25% hydrogen.
• 30HCNG – The highest hydrogen content tested in this study, composed of 30% H₂ and 70% CNG.

The blends were prepared using precision gas mixing units equipped with mass flow controllers, ensuring consistent volumetric ratios. The gases were stored in ISO 9809-compliant high-pressure cylinders, filtered for moisture and particulates, and validated using gas chromatography, as per standard HCNG blending practices.

Table 2. Detailed CNG composition

2.3 Engine dynamometer and data acquisition systems

The engine was coupled to a 150kW eddy current dynamometer equipped with a load cell for torque measurement and a high-precision shaft encoder for crank angle resolution (up to 0.1° CA). The dynamometer was controlled through a programmable electronic controller in the test cell Automation system, capable of maintaining constant engine speed and varying engine load with high repeatability. The schematic engine test setup is shown in Figure 1.

1.png

Figure 1. Experimental test setup with HSDA system for in-cylinder combustion analysis

An AVL GH15D piezoelectric pressure transducer was installed in one cylinder to measure real-time in-cylinder pressure data. The sensor was linked to an AVL FI Piezo charge amplifier, which transmitted signals to a high-speed data acquisition system (HSDA System), AVL Indimicro. Pressure signals were logged over 300 consecutive engine cycles at each operating condition to account for cycle-to-cycle variation. The start of combustion and heat release analysis were determined using the first derivative of pressure with respect to crank angle and applying the Rassweiler-Withrow method for HRR calculations. The complete instrumentation is given in Table 3.

Table 3. Instrumentation for test data acquisition

2.4 Test conditions: Single-speed operation with varying load levels

The engine was operated at a constant speed of 1500 RPM. For each fuel blend, the engine was subjected to four load levels represented by BMEP as 1.5 bar, 3.7 bar, 6 bar and 7.5 bar, to assess the impact of engine loading on combustion performance and stability.

At each operating point, the engine was allowed to reach thermal steady-state, confirmed when coolant and oil temperatures stabilized within ±2℃ for at least 10 minutes. All data collection was initiated after stabilization to minimize thermal transients and ensure repeatability.

Ignition timing and throttle opening were kept constant across blends to isolate the effect of hydrogen enrichment. Fuel mass flow rates were adjusted to maintain consistent equivalence ratios (λ ≈ 1) across test cases, measured using a gas flowmeter and corrected for ambient pressure and temperature.

4.1 Rate of Pressure Rise (RoPR)

The Rate of Pressure Rise (RoPR), expressed in bar per crank angle degree (bar/°CA), is a key indicator of combustion intensity and flame development speed within SI engines. It reflects how rapidly in-cylinder pressure increases as combustion progresses and is particularly important when analysing the effects of fuel reactivity and ignition timing. For HCNG-fueled engines, RoPR is influenced by hydrogen concentration, load condition, and combustion phasing, making it a sensitive metric for assessing combustion quality and knock propensity. The RoPR for the CNG and HCNG fuel blends at various BMEP levels is shown in Figure 2.

2.png

Figure 2. RoPR for CNG and HCNG blends under various BMEP conditions

These results clearly demonstrate that hydrogen enrichment significantly elevates the peak pressure rise rate, particularly in mid- to high-load regimes. The largest relative increase in RoPR is observed at low load (1.5 BMEP), where hydrogen improves flame kernel stability and accelerates early combustion, effectively overcoming CNG’s limitations in slow ignition and lean burn instability.

At higher loads (6.0 and 7.5 bar BMEP), all blends produce more intense combustion due to an increase in in-cylinder pressure and temperature. However, the incremental effect of hydrogen begins to plateau, suggesting that the combustion environment is already favourable and additional reactivity contributes less proportionally. The findings indicate that while hydrogen enrichment enhances combustion characteristics at lower loads, its benefits may diminish in more favourable combustion environments, necessitating further investigation into optimal blend ratios across varying operational conditions.

Figure 2 presents the variation in the rate of pressure rise (RoPR) as a function of crank angle for different HCNG blends (18%, 25%, and 30% hydrogen by volume) compared to conventional CNG, at four BMEP levels: 1.5, 3.7, 6.0, and 7.5 bar.

At low load (BMEP 1.5 bar), the peak RoPR for CNG is modest at 0.59 bar/°CA, reflecting the slow and incomplete combustion associated with lean burn and low in-cylinder temperatures. With hydrogen enrichment, RoPR increases to 0.69, 0.74, and 0.76 bar/°CA for 18HCNG, 25HCNG, and 30HCNG, respectively. These results confirm hydrogen’s role in improving flame propagation and early-stage combustion under dilute conditions.

In the mid-load condition (BMEP 3.7 bar), RoPR grows significantly. CNG peaks at 1.10 bar/°CA, while 30HCNG reaches 1.54 bar/°CA—marking a ~40% increase. The flame speed enhancement due to hydrogen is more effective in this regime, where pressure and temperature conditions are sufficient to support rapid energy release. Similar enhancements have been documented by Bhasker and Porpatham in lean burn HCNG engines [31].

At 6.0 bar BMEP, combustion intensity becomes more robust for all blends, with RoPR values rising to 1.76, 2.05, 2.12, and 2.19 bar/°CA for CNG, 18HCNG, 25HCNG, and 30HCNG, respectively. While the increment continues with hydrogen content, the marginal increase diminishes, suggesting that the combustion environment is nearing its optimal reactivity threshold. This trend aligns with De Simio et al., who observed a saturation effect in RoPR enhancement at high hydrogen fractions [18].

At full load (BMEP 7.5 bar), RoPR peaks at 2.67 bar/°CA for 30HCNG compared to 2.46 for CNG. The rate of increase is comparatively small, indicating diminishing returns due to already favourable ignition conditions. At this point, further increases in hydrogen concentration may necessitate spark retard or EGR strategies to avoid knock, as warned by Nitnaware and Suryawanshi [32].

Overall, RoPR increases with both hydrogen enrichment and BMEP. However, the influence of hydrogen is most pronounced at low and mid-loads and plateaus at high loads, underscoring the need for adaptive combustion control strategies for high hydrogen blends.

Knock margin analysis based on the rate of pressure rise

The RoPR is a crucial metric for assessing combustion intensity and knock susceptibility in spark-ignition engines. It serves as an early indicator of pressure oscillations and potential autoignition in the end-gas region. In this study, RoPR values were carefully analyzed across varying hydrogen blending ratios and engine loads, particularly focusing on the knock-prone full-load condition (BMEP = 7.5 bar).

Contrary to general concern for high hydrogen blends, the measured RoPR for the 30HCNG blend at 7.5 bar BMEP was 2.67 bar/°CA, only marginally higher than that of CNG (2.46 bar/°CA) under the same conditions. These values fall significantly below the typical knock-onset thresholds (8–10 bar/°CA) cited in literature for SI engines operating without EGR or knock-suppression strategies [33, 34]. Therefore, it is evident that even at full load, the 30HCNG blend exhibits stable combustion without entering the knock-limited regime.

The slightly higher RoPR for HCNG blends can be attributed to the enhanced laminar flame speed and thermal diffusivity of hydrogen, which accelerates combustion and leads to a steeper but still controlled pressure rise. This behaviour is further substantiated by the earlier MBF50 phasing observed in Section 4.3 and the elevated, yet stable, heat release rates and combustion temperatures discussed in Sections 4.2 and 4.4. Importantly, the pressure development remains within safe operational margins, demonstrating that hydrogen enrichment up to 30% vol. in CNG does not inherently induce knocking, at least under the current ignition phasing and mixture strategy.

Thus, the knock margin for all tested HCNG blends, including 30HCNG, is considered sufficient under the current combustion settings. These findings validate the potential of hydrogen addition in improving combustion dynamics without compromising knock resistance, provided that optimal spark timing and air–fuel ratios are maintained.

4.2 Heat Release Rate (HRR)

The HRR represents the rate at which chemical energy from the air–fuel mixture is converted into thermal energy within the combustion chamber. It is directly influenced by the fuel composition, ignition timing, flame speed, and in-cylinder turbulence. In SI engines, HRR serves as a critical metric for evaluating combustion intensity, speed, and efficiency [35]. Particularly for HCNG blends, HRR trends reveal important insights into how hydrogen affects the overall combustion process across varying engine loads [36].

The HRR is a key indicator of the energy conversion process within the combustion chamber. It reflects the rate at which chemical energy is transformed into thermal energy during combustion. In this study, the HRR was analyzed for four fuel blends—CNG, 18HCNG, 25HCNG, and 30HCNG—across four BMEP levels.

The normalised HRR trends, as illustrated in Figure 3, demonstrate a consistent and quantifiable increase in peak HRR values with rising hydrogen content across all load conditions. At 7.5 bar BMEP, 30HCNG showed the highest normalized HRR, followed by 25HCNG, 18HCNG, and CNG. A similar progression was observed at 6 bar, 3.7 bar and 1.5 bar BMEP levels, indicating that hydrogen enrichment consistently enhances the combustion intensity.

3.png

Figure 3. HRR for CNG and HCNG blends under various BMEP conditions

This improvement is attributable to the superior diffusivity, flame speed, and lower ignition energy of hydrogen, which enhance the overall combustion rate and shorten the combustion duration. Consequently, the combustion process becomes more complete and efficient, especially under lean and lower load conditions where CNG-only blends typically struggle due to slower flame propagation.

At a BMEP of 7.5 bar, the 30HCNG blend exhibited a pronounced spike in the HRR, a characteristic indicative of the highly reactive combustion dynamics inherent to hydrogen-enriched fuel mixtures under elevated load conditions. The inclusion of 30% hydrogen in the CNG blend significantly reduces ignition delay and enhances the laminar flame speed due to hydrogen's fundamental combustion properties—namely, low activation energy, high diffusivity, and high flame propagation rates. These features result in a rapid and concentrated combustion event near TDC, causing an abrupt and elevated HRR profile.

While such rapid combustion can enhance thermal efficiency, it also raises the risk of elevated pressure gradients or knock tendencies if not managed through appropriate spark timing or mixture control. These findings align with the established understanding of hydrogen combustion in spark ignition engines, wherein fuel reactivity increases disproportionately with higher hydrogen ratios, especially under high load conditions [33].

However, it is crucial to balance higher HRR with mechanical and thermal stress considerations. While higher HRR correlates positively with thermal efficiency, it also increases in-cylinder temperature and pressure gradients, which could risk engine knocking and increased NOₓ emissions at elevated hydrogen concentrations. Studies by Ma et al. [37] affirm that optimal HRR supports improved indicated thermal efficiency but caution against excessive enrichment beyond stoichiometric limits due to potential detonation and durability concerns.

Based on the collected data and corroborated literature, an optimum HRR is likely achieved with 25–30% HCNG blends under medium to full-load conditions. This range provides a favourable trade-off between combustion stability, energy conversion, and mechanical safety.

Therefore, 25HCNG can be considered the optimal blend for most conditions, providing enhanced combustion without excessive pressure or thermal gradients. It enables improved engine efficiency, reduced misfire potential at low loads, and robust operation at full load without exceeding safe limits for pressure rise or temperature.

4.3 Combustion phasing – MBF50

MBF50 is the crank angle (or time) at which 50% of the fuel mass has been combusted, serving as a key indicator of combustion phasing. Accurate phasing ensures that the bulk of the energy release occurs during the most efficient portion of the expansion stroke, directly influencing torque output, thermal efficiency, and knock resistance [38].

A shift in MBF50 closer to TDC typically indicates a faster and more centered combustion event, contributing to higher thermal efficiency [39]. However, excessively advanced MBF50 can cause over-rapid pressure rise and increased knock propensity, while delayed MBF50 suggests inefficient combustion and greater heat losses [40].

An ideal MBF50 point often lies within 5–15° After Top Dead Center (ATDC) for optimum thermal efficiency and reduced emissions. In this study, MBF50 was measured across four different brake BMEP levels—1.5, 3.7, 6.0, and 7.5 bar—for CNG and three hydrogen-enriched HCNG blends (18HCNG, 25HCNG, and 30HCNG). The results are shown in Figure 4.

4.png

Figure 4. MBF50% in μs for CNG and HCNG blends under various BMEP conditions

The results clearly demonstrate a consistent shift of MBF50 towards earlier phasing as hydrogen content increases. This trend becomes more prominent with increasing engine load. For instance, at BMEP 7.5 bar, MBF50 for CNG occurred at approximately 550 µs, whereas it advanced to around 430 µs for 30HCNG. To study combustion phasing analysis and to enable direct comparison with standard literature, MBF50 can be represented as crank angle degrees (°CA). For the tests conducted at 1500 RPM, 1°CA corresponds to 111 μs. Therefore, an MBF50 of 550 μs is equivalent to approximately 5°CA ATDC.

At low loads (1.5 bar), the combustion phasing was delayed for all fuels due to slower flame propagation, with CNG showing MBF50 at 1050 µs and 30HCNG at 700 µs. The enriched hydrogen content enhanced flame propagation speed and reduced ignition delay, aligning with the findings of Hora and Agarwal, who reported improved combustion characteristics with HCNG blends under varying compression ratios [41].

This advancement in MBF50 is attributed to hydrogen’s high diffusivity and low ignition energy, enabling faster kernel development and homogeneous flame propagation throughout the combustion chamber [42]. However, excessively advanced combustion phasing must be avoided, as it can lead to knocking tendencies or increased NOₓ formation due to peak in-cylinder temperatures occurring too close to top dead center [43].

From a control strategy perspective, optimizing MBF50 through tailored spark timing becomes critical when varying HCNG ratios. Previous studies suggest that maintaining MBF50 within the window of 555µs–1333µs (i.e., 5–12° ATDC) is optimal for most HCNG engines to balance efficiency and combustion stability [44].

The MBF50 analysis confirms that hydrogen enrichment advances combustion phasing, improving efficiency and responsiveness. The 25HCNG blend offers the best compromise, delivering early yet controlled combustion across all load conditions. These results reinforce the importance of MBF50 as a combustion tuning metric, particularly when integrating alternative fuel blends into conventional SI engine platforms [45].

4.4 In-cylinder temperature

In-cylinder temperature is a critical outcome of the combustion process in SI engines. Though difficult to measure directly during engine operation, it is reliably inferred from combustion pressure traces, HRR profiles, and engine boundary conditions [45]. In the context of this study, in Figure 5, illustrates the variation of the in-cylinder temperature (Temp [K]) with crank angle for CNG and different hydrogen-enriched CNG (HCNG) blends. It also highlights how hydrogen enrichment in CNG influences both the magnitude and timing of the maximum combustion temperature under varying engine operating conditions. The normalized in-cylinder temperature trends were estimated based on the HRR profiles corresponding to four engine load conditions: 1.5, 3.7, 6.0, and 7.5 bar BMEP.

5.png

Figure 5. Combustion temperature for CNG and HCNG blends under various BMEP conditions

The combustion of HCNG blends leads to elevated HRR peaks as hydrogen content increases. Given that hydrogen possesses a higher adiabatic flame temperature than methane and ignites more rapidly, this inherently leads to greater thermal energy generation within a shorter crank angle window [46]. The faster energy release contributes to a rise in peak and average in-cylinder temperatures, particularly at higher BMEPs.

Thermodynamic model for in-cylinder gas temperature estimation

The accurate estimation of mean in-cylinder gas temperature is essential for understanding the thermodynamic behaviour of combustion and assessing energy transfer characteristics in internal combustion engines, particularly when evaluating the impact of hydrogen-enriched fuels. In the present study, a quasi-steady, pressure-based thermodynamic model is implemented to determine the instantaneous mean gas temperature (T_i) from measured pressure and calculated in-cylinder volume profiles. The approach excludes surface heat losses and utilizes a variable polytropic exponent to reflect real-time changes in gas thermodynamic properties during combustion.

The procedure begins with calculating the in-cylinder gas mass mmm, derived using the ideal gas law based on intake manifold conditions. The expression for inducted charge is:

$m=I . V_H \frac{P_s}{R . T_s}$           (1)

where, I is the volumetric efficiency (typically 0.9), V_H the swept volume, P_s and T_s the intake pressure and temperature, and R the specific gas constant (287.12 J/kg·K). This standard intake-based estimation aligns with formulations established in foundational thermodynamic texts, including Heywood [47].

To further resolve the thermodynamic state of the cylinder, the apparent heat release Q_i is estimated using a single-zone formulation that incorporates a variable polytropic exponent K_i, improving the fidelity of HRR analysis over fixed-k models:

$Q_i=\frac{\kappa_i}{\kappa_i-1} V_{i+\frac{n}{2}}\left\lceil P_{i+\frac{n}{2}}-P_{i-\frac{n}{2}}\left(\frac{V_i}{V_{i+\frac{n}{2}}}\right)^{\kappa_i}\right\rceil\left(X_i+1\right)$             (2)

Here, X_i denotes an engine-type factor (0 for spark-ignition engines, 1 for compression-ignition), and n is the crank angle interval (1°CA in this case). This approach is grounded in thermodynamic principles detailed by Brunt and Platts, who emphasize its applicability for high-resolution combustion analysis in diesel engines [48].

A crucial feature of the model is the dynamic computation of the polytropic coefficient k_i, which adapts to temperature-induced variations in specific heat at constant volume (C_vi). This coefficient is modelled empirically as:

$C_{v i}=0.7+T_i \cdot\left(0.155+A_i\right) \cdot 10^{-3}$            (3)

$\kappa_i=\frac{0.2888}{c_{v i}}+1$          (4)

where, A_i is a correction factor depending on engine type, as described by Krieger and Borman [49]. This dynamic approach captures gas-specific heat variations due to combustion product composition and temperature levels.