Skip to Content
EnergiesEnergies
  • Article
  • Open Access

23 June 2026

26 Pages

Flame Propagation Characteristics of Premixed H2-O2 Combustion in an Ultra-High-Pressure Constant-Volume Chamber

,
,
,
and
1
Naval University of Engineering, Wuhan 430033, China
2
Jianghan University, Wuhan 430056, China
*
Authors to whom correspondence should be addressed.
This article belongs to the Section I2: Energy and Combustion Science

Abstract

To investigate the early-stage flame propagation and pressure response of premixed H2-O2 combustion under ultra-high-pressure constant-volume conditions, a transient CFD model was developed for a large-volume confined chamber. The numerical framework combines a density-based solver, the Peng–Robinson real equation of state, large eddy simulation, and a reduced H2-O2 chemical kinetic mechanism. Simulations were conducted at initial pressures of 30 and 40 MPa, H2/O2 molar ratios of 8:1 and 12:1, and three-, four-, and five-point ignition configurations. The results show that increasing the initial pressure from 30 MPa to 40 MPa advances the pressure rise onset from approximately 1.65 ms to 1.28 ms and increases the maximum pressure rise rate from 18.6 MPa·ms−1 to 27.4 MPa·ms−1 under the H2/O2 = 8:1 and three-point ignition condition. Under the investigated fuel-rich conditions, increasing the H2/O2 molar ratio from 8:1 to 12:1 delays the pressure rise onset from approximately 1.28 ms to 1.46 ms and reduces the maximum pressure rise rate from 27.4 MPa·ms−1 to 21.1 MPa·ms−1. For the 30 MPa and H2/O2 = 8:1 cases, the four-point ignition case produces the largest pressure rise rate of approximately 23.5 MPa·ms−1, whereas the five-point ignition case shows a lower pressure fluctuation amplitude of approximately 3.6 MPa. The present conclusions are based on CFD quantitative engineering predictions and should be further validated using quantitative experimental measurements.

1. Introduction

Amid the global carbon neutrality trend, the energy transition faces structural and technological challenges. Hydrogen energy, as a key carrier for achieving carbon neutrality, is rapidly expanding its applications beyond transportation sectors (e.g., hydrogen fuel cells) to include power generation, energy storage, aerospace, and combustion-based energy systems [1]. In engineering applications, combustion reactions represent the most direct method for hydrogen utilization, with current research focusing primarily on the stability of combustion processes [2].
Hydrogen has distinctive physicochemical properties, including low molecular weight, rapid diffusion, high thermal conductivity, and fast chain-branching reactions. These properties lead to a higher flame propagation tendency than that of many conventional fuels. In confined premixed combustion, rapid heat release may cause strong changes in pressure, temperature, flow velocity, and flame morphology. Under sufficiently strong confinement and favorable pressure wave/flame interaction, flame acceleration may develop toward deflagration-to-detonation transition (DDT). However, whether a fully developed detonation is formed must be evaluated using quantitative criteria, such as flame speed, shock speed, pressure jump, and comparison with the Chapman-Jouguet condition. Since these criteria are not fully evaluated in the present work, this study focuses on flame propagation, flame acceleration, and pressure wave/flame interaction rather than confirmed detonation formation [3,4]. Rapid combustion can still generate strong pressure wave loads that may be detrimental in engineering applications. Therefore, understanding and controlling flame acceleration in confined hydrogen combustion remain important for combustion stability and safety analysis [5].
As previously discussed, accurately modeling flame propagation and pressure variations of hydrogen under different initial conditions is fundamental to studying combustion stability. Constant-volume combustion, a typical premixed combustion regime, is characterized by occurring in rigid, sealed containers with constant volume. The energy released during combustion is entirely utilized to increase internal system temperature and pressure, with no external work performed. When thermal losses are neglected, this process constitutes an adiabatic process [6].
Constant-volume combustion simplifies the flow characteristics of premixed gases, converting microscopic thermodynamic responses during combustion into macroscopic pressure fluctuations. The combustion chamber structure can be modeled as an acoustic resonator, providing an effective approach for studying thermo-acoustic oscillation effects in combustion processes. By adjusting various initial conditions—including initial pressure, temperature, equivalence ratio, fuel composition, ignition energy, and geometric constraints—researchers can effectively investigate transition conditions from stable combustion to explosive combustion, pressure rise dynamics in combustion chambers, and variations in flame morphology and propagation characteristics. These findings hold fundamental significance for elucidating combustion mechanisms, advancing combustion theory, guiding burner design, and evaluating the of explosion phenomena [7].
Experimental research remains the primary approach for obtaining fundamental combustion parameters and understanding combustion phenomena. Utilizing constant-volume combustion chamber structures, researchers have accurately measured critical parameters such as laminar combustion velocity, Markstein length, and flame instability for various fuels. Bai investigated the propagation trajectory and pressure rise rate of methane/air premixed flames in the presence of wall oil films [8]. Yoon demonstrated that increasing ignition energy can effectively expand the stable combustion limit of propane/air mixtures while accelerating initial flame development [9]. Song’s system revealed the synergistic enhancement effects of hydrogen blending ratios, initial pressure, and temperature on natural gas laminar combustion velocity, as well as their impacts on flame stability [10]. Li found that elevated initial pressure exacerbates cellular flame instability, while hydrogen addition significantly increases combustion velocity and alters instability thresholds [11]. Su’s study indicated that flame propagation at millimeter and submillimeter scales is predominantly governed by wall quenching effects, potentially leading to periodic oscillations or combustion state transitions, with rising initial pressure critically influencing these phenomena [12]. Gao observed hydrogen/air flame interactions with self-induced pressure waves in obstacle-equipped constant-volume burners in obstacle-equipped constant-volume burners, revealing how initial conditions modulate pressure oscillation frequency and amplitude [13]. Li systematically investigated flame morphological evolution and propagation patterns under varying initial turbulence intensities [14].
Numerical simulations and theoretical analysis serve as powerful tools for understanding the physicochemical mechanisms underlying experimental phenomena. Tsunemi utilized high-fidelity simulations such as direct numerical simulations to meticulously analyze transient processes of turbulent premixed flame/wall interactions in constant-volume containers, quantifying the impact of wall heat loss on localized quenching and overall pressure development [15]. Najim combined computational fluid dynamics simulations with experiments to reveal how centrifugal forces generated by curved channels interact with pressure gradients, distorting flame surfaces and altering propagation velocities [16]. Lin developed a multi-group discrete Boltzmann model, providing novel tools for simulating reactive flows under non-equilibrium conditions—particularly constant-volume combustion processes involving complex dynamics [17]. In addition, significant progress has been made in fundamental research on isochoric combustion of hydrogen and syngas fuels. Researchers systematically characterized the combustion process of H2-CO syngas mixtures in isochoric combustion cells, providing an experimental benchmark for understanding flame propagation in complex fuels [18]. In terms of numerical simulation methods, a comprehensive modeling framework for hydrogen combustion has been developed, ranging from detailed chemical reactor (0D/1D) evaluations to full three-dimensional computational fluid dynamics (3D-CFD) engine simulations, providing methodological guidance for multiscale simulations [19]. Regarding high-pressure combustion mechanisms, studies evaluated the applicability of simplified and detailed hydrogen combustion kinetic models under supercritical CO2 or argon dilution conditions, providing valuable references for simulations under high-pressure conditions [20]. Numerical studies on the evolution of complex flame morphologies, such as the tulip flame in confined hydrogen/air premixed flames [21], and simulations of unsteady phenomena in turbulent hydrogen–air flames [22], have deepened the understanding of hydrogen combustion kinetics. Luijten proposed an analytical model correlating pressure rise with ignited mass fraction, improving the derivation of laminar combustion velocity from constant-volume pressure data [23]. Feidt established detailed chemical kinetic models to evaluate combustion characteristics of mixed gases under varying hydrogen blending ratios, providing theoretical input for engine optimization [24].
However, systematic investigations into flame acceleration and flame interaction in H2/O2 mixtures under ultra-high-pressure conditions of 30–40 MPa remain limited. In particular, the combined effects of initial pressure, fuel-rich mixture composition, and multi-point ignition configuration on flame development and pressure evolution have not been sufficiently quantified. This gap motivates the present numerical study.
Although previous studies have provided valuable information on constant-volume combustion, flame instability, and hydrogen-containing mixtures, several limitations remain. First, most existing constant-volume combustion studies focus on atmospheric or moderately elevated pressures, while fewer studies address ultra-high-pressure H2/O2 combustion at 30–40 MPa. Second, many available investigations concern H2/air or hydrocarbon–air mixtures, whereas oxygen-rich or fuel-rich H2/O2 systems exhibit different chemical time scales and heat release characteristics. Third, the combined effects of initial pressure, highly fuel-rich mixture composition, and multi-point ignition layout on flame interaction remain insufficiently quantified. Therefore, the present work focuses on the numerical comparison of flame development and pressure evolution under selected ultra-high-pressure H2/O2 constant-volume conditions. The present study assumes a fully premixed H2/O2 mixture and a constant-volume chamber. Effects of non-uniform premixing, concentration gradients, temperature gradients, and variable-volume boundary motion are not considered and are left for future work.
This paper examines the structure of a specific type of large-volume fixed-volume burner. Using numerical simulation techniques, it investigates the effects of ultra-high pressure (30 MPa and 40 MPa), rich mixtures (H2-O2 molar ratios of 8:1 and 12:1), and different ignition configurations on the combustion characteristics of the H2-O2 mixture. The study focuses on the deflagration process of the H2-O2 mixture within a sealed combustion chamber, on the mechanisms of energy release, and the flame interaction patterns, providing theoretical support and engineering references for subsequent efforts to enhance the operational stability of H2-O2 combustion and improve combustion chamber safety.
The results are expected to provide preliminary numerical references for the safety assessment and early-stage design of high-pressure hydrogen–oxygen combustion systems, including confined constant-volume combustion devices, hydrogen safety vessels, propulsion-related high-pressure combustors, and combustion-driven gas generation systems. However, the present results are intended for mechanism analysis and comparative evaluation rather than direct design certification.

2. Establishment of Theoretical Analysis Model

In this paper, CVCC denotes constant-volume combustion chamber, LES denotes large eddy simulation, SGS denotes sub-grid scale, PR EOS denotes the Peng-Robinson equation of state, DRGEP denotes directed relation graph with error propagation, and DDT denotes deflagration-to-detonation transition.

2.1. Computational Geometry and Initial Conditions

This paper uses the structure of a specific type of large-volume fixed-volume burner as the basic computational model to conduct simulation and computational analysis under different initial conditions regarding initial pressure, equivalence ratio, and ignition methods. The dimensions of the computational model are shown in Figure 1. The main body of the combustion chamber consists of a constant-diameter cylindrical gas chamber with a diameter of 400 mm and a length of 1494 mm, featuring a 120° inclined taper at the front end of the cylinder. The mixture is ignited by spherical pre-burned zones with a diameter of 5 mm and a temperature of 1500 K, with ignition points evenly distributed along the cylinder’s axis. The initial temperature for all operating conditions in this study is set to 298 K (room temperature) to focus on the effects of the two key variables: ultra-high pressure and rich combustion conditions. The specific design of the initial conditions is shown in Table 1.
Figure 1. Geometry and ignition point arrangement of the constant-volume combustion chamber.
Table 1. Initial conditions for premixed H2/O2 constant-volume combustion simulations.
The initial pressures of 30 and 40 MPa were selected to represent ultra-high-pressure confined H2-O2 combustion conditions relevant to high-pressure combustion devices and safety-critical hydrogen systems. The H2/O2 molar ratios of 8:1 and 12:1 correspond to highly fuel-rich mixtures, which were selected to examine whether oxygen availability becomes a controlling factor under ultra-high-pressure conditions. The three-, four-, and five-point ignition configurations were used to evaluate the sensitivity of early flame development and reflected interaction to ignition layout.
The chamber geometry was selected as a representative large-volume confined combustion structure with a long cylindrical section and a front contraction section. The long cylindrical section allows flame propagation and reflection to develop over a sufficient axial distance, while the inclined contraction section promotes wave reflection and local intensification. Therefore, this geometry is suitable for investigating flame interaction and ignition layout effects in confined ultra-high-pressure H2/O2 combustion. The present geometry should be regarded as a representative configuration rather than a universal chamber design.
All chamber walls were treated as no-slip and adiabatic boundaries to focus on the coupled effects of flame propagation, reflection, and chamber confinement during the early combustion stage. Wall heat loss and surface reactions were not included in the present model; therefore, the results represent an idealized confined combustion condition and should be interpreted accordingly.

2.2. Governing Equations

The density-based transient solver in ANSYS 2021R2 Fluent was employed to model the compressible reacting flow characteristics of premixed H2/O2 combustion in a closed constant-volume chamber. This solver is appropriate for the present problem because the combustion process involves strong density variation, rapid propagation, transient heat release, and large pressure gradients under ultra-high initial pressure. Compared with a pressure-based incompressible flow formulation, the density-based solver is more suitable for resolving compressible wave motion and pressure evolution in high-pressure reacting flows. The governing equations include the conservation equations of mass, momentum, and total energy, which can be written as follows:
∂ ρ ∂ t + ∇ ⦁ ρ U = 0 ∂ ρ U ∂ t + ∇ ⦁ ρ U ⊗ U + ∇ p = ∇ ⦁ μ ∇ U + ∇ U T + ρ g ∂ ρ E ∂ t + ∇ ⦁ ρ U E ⊗ U + ∇ p = ∇ ⦁ k ∇ T + ϕ + ρ U ⦁ g
where ρ is the fluid density, u is the velocity vector, p is the pressure, μ is the dynamic viscosity, g is the body force acceleration, E is the total energy per unit mass, k is the thermal conductivity, T is the temperature, and Φ is the viscous dissipation term. The operator ∇ denotes the spatial gradient operator. The three equations represent mass conservation, momentum conservation, and total energy conservation, respectively.
In the present study, the solver is used to analyze flame propagation, flame acceleration, and pressure evolution rather than to directly identify fully developed detonation. The finite-rate chemical kinetic model is coupled with the compressible flow solver to describe transient heat release, species conversion, and flame interaction during confined H2/O2 combustion. Therefore, the numerical framework is consistent with the main objective of this work, namely to investigate how initial pressure, mixture ratio, and ignition configuration affect flame acceleration and pressure response in an ultra-high-pressure constant-volume chamber.

2.3. Real Equation of State

Regarding the equation of state, the Peng–Robinson equation of state was adopted to account for real compressibility under ultra-high initial pressure. Compared with the ideal gas assumption, the PR EOS provides a more appropriate description of non-ideal gas behavior at the initial pressures of 30–40 MPa considered in this study. In the present work, the PR EOS is used only for the gas-phase H2/O2 mixture, and no gas–liquid phase transition is involved. It should also be noted that the combustion process involves high temperature, strong chemical reaction, and possible dissociation of intermediate species; therefore, the PR EOS should be regarded as an engineering approximation rather than an exact thermodynamic description for all reacting flow states. Its main role in this study is to improve the representation of high-pressure gas compressibility during the early stage of flame propagation and development. Its mathematical formulation is as follows:
p = R T V − b − a V V + b + b V − b
where p is the pressure, T is the temperature, V is the molar volume, R is the universal gas constant, and a and b are the attraction and co-volume parameters of the PR EOS, respectively. The parameters a and b are calculated from the critical temperature Tc, critical pressure Pc, and acentric factor-related correction coefficient κ . The calculation formulas are as follows:
a = 0.45724 R 2 T c 2 P c 1 + κ 1 − T / T c 2
b = 0.07780 R T c P c
In Equations (3) and (4), Tc and Pc denote the critical temperature and critical pressure of the gas species, respectively. The parameter κ is related to the acentric factor and is used to correct the temperature dependence of the attractive term. For gas mixtures, the mixture parameters are evaluated using the mixing rules implemented in ANSYS Fluent.
Therefore, the influence of the EOS selection on the predicted pressure history remains a source of modeling uncertainty and should be further evaluated in future work by comparing PR EOS results with ideal and other real EOS formulations.

2.4. LES Turbulence Model and Turbulence–Combustion Treatment

Under ultra-high-pressure confined H2/O2 combustion conditions, rapid heat release, flame expansion, and reflection generate strongly unsteady flow structures and large local velocity gradients. In order to resolve the transient deformation and the interaction between pressure waves and large-scale vortical motion, large eddy simulation (LES) was employed in the present study.
The filtered governing equation used in LES can be written as
∂ u i ¯ ∂ t + u j ¯ ∂ u i ¯ ∂ x j = − ∂ p ¯ ∂ x i + v ∂ 2 u i ¯ ∂ x j 2 + f i + S G S T e r m
where u ¯ i is the filtered velocity component in the i-direction, p ¯ is the filtered pressure, ρ is the density, ν is the molecular kinematic viscosity, and f i is the body force term. In the LES framework, the effects of unresolved sub-grid turbulent structures are represented through an SGS eddy-viscosity model.
In the present simulations, the wall-adapting local eddy-viscosity (WALE) model was adopted as the SGS closure model. Compared with the classical Smagorinsky model, the WALE model provides improved near-wall behavior because the eddy viscosity naturally approaches zero near solid boundaries without requiring additional wall-damping functions. Therefore, the WALE model is suitable for the present confined reacting flow problem involving strong wall interaction, transient shear layer formation, and flame interaction.
The SGS eddy viscosity in the WALE model is expressed as
μ t = ρ Δ 2 S i j d S i j d 3 / 2 S ¯ i j S ¯ i j 5 / 2 + S i j d S i j d 5 / 4
where μ t is the SGS eddy viscosity, Δ is the local filter width related to the computational grid, S ¯ i j is the resolved strain rate tensor, and S i j d is the traceless symmetric part of the square of the velocity gradient tensor.
The initial gas mixture was assumed to be homogeneous and quiescent, with an initial velocity of 0 m/s. No prescribed initial turbulence intensity or artificial turbulent fluctuation was imposed at the initial condition. Therefore, the unsteady flow structures observed during the calculation were generated by flame expansion, wall reflection, interaction, and shear layer instability during the transient combustion process. Turbulence–chemistry interaction was treated using finite-rate chemistry directly coupled with the LES-resolved flow field. No flamelet, presumed probability density function (PDF), or flame surface density model was introduced. The local reaction rates were calculated using the reduced H2/O2 kinetic mechanism together with the local temperature, pressure, and species concentrations obtained from the LES-resolved transient flow field.
It should be noted that the present LES calculation resolves only the large-scale turbulent structures, while the smallest dissipative turbulent scales remain modeled. Therefore, the predicted wrinkling and pressure oscillations may still depend on the mesh resolution, SGS model selection, and temporal resolution [22].

2.5. Chemical Kinetic Mechanism

Regarding the selection of reaction mechanisms for flame acceleration processes, this study references the research findings of Zhou Fei et al. [25] and adopts the modified H2-O2 detailed reaction mechanism proposed by Burke et al. [26,27], which comprises 8 components and 32 elementary reactions. To enhance computational efficiency, the 19-step reaction mechanism was simplified using the DRGEP algorithm. The resulting skeletal mechanism, as shown in Table 2, comprises 8 components and 19 reaction steps. Comparative calculations were conducted between the skeletal model and the detailed mechanism model using a zero-dimensional closed reactor, with results presented in Figure 2 and Figure 3. The computational results demonstrate that the simplified mechanism exhibits high consistency with the detailed model in terms of sensitivity to key combustion parameters, combustion temperature, and reaction rates, with maximum errors below 6%. This indicates appropriate selection of the mechanism simplification algorithm, confirming that the skeletal model maintains excellent applicability and computational accuracy.
Table 2. Detailed H2/O2 reaction mechanism based on Burke et al. [26].
Figure 2. Sensitivity analysis of different reaction mechanisms. The gray bars represent the detailed chemical reaction mechanism, and the pink bars represent the simplified chemical reaction mechanism. Positive and negative values indicate promoting and inhibiting effects on the selected target variable, respectively. (a) Analysis of OH radical sensitivity, (b) Analysis of reaction temperature sensitivity.
Figure 3. Time-dependent curves of combustion parameters under different reaction mechanisms. (a) Temperature change curve of the reaction, (b) H2 concentration change curve, (c) O2 concentration change curve, (d) OH concentration change curve, (e) H2O concentration change curve.
This reaction mechanism demonstrates high computational accuracy for flame propagation rate calculations of H2-O2 premixed gases. The validation of its precision is demonstrated in the related research by Zhou Fei et al. [25] and will not be elaborated further in this paper.

2.6. Numerical Implementation in ANSYS Fluent

The simulations were performed using ANSYS Fluent 2021 R2 (ANSYS Inc., Canonsburg, PA, USA) with the density-based transient solver for compressible reacting flow. The governing equations were discretized using second-order spatial schemes, and time advancement was performed using an implicit transient formulation. The density-based formulation was selected because the cases involve rapid heat release, large density variation, and propagation under ultra-high initial pressure.
The chamber wall was treated as a no-slip and adiabatic boundary. The computational domain was initialized using a homogeneous premixed H2/O2 mixture at 298 K and the specified initial pressure. Ignition was modeled by spherical high-temperature kernels with a diameter of 5 mm and an initial temperature of 1500 K.
The initial velocity of the premixed gas was set to 0 m/s, and no initial turbulence intensity was prescribed. Thus, the LES calculation did not rely on imposed inlet turbulence or artificial initial perturbations. The large-scale unsteady structures developed during the simulation as a result of flame expansion, reflection, wall interaction, and shear layer formation.
The finite-rate chemistry model was coupled directly with the LES-resolved flow field. Radiation heat transfer, wall heat loss, and surface reactions were neglected in the present calculation because the present work mainly focuses on the early-stage flame propagation and pressure evolution under ultra-high-pressure conditions.
The transient time step was selected according to the Courant-Friedrichs-Lewy (CFL) condition for compressible reacting flow. During the main combustion stage, the CFL number was maintained below 1 in the refined flame propagation region to ensure the numerical stability and sufficient temporal resolution of propagation. Pressure histories were monitored at the contraction section, and characteristic quantities, including onset time, maximum rate, and pressure fluctuation amplitude, were extracted from these histories for quantitative comparison.
Residual convergence was monitored during each time step. The residual criterion for continuity, momentum, energy, and species equations was set to 10−5.
Species transport equations were solved together with the finite-rate chemical reaction model. The energy equation was enabled for all simulations. Pressure, momentum, energy, and species transport equations were solved in a fully coupled density-based framework.
For clarity and reproducibility, the principal numerical settings used in the present CFD simulations are summarized in Table 3, including the solver configuration, turbulence model, combustion treatment, temporal discretization, and convergence criteria.
Table 3. Main numerical settings used in the CFD simulations.

3. Model Validation and Numerical Independence Verification

The present simulations focus on flame propagation, development, and flame acceleration in a confined ultra-high-pressure chamber. Since quantitative detonation criteria, such as local shock speed, CJ speed comparison, and von Neumann pressure state, are not fully evaluated in this work, the term “detonation” is avoided unless directly supported by measurable indicators.

3.1. Validation of Flame Propagation and Pressure Evolution Trends

The quantitative comparison indicates that the predicted time agrees reasonably well with the reference value reported by Zhou Fei et al. [25], with a relative error of approximately 3.8% for the selected 0.2 mm grid.
It should be noted that the available reference data do not provide complete time-resolved pressure histories, local shock wave information, or detailed flame speed measurements under exactly the same geometry and operating conditions as the present study. Therefore, the present validation is limited mainly to flame propagation behavior and global pressure evolution trends.
Consequently, the present CFD framework should be regarded as a partially validated numerical model suitable for comparative analysis of different operating conditions, rather than as a fully experimentally calibrated predictive tool for quantitative detonation analysis.
The quantitative validation metrics reported in Table 4 provide a direct comparison between the reference data and the present simulation, thereby improving the transparency of model assessment.
Table 4. Quantitative validation of the numerical framework.

3.2. Mesh Independence Verification

A mesh independence assessment was performed to evaluate the influence of spatial resolution on flame propagation and pressure evolution. Three grid resolutions, 0.2 mm, 0.4 mm, and 0.8 mm, were tested using the same 900 mL constant-volume combustion chamber configuration. The validation case was based on the high-pressure H2/O2 combustion configuration reported by Zhou Fei et al. [25].
The validation model was discretized using a two-dimensional axisymmetric structured quadrilateral mesh. Local refinement was applied near the ignition region, chamber wall, and contraction section, where strong deformation and variation are expected. The nominal grid sizes of the three tested meshes were 0.2 mm, 0.4 mm, and 0.8 mm, respectively.
The chamber wall was treated as a no-slip and adiabatic boundary. Since the present study focuses mainly on flame propagation and flame interaction rather than wall-resolved heat transfer quantitative engineering predictions, y+ was not used as the primary grid selection criterion. Instead, grid suitability was evaluated using time, onset, maximum pressure, and resolution. The near-wall cell size was kept consistent with the corresponding nominal grid size in each mesh case.
Figure 4 compares the pressure histories at the contraction section predicted using different grid resolutions. The results indicate that the grid resolution has a clear influence on the predicted onset, sharpness, and oscillation amplitude.
Figure 4. Pressure histories at the contraction section obtained using different grid resolutions. The red, cyan, and pink curves represent the results obtained using grid sizes of 0.2 mm, 0.4 mm, and 0.8 mm, respectively.
In general, a coarser grid introduces stronger spatial numerical dissipation, which tends to smooth local pressure gradients and weaken high-frequency components. Therefore, the smoother pressure response and delayed onset predicted by the 0.8 mm grid are mainly attributed to the insufficient spatial resolution of the flame front and region.
Compared with the 0.8 mm grid, the 0.4 mm grid improves the resolution of propagation and produces a pressure history closer to that of the 0.2 mm grid. The 0.2 mm grid provides the sharpest process and better captures the transient interaction between the flame front and reflected pressure waves.
It should be emphasized that the pressure oscillations predicted by different grids should not be interpreted solely in terms of oscillation amplitude. For coarse grids, pressure fluctuations may be distorted by numerical dissipation and insufficient resolution. Therefore, the 0.2 mm grid was selected because it provides more reliable resolution of the coupled and structures, rather than simply because it produces larger or smaller pressure oscillations.
The use of multiple quantitative indicators, including time, maximum pressure, onset, and relative error, enables a more comprehensive assessment of mesh independence than reliance on a single parameter.
The differences in maximum pressure and onset among the three grids are mainly caused by the different levels of spatial numerical dissipation. The 0.8 mm grid smooths the field and delays propagation, leading to a later onset and lower maximum pressure. The 0.2 mm grid reduces this numerical dissipation and better resolves the interaction. Therefore, the pressure oscillation behavior obtained using the 0.2 mm grid is considered more reliable for the following analysis.
Although the 0.4 mm grid produces a time very close to the reference value, mesh selection in the present LES calculation should not be based on a single global indicator. The 0.2 mm grid provides better resolution of local deformation, structure, and flame interaction, which are the primary focuses of this study.
Table 5 summarizes the quantitative comparison among the three grid resolutions. The flame arrival time predicted using the 0.2 mm grid is approximately 2.50 ms, which is close to the reference value of 2.6 ms reported in Ref. [25]. The 0.4 mm grid also gives a time close to the reference value; however, its onset is slightly delayed and the maximum pressure is lower than that predicted using the 0.2 mm grid.
Table 5. Mesh parameters and quantitative comparison of mesh independence results.
The 0.8 mm grid shows the largest deviation in time and produces the lowest maximum pressure, indicating that this grid is too coarse to resolve the transient flame interaction adequately. Although the 0.4 mm grid provides an acceptable estimate of the global time, the 0.2 mm grid better resolves the structure and deformation during the later combustion stage. Therefore, the 0.2 mm grid was selected for the subsequent simulations as a compromise between numerical accuracy and computational cost.
Figure 5 compares the flame temperature distributions obtained using the three grid resolutions. The 0.2 mm grid resolves a clearer structure and captures more detailed flame deformation during the flame acceleration stage. The flame reaches the contraction section at approximately 2.50 ms using the 0.2 mm grid, which is consistent with the reference time scale of 2.6 ms reported by Zhou Fei et al. [25].
Figure 5. Flame temperature distributions predicted using different grid resolutions. (a) 0.2 mm, (b) 0.4 mm, (c) 0.8 mm.
The 0.4 mm grid produces a similar global flame propagation trend, but the local deformation and high-temperature region are less sharply resolved. The 0.8 mm grid produces a smoother flame structure and delayed flame arrival, indicating that excessive spatial numerical dissipation affects the quantitative engineering predictions of transient flame propagation.
Figure 6 shows the instantaneous velocity field distribution during the flame acceleration stage. A high-speed flow region is formed near the leading reaction zone, indicating that flame expansion and motion induce strong local acceleration of the unburned gas. This velocity structure reflects the coupling between the flame front, pressure waves, and confined chamber geometry. The enlarged view shown in Figure 6 also facilitates identification of the local high-velocity region formed ahead of the leading reaction zone.
Figure 6. Instantaneous velocity field distribution near the leading reaction zone during the flame acceleration stage. The color contour represents the velocity magnitude. The red rectangle marks the local high-velocity region near the leading reaction zone, and the red lines indicate the enlarged view of this region.
The concentrated high-velocity region ahead of the reaction zone suggests that motion contributes to local flow acceleration during flame propagation. However, because quantitative detonation criteria such as local shock speed, Chapman-Jouguet velocity comparison, and von Neumann pressure state are not evaluated in the present work, this velocity field result is interpreted as evidence of flame acceleration and flame interaction rather than definitive proof of fully developed detonation.
Overall, the mesh independence analysis indicates that the 0.2 mm grid provides improved resolution of deformation and interaction compared with the 0.4 mm and 0.8 mm grids. The 0.8 mm grid is insufficient for resolving the transient combustion process, while the 0.4 mm grid gives acceptable global time but less detailed and structures. Based on the above comparison, the difference in pressure oscillation behavior among the three grids is mainly attributed to spatial numerical dissipation and the resolution of interaction. The 0.2 mm grid provides the most suitable resolution among the tested grids and is therefore adopted in the subsequent simulations [28,29].

3.3. Time-Step Independence Verification

In addition to the mesh independence assessment, a time-step independence analysis was performed to evaluate the influence of temporal resolution on the predicted pressure evolution and flame propagation behavior. Three transient time steps, 5 × 10−8, 1 × 10−7, and 2 × 10−7, were tested using the 0.2 mm grid under the same operating condition.
The quantitative comparison results are summarized in Table 6. The results show that the onset time and the main pressure evolution trend predicted using 1 × 10−7 s and 5 × 10−8 s are very similar. In contrast, the larger time step of 2 × 10−7 s produces slightly delayed propagation and smoother pressure fluctuations because of increased temporal numerical dissipation.
Table 6. Time-step independence analysis.
The relative differences in onset time and maximum pressure between 5 × 10−8 s and 1 × 10−7 s remain below 2%, indicating acceptable temporal convergence. Therefore, the time step of 1 × 10−7 s was selected as a compromise between computational accuracy and computational cost in the present simulations.

4. Effects of Initial Conditions on Flame Propagation and Pressure Evolution

To reduce the dependence on qualitative contour interpretation, several characteristic quantities were extracted from the simulations, including onset time, flame coalescence time, maximum rate, and pressure fluctuation amplitude. These indicators provide quantitative support for comparing the effects of initial pressure, mixture ratio, and ignition configuration. In this study, the onset time refers to the time at which a sustained pressure increase is observed in the monitored pressure history after ignition; the maximum rate is obtained from the steepest segment of the pressure/time curve; and the pressure fluctuation amplitude is used to characterize the intensity of oscillatory pressure response after flame acceleration.
In the following discussion, a distinction is made between direct numerical observations and physical interpretations. The temperature fields, distributions, pressure histories, and quantitative indicators are treated as numerical observations. Explanations related to flame interaction, oxygen-limited heat release, and large-scale unsteady flow development are interpreted as possible physical interpretation inferred from the CFD results. These interpretations by additional experimental measurements or dedicated diagnostic simulations.
The following discussions are based on comparative numerical analysis using the partially validated CFD framework described in Section 3.

4.1. Analysis of Initial Pressure Effects on Flame Propagation Mechanisms

Figure 7 and Figure 8 illustrate the combustion chamber flame acceleration process under initial pressure conditions of 30/40 MPa, H2-O2 molar ratio of 8:1, and three-point ignition conditions.
Figure 7. Combustion chamber flame acceleration process under 30 MPa pressure, H2/O2 molar ratio of 8:1, and three-point ignition conditions. The time labels in each panel indicate the elapsed time after ignition. In each time-resolved image, the upper contour field represents the pressure-gradient distribution, and the lower contour field represents the temperature distribution. The color bars indicate the corresponding magnitudes of pressure gradient and temperature, respectively. Warmer colors indicate regions with higher pressure gradient or higher temperature, while cooler colors indicate lower values.
Figure 8. Combustion chamber flame acceleration process under 40 MPa pressure, H2/O2 molar ratio of 8:1, and three-point ignition conditions. The time labels denote the elapsed time after ignition. The upper contour field in each panel shows the pressure-gradient distribution, while the lower contour field shows the temperature distribution. The color scales represent the magnitudes of pressure gradient and temperature, respectively. Warmer colors correspond to higher values, and cooler colors correspond to lower values.
When the initial pressure reaches 30 MPa, three distinct and nearly symmetrical flame nuclei become visible at 0.05 ms, maintaining morphological characteristics similar to laminar flames. The corresponding pressure gradient exhibits a clear concentric ring distribution, indicating that rapid heat release during ignition initially triggers outward propagating compression waves. By 0.5 ms, pressure gradient waves from different ignition sources undergo reflection and superposition at wall surfaces and geometric boundaries, forming pronounced cross-wave structures. These compression waves induce periodic pre-compression and preheating in the unburned mixture, creating favorable conditions for subsequent flame acceleration. At this stage, the flame remains predominantly in a “separate expansion” phase. The instability and complex morphological evolution of this flame front share similar fluid dynamic instability origins with classic phenomena such as the “tulip flame” generated by hydrogen/air premixed flames in confined ducts at lower pressures [21].
Between 1.0 and 1.2 milliseconds, the flame core significantly enlarges and begins to converge, with the flame front transitioning from a smooth morphology to a wavy, wrinkled structure. Concurrently, the pressure gradient intensifies markedly in the narrow region between the flame front and the flame body, indicating localized focusing of compressive waves at the flame fusion zone. The resulting high-pressure region coincides spatially with the flame front, suggesting that the pressure wave has evolved from passive propagation to actively driving flow acceleration, flame stretching, and entrainment. This establishes a positive feedback mechanism of heat release/compression/acceleration, marking the critical phase for rapid flame acceleration.
After entering the 1.7 ms phase, the flames have merged into a continuous flame band extending along the combustion chamber, with the front edge exhibiting distinct wavy and undulating structures. Concurrently, the pressure gradient in the lower half evolves from an initially regular corrugated pattern into a highly turbulent high-gradient network, indicating that the LES-resolved temperature and fields show increasingly distorted flame fronts and irregular high-gradient structures, indicating the development of large-scale unsteady vortical motion induced by flame expansion and reflected pressure waves. However, quantitative turbulence statistics such as resolved turbulent kinetic energy and vorticity magnitude were not extracted in the present work; therefore, the discussion of enhanced large-scale unsteady motion remains mainly qualitative. During this stage, near-wall reflected waves interact with large-scale vortices induced by flame expansion, continuously entraining unburned mixtures into the reaction zone. Flame propagation transitions from ignition source control to unsteady flame propagation dominated by large-scale flow motion, resulting in significant increases in both the overall heat release rate and average pressure within the combustion chamber.
Compared to 30 MPa conditions, the transition from ignition core development to accelerated flame propagation occurs earlier at 40 MPa. The pressure gradient field evolves more rapidly from regular reflected wave ripples into a dense high-gradient network, indicating that elevated initial pressure significantly enhances compressive wave intensity and shortens local chemical induction time. This strengthens wave/flame coupling and turbulent flow generation, accelerating the flame propagation mechanism from wave-induced acceleration to strong unsteady flame propagation dominated by large-scale flow motion. The increased initial pressure thus drives the flame acceleration process through three synergistic effects: accelerated chemical reactions, enhanced compressibility coupling, and premature turbulent instability development. These mechanisms collectively increase the probability of localized high compression and hot spot formation during flame propagation.
The quantitative indicators in Table 7 further support this observation. When the initial pressure increases from 30 MPa to 40 MPa at the same H2/O2 molar ratio of 8:1 and three-point ignition, the onset advances from approximately 1.65 ms to 1.28 ms, and the maximum rate increases from 18.6 MPa·ms−1 to 27.4 MPa·ms−1. This indicates that the higher initial pressure accelerates the global pressure response and strengthens the coupling between heat release and compressible wave motion.
Table 7. Quantitative indicators extracted from the numerical results.
In the present work, enhanced large-scale unsteady motion is not evaluated using time-averaged turbulence statistics. Instead, it is inferred from the LES-resolved deformation, irregular structures, and increased rate. Therefore, the discussion should be interpreted as a qualitative assessment of large-scale unsteady flow development rather than a direct measurement of turbulent kinetic energy.
Figure 9 presents pressure fluctuation curves at the nozzle throat of premixed gas combustion chambers under varying initial pressures and H2/O2 molar ratios. Comparative analysis between initial pressures of 30 MPa and 40 MPa reveals that all operating conditions initially exhibit a plateau phase with stable pressure levels, corresponding to flame nucleation and early development stages where overall heat release remains insufficient to significantly elevate chamber average pressure. As combustion progresses, pressure rise characteristics demonstrate distinct variations across different initial pressure conditions.
Figure 9. Pressure histories at the contraction section under different initial pressures and H2/O2 molar ratios. The pink curve represents the 30 MPa, H2/O2 = 8:1 case; the orange curve represents the 40 MPa, H2/O2 = 8:1 case; and the red curve represents the 40 MPa, H2/O2 = 12:1 case. The horizontal axis denotes time after ignition, and the vertical axis denotes pressure at the monitored contraction section.
Compared to the 30 MPa initial pressure condition, the pressure rise initiation under the 40 MPa condition occurs significantly earlier, with a markedly increased growth rate during the rapid pressurization phase. While the condition requires approximately 1.6–1.8 ms to enter the distinct pressurization stage, the 40 MPa condition initiates sustained pressure escalation within 1.2–1.4 ms and 30 MPa reaches high-pressure levels in a shorter timeframe. This demonstrates that increasing initial pressure effectively shortens the induction phase of flame propagation reactions and accelerates the conversion of total heat release into combustion chamber pressure. The underlying mechanisms include: higher initial pressure leading to increased gas mixture density and molecular collision frequency, facilitating rapid establishment of H2-O2 chain reactions; and enhanced compressibility effects that amplify early thermal release, thereby more readily triggering global pressure responses.
From the numerical results, the 40 MPa case shows earlier deformation, denser structures, and an earlier onset than the 30 MPa case. These are direct observations obtained from the LES temperature fields, contours, and pressure history data.
The corresponding physical interpretation is that the higher initial pressure increases the gas density and shortens the chemical induction time, which may strengthen the coupling between heat release and compressible wave motion. This coupling may further promote large-scale unsteady flow development and distortion. However, this mechanism is inferred from the present CFD results and has not been independently verified by direct measurements of local shock speed, local heat release rate, or turbulence statistics.

4.2. Analysis of the Effect of Gas Mixing Ratio on Flame Propagation Mechanism

Figure 10 illustrates the flame acceleration process under 40 MPa initial pressure, H2/O2 molar ratio of 12:1, and three-point ignition conditions. Compared with the 40 MPa and H2/O2 = 8:1 case, increasing the H2/O2 molar ratio to 12:1 delays the onset from approximately 1.28 ms to 1.46 ms and reduces the maximum rate from 27.4 MPa·ms−1 to 21.1 MPa·ms−1. The pressure fluctuation amplitude also decreases from 5.6 MPa to 3.8 MPa. These quantitative results indicate that the richer mixture produces a slower and smoother pressure response.
Figure 10. Flame acceleration process under 40 MPa initial pressure, H2/O2 molar ratio of 12:1, and three-point ignition conditions. The time labels indicate the elapsed time after ignition. The upper contour field in each panel represents the pressure-gradient distribution, and the lower contour field represents the temperature distribution. The color bars indicate the corresponding magnitudes of pressure gradient and temperature. Warmer colors denote higher pressure-gradient or temperature regions, whereas cooler colors denote lower-value regions.
The numerical results show that the 12:1 case exhibits delayed flame coalescence and weaker pressure growth behavior than the 8:1 case under the same initial pressure. These observations are directly supported by the pressure history data in Figure 9 and the quantitative indicators listed in Table 7.
The inferred mechanism is that increasing the H2/O2 molar ratio under the investigated fuel-rich condition reduces the effective oxygen availability, thereby weakening the volumetric heat release intensity and delaying flame coalescence. This explanation is consistent with the observed pressure response, but it should be regarded as a mechanism inferred from the numerical results rather than a directly measured chemical/kinetic pathway.
Overall, the H2/O2 molar ratio regulates the smoothness and unsteadiness of pressure evolution by influencing effective oxygen availability and heat release rate. These results indicate that mixture composition is an important control parameter for flame acceleration and pressure evolution in premixed H2/O2 constant-volume combustion.

4.3. Analysis of Flame Propagation Mechanism in Multi-Point Ignition Systems

The flame development during four-point ignition within the 0–140 μs period is illustrated in Figure 11. During the 0–40 μs period, multiple ignition sources simultaneously form independent flame kernels, with corresponding pressure gradients expanding outward from each ignition location. As time progresses, the zones interact with the chamber wall and with neighboring ignition kernels, promoting early flame deformation and local acceleration.
Figure 11. Initial flame development under four-point ignition conditions. The time labels indicate the elapsed time after ignition. The contour colors represent the pressure-gradient distribution generated by the four ignition kernels and their reflected waves. Warmer colors indicate regions with stronger pressure-gradient intensity, while cooler colors indicate weaker pressure-gradient regions. The four bright circular regions at the early stage correspond to the initial ignition kernels.
Due to the higher initial heat release rate associated with multi-point simultaneous ignition, pressure gradient fluctuations become more pronounced, facilitating reflection and superposition at the chamber wall and inclined contraction section. Within 60–140 μs, the compression wave field shows clear convergence and reflection patterns, suggesting that multi-point compression waves form localized wave focusing and strong compression zones after wall reflection.
This process has two effects: it intensifies pre-compression and preheating of the unburned gas, and it induces velocity fluctuations that enhance wrinkling. As a result, the flame morphology changes from a smooth leading edge to a wrinkled structure. This phase represents a critical stage for early flame acceleration and flame interaction.
The mid-to-late flame development phase (0.5–3 ms) is illustrated in Figure 12. The flame front evolves from multiple flame points merging into a continuous combustion zone extending along the channel, exhibiting distinct undulations and wavy structures. The field transitions from initially regular radiated/reflected wave patterns to an irregular high-gradient network, indicating the development of a strongly unsteady flow state with intensified distortion and fluctuation, indicating that the flow field has entered a strongly unsteady flow state with intensified distortion and fluctuation. This dynamic process drives rapid entrainment of unburned gases into the reaction zone, thereby sustaining flame acceleration.
Figure 12. Development of the flame during the middle and late stages under four ignition conditions. The time labels indicate the elapsed time after ignition. The upper contour field in each panel represents the pressure-gradient distribution, and the lower contour field represents the temperature distribution. The color scales show the magnitudes of pressure gradient and temperature, respectively. Warmer colors correspond to higher pressure-gradient or higher-temperature regions, while cooler colors represent lower values. The light-colored regions indicate the main high-temperature flame zone formed after flame-kernel expansion and coalescence.
Figure 13 and Figure 14 illustrate the initial and mid-to-late-stage flame development under five-point ignition conditions. Compared to four-point ignition, the total flame surface area expands more rapidly, with cross-grid-like high-gradient ripples emerging earlier in the field and forming more continuous high-gradient zones between ignition points. This indicates that multi-source compression waves are more prone to phase superposition and reflection coupling between adjacent ignition points, thereby establishing a more efficient pressure gradient network across the channel. The increased number and more uniform distribution of ignition sources typically result in denser superimposed pressure gradient wave systems, accelerating the formation of extensive shear layer disturbances and vortex structures within the channel. These phenomena enhance the entrainment and mixing of unburned gases, ultimately improving flame development continuity.
Figure 13. Initial flame development under five-point ignition conditions. The time labels indicate the elapsed time after ignition. The contour colors represent the pressure-gradient distribution generated by the five ignition kernels and their reflected waves. Warmer colors correspond to stronger pressure-gradient regions, while cooler colors correspond to weaker regions. The five bright circular regions at the early stage represent the initial ignition kernels.
Figure 14. Development of the flame during the mid-to-late stages under five-point ignition conditions. The time labels denote the elapsed time after ignition. The upper contour field in each panel represents the pressure-gradient distribution, and the lower contour field represents the temperature distribution. The color bars indicate the corresponding magnitudes of pressure gradient and temperature. Warmer colors indicate higher pressure-gradient or higher-temperature regions, while cooler colors indicate lower-value regions. The light-colored regions correspond to the high-temperature flame zone after flame expansion and interaction.
Figure 15 compares the pressure histories at the contraction section of the combustion chamber for three-, four-, and five-point ignition configurations under 30 MPa initial pressure and H2/O2 molar ratio of 8:1. During the early stage of 0–1.5 ms, the three pressure curves are close to each other because the flame kernels are still developing independently and the total heat release is insufficient to cause a significant global pressure rise.
Figure 15. Pressure histories at the contraction section for premixed H2/O2 combustion under 30 MPa initial pressure, H2/O2 = 8:1, and different ignition configurations. The pink curve represents the three-point ignition case, the orange curve represents the four-point ignition case, and the red curve represents the five-point ignition case. The horizontal axis denotes time after ignition, and the vertical axis denotes pressure at the monitored contraction section.
After approximately 1.8 ms, clear differences appear among the three ignition configurations. These differences are mainly associated with the interaction among expansion, reflection, and flame coalescence. Compression waves generated by rapid heat release are repeatedly reflected between the chamber wall, adjacent ignition kernels, and the contraction section. Depending on the ignition point spacing, these reflected waves may undergo constructive or destructive superposition, resulting in different rates and oscillation amplitudes.
In the present chamber geometry, the four-point ignition case shows a stronger response. This may be attributed to a more favorable phase relationship between reflected compression waves and expanding flame fronts. In contrast, the five-point ignition case distributes the initial heat release more uniformly along the chamber axis, which tends to smooth the pressure buildup and reduce the pressure fluctuation amplitude.
The numerical results indicate that the ignition point number affects onset, maximum rate, and pressure fluctuation amplitude. The pressure oscillations mainly arise from repeated reflection and superposition of compression waves between the chamber wall, adjacent ignition kernels, and the contraction section. As the flame fronts expand and merge, the spatial distribution of heat release changes continuously, which further modifies the phase relationship between pressure waves and flame fronts.
A possible interpretation is that the four-point ignition layout produces a more favorable coupling between reflected compression waves and expanding flame fronts, thereby enhancing local flame interaction. In contrast, the five-point ignition layout may distribute the initial heat release more uniformly, leading to smoother pressure buildup. This interpretation is geometry-dependent and is inferred from the present pressure histories and flame contour evolution. It should not be generalized to other chamber configurations without further validation.
The geometric dependence mainly arises from the fact that reflection, wave focusing, and merging are strongly affected by chamber length, chamber diameter, contraction angle, and ignition point spacing. In another chamber geometry, the phase relationship among reflected compression waves may change, and the ignition layout producing the strongest response may also be different. Therefore, the present conclusion that four-point ignition produces the strongest behavior should be interpreted as a configuration-specific result for the present chamber.
Table 7 shows that changing the ignition point number affects both the rate and pressure fluctuation amplitude. For the 30 MPa and 8:1 cases, increasing the ignition point number from three to four advances the onset from approximately 1.65 ms to 1.54 ms and increases the maximum rate from 18.6 MPa·ms−1 to 23.5 MPa·ms−1. This suggests that the four-point ignition layout promotes stronger early flame surface growth and interaction in the present chamber geometry.
However, further increasing the ignition point number to five does not produce a monotonic increase in intensity. The five-point case shows a maximum rate of approximately 20.2 MPa·ms−1 and a lower pressure fluctuation amplitude of 3.6 MPa. This indicates that the five-point ignition layout produces a more distributed and smoother combustion process rather than the strongest pressure buildup.
Since resolved turbulent kinetic energy and vorticity were not quantitatively extracted, the role of turbulence is discussed indirectly through flame coalescence time, rate, and pressure fluctuation amplitude. The conclusion regarding ignition layout effects should therefore be interpreted as geometry-dependent and based on the present numerical indicators.
From an engineering perspective, the non-monotonic effect of ignition point number should not be used directly as a universal ignition design rule. The present result indicates that ignition layout interacts with the chamber geometry through reflection and flame coalescence. Additional simulations with different chamber lengths, diameters, contraction angles, and ignition point spacings are required before extending the ignition configuration conclusion to other confined combustion systems.
It should be noted that the present discussion of turbulence-related behavior is mainly based on LES-resolved instantaneous flame structures, distributions, rates, and pressure fluctuation amplitudes. Quantitative turbulence statistics, such as resolved turbulent kinetic energy, vorticity magnitude, and SGS kinetic energy, were not extracted in the present work. Therefore, statements related to turbulence enhancement are interpreted as qualitative physical explanations rather than direct turbulence measurements.
Overall, the results presented in this section include both direct numerical observations and inferred physical interpretations. The onset, flame coalescence time, rate, and pressure fluctuation amplitude are directly extracted from the numerical data. In contrast, explanations involvingflame coupling, oxygen-limited heat release, and ignition layout-dependent wave focusing are inferred mechanisms. These interpretations are physically plausible but should be further validated using local heat release rate fields, shock speed measurements, turbulence statistics, or experimental diagnostics.

4.4. Advantages and Limitations of the Numerical Framework

Compared with zero-dimensional models or simplified reactor calculations, the present CFD framework can resolve spatial flame development, reflection, and ignition layout effects in a confined chamber. Compared with Reynolds-averaged turbulence models, the LES-WALE approach provides a better description of transient large-scale unsteady structures generated by flame expansion and wall reflection.
Nevertheless, several limitations remain. First, only one chamber geometry is considered. Second, the initial mixture is assumed to be homogeneous and quiescent. Third, wall heat loss, radiation, and surface reactions are neglected. Fourth, only pure H2/O2 mixtures are considered. Fifth, the validation remains partial because complete time-resolved experimental pressure histories and local flame speed measurements under the same ultra-high-pressure conditions are not available. These limitations should be addressed in future experimental and numerical studies.

5. Conclusions

In this study, premixed H2/O2 combustion in a large-volume constant-volume chamber was numerically investigated under ultra-high initial pressures of 30–40 MPa. The effects of initial pressure, H2/O2 molar ratio, and ignition point number on flame propagation and pressure evolution were analyzed using a transient CFD framework combining the Peng–Robinson equation of state, LES with the WALE SGS model, and a reduced H2/O2 finite-rate chemical kinetic mechanism. Based on the numerical results obtained in the present chamber geometry, the following conclusions can be drawn.
Increasing the initial pressure from 30 MPa to 40 MPa advances the onset from approximately 1.65 ms to 1.28 ms and increases the maximum rate from 18.6 MPa·ms−1 to 27.4 MPa·ms−1.
Increasing the H2/O2 molar ratio from 8:1 to 12:1 delays the onset from approximately 1.28 ms to 1.46 ms and reduces the maximum rate from 27.4 MPa·ms−1 to 21.1 MPa·ms−1.
For the 30 MPa and H2/O2 = 8:1 cases in the present chamber geometry, the four-point ignition case gives the largest maximum rate, approximately 23.5 MPa·ms−1, while the five-point case shows a lower pressure fluctuation amplitude. This conclusion is geometry-dependent and should not be directly generalized to other chamber configurations without additional simulations.
The numerical framework used in this study provides a partially validated CFD-based description of flame propagation, interaction, and large-scale unsteady flow development under ultra-high-pressure constant-volume conditions. However, the present results are still subject to modeling assumptions associated with the LES SGS model, reduced reaction mechanism, adiabatic wall condition, homogeneous premixing, and single chamber geometry.
Overall, the present study provides numerical insight into the effects of initial pressure, mixture ratio, and ignition configuration on ultra-high-pressure H2/O2 constant-volume combustion. The conclusions should be interpreted as CFD-based findings for the investigated configuration rather than universal design rules. In particular, the ignition point number effect may vary with chamber geometry because reflection and wave focusing are sensitive to chamber length, diameter, contraction angle, and ignition point spacing. Future work should include quantitative experimental validation, local shock/flame speed measurements, turbulence statistics, wall heat loss effects, and additional chamber geometries to further verify the inferred physical interpretation. Future work will focus on quantitative experimental validation and additional turbulence statistics to further improve the predictive capability of the numerical framework.

Author Contributions

Conceptualization, C.L., W.L. and S.S.; Methodology, C.L., W.L. and S.S.; Software, C.L., W.L. and X.Z.; Validation, W.L.; Formal analysis, W.L.; Investigation, W.L., X.Z. and S.S.; Resources, Y.Z.; Data curation, C.L., X.Z. and Y.Z.; Writing—original draft, C.L.; Writing—review & editing, C.L., W.L., X.Z., Y.Z. and S.S.; Visualization, C.L., C and X.Z.; Supervision, X.Z.; Project administration, X.Z., Y.Z. and S.S.; Funding acquisition, W.L., Y.Z. and S.S. All authors have read and agreed to the published version of the manuscript.

Funding

The author gratefully acknowledges the financial support from Hubei Provincial Natural Science Foundation, 2023AFB900.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding authors.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Bloomberg New Energy Finance. Energy Storage System Costs Survey 2024; Bloomberg Finance LP: New York, NY, USA, 2024. [Google Scholar]
  2. International Energy Agency. Renewables 2024: Analysis and Forecast to 2029; IEA: Paris, France, 2024; Available online: https://www.iea.org/reports/renewables-2024 (accessed on 19 March 2026).
  3. Global Carbon Project. Global Carbon Budget 2024, 2024. Available online: https://www.globalcarbonproject.org/carbonbudget (accessed on 19 March 2026).
  4. Winter, C.J.; Nitsch, J. Hydrogen as an Energy Carrier: Technologies, Systems, Economy; Springer: Berlin/Heidelberg, Germany, 2018. [Google Scholar]
  5. International Renewable Energy Agency. Green Hydrogen Cost Reduction: Scaling up Electrolysists to Meet the 1.5 °C Climate Goal; IRENA: Abu Dhabi, United Arab Emirates, 2024; Available online: https://www.irena.org/-/media/Files/IRENA/Agency/Publication/2020/Dec/IRENA_Green_hydrogen_cost_2020.pdf (accessed on 19 March 2026).
  6. Wang, Y.; Leung, D.Y.C.; Xuan, J. A review on unitized regenerative fuel cell technologies, part B: Unitized regenerative alkaline fuel cell, solid oxide fuel cell, and microfluidic fuel cell. Renew. Sustain. Energy Rev. 2021, 151, 111585. [Google Scholar] [CrossRef] [Scilit]
  7. Bloomberg New Energy Finance. Hydrogen Economy Outlook: Key Insights 2024; Bloomberg Finance LP: New York, NY, USA, 2024. [Google Scholar]
  8. Tong, D.; Nguyen, T.; Nguyen, L. Experimental and Modeling Study of Wall Film Effect on Combustion Characteristics of Premixed Flame in a Constant Volume Combustion Bomb. In Proceedings of the 3rd Annual International Conference on Material, Machines and Methods for Sustainable Development (MMMS2022), Can Tho, Vietnam, 13 October 2022. [Google Scholar] [CrossRef] [Scilit]
  9. Yoon, T.; Kim, K.; Choi, D. Research on characteristics and effects of combustion performance by amplified ignition energy in CVCC system. J. Mech. Sci. Technol. 2018, 32, 5989–5998. [Google Scholar] [CrossRef] [Scilit]
  10. Song, Z.; Zhang, X.; Hou, X.; Li, M. Effect of initial pressure, temperature and equivalence ratios on laminar combustion characteristics of hydrogen enriched natural gas. J. Energy Inst. 2017, 91, 887–893. [Google Scholar] [CrossRef] [Scilit]
  11. Huizhen, L.; Huahua, X.; Jinhua, S. Laminar burning velocity, Markstein length, and cellular instability of spherically propagating NH3/H2/Air premixed flames at moderate pressures. Combust. Flame 2022, 241, 112079. [Google Scholar] [CrossRef] [Scilit]
  12. Su, H.; Huo, J.; Wang, X.; Jiang, L.; Song, Q.; Zhao, D. Flame Propagation and Oscillation in a Millimeter-scale Constant Volume Space. Combust. Sci. Technol. 2020, 193, 1747–1765. [Google Scholar] [CrossRef] [Scilit]
  13. Dongzhi, G.; Haiqiao, W.; Lei, Z. Experimental study of flame-shock wave interaction and cylinder pressure oscillation in confined space. Infrared Laser Eng. 2017, 46, 239004. [Google Scholar] [CrossRef] [Scilit]
  14. Wang, Y.; Zhang, X.; Li, Y.; Qin, S. Experimental study on premixed turbulent flame of methane/air in a spherical closed chamber. Energy Convers. Manag. 2020, 222, 113219. [Google Scholar] [CrossRef] [Scilit]
  15. Tsunemi, A.; Horiko, Y.; Shimura, M. DNS of Flame-Wall Interaction and Heat Transfer in a Constant Volume Vessel. In Proceedings of the 14th International Heat Transfer Conference, Washington, DC, USA, 8 August 2010. [Google Scholar] [CrossRef] [Scilit]
  16. Najim, Y.M.; Mueller, N.; Wichman, I.S. On premixed flame propagation in a curved constant volume channel. Combust. Flame 2015, 162, 3980–3990. [Google Scholar] [CrossRef] [Scilit]
  17. Lin, C.; Luo, K.H.; Fei, L.; Succi, S. A multi-component discrete Boltzmann model for nonequilibrium reactive flows. Sci. Rep. 2017, 7, 2045–2322. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  18. Reyes, M.; Tinaut, F.V.; Giménez, B.; Camaño, A. Combustion and Flame Front Morphology Characterization of H2-CO Syngas Blends in Constant Volume Combustion Bombs. Energy Fuels 2021, 35, 3497−3511. [Google Scholar] [CrossRef] [Scilit]
  19. Gal, T.; Schmelcher, R.; Vacca, A.; Cupo, F.; Chiodi, M.; Casal Kulzer, A. Modeling of Hydrogen Combustion from a 0D/1D Analysis to Complete 3D-CFD Engine Simulations. Energies 2024, 17, 5543. [Google Scholar] [CrossRef] [Scilit]
  20. Ole’s, S.; Mularski, J.; Pawlak-Kruczek, H.; Singh, A.K.; Pozarlik, A. Assessment of Global and Detailed Chemical Kinetics in Supercritical Combustion for Hydrogen Gas Turbines. Energies 2025, 18, 3380. [Google Scholar] [CrossRef] [Scilit]
  21. Qian, C.; Liberman, M.A. Influence of Chemical Kinetics on Tulip Flame Formation in Highly Reactive (H2/Air) and Low Reactive (CH4/Air) Mixtures. Energies 2025, 18, 885. [Google Scholar] [CrossRef] [Scilit]
  22. Boukhelef, M.; Senouci, M.; Alliche, M.; Merouane, H.; Bounif, A. Predictive URANS/PDF Modeling of Unsteady-State Phenomena in Turbulent Hydrogen–Air Flames. Fluids 2025, 10, 258. [Google Scholar] [CrossRef] [Scilit]
  23. Luijten, C.C.M.; Doosje, E.; de Goey, L.P.H. Accurate analytical models for fractional pressure rise in constant volume combustion. Int. J. Therm. Sci. 2009, 48, 1213–1222. [Google Scholar] [CrossRef] [Scilit]
  24. Feidt, M.; Dumitrascu, G.; Lupu, A. Chemical Modeling of Constant-Volume Combustion of the Mixture of Methane and Hydrogen Used in Spark Ignition Otto Cycles. Energies 2023, 16, 4578. [Google Scholar] [CrossRef] [Scilit]
  25. Fei, Z. Mechanism Study of the flame propagation Phenomenon and the Stable Combustion Control of the Combustion Light Gas Gun. Ph.D. Thesis, Nanjing University of Science & Technology, Nanjing, China, 2019. [Google Scholar]
  26. Burke, M.P.; Chaos, M.; Ju, Y. Comprehensive H2/O2 kinetic model for high-pressure combustion. Int. J. Chem. Kinet. 2012, 44, 444–474. [Google Scholar] [CrossRef] [Scilit]
  27. UTRON, Inc. Combustion Light Gas Gun Technology Demonstration Final Report; Report for Contract No. N00014-02-C-0419; Sponsored by the Office of Naval Research; UTRON, Inc.: Manassas, VA, USA, 2006. [Google Scholar]
  28. Emami, S.; Mazaheri, K.; Shamooni, A. LES of flame acceleration and DDT in hydrogen-air mixture using artificially thickened flame approach and detailed chemical kinetics. Int. J. Hydrogen Energy 2015, 40, 7395–7408. [Google Scholar] [CrossRef] [Scilit]
  29. Gamezo, V.N.; Ogawa, T.; Oran, E.S. Numerical simulations of flame propagation and DDT in obstructed channels filled with hydrogen-air mixture. Proc. Combust. Inst. 2007, 31, 2463–2471. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.