Study on Combustion Simplification Mechanism and 3D Simulation of Ammonia/Diesel Dual-Fuel Engine

Abstract

To accurately describe the combustion process of ammonia/diesel dual-fuel, this paper develops a simplified kinetic mechanism for ammonia/diesel dual-fuel based on the decoupling method and a modular approach, comprising 212 components and 620 elementary reactions. The diesel component is represented by four components: n-heptane, n-hexadecane, isohexadecane, and α-methylnaphthalene. The mechanism was validated using shock tube experimental data. The results indicate that the developed mechanism can accurately predict key parameters such as ignition delay time and laminar flame speed under different ammonia-blending ratios, showing good agreement with experimental values. Single-component ignition delay prediction error ≤ 6%; laminar flame speed deviation error ≤ 2%; CFD validation metrics (e.g., peak cylinder pressure error within 1.35%) Furthermore, the mechanism was coupled with 3D CFD software to validate the cylinder pressure and heat release rates, using a six-cylinder, heavy-duty diesel engine with a bore of 114 mm, a stroke of 145 mm, a displacement of 8.9 L, and a compression ratio of 16.6 as the study subject. Based on the validation of the model and the feasibility of the mechanism, further studies were conducted on combustion and emission patterns under different load conditions and ammonia substitution rates. The results indicate that at low ammonia substitution rates, as the load decreases, the combustion rate slows down and thermal efficiency declines, while the indicated thermal efficiency first decreases and then increases; load primarily influences ignition and combustion processes by altering the thermodynamic state within the cylinder. At ammonia substitution rates of 20–60%, the heat release rate exhibits a “bimodal” pattern under different load conditions. NO, NO₂, and N₂O emissions first increase and then decrease with increasing ammonia substitution rate, peaking in the 40–60% range; CO₂ emissions gradually decrease as the ammonia substitution rate increases.

1. Introduction

Under the backdrop of carbon neutrality, internal combustion engines are facing severe challenges. To meet the requirements of carbon neutrality, internal combustion engines must be able to burn carbon-neutral fuels in the future and break through the technology for burning carbon-neutral fuels. As a zero-carbon fuel, ammonia (NH₃) holds significant potential for the low-carbon transition of heavy-duty power applications [1]. Ammonia–diesel-blended combustion technology addresses the poor ignitability of ammonia by using diesel to initiate combustion. However, this technology still faces two major challenges: first, the uneven mixture in large-bore engines leads to worsened NO_x emissions; second, existing diesel combustion mechanisms are mostly simplified single-component models, making it difficult to accurately characterize actual fuel properties, which results in discrepancies between the simulation results and engineering practice. The combustion properties of ammonia differ substantially from those of diesel, presenting a fundamental challenge for ammonia/diesel dual-fuel combustion. The key physicochemical and combustion characteristics of ammonia and diesel are compared in

Table 1. Fuel properties [2].

Property	Ammonia (NH3)	Diesel
Lower heating value (MJ/kg)	18.8	43.4
Auto-ignition temperature (K)	930	527–558
Laminar flame speed (cm/s)	10	33
Flammability limit (vol%)	16–25	0.6–5.5
Research octane number (RON)	110	N/A
Latent Heat of Vaporization, kJ/kg	1370	270

. Therefore, to achieve low-carbon emission reduction targets, it is necessary to develop a highly accurate model of diesel combustion mechanisms.

Dai et al. [3] revealed the influence of ammonia blending with methane on ignition delay through self-ignition experiments of NH₃/CH₄ mixtures under high-pressure conditions. Regarding laminar flame propagation, Okafor et al. [4] and Mei et al. [5], respectively, explored the laminar flame propagation speed of ammonia from different perspectives, providing fundamental data for the combustion organization of ammonia fuel. In terms of nitrogen and oxygen generation, Mendiara et al. [6] and Bykov et al. [7] analyzed the transformation pathways of nitrogen elements during ammonia oxidation from a kinetic perspective. Regarding flame stretch extinction, Colson et al. [8] explained phenomena such as the extinction characteristics of ammonia/air-premixed flames at different pressures from a chemical kinetics perspective. Reiter et al. [9] investigated the flame stretch extinction characteristics of ammonia and diesel during combustion under engine-operating conditions. In terms of flame structure, Brackmann et al. [10] systematically analyzed the structural characteristics of premixed ammonia flames at atmospheric pressure by combining laser diagnostics and kinetic simulations. The focus is on revealing the path-dominating role of specific intermediate products during ammonia oxidation. Chai et al. [11] found that ammonia transforms into HNO under lean conditions and into N₂NH under rich conditions, and these intermediate products can effectively regulate the formation of fuel-type NO_x. Song et al. [12] conducted fundamental experiments on ammonia combustion at 30 bar and 100 bar, analyzed the evolution laws of key intermediate products, and established a complete combustion mechanism of ammonia under high-pressure conditions. The results indicated that the NH₂ to H₂NO transformation pathway, which is non-dominant at low pressure, becomes the core reaction channel for ammonia oxidation at high pressure. Hurault et al. [13] measured the ignition delay time of pure ammonia at high pressure (40–70 bar) and medium–low temperature (950–1150 K) using a rapid compression machine and found that the ignition delay time decreases with increasing equivalence ratio, temperature, and pressure.

In existing studies, to simplify calculations, most ammonia–diesel combustion models use n-heptane as a single-component substitute for diesel fuel, which is directly coupled with the ammonia mechanism. However, actual diesel is a mixture composed of n-saturated alkanes, iso-saturated alkanes, aromatics, and naphthenes; therefore, n-heptane-type alkanes alone cannot be used to represent diesel components.

Table 2. Composition and construction methods of existing mechanisms.

Mechanism	Species	Reactions	Construction Method	Diesel Components
Xu [14]	69	389	Decoupling + GA optimization	n-heptane
Wang B [15]	74	495	DRGEP	n-heptane
Wang [16]	84	422	Sensitivity analysis + rate constant optimization	n-heptane
Sun [17]	227	937	Multi-objective GA	5-component
Shang [18]	138	963	Merging of existing mechanisms	n-heptane
Cai [19]	162	755	Sensitivity analysis	n-heptane
Jiang [20]	90	400	Modularization methods	4 component
Yu [21]	1376	6499	Merging of existing mechanisms	n-heptane
Dong [22]	2854	11,790	New detailed	n-heptane

summarizes recent literature based on two dimensions: construction methods and mechanistic scale. In terms of construction methods, the development of ammonia/diesel dual-fuel mechanisms currently follows three main technical approaches: (1) Construction from the skeleton level using decoupling and modularization methods; for example, Xu et al. [14] constructed a 69-component, 389-reaction mechanism for an ammonia/n-heptane dual-fuel system based on the decoupling method. (2) Starting from detailed mechanisms and simplifying them using methods such as DRG, DRGEP, and sensitivity analysis, as demonstrated in the studies by Dong et al. [15] and Wang [16]. The advantage of this approach lies in its ability to retain the chemical information of detailed mechanisms; however, the removal of components and reactions during simplification may affect the prediction accuracy under specific operating conditions; (3) Directly merging existing ammonia oxidation mechanisms with diesel substitute mechanisms, such as the ammonia/n-heptane mechanism developed by Shang et al. [17] by combining the ammonia/C1 radical mechanism, the n-heptane oxidation radical mechanism, and the NH₂/CN radical mechanism.

The ammonia energy replacement rate is a key parameter that affects the combustion mode and emission characteristics of ammonia/diesel dual-fuel engines. Nadimi et al. [23] found through simulation that as the ammonia replacement rate increases, the combustion mode changes from diesel diffusion combustion to dual-fuel premixed combustion; at a 61.6% replacement rate, the cylinder pressure rise rate reaches 9.5 bar/deg, an increase of 76.5% compared to pure diesel. In terms of emissions, CO₂ significantly decreases, but NO, NO₂, N₂O, and unburned ammonia emissions all increase; among them, the greenhouse effect of N₂O is 298 times that of CO₂, but the total greenhouse gas emissions still decrease with the increase in the replacement rate, from 727 g/kWh to 243 g/kWh at the highest replacement rate. Almanzalawy et al. [24] pointed out that a 54% replacement rate can achieve stable operation, but a higher rate will cause misfire. Hu Shengqi [25] and Yang Jianqiang [26], respectively, verified from the perspectives of NO_x reduction mechanism, emission reduction rate, and load impact, indicating that at high ammonia replacement rates, combustion efficiency decreases and flame development is delayed. Wang et al. [27] found that in ammonia/diesel dual-fuel engines, increasing the ammonia energy replacement rate (AER) reduces the in-cylinder pressure, temperature, and heat release rate, while significantly reducing CO₂ and NO_x emissions. However, N₂O and unburned ammonia emissions increase, partially offsetting the greenhouse gas reduction effect. In the future, combustion and emission performance can be improved by optimizing the injection strategy.

Although previous studies have made significant progress in understanding the fundamental combustion characteristics of ammonia/diesel mixtures and the influence of injection strategies, systematic research on the combustion and emission behavior of large-bore diesel engines across a wide load range and under different ammonia substitution rates remains insufficient. To address this, this study employed decoupling and modularization methods [28,29] to develop a simplified dynamical model of the ammonia/diesel dual-fuel simplified kinetic model, in which the diesel engine is modeled as a four-component system. The model was validated using experimental data on the shock tube ignition delay and laminar flame speed. It was subsequently coupled with three-dimensional CFD software to predict combustion and emission characteristics under different load conditions and ammonia substitution rates, thereby providing a theoretical foundation for the optimization of combustion systems and low-carbon operation in ammonia/diesel dual-fuel engines.

2. Materials and Methods

2.1. Development of Combustion Mechanisms

To accurately characterize the physicochemical properties of actual diesel fuel, a diesel surrogate was prepared using four components: n-heptane, n-hexadecane, isohexadecane, and α-methylnaphthalene. Among these, n-heptane represents the light alkane fraction; n-hexadecane and isohexadecane work together to adjust the cetane number to match the ignition characteristics of actual diesel fuel; and α-methylnaphthalene is used to characterize the aromatic hydrocarbon content and its effect on soot formation. The composition of the diesel substitute was determined using the Surrogate Blend Optimizer submodule of the CHEMKIN (version 2021) simulation software [28]. After multiple rounds of optimization, the parameters of the substitute model fell within the range of those for actual diesel; the comparison of ranges is shown in

Table 3. Comparison of target fuel properties: diesel surrogate vs. real diesel.

Property	Real Diesel	Diesel Surrogate
Cetane number	40–55	52.09
Aromatic hydrocarbon mass fraction	20~30%	23.64%
Lower heating value (MJ/kg)	42.5~44.40	43.47

. Figure 1 illustrates the molar percentage of each component.

Based on the principles of the decoupling method [29], this mechanism uses a simplified ammonia–diesel mechanism [30] as its basic framework. This mechanism includes reactions related to nitrogen-based components, carbon-based components of C7 and below, and the formation and consumption of NO_x. Building upon this foundation, a modular construction approach is adopted to expand the diesel component.

The ammonia/diesel mechanism was developed using Chemkin (version 2021) software; the schematic diagram of the simplified ammonia/diesel mechanism is shown in Figure 2. First, the isodecane mechanism from Chang et al. [29] and the α-methylnaphthalene mechanism from Qiu et al. [31] were extracted and integrated with the simplified ammonia–diesel mechanism to form a new mechanism. When merging the isodecane mechanism with the simplified ammonia–diesel mechanism, the isodecane mechanism was used as the host mechanism. This ensures that none of the species in the mechanism conflict with the host mechanism; instead, they are treated as a completely new, independent chemical system and added in their entirety to the merged mechanism. The decomposition pathways of isodecane are thus preserved, leading to more reliable computational results. This merged mechanism is named merge1. Next, merge1 is combined with the extracted α-methylnaphthalene mechanism; similarly, using merge1 as the host mechanism yields merge2. Finally, n-hexadecane is extracted from the wide-range fuel mechanism [32] for coupling, resulting in merge3. Merge3 is then coupled with the cross-reactions between the nitrogen-containing components and the model diesel fuel components reported by Wang et al. [33], yielding merge4. At this point, the ammonia–diesel composite mechanism is complete. Through the above process, a simplified ammonia/diesel composite combustion mechanism comprising 212 components and 620 elementary reactions has been obtained.

2.2. Correction of Combustion Mechanisms

2.2.1. Calculation of Ignition Delay

The combustion mechanism developed above was used to conduct a preliminary validation of the ignition delay times for single-component fuels consisting of ammonia, isohexadecane, α-methylnaphthalene, and n-heptane. This validation was based on shock tube experimental data published in the literature [32,34,35,36,37]; the simulation conditions were uniformly set to an equivalence ratio of 0.5 and a pressure of 20 atm, with the temperature ranges for each component selected according to the original data in the literature. Since n-heptane and n-hexadecane both belong to the n-alkane class and share similar chemical kinetic characteristics, n-heptane was selected as a representative to validate this class of components. The comparison between simulated and experimental values is shown in Figure 3. The calculated ignition delay time for ammonia is slightly higher than the experimental values overall; however, the ignition delay time for isohexadecane matches the experimental results, so no adjustment is necessary. For n-heptane, the simulation results match the experimental data well in the medium-temperature range, but the simulated values are slightly lower than the experimental values in the high-temperature range and slightly higher in the low-temperature range, requiring adjustment. The ignition delay time for α-methylnaphthalene is generally slightly higher than the experimental results and requires adjustment; therefore, appropriate corrections and adjustments need to be made to the combustion mechanisms of the ammonia, n-heptane, and α-methylnaphthalene fuels.

2.2.2. Calculation of Temperature Sensitivity

A temperature-sensitivity analysis of the ignition delay period was conducted for ammonia, isodecane, α-methylnaphthalene, and n-heptane under a combined ammonia/diesel ignition mechanism.

Table 4. Equations for some important elementary reactions.

Reaction No.	Elementary Reaction	Reaction No.	Elementary Reaction
R15	NH2 + HO2 <=> H2NO + OH	R498	hmn + HO2 <=> hmnr + H2O2
R18	NH2 + O2 <=> H2NO + O	R510	ic4h8 <=> ic4h7 + H
R24	NH2 + NO <=> N2 + H2O	R512	ic4h8 + OH <=> ic4h7 + H2O
R25	NH2 + NO <=> N2 + H2O	R513	ic4h7 + HO2 <=> ic4h7o + OH
R26	NH2 + NO <=> NNH + OH	R515	ic3h5cho + OH <=> ic3h5co + H2O
R27	NH2 + HONO <=> NH3 + NO2	R533	C3H5-A + HO2 <=> C3H5O + OH
R77	H2NO + O2 <=> HNO + HO2	R549	C2H4 + OH <=> C2H3 + H2O
R78	H2NO + NH2 <=> HNO + NH3	R556	C2H3 + O2 <=> CH2CHO + O
R142	HNOH + NH2 <=> H2NN + H2O	R568	CH3 + HO2 <=> CH3O + OH
R143	2NH2 (+M) <=> N2H4 (+M)	R571	CH2O + OH <=> HCO + H2O
R211	2CH3 (+M) <=> C2H6 (+M)	R573	CH2O + O <=> HCO + OH
R461	C7H15O2-3 <=> C7H14-3 + HO2	R581	CH3 + HO2 <=> CH4 + O2
R464	C7H15O2-2 <=> C7H14OOH2-4	R600	H + O2 <=> O + OH
R465	C7H15O2-3 <=> C7H14OOH3-5	R606	H + O2 (+M) <=> HO2 (+M)
R479	C7H14OOH2-4O2 <=> NC7KET24 + OH	R607	HO2 + O <=> OH + O2
R480	C7H14OOH3-5O2 <=> NC7KET35 + OH	R610	HO2 + OH <=> H2O + O2
R481	C7H14OOH4-2O2 <=> NC7KET42 + OH	R612	2OH (+M) <=> H2O2 (+M)
R496	hmn + H <=> hmnr + H2	R620	H2O2 + OH <=> H2O + HO2

lists some of the key elementary reaction equations. Figure 4 shows the top 10 elementary reactions with the highest sensitivity coefficients during the combustion process for ammonia, isohexadecane, α-methylnaphthalene, and diesel fuel at an equivalence ratio of 0.5, a pressure of 20 atm, and under various temperature conditions. Equation (1) represents the temperature sensitivity coefficient.

$$S_i={\displaystyle \frac{\mathit{\partial }\ln T}{\mathit{\partial }\ln k_i}}={\displaystyle \frac{T\left(2k_i\right)-T\left(0.5k_i\right)}{1.5 T\left(k_i\right)}}$$

(1)

In Equation (1), $S_i$ is the ignition delay time sensitivity coefficient for the i-th reaction; $k_i$ is the reaction rate constant for the i-th reaction.

Sensitivity analysis identified several key reactions that significantly influence the ignition delay time; the signs and magnitudes of the sensitivity coefficients directly reflect whether these reactions promote or inhibit the ignition process. A temperature sensitivity analysis of Figure 4a for ammonia at T = 1220 K reveals that the temperature sensitivity of ammonia combustion is dominated by the competing reactions involving the conversion of NH₂ + NO and H₂NO. The key reactions promoting ignition are R24 (NH₂ + NO → N₂ + H₂O) and R78 (H₂NO + NH₂ → HNO + NH₃). R24 indirectly promotes chain-branching reactions by converting NO into stable N₂, thereby preventing the reaction of NO with active radicals; R78 generates the radical precursor HNO, which serves as a source for the subsequent formation of reactive radicals such as OH and H, thereby directly driving the ignition process; the primary reactions that inhibit ignition are R77 (H₂NO + O₂ → HNO + HO₂) and R26 (NH₂ + NO → NNH + OH), which produce the less reactive HO₂ and the stable intermediate NNH, respectively.

A comparative analysis of the sensitivity of n-heptane in the high-temperature (1250 K) and low-temperature (909 K) regions, as shown in Figure 4b, reveals that at high temperatures, the sensitivity coefficients for all reactions are less than 0.08, indicating overall weak sensitivity. At low temperatures, R612 (2OH ⇌ H₂O₂) exhibits the strongest positive sensitivity; this reaction significantly delays the ignition process by converting highly reactive OH radicals into relatively stable H₂O₂. Reactions that promote ignition include R461(C₇H₁₅O₂-3 ⇌ C₇H₁₄-3 + HO₂), R610 (HO₂ + OH ⇌ H₂O + O₂), and R620 (H₂O₂ + OH ⇌ H₂O + HO₂).

The sensitivity analysis of α-methylnaphthalene at three characteristic temperatures—high (1333 K), medium (1176 K), and low (1053 K)—as shown in Figure 4c, indicates that ignition in the high-temperature region (1333 K) is dominated by classical small-molecule radical reactions (e.g., R600: H + O₂ ⇌ O + OH), whereas the sensitivity in the medium- and low-temperature regions (1176 K, 1053 K) shifts significantly toward reactions of intermediates (e.g., C₄H₈, C₄H₇) with radicals such as OH and HO₂ (e.g., R512: C₄H₈ + OH ⇌ C₄H₇ + H₂O), revealing the critical role of these low-temperature chemical pathways.

Chemical kinetics describes the rates of elementary reaction steps at the molecular level and the patterns of change in the concentrations of reactants and products over time. Kinetic models of reaction mechanisms include several parameters: reaction rate constants, elementary reaction equations, and corresponding kinetic parameters. Therefore, after extensive experimentation and validation, Arrhenius proposed the following equation:

$$k_{if}={\mathrm{A}}_i{\mathrm{T}}^{{\beta }_i}exp\left(-{\displaystyle \frac{E_i}{RT}}\right)$$

(2)

In Equation (2), $k_{if}$ is the forward reaction rate constant of the $i$-th reaction; $A_i$, ${\beta }_i$, and $E_i$ are the pre-exponential factor, temperature exponent, and activation energy, respectively; $T$ is the temperature; and $R$ is the molar gas constant.

To improve the accuracy of predicting ignition delay times for ammonia, isohexadecane, α-methylnaphthalene, and diesel fuel in the ammonia/diesel mechanism, the first 13 elementary reactions that significantly influence the ignition process were selected based on the modified Arrhenius equation, and the pre-exponential factor A for these 13 key elementary reactions was adjusted. The final value of A was determined by comparing the ignition delay times before and after adjusting A. When A is sufficiently large, adjusting the A value may cause oscillations or instability in the numerical solver. In such cases, it is necessary to adjust the activation energy. Lowering the activation energy selectively accelerates the forward reaction while having a relatively minor effect on the reverse reaction, which helps shift the equilibrium and thereby effectively promotes ignition.

Table 5. Comparison before and after adjustment of the key elementary reaction.

Reaction No.	Before	After	Reaction No.	Before	After
R481	1.25 × 1010	6.25 × 1011	R143	5.60 × 1014	2.80 × 1015
R480	1.25 × 1010	6.25 × 1011	R26	4.29 × 1010	4.30 × 1011
R479	1.25 × 1010	6.25 × 1011	R77	2.30 × 102	1.18 × 103
R465	2.50 × 1010	3.00 × 1011	R512	3.4 × 104	5.44 × 106
R464	2.50 × 1010	3.00 × 1011	R513	7.00 × 1012	7.00 × 1012
R600	1.97 × 1014	3.07 × 1014	R510	Ea = 1.143 × 105	Ea = 1.093 × 105
R606	1.48 × 1012	1.18 × 1012

presents a comparison of the pre-exponential factors and activation energies for the key elementary reactions before and after adjustment.

Following the aforementioned modifications and adjustments, a composite ammonia/diesel mechanism comprising 212 components and 620 elementary reactions was obtained. This mechanism is capable of describing the ignition process of ammonia–diesel composite fuel. The ignition delay times for ammonia, isohexadecane, α-methylnaphthalene, and diesel were recalculated using the modified mechanism.

Figure 5 shows a comparison of the simulated and experimental ignition delay times for ammonia, isohexane, α-methylnaphthalene, and diesel fuel under the ammonia/diesel composite mechanism at equivalent ratios of 0.5 and 1.0 under various temperature and pressure conditions. Clearly, the adjusted mechanism can more accurately predict the ignition delay times of these fuels.

2.3. Verification of Combustion Mechanisms

2.3.1. Ignition Delay Verification

Figure 5a–d present the comparison results of ignition delay times for ammonia, isohexadecane, n-heptane, and α-methylnaphthalene, respectively. It can be seen that under different pressure, equivalence ratio, and temperature conditions, the ignition delay times of each single-component fuel simulated by the reduced mechanism are very close to the experimental values [33,34,35,36,37], and the trends with respect to temperature remain consistent. Overall, the ammonia/diesel-reduced mechanism can predict the ignition delay characteristics of these single-component fuels with reasonable accuracy.

2.3.2. Verification of Laminar Combustion Velocity

The laminar flame propagation velocity is a key parameter for measuring the spatial movement rate of the flame front during combustion, and it is one of the most important intrinsic properties of a combustible mixture [38]. Therefore, it is crucial that the mechanism can reasonably predict the laminar flame propagation velocity.

Figure 6a shows a comparison of simulated and experimental laminar flame speeds for ammonia. Under conditions of an equivalence ratio of 0.7–1.6 and 1 atm, the simulated and experimental values show a high degree of agreement. The predicted peak flame velocity occurs under rich conditions (φ ≈ 1.05–1.10), consistent with the trends revealed by most experimental data. Near the stoichiometric range (φ = 0.9–1.2), the simulated values show the best agreement with the experimental results of Pfahl [39] and Hayakawa [40], with differences within ±2 cm/s. In the lean and rich combustion regions, although the simulated values deviate from some data, they generally fall within the dispersion range of the experimental data and show a high degree of agreement with the independent simulation results of Takizawa [41]. Therefore, this mechanism demonstrates good predictive consistency and reliability across the entire flammability range and is suitable for subsequent engineering applications and combustion characteristic studies.

Figure 6b presents a comparison of experimental and simulated laminar combustion velocities for n-heptane at 298 K and 398 K, with equivalence ratios ranging from 0.6 to 1.6 [36]. The simulation results are generally consistent with the experimental data and show good agreement, effectively validating the accuracy of the laminar burning velocity predictions for n-heptane. While the simulated values are slightly higher than the experimental ones in some equivalence ratio ranges (especially at 398 K and equivalence ratios of about 1.0–1.1), with a maximum relative error within +2 cm/s. This deviation may arise from simplifications in the chemical kinetic mechanism, heat transfer boundary conditions, or experimental measurement uncertainties. However, it does not affect the qualitative and semi-quantitative analysis of the combustion behavior.

The discrepancy could stem from factors such as simplifications in chemical kinetics, simplifications in heat transfer boundary conditions, or uncertainties in experimental measurements.

2.3.3. Verification of Ignition Delay for Different Blending Ratios of Ammonia–Diesel Mixtures

Building on the validation of ignition delay and laminar flow velocity for single-component fuels, we further validated the mechanism’s ability to predict ignition delay times for ammonia/diesel blends with different mixing ratios, in order to assess its applicability in dual-fuel combustion simulations.

Figure 7 illustrates the correlation between ignition delay times for ammonia/diesel blended fuels (with blending ratios of 10/90, 20/80, 30/70, 40/60, and 60/40). Figure 7 shows the relationship between ignition delay time and temperature for different ammonia–diesel blending ratios (N/D). Overall, as the temperature increases, the ignition delay time for all operating conditions shortens significantly, consistent with the Arrhenius law. A comparison of different blending ratios reveals that as the ammonia ratio increases from 10% to 60%, the ignition delay time lengthens markedly, indicating that ammonia inhibits the ignition process of the blended fuel. The simulated results agree well with experimental data [45,46] over most of the temperature range, indicating that the reaction mechanism used can accurately predict the ignition characteristics of the blended fuel; however, some discrepancies exist at high ammonia ratios (e.g., N/D = 60/40) and in the low-temperature region, suggesting that the relevant reaction pathways still require further optimization.

Figure 8 shows the variation in ignition delay (ms) as a function of the reciprocal of temperature for different ammonia/diesel blend ratios (90/10, 80/20, 70/30), along with a comparison to experimental data. The results indicate that as temperature decreases, the ignition delay for all mixture ratios increases exponentially, reflecting the inhibitory effect of low temperatures on fuel radical reactions. Additionally, under the same temperature conditions, the higher the ammonia mixture ratio, the longer the ignition delay, which is consistent with ammonia’s high ignition temperature and low chemical reactivity. Furthermore, the simulated curves show a high degree of agreement with the experimental data [47], validating the reliability of numerical simulation methods in studying the ignition characteristics of this type of blended fuel.

2.4. CFD Model Validation and Prediction

After constructing and validating the chemical reaction mechanism, the validated mechanism was coupled with the CFD software Converge (version 3.0). Using a specific type of automotive diesel engine [48] as the subject of study, the predictive capability of the mechanism under actual engine operating conditions was verified using experimental data, and the combustion process and emission characteristics of the ammonia/diesel dual-fuel system under different load conditions were further investigated. The main structural parameters and initial calculation parameters of this diesel engine are shown in

Table 6. Basic engine specifications and boundary conditions.

Parameter	Value	Parameter	Value
Bore (mm)	114	Nozzle protrusion into cylinder (mm)	3.5
Stroke (mm)	145	Fuel supply advance angle (°CA)	14
Connecting rod length (mm)	216	Injection duration (°CA)	31
Fuel injection quantity per cycle (g)	0.1365	Number of nozzle holes/hole diameter (mm)	7/0.25
Initial pressure (bar)	2.86	Initial temperature(K)	396
Compression ratio	16.6	Kinetic energy (m2/s2)	42.4
Injected fuel temperature (K)	350

. Figure 9 displays a cross-sectional view of the combustion chamber. The combustion chamber was modeled using SolidWorks (version 2021), exported in STL format, and configured in Converge. Figure 10 shows the single-cylinder geometric model. The main physical models are listed in

Table 7. Validation of key physical models in 3D.

Model Type	Name
Turbulence model	RNG-k-ε
Droplet collision model	NTC collision model
Spray breakup model	KH-RT
Evaporation model	Frossling
Combustion model	SAGE
Heat transfer	O’Rourke and Amsden

. This study primarily simulates the in-cylinder combustion process and emission characteristics of the dual-fuel system. To reduce computational costs, the effects of intake and exhaust gas flow were neglected; therefore, it was not necessary to model the valves or intake and exhaust ports, thereby simplifying the model. The specific simplifications are as follows:

(1)

In the 3D model, the cylinder head and valve structures were simplified into a single plane;

(2)

The 0 °CA was defined as the top dead center (TDC) of ignition. The computational time interval spans from the closing of the intake valve to the opening of the exhaust valve, studying the process from compression to expansion in the diesel engine. The corresponding crankshaft angle range is −150 °CA to 132 °CA.

Figure 11 shows a comparison of the mesh-independent cylinder pressure simulation results with experimental data. Regarding the meshing strategy, Level 2 adaptive mesh refinement (AMR) was applied to critical areas, such as the nozzle region, boundary layer, and cylinder head. To verify mesh independence, simulations under pure diesel operating conditions were performed using three base mesh sizes: 5 mm, 4 mm, and 3 mm. The results show that the cylinder pressure curves for the 4 mm and 3 mm grids are essentially identical, and both agree well with the experimental values, indicating that the 4 mm grid has achieved result convergence. Balancing computational accuracy and cost, a 4 mm base grid was used for all subsequent simulations.

Figure 12 shows a comparison of the simulated and experimental results for cylinder pressure and heat release rate under initial operating conditions. The simulated values follow the same trend as the experimental values, with an error of 1.35% in the peak pressure. The overall trend of the heat release rate curve matches the experimental data, indicating that the simulation model accurately reflects the actual combustion process and provides a reliable basis for subsequent analysis.

Table 8. Simulation case setup.

Case	Engine Load	Engine Speed (rpm)	Ammonia Substitution Rate
Case1-12345	25%	2200	0, 20%, 40%, 60%, 80%
Case2-1234	50%	2200	20%, 40%, 60%, 80%
Case3-1234	100%	2200	20%, 40%, 60%, 80%

summarizes the main simulation cases. The effects of the ammonia substitution rate (ranging from 20% to 80%) on the combustion and emission characteristics of an ammonia/diesel dual-fuel engine are investigated under different engine loads. The engine load was varied by adjusting the per-cycle diesel injection quantity at a fixed speed of 2200 rpm, covering 25%, 50%, and 100% load conditions.

The ammonia energy substitution ratio (ESR) is defined as the fraction of the total fuel energy supplied by ammonia, calculated based on the lower heating values of ammonia and diesel, as shown in Equation (3).

$$\mathrm{AER}={\displaystyle \frac{m_{{\mathrm{NH}}_3}·{\mathit{LHV}}_{{\mathrm{NH}}_3}}{m_{{\mathrm{NH}}_3}·{\mathit{LHV}}_{{\mathrm{NH}}_3}+m_{\mathrm{diesel}}·{\mathit{LHV}}_{\mathrm{diesel}}}}\times 100$$

(3)

In Equation (3), $m_{{\mathrm{NH}}_3}$ and $m_{\mathrm{diesel}}$ are the mass flow rates (kg/h) of ammonia and diesel, ${\mathit{LHV}}_{{\mathrm{NH}}_3}$ and ${\mathit{LHV}}_{\mathrm{diesel}}$ are their corresponding lower heating values (MJ/kg).

3. Results

3.1. Analysis of Combustion Characteristics at Different Loads for Low-Ammonia Substitution Rates

Figure 13a compares the in-cylinder pressure and heat release rate of pure diesel (N0) at full load and 20% ammonia substitution (N20) under different loads. The in-cylinder pressure curves show that ammonia addition exerts a pronounced inhibiting effect on combustion. At full load, the peak pressure drops from approximately 15 MPa for pure diesel to about 14 MPa with 20% ammonia, accompanied by a more gradual pressure rise. This inhibiting effect intensifies at medium and low loads, where the N20 pressure traces exhibit a distinctly flattened profile with substantially reduced peaks. Regarding the crank angle of peak pressure, the full-load N0 peak occurs at approximately 8 °CA ATDC, which advances to about 7 °CA ATDC with ammonia addition. As the load decreases, θPmax for N20 further advances to approximately 0.9 °CA ATDC at medium load and 0.4 °CA ATDC at low load. This advancing trend is consistent with the diesel pilot-ignition mechanism: the reduction in diesel injection quantity at higher ammonia substitution rates shortens the injection duration, leading to faster fuel vaporization and earlier diesel auto-ignition, which in turn advances the overall combustion phasing. At low loads, the diminished late-stage heat release from ammonia’s slow combustion causes the pressure to peak near TDC and decline rapidly thereafter. The heat release rate curves further illustrate the changes in energy release. At full load, pure diesel exhibits a characteristic bimodal structure—a premixed combustion peak near TDC followed by a diffusion combustion peak at approximately 10 °CA ATDC—with concentrated and high-peak heat release. With 20% ammonia, the bimodal structure is retained, though both peaks are reduced in magnitude; the premixed peak position remains essentially unchanged, while the second peak advances slightly to about 9 °CA ATDC, and the overall heat release duration is extended. At medium load, the second peak of N20 shifts later to approximately 15 °CA ATDC, with a further reduced peak and a broadened profile. At low load, the bimodal structure of N20 disappears entirely, evolving into a broad single peak with distinct staged combustion features, reflecting unstable ammonia combustion and insufficient mixture preparation under low-temperature conditions.

Figure 13b shows the average in-cylinder temperature curves under different loads and ammonia substitution rates. Under full-load conditions, the peak temperature for pure diesel (N0) is approximately 1680 K; after blending with 20% ammonia (N20), it drops to approximately 1440 K, a decrease of about 14.3%. At the same ammonia substitution rate, when the load is reduced from 100% to 25%, the peak temperature further decreases to approximately 1050 K, representing a reduction of about 27.1%. During the compression stroke (−150–0 °CA), the temperature curves for all operating conditions almost coincide, indicating good consistency in the simulation of the compression process. Differences primarily appear during the combustion phase (0–40 °CA), where both the load and ammonia substitution rate have significant effects on the peak temperature.

Figure 14a shows the variations in CA10 and CA50 at different loads for a low ammonia substitution ratio (20%). Here, CA10 and CA50 refer to the crankshaft angles at which the cumulative heat release reaches 10% and 50% of the total heat release, respectively. Under full-load conditions with pure diesel, CA10 occurs earlier, indicating a short ignition delay and rapid combustion onset. Meanwhile, CA50 is relatively centered, reflecting a concentrated combustion. After adding 20% ammonia, the CA10 of the full-load N20 case is noticeably delayed, indicating an extended ignition delay. Quantitatively, at full load, CA10 is delayed by approximately 1.1 °CA. This is attributed to ammonia’s high auto-ignition temperature and its endothermic decomposition characteristics. As the load decreases, the CA10 of N20 under medium and low loads is further delayed, with the most significant delay occurring at low load. This suggests that ammonia’s ignition performance deteriorates at lower cylinder temperatures. Regarding CA50, the CA50 of full-load N20 shifts later than that of neat diesel, but the shift is smaller than that of CA10; at full load, CA50 is delayed by approximately 4.5 °CA, indicating a relatively stable combustion process after ignition. Although the CA50 of N20 at low loads also shifts later, the interval between CA50 and CA10 (i.e., the combustion duration) does not increase significantly; in fact, it is slightly shorter under certain operating conditions. This may imply that under ammonia-blended conditions, although ignition occurs later, the energy release is more concentrated during the middle and late stages of combustion, suggesting that the ammonia–diesel mixture exhibits certain combustion acceleration characteristics after ignition.

Figure 14b shows the indicated thermal efficiency (ITE) and combustion efficiency at different loads under a 20% ammonia substitution rate (N₂O). The addition of ammonia has a dual effect on engine performance. On the one hand, ammonia inhibits combustion, causing combustion efficiency to decrease monotonically from 92.0% at full load to 75.0% at low load; the high auto-ignition temperature and low flame speed of ammonia are further amplified at low loads. On the other hand, the addition of ammonia causes the ITE to exhibit a non-monotonic variation: at full load, the ITE rises from 32.6% for pure diesel to 36.4%; at medium load, it drops sharply to 19.0%; and at low load, it rebounds to 24.2%. Ammonia’s low-temperature flame-retardant properties cause the ITE to rebound at low loads, reducing peak cylinder temperatures and heat transfer losses. The shift in the heat release phase optimizes thermal-to-mechanical conversion efficiency, and these positive effects outweigh the negative impact of reduced combustion efficiency.

Based on Figure 15, this paper defines the ignition timing as the crankshaft angle CA2 corresponding to the point where the cumulative heat release reaches 2% of the total heat release. Figure 16 shows the in-cylinder temperature distribution at the CA2, CA10, and CA20 timings under different load conditions; this cross-section of the combustion chamber is a horizontal top view. As shown in Figure 16, when comparing different load conditions, the flame cores are all ignited at the CA2 time point; however, the high-temperature zone under low load is more concentrated and has a smaller extent. At the CA20 timing, the high-temperature flame front under full-load conditions has propagated near the combustion chamber wall, almost filling the entire cross-section, while the high-temperature zone under low-load conditions remains primarily confined to the central region of the combustion chamber, exhibiting a distinct core-concentrated combustion pattern with limited expansion and a relatively slow progression. The higher local temperatures observed at full load indicate that, under full-load conditions, fuel-air mixing is more thorough, the formation of the combustible mixture occurs more rapidly, and ignition conditions are more favorable. This is because, at full load, the fuel injection volume is high, resulting in high mixing energy; the resulting combustible mixture covers a wide area with an appropriate concentration, leading to rapid flame propagation and achieving global combustion in a short time. In contrast, at low load, the fuel injection volume is small, the overall temperature and pressure within the cylinder are lower, and the formation rate of the mixture is slow, potentially resulting in local areas that are too lean or too rich. This leads to a slow flame propagation speed, causing the combustion reaction to remain stable only in the local core area with the most optimal conditions, making it difficult for the flame to spread toward the periphery.

The addition of ammonia suppresses in-cylinder pressure, heat release rate, and temperature, with the extent of this effect increasing as the load decreases. Ammonia’s lower calorific value reduces the total heat released during combustion, directly leading to a decrease in peak cylinder pressure and temperature. Ammonia’s higher autoignition temperature extends the ignition delay period, causing the peak pressure and peak heat release rate to occur later in the cycle. Ammonia’s lower laminar flame speed slows the combustion rate, resulting in a lower peak heat release rate and a prolonged heat release process [49,50]. The combined effect of these three factors results in a more gradual combustion process and a decrease in peak temperature. As the load decreases, the base temperature inside the cylinder drops further, amplifying ammonia’s low reactivity. At the same time, relative heat loss increases, making the aforementioned inhibitory effects even more pronounced at low loads [51,52,53].

Therefore, at low ammonia substitution rates, the load primarily influences ignition and combustion processes by altering the thermodynamic conditions within the cylinder; this differs fundamentally from the changes in combustion characteristics driven by fuel chemistry at high ammonia substitution rates, as discussed below.

3.2. Analysis of Combustion Characteristics Under Different Loads at Medium- and High-Ammonia Concentrations

To investigate the combustion characteristics of diesel–ammonia dual-fuel mixtures at medium-to-high ammonia substitution rates, this study compared the in-cylinder pressure and heat release rate curves, as well as the in-cylinder average temperature curves, of pure diesel (N0) with diesel blends containing 40%, 60%, and 80% ammonia substitution rates under fixed conditions of rotational speed, intake air temperature, intake air pressure, and injection timing. The results are shown in Figure 16.

As shown in Figure 16a, under full-load and medium-load conditions, the heat release rate curve for ammonia substitution rates of 20–80% exhibits a “bimodal” heat release pattern, whereas under light-load conditions, the heat release rate curve exhibits only a “unimodal” heat release pattern. This is because, under low-to-medium load conditions, the amount of pre-mixed ammonia fuel in the engine cylinder decreases, and the equivalence ratio gradually moves away from the combustion limits of ammonia. Consequently, the combustion rate of ammonia slows down, the heat release decreases, and the peak heat release from ammonia becomes less pronounced. In quantitative terms, compared with in-cylinder combustion of diesel alone, the second peak of the bimodal heat release pattern occurs 4–5 °CA later. Under different load conditions, the peak of the cylinder pressure curve decreases as the ammonia substitution rate increases. This is because, at low loads, as the ammonia substitution rate rises, the increase in the ammonia equivalence ratio is small. The effect of promoting fuel combustion is weaker than the impact of the slow ammonia flame propagation speed on combustion; consequently, ammonia combustion efficiency decreases, the heat released by ammonia decreases, and the total heat released by the fuel decreases. In addition, an increase in the ammonia substitution ratio causes the peak of the cylinder pressure curve to occur earlier. This is because, as the ammonia substitution ratio increases, the diesel injection volume decreases; consequently, the diesel mixes more uniformly with the ammonia–air mixture after injection, resulting in more efficient combustion and an earlier release of combustion heat. Specifically, under full-load and medium-load conditions, the peak phase of cylinder pressure advances by 1–2 °CA. However, under light-load conditions, this advance is not obvious.

As shown in Figure 16b, which depicts the average in-cylinder temperature curves at different ammonia substitution rates, the average in-cylinder temperature decreases as the ammonia substitution rate increases. As the ammonia substitution rate increases, the effect of ammonia’s equivalence ratio in promoting fuel combustion is outweighed by the impact of ammonia’s slow flame propagation speed on fuel combustion, resulting in a decrease in ammonia combustion efficiency. Furthermore, under medium-to-low-load conditions, the peak phase of the average in-cylinder temperature is progressively advanced as the ammonia substitution rate increases. This is because of the low combustion efficiency of ammonia; diesel plays a dominant role in combustion and heat release. Consequently, the phase of the peak in the average in-cylinder temperature is more strongly influenced by diesel combustion. At this point, as the ammonia substitution rate increases, a smaller amount of diesel is injected over a shorter injection duration. This results in faster mixing with air and a faster rate of combustion and heat release, causing the combustion reaction to occur earlier.

To further quantify the effect of load on combustion characteristics at medium-to-high ammonia substitution rates, this study uses a 60% ammonia substitution rate as an example to analyze the comparison of different load conditions at the same substitution rate. Compared to low-ammonia conditions (20%), the impact of load changes on the peak cylinder pressure and heat release rate is somewhat reduced at a 60% ammonia substitution rate. When the load is reduced from full load to low load, the peak cylinder pressure decreases by approximately 15%, whereas under the same load change, the decrease at a 20% ammonia substitution rate exceeds 25% (see Figure 13). Regarding the heat release rate, the decline in the peak heat release rate with decreasing load is more pronounced under low-ammonia conditions. This indicates that as the ammonia substitution rate increases, the chemical effects of the fuel gradually replace the thermodynamic effects, becoming the dominant factor influencing the combustion process. Therefore, under medium-to-high ammonia substitution rates, the influence of load on combustion characteristics is significantly reduced compared to low-ammonia conditions.

3.3. Analysis of Emissions at Different Load Levels for the Low-Ammonia Replacement Rate

Figure 17 clearly illustrates the shift in emission characteristics of an ammonia/diesel dual-fuel engine. Compared to conventional diesel, NO_x emissions increase at a 20% ammonia substitution rate under full load. In pure diesel combustion, NO primarily originates from the thermal pathway, which is highly dependent on high temperatures. When 20% ammonia is added, the combustion temperature decreases slightly, reducing thermal NO, but the ammonia provides an additional nitrogen source. Figure 17 also indicates that NH₃ slip occurs across all load conditions, yet unburned NH₃ increases conversely, with the most severe occurrence at low loads. N₂O emissions peak at medium loads, signaling a shift in combustion mode from carbon-based to nitrogen-based fuels, which introduces more complex challenges for multi-component emission control.

Figure 18 illustrates the distribution of NO emissions during combustion. Compared to pure diesel combustion, after the addition of ammonia, at the CA2 timing, the NO from N0 and N20 at full load has not yet activated fuel-derived NO due to the relatively low internal temperature; however, at medium and low loads, because of the lower initial temperature inside the cylinder, fuel evaporation and mixing are slower. The longer ignition delay period allows ample time for low-temperature chemical reactions between ammonia and diesel, resulting in NO concentration around the flame. As the crankshaft angle changes, NO initially concentrates around the flame core and subsequently diffuses throughout the combustion chamber. For pure diesel, this is primarily due to the thermochemical NO pathway, which relies on localized high temperatures. For ammonia-blended conditions, although the overall temperature is slightly reduced, the HNO derived from ammonia encounters abundant OH radicals in the flame zone and rapidly generates NO via the reaction HNO + OH → NO + H₂O, leading to the accumulation of NO throughout the entire cylinder.

3.4. Emission Analysis for Medium-to-High Ammonia Substitution Rates

Figure 19, Figure 20 and Figure 21 show the variation in NO_x emissions with ammonia substitution rates under full-load, medium-load, and low-load conditions, respectively. The trends observed in all three cases are broadly similar: emissions of nitrogen oxides (NO, NO₂, N₂O) initially increase and then decrease as the ammonia substitution rate rises, peaking in the medium substitution rate range (40–60%). As the load decreases from high to low, the in-cylinder combustion temperature and pressure decrease accordingly, leading to a sequential reduction in the overall emission levels of various pollutants. However, the trends of the change curves remain largely consistent, indicating that the moderate ammonia substitution rate range is the critical zone for NO_x formation, unaffected by changes in engine load. Therefore, in practical applications, attention should be focused on combustion optimization and control within this substitution rate range.

As shown in Figure 22, the instantaneous CO₂ generation rate indicates that, under different load conditions, the CO₂ generation rate gradually decreases as the peak ammonia substitution rate increases, demonstrating that ammonia, as a zero-carbon fuel, holds significant potential for carbon reduction.

Figure 23 presents the NO formation and consumption pathways. A reaction path analysis was performed using Chemkin-Pro at 1200 K and 1 atm for two typical ammonia substitution ratios, 20% and 80%, with a focus on the formation and consumption pathways of NO. At the 20% ammonia substitution ratio (rising NO_x stage), NO is mainly formed via HNO→NO and N₂O→NO. The ammonia oxidation chain (NH₃→NH₂→NH→HNO) also contributes moderately. In terms of consumption, NO is primarily oxidized to NO₂, while direct reduction to N₂ or to NNH is relatively weak. Because NO is mainly converted to another NO_x species (NO₂) rather than being completely removed from the system, the net NO_x emission remains high. At the 80% ammonia substitution ratio (falling NO_x stage), although some formation pathways become more prominent (especially HNO→NO and N₂O→NO), the consumption pattern changes fundamentally: the NO→NO₂ pathway almost disappears, whereas the reduction pathways NO→N₂ and NO→NNH are significantly enhanced. These reduction routes convert NO directly into N₂, so that the net removal of NO exceeds its formation, leading to a decrease in NO_x emissions.

The results of this pathway analysis are consistent with the emission trends obtained from CFD simulations, confirming that the mechanism is capable of reproducing the experimentally observed NO_x behavior.

4. Conclusions

This paper first establishes and validates a combustion mechanism for ammonia/diesel dual-fuel engines, and then applies the validated mechanism to investigate the combustion and emission characteristics of such engines. The main conclusions are as follows:

(1)

Using decoupling methods and a modular construction strategy, a simplified ammonia/diesel mechanism comprising 212 species and 620 elementary reactions was developed, with n-heptane, n-hexadecane, isohexadecane, and α-methylnaphthalene as diesel surrogate components; its scale is approximately one-tenth that of a detailed mechanism. The mechanism was validated using published shock tube ignition delay data (φ = 0.5 and 1.0, 20 atm) and laminar burning velocity data (ammonia: φ = 0.7–1.6, 1 atm; n-heptane: φ = 0.6–1.6, 298 K and 398 K). The results show good agreement, indicating that the model can be reliably used for subsequent CFD simulations.

(2)

At a low ammonia replacement rate (20%), as the load decreases, the peak cylinder pressure drops, the heat release rate decreases, and the heat release process is prolonged, resulting in a slower combustion rate. Compared to pure diesel, the crankshaft angles corresponding to CA10 and CA50 increase as the load decreases after ammonia blending; combustion efficiency decreases, and the indicated thermal efficiency follows a trend of first decreasing and then increasing. Regarding flame development, at low loads, the high-temperature zone is more concentrated and propagates more slowly; at full load, the flame front has essentially filled the combustion chamber. In terms of emissions, compared to pure diesel, all emissions increase. As the load decreases, nitrogen oxides gradually decrease, soot emissions first decrease and then increase, and unburned ammonia gradually increases. Load affects the ignition and combustion processes by altering the thermodynamic state within the cylinder.

(3)

At medium to high ammonia substitution rates, the influence of load on combustion characteristics weakens, and the ammonia substitution rate becomes the primary influencing factor. At medium to high loads, the heat release rate exhibits a “bimodal” pattern for ammonia substitution rates ranging from 20% to 60%; at low loads, all heat release rates follow a “unimodal” distribution. As the ammonia substitution rate increases, the peak cylinder pressure decreases, the average in-cylinder temperature drops, and the phase of the temperature peak advances at medium and low loads. NO, NO₂, and N₂O emissions follow a trend of first increasing and then decreasing with increasing ammonia substitution rate, peaking in the 40–60% range; CO₂ emissions gradually decrease as the ammonia substitution rate rises.

(4)

Reaction pathway analysis at 1200 K and 1 atm reveals the mechanism behind the non-monotonic NO_x trend with increasing ammonia substitution. At 20% ammonia, NO is mainly oxidized to NO₂; at 80% ammonia, reduction pathways (NO→N₂ and NO→NNH) become dominant, converting NO directly to harmless N₂. This shift from oxidation-dominated to reduction-dominated NO consumption explains the observed rise-then-fall NO_x emissions. The pathway results are consistent with CFD emission predictions, further confirming the mechanism’s reliability.