Unstable Flame Research Articles

Prognostics of Combustion Instabilities from Hi-speed Flame Video using A Deep Convolutional Selective Autoencoder

The thermo-acoustic instabilities arising in combustion processes cause significant deterioration and safety issues in various human-engineered systems such as land and air based gas turbine engines. The phenomenon is described as selfsustaining and having large amplitude pressure oscillations with varying spatial scales of periodic coherent vortex shedding. Early detection and close monitoring of combustion instability are the keys to extending the remaining useful life (RUL) of any gas turbine engine. However, such impending instability to a stable combustion is extremely difficult to detect only from pressure data due to its sudden (bifurcationtype) nature. Toolchains that are able to detect early instability occurrence have transformative impacts on the safety and performance of modern engines. This paper proposes an endto- end deep convolutional selective autoencoder approach to capture the rich information in hi-speed flame video for instability prognostics. In this context, an autoencoder is trained to selectively mask stable flame and allow unstable flame image frames. Performance comparison is done with a wellknown image processing tool, conditional random field that is trained to be selective as well. In this context, an informationtheoretic threshold value is derived. The proposed framework is validated on a set of real data collected from a laboratory scale combustor over varied operating conditions where it is shown to effectively detect subtle instability features as a combustion process makes transition from stable to unstable region.

Large eddy simulation of the unstable flame structure and gas-to-liquid thermal feedback in a medium-scale methanol pool fire

The objective of the present study is to perform fine-grained Large Eddy Simulations (LES) of a canonical 30-cm-diameter methanol pool fire configuration in order to evaluate the ability of current fire models to predict the flame structure and the rate of heat transfer to the liquid fuel surface. The simulations are performed using an LES solver called FireFOAM, a prescribed fuel evaporation rate, and a “wall-resolved” approach; the simulations also include a description of the fuel pan lip. Great attention is paid to the design of the computational grid and to the control of spatial resolution. Less attention is paid to the LES model formulation: except for the treatment of radiation for which we consider both a simplified model using a prescribed global radiant fraction and a more advanced model based on the Weighted-Sum-of-Gray-Gases approach, the simulations use baseline FireFOAM treatments to describe subgrid-scale turbulence and combustion. Results suggest that grid-converged solutions are achieved at a spatial resolution of 2 mm in the near-pool-surface region and 5 mm in the bulk of the flame region. Consistent with experimental observations, the simulated flames feature a strong instability characterized by the cyclic formation of thin boundary layer flames at the liquid pool surface and vortex rings at the burner rim that grow into large puffs and determine the structure of the entire flame. Predictions of the mean and root-mean-square temperature and velocity are found to be in good agreement with measurements taken from the literature. Equally encouraging results are obtained for the mean radiative and convective heat fluxes near the pool surface; quantitatively, simulations are found to overestimate the intensity of the thermal feedback by 25%. These results suggest that provided that the computational grid is fine enough, current fire models can be used to predict the gas-to-fuel thermal feedback.

Assessing the impact of multicomponent diffusion in direct numerical simulations of premixed, high-Karlovitz, turbulent flames

Implementing multicomponent diffusion models in numerical combustion studies is computationally expensive; to reduce cost, numerical simulations commonly use mixture-averaged diffusion treatments or simpler models. However, the accuracy and appropriateness of mixture-averaged diffusion has not been verified for three-dimensional, turbulent, premixed flames. In this study we evaluated the role of multicomponent mass diffusion in premixed, three-dimensional high Karlovitz-number hydrogen, n-heptane, and toluene flames, representing a range of fuel Lewis numbers. We also studied a premixed, unstable two-dimensional hydrogen flame due to the importance of diffusion effects in such cases. Our comparison of diffusion flux vectors revealed differences of 10–20% on average between the mixture-averaged and multicomponent diffusion models, and greater than 40% in regions of high flame curvature. Overall, however, the mixture-averaged model produces small differences in diffusion flux compared with global turbulent flame statistics. To evaluate the impact of these differences between the two models, we compared normalized turbulent flame speeds and conditional means of species mass fraction and source term. We found differences of 5–20% in the mean normalized turbulent flame speeds, which seem to correspond to differences of 5–10% in the peak fuel source terms. Our results motivate further study into whether the mixture-averaged diffusion model is always appropriate for DNS of premixed turbulent flames.

Propagation of Darrieus–Landau unstable laminar and turbulent expanding flames

The propagation of laminar and turbulent expanding flames subjected to Darrieus–Landau (DL), hydrodynamic instability was experimentally studied by employing stoichiometric H2/O2/N2 flames under quiescent and turbulent conditions performed in a newly developed medium-scale, fan-stirred combustion chamber. In quiescent environment, DL unstable laminar flame exhibits three-stage propagation, i.e. smooth expansion, transition acceleration, and self-similar acceleration. The self-similar acceleration is characterized by a power-law growth of acceleration exponent, α, with normalized Peclet number, which is different from the usually suggested self-similar propagation with a constant α. The imposed turbulence advances the onset of both transition acceleration and self-similar acceleration stages and promotes the strength of flame acceleration as additional wrinkles are invoked by turbulence eddies. A DL–turbulent interaction regime is confirmed to be the classical corrugated flamelets regime. Furthermore, the DL instability significantly facilitates the propagation of expanding flames in medium and even intense turbulence. The development of DL cells is not suppressed by turbulence eddies, and it needs to be considered in turbulent combustion.

Near-limit oscillatory behaviors on wick flames of dimethyl carbonate with trimethyl phosphate additions

The near-limit oscillatory behaviors on wick flames of dimethyl carbonate (DMC) with trimethyl phosphate (TMP) additions have been investigated experimentally. The experiments were conducted under a wick burner in conjunction with the limiting oxygen concentration test (wick-LOC), and the fuels were selected as typical examples of electrolyte solvents and organophosphorus compounds (OPCs) for lithium-ion batteries. The near-limit oscillating flames on the wick configuration were observed into two main characters: side-to-wake or side oscillation in for the side-stabilized flame (full flame) and wake oscillation for the wake-stabilized flame (wake flame). By tracing the flame base movement using a 240-fps camera, stable limit cycle oscillations were found in full flame cases, while the wake flame cannot sustain the oscillation for long and finally leads to global extinction. With the addition of TMP in DMC, the transition from the side-to-wake oscillation to the side oscillation was found in the unstable full flames with a linear increase in frequency and a significant drop in amplitude, while the wake oscillating flames performed increased trends for both. To clarify the dominant mechanism of TMP-added flame oscillations, laminar burning velocities of DMC+TMP mixtures were calculated at the oxygen level of each near-limit oscillating flame. The weakened flame speed with TMP addition revealed the inadequacy of buoyancy-driven mechanisms for the OPC added wick-flame. Then the wick surface temperature was measured adopting a special thermocouple arrangement to validate the thermal-diffusive promotion by TMP addition. Results showed that the heat feedback from TMP-added flames compensated for the low reactivity and provided a faster oscillation near the extinction limit.

Dynamics and stability of premixed hydrogen-air flames in square microchannels with wall temperature gradients

The dynamics and stability of premixed hydrogen-air flames in square microchannels with heated walls were investigated through three-dimensional direct numerical simulations. The inlet velocity and equivalence ratio were 1.5 m/s and 0.5. The effect of the wall temperature gradient characteristics on the flame dynamics and stability was examined varying the width and location of the wall temperature gradient for a channel height of 1.5 mm. Five distinct flame modes were observed at different wall temperature profiles: flame with repetitive extinction-ignition (FREI), pulsating flame, laterally oscillating flame, spinning flame, and steady flame modes. Furthermore, transitions between these flame modes were observed for specific inflow and boundary conditions. The effect of the channel height on the flame stability was investigated by varying the channel height from 1.0 mm to 1.677 mm for a fixed wall temperature gradient. As the channel height was increased, four of the flame modes, namely, FREI, laterally oscillating flame, spinning flame, and steady flame modes appeared sequentially. To determine whether this sequential appearance was associated with the variation of the wall heat loss, the maximum wall temperature was changed by small amounts. For a lower wall temperature, the laterally oscillating flame mode transitioned to the FREI mode, and for a higher wall temperature, unstable flame modes such as the FREI and laterally oscillating flame modes disappeared, resulting in stable flame.

Laminar burning velocities of 2-methyltetrahydrofuran at elevated pressures

The laminar burning velocities (LBVs) and cellular instability of 2-methyltetrahydrofuran (2-MTHF) were investigated at the unburned temperature of 423 K and pressures from 1 to 10 atm in a cylindrical constant-volume vessel. The LBVs of 2-MTHF/air flame exhibit a notably dropping with increasing pressure. The cellular instability analysis indicates that the critical flame radius of 2-MTHF/air mixture monotonically increases with increasing pressure and the flame surface suffers more badly cellularity under higher pressures. The critical flame radius exhibits non-monotonic variation versus ϕ and the most unstable flames appear at ϕ ≈ 1.3. It is observed that the measured Markstein length of 2-MTHF/air mixture decreases with increasing ϕ and Pu, leading to an earlier formation of wrinkling and cracks with respect to preferential-diffusional instability. Further investigation found that by using a mixture of 14.2% oxygen with 85.8% helium in place of air as bath gas at 10 atm can effectively suppress the cellular instability. Two recently developed models were used to simulate the experimental results and explore the chemical kinetic effects on LBV. Reaction path analysis reveals that the most consumption of 2-MTHF/air at stoichiometric conditions is through the abstraction of H-atom to form radical C5H9O-5. While the competitiveness of decomposition by CC scission yielding CH3 and tetrahydrofuran radical is relatively weak. Sensitivity analysis illustrates that small-species reactions show a controlling effect on LBV. The increasing pressure leads to an evident increase in the sensitivity coefficient of the recombination reaction H + O2 (+M)=HO2 (+M). The reduction of H atom concentration will cause competition to the initiation reaction H + O2 = O+OH. This could lower the overall oxidation rate and reduce the burning velocity.

Control of intrinsic thermoacoustic instabilities using hydrogen fuel

This conceptual numerical study explores a possible strategy for control of unstable intrinsic thermoacoustic (ITA) modes. Control of the ITA feedback loop has not been discussed yet. We propose addition of hydrogen fuel as a means of control and investigate the effect of incremental addition to the fuel mixture on the stability of a laminar slit flame using a variety of approaches: first, the ITA frequencies are estimated by a chemical kinetics solver for hydrogen fuel content up to 50% of the total fuel mass flow. Second, we perform DNS for each case and compute the Flame Transfer Function (FTF), from which the ITA mode frequencies and their stability can be estimated. Additionally, for each case the unstable flame is computed to confirm the estimated ITA mode frequencies. Third, an acoustic network model is employed, which uses the FTFs and predicts the stability limits observed by the DNS. Comparison with DNS data features very good agreement and allows to model the impact of hydrogen on the stability of the combustor.

Mitigation of Darrieus–Landau instability effects on turbulent premixed flames

Theoretical considerations on the competition between the most amplified modes for Darrieus–Landau (DL) hydrodynamic instability and turbulence timescales, show that, two extremal regimes can be identified: the instability-dominated and turbulence-dominated regimes. In the latter, also denoted as unified regime, both experiments and numerical simulations give evidence showing how the large scale, cusp-like structures of the flame front surface, typical of DL instability, are hindered by turbulent fluctuations. The result is that quantities such as turbulent flame propagation and front curvature statistics, which in the instability dominated regime are enhanced or modified by the overwhelming presence of hydrodynamic instability, are now mitigated and a unified regime is reached in which the characteristics of DL unstable and stable flame configurations become indistinguishable. In this work we analyze the concealing effects of increasing level of turbulence over the hydrodynamic Darrieus–Landau instability, and we show that, although some global indices such as the skewness of the curvature p.d.f. suggest that a unified regime is reached, others show the persistence of residual differences: in particular, the power spectral density of the flame front curvature. We use both experimental and numerical datasets of stable and unstable (based on linear stability analysis) flames, in conditions ranging from quasi laminar to significantly turbulent regimes.

Effect of Flow Parameters and Injection Flow Area on Non-Premixed Methane/Oxygen Diffusion Flame Stability

ABSTRACT The focus of this research was the design, fabrication, and testing of an experimental non-premixed flame burner to determine the effect of gaseous flow parameters and flow geometry on the stability of methane/oxygen combustion. The non-premixed burner consisted of a horizontally mounted, rectangular combustion chamber with a single coaxial injector, capable of introducing reactants at a specified impingement angle of 30° at the exit plane. Gaseous oxygen was the primary flow and gaseous methane was the secondary flow. The non-premixed burner was equipped with optical windows on both sides of the chamber parallel to the axis of the flame, which allowed for clear viewing of the product flame. Ignition of the reactants was achieved by a retractable spark plug. Stability maps of the resultant diffusion flame were created for three different injection flow areas based upon reactant equivalence ratio and primary reactant Reynolds numbers. The results showed that decreasing the primary flow area resulted in a more stable flame condition over a broader range of Reynolds numbers (laminar flow to over 40000) and equivalence ratios (fuel-lean 0.27 to fuel-rich 4.94). Increasing Reynolds number also resulted in a transition from a stable, anchored flame to a detached, unstable flame.

Role of H2/CO Addition to Flame Instabilities and Their Control in a Stepped Microcombustor

ABSTRACT The effect of H2/CO addition on flame dynamics of propane-air mixtures in a stepped microcombustor with heat recirculation was investigated experimentally. Unstable flame propagation regimes and their transition modes were observed for a range of mixture velocities (5–8 m/s) and equivalence ratios (0.6–0.95). Up to 60% H2 addition to lean propane-air mixtures helps improve the flame stability range in the combustor. The suppression of these flame instabilities with H2 addition is attributed to the increase in the overall mixture reactivity and burning velocity. The enhancement in OH intensity with H2 addition indicated a change in the elementary reactions, thereby affecting the flame dynamics. Contrary to H2 addition, CO addition does not show any noticeable effect on the formation of stable flame modes. The change in mixture properties and burning velocity leads to a change in the flame behavior and alters the heat release rate due to H2 addition to propane-air mixtures. A higher percentage of H2 addition led to the reappearance of the unstable (rotating) flames in the microcombustor. Intrinsic flame instabilities appeared at higher H2 percentages due to differential diffusion and increased thermal-wall coupling effects.

Continuous 500-Hz OH-PLIF Measurements in a Hydrogen-Fueled Scramjet Combustor

In recent years, significant research attention has focused on scramjet bench-testing using optical measurement techniques to obtain significant experimental data. The work aims to verify the environmental reliability of high-speed planar laser-induced fluorescence (PLIF) technique and satisfy bench-test requirements. A high-speed OH-PLIF system with ~1.8 mJ/pulse at 500 Hz is developed, and conducted tests at a direct-connect supersonic combustion facility. In this paper, OH-PLIF signals and pressure were measured simultaneously, and flame dynamics during pressure buildup were investigated, with a special focus on the ignition and flameout processes. In addition, the flame trajectory of the ignition process is depicted. The results show that the time scale of flame development is consistent with that of pressure stability, whereas the pressure change lags the change in flame intensity. The flame structure is discussed in both scramjet and ramjet modes. An unstable flame structure develops in ramjet mode, and the flame area decreases similarly in both combustion modes. Flame oscillations are captured for three hydrogen equivalence ratios.

Propagation of premixed flames in the presence of Darrieus–Landau and thermal diffusive instabilities

We study the propagation of premixed flames, in the absence of external turbulence, under the effect of both hydrodynamic (Darrieus–Landau) and thermodiffusive instabilities. The Sivashinsky equation in a suitable parameter space is initially utilized to parametrically investigate the flame propagation speed under the potential action of both kinds of instability. An adequate variable transformation shows that the propagation speed can collapse on a universal scaling law as a function of a parameter related to the number of unstable wavelengths within the domain nc. To assess whether this picture can persist in realistic flames, a DNS database of large scale, two-dimensional flames is presented, embracing a range of nc values and subject to either purely hydrodynamic instability (DL) or both kinds of instability (TD). With the aid of similar DNS databases from the literature we observe that when adequately rescaled, propagation speeds follow two distinct scaling laws, depending on the presence of thermodiffusive instability or lack thereof. We verify the presence of secondary cutoff values for nc identifying (a) the insurgence of secondary wrinkling in purely hydrodynamically unstable flames and (b) the attainment of domain independence in thermodiffusively unstable flames. A possible flame surface density based model for the subgrid wrinkling is also proposed.

Chemiluminescence-based characterization of tail biogas combustion stability under syngas and oxygen-enriched conditions

After the biogas is upgraded to CH4, a waste stream (tail biogas) is obtained, which contains methane in the order of few to tens vol.%. However, the composition of tail biogases can be unstable, therefore, safety should be maintained during the onsite combustion processes. Accordingly, a process control system with various sensors needs to be introduced. This study aims to analyze the experimental limits of flame stability using the chemiluminescence technique during the premixed combustion of low calorific tail biogas (mixture of CH4 with 15–30 vol% concentration in CO2) by adding oxygen or syngas. For the experimental investigation, a flat flame burner mounted in an enclosed combustion chamber was used. The stability of the flame was analyzed via the lift-off height by capturing the flame chemiluminescence from the excited OH* (282.9 nm), CH* (387.1 nm), and C2* (514.0 nm) radicals.This study reveals that the chemiluminescence spatial intensity of species such as OH*, CH*, and C2* is enhanced with an increase in the addition of syngas and oxygen. The peak position of the emission intensities in the flame shifts depending on the composition of gases, showing the existence of unstable flame that can reach its blow-off limit. When syngas is added, an excess of air leads to an increase in flame lifting; however, an O2 surplus results in more stable flame with minimized lifting. Moreover, it was determined that the lifting height of the flame is higher when syngas is added as compared to O2 enrichment. Finally, the flue gas emissions, such as CO and NOx, were examined and the most influential formation parameter was determined.

Aerodynamic Studies on Non-Premixed Oxy-Methane Flames and Separated Oxy-Methane Cold Jets

Both cold and flame jets find numerous applications in different fields, ranging from domestic applications to aerospace and space technology. Indeed, the applications of isothermal and non-isothermal jets in the flame heating industry fascinated the researchers to gain an in-depth understanding. Nevertheless, these benefits are not standalone, rather, they are associated with major disadvantages such as improper jet mixing and flame instabilities that require careful remedies. In the present investigation, three-inline jets, having methane jet at the center and two oxygen jets at the periphery, are studied computationally in a three-dimensional domain, with and without considering the effects of combustion. To study the mixing characteristics of cold jets, the radial velocity distributions at different streamwise locations have been analyzed at the jet inlet velocity of 27 m/s. However, for oxygen and methane flame jets, inlet velocities are varied as 27 m/s and 54 m/s. Moreover, to investigate the effects of temperature variation on mixing characteristics at a particular jet velocity, the inlet temperatures of reactants are varied as 300 K, 500 K, and 700 K, at the jet inlet velocity of 27 m/s. Combustion is found to have a profound impact on the mixing characteristics. At the inlet temperature of 300 K, a higher centerline velocity decay is observed for non-reactive jets as compared to reactive flame jets. Moreover, the turbulent kinetic energy distribution is relatively higher in the case of non-reactive jets, which is the direct evidence of an augmented mixing. As is obvious, the turbulent kinetic energy at the jet shear layer is maximum due to the formation of large-scale coherent eddies. The decay in centerline velocity is found to be increasing with an increase of inlet temperature. Additionally, with an increase of jet velocity, the radial velocity profiles become more asymmetrical, consequently yielding an unstable flame.

Effect of differential diffusion on turbulent lean premixed hydrogen enriched flames through structure analysis

This study presents the flame structure influenced by the differential diffusion effects and evaluates the structural modifications induced by the turbulence, thus to understand the coupling effects of the diffusively unstable flame fronts and the turbulence distortion. Lean premixed CH4/H2/air flames were conducted using a piloted Bunsen burner. Three hydrogen fractions of 0, 30% and 60% were adopted and the laminar flame speed was kept constant. The turbulence was generated with a single-layer perforated plate, which was combined with different bulk velocities to obtain varied turbulence intensities. Quasi-laminar flames without the plate were also performed. Explicit flame morphology was obtained using the OH-PLIF. The curvature, flame surface density and turbulent burning velocity were measured. Results show that the preferential transport of hydrogen produces negatively curved cusps flanked with positively curved bulges, which are featured by skewed curvature pdfs and consistent with the typical structure caused by the Darrieus-Landau instability. Prevalent bulge-cusp like wrinkles remain with relatively weak turbulence. However, stronger turbulence can break the bulges to be finer, and induce random positively curved cusps, therefore to destroy the bulge-cusp structures. Evident positive curvatures are generated in this process modifying the skewed curvature pdfs to be more symmetric, while the negative curvatures are not affected seriously. From low to high turbulence intensities, the hydrogen addition always strengthens the flame wrinkling. The augmentation of flame surface density and turbulent burning velocity with hydrogen is even more obvious at higher turbulence intensity. It is suggested that the differential diffusion can persist and even be strengthened with strong turbulence.

Interaction of flame instabilities and pressure behavior of hydrogen-propane explosion

The flame destabilization mechanism of hydrogen-propane-air mixture is firstly revealed. The effects of unstable flame formation on pressure rise rate and burning rate are quantified. Finally, the theoretical prediction of explosion pressure behavior is performed by considering diffusive-thermal and hydrodynamic instability. The results demonstrated that before the explosion pressure starts to climbe, as the propane fraction increases, the effective Lewis number of lean and stoichiometric mixture undergoes the transition from Leeff < 1.0 to Leeff > 1.0, the stabilizing effect of diffusive-thermal instability continues to reduce for the rich mixture. After the explosion pressure starts to climbe, the hydrogen-propane flame becomes more unstable, which is mainly attributed to enhancing hydrodynamic instability. The maximum rate of pressure rise and burning rate should be augmented by unstable flame formation, the flame instabilities must be considered in the explosion pressure evaluation.

CFD simulations of hydrogen deflagration in slow and fast combustion regime

This paper presents CFD modelling of hydrogen-air deflagrations. The proposed mathematical models are validated with the experiments provided to SARNET2 (Severe Accident Research NETwork of Excellence 2) project [A. Bentaib, A. Bleyer, N. Chaumeix, B. Schramm, P. Kostka, M. Movahed, H.S. Kang and M. Povilaitis, Final results of the SARNET Hydrogen deflagration Benchmark Effect of turbulence on flame acceleration. (2012), pp. 1–15; A. Bentaib, A. Bleyer, N. Meynet, N. Chaumeix, B. Schramm, M. Höhne, P. Kostka, M. Movahed, S. Worapittayaporn, T. Brähler, H. Seok-Kang, M. Povilaitis, I. Kljenak and P. Sathiah, SARNET hydrogen deflagration benchmarks: Main outcomes and conclusions. Ann. Nucl. Energy 74 (2014), pp. 143–152. doi:10.1016/j.anucene.2014.07.012]. The goal of SARNET2 project was to investigate the effect of blockage ratio on flame propagation and pressure evolution during deflagration of a lean mixture of hydrogen and air. Three blockage ratios were tested: BR = 0, 0.33 and 0.63. The experiments were carried out in the ENACCEF facility. The proposed mathematical models test the effect of turbulence models and radiative heat losses on flame front position and absolute pressure in the testing facility. A good agreement between the calculated and experimental results for the three BRs is found for the SST k-ω model alone and in combination with the γ transition model when radiative heat losses are accounted for in the modelling. In addition, an unstable tulip flame is evidenced in our CFD simulations what corroborates previous reports on premixed combustion in a closed tube.

Chemiluminescence as a diagnostic in studying combustion instability in a practical combustor

Prediction of combustion instability is essential in reducing development costs of large scale rocket engines. Recently, high fidelity CFD models have shown some ability to represent the stability characteristics of sub-scale experimental combustors. A challenge remains in acquiring experimental measurements of critical combusting flow properties, especially measurements of the unsteady heat addition field, for comparison to models. The chemiluminescent emission of radicals OH* and CH*, although commonly employed, may fall short due to its known dependence on pressure, strain rate, equivalence ratio, and turbulence level. Since these flame properties have large variations in an unstable rocket combustor, it is difficult to decouple the emission from all other variables except the heat release. Computations capture the local variations, and therefore can be used for a direct comparison with the measurements. The present study investigates the relationship between chemiluminescent species and heat release rate of the flame using simulations that employ detailed kinetics including the emitting species. This relationship is then compared to experimental spectral measurements captured with a fiber optic probe to check the validity of the model and to provide insight on the ability of the chemiluminescent light emission to indicate heat release in a high pressure unstable flame.

Hydrogen effect on lean flammability limits and burning characteristics of an isooctane–air mixture

The flammability limit is a hurdle to realizing ultra-lean combustion and achieving high thermal efficiency in spark ignition engines. Hydrogen, known for its wide explosion limits, has been utilized in gasoline to improve the lean limit. The influence of hydrogen on lean flammability limits of isooctane was studied at initial temperatures of 393–513 K and an ignition energy (IE) of 3000 mJ, while the burning characteristics were studied at an initial temperature of 393 K and IE = 37 mJ. It is found that a spherical flame appeared when the isooctane–air–H2 mixture was ignited from λ = 0.8 to λ = 1.2 at IE = 37 mJ, whereas a leaner mixture was burnt at IE = 3000 mJ with unstable flame growth during the early stage. Flame instability eventually evanesced and buoyancy was observed near the lean limits. The lean limits of isooctane–air increased to 2.1λ with 0–90% hydrogen. Raising the initial temperature from 393 to 513 K increased the lean limits of isooctane–air–H2 to 0.2λ. Hydrogen obtained from CH2O (C, H, and O species) makes a considerable contribution to improving the burning near the lean limits. The mole fraction of the short-lived radicals H, OH, and O decreased when the excess air coefficient moved toward the lean limits and increased by introducing hydrogen, which improved the lean flammability limits and burning velocity.