Search

Within the scope of reacting flow simulations, the real-time direct integration (DI) of stiff ordinary differential equations (ODE) for the computation of chemical kinetics stands as the primary demand on computational resources. Meanwhile, as the number of transport equations that need to be solved increases, the computational cost grows more substantially, particularly for those combustion models involving direct coupling of chemistry and flow such as the transported probability density function model. In the current study, an integrated Graphics Processing Unit-Artificial Neural Network (GPU-ANN) framework is introduced to comply with heavy computational costs while maintaining high fidelity. Within this framework, a GPU-based solver is employed to solve partial differential equations and compute thermal and transport properties, and an ANN is utilized to replace the calculation of reaction rates. Large eddy simulations of two swirling flames provide a robust validation, affirming and extending the GPU-ANN approach's applicability to challenging scenarios. The simulation results demonstrate a strong correlation in the macro flame structure and statistical characteristics between the GPU-ANN approach and the traditional Central Processing Unit (CPU)-based solver with DI. This comparison indicates that the GPU-ANN approach is capable of attaining the same degree of precision as the conventional CPU-DI solver, even in more complex scenarios. In addition, the overall speed-up factor for the GPU-ANN approach is over two orders of magnitude. This study establishes the potential groundwork for widespread application of the proposed GPU-ANN approach in combustion simulations, addressing various and complex scenarios based on detailed chemistry, while significantly reducing computational costs.

In the stratified or partially premixed piloted jet flames, previous experimental and priori studies have identified a strong correlation between mixture fraction and progress variable. In the framework of large-eddy simulation (LES) and flamelet-generated manifolds (FGM) approach, a joint probability density function (PDF) method is constructed to characterize subgrid correlations. To pave the way for high dimensional tabulation modeling, a deep residual network (ResNet) is trained, dramatically reducing the memory footprint of tabulation. The Message Passing Interface (MPI) shared memory technique is applied to load the original chemical table during parallel computations. Application of LES to a partially pre-vaporized ethanol spray flame demonstrates good agreement with experimental results. Consideration of the subgrid correlation results in a noticeable improvement in temperature prediction. Calculations using ResNet show a notable consistency with those using chemical tables. Visualization of enthalpy highlights the significance of non-adiabatic tabulation in modeling liquid fuel combustion. The unscaled progress variable is selected to better describe the chemical reaction rate in the blending zone of an air stream and a pilot stream with the product of a fully burnt lean fuel mixture. The impact of the source term due to evaporation in the transport equation of the progress variable is validated. The correlation coefficient is found to significantly influence the chemical reaction rate. The subgrid-scale interaction between liquid fuel evaporation and subgrid correlation is elucidated.

The application of deep neural networks (DNNs) holds considerable promise as a substitute for the direct integration of chemical source terms in combustion simulations. However, challenges persist in ensuring high precision and generalisation across various different fuels and flow conditions. In this study, we propose and validate a consistent DNN approach for chemistry integration in a range of fuels and premixed flame configurations. This approach generates thermochemical base state from a set of low-dimensional laminar flames, followed by an effective perturbation strategy to enhance the coverage of the composition space for higher generalisation ability. A constraint criterion based on heat release rate is then employed to remove the nonphysical perturbed states for improved accuracy.Without specific tuning, three DNNs are consistently trained for three representative fuels, i.e., hydrogen, ethylene and Jet-A. Comprehensive validations are conducted using 1-D laminar flames and two typical turbulent premixed flames. The DNN model predictions on various physical characteristics, including laminar and turbulent flame speeds, dynamic flame structures influenced by turbulence-chemistry interactions, and conditional scalar profiles, all exhibit good agreement with the results obtained from direct integration. This demonstrates the exceptional accuracy and generalisation ability of the proposed DNN approach. Furthermore, when the DNN is used in the simulation, a significant speed-up for the chemistry integration is achieved, approximately 50 for the ethylene/air flame and 90 for the Jet-A/air flame.

Recent progress in artificial intelligence (AI) and high-performance computing (HPC) have brought potentially game-changing opportunities in accelerating reactive flow simulations. In this study, we introduce an open-source computational fluid dynamics (CFD) framework that integrates the strengths of machine learning (ML) and graphics processing unit (GPU) to demonstrate their combined capability. Within this framework, all computational operations are solely executed on GPU, including ML-accelerated chemistry integration, fully-implicit solving of PDEs, and computation of thermal and transport properties, thereby eliminating the CPU-GPU memory copy overhead. Optimisations both within the kernel functions and during the kernel launch process are conducted to enhance computational performance. Strategies such as static data reorganisation and dynamic data allocation are adopted to reduce the GPU memory footprint. The computational performance is evaluated in two turbulent flame benchmarks using quasi-DNS and LES modelling, respectively. Remarkably, while maintaining a similar level of accuracy to the conventional CPU/CVODE-based solver, the GPU/ML-accelerated approach shows an overall speedup of over two orders of magnitude for both cases. This result highlights that high-fidelity turbulent combustion simulation with finite-rate chemistry that requires normally hundreds of CPUs can now be performed on portable devices such as laptops with a medium-end GPU.

Stratified premixed combustion, known for its capability to expand flammability limits and reduce overall-lean combustion instability, has been widely adopted to comply with increasingly stringent environmental regulations. Numerous numerical simulations with different combustion models and mesh resolutions have been conducted on laboratory-scale flames to further understand the stratified premixed combustion. However, the trade-off between the high-fidelity and low computational cost for simulating laboratory-scale flames still remains, particularly for those combustion models involving direct coupling of chemistry and flow. In the present study, a GPU-based solver is employed to solve partial differential equations and calculate the thermal and transport properties, while an artificial neural network (ANN) is introduced to replace reaction rate calculation. Particular emphasis is placed on evaluating the proposed GPU-ANN approach through the large eddy simulation of the Cambridge stratified flame. The simulation results show good agreement for the flow and flame statistics between the GPU-ANN approach and the conventional CPU-based solver with direct integration (DI). The comparison suggests that the GPU-ANN approach can achieve the same level of accuracy as the conventional CPU-DI solver. In addition, the overall speed-up factor for the GPU-ANN approach is over two orders of magnitude. This study lays the potential groundwork for fully resolved laboratory-scale flame simulations based on detailed chemistry with much more affordable computational cost.

Accurate and affordable simulation of supercritical reacting flow is of practical importance for developing advanced engine systems for liquid rockets, heavy-duty powertrains, and next-generation gas turbines. In this work, we present detailed numerical simulations of LOX/GCH4 flame-vortex interaction under supercritical conditions. The well-established benchmark configuration of three-dimensional Taylor-Green vortex (TGV) embedded with a diffusion flame is modified for real fluid simulations. Both ideal gas and Peng-Robinson (PR) cubic equation of states are studied to reveal the real fluid effects on the TGV evolution and flame-vortex interaction. The results show intensified flame stretching and quenching arising from the intrinsic large density gradients of real gases, as compared to that for the idea gases. Furthermore, to reduce the computational cost associated with real fluid thermophysical property calculations, a machine learning-based strategy utilising deep neural networks (DNNs) is developed and then assessed using the three-dimensional reactive TGV. Generally good prediction accuracy is achieved by the DNN, meanwhile providing a computational speed-up of 13 times over the convectional approach. The profound physics involved in flame-vortex interaction under supercritical conditions demonstrated by this study provides a benchmark for future related studies, and the machine learning modelling approach proposed is promising for practical high-fidelity simulation of supercritical combustion.

In this work, supersonic premixed H2/air combustion stabilised by a cavity flame-holder in a model scramjet combustor with a Mach 1.5 inflow is studied via direct numerical simulations. By applying turbulent and laminar inlet condition separately, we compare the flame stabilisation characteristics between these two cases to study the effect of inflow turbulence. The ignition and flame propagation process are similar in both cases.After combustion in the area near the cavity reaches quasi-stationary state, the results show that with inflow turbulence more robust vortices are developed in cavity shear layer and near the aft wall, resulting in a thicker reaction zone and larger flame surface which present more efficient combustion. In addition, the inlet turbulence enhances the entry of fresh mixture and this further supports flame stabilisation near the cavity. However, the fully developed cavity shear layer also leads to greater cavity resistance. As for the flame stabilization, with inlet turbulence more robust recirculation and stronger mass exchange process is observed and promote the radical transport. These results give a preliminary insight into the flame stabilization mechanism. The distribution of progress variable indicates that the flame stabilized toward products side with turbulent inlet.

Understanding the interactions between hydrogen flame and turbulent vortices is important for developing the next-generation carbon neutral combustion systems. In the present work, we perform several direct numerical simulation (DNS) cases to study the dynamics of a hydrogen diffusion flame embedded in the Taylor-Green Vortex (TGV). The evolution of flame and vortex is investigated for a range of initial Reynolds numbers up to 3200 with different mass diffusion models. We show that the vortices dissipate rapidly in cases at low Reynolds numbers, while the consistent stretching, splitting and twisting of vortex tubes are observed in cases with evident turbulence transition at high Reynolds numbers. Regarding the interactions between the flame and vortex, it is demonstrated that the heat release generated by the flame has suppression effects on the turbulence intensity and its development of the TGV. Meanwhile, the intense turbulence provides abundant kinetic energy, accelerating the mixing of the diffusion flame with a contribution to higher strain rate and larger curvatures of the flame. Considering the effects of non-unity Lewis number, it is revealed that the flame strength is more intense in the cases with mixture averaged model. However, this effect is relatively suppressed under the impacts of the intense turbulence.

In this work, we introduce DeepFlame, an open-source C++ platform with the capabilities of utilising machine learning algorithms and pre-trained models to solve for reactive flows. We combine the individual strengths of the computational fluid dynamics library OpenFOAM, machine learning framework Torch, and chemical kinetics program Cantera. The complexity of cross-library function and data interfacing (the core of DeepFlame) is minimised to achieve a simple and clear workflow for code maintenance, extension and upgrading. As a demonstration, we apply our recent work on deep learning for predicting chemical kinetics (Zhang et al. Combust. Flame vol. 245 pp. 112319, 2022) to highlight the potential of machine learning in accelerating reacting flow simulation. A thorough code validation is conducted via a broad range of canonical cases to assess its accuracy and efficiency. The results demonstrate that the convection-diffusion-reaction algorithms implemented in DeepFlame are robust and accurate for both steady-state and transient processes. In addition, a number of methods aiming to further improve the computational efficiency, e.g. dynamic load balancing and adaptive mesh refinement, are explored. Their performances are also evaluated and reported. With the deep learning method implemented in this work, a speed-up of two orders of magnitude is achieved in a simple hydrogen ignition case when performed on a medium-end graphics processing unit (GPU). Further gain in computational efficiency is expected for hydrocarbon and other complex fuels. A similar level of acceleration is obtained on an AI-specific chip - deep computing unit (DCU), highlighting the potential of DeepFlame in leveraging the next-generation computing architecture and hardware.

Machine learning has long been considered as a black box for predicting combustion chemical kinetics due to the extremely large number of parameters and the lack of evaluation standards and reproducibility. The current work aims to understand two basic questions regarding the deep neural network (DNN) method: what data the DNN needs and how general the DNN method can be. Sampling and preprocessing determine the DNN training dataset, further affect DNN prediction ability. The current work proposes using Box-Cox transformation (BCT) to preprocess the combustion data. In addition, this work compares different sampling methods with or without preprocessing, including the Monte Carlo method, manifold sampling, generative neural network method (cycle-GAN), and newly-proposed multi-scale sampling. Our results reveal that the DNN trained by the manifold data can capture the chemical kinetics in limited configurations but cannot remain robust toward perturbation, which is inevitable for the DNN coupled with the flow field. The Monte Carlo and cycle-GAN samplings can cover a wider phase space but fail to capture small-scale intermediate species, producing poor prediction results. A three-hidden-layer DNN, based on the multi-scale method without specific flame simulation data, allows predicting chemical kinetics in various scenarios and being stable during the temporal evolutions. This single DNN is readily implemented with several CFD codes and validated in various combustors, including (1). zero-dimensional autoignition, (2). one-dimensional freely propagating flame, (3). two-dimensional jet flame with triple-flame structure, and (4). three-dimensional turbulent lifted flames. The results demonstrate the satisfying accuracy and generalization ability of the pre-trained DNN. The Fortran and Python versions of DNN and example code are attached in the supplementary for reproducibility.

Self-excited spinning mode azimuthal instability in an annular combustor with non-swirling flow is investigated using large eddy simulation (LES). Compressible Navier-Stoke equations are solved with a flamelet combustion model to describe the subgrid chemistry$-$turbulence interactions. Two flamelet models, with and without heat loss effects, are compared to elucidate the non-adiabatic wall effects on the thermoacoustic instability. The azimuthal modes are captured well by both models with only marginal differences in the computed frequencies and amplitudes. By comparing with the experimental measurements, the frequencies given by the LES are approximately 10\% higher and the amplitudes are well predicted. Further analysis of the experimental and LES data shows a similar dominant anti-clockwise spinning mode, under which a good agreement is observed for the phase-averaged heat release rate fluctuations. Dynamic mode decomposition (DMD) is applied to shed more light on this spinning mode. The LES and experimental DMD modes reconstructed for their azimuthal mode frequencies agree very well for the heat release fluctuations. The DMD mode structure for the acoustic pressure from the LES shows a considerable non-zero profile at the combustor outlet, which could be essential for azimuthal modes to establish in this annular combustor. Finally, a low-order modelling study was conducted using an acoustic network combined with the flame transfer function extracted from LES. The results show that the dominant mode is associated with the plenum showing a first longitudinal and azimuthal mixed mode structure. By tuning the plenum length to match the effective volume, the predicted frequency becomes very close to the measured value.

Large Eddy Simulation with a Perfectly Stirred Reactor model (LES-PSR) is developed to simulate supersonic combustion with high-enthalpy flow conditions. The PSR model considers the viscous heating and compressibility effects on the thermo-chemical state, through correcting the chemical source term for progress variable and incorporating absolute enthalpy as the control variable for the look-up table. It is firstly validated by using a priori analysis of the viscous heating and compressibility effects. Then an auto-igniting hydrogen flame stabilized in supersonic vitiated co-flowing jet is simulated with LES-PSR method. The results show that the shock wave structure, overall flame characteristics, flame-shock interaction and lift-off height are accurately captured. Good agreements of the velocity and mixture fraction statistics with the experimental data are observed. The results also show that the LES-PSR model can predict the mean temperature and mole fractions of major species quite well in both flame induction and stabilization zones. However, there are some under-predictions of temperature RMS by about 100-150 K, which may be due to the chemical non-equilibrium in the H2/O2-enriched combustion product of the co-flowing jet. The scatter plots of two probe locations respectively from induction and flame zones show that the respective flame structures in mixture fraction space are captured well. However, the flucturations of the temperature and species mole fractions are under-predicted in the flame zone. The shock-induced auto-igniting spots are captured by the PSR model. These spots are highly unsteady and play an important role in flame stabilization. It is also shown that the intense reactions are initiated at mixture fractions around the stoichiometry or fuel-lean values, corresponding to local elevated pressure (1.5-2.0 atm) due to shock compression.