An adaptive time-integration scheme for stiff chemistry based on computational singular perturbation and artificial neural networks

Galassi, Riccardo Malpica; Ciottoli, Pietro Paolo; Valorani, Mauro; Im, Hong G.

doi:10.1016/j.jcp.2021.110875

Cited by 17 publications

(3 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…52 To date, a TSR-based timescale was never used in the context of reactor-based models. We are aware that the TSR estimate comes along with a computational cost, which is orders of magnitude higher than the simpler approximations presented above, even though this cost may be substantially alleviated resorting to Machine Learning techniques to train a surrogate model for the eigen-system, as shown in Malpica Galassi et al 53 However, in this work, we are mostly concerned with the accuracy of the s c submodels, rather than their efficiency. As a final remark on the chemical timescale sub-models, we point out that all the thermo-chemical quantities in Eqs.…”

Section: Formation Rates Time Scales (Fr)mentioning

confidence: 99%

The partially stirred reactor model for combustion closure in large eddy simulations: Physical principles, sub-models for the cell reacting fraction, and open challenges

et al. 2022

Self Cite

View full text Add to dashboard Cite

For their ability to account for finite-rate chemistry, reactor-based models are well suited Turbulence–Chemistry Interactions (TCI) Sub-Grid Scale (SGS) closures for Large Eddy Simulations (LES). The SGS closure in the Partially Stirred Reactor (PaSR) model relies on the determination of the reacting fraction of each computational cell, whose definition is based on estimates of the characteristic mixing and chemical time scales. Direct Numerical Simulations (DNS) of turbulent combustion can supply key information on TCI for the development, validation, and comparison of combustion models. In particular, a priori testing allows the direct validation of model assumptions. In the present work, an a priori assessment of the PaSR model is conducted. Its ability to reconstruct thermo-chemical quantities of interest is investigated along with model assumptions. Sub-grid quantities are extracted from the DNS to investigate the role of the cell reacting fraction. Various definitions are then proposed to estimate the characteristic chemical timescale in the PaSR model. Modeled chemical source terms and heat release rates are compared against the filtered quantities from DNS data of a two-dimensional, spatially developing, turbulent nonpremixed jet flame with detailed kinetics. The results demonstrate the importance of accounting for the fine structures quantities in the context of reactor-based models. A new formulation of the chemical timescale is proposed and provides improved overall predictions. Several issues are raised in the discussion, representing realistic prospects for further developments of the PaSR model as a SGS combustion closure for LES.

show abstract

Section: Formation Rates Time Scales (Fr)mentioning

confidence: 99%

The partially stirred reactor model for combustion closure in large eddy simulations: Physical principles, sub-models for the cell reacting fraction, and open challenges

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…The reactive system is then characterized in the ROM description by employing a limited set of composition variables, such as significant species or latent variables. The equations governing the evolution of these reduced composition variables are then solved, as in the G-Scheme [21,22] and the dynamically-informed autoencoder with neural ODE [23].…”

Section: Introductionmentioning

confidence: 99%

Analysis of droplet evaporation dynamics using computational singular perturbation and tangential stretching rate

Angelilli,

Galassi,

Ciottoli

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Computational singular perturbation (CSP) has been successfully used in the analysis of complex chemically reacting flows by systematically identifying the intrinsic timescales and slow invariant manifolds that capture the essential subprocesses driving the dynamics of the system. In this article, the analytical and computational framework is applied for the first time to analyze the Lagrangian droplets undergoing evaporation and dispersion in the surrounding gases. First, a rigorous mathematical formulation is derived to adapt the CSP tools into the droplet dynamics equations, including the formal definition of the tangential stretching rate (TSR) that represents the explosive/dissipative nature of the system. A steady ammonia and a falling water droplet studies are then conducted to demonstrate the utility of the CSP methodology in identifying various physical mechanisms driving the evolution of the system, such as the distinction of thermal-driven and mass-driven regimes. Various definitions of the importance indices are also examined to provide in-depth analysis of different subprocesses and their interactions in modifying the droplet dynamics.

show abstract

“…Variety methodologies in mechanism reduction have been developed in decades of research to minimize the mechanism sizes and stiffness. It can be characterized into roughly four primary groups based on its functions, including skeletal (Xue et al, 2020;Wu et al, 2020), lumping (Till et al, 2019;Brunialti et al, in press), time-scale analysis (Koniavitis et al, 2017;Chang et al, 2020), and stiffness reduction (Felden et al, 2019;Malpica Galassi et al, 2022), in which they employ the user-specified tolerance for regulating accuracy. Several public mechanisms are usually constructed in a restricted combustion condition, such as a flow regime, to obtain a more downsize, resulting in a narrow range of usability (Zhao et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

Assessment and development of a novel reduced mechanism for methane combustion in computational fluid dynamics modeling

Kamma

Suvanjumrat

2022

Preprint

View full text Add to dashboard Cite

The main objective of this work is to obtain the reduced reaction mechanism, which is consistent with a benchmark case in modeling a 0-D ignition delay, 1-D laminar flame speed, and 2-D simulated flame result and spent less processing time. In achieving this, the ten reduced reaction mechanisms developed for methane combustion were assessed, whereas the GRI-Mech 3.0 is considered a Benchmark. The result showed that only a reaction mechanism named SK30 was satisfactory. Still, the processing time in simulating the simple 2-D of a premixed model at the microscale was overly substantial. Subsequently, SK30 was further lessened using the two reduction steps. Firstly, the automatic algorithm based on a direct relation graph with the error propagation aided sensitivity analysis using ignition delays as a criterion in automatic reduction was applied. By doing this, the accuracy of ignition delays was maintained, and the flame speed was distorted. Accordingly, sensitivity analysis was employed to obtain the influential reaction in the benchmarking mechanism in the second step. The significant species and reactions on flame speed but less in ignition delay, which was missing in the current development, were considered to retrieve back manually as few as possible. Finally, the novel mechanism consisting of 25 species 132 reactions was proposed for methane-air combustion. In validation, the 1-D flame speed and the 2-D premixed flame model were agreement with the benchmark model. In addition, the processing time of this reduced mechanism was 50% faster than the SK30.

show abstract

An adaptive time-integration scheme for stiff chemistry based on computational singular perturbation and artificial neural networks

Cited by 17 publications

References 36 publications

The partially stirred reactor model for combustion closure in large eddy simulations: Physical principles, sub-models for the cell reacting fraction, and open challenges

The partially stirred reactor model for combustion closure in large eddy simulations: Physical principles, sub-models for the cell reacting fraction, and open challenges

Analysis of droplet evaporation dynamics using computational singular perturbation and tangential stretching rate

Assessment and development of a novel reduced mechanism for methane combustion in computational fluid dynamics modeling

Contact Info

Product

Resources

About