Co-optimized machine-learned manifold models for large eddy simulation of turbulent combustion

Perry, Barbara; Frahan, Marc Henry de; Yellapantula, Shashank

doi:10.1016/j.combustflame.2022.112286

Cited by 10 publications

(4 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This result can explain the improved ROM performance reported in the literature. 24 , 25 , 26 In Figure 3 A, we demonstrate this using three reacting flow datasets for the combustion of hydrogen, methane, and ethylene in air. We generate 2D and 3D projections of high-dimensional thermochemical state spaces.…”

Section: Resultsmentioning

confidence: 91%

“…This approach naturally penalizes projections that exhibit nonuniqueness or large gradients, since any difficulties in representing QoIs on a projection immediately increase the mean-squared-error (MSE) loss function during training. Evidence from the existing research 24 , 25 , 26 suggests that such a joint encoding-decoding 27 , 28 , 29 , 30 , 31 approach provides improvements to the topology of a low-dimensional data projection. Here, we demonstrate quantitatively that this is the case.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improving reduced-order models through nonlinear decoding of projection-dependent outputs

Zdybał,

Parente,

Sutherland

2023

Patterns

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 91%

Section: Introductionmentioning

confidence: 99%

Improving reduced-order models through nonlinear decoding of projection-dependent outputs

Zdybał,

Parente,

Sutherland

2023

Patterns

View full text Add to dashboard Cite

“…Its unique features consist in combining an AMR approach with a low Mach number formulation to achieve high performances from a small desktop station to the world's largest supercomputer. Recent code developments focused on enabling massively parallel simulations at scale on high-performance accelerated computer architectures to tackle the challenging requirements of fundamental and applied combustion research, as well as extending the solver modeling capabilities by including Large Eddy Simulation (LES) closure models and support for data-driven combustion models (Perry et al, 2022).…”

Section: Statement Of Needmentioning

confidence: 99%

PeleLMeX: an AMR Low Mach Number Reactive Flow Simulation Code without level sub-cycling

Esclapez,

Day,

Bell

et al. 2023

JOSS

Self Cite

View full text Add to dashboard Cite

PeleLMeX simulates chemically reacting low Mach number flows with block-structured adaptive mesh refinement (AMR). The code is built upon the AMReX (Zhang et al., 2019) library, which provides the underlying data structures and tools to manage and operate on them across massively parallel computing architectures. PeleLMeX algorithmic features are inherited from its predecessor PeleLM (PeleLM Team, 2022) but key improvements allow representation of more complex physical processes. Together with its compressible flow counterpart PeleC (Henry de Frahan et al., 2023), the thermo-chemistry library PelePhysics and the multi-physics library PeleMP, it forms the Pele suite of open-source reactive flow simulation codes.

show abstract

“…The neural network trained from such data set has been confirmed to be accurate and efficient in predicting reaction rates under various conditions. Besides, DNN was also extended to the high-dimensional tabulation of flamelets and effectively reduced the memory requirement [9][10][11][12].…”

Section: Introductionmentioning

confidence: 99%

DeepFlame: A deep learning empowered open-source platform for reacting flow simulations

Mao¹,

Lin²,

Zhang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Recent developments in deep learning have brought many inspirations for the scientific computing community and it is perceived as a promising method in accelerating the computationally demanding reacting flow simulations. 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.

show abstract

Co-optimized machine-learned manifold models for large eddy simulation of turbulent combustion

Cited by 10 publications

References 66 publications

Improving reduced-order models through nonlinear decoding of projection-dependent outputs

Improving reduced-order models through nonlinear decoding of projection-dependent outputs

PeleLMeX: an AMR Low Mach Number Reactive Flow Simulation Code without level sub-cycling

DeepFlame: A deep learning empowered open-source platform for reacting flow simulations

Contact Info

Product

Resources

About