Sparse identification of multiphase turbulence closures for coupled fluid–particle flows

Beetham, Sarah; Fox, Rodney O.; Capecelatro, Jesse

doi:10.1017/jfm.2021.53

Cited by 55 publications

(25 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The work of Ling et al [30] designed a custom neural network layer that enforced Galilean invariance in the Reynolds stress tensors that they were modeling. Related Reynolds stress models have been developed using the SINDy sparse modeling approach [87][88][89]. Hybrid models that combine linear system identification and nonlinear neural networks have been used to model complex aeroelastic systems [90].…”

Section: Examples In Fluid Mechanicsmentioning

confidence: 99%

“…SINDy has been used to generate reduced-order models for how dominant coherent structures evolve in a flow for a range of configurations [100,[102][103][104][105]. These models have also been extended to develop compact closure models [87][88][89]. Recently, the physical notion of boundedness of solutions, which is a fundamental concept in reduced-order models of fluids [106], has been incorporated into the SINDy modeling framework as a novel loss function.…”

Section: The Loss Functionmentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning for Fluid Mechanics

Brunton

Noack

Koumoutsakos

2020

Annu. Rev. Fluid Mech.

1,969

742

View full text Add to dashboard Cite

The field of fluid mechanics is rapidly advancing, driven by unprecedented volumes of data from experiments, field measurements, and large-scale simulations at multiple spatiotemporal scales. Machine learning presents us with a wealth of techniques to extract information from data that can be translated into knowledge about the underlying fluid mechanics. Moreover, machine learning algorithms can augment domain knowledge and automate tasks related to flow control and optimization. This article presents an overview of past history, current developments, and emerging opportunities of machine learning for fluid mechanics. We outline fundamental machine learning methodologies and discuss their uses for understanding, modeling, optimizing, and controlling fluid flows. The strengths and limitations of these methods are addressed from the perspective of scientific inquiry that links data with modeling, experiments, and simulations. Machine learning provides a powerful information processing framework that can augment, and possibly even transform, current lines of fluid mechanics research and industrial applications.

show abstract

Section: Examples In Fluid Mechanicsmentioning

confidence: 99%

Section: The Loss Functionmentioning

confidence: 99%

Machine Learning for Fluid Mechanics

Brunton

Noack

Koumoutsakos

2020

Annu. Rev. Fluid Mech.

1,969

742

View full text Add to dashboard Cite

show abstract

“…ANN is able to process big data and produce good predictions in the training step, which makes it a powerful data mining technique 41 . ANN models have been designed and trained to promote advances in data‐driven flow turbulence modeling, 48,49 recognize the flow regime transition in trickle bed 50 and so on 51‐53 . The ANN model can also be used to predict the relationship between the input markers and the output without requiring a specific function expression, which is very flexible for modeling in fTFM.…”

Section: Introductionmentioning

confidence: 99%

Data‐driven modeling of mesoscale solids stress closures for filtered two‐fluid model in gas–particle flows

2021

View full text Add to dashboard Cite

This study performs data‐driven modeling of mesoscale solids stress closures for filtered two‐fluid model (fTFM) in gas–particle flows via an artificial neural network (ANN) based machine learning method. The data used for developing the ANN‐based predictive data‐driven modeling framework is systematically filtered from fine‐grid simulations. The loss function optimization result reveals that coupling two loss functions promotes more accurate predictions of the mesoscale solids stresses than using a single loss function. Further comprehensive assessments of closure markers demonstrate a systematic dependence of the mesoscale solids stresses on the filtered particle velocity and its gradient as additional anisotropic markers, instead of using the conventional isotropic filtered rate of solid phase deformation as a closure marker. An optimal three‐marker mesoscale closure is thus proposed. Comparative analysis of the conventional filtered model and present three‐marker model shows that the data‐driven model can substantially enhance the prediction accuracy.

show abstract

“…Moreover, validation of the proposed heat transfer correction for the coarse‐grid simulations is needed. Additionally, it is worth developing an explicit algebraic form that can robustly predict mesoscale flow and transport data using alternative DDM techniques from the perspective of ease‐of‐use and practicality 21,55 …”

Section: Discussionmentioning

confidence: 99%

Conventional anddata‐drivenmodeling of filtered drag, heat transfer, and reaction rate ingas–particleflows

Zhu

Ouyang

et al. 2021

AIChE Journal

View full text Add to dashboard Cite

This study presents conventional and artificial neural network‐based data‐driven modeling (DDM) methods to model simultaneously the filtered mesoscale drag, heat transfer and reaction rate in gas–particle flows. The dataset used for developing the DDM is filtered from highly resolved simulations closed by our recently formulated microscopic drag and heat transfer coefficients (HTCs). Results reveal that the filtered drag correction is nearly independent of filter size when including the filtered gas phase pressure gradient. We further find that the filtered HTC correction critically depends on the added filtered temperature difference marker while the filtered reaction rate correction shows weak dependence on the additional markers. Moreover, compared with conventional correlations, DDM predictions agree better with filtered resolved data. Comparative analysis is also conducted between existing HTC corrections and our work. Finally, the applicability of conventional and data‐driven models coupled with coarse‐grid computational fluid dynamics simulations for pilot‐scale (reactive) gas–particle flows is validated comprehensively.

show abstract

Sparse identification of multiphase turbulence closures for coupled fluid–particle flows

Abstract: Abstract

Cited by 55 publications

References 69 publications

Machine Learning for Fluid Mechanics

Machine Learning for Fluid Mechanics

Data‐driven modeling of mesoscale solids stress closures for filtered two‐fluid model in gas–particle flows

Conventional anddata‐drivenmodeling of filtered drag, heat transfer, and reaction rate ingas–particleflows

Contact Info

Product

Resources

About