JAX-Fluids: A fully-differentiable high-order computational fluid dynamics solver for compressible two-phase flows

Bezgin, Deniz A.; Buhendwa, Aaron B.; Adams, Nikolaus A.

doi:10.1016/j.cpc.2022.108527

Cited by 34 publications

(13 citation statements)

References 80 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…32,35,47 The most common way to account for fluid compressibility is to introduce an equation of state (EOS). [48][49][50] Therefore, the locally averaged compressible Navier-Stokes equations with the same additional EOS as expressed in reference 34 are applied to investigate the influence of fluid compressibility on the mechanical behaviors of geomaterials. These equations can be combined as follows:…”

Section: Establishment and Solution Principle Of The Fluid Modulementioning

confidence: 99%

“…Numerous works considering the impact of fluid compressibility have been previously carried out with some meaningful results obtained 32,35,47 . The most common way to account for fluid compressibility is to introduce an equation of state (EOS) 48–50 . Therefore, the locally averaged compressible Navier—Stokes equations with the same additional EOS as expressed in reference 34 are applied to investigate the influence of fluid compressibility on the mechanical behaviors of geomaterials.…”

Section: Principles and Formulationsmentioning

confidence: 99%

See 1 more Smart Citation

A novel unresolved/semi‐resolved CFD‐DEM coupling method with dynamic unstructured mesh

He,

Li,

Chen

2024

Num Anal Meth Geomechanics

View full text Add to dashboard Cite

The greatest challenge when performing large deformation simulations using the CFD‐DEM coupling method lies in the dynamical update of the fluid meshes. To address this problem, a novel CFD‐DEM coupling method integrated with the dynamic unstructured grid is proposed in this work. The mesh initialization and reconstruction are performed by the Constrained Delaunay triangulation (CDT) implemented by the tetgen algorithm. Moreover, the equation of state (EOS) is introduced to consider the impact of fluid compressibility. The reliability of this coupling method is preliminarily verified by the particle sedimentation test. Additionally, undrained triaxial shear and one‐dimensional consolidation tests are conducted to examine the applicability of the proposed method in simulating geotechnical cases. From our analysis, it is found that the shear responses in the undrained triaxial shear test obtained from the proposed method with dynamic mesh are in direct accordance with those led by the constant volume (CV) method. However, difficulties arise in accurately describing the moving boundary by the fixed mesh, leading to unreliable results. Moreover, in the one‐dimensional consolidation test, both the dynamic and fixed grids can capture the Mandel—Cryer effect, which is more pronounced under the dynamic grid, mainly due to the dual impact of the loading device and particle skeleton on the fluid. A greater compressibility specified under the dynamic grid led to a weakening of this effect, whereas its duration is increased. Our work provides valuable insights for effectively dealing with dynamic fluid boundaries.

show abstract

Section: Establishment and Solution Principle Of The Fluid Modulementioning

confidence: 99%

Section: Principles and Formulationsmentioning

confidence: 99%

A novel unresolved/semi‐resolved CFD‐DEM coupling method with dynamic unstructured mesh

He,

Li,

Chen

2024

Num Anal Meth Geomechanics

View full text Add to dashboard Cite

show abstract

“…To enable the use of a quantity for which a model cannot be directly supervised, we choose to develop a differentiable fluid simulator for plasma physics called ADEPT 4 . While DSs have been used to construct models for computational fluid dynamics [6,[21][22][23][24][25], and for physics discovery in plasma kinetics [26], to our knowledge, this is the first DS of computational plasma fluid dynamics. The code is publicly available along with the dataset introduced in this work.…”

Section: Inclusion Of Kinetic Effects In a Fluid Codementioning

confidence: 99%

Machine learning of hidden variables in multiscale fluid simulation

Joglekar,

Thomas

2023

Mach. Learn.: Sci. Technol.

View full text Add to dashboard Cite

Solving fluid dynamics equations often requires the use of closure relations that account for missing microphysics. For example, when solving equations related to fluid dynamics for systems with a large Reynolds number, sub-grid effects become important and a turbulence closure is required, and in systems with a large Knudsen number, kinetic effects become important and a kinetic closure is required. By adding an equation governing the growth and transport of the quantity requiring the closure relation, it becomes possible to capture microphysics through the introduction of ``hidden variables'' that are non-local in space and time. The behavior of the ``hidden variables'' in response to the fluid conditions can be learned from a higher fidelity or ab-initio model that contains all the microphysics. In our study, a partial differential equation simulator that is end-to-end differentiable is used to train judiciously placed neural networks against ground-truth simulations. We show that this method enables an Euler equation based approach to reproduce non-linear, large Knudsen number plasma physics that can otherwise only be modeled using Boltzmann-like equation simulators such as Vlasov or Particle-In-Cell modeling.

show abstract

“…Applying reverse-mode AD to general purpose numerical programs is often termed differentiable programming. It has been applied extensively in, for example, the geosciences [13], in computational fluid dynamics [14][15][16][17], and in accelerator physics [18,19]. Nascent applications toward plasma physics include the development of theoretical plasma physics [20] and the implementation of reduced models in multiscale plasma physics [21].…”

Section: Introductionmentioning

confidence: 99%

Qualitative and quantitative enhancement of parameter estimation for model-based diagnostics using automatic differentiation with an application to inertial fusion

Milder,

Joglekar,

Rozmus

et al. 2024

Mach. Learn.: Sci. Technol.

View full text Add to dashboard Cite

Parameter estimation using observables is a fundamental concept in the experimental sciences. Mathematical models that represent the physical processes can enable reconstructions of the experimental observables and greatly assist in parameter estimation by turning it into an optimization problem which can be solved by gradient-free or gradient-based methods. In this work, the recent rise in flexible frameworks for developing differentiable scientific computing programs is leveraged in order to dramatically accelerate data analysis of a common experimental diagnostic relevant to laser–plasma and inertial fusion experiments, Thomson scattering. A differentiable Thomson-scattering data analysis tool is developed that uses reverse-mode automatic differentiation (AD) to calculate gradients. By switching from finite differencing to reverse-mode AD, three distinct outcomes are achieved. First, gradient descent is accelerated dramatically to the extent that it enables near real-time usage in laser–plasma experiments. Second, qualitatively novel quantities which require O ( 10 3 ) parameters can now be included in the analysis of data which enables unprecedented measurements of small-scale laser–plasma phenomena. Third, uncertainty estimation approaches that leverage the value of the Hessian become accurate and efficient because reverse-mode AD can be used for calculating the Hessian.

show abstract

JAX-Fluids: A fully-differentiable high-order computational fluid dynamics solver for compressible two-phase flows

Cited by 34 publications

References 80 publications

A novel unresolved/semi‐resolved CFD‐DEM coupling method with dynamic unstructured mesh

A novel unresolved/semi‐resolved CFD‐DEM coupling method with dynamic unstructured mesh

Machine learning of hidden variables in multiscale fluid simulation

Qualitative and quantitative enhancement of parameter estimation for model-based diagnostics using automatic differentiation with an application to inertial fusion

Contact Info

Product

Resources

About