Potential Flow Generator With L 2 Optimal Transport Regularity for Generative Models
Not provided.
Computer Science↗
SEARCH · Search NASA
Results for “generative flow model”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
RockFlow: Fast Generation of Synthetic Source Rock Images Using Generative Flow Models
Image-based evaluation methods are a valuable tool for source rock characterization. The time and resources needed to obtain images has spurred development of machine-learning generative models to create synthetic images of pore structure and rock fabric from limited image data. While generative models have shown success, existing methods for generating 3D volumes from 2D training images are restricted to binary images and grayscale volume generation requires 3D training data. Shale characterization relies on 2D imaging techniques such as scanning electron microscopy (SEM), and grayscale values carry important information about porosity, kerogen content, and mineral composition of the shale. Here, we introduce RockFlow, a method based on generative flow models that creates grayscale volumes from 2D training data. We apply RockFlow to baseline binary micro-CT image volumes and compare performance to a previously proposed model. We also show the extension of our model to 2D grayscale data by generating grayscale image volumes from 2D SEM and dual modality nanoscale shale images. The results show that our method underestimates the porosity and surface area on the binary baseline datasets but is able to generate realistic grayscale image volumes for shales. With improved binary data preprocessing, we believe that our model is capable of generating synthetic porous media volumes for a very broad class of rocks from shale to carbonates to sandstone.
58 GEOSCIENCES↗
Learning turbulent flows with generative models for super resolution and sparse flow reconstruction
Neural operators are promising surrogates for dynamical systems but when trained with standard L 2 losses they tend to oversmooth fine-scale turbulent structures. Here, we show that combining operator learning with generative modeling overcomes this limitation. We consider three practical turbulent-flow challenges where conventional neural operators fail: spatio-temporal super-resolution, forecasting, and sparse flow reconstruction. For Schlieren jet super-resolution, an adversarially trained neural operator (adv-NO) reduces the energy-spectrum error by 15 × while preserving sharp gradients at neural operator-like inference cost. For 3D homogeneous isotropic turbulence, adv-NO trained on only 160 timesteps from a single trajectory forecasts accurately for five eddy-turnover times and offers 114 × wall-clock speed-up at inference than the baseline diffusion-based forecasters, enabling near-real-time rollouts. For reconstructing cylinder wake flows from highly sparse Particle Tracking Velocimetry-like inputs, a conditional generative model infers full 3D velocity and pressure fields with correct phase alignment and statistics. These advances enable accurate reconstruction and forecasting at low compute cost, bringing near-real-time analysis and control within reach in experimental and computational fluid mechanics.
Fluid dynamics↗
A Deep Gravity model for mobility flows generation
The movements of individuals within and among cities influence critical aspects of our society, such as well-being, the spreading of epidemics, and the quality of the environment. When information about mobility flows is not available for a particular region of interest, we must rely on mathematical models to generate them. In this work, we propose Deep Gravity, an effective model to generate flow probabilities that exploits many features (e.g., land use, road network, transport, food, health facilities) extracted from voluntary geographic data, and uses deep neural networks to discover non-linear relationships between those features and mobility flows. Our experiments, conducted on mobility flows in England, Italy, and New York State, show that Deep Gravity achieves a significant increase in performance, especially in densely populated regions of interest, with respect to the classic gravity model and models that do not use deep neural networks or geographic data. Deep Gravity has good generalization capability, generating realistic flows also for geographic areas for which there is no data availability for training. Finally, we show how flows generated by Deep Gravity may be explained in terms of the geographic features and highlight crucial differences among the three considered countries interpreting the model’s prediction with explainable AI techniques.
97 MATHEMATICS AND COMPUTING↗
MENT-Flow: maximum-entropy phase space tomography using normalizing flows
Generative models can be trained to reproduce low-dimensional projections of high-dimensional phase space distributions. Normalizing flows are generative models that parameterize invertible transformations, allowing exact probability density evaluation and sampling. Consequently, flows are unbiased entropy estimators and could be used to solve the high-dimensional maximum-entropy tomography (MENT) problem. In this work, we evaluate a flow-based MENT solver (MENT-Flow) against exact maximum-entropy solutions and Minerbo's iterative MENT algorithm in two dimensions.
Hoover, Austin↗
Simulation and Modeling of Flow Generated by Gust Generator in a Wind Tunnel
A computational study of a gust field generated by a gust generator in a low-speed wind tunnel. The gust generator is designed for the University of Washington Aeronautical Laboratory (UWAL) Kirsten wind tunnel for a gust load alleviation (GLA) control experiment of a Common Research Model (CRM) flexible wing utilizing the Variable Camber Continuous Trailing Edge Flap (VCCTEF). The gust generator comprises four horizontal NACA 0015 gust generator vanes placed upstream of the test section. Computational fluid dynamics simulations using a two-dimensional (2D) Unsteady-Reynolds-Averaged-Navier-Stokes (URANS) with k-ω Shear Stress Transport (SST) turbulence model provide detailed time-resolved information about the generated flow by the gust generator under prescribed sinusoidal motion. The characteristics of the induced flow by the gust generator are analyzed. A gust propagation model of the gust field is investigated. An unsteady lift model is developed using a varying-fidelity approach which includes a 2D interference aerodynamic model of the combined gust generator-wing system. The computed integrated unsteady lift is compared to experimental data for validation of the unsteady lift model. Both the amplitude and transport delay are found to be accurately captured by the unsteady lift model.
Nhan, Nguyen↗
One- and two-dimensional Stirling machine simulation using experimentally generated flow turbulence models
Investigations of one- and two-dimensional (1- or 2-D) simulations of Stirling machines centered around experimental data generated by the U. of Minnesota Mechanical Engineering Test Rig (METR) are covered. This rig was used to investigate oscillating flows about a zero mean with emphasis on laminar/turbulent flow transitions in tubes. The Space Power Demonstrator Engine (SPDE) and in particular, its heater, were the subjects of the simulations. The heater was treated as a 1- or 2-D entity in an otherwise 1-D system. The 2-D flow effects impacted the transient flow predictions in the heater itself but did not have a major impact on overall system performance. Information propagation effects may be a significant issue in the simulation (if not the performance) of high-frequency, high-pressure Stirling machines. This was investigated further by comparing a simulation against an experimentally validated analytic solution for the fluid dynamics of a transmission line. The applicability of the pressure-linking algorithm for compressible flows may be limited by characteristic number (defined as flow path information traverses per cycle); this warrants further study. Lastly the METR was simulated in 1- and 2-D. A two-parameter k-w foldback function turbulence model was developed and tested against a limited set of METR experimental data.
Goldberg, Louis F.↗
Active learning for the design of polycrystalline textures using conditional normalizing flows
Generative modeling has opened new avenues for solving previously intractable materials design problems. However, these new opportunities are accompanied by a drastic increase in the required amount of training data. This is in stark juxtaposition to the high expense and difficulty in curating such large materials datasets. In this work, we propose a novel framework for integrating generative models within an active learning loop. Further, this enables the training of generative models with datasets significantly smaller than what has previously been demonstrated, providing a direct route for their application in data constrained environments. The functionality of this framework is then demonstrated by addressing the challenge of designing polycrystalline textures associated with target anisotropic mechanical properties. The developed protocol exhibited a cost reduction between 14 to 18 times over a randomly sampled experimental design.
36 MATERIALS SCIENCE↗
Application of a computational model for vortex generators in subsonic internal flows
A model for the analysis of vortex generators in a fully viscous subsonic internal flow is evaluated. A vorticity source term is used in a modified form of the Parabolized Navier-Stokes equations to model the shed vortex. Computed results are compared with idealized flow vortex paths, and with experimental data for vortex generators embedded in a thick turbulent boundary layer. The analysis is also compared with experimental data for a separated diffusing S-duct and for a diffusing S-duct with vortex generators. Quantitative comparisons are shown for the latter three cases. Emphasis is placed on verifying the ability of the model to predict global distortions in the flow field.
Kunik, W. G.↗
Application of a computational model for vortex generators in subsonic internal flows
A model for the analysis of vortex generators in a fully viscous subsonic internal flow is evaluated. A vorticity source term is used in a modified form of the Parabolized Navier-Stokes equations to model the shed vortex. Computed results are compared with idealized flow vortex paths, and with experimental data for vortex generators embedded in a thick turbulent boundary layer. The analysis is also compared with experimental data for a separated diffusing S-duct and for a diffusing S-duct with vortex generators. Quantitative comparisons are shown for the latter three cases. Emphasis is placed on verifying the ability of the model to predict global distortions in the flow field.
Kunik, W. G.↗
Reconstructing porous media using generative flow networks
One area of intense scientific interest for the study of sandstones, carbonates, and shale at the pore scale is the use of limited image and petrophysical data to generate multiple realizations of a rock’s pore structure. Such images aid efforts to quantify uncertainty in petrophysical properties, including porosity–permeability transforms. We develop and evaluate a deep learning-based method to synthesize porous media volumes using a so-called generative flow model trained on x-ray microscope images of rock texture and pore structure. These models are optimized on a log-likelihood objective and they synthesize large and realistic three-dimensional images. Further, we demonstrate the rapid generation of sandstone image volumes that display realism as gauged by quantitative comparison of topological features using Minkowski functionals of porosity, specific surface area, and the Euler–Poincaré characteristic (i.e., pore connectivity). We also evaluate the single-phase permeability using Navier–Stokes and lattice Boltzmann methods and show that transport properties of the generated samples match measured trends.
58 GEOSCIENCES↗
Applications of flow models to the generation of correlated lattice QCD ensembles
Machine-learned normalizing flows can be used in the context of lattice quantum field theory to generate statistically correlated ensembles of lattice gauge fields at different action parameters. This work demonstrates how these correlations can be exploited for variance reduction in the computation of observables. Three different proof-of-concept applications are demonstrated using a novel residual flow architecture: continuum limits of gauge theories, the mass dependence of QCD observables, and hadronic matrix elements based on the Feynman–Hellmann approach. In all three cases, it is shown that statistical uncertainties are significantly reduced when machine-learned flows are incorporated as compared with the same calculations performed with uncorrelated ensembles or direct reweighting. Published by the American Physical Society 2024
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
A Combined Geometric Approach for Computational Fluid Dynamics on Dynamic Grids
A combined geometric approach for computational fluid dynamics is presented for the analysis of unsteady flow about mechanisms in which its components are in moderate relative motion. For a CFD analysis, the total dynamics problem involves the dynamics of the aspects of geometry modeling, grid generation, and flow modeling. The interrelationships between these three aspects allow for a more natural formulation of the problem and the sharing of information which can be advantageous to the computation of the dynamics. The approach is applied to planar geometries with the use of an efficient multi-block, structured grid generation method to compute unsteady, two-dimensional and axisymmetric flow. The applications presented include the computation of the unsteady, inviscid flow about a hinged-flap with flap deflections and a high-speed inlet with centerbody motion as part of the unstart / restart operation.
Slater, John W.↗
Transonic, turbulent boundary-layer separation generated on an axisymmetric flow model
Experimental data describing the transonic, turbulent, separated flow generated by an axisymmetric flow model are presented. The model consisted of a circular-arc bump affixed to a straight, circular cylinder aligned with the flow direction. Measurements of the mean velocity, turbulence intensity, and Reynolds shear-stress profiles were made in the separated flow. These data revealed dramatic changes in the shear-stress levels as the flow passed through the interaction to reattachment. Behavior of the turbulence reaction to the imposed pressure gradients was examined in terms of the mixing length and the excursions of the turbulence from equilibrium.
Bachalo, W. D.↗
An investigation of transonic turbulent boundary layer separation generated on an axisymmetric flow model
Experimental data are presented describing the transonic turbulent separated flow generated by an axisymmetric flow model. The model consisted of a circular-arc bump affixed to a straight circular cylinder aligned with the flow direction. Measurements of the mean velocity, turbulence intensity, and Reynolds shear stress profiles were made in the separated flow. These data revealed the dramatic changes in the shear stress levels as the flow passed from the interaction through to reattachment. Information on the behavior of the turbulence reaction to the imposed pressure gradients, as presented in this investigation, will be required for the development of the turbulence models used in predicting nonequilibrium turbulent flow fields.
Bachalo, W. D.↗
A source-sink model of the generation of plate tectonics from non-Newtonian mantle flow
A model of mantle convection which generates plate tectonics requires strain rate- or stress-dependent rheology in order to produce strong platelike flows with weak margins as well as strike-slip deformation and plate spin (i.e., toroidal motion). Here, we employ a simple model of source-sink driven surface flow to determine the form of such a rheology that is appropriate for Earth's present-day plate motions. In this model, lithospheric motion is treated as shallow layer flow driven by sources and sinks which correspond to spreading centers and subduction zones, respectively. Two plate motion models are used to derive the source sink field. As originally implied in the simpler Cartesian version of this model, the classical power law rheologies do not generate platelike flows as well as the hypothetical Whitehead-Gans stick-slip rheology (which incorporates a simple self-lubrication mechanism). None of the fluid rheologies examined, however, produce more than approximately 60% of the original maximum shear. For either plate model, the viscosity fields produced by the power law rheologies are diffuse, and the viscosity lows over strike-slip shear zones or pseudo-margins are not as small as over the prescribed convergent-divergent margins. In contrast, the stick-slip rheology generates very platelike viscosity fields, with sharp gradients at the plate boundaries, and margins with almost uniformly low viscosity. Power law rheologies with high viscosity contrasts, however, lead to almost equally favorable comparisons, though these also yield the least platelike viscosity fields. This implies that the magnitude of toroidal flow and platelike strength distributions are not necessarily related and thus may present independent constraints on the determination of a self-consistent plate-mantle rheology.
Bercovici, David↗
Simultaneous three-dimensional velocity and mixing measurements by use of laser Doppler velocimetry and fluorescence probes in a water tunnel
A water tunnel investigation was conducted to demonstrate the capabilities of a laser-based instrument that can measure velocity and fluorescence intensity simultaneously. Fluorescence intensity of an excited fluorescent dye is directly related to concentration level and is used to indicate the extent of mixing in flow. This instrument is a three-dimensional laser Doppler velocimeter (LDV) in combination with a fluorometer for measuring fluorescence intensity variations. This capability allows simultaneous flow measurements of the three orthogonal velocity components and mixing within the same region. Two different flows which were generated by two models were studied: a generic nonaxisymmetric nozzle propulsion simulation model with an auxiliary internal water source that generated a jet flow and an axisymmetric forebody model with a circular sector strake that generated a vortex flow. The off-body flow fields around these models were investigated in the Langley 16- by 24-Inch Water Tunnel. The experimental results were used to calculate 17 quantities that included mean and fluctuating velocities, Reynolds stresses, mean and fluctuating dye fluorescence intensities (proportional to concentration), and fluctuating velocity and dye concentration correlations. An uncertainty analysis was performed to establish confidence levels in the experimental results. In general, uncertainties in mean velocities varied between 1 and 7 percent of free-stream velocity; uncertainties in fluctuating velocities varied between 1 and 5 percent of reference values. The results show characteristics that are unique to each type of flow.
Neuhart, Dan H.↗