Search NASA⌕ Search

SEARCH · Search NASA

Results for “normalizing flows”

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.

At least 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
NASA NTRS1959

An Investigation of the Performance of Various Reaction Control Devices

An investigation of a small-scale reaction control devices in still air with both subsonic and supersonic internal flows has shown that lateral forces approaching 70 percent of the resultant force of the undeflected jet can be obtained. These results were obtained with a tilted extension at a deflection of 40 deg. The tests of tilted extensions indicated an optimum length-to-diameter ratio of approximately 0.75 to 1.00, dependent upon the deflection angle. For the two geometric types of spoiler tabs tested, blockage-area ratio appears to be the only variable affecting the lateral force developed. Usable values of lateral force were developed by the full-eyelid type of device with reasonably small losses in the thrust and weight flow. Somewhat larger values of lateral force were developed by injecting a secondary flow normal to the primary jet, but for conditions of these tests the losses in thrust and weight flow were large. Relatively good agreement with other investigations was obtained for several of the devices. The agreement of the present results with those of an investigation made with larger-scale equipment indicates that Reynolds number may not be critical for these tests. In as much as the effects of external flow could influence the performance and other factors affecting the choice of a reaction control for a specific use, it would appear desirable to make further tests of the devices described in this report in the presence of external flow.

Hunter, Paul A.↗
NASA NTRS2000

Phase Shift Interferometer and Growth Set Up to Step Pattern Formation During Growth From Solutions. Influence of the Oscillatory solution Flow on Stability

We have assembled an experimental setup based on Michelson interferometry with the growing crystal surface as one of the reflective surfaces. The crystallization part of the device allows optical monitoring of a face of a crystal growing at temperature stable within 0.05 C in a flow of solution of controlled direction and speed. The reference arm of the interferometer contains a liquid crystal element that allows controlled shifts of the phase of the interferograms. We employ an image-processing algorithm, which combines five images with a pi/2 phase difference between each pair of images. The images are transferred to a computer by a camera capable of capturing 60 frames per second. The device allows data collection on surface morphology and kinetics during the face layers growth over a relatively large area (approximately 4 sq. mm) in situ and in real time during growth. The estimated depth resolution of the phase shifting interferometry is approximately 50 Angstroms. The data will be analyzed in order to reveal and monitor step bunching during the growth process. The crystal chosen as a model for study in this work is KH2PO4 (KDP). This optically non-linear material is widely used in frequency doubling applications. There have been a number of studies of the kinetics of KDP crystallization that can serve as a benchmark for our investigations. However, so far, systematic quantitative characteristics of step interaction and bunching are missing. We intend to present our first quantitative results on the onset, initial stages and development of instabilities in moving step trains on vicinal crystal surfaces at varying supersaturation, flow rate, and flow direction. Behavior of a vicinal face growing from solution flowing normal to the steps and periodically changing its direction in time was considered theoretically. It was found that this oscillating flow reduces both stabilization and destabilization effects resulted from the unidirectional solution flow directed up the step stream and down the step stream. This reduction of stabilization and destabilization comes from effective mixing which entangles the phase shifts between the spatially periodic interface perturbation and the concentration wave induced by this perturbation. Numerical results and simplified mixing criterion will be discussed.

Chernov, Alex A.↗
NASA NTRS1973

An integral turbulent kinetic energy analysis of free shear flows

Mixing of coaxial streams is analyzed by application of integral techniques. An integrated turbulent kinetic energy (TKE) equation is solved simultaneously with the integral equations for the mean flow. Normalized TKE profile shapes are obtained from incompressible jet and shear layer experiments and are assumed to be applicable to all free turbulent flows. The shear stress at the midpoint of the mixing zone is assumed to be directly proportional to the local TKE, and dissipation is treated with a generalization of the model developed for isotropic turbulence. Although the analysis was developed for ducted flows, constant-pressure flows were approximated with the duct much larger than the jet. The axisymmetric flows under consideration were predicted with reasonable accuracy. Fairly good results were also obtained for the fully developed two-dimensional shear layers, which were computed as thin layers at the boundary of a large circular jet.

Peters, C. E.↗
NASA NTRS1991

The Development of High Order Numerical Techniques for Reentry Simulation of Hypersonic Spacecraft

The primary difficulty encountered when simulating hypersonic flow is that the flow normally includes strong nonlinear discontinuities. These discontinuities fall into three broad classes: shocks, slip-lines, and rarefaction waves. Moreover, in the hypersonic flow regime, the chemistry of hot gases plays a vital role and can not be neglected. These facts combine to make the numerical treatment of spacecraft reentry a most challenging problem. In this work, we develop a class of finite difference schemes that accurately resolve discontinuous solutions to spacecraft reentry flow and are simple to incorporate into existing spacecraft reentry codes.

Sanders, Richard↗
NASA NTRS1995

Non-radial flow in the solar wind

Although the radial component of the solar wind dominates the solar wind speed, significant non-radial velocity components are also present. These flows are more difficult to measure accurately, but we now have data sets including the east-west (tangential) and north-south (normal) flows from PVO at Venus, IMP 8 at Earth, and Voyagers 1 and 2 from 1 to 45 AU. We compare the non-radial flow observations from these spacecraft. One of the more interesting features is that the north-south flow angle observed at Earth and Venus oscillates with the period of a local (Earth or Venus) year. These oscillations occur throughout two solar cycles in the IMP 8 data set and are very apparent in the PVO data from 1978 to 1986 but less obvious after this. We will report on the origin of this feature. The tangential flow observed by both IMP 8 and Voyager is on average slightly positive (approximately 1.75 km/s). The magnitudes of the nonradial velocity components decrease with distance from the Sun.

Richardson, J. D.↗
DOE OSTI2026

Conditional deep generative models for simultaneous simulation and reconstruction of entire events

We extend the particle-flow neural assisted simulations (arnassus) framework of fast simulation and reconstruction to entire collider events. In particular, we use two generative artificial intelligence tools, continuous normalizing flows and diffusion models, to create a set of reconstructed particle-flow objects conditioned on truth-level particles from CMS Open Simulations. While previous work focused on jets, our updated methods now can accommodate all particle-flow objects in an event along with particle-level attributes like particle type and production vertex coordinates. This approach is fully automated, entirely written in Python, and GPU-compatible. Using a variety of physics processes at the LHC, we show that the extended arnassus is able to generalize beyond the training dataset and outperforms the standard, public tool elphes.

Dreyer, Etienne [Weizmann Institute of Science, Re↗
DOE OSTI2025

Flow matching beyond kinematics: Generating jets with particle identification and trajectory displacement information

We introduce the first generative model trained on the etlass dataset. Our model generates jets at the constituent level, and it is a permutation-equivariant continuous normalizing flow (CNF) trained with the flow matching technique. It is conditioned on the jet type, so that a single model can be used to generate the ten different jet types of etlass. For the first time, we also introduce a generative model that goes beyond the kinematic features of jet constituents. The etlass dataset includes more features, such as particle-ID and track impact parameter, and we demonstrate that our CNF can accurately model all of these additional features as well. Our generative model for etlass expands on the versatility of existing jet generation techniques, enhancing their potential utility in high-energy physics research, and offering a more comprehensive understanding of the generated jets. Published by the American Physical Society 2025

Birk, Joschka (ORCID:0000000219310127)↗
NASA NTRS1976

Reduction of wind-tunnel wall interference by controlled wall flow

Corrections for wind tunnel wall interferences are applied successfully to high lift models when the model to tunnel size ratio is small. The accuracy of the corrections becomes poorer when larger models are tested. An alternate method of testing was developed in which flow through the porous walls of the tunnel was actively controlled so as to approximate free air conditions in the neighborhood of the model during the test. The amount and distribution of the controlled flow through the walls is computed using a potential flow representation of the model based on the measured lift. Theoretical analysis is presented to prove the convergence of the method to free air conditions and to substantiate the general three dimensional theory of operation when the normal flow distribution is continuous. A two dimensional tunnel was constructed to evaluate the concept. Results show that substantial reduction of wall interference may be achieved with relatively low values of porosity of actively controlled walls.

Bernstein, S.↗
NASA NTRS1975

Reduction of wind tunnel wall interference by controlled wall flow

An alternate method of testing was developed in which flow through the porous walls of the tunnel was actively controlled so as to approximate free air conditions in the neighborhood of the model during the test. The amount and distribution of the controlled flow through the walls is computed using a potential flow representation of the model based on the measured lift. Theoretical analysis is presented to prove the convergence of the method to free air conditions and to substantiate the general three-dimensional theory of operation when the normal flow distribution is continuous. A two-dimensional tunnel was constructed to evaluate the concept. Results show that substantial reduction of wall interference may be achieved with relatively low values of porosity of actively controlled walls.

Bernstein, S.↗
NASA NTRS1994

Interactive computer graphics applications for compressible aerodynamics

Three computer applications have been developed to solve inviscid compressible fluids problems using interactive computer graphics. The first application is a compressible flow calculator which solves for isentropic flow, normal shocks, and oblique shocks or centered expansions produced by two dimensional ramps. The second application couples the solutions generated by the first application to a more graphical presentation of the results to produce a desk top simulator of three compressible flow problems: 1) flow past a single compression ramp; 2) flow past two ramps in series; and 3) flow past two opposed ramps. The third application extends the results of the second to produce a design tool which solves for the flow through supersonic external or mixed compression inlets. The applications were originally developed to run on SGI or IBM workstations running GL graphics. They are currently being extended to solve additional types of flow problems and modified to operate on any X-based workstation.

Benson, Thomas J.↗
NASA NTRS1989

Euler procedure for three-dimensional transonic wall interference

Based on an optimization formulation, a procedure has been developed to evaluate Mach number and angle-of-attack corrections. The Euler equations are assumed to be the flow governing equations. To obtain efficient solutions for the optimization problem, the iterative solutions for the flow variables and the design parameters are simultaneously updated. In addition to the model lift and geometry, the procedure requires pressure measurements near the tunnel walls. The tunnel boundary conditions are based on the introduction of Reimann invariants for a one-dimensional flow normal to the boundary. Computations are performed to verify that the errors introduced by this approximate boundary-condition formulation are acceptably small. The correction scheme is applied to an aircraft configuration in an open jet. The results indicate that the optimization scheme is highly efficient with the rate of convergence of the flow solution nearly equal to the corresponding rate of a regular analysis problem.

Rizk, Magdi H.↗
NASA NTRS1976

Influence of particle drag coefficient on particle motion in high-speed flow with typical laser velocimeter applications

The effect of using different particle drag coefficient C sub D equations for computing the velocity of seeded particles in high-speed gas flows was investigated. The C sub D equations investigated included the Stokes equation, a second incompressible equation valid for higher relative Reynolds numbers, and six equations that account for the effects of compressibility together with the effects of relative Reynolds numbers greater than one. The flows investigated were center-line nozzle flows, normal shocks, and oblique shocks for free-stream Mach numbers of 1.6 to 6 and stagnation pressures of 1 and 3.4 atmospheres. The net result was empirical C sub D equation based on the latest sphere C sub D data for the low relative Mach number and Reynolds number conditions that are encountered in supersonic flows.

Walsh, M. J.↗
NASA NTRS1953

Method for Calculation of Laminar Heat Transfer in Air Flow Around Cylinders of Arbitrary Cross Section (including Large Temperature Differences and Transpiration Cooling)

The solution of heat-transfer problems has become vital for many aeronautical applications. The shapes of objects to be cooled can often be approximated by cylinders of various cross sections with flow normal to the axis as, for instance heat transfer on gas-turbine blades and on air foils heated for deicing purposes. A laminar region always exists near the stagnation point of such objects. A method previously presented by E. R. G. Eckert permits the calculation of local heat transfer around the periphery of cylinders of arbitrary cross section in the laminar region for flow of a fluid with constant property values with an accuracy sufficient for engineering purposes. The method is based on exact solutions of the boundary-layer equations for incompressible wedge-type flow and on the postulate that at any point on the cylinder the boundary-layer growth is the same as that on a wedge with comparable flow conditions. This method is extended herein to take into account the influence of large temperature differences between the cylinder wall and the flow as well as the influence of transpiration cooling when the same medium as the outside flow is used as coolant.

Eckert, E R↗
NASA NTRS1991

The interaction of superfluid vortex filaments with a normal fluid channel flow

Vortex filaments in superfluids such as helium 2 may provide new insights into very high Reynolds number flows. The behavior of a superfluid vortex ring interacting with a normal fluid shear flow, specifically channel flow, is simulated. The vortex ring evolves into a stable horseshoe configuration which propagates without further change of form. In this simulation, a boundary layer behavior in a superfluid through the coupling of the superfluid and the normal fluid is demonstrated.

Samuels, David C.↗
NASA NTRS1981

Aeroelastic instability stoppers for wind tunnel models

A mechanism for diverting the flow in a wind tunnel from the wing of a tested model is described. The wing is mounted on the wall of a tunnel. A diverter plate is pivotally mounted on the tunnel wall ahead of the model. An actuator fixed to the tunnel is pivotably connected to the diverter plate, by plunger. When the model is about to become unstable during the test the actuator moves the diverter plate from the tunnel wall to divert maintaining stable model conditions. The diverter plate is then retracted to enable normal flow.

Doggett, R. V., Jr.↗
NASA NTRS1994

Experimental Investigation of Jet-Induced Mixing of a Large Liquid Hydrogen Storage Tank

Experiments have been conducted to investigate the effect of fluid mixing on the depressurization of a large liquid hydrogen storage tank. The test tank is approximately ellipsoidal, having a volume of 4.89 m(exp 3) and an average wall heat flux of 4.2 W/m(exp 2) due to external heat input. A mixer unit was installed near the bottom of the tank to generate an upward directed axial jet flow normal to the liquid-vapor interface. Mixing tests were initiated after achieving thermally stratified conditions in the tank either by the introduction of hydrogen gas into the tank or by self-pressurization due to ambient heat leak through the tank wall. The subcooled liquid jet directed towards the liquid-vapor interface by the mixer induced vapor condensation and caused a reduction in tank pressure. Tests were conducted at two jet submergence depths for jet Reynolds numbers from 80,000 to 495,000 and Richardson numbers from 0.014 to 0.52. Results show that the rate of tank pressure change is controlled by the competing effects of subcooled jet flow and the free convection boundary layer flow due to external tank wall heating. It is shown that existing correlations for mixing time and vapor condensation rate based on small scale tanks may not be applicable to large scale liquid hydrogen systems.

Lin, C. S.↗
DOE OSTI2025

Reweighting configurations generated by transferable, machine learned models for protein sidechain backmapping

Multiscale modeling requires the linking of models at different levels of detail, with the goal of gaining accelerations from lower fidelity models while recovering fine details from higher resolution models. Communication across resolutions is particularly important in modeling soft matter, where tight couplings exist between molecular-level details and mesoscale structures. While multiscale modeling of biomolecules has become a critical component in exploring their structure and self-assembly, backmapping from coarse-grained to fine-grained, or atomistic, representations presents a challenge, despite recent advances through machine learning. A major hurdle, especially for strategies utilizing machine learning, is that backmappings can only approximately recover the atomistic ensemble of interest. We demonstrate conditions for which backmapped configurations may be reweighted to exactly recover the desired atomistic ensemble. By training separate decoding models for each sidechain type, we develop an algorithm based on normalizing flows and geometric algebra attention to autoregressively propose backmapped configurations for any protein sequence. Critical for reweighting with modern protein force fields, our trained models include all hydrogen atoms in the backmapping and make probabilities associated with atomistic configurations directly accessible. We also demonstrate, however, that reweighting is extremely challenging despite state-of-the-art performance on recently developed metrics and generation of configurations with low energies in atomistic protein force fields. Through detailed analysis of configurational weights, we show that machine-learned backmappings must not only generate configurations with reasonable energies, but also correctly assign relative probabilities under the generative model. These are broadly important considerations in generative modeling of atomistic molecular configurations.

Monroe, Jacob I. [Univ. of Arkansas, Fayetteville,↗
DOE OSTI2025

ForestFlow: predicting the Lyman-α forest clustering from linear to nonlinear scales

On large scales, the Lyman-α forest provides insights into the expansion history of the Universe, while on small scales, it imposes strict constraints on the growth history, the nature of dark matter, and the sum of neutrino masses. This work introduces ForestFlow, a novel framework that bridges the gap between large- and small-scale analyses, which have traditionally relied on distinct modeling approaches. Using conditional normalizing flows, ForestFlow predicts the two Lyman-α linear biases (b δ and b η ) and six parameters describing small-scale deviations of the three-dimensional flux power spectrum (P 3D ) from linear theory as a function of cosmology and intergalactic medium physics. These are then combined with a Boltzmann solver to make consistent predictions, from arbitrarily large scales down to the nonlinear regime, for P 3D and any other statistics derived from it. Trained on a suite of 30 fixed-and-paired cosmological hydrodynamical simulations spanning redshifts from z = 2 to 4.5, ForestFlow achieves 3 and 1.5% precision in describing P 3D and the one-dimensional flux power spectrum (P 1D ) from linear scales to k = 5 Mpc −1 and k ∥ = 4 Mpc −1 , respectively. Thanks to its conditional parameterization, ForestFlow shows similar performance for ionization histories and two ΛCDM model extensions – massive neutrinos and curvature – even though none of these are included in the training set. This framework will enable full-scale cosmological analyses of Lyman-α forest measurements from the DESI survey.

79 ASTRONOMY AND ASTROPHYSICS↗