Search NASA⌕ Search

SEARCH · Search NASA

Results for “embedding 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.

At least 307 records · Page 17
All sourcesNASA NTRSDOE OSTI
NASA NTRS2016

A Continuum Damage Mechanics Model to Predict Kink-Band Propagation Using Deformation Gradient Tensor Decomposition

A new model is proposed that represents the kinematics of kink-band formation and propagation within the framework of a mesoscale continuum damage mechanics (CDM) model. The model uses the recently proposed deformation gradient decomposition approach to represent a kink band as a displacement jump via a cohesive interface that is embedded in an elastic bulk material. The model is capable of representing the combination of matrix failure in the frame of a misaligned fiber and instability due to shear nonlinearity. In contrast to conventional linear or bilinear strain softening laws used in most mesoscale CDM models for longitudinal compression, the constitutive response of the proposed model includes features predicted by detailed micromechanical models. These features include: 1) the rotational kinematics of the kink band, 2) an instability when the peak load is reached, and 3) a nonzero plateau stress under large strains.

Bergan, Andrew C.↗
NASA NTRS2007

Numerical Simulations of Flow Separation Control in Low-Pressure Turbines using Plasma Actuators

A recently introduced phenomenological model to simulate flow control applications using plasma actuators has been further developed and improved in order to expand its use to complicated actuator geometries. The new modeling approach eliminates the requirement of an empirical charge density distribution shape by using the embedded electrode as a source for the charge density. The resulting model is validated against a flat plate experiment with quiescent environment. The modeling approach incorporates the effect of the plasma actuators on the external flow into Navier Stokes computations as a body force vector which is obtained as a product of the net charge density and the electric field. The model solves the Maxwell equation to obtain the electric field due to the applied AC voltage at the electrodes and an additional equation for the charge density distribution representing the plasma density. The new modeling approach solves the charge density equation in the computational domain assuming the embedded electrode as a source therefore automatically generating a charge density distribution on the surface exposed to the flow similar to that observed in the experiments without explicitly specifying an empirical distribution. The model is validated against a flat plate experiment with quiescent environment.

Suzen, Y. B.↗
NASA NTRS1993

Session on validation of coupled models

The session on validation of coupled models is reviewed. The current use of a mesoscale model with a grid size of 20-km during STORM-FEST in 1992 has proven to be extremely valuable. The availability of forecast products at a much higher temporal and spatial resolution was very helpful for mesoscale forecasting, mission planning, and the guidance of research aircraft. Recent numerical simulation of ocean cyclones and mesoscsle convective systems using nonhydrostatic cloud/mesoscale models with a grid size as small as 2-km have demonstrated the potential of these models for predicting mesoscale convective systems, squall lines, hurricane rainbands, mesoscale gravity waves, and mesoscale frontal structures embedded within an extratropical cyclone. Although mesoscale/cloud scale models have demonstrated strong potential for use in operational forecasting, very limited quantitative evaluation (and verification) of these models were performed. As a result, the accuracy, the systematic biases, and the useful forecasts limits were not properly defined for these models. Also, no serious attempts were made to use these models for operational prediction of mesoscale convective systems.

Kuo, Bill↗
NASA NTRS1994

Radiative acceleration in outflows from broad absorption line quasi-stellar objects. 2: Wind models

We investigate the dynamics of radiatively driven broad absorption-line (BAL) outflows in quasi-stellar objects (QSOs) by developing radial and time-independent numerical models. Two limits are explored. The first assumes that the absorbing matter is not forced to comove with the substrate, which provides pressure confinement. This assumption allows us to explore in detail a case in which the acceleration is entirely due to radiation pressure. Using the parameters inferred from observations, we find that under these conditions radiative acceleration (mainly due to resonance line scattering) can readily accelerate the flow to the observed velocities. An important feature of the noncoupled flow is that the line profiles tend to stay relatively flat throughout the velocity interval covered by the line. We discuss how relaxing the assumptions of radial symmetry and time independent may help to explain the structures observed in BALs. In the second class of models, the absorbing flow is assumed to be completely coupled to the substrate in which it is embedded. Aside from being more plausible physically, these models produce line profiles that trail off at higher velocities, a behavior observed in some BALs. We show that, even if the substrate is massless, we have to assume a starting radius very close to the inferred radius of the broad emission-line region (approximately 0.1 pc) in order to obtain a significant contribution from radiative acceleration, given a typical active galactic nucleus (AGN) spectrum. The reason is that the energy input needed to pressurize the substrate, allowing the flow to become supersonic and to retain a reasonable ionization equilibrium, at the same time contributes appreciably to the acceleration. A way to relax the small starting radius constraint is to use a softer ionizing spectrum.

Arav, Nahum↗
NASA NTRS2006

Software Tools for Developing and Simulating the NASA LaRC CMF Motion Base

The NASA Langley Research Center (LaRC) Cockpit Motion Facility (CMF) motion base has provided many design and analysis challenges. In the process of addressing these challenges, a comprehensive suite of software tools was developed. The software tools development began with a detailed MATLAB/Simulink model of the motion base which was used primarily for safety loads prediction, design of the closed loop compensator and development of the motion base safety systems1. A Simulink model of the digital control law, from which a portion of the embedded code is directly generated, was later added to this model to form a closed loop system model. Concurrently, software that runs on a PC was created to display and record motion base parameters. It includes a user interface for controlling time history displays, strip chart displays, data storage, and initializing of function generators used during motion base testing. Finally, a software tool was developed for kinematic analysis and prediction of mechanical clearances for the motion system. These tools work together in an integrated package to support normal operations of the motion base, simulate the end to end operation of the motion base system providing facilities for software-in-the-loop testing, mechanical geometry and sensor data visualizations, and function generator setup and evaluation.

Bryant, Richard B., Jr.↗
NASA NTRS2019

Computational Micromechanics Model for the Analysis of Fiber Kinking in Unidirectional FRP

A computational micromechanics (CMM) model is developed to analyze fiber kinking, which is a failure mechanism that takes place in fiber-reinforced composites when they are loaded under longitudinal compression. The CMM model consists of a single AS4 carbon fiber with an initial misalignment embedded in an 8552 polymer matrix. The deformation of the model is governed by periodic boundary conditions (PBC). The relatively simple CMM model enables the evaluation of the role played by initial misalignment of the fiber, shear yielding of the matrix and fiber-matrix debonding. A novel microscale experimental technique devoted to the characterization of the longitudinal compressive strength of the fibers, X(sup f)(sub c) , was developed. By exercising the model and comparing it with several existing models in the literature, the nonlinear shear response of the composite lamina was shown to play a fundamental role not only in the prediction of the compressive strength X(sub c), but also during the post-peak regime in terms of residual stress σ(sub r) and fiber rotation φ. Finally, the influence of the fiber-matrix interface damage (not considered in most other fiber kinking models) on the fiber kinking phenomenon was assessed through a parametric study.

Polymer-matrix composites (PMC)↗
DOE OSTI2024

Enhancing Network Anomaly Detection Using Graph Neural Networks

In the world of Internet of Things (IoT) networks, where devices are constantly communicating, keeping them secure from cyber threats is critical. This paper introduces a novel approach to detecting unusual and potentially harmful activities in these networks using graph neural networks (GNNs). We combine two specific types of GNNs-GraphSAGE and graph attention networks (GAT)-to create a model that understands and represents the behaviors and interactions in a network. GraphSAGE creates an embedding of network activities by examining local data interactions, while GAT directs the model's focus to the most critical interactions. By integrating these two methods in a single model that considers different types of interactions (both host and flow nodes), we aim to create a system that accurately represents the current state of a network and can also spot anomalies effectively while reducing false positives and negatives. Our innovative approach has demonstrated promising results, achieving an accuracy of 98% on the UNSW-NB15 dataset, significantly outperforming standalone GraphSAGE and GAT models. This underscores its potential as a robust framework for securing IoT networks against cyber threats and anomalies.

Marfo, William↗
DOE OSTI2024

Embedded symmetric positive semi-definite machine-learned elements for reduced-order modeling in finite-element simulations with application to threaded fasteners

Here, we present a machine-learning strategy for finite element analysis of solid mechanics wherein we replace complex portions of a computational domain with a data-driven surrogate. In the proposed strategy, we decompose a computational domain into an “outer” coarse-scale domain that we resolve using a finite element method (FEM) and an “inner” fine-scale domain. We then develop a machine-learned (ML) model for the impact of the inner domain on the outer domain. In essence, for solid mechanics, our machine-learned surrogate performs static condensation of the inner domain degrees of freedom. This is achieved by learning the map from displacements on the inner-outer domain interface boundary to forces contributed by the inner domain to the outer domain on the same interface boundary. We consider two such mappings, one that directly maps from displacements to forces without constraints, and one that maps from displacements to forces by virtue of learning a symmetric positive semi-definite (SPSD) stiffness matrix. We demonstrate, in a simplified setting, that learning an SPSD stiffness matrix results in a coarse-scale problem that is well-posed with a unique solution. We present numerical experiments on several exemplars, ranging from finite deformations of a cube to finite deformations with contact of a fastener-bushing geometry. We demonstrate that enforcing an SPSD stiffness matrix drastically improves the robustness and accuracy of FEM–ML coupled simulations, and that the resulting methods can accurately characterize out-of-sample loading configurations with significant speedups over the standard FEM simulations.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1988

Mesoscale spiral vortex embedded within a Lake Michigan snow squall band - High resolution satellite observations and numerical model simulations

It is known that Great Lakes snow squall convection occurs in a variety of different modes depending on various factors such as air-water temperature contrast, boundary-layer wind shear, and geostrophic wind direction. An exceptional and often neglected source of data for mesoscale cloud studies is the ultrahigh resolution multispectral data produced by Landsat satellites. On October 19, 1972, a clearly defined spiral vortex was noted in a Landsat-1 image near the southern end of Lake Michigan during an exceptionally early cold air outbreak over a still very warm lake. In a numerical simulation using a three-dimensional Eulerian hydrostatic primitive equation mesoscale model with an initially uniform wind field, a definite analog to the observed vortex was generated. This suggests that intense surface heating can be a principal cause in the development of a low-level mesoscale vortex.

Lyons, Walter A.↗
NASA NTRS1993

Unsteady evaporation and combustion of a drop cluster inside a vortex

A model has been developed which describes the evaporation, ignition and combustion of a drop cluster embedded in a large vortex. The purpose of this model is to simulate the behavior of drops in large coherent vortices produced in the shear layer of a jet. The model treats the dynamic interactions between the drops and the vortex, and also takes into account the drop proximity to calculate the heat and mass transfers between drops and ambient gas. The gas phase outside the cluster is treated as an unsteady, reacting phase, whereas quasi-steadiness is assumed between the drops and surrounding gas inside the cluster. The results show a very complex interaction between the dynamics of the drop-loaded vortex, the flame, and the evaporation process.

Fichot, F.↗
DOE OSTI2025

CFD modeling and simulation for corrosive wear of refractory in molten slag

The development of high-wear resistant refractories having minimal production costs is facilitated by characterizing the wear mechanisms associated with their corrosive wear. Static cup testing is a commonly used method for comparing the corrosion resistance performance of two or more refractory materials. Although the static cup test conditions are not as severe as dynamic tests, this study shows that the thermal gradient present within the system during heating and cooling stages serves to generate movement of the slag leading to mechanical wear. The thermal gradient within the refractory, and between the slag and the refractory, occurs during the ramp stage of the test and lasts until the soaking stage is reached bringing the system to a thermal equilibrium. Using computational fluid dynamics (CFD) capabilities embedded within ANSYS software, this study modelled and quantified the convection currents within the slag and associated shear stresses generated on the refractory walls due to the thermal gradient. A traditional ladle furnace was employed as a case study to verify the results of the studied CFD model. The corrosion rate of the refractory lining was found to depend on the mass transfer coefficient of the refractory dissolution into the slag, and a velocity term which governs the extent of corrosion at any given location. This velocity term is a function of slag viscosity, as well as the concentration gradient and/or temperature gradient at the triple points. In this study, wall shear stress was used as a reliable proxy for identifying high-velocity regions prone to excessive corrosive wear. Elevated wall shear stress near the slag/air and slag/molten steel interfaces align with observed corrosion grooves, which reflects the intensified corrosive wear at these locations.

Ramteke, Rajat Rajat Durgesh Ramteke [University o↗
NASA NTRS2007

Evaluation of a Cloud Resolving Model Using TRMM Observations for Multiscale Modeling Applications

The climate change simulation community is moving toward use of global cloud resolving models (CRMs), however, current computational resources are not sufficient to run global CRMs over the hundreds of years necessary to produce climate change estimates. As an intermediate step between conventional general circulation models (GCMs) and global CRMs, many climate analysis centers are embedding a CRM in each grid cell of a conventional GCM. These Multiscale Modeling Frameworks (MMFs) represent a theoretical advance over the use of conventional GCM cloud and convection parameterizations, but have been shown to exhibit an overproduction of precipitation in the tropics during the northern hemisphere summer. In this study, simulations of clouds, precipitation, and radiation over the South China Sea using the CRM component of the NASA Goddard MMF are evaluated using retrievals derived from the instruments aboard the Tropical Rainfall Measuring Mission (TRMM) satellite platform for a 46-day time period that spans 5 May - 20 June 1998. The NASA Goddard Cumulus Ensemble (GCE) model is forced with observed largescale forcing derived from soundings taken during the intensive observing period of the South China Sea Monsoon Experiment. It is found that the GCE configuration used in the NASA Goddard MMF responds too vigorously to the imposed large-scale forcing, accumulating too much moisture and producing too much cloud cover during convective phases, and overdrying the atmosphere and suppressing clouds during monsoon break periods. Sensitivity experiments reveal that changes to ice cloud microphysical parameters have a relatively large effect on simulated clouds, precipitation, and radiation, while changes to grid spacing and domain length have little effect on simulation results. The results motivate a more detailed and quantitative exploration of the sources and magnitude of the uncertainty associated with specified cloud microphysical parameters in the CRM components of MMFs.

Posselt, Derek J.↗
NASA NTRS2014

Model-Based Systems Engineering With the Architecture Analysis and Design Language (AADL) Applied to NASA Mission Operations

The potential of Model Model Systems Engineering (MBSE) using the Architecture Analysis and Design Language (AADL) applied to space systems will be described. AADL modeling is applicable to real-time embedded systems- the types of systems NASA builds. A case study with the Juno mission to Jupiter showcases how this work would enable future missions to benefit from using these models throughout their life cycle from design to flight operations.

MBSE↗
NASA NTRS2009

Embedded passives emulator FY2003 interim report

The National Aeronautics and Space Administration Electronic Parts and Packaging (NEPP) Program has been evaluating state-of-the-art Embedded Passive Technology for the printed circuit board (PCB) industry for the past several years and has conducted its own testing on embedded passive coupons (resistors and capacitors). This document is a progress report of the work performed through fiscal year 2008 by NASA and the Navy to convert an existing PWB (printed wiring board) design to a design that incorporates embedded passive components. The work is being performed through a memorandum of understanding (MOU) between the NEPP Program (NASA) and the Naval Surface Warfare Center (NSWC) Crane Emerging Critical Interconnect Technology (ECIT) Program. NSWC Crane’s ECIT Program has funding to convert two PCB designs from the military and aerospace industry to a PCB design with embedded passives. ECIT will provide a design team that will take a well characterized design and convert it to an embedded design. They will also provide electrical modeling and layout expertise as well as build the PWB in their facility in Indiana. NEPP, through the Jet Propulsion Laboratory (JPL), will assemble any passives deemed not suitable for embedding and the active components and electrically test the new design to the existing (non-embedded) design

Gerke, David↗
DOE OSTI2024

Development of message passing-based graph convolutional networks for classifying cancer pathology reports

Abstract Background Applying graph convolutional networks (GCN) to the classification of free-form natural language texts leveraged by graph-of-words features (TextGCN) was studied and confirmed to be an effective means of describing complex natural language texts. However, the text classification models based on the TextGCN possess weaknesses in terms of memory consumption and model dissemination and distribution. In this paper, we present a fast message passing network (FastMPN), implementing a GCN with message passing architecture that provides versatility and flexibility by allowing trainable node embedding and edge weights, helping the GCN model find the better solution. We applied the FastMPN model to the task of clinical information extraction from cancer pathology reports, extracting the following six properties: main site, subsite, laterality, histology, behavior, and grade. Results We evaluated the clinical task performance of the FastMPN models in terms of micro- and macro-averaged F1 scores. A comparison was performed with the multi-task convolutional neural network (MT-CNN) model. Results show that the FastMPN model is equivalent to or better than the MT-CNN. Conclusions Our implementation revealed that our FastMPN model, which is based on the PyTorch platform, can train a large corpus (667,290 training samples) with 202,373 unique words in less than 3 minutes per epoch using one NVIDIA V100 hardware accelerator. Our experiments demonstrated that using this implementation, the clinical task performance scores of information extraction related to tumors from cancer pathology reports were highly competitive.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Hierarchical-embedding autoencoder with a predictor as efficient architecture for learning time-evolution in multi-scale turbulent flows

We introduce a scale-aware, data-driven deep learning modeling framework for accurately predicting the time evolution of multi-scale turbulent plasma and liquid flows. The approach is motivated by the idea of scale separation. Structures of vastly different length scales emerge in these systems, and interactions between these structures occur only locally. To exploit this structure, the flow state is transformed by a hierarchical, fully convolutional autoencoder, not into a single embedding layer as in conventional convolutional surrogate models, but into a series of embedding layers. A stepwise training strategy ensures that fine-scale features are encoded on a high-resolution grid, while larger structures are represented on progressively coarser layers. The time evolution predictor advances all embedding layers in sync, capturing local interactions between features at the same scale as well as between all scales. This approach enables efficient modeling of multi-scale systems since negligible interactions between distant, small-scale structures do not need to be directly modeled. Our hierarchical-embedding autoencoder with a predictor framework is evaluated on canonical examples of multi-scale turbulence: two-dimensional Kolmogorov flow and Hasegawa–Wakatani plasma turbulence. In both cases, the proposed framework significantly improves predictive accuracy relative to conventional convolutional network architectures. A significant improvement in prediction accuracy was observed for crucial statistical characteristics of the Hasegawa–Wakatani plasma as well as for individual trajectories of the Kolmogorov flow turbulence. Importantly, the model's rollout for the Hasegawa–Wakatani problem demonstrates a four-order-of-magnitude speedup compared to traditional numerical solvers.

Khrabry, Alexander I. [Princeton Univ., NJ (United↗
NASA NTRS1997

X-Ray Gas Temperatures in the Arc Clusters MS0440+204 and MS0302+1658

The cluster of galaxies MS0440+02, originally discovered through its X-ray emission, was part of an optical observational program to search for arcs and arclets in a complete sample of X-ray luminous, medium-distant clusters of galaxies. Mauna Kea CCD images of MS0440+02 showed a remarkable optical morphology. The core of the cluster contains 6 bright galaxies and numerous fainter ones embedded in a low surface brightness halo. Besides, MS0440+02 is the most spectacular example that we have found of an arc system in a compact condensed cluster, with arcs symmetrically distributed to draw almost perfect circles around the cluster center. Giant arcs are magnified images of distant galaxies, gravitationally distorted by massive foreground clusters. It is of great importance to compare the results of the lensing studies with those derived from X-ray observations, as the two are independent methods of studying the mass distribution. Thus MS0440+02 was the ideal target to obtain temperature measurement with ASCA and good spatial resolution X-ray observations with ROSAT. The X-ray data have been used in conjunction with Hubble Space Telescope observations to put more stringent constrains on the mass estimates. Most of the different wavelength datasets have been reduced and analyzed. Mass determinations have been separately obtained from galaxy virial motions and X-ray profile fitting using the cluster gas temperature as measured by the ASCA satellite. Assuming that the hot gas is in hydrostatic equilibrium and in a spherical potential, we find from the X-ray data a mass distribution profile that is well described by a Beta model. From the multiple images formed by gravitational lensing (HST data) using the modelling of the gravitational lensed arcs, we have derived Beta model. To reconcile the mass estimates we have explored the possibility of having a supercluster surrounding the MOS0440 cluster, that is a model with two isothermal spheres, one embedded inside the other. These results have been published or are in press.

Gioia, Isabella M.↗