Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Embedding”

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 343 records · Page 19

Understanding Generative AI Content with Embedding Models

The construction of high-quality numerical features is critical to any quantitative data analysis. Feature engineering has been historically addressed by carefully hand-crafting data representations based on domain expertise. This work views the internal representations of modern deep neural networks (DNNs), called embeddings, as an implicit form of traditional feature engineering. For trained DNNs, we show that these embeddings can reveal interpretable, high-level concepts in unstructured sample data. We use these embeddings in natural language and computer vision tasks to uncover both inherent heterogeneity in the underlying data and human-understandable explanations for it. In particular, we find empirical evidence that there is inherent separability between real data and those generated from AI models.

Vargas, Max↗

Data and Code for Understanding Generative AI Content with Embedding Models

This repository contains code for the experiments in the paper "Understanding Generative AI Content with Embedding Models". Constructing high-quality features is critical to any quantitative data analysis. While feature engineering was historically addressed by carefully hand-crafting data representations based on domain expertise, deep neural networks (DNNs) now offer a radically different approach. DNNs implicitly engineer features by transforming their input data into hidden feature vectors called embeddings. For embedding vectors produced by foundation models -- which are trained to be useful across many contexts -- we demonstrate that simple and well-studied dimensionality-reduction techniques such as Principal Component Analysis uncover inherent heterogeneity in input data concordant with human-understandable explanations. Of the many applications for this framework, we find empirical evidence that there is intrinsic separability between real samples and those generated by artificial intelligence (AI).

Vargas, Max [Pacific Northwest National Laboratory↗

Direct Ink-Write Printing of Ceramic Clay with an Embedded Wireless Temperature and Relative Humidity Sensor

This research presents a simple method to additively manufacture Cone 5 porcelain clay ceramics by using the direct ink-write (DIW) printing technique. DIW has allowed the application of extruding highly viscous ceramic materials with relatively high-quality and good mechanical properties, which additionally allows a freedom of design and the capability of manufacturing complex geometrical shapes. Clay particles were mixed with deionized (DI) water at different ratios, where the most suitable composition for 3D printing was observed at a 1:5 w/c ratio (16.2 wt.%. of DI water). Differential geometrical designs were printed to demonstrate the printing capabilities of the paste. In addition, a clay structure was fabricated with an embedded wireless temperature and relative humidity (RH) sensor during the 3D printing process. The embedded sensor read up to 65% RH and temperatures of up to 85 °F from a maximum distance of 141.7 m. The structural integrity of the selected 3D printed geometries was confirmed through the compressive strength of fired and non-fired clay samples, with strengths of 70 MPa and 90 MPa, respectively. This research demonstrates the feasibility of using the DIW printing of porcelain clay with embedded sensors, with fully functional temperature- and humidity-sensing capabilities.

42 ENGINEERING↗

Common Envelope Wind Tunnel: Coefficients of Drag and Accretion in a Simplified Context for Studying Flows around Objects Embedded within Stellar Envelopes

This paper examines the properties of flows around objects embedded within common envelopes in the simplified context of a “wind tunnel.” We establish characteristic relationships between key common envelope flow parameters like the Mach number and density scale height. Our wind tunnel is a three-dimensional, Cartesian geometry hydrodynamic simulation setup that includes the gravity of the primary and secondary stars and allows us to study the coefficients of drag and accretion experienced by the embedded object. Accretion and drag lead to a transformation of an embedded object and its orbit during a common envelope phase. We present two suites of simulations spanning a range of density gradients and Mach numbers—relevant for flow near the limb of a stellar envelope to the deep interior. In one suite, we adopt an ideal gas adiabatic exponent of γ=5/3, in the other, γ=4/3. We find that coefficients of drag rise in flows with steeper density gradients and that coefficients of drag and accretion are consistently higher in the more compressible, γ=4/3 flow. We illustrate the impact of these newly derived coefficients by integrating the inspiral of a secondary object through the envelopes of 3M{sub ⊙} (γ≈5/3) and 80M{sub ⊙} (γ≈4/3) giants. In these examples, we find a relatively rapid initial inspiral because, near the stellar limb, dynamical friction drag is generated mainly from dense gas focused from deeper within the primary-star’s envelope. This rapid initial inspiral timescale carries potential implications for the timescale of transients from early common envelope interaction.

79 ASTRONOMY AND ASTROPHYSICS↗

Enhanced thermal conductance of polymer composites through embedding aligned carbon nanofibers

The focus of this work is to find a more efficient method of enhancing the thermal conductance of polymer thin films. This work compares polymer thin films embedded with randomly oriented carbon nanotubes to those with vertically aligned carbon nanofibers. Thin films embedded with carbon nanofibers demonstrated a similar thermal conductance between 40–60 μm and a higher thermal conductance between 25–40 μm than films embedded with carbon nanotubes with similar volume fractions even though carbon nanotubes have a higher thermal conductivity than carbon nanofibers

36 MATERIALS SCIENCE↗

Normal and radial impact of composites with embedded penny-shaped cracks

A method is developed for the dynamic stress analysis of a layered composite containing an embedded penny-shaped crack and subjected to normal and radial impact. The material properties of the layers are chosen such that the crack lies in a layer of matrix material while the surrounding material possesses the average elastic properties of a two-phase medium consisting of a large number of fibers embedded in the matrix. Quantitatively, the time-dependent stresses near the crack border can be described by the dynamic stress intensity factors. Their magnitude depends on time, on the material properties of the composite and on the relative size of the crack compared to the composite local geometry. Results obtained show that, for the same material properties and geometry of the composite, the dynamic stress intensity factors for an embedded (penny-shaped) crack reach their peak values within a shorter period of time and with a lower magnitude than the corresponding dynamic stress intensity factors for a through-crack.

Sih, G. C.↗

A flexible grid embedding technique with application to the Euler equations

An automated grid embedding procedure for solution of flows about complex geometries is described. The physical domain is subdivided into regions that can accommodate easily generated grids. The grids are organized in a hierarchical structure, and communication among grids is accomplished by interpolation of the flow variables at mesh boundaries. Algorithms for locating embedded boundaries, special treatment of the embedded grids, and the data structures required for manipulating the solution data are described. The method is demonstrated by solution of the Euler equations for transonic flow about a supercritical airfoil and a flapped airfoil.

Benek, J. A.↗

An embedding method for the steady Euler equations

Certain difficulties arise in connection with the numerical solution of a direct finite difference representation of the steady Euler equations. Johnson (1979, 1981, 1982) has, therefore, proposed a surrogate-equation technique, in which the first-order steady Euler equations are embedded in a certain second-order system of equations. The present paper is concerned with the theoretical justification for such an embedding approach. For the numerical solution of the two-dimensional steady Euler equations, it is shown that, under a continuity restriction, it is possible to solve a second-order embedded system together with appropriate additional boundary conditions. The result indicates that a more direct and potentially more efficient approach to the steady solutions exists than the alternative of solving the unsteady equations.

Chang, S.-H.↗

EOS: A project to investigate the design and construction of real-time distributed Embedded Operating Systems

Project EOS is studying the problems of building adaptable real-time embedded operating systems for the scientific missions of NASA. Choices (A Class Hierarchical Open Interface for Custom Embedded Systems) is an operating system designed and built by Project EOS to address the following specific issues: the software architecture for adaptable embedded parallel operating systems, the achievement of high-performance and real-time operation, the simplification of interprocess communications, the isolation of operating system mechanisms from one another, and the separation of mechanisms from policy decisions. Choices is written in C++ and runs on a ten processor Encore Multimax. The system is intended for use in constructing specialized computer applications and research on advanced operating system features including fault tolerance and parallelism.

Campbell, R. H.↗

Implementing embedded artificial intelligence rules within algorithmic programming languages

Most integrations of artificial intelligence (AI) capabilities with non-AI (usually FORTRAN-based) application programs require the latter to execute separately to run as a subprogram or, at best, as a coroutine, of the AI system. In many cases, this organization is unacceptable; instead, the requirement is for an AI facility that runs in embedded mode; i.e., is called as subprogram by the application program. The design and implementation of a Prolog-based AI capability that can be invoked in embedded mode are described. The significance of this system is twofold: Provision of Prolog-based symbol-manipulation and deduction facilities makes a powerful symbolic reasoning mechanism available to applications programs written in non-AI languages. The power of the deductive and non-procedural descriptive capabilities of Prolog, which allow the user to describe the problem to be solved, rather than the solution, is to a large extent vitiated by the absence of the standard control structures provided by other languages. Embedding invocations of Prolog rule bases in programs written in non-AI languages makes it possible to put Prolog calls inside DO loops and similar control constructs. The resulting merger of non-AI and AI languages thus results in a symbiotic system in which the advantages of both programming systems are retained, and their deficiencies largely remedied.

Feyock, Stefan↗

Embedding triple-modular redundancy into a hypercube architecture

This paper describes an embedding of Triple Modular Redundancy (TMR) into a binary hypercube. The goal is to improve fault tolerance by masking any single-point faults. Each module of an application task is triplicated and executed in parallel on three nodes of a 2-dimensional subcube (Q2) of the hypercube. Each of these nodes also executes a voter process. The remaining node is used for message passing only. All outputs from the triplicated modules are voted on, and the voting results are transmitted to the appropriate destination. Thus, all interunit messages are also triplicated. We propose an embedding of TMR into a hypercube which can be implemented in a manner transparent to the application program. Subcubes are allocated so that the address space for the TMR units is also a hypercube. Hence, the subcube allocation and intermodule communication schemes are defined to be analogous to the schemes used in the nonredundant system. The embedded system is proven to mask all single-point faults.

Kiskis, Daniel L.↗

Run-time implementation issues for real-time embedded Ada

A motivating factor in the development of Ada as the department of defense standard language was the high cost of embedded system software development. It was with embedded system requirements in mind that many of the features of the language were incorporated. Yet it is the designers of embedded systems that seem to comprise the majority of the Ada community dissatisfied with the language. There are a variety of reasons for this dissatisfaction, but many seem to be related in some way to the Ada run-time support system. Some of the areas in which the inconsistencies were found to have the greatest impact on performance from the standpoint of real-time systems are presented. In particular, a large part of the duties of the tasking supervisor are subject to the design decisions of the implementer. These include scheduling, rendezvous, delay processing, and task activation and termination. Some of the more general issues presented include time and space efficiencies, generic expansions, memory management, pragmas, and tracing features. As validated compilers become available for bare computer targets, it is important for a designer to be aware that, at least for many real-time issues, all validated Ada compilers are not created equal.

Maule, Ruth A.↗

Strain and dynamic measurements using fiber optic sensors embedded into graphite/epoxy tubes

Graphite/epoxy tubes were fabricated with embedded optical fibers to evaluate the feasibility of monitoring strains with a fiber optic technique. Resistance strain gauges were attached to the tubes to measure strain at four locations along the tube for comparison with the fiber optic sensors. Both static and dynamic strain measurements were made with excellent agreement between the embedded fiber optic strain sensor and the strain gauges. Strain measurements of 10(exp -7) can be detected with the optical phase locked loop (OPLL) system using optical fiber. Because of their light weight, compatibility with composites, immunity to electromagnetic interference, and based on the static and dynamic results obtained, fiber optic sensors embedded in composites may be useful as the sensing component of smart structures.

Dehart, D. W.↗

Fiber optic strain measurements in filament-wound graphite-epoxy tubes containing embedded fibers

Filament-wound graphite-epoxy tubes fabricated with embedded fiber optic sensors were tested at NASA Langley Research Center to evaluate the feasibility of monitoring stress with a fiber optic technique. Resistance strain gauges were attached to the tubes to measure strain at four locations along the tubes. Both static and dynamic strain measurements were made with an excellent agreement between the embedded fiber optic strain sensor and the strain gauges. The results indicate that fiber optic sensors embedded in composites may be useful as the sensing component of smart structures.

Rogowski, R. S.↗

Embedded function methods for supersonic turbulent boundary layers

The development of embedded functions to represent the mean velocity and total enthalpy distributions in the wall layer of a supersonic turbulent boundary layer is considered. The asymptotic scaling laws (in the limit of large Reynolds number) for high speed compressible flows are obtained to facilitate eventual implementation of the embedded functions in a general prediction method. A self-consistent asymptotic structure is derived, as well as a compressible law of the wall in which the velocity and total enthalpy are logarithmic within the overlap zone, but in the Howarth-Dorodnitsyn variable. Simple outer region turbulence models are proposed (some of which are modifications of existing incompressible models) to reflect the effects of compressibility. As a test of the methodology and the new turbulence models, a set of self-similar outer region profiles is obtained for constant pressure flow; these are then coupled with embedded functions in the wall layer. The composite profiles thus obtained are compared directly with experimental data and good agreement is obtained for flows with Mach numbers up to 10.

He, J.↗

An embedded grid formulation applied to a delta wing

Applications using a three-dimensional embedded grid scheme are made to high angle of attack viscous flow over two bodies: a slender cone using the conical approximation and a 75 deg swept delta wing. The embedded grids are used principally to improve the numerical resolution of the separated vortical flow above the body. Detailed comparisons are made with experimental measurements of the velocity field over the delta wing. The prediction of the maximum steamwise velocity is improved using two levels of embedded grid refinement but is still less than the experimental measurements available from a laser velocimeter.

Krist, Sherrie L.↗

The structure and development of streamwise vortex arrays embedded in a turbulent boundary layer

An investigation of the structure and development of streamwise vortices embedded in a turbulent boundary layer was conducted. The vortices were generated by a single spanwise row of rectangular vortex generator blades. A single embedded vortex was examined, as well as arrays of embedded counter rotating vortices produced by equally spaced vortex generators. Measurements of the secondary velocity field in the crossplane provided the basis for characterization of vortex structure. Vortex structure was characterized by four descriptors. The center of each vortex core was located at the spanwise and normal position of peak streamwise vorticity. Vortex concentration was characterized by the magnitude of the peak streamwise vorticity, and the vortex strength by its circulation. Measurements of the secondary velocity field were conducted at two crossplane locations to examine the streamwise development of the vortex arrays. Large initial spacings of the vortex generators produced pairs of strong vortices which tended to move away from the wall region while smaller spacings produced tight arrays of weak vortices close to the wall. A model of vortex interaction and development is constructed using the experimental results. The model is based on the structure of the Oseen Vortex. Vortex trajectories are modelled by including the convective effects of neighbors.

Wendt, Bruce J.↗

The structure and development of streamwise vortex arrays embedded in a turbulent boundary layer

The results of an experimental investigation of the structure and development of streamwise vortices embedded in a turbulent boundary layer are presented. Measurements of secondary velocity in the crossplane are used to characterize the vortex array structure. Measurements in the crossplane at two streamwise locations characterize the influence of interactions among the vortices on the array structure when the initial spacing between vortices is varied. Evidence of the merging of counter-rotating cores is found in embedded arrays of closely spaced vortices. A model of vortex interaction and development is constructed from the experimental results. This model is based on the structure of the two dimensional Ossen vortex. The decay of vortex circulation due to the merging of the cores is correlated with the crossplane gradient in streamwise vorticity occurring between an embedded vortex and its adjacent counter-rotating neighbors.

Wendt, Bruce J.↗