Search NASA⌕ Search

SEARCH · Search NASA

Results for “edge computer”

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 811 records · Page 45

Computational study of the aerodynamics and control by blowing of asymmetric vortical flows over delta wings

Some of the work is described which was done in a study of the flow field produced by tangential leading edge blowing on a 60 deg. delta wing. The flow is studied computationally by solving the Thin Layer Navier-Stokes equations. Steady state flow fields are calculated for various angles of attack and yaw, with and without the presence of tangential leading edge blowing. The effectiveness of blowing as a rolling moment control mechanism to extend the envelope of controllability is illustrated at pre- and post-stall angles of attack. The numerical grid is generated using algebraic grid generation and various interpolation and blending techniques. The jet emanates from a slot with linearly varying thickness and is introduced into the flow field using the concept of an actuator plane, thereby not requiring resolution of the jet slot geometry. The Baldwin-Lomax algebraic turbulence model is used to provide turbulent closure. The computational results are compared with those of experiments.

Craig, Ken↗

Influence matrix program for aerodynamic lifting surface theory

A users manual is described for a USA FORTRAN 4 computer program which computes an aerodynamic influence matrix and is one of several computer programs used to analyze lifting, thin wings in steady, subsonic flow according to a kernel function method lifting surface theory. The most significant features of the program are that it can treat unsymmetrical wings, control points can be placed on the leading and/or trailing edges, and a stable, efficient algorithm is used to compute the influence matrix.

Medan, R. T.↗

Accelerating computing for the future electric grid (CRADA Final Report)

As a participant in the Cyclotron Road Lab-Embedded Entrepreneurship Program (LEEP), Vellex Computing, Inc. has successfully validated the "Vellex Computing Stack," a breakthrough Analog Neural Computer (ANC) specifically designed for high-performance edge optimization. This project achieved critical milestones in mixed-signal circuit stability and software-hardware co-design, directly addressing national priorities in semiconductor resiliency. The success of this work is deeply rooted in the support from the Cyclotron Road LEEP, which provided the essential "hard tech" runway—funding, mentorship, and access to Lawrence Berkeley National Laboratory’s world-class characterization facilities—allowing Vellex to overcome the "Valley of Death" often faced by deep-tech hardware startups. By leveraging LBNL’s advanced testing infrastructure, Vellex was able to rigorously benchmark the ANC architecture against state-of-the-art digital solutions, a feat that would have been resource-prohibitive independently. This collaboration has not only advanced American leadership in analog computing but has also matured Vellex’s technology to a stage ripe for private sector commercialization.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Numerical solutions of the Navier-Stokes equations for the supersonic laminar flow over a two-dimensional compression corner

Numerical solutions have been obtained for the supersonic, laminar flow over a two-dimensional compression corner. These solutions were obtained as steady-state solutions to the unsteady Navier-Stokes equations using the finite difference method of Brailovskaya, which has second-order accuracy in the spatial coordinates. Good agreement was obtained between the computed results and wall pressure distributions measured experimentally for Mach numbers of 4 and 6.06, and respective Reynolds numbers, based on free-stream conditions and the distance from the leading edge to the corner. In those calculations, as well as in others, sufficient resolution was obtained to show the streamline pattern in the separation bubble. Upstream boundary conditions to the compression corner flow were provided by numerically solving the unsteady Navier-Stokes equations for the flat plate flow field, beginning at the leading edge. The compression corner flow field was enclosed by a computational boundary with the unknown boundary conditions supplied by extrapolation from internally computed points.

Carter, J. E.↗

Simulation of Acoustic Scattering from a Trailing Edge

Three model problems were examined to assess the difficulties involved in using a hybrid scheme coupling flow computation with the the Ffowcs Williams and Hawkings equation to predict noise generated by vortices passing over a sharp edge. The results indicate that the Ffowcs Williams and Hawkings equation correctly propagates the acoustic signals when provided with accurate flow information on the integration surface. The most difficult of the model problems investigated inviscid flow over a two-dimensional thin NACA airfoil with a blunt-body vortex generator positioned at 98 percent chord. Vortices rolled up downstream of the blunt body. The shed vortices possessed similarities to large coherent eddies in boundary layers. They interacted and occasionally paired as they convected past the sharp trailing edge of the airfoil. The calculations showed acoustic waves emanating from the airfoil trailing edge. Acoustic directivity and Mach number scaling are shown.

Singer, Bart A.↗

DataSpaces: an interaction and coordination framework for coupled simulation workflows

Emerging high-performance distributed computing environments are enabling new end-to-end formulations in science and engineering that involve multiple interacting processes and data-intensive application workflows. For example, current fusion simulation efforts are exploring coupled models and codes that simultaneously simulate separate application processes, such as the core and the edge turbulence. These components run on different high performance computing resources, need to interact at runtime with each other and with services for data monitoring, data analysis and visualization, and data archiving. As a result, they require efficient and scalable support for dynamic and flexible couplings and interactions, which remains a challenge. This paper presents DataSpaces a flexible interaction and coordination substrate that addresses this challenge. DataSpaces essentially implements a semantically specialized virtual shared space abstraction that can be associatively accessed by all components and services in the application workflow. It enables live data to be extracted from running simulation components, indexes this data online, and then allows it to be monitored, queried and accessed by other components and services via the space using semantically meaningful operators. The underlying data transport is asynchronous, low-overhead and largely memory-to-memory. The design, implementation, and experimental evaluation of DataSpaces using a coupled fusion simulation workflow is presented.

Computer Science↗

Prediction of vortical flows on wings using incompressible Navier-Stokes equations

Grid-density and Reynolds-number effects on integral values and static pressure distributions are presently studied in the steady-state Navier-Stokes solutions obtained for the flow around a round-edged double delta wing using an upwind-relaxation finite-difference algorithm. It is found that while the computed longitudinal aerodynamic coefficients are in good agreement with extant experimental data, the magnitudes of suction-pressure peaks are underpredicted in the vicinity of the trailing edge. It is judged that additional computations employing finer-grid solutions in the vortical flow region are required.

Hsu, C.-H.↗

Influence of plasma pedestal profiles on access to ELM-free regimes in ITER

The influence of current density and pressure gradient profiles in the pedestal on the access to the regimes free from edge localized modes (ELMs) like quiescent H-mode in ITER is investigated. Using the simulator of MHD modes localized near plasma boundary based on the KINX code, calculations of the ELM stability were performed for the ITER plasma in scenarios 2 and 4 under variations of density and temperature profiles with the self-consistent bootstrap current in the pedestal. Low pressure gradient values at the separatrix, the same position of the density and temperature pedestals and high poloidal beta values facilitate reaching high current density in the pedestal and a potential transition into the regime with saturated large scale kink modes. New version of the localized MHD mode simulator allows one to compute the growth rates of ideal peeling-ballooning modes with different toroidal mode numbers and to determine the stability region taking into account diamagnetic stabilization. The edge stability diagrams computations and sensitivity studies of the stability limits to the value of diamagnetic frequency show that diamagnetic stabilization of the modes with high toroidal mode numbers can help to access the quiescent H-mode even with high plasma density but only with low pressure gradient values at the separatrix. The limiting pressure at the top of the pedestal increases for higher plasma density. With flat density profile the access to the quiescent H-mode is closed even with diamagnetic stabilization taken into account, while toroidal mode numbers of the most unstable peeling-ballooning mode decrease from n = 10−40 to n = 3−20.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Multiblock Navier-Stokes solutions about the F/A-18 wing-LEX-fuselage configuration

Three-dimensional thin-layer Navier-Stokes computations are presented for the F/A-18 configuration. The modeled configuration includes an accurate surface representation of the fuselage, leading-edge extension (LEX), and wing, both with and without leading-edge flap deflection. A multiblock structured grid strategy is employed to decompose the computational flowfield domain around the subject configuration. Steady-state solutions are obtained from an algorithm that solves the compressible Navier-Stokes equations with an upwind-biased, flux-difference splitting approach. The results presented are based on a fully turbulent flow assumption, simulating the high Reynolds number flow conditions that correspond to a recent FI A-18 flight experiment. Good agreements between the computations and the flight test results are obtained for both surface flow patterns as well as surface pressure distributions. Furthermore, a correlation between the computed LEX vortex-core and the flight test results, observed by way of smoke visualization, is also presented.

Aircraft configurations↗

Microstructural characterization of in situ MXCT images of high density foams under large strains

Foams are used in numerous applications, such as vibration damping and energy mitigation (e.g., packaging and helmets), wherein their mechanical properties are of critical importance. A typical compressive response of a high density elastomeric foam, shown in Fig 1, generally contains three regions of interest: (I) a linear-elastic region, governed by strut bending; (II) a relatively flat, or slowly increasing stress-strain response, accompanied by strut buckling and the localized collapse of pores; and (III) an exponentially increasing stress-strain curve wherein the collapse of the pore matrix leads to densification. Two material properties of interest, upon which considerable research has focused are the foam's Young's modulus, E{sub f}, defined as the slope of the stress-strain response in region I, and the collapse stress, {sigma}{sub f}, defined as the stress at which the response begins to deviate from linearity. It has been observed [1, 2, 3] that Young's modulus and the collapse stress are dependent on the material properties of the strut material and the non-dimensional relative-density of the foam, {rho}* = {rho}{sub f}/{rho}{sub m}, where {rho}{sub r} is the gross density of the foam and {rho}{sub m} is the density of the strut, or matrix, material. For foam of low relative-density, i.e, {rho}* < 0.1, the collapse stress and Young's modulus obey the relations {sigma}{sub c}/E{sub m} {proportional_to} ({rho}*){sup m} and E{sub f}/E{sub m} {proportional_to} ({rho}*){sup n} where E{sub m} is Young's modulus of the strut material and the bounds on the parameters m and n are 0.05 {le} m {le} 0.2 and 1 {le} n {le} 4 [4]. For open-celled foams, n = 2, whereas for closed-celled foams, n = 3. Theoretically, n = 1 for foams with an ''ideal strut'' configuration [6]. Foams of high relative-density ({rho}* > 0.1) require correcting terms to account for the axial contributions of the ''thick'' struts [5]. The above equations relate important foam properties to the relative-density of the foam; however, there exists a gap in the understanding of how the foam microstructure affects the mechanical response of the foam. This is due in large part to the difficulty of characterizing foam structures in 3D, especially foams of high relative-density. Most elastomeric foams are manufactured by the introduction of a gas into a cross-linking polymer. The developing foam microstructure has a complex dependence on the polymer viscosity and rate of polymerization, resulting in a randomly arranged pore structure with a large distribution of pore sizes. One approach is to characterize foam microstructures solely in terms of the cross-sectional shape and vector arrangement of the strut matrix, since it is this matrix that supports the stresses upon loading of the foam; yet as the density of a foam is increased, the very definition of what constitutes a strut brakes down. Another, perhaps easier to visualize, characterization of foam microstructure can come from a description of the pore shape and arrangement. Given the random nature of the microstructures of blown foam, both approaches are useful and valid. This paper describes our work aimed at linking the mechanical response and microstructural evolution of high relative-density foam as it undergoes large deformation. This work consists of several inter-related parts, including (i) measuring the compressive stress-strain response, as illustrated in Fig. 1, (ii) obtaining in situ micro X-ray computed tomography (MXCT) images of high relative-density foams undergoing large strains, and (iii) developing mathematical, computer aided, methodologies to perform image analysis and calculations of parameters that characterize the pores and struts. By using MXCT, a non-invasive technique for imaging the internal structure of materials, we are able to observe, internally, individual struts and pores as they undergo large deformation. Here we describe our computer aided image analysis methodologies and present examples of their application to the MXCT images of foams. Operations of the computer program include noise removal (smoothing), thresholding [7], edge detection [8], distance transformation [9, 10], pore detection via a Hough transformation [11], and skeletonization [12]. Strut parameters, such as the strut cross-sectional area and strut arrangement, are calculated from a skeleton-based model of the MXCT images. The Hough transform is a feature detection method used primarily in image analysis applications. We use this transform to detect and calculate pore parameters, such as pore arrangement and pore size distribution. The experiments and calculations are performed in order to develop data for the formulation of theory that links foam mechanical behavior specifically to the microstructure of high relative-density foams, not just simply to the relative-density of the foam.

36 MATERIALS SCIENCE↗

Bone mineral computation with a rectilinear scanner

A portable rectilinear transmission scanner and associated computerized data reduction techniques for estimating bone mineral content are described. This unit can be easily disassembled for transport to various measurement sites and has been used to estimate the bone mineral content of the os calcis, radius, and ulna in the Apollo and Skylab astronauts. The scanner is used to obtain multiple rows of data from which a bone profile is derived. Bone edges are determined with the aid of a digital computer program which employs an algorithm that determines the greatest rate of change of the counting rate.

Ullman, J.↗

An edge-based solution-adaptive method applied to the AIRPLANE code

Computational methods to solve large-scale realistic problems in fluid flow can be made more efficient and cost effective by using them in conjunction with dynamic mesh adaption procedures that perform simultaneous coarsening and refinement to capture flow features of interest. This work couples the tetrahedral mesh adaption scheme, 3D_TAG, with the AIRPLANE code to solve complete aircraft configuration problems in transonic and supersonic flow regimes. Results indicate that the near-field sonic boom pressure signature of a cone-cylinder is improved, the oblique and normal shocks are better resolved on a transonic wing, and the bow shock ahead of an unstarted inlet is better defined.

Biswas, Rupak↗

Theoretical and Experimental Unsteady Aerodynamics Compared for a Linear Oscillating Cascade With a Supersonic Leading-Edge Locus

Experimental data were obtained to help validate analytical and computational fluid dynamics (CFD) codes used to compute unsteady cascade aerodynamics in a supersonicaxial- flow regime. Results from two analytical codes and one CFD code were compared with experimental data. One analytical code did not account for airfoil thickness or camber; another, using piston theory (piston code), accounted for thickness and camber upstream of the first shockwave/airfoil impingement locations. The Euler CFD code accounted fully for airfoil shape.

Ramsey, John K.↗

Effect of Reaction Kinetics on Graphene-Edge Morphology and Composition

New simulations of graphene growth in flame environments are presented. The simulations employed a kinetic Monte Carlo (KMC) algorithm coupled to molecular mechanics (MM) geometry optimization to track individual graphenic species as they evolve. Focus was given to incorporation of five-member rings and resulting curvature and edge defects. The model code was re-written to be more computationally efficient enabling a larger set of simulations to be run, decreasing stochastic fluctuations in the averaged results. The model also included updated rate coefficients for graphene edge reactions recently published in the literature. Sensitivity analysis, enabled by the more efficient coding, confirmed the importance of the updated rates as well as providing further insight on key reactions controlling layer growth and morphology. The new KMC code was applied to simulate graphene-layer growth in the environments of laminar premixed flames. Finally, in these flame simulations, C-H ratios of evolving structures were computed and compared to those from experiment and from an alternate model.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

The spiral-compact galaxy pair AM 2208-251: Computer simulations versus observations

The system AM2208-251 is a roughly edge-on spiral extending east-west with a smaller round compact E system about 60 arcsec east of the spiral nucleus along the major axis of the spiral. Bertola, Huchtmeier, and Zeilinger (1990) have presented optical spectroscopic as well as single dish 21 cm observations of this system. Their spectroscopic data show, via emission lines lambda lambda 3727-29A, a rising rotation curve near the nucleus. These spectroscopic observations may indicate a tidal interaction in the system. In order to learn more about such pairs, the authors simulated the interaction using the computer model developed by Miller (1976 a,b, 1978) and modified by the authors (Byrd 1986, 1987, 1988). To do the simulation they need an idea of the mutual orbits of the two galaxies. Their computer model is a two-dimensional polar N-body program. It consists of a self-gravitating disk of particles, within an inert axially symmetric stabilizing halo potential. The particles are distributed in a 24(radial) by 36(azimuthal) polar grid. Self consistent calculations can be done only within the grid area. The disk is modeled with a finite Mestel disk, where all the particles initially move in circular orbits with constant tangential velocities (Mestel 1963), resulting in a flat rotation curve. The gas particles in the spiral's disk, which make up 30 percent of its mass, collide in the following manner. The number of particles in each bin of the polar grid is counted every time step. If it is greater than a given critical density, all the particles in the bin collide, obtaining in the result the same velocities, equal to the average for the bin. This process produces clumps of gas particles-the star formation sites. The authors suppress the collision in the inner part of the disk (within the circle r = 6) to represent the hole seen in the gas in the nuclear bulge of spirals. They thus avoid spurious effects due to collisions in that region. They also varied the size of the collisional bins, which did not affect their conclusions.

Klaric, Mario↗

Baseline Validation of Unstructured Grid Reynolds-Averaged Navier-Stokes Toward Flow Control

The value of the use of the Reynolds-averaged Navier-Stokes methodology for active flow control applications is assessed. An experimental flow control database exists for a NACA0015 airfoil modified at the leading edge to implement a fluidic actuator; hence, this configuration is used. Computational results are documented for the baseline wing configuration (no control) with the experimental results and assumes two-dimensional flow. The baseline wing configuration has discontinuities at the leading edge, trailing edge, and aft of midchord on the upper surface. A limited number of active flow control applications have been tested in the laboratory and in flight. These applications include dynamic stall control using a deformable leading edge, separation control for takeoff and landing flight conditions using piezoelectric devices, pulsed vortex generators, zero-net-mass oscillations, and thrust vectoring with zero-net-mass piezoelectric-driven oscillatory actuation. As yet, there is no definitive comparison with experimental data that indicates current computational capabilities can quantitatively predict the large aerodynamic performance gains achieved with active flow control in the laboratory. However, one study using the Reynolds-averaged Navier-Stokes (RANS) methodology has shown good quantitative agreement with experimental results for an isolated zero-net-mass actuator. In addition, some recent studies have used RANS to demonstrate qualitative performance gains compared with the experimental data for separation control on an airfoil. Those quantitative comparisons for both baseline and flow control cases indicated that computational results were in poor quantitative agreement with the experiments. The current research thrust will investigate the potential use of an unstructured grid RANS approach to predict aerodynamic performance for active flow control applications building on the early studies. First the computational results must quantitatively match experiments for the no-control case before proceeding to the time-dependent flow control case. This paper documents the baseline (no-control) case using an unswept airfoil configuration. The next section describes the configurations used for the computations and the experimentals. The computational approach is then described followed by results and concluding remarks.

Joslin, Ronald D.↗

Computation and parallel implementation for early vision

The problem of early vision is to transform one or more retinal illuminance images-pixel arrays-to image representations built out of such primitive visual features such as edges, regions, disparities, and clusters. These transformed representations form the input to later vision stages that perform higher level vision tasks including matching and recognition. Researchers developed algorithms for: (1) edge finding in the scale space formulation; (2) correlation methods for computing matches between pairs of images; and (3) clustering of data by neural networks. These algorithms are formulated for parallel implementation of SIMD machines, such as the Massively Parallel Processor, a 128 x 128 array processor with 1024 bits of local memory per processor. For some cases, researchers can show speedups of three orders of magnitude over serial implementations.

Gualtieri, J. Anthony↗

PCS: A Productive Computational Science Platform

As modern supercomputers continue to be increasingly heterogeneous with diverse computational accelerators, software becomes a critical aspect necessary to fully exploit these accelerators while not increasing the effort and time required to make variable scientific discovery in the face of complicated programming environments introduced by these accelerators. This paper presents PCS, a productive computational science platform for cluster-scale heterogeneous computing. The PCS platform by design follows a basic principle that, multiple programming models in a unified programming environment, is essential to facilitate large-scale, accelerator-aware, heterogeneous computing for next-generation scientific applications. We propose a new scheme — Edge Neural Network — for managing regularity in large dataflow graphs using higher-order bulk synchronous operators, providing a foundation for formalizing computational models for cluster-wide dataflow programming of heterogeneous clusters.

Ojika, David↗