SEARCH · Search NASA

Results for “Acceleration methods”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 361 records · Page 20

All sources NASA NTRS DOE OSTI

NASA NTRS1985

Accelerated solution of the steady Euler equations

The present paper is concerned with two methods for the accelerated solution of the steady Euler equations. One method makes use of a second-order embedding to facilitate the derivation of the relaxation solution of the steady equations of motion, while the other method employs a multile-gridding concept to accelerate the convergence of a simple, explicit, time-marching scheme applied to the unsteady equations. It is pointed out that the surrogate equation technique provides a means for formulating problems involving the full steady Euler equations in such a way as to allow the use of relaxation solution procedures. It is, therefore, possible to solve either irrotational or rotational flow problems spanning the entire spectrum of subsonic, transonic, and supersonic conditions. The solutions can be obtained without an employement of either derived dependent variables, semidirect methods, or an unsteady formulation.

Johnson, G. M.↗

NASA NTRS2017

GPU Accelerated Prognostics

Prognostic methods enable operators and maintainers to predict the future performance for critical systems. However, these methods can be computationally expensive and may need to be performed each time new information about the system becomes available. In light of these computational requirements, we have investigated the application of graphics processing units (GPUs) as a computational platform for real-time prognostics. Recent advances in GPU technology have reduced cost and increased the computational capability of these highly parallel processing units, making them more attractive for the deployment of prognostic software. We present a survey of model-based prognostic algorithms with considerations for leveraging the parallel architecture of the GPU and a case study of GPU-accelerated battery prognostics with computational performance results.

Prognostics↗

NASA NTRS1979

Space shuttle active-pogo-suppressor control design using linear quadratic regulator techniques

Two methods of active pogo suppression (stabilization) for the space shuttle vehicle were studied analytically. The basis for both approaches was the linear quadratic regulator, state space technique. The first approach minimized root-mean-square pump inlet pressure by using either fullstate feedback, partial-state feedback, or output feedback with a Kalman filter. The second approach increased the modal damping associated with the critical structural modes by using either full-state feedback or reconstructed state feedback. A number of implementable controls were found by both approaches. The designs were analyzed with respect to sensitivity, complexity, and controller energy requirements, as well as controller performance. Practical controllers resulting from the two design approaches tended to use pressure and flow as feedback variables for the minimum-rms method and structural accelerations or velocities for the modal control method. Both approaches are suitable for the design of active pogo-suppression controllers.

Lehtinen, B.↗

NASA NTRS1985

Investigation of test methods, material properties and processes for solar cell encapsulants

The historical development of ethylene vinyl acetate (EVA) is presented, including the functional requirements, polymer selection, curing, stabilization, production and module processing. The construction and use of a new method for the accelerated aging of polymers is detailed. The method more closely resembles the conditions that may be encountered in actual module field exposure and additionally may permit service life to be predicted accurately. The use of hardboard as a low cost candidate substrate material is studied. The performance of surface antisoiling treatments useful for imparting a self cleaning property to modules is updated.

Willis, P. B.↗

NASA NTRS1991

Compensating sampling errors in stabilizing helmet-mounted displays using auxiliary acceleration measurements

A method based on complementary filtering is shown to be effective in compensating for the image stabilization error due to sampling delays of HMD position and orientation measurements. These delays would otherwise have prevented the stabilization of the image in HMDs. The method is also shown to improve the resolution of the head orientation measurement, particularly at low frequencies, thus providing smoother head control commands, which are essential for precise head pointing and teleoperation.

Merhav, S.↗

NASA NTRS1993

Time-accurate Navier-Stokes calculations with multigrid acceleration

An efficient method for calculating unsteady flows is presented, with emphasis on a modified version of the thin-layer Navier-Stokes equations. Fourier stability analysis is used to illustrate the effect of treating the source term implicitly instead of explicity, as well as to illustrate other algorithmic choices. A 2D circular cylinder (with a Reynolds number of 1200 and a Mach number of 0.3) is calculated. The present scheme requires only about 10 percent of the computer time required by global minimum time stepping.

Melson, N. D.↗

DOE OSTI2022

Predicting beam transmission using 2-dimensional phase space projections of hadron accelerators

We present a method to compress the 2D transverse phase space projections from a hadron accelerator and use that information to predict the beam transmission. This method assumes that obtaining at least three projections of the 4D transverse phase space is possible and that an accurate simulation model is available for the beamline. Using a simulated model, we show that—a computer can train a convolutional autoencoder to reduce phase-space information which can later be used to predict the beam transmission. Finally, we argue that although using projections from a realistic nonlinear distribution produces less accurate results, the method still generalizes well.

43 PARTICLE ACCELERATORS↗

DOE OSTI2021

Diagnosing nuclear power plant pipe wall thinning due to flow accelerated corrosion using a passive, thermal non-destructive evaluation method: Feasibility assessment via numerical experiments

Flow accelerated corrosion (FAC) in nuclear power plant pipes is one of the leading causes of accidents, fatalities, damage and outages. Current FAC identification methods employ expensive sensing technology and are “active” methods, where the response of the piping system to an externally-generated thermal, mechanical or optical excitation must be measured. As a result, these techniques require a disruptive and time-consuming setup. Here we propose a method that utilizes pipe surface temperature measurements to passively monitor for FAC-induced pipe wall thinning without the need for expensive equipment or post-installation setup time. This diagnostic method utilizes a simulation data-driven diagnostic model to estimate the amount of thickness reduction in a pipe based on changes in measured steady-state pipe temperatures. In order to reduce the computational burden of generating large, simulation-based datasets, the behavior of the insulation of the pipe was modeled using a suitably calibrated heat transfer boundary parameter. Additionally, global sensitivity analysis was performed to determine system parameter(s), such as the temperature of water flowing inside the pipe, which significantly affect the steady state pipe wall temperature and could cause errors in diagnosis. Two diagnostic models, one using only the change in steady-state temperature as an indicator for FAC-induced pipe wall thinning and the other using water temperature as an additional diagnostic model input were evaluated for their ability to estimate thickness reductions in a pipe using simulated pipe wall temperature data. For the numerical experiments conducted in this work, both models estimated wall thickness with errors within 0.5 mm, indicating that the proposed technique can potentially be used as a low-cost, first-pass method for FAC monitoring.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

NASA NTRS1990

Implementation of two projection methods on a shared memory multiprocessor - DEC VAX 6240

The relative performance of two iterative schemes, based on projection techniques, is compared on a shared memory multiprocessor - VAX 6240. The CG accelerated Block-SSOR method and the CG accelerated Symmetric-Kaczmarz method are considered for the solution of large sparse nonsymmetric systems of linear equations. It is shown that the regular structure of many matrices can be exploited by the CG-accelerated Block-SSOR method to provide good speedup in a multiprocessing environment. However, the CG accelerated Symmetric-Kaczmarz method, while being a viable alternative on a scalar machine, is unable to benefit from multiprocessing.

Kamath, C.↗

DOE OSTI2021

Integrating Advanced Modeling and Accelerated Testing for a Modernized Fuel Qualification Paradigm

With the increasing interest in sodium fast reactor technology, as seen by applications to the U.S. Nuclear Regulatory Commission for the OKLO Aurora plant, fuel testing for the TerraPower Traveling Wave Reactor, and the impending construction and startup of the versatile test reactor (VTR), a modernized, accelerated approach to fuel qualification is needed. To guide this effort, a Phenomena Identification Ranking Table–styled analysis was performed for a U-Pu-Zr sodium-free annular fuel system. This analysis evaluated a series of fuel design properties and parameters against their contributions to key fuel performance phenomena. The resulting priority parameters were then reviewed against existing modeling and experimental capabilities to support investigation of the highest-priority parameters. A pathway for qualification was then established using highthroughput, high-volume experiments from MiniFuel and FAST in parallel with advanced physics-based model development. This effort outlines how the first stages of qualification can be reduced from the typical 20+-year development cycle to 5 to 7 years by deploying accelerated irradiation testing platforms. As with any accelerated test, these methods are prototypic in some aspects and less so in others; however, by coupling with advanced fuel performance modeling and simulation capabilities, the larger space of irradiation parameters and material response provided offers advantages for the validation of physics-based models supporting the deployment of novel fuel designs. As a test case, this paper utilizes a proposed Mark II fuel system for the upcoming VTR. Thus, an accelerated qualification method can be tested for the development of MARK II driver fuel so that by the time of VTR startup, lead test assemblies for a Mark II fuel can be initiated.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2024

Toward a Holistic Performance Evaluation of Large Language Models Across Diverse AI Accelerators

Artificial intelligence (AI) methods have become critical in scientific applications to help accelerate scientific discovery. Large language models (LLMs) are being considered a promising approach to address some challenging problems because of their superior generalization capabilities across domains. The effectiveness of the models and the accuracy of the applications are contingent upon their efficient execution on the underlying hardware infrastructure. Specialized Al accelerator hardware systems have recently become available for accelerating Al applications. However, the comparative performance of these AI accelerators on large language models has not been previously studied. In this paper, we systematically study LLMs on multiple AI accelerators and GPUs and evaluate their performance characteristics for these models. We evaluate these systems with (i) a micro-benchmark using a core transformer block, (ii) a GPT-2 model, and (iii) an 1,I,M-driven science use case, GenSLM. We present our findings and analyses of the models' performance to better understand the intrinsic capabilities of AI accelerators. Furthermore, our analysis takes into account key factors such as sequence lengths, scaling behavior, and sensitivity to gradient accumulation steps.

Emani, Murali↗

NASA NTRS1965

Energy addition methods for hyper-velocity true flight simulation

Arc heating and magnetohydrodynamic acceleration energy addition methods for wind tunnel hypervelocity flow true flight simulation

FLIGHT SIMULATION↗

NASA NTRS1991

Spaceborne centrifugal relays for spacecraft propulsion

Acceleration using centrifugal relays is a recently discovered method for the acceleration of spaceborne payloads to high velocity at high thrust. Centrifugal relays are moving rotors which progressively accelerate reaction mass to higher velocities. One important engineering problem consists of accurately tracking the position of the projectiles and rotors and guiding each projectile exactly onto the appropriate guide tracks on each rotor. The topics of this research are the system kinematics and dynamics and the computerized guidance system which will allow the projectile to approach each rotor with exact timing with respect to the rotor rotation period and with very small errors in lateral positions. Kinematics studies include analysis of rotor and projectile positions versus time and projectile/rotor interactions. Guidance studies include a detailed description of the tracking mechanism (interrupt of optical beams) and the aiming mechanism (electromagnetic focusing) including the design of electromagnetic deflection coils and the switching circuitry.

Ouzidane, Malika↗

NASA NTRS1999

Efficient Multi-Stage Time Marching for Viscous Flows via Local Preconditioning

A new method has been developed to accelerate the convergence of explicit time-marching, laminar, Navier-Stokes codes through the combination of local preconditioning and multi-stage time marching optimization. Local preconditioning is a technique to modify the time-dependent equations so that all information moves or decays at nearly the same rate, thus relieving the stiffness for a system of equations. Multi-stage time marching can be optimized by modifying its coefficients to account for the presence of viscous terms, allowing larger time steps. We show it is possible to optimize the time marching scheme for a wide range of cell Reynolds numbers for the scalar advection-diffusion equation, and local preconditioning allows this optimization to be applied to the Navier-Stokes equations. Convergence acceleration of the new method is demonstrated through numerical experiments with circular advection and laminar boundary-layer flow over a flat plate.

Kleb, William L.↗

NASA NTRS1999

Efficient Multi-Stage Time Marching for Viscous Flows via Local Preconditioning

Kleb, William L.↗

NASA NTRS1995

Convergence acceleration of the Proteus computer code with multigrid methods

This report presents the results of a study to implement convergence acceleration techniques based on the multigrid concept in the two-dimensional and three-dimensional versions of the Proteus computer code. The first section presents a review of the relevant literature on the implementation of the multigrid methods in computer codes for compressible flow analysis. The next two sections present detailed stability analysis of numerical schemes for solving the Euler and Navier-Stokes equations, based on conventional von Neumann analysis and the bi-grid analysis, respectively. The next section presents details of the computational method used in the Proteus computer code. Finally, the multigrid implementation and applications to several two-dimensional and three-dimensional test problems are presented. The results of the present study show that the multigrid method always leads to a reduction in the number of iterations (or time steps) required for convergence. However, there is an overhead associated with the use of multigrid acceleration. The overhead is higher in 2-D problems than in 3-D problems, thus overall multigrid savings in CPU time are in general better in the latter. Savings of about 40-50 percent are typical in 3-D problems, but they are about 20-30 percent in large 2-D problems. The present multigrid method is applicable to steady-state problems and is therefore ineffective in problems with inherently unstable solutions.

Demuren, A. O.↗

DOE OSTI2021

Scalable HPC & AI infrastructure for COVID-19 therapeutics

COVID-19 has claimed more than 2.7 × 10^6 lives and resulted in over 124 × 10^6 infections. There is an urgent need to identify drugs that can inhibit SARS-CoV-2. We discuss innovations in computational infrastructure and methods that are accelerating and advancing drug design. Specifically, we describe several methods that integrate artificial intelligence and simulation-based approaches, and the design of computational infrastructure to support these methods at scale. We discuss their implementation, characterize their performance, and highlight science advances that these capabilities have enabled.

Lee, Hyungro↗

NASA NTRS2017

CME Velocity and Acceleration Error Estimates Using the Bootstrap Method

The bootstrap method is used to determine errors of basic attributes of coronal mass ejections (CMEs) visually identified in images obtained by the Solar and Heliospheric Observatory (SOHO) mission's Large Angle and Spectrometric Coronagraph (LASCO) instruments. The basic parameters of CMEs are stored, among others, in a database known as the SOHO/LASCO CME catalog and are widely employed for many research studies. The basic attributes of CMEs (e.g. velocity and acceleration) are obtained from manually generated height-time plots. The subjective nature of manual measurements introduces random errors that are difficult to quantify. In many studies the impact of such measurement errors is overlooked. In this study we present a new possibility to estimate measurements errors in the basic attributes of CMEs. This approach is a computer-intensive method because it requires repeating the original data analysis procedure several times using replicate datasets. This is also commonly called the bootstrap method in the literature. We show that the bootstrap approach can be used to estimate the errors of the basic attributes of CMEs having moderately large numbers of height-time measurements. The velocity errors are in the vast majority small and depend mostly on the number of height-time points measured for a particular event. In the case of acceleration, the errors are significant, and for more than half of all CMEs, they are larger than the acceleration itself.

Michalek, Grzegorz↗