Results for “Approximate computing” — Page 79

NASA NTRS1991

Hierarchic Extensions in the Static and Dynamic Analysis of Elastic Beams

Approximate solutions of static and dynamic beam problems by the p-version of the finite element method are investigated. Within a hierarchy of engineering beam idealizations, rigorous formulations of the strain and kinetic energies for straight and circular beam elements are presented. These formulations include rotating coordinate system effects and geometric nonlinearities to allow for the evaluation of vertical axis wind turbines, the motivating problem for this research. Hierarchic finite element spaces, based on extensions of the polynomial orders used to approximate the displacement variables, are constructed. The developed models are implemented into a general purpose computer program for evaluation. Quality control procedures are examined for a diverse set of sample problems. These procedures include estimating discretization errors in energy norm and natural frequencies, performing static and dynamic equilibrium checks, observing convergence for qualities of interest, and comparison with more exacting theories and experimental data. It is demonstrated that p-extensions produce exponential rates of convergence in the approximation of strain energy and natural frequencies for the class of problems investigated.

Watson, Robert A.↗

NASA NTRS1995

Combining Comparison Functions and Finite Element Approximations in CFD

In a variety of potential flow applications, the modal element method has been shown to significantly reduce the numerical grid, employ a more precise grid termination boundary condition, and give theoretical insight to the flow physics. The method employs eigenfunctions to replace the numerical grid over significant portions of the flow field. Generally, a numerical grid is employed around obstacles with complex geometry while eigenfunctions are applied to regions in the flow field where the boundary conditions can easily be satisfied. To handle a wider class of computational fluid dynamics (CFD) problems, the present paper extends the modal element to include function approximations which do not satisfy the governing differential equation. To accomplish this task, a double modal series approximation and weighted residual constraints are developed to force the comparison functions to satisfy the governing differential equation and to interface properly with the finite element solution. As an example, the method is applied to the problem of potential flow in a channel with two-dimensional cylindrical like obstacles. The calculated flow fields are in excellent agreement with exact analytical solutions.

Baumeister, Kenneth J.↗

NASA NTRS2005

Computational Materials Program for Alloy Design

The research program sponsored by this grant, "Computational Materials Program for Alloy Design", covers a period of time of enormous change in the emerging field of computational materials science. The computational materials program started with the development of the BFS method for alloys, a quantum approximate method for atomistic analysis of alloys specifically tailored to effectively deal with the current challenges in the area of atomistic modeling and to support modern experimental programs. During the grant period, the program benefited from steady growth which, as detailed below, far exceeds its original set of goals and objectives. Not surprisingly, by the end of this grant, the methodology and the computational materials program became an established force in the materials communitiy, with substantial impact in several areas. Major achievements during the duration of the grant include the completion of a Level 1 Milestone for the HITEMP program at NASA Glenn, consisting of the planning, development and organization of an international conference held at the Ohio Aerospace Institute in August of 2002, finalizing a period of rapid insertion of the methodology in the research community worlwide. The conference, attended by citizens of 17 countries representing various fields of the research community, resulted in a special issue of the leading journal in the area of applied surface science. Another element of the Level 1 Milestone was the presentation of the first version of the Alloy Design Workbench software package, currently known as "adwTools". This software package constitutes the first PC-based piece of software for atomistic simulations for both solid alloys and surfaces in the market.Dissemination of results and insertion in the materials community worldwide was a primary focus during this period. As a result, the P.I. was responsible for presenting 37 contributed talks, 19 invited talks, and publishing 71 articles in peer-reviewed journals, as detailed later in this Report.

Bozzolo, Guillermo↗

NASA NTRS1993

Efficient and accurate computation of the incomplete Airy functions

The incomplete Airy integrals serve as canonical functions for the uniform ray optical solutions to several high-frequency scattering and diffraction problems that involve a class of integrals characterized by two stationary points that are arbitrarily close to one another or to an integration endpoint. Integrals with such analytical properties describe transition region phenomena associated with composite shadow boundaries. An efficient and accurate method for computing the incomplete Airy functions would make the solutions to such problems useful for engineering purposes. In this paper a convergent series solution for the incomplete Airy functions is derived. Asymptotic expansions involving several terms are also developed and serve as large argument approximations. The combination of the series solution with the asymptotic formulae provides for an efficient and accurate computation of the incomplete Airy functions. Validation of accuracy is accomplished using direct numerical integration data.

Constantinides, E. D.↗

NASA NTRS2010

Guidance and Control System for a Satellite Constellation

A distributed guidance and control algorithm was developed for a constellation of satellites. The system repositions satellites as required, regulates satellites to desired orbits, and prevents collisions. 1. Optimal methods are used to compute nominal transfers from orbit to orbit. 2. Satellites are regulated to maintain the desired orbits once the transfers are complete. 3. A simulator is used to predict potential collisions or near-misses. 4. Each satellite computes perturbations to its controls so as to increase any unacceptable distances of nearest approach to other objects. a. The avoidance problem is recast in a distributed and locally-linear form to arrive at a tractable solution. b. Plant matrix values are approximated via simulation at each time step. c. The Linear Quadratic Gaussian (LQG) method is used to compute perturbations to the controls that will result in increased miss distances. 5. Once all danger is passed, the satellites return to their original orbits, all the while avoiding each other as above. 6. The delta-Vs are reasonable. The controller begins maneuvers as soon as practical to minimize delta-V. 7. Despite the inclusion of trajectory simulations within the control loop, the algorithm is sufficiently fast for available satellite computer hardware. 8. The required measurement accuracies are within the capabilities of modern inertial measurement devices and modern positioning devices.

Bryson, Jonathan Lamar↗

NASA NTRS1999

Neural Network and Regression Approximations Used in Aircraft Design

NASA Lewis Research Center's CometBoards Test Bed was used to create regression and neural network models for a High-Speed Civil Transport (HSCT) aircraft. Both approximation models that replaced the actual analysis tool predicted the aircraft response in a trivial computational effort. The models allow engineers to quickly study the effects of design variables on constraint and objective values for a given aircraft configuration. For example, an engineer can change the engine size by 1000 pounds of thrust and quickly see how this change affects all the output values without rerunning the entire simulation. In addition, an engineer can change a constraint and use the approximation models to quickly reoptimize the configuration. Generating the neural network and the regression models is a time-consuming process, but this exercise has to be carried out only once. Furthermore, an automated process can reduce calculations substantially.

Patnaik, Surya N.↗

NASA NTRS1996

A-priori testing of sub-grid models for chemically reacting nonpremixed turbulent shear flows

The beta-assumed-pdf approximation of (Cook & Riley 1994) is tested as a subgrid model for the LES computation of nonpremixed turbulent reacting flows, in the limit of cold infinitely fast chemistry, for two plane turbulent mixing layers with different degrees of intermittency. Excellent results are obtained for the computation of integrals properties such as product mass fraction, and the model is applied to other quantities such as powers of the temperature and the pdf of the scalar itself. Even in these cases the errors are small enough to be useful in practical applications. The analysis is extended to slightly out of equilibrium problems such as the generation of radicals, and formulated in terms of the pdf of the scalar gradients. It is shown that the conditional gradient distribution is universal in a wide range of cases whose limits are established. Within those limits, engineering approximations to the radical concentration are also possible. It is argued that the experiments in this paper are essentially in the limit of infinite Reynolds number.

Jimenez, J.↗

NASA NTRS2006

Shuttle Debris Impact Tool Assessment Using the Modern Design of Experiments

Computational tools have been developed to estimate thermal and mechanical reentry loads experienced by the Space Shuttle Orbiter as the result of cavities in the Thermal Protection System (TPS). Such cavities can be caused by impact from ice or insulating foam debris shed from the External Tank (ET) on liftoff. The reentry loads depend on cavity geometry and certain Shuttle state variables, among other factors. Certain simplifying assumptions have been made in the tool development about the cavity geometry variables. For example, the cavities are all modeled as shoeboxes , with rectangular cross-sections and planar walls. So an actual cavity is typically approximated with an idealized cavity described in terms of its length, width, and depth, as well as its entry angle, exit angle, and side angles (assumed to be the same for both sides). As part of a comprehensive assessment of the uncertainty in reentry loads estimated by the debris impact assessment tools, an effort has been initiated to quantify the component of the uncertainty that is due to imperfect geometry specifications for the debris impact cavities. The approach is to compute predicted loads for a set of geometry factor combinations sufficient to develop polynomial approximations to the complex, nonparametric underlying computational models. Such polynomial models are continuous and feature estimable, continuous derivatives, conditions that facilitate the propagation of independent variable errors. As an additional benefit, once the polynomial models have been developed, they require fewer computational resources to execute than the underlying finite element and computational fluid dynamics codes, and can generate reentry loads estimates in significantly less time. This provides a practical screening capability, in which a large number of debris impact cavities can be quickly classified either as harmless, or subject to additional analysis with the more comprehensive underlying computational tools. The polynomial models also provide useful insights into the sensitivity of reentry loads to various cavity geometry variables, and reveal complex interactions among those variables that indicate how the sensitivity of one variable depends on the level of one or more other variables. For example, the effect of cavity length on certain reentry loads depends on the depth of the cavity. Such interactions are clearly displayed in the polynomial response models.

DeLoach, R.↗

NASA NTRS1985

Convergence acceleration for a three-dimensional Euler/Navier-Stokes zonal approach

A fast diagonal algorithm is coupled with a zonal approach to solve the three-dimensional Euler/Navier-Stokes equations. Transonic viscous solutions are obtained on a 150,000 point mesh for a NACA 0012 wing. The new computational approach yields a speedup by as much as a factor of 40 over the standard Beam-Warming algorithm/zonal method originally coded. A three-order-of-magnitude drop in the L2-norm of the residual requires approximately 500 iterations, which takes about 45 min of CPU time on a Cray-XMP. The numerically computed solutions are in good agreement with experimental results. Effects on convergence rate owing to increasing the zonal boundary overlap regions, different stretching distributions in the viscous regions, and different CFL values are also explored.

Flores, J.↗

NASA NTRS1997

Global Magnetohydrodynamic Modeling of the Solar Corona

Under this contract, we have continued our investigations of the large scale structure of the solar corona and inner heliosphere using global magnetohydrodynamic (MHD) simulations. These computations have also formed the basis for studies of coronal mass ejections (CMES) using realistic coronal configurations. We have developed a technique for computing realistic magnetohydrodynamic (MHD) computations of the solar corona and inner heliosphere. To perform computations that can be compared with specific observations, it is necessary to incorporate solar observations into the boundary conditions. We have used the Wilcox Solar Observatory synoptic maps (collected during a solar rotation by daily measurements of the line-of-sight magnetic field at central meridian) to specify the radial magnetic field (B,) at the photosphere. For the initial condition, we use a potential magnetic field consistent with the specified distribution of B, at the lower boundary, and a wind solution consistent with the specified plasma density and temperature at the solar surface. Together this initial condition forms a (non-equilibrium) approximation of the state of the solar corona for the time-dependent MHD computation. The MHD equations are then integrated in time to steady state. Here we describe solutions relevant to a recent solar eclipse, as well as Ulysses observations. We have also developed a model configuration of solar minimum, useful for studying CME initiation and propagation.

Linker, Jon A.↗

NASA NTRSDate not supplied

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

High-performance computing (HPC) architectures are trending toward manycore paradigms such as graphics processing units (GPUs). Approximately half of the top 100 publicly disclosed supercomputers in the world utilize GPU accelerators for performance. This is in contrast to a decade ago, where there were only a few such machines in the top 100. It is not currently possible to compile and run legacy central processing unit (CPU) software efficiently on GPUs without significant refactoring. Though a number of frameworks offering performance portability exist, none offer a standardized specification that is supported by all major hardware vendors. Additionally, experiences show that obtaining a high percentage of peak performance often requires architecture-specific code. This work details a pragmatic multi-architecture computational fluid dynamics library focused on aerospace problems across the speed range from low subsonic to hypersonic flows involving thermochemical nonequilibrium. A thin abstraction layer above NVIDIA CUDA C++ is utilized, which enables primarily single-source software currently capable of running efficiently on multicore CPUs, NVIDIA GPUs, AMD GPUs, and Intel GPUs. Results on various problems of interest across the speed range are presented and performance is compared between various architectures.

GPU↗

NASA NTRSDate not supplied

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

High-performance computing (HPC) architectures are trending toward manycore paradigms such as graphics processing units (GPUs). Approximately half of the top 100 publicly disclosed supercomputers in the world utilize GPU accelerators for performance. This is in contrast to a decade ago, where there were only a few such machines in the top 100. It is not currently possible to compile and run legacy central processing unit (CPU) software efficiently on GPUs without significant refactoring. Though a number of frameworks offering performance portability exist, none offer a standardized specification that is supported by all major hardware vendors. Additionally, experiences show that obtaining a high percentage of peak performance often requires architecture-specific code. This work details a pragmatic multi-architecture computational fluid dynamics library focused on aerospace problems across the speed range from low subsonic to hypersonic flows involving thermochemical nonequilibrium. A thin abstraction layer above NVIDIA CUDA C++ is utilized, which enables primarily single-source software currently capable of running efficiently on multicore CPUs, NVIDIA GPUs, AMD GPUs, and Intel GPUs. Results on various problems of interest across the speed range are presented and performance is compared between various architectures.

GPU↗

NASA NTRS1994

Three-Dimensional Navier-Stokes Method with Two-Equation Turbulence Models for Efficient Numerical Simulation of Hypersonic Flows

A new computational efficient 3-D compressible Reynolds-averaged implicit Navier-Stokes method with advanced two equation turbulence models for high speed flows is presented. All convective terms are modeled using an entropy satisfying higher-order Total Variation Diminishing (TVD) scheme based on implicit upwind flux-difference split approximations and arithmetic averaging procedure of primitive variables. This method combines the best features of data management and computational efficiency of space marching procedures with the generality and stability of time dependent Navier-Stokes procedures to solve flows with mixed supersonic and subsonic zones, including streamwise separated flows. Its robust stability derives from a combination of conservative implicit upwind flux-difference splitting with Roe's property U to provide accurate shock capturing capability that non-conservative schemes do not guarantee, alternating symmetric Gauss-Seidel 'method of planes' relaxation procedure coupled with a three-dimensional two-factor diagonal-dominant approximate factorization scheme, TVD flux limiters of higher-order flux differences satisfying realizability, and well-posed characteristic-based implicit boundary-point a'pproximations consistent with the local characteristics domain of dependence. The efficiency of the method is highly increased with Newton Raphson acceleration which allows convergence in essentially one forward sweep for supersonic flows. The method is verified by comparing with experiment and other Navier-Stokes methods. Here, results of adiabatic and cooled flat plate flows, compression corner flow, and 3-D hypersonic shock-wave/turbulent boundary layer interaction flows are presented. The robust 3-D method achieves a better computational efficiency of at least one order of magnitude over the CNS Navier-Stokes code. It provides cost-effective aerodynamic predictions in agreement with experiment, and the capability of predicting complex flow structures in complex geometries with good accuracy.

Bardina, J. E.↗

NASA NTRS1986

Upwind Navier-Stokes solutions for separated periodic flows

The application of an upwind implicit approximate factorization Navier-Stokes algorithm to highly separated flow is described. Using both the thin-layer and complete forms of the Navier-Stokes equations, the low Reynolds number laminar flow around a circular cylinder with periodic shedding is solved. The effect of grid density, grid extent, and time step on the Strouhal number is shown. Results from both sets of equations agree within the experimental data band. Unsteady, laminar flow computations around inclined plates and separated airfoils are also described. Strouhal numbers agree to within 5 percent of experiments for inclined plates. Differences between the complete equations and the thin-layer approximation for separated periodic flows are discussed. Computations of an impulsively started circular cylinder and airfoil yield time-accurate flowfield shapes in good agreement with experimental flow visualizations. The turbulent computation of an airfoil at a high angle-of-attack is massively separated, but shows no evidence of periodicity.

Rumsey, C. L.↗

NASA NTRS1981

Numerical analysis of spectral properties of coupled oscillator Schroedinger operators. I - Single and double well anharmonic oscillators

Several methods for computing many eigenvalues and eigenfunctions of a single anharmonic oscillator Schroedinger operator whose potential may have one or two minima are described. One of the methods requires the solution of an ill-conditioned generalized eigenvalue problem. This method has the virtue of using a bounded amount of work to achieve a given accuracy in both the single and double well regions. Rigorous bounds are given, and it is proved that the approximations converge faster than any inverse power of the size of the matrices needed to compute them. The results of computations for the g:phi(4):1 theory are presented. These results indicate that the methods actually converge exponentially fast.

Isaacson, D.↗

NASA NTRS1993

Mathematical description of complex chemical kinetics and application to CFD modeling codes

A major effort in combustion research at the present time is devoted to the theoretical modeling of practical combustion systems. These include turbojet and ramjet air-breathing engines as well as ground-based gas-turbine power generating systems. The ability to use computational modeling extensively in designing these products not only saves time and money, but also helps designers meet the quite rigorous environmental standards that have been imposed on all combustion devices. The goal is to combine the very complex solution of the Navier-Stokes flow equations with realistic turbulence and heat-release models into a single computer code. Such a computational fluid-dynamic (CFD) code simulates the coupling of fluid mechanics with the chemistry of combustion to describe the practical devices. This paper will focus on the task of developing a simplified chemical model which can predict realistic heat-release rates as well as species composition profiles, and is also computationally rapid. We first discuss the mathematical techniques used to describe a complex, multistep fuel oxidation chemical reaction and develop a detailed mechanism for the process. We then show how this mechanism may be reduced and simplified to give an approximate model which adequately predicts heat release rates and a limited number of species composition profiles, but is computationally much faster than the original one. Only such a model can be incorporated into a CFD code without adding significantly to long computation times. Finally, we present some of the recent advances in the development of these simplified chemical mechanisms.

Bittker, D. A.↗

NASA NTRS1993

Mathematical Description of Complex Chemical Kinetics and Application to CFD Modeling Codes

A major effort in combustion research at the present time is devoted to the theoretical modeling of practical combustion systems. These include turbojet and ramjet air-breathing engines as well as ground-based gas-turbine power generating systems. The ability to use computational modeling extensively in designing these products not only saves time and money, but also helps designers meet the quite rigorous environmental standards that have been imposed on all combustion devices. The goal is to combine the very complex solution of the Navier-Stokes flow equations with realistic turbulence and heat-release models into a single computer code. Such a computational fluid-dynamic (CFD) code simulates the coupling of fluid mechanics with the chemistry of combustion to describe the practical devices. This paper will focus on the task of developing a simplified chemical model which can predict realistic heat-release rates as well as species composition profiles, and is also computationally rapid. We first discuss the mathematical techniques used to describe a complex, multistep fuel oxidation chemical reaction and develop a detailed mechanism for the process. We then show how this mechanism may be reduced and simplified to give an approximate model which adequately predicts heat release rates and a limited number of species composition profiles, but is computationally much faster than the original one. Only such a model can be incorporated into a CFD code without adding significantly to long computation times. Finally, we present some of the recent advances in the development of these simplified chemical mechanisms.

Bittker, D. A.↗

NASA NTRS1978

Effect of molecular anisotropy on beam scattering measurements

Within the energy sudden approximation, the total integral and total differential scattering cross sections are given by the angle average of scattering cross sections computed at fixed rotor orientations. Using this formalism the effect of molecular anisotropy on scattering of He by HCl and by CO is examined. Comparisons with accurate close coupling calculations indicate that this approximation is quite reliable, even at very low collision energies, for both of these systems. Comparisons are also made with predictions based on the spherical average of the interaction. For HCl the anisotropy is rather weak and its main effect is a slight quenching of the oscillations in the differential cross sections relative to predictions of the spherical averaged potential. For CO the anisotropy is much stronger, so that the oscillatory pattern is strongly quenched and somewhat shifted. It appears that the sudden approximation provides a simple yet accurate method for describing the effect of molecular anisotropy on scattering measurements.

Goldflam, R.↗