Search NASA⌕ Search

SEARCH · Search NASA

Results for “numerical solver”

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 919 records · Page 51
All sourcesNASA NTRSDOE OSTI
NASA NTRS1993

Computer modeling of pulsed CO2 lasers for lidar applications

The object of this effort is to develop code to enable the accurate prediction of the performance of pulsed transversely excited (TE) CO2 lasers prior to their construction. This is of particular benefit to the NASA Laser Atmospheric Wind Sounder (LAWS) project. A benefit of the completed code is that although developed specifically for the pulsed CO2 laser much of the code can be modified to model other laser systems of interest to the lidar community. A Boltzmann equation solver has been developed which enables the electron excitation rates for the vibrational levels of CO2 and N2, together with the electron ionization and attachment coefficients to be determined for any CO2 laser gas mixture consisting of a combination of CO2, N2, CO, He and CO. The validity of the model has been verified by comparison with published material. The results from the Boltzmann equation solver have been used as input to the laser kinetics code which is currently under development. A numerical code to model the laser induced medium perturbation (LIMP) arising from the relaxation of the lower laser level has been developed and used to determine the effect of LIMP on the frequency spectrum of the LAWS laser output pulse. The enclosed figures show representative results for a laser operating at 0.5 atm. with a discharge cross-section of 4.5 cm to produce a 20 J pulse with aFWHM of 3.1 microns. The first four plots show the temporal evolution of the laser pulse power, energy evolution, LIMP frequency chirp and electric field magnitude. The electric field magnitude is taken by beating the calculated complex electric field and beating it with a local oscillator signal. The remaining two figures show the power spectrum and energy distribution in the pulse as a function of the varying pulse frequency. The LIMP theory has been compared with experimental data from the NOAA Windvan Lidar and has been found to be in good agreement.

Spiers, Gary D.↗
NASA NTRS1996

Progress Toward an Efficient and General CFD Tool for Propulsion Design/Analysis

The simulation of propulsive flows inherently involves chemical activity. Recent years have seen substantial strides made in the development of numerical schemes for reacting flowfields, in particular those involving finite-rate chemistry. However, finite-rate calculations are computationally intensive and require knowledge of the actual kinetics, which are not always known with sufficient accuracy. Alternatively, flow simulations based on the assumption of local chemical equilibrium are capable of obtaining physically reasonable results at far less computational cost. The present study summarizes the development of efficient numerical techniques for the simulation of flows in local chemical equilibrium, whereby a 'Black Box' chemical equilibrium solver is coupled to the usual gasdynamic equations. The generalization of the methods enables the modelling of any arbitrary mixture of thermally perfect gases, including air, combustion mixtures and plasmas. As demonstration of the potential of the methodologies, several solutions, involving reacting and perfect gas flows, will be presented. Included is a preliminary simulation of the SSME startup transient. Future enhancements to the proposed techniques will be discussed, including more efficient finite-rate and hybrid (partial equilibrium) schemes. The algorithms that have been developed and are being optimized provide for an efficient and general tool for the design and analysis of propulsion systems.

Cox, C. F.↗
NASA NTRS2011

Numerical Simulations of a 20-kW Class Hall Thruster Using the Magnetic-Field-Aligned-Mesh Code Hall2De

This paper reports on numerical simulations of the NASA-300M, a 20-kW class Hall thruster developed at the NASA Glenn Research Center (GRC). The numerical simulations have been performed with a 2-D axisymmetric, magnetic field-aligned-mesh (MFAM) plasma solver developed at the Jet Propulsion Laboratory (JPL). The main objective of the collaborative effort is to combine physics-based simulation, plasma diagnostics and recent findings on erosion physics to design and demonstrate a high-power, high-performance Hall thruster that exceeds the life of state-of-the-art Hall thrusters by more than one order of magnitude. The thruster simulations have been carried out at a discharge voltage of 500 V and discharge current of 40 A. The results indicate that although the impact energy of ions may attain values that are comparable to the discharge voltage along the downstream portions of the channel, a withdrawn ionization region and significant ion focusing combine to sustain erosion rates below 1 mm/kh. A more extensive evaluation of the baseline NASA-300M configuration and re-design of this thruster with magnetically shielded walls constitute the main focus of our work in the coming months.

plasma simulations↗
NASA NTRS2010

Magnetic Shielding of the Acceleration Channel Walls in a Long-Life Hall Thruster

In a Qualification Life Test (QLT) of the BPT-4000 Hall thruster that recently accumulated greater than 10,000 h it was found that the erosion of the acceleration channel practically stopped after approximately 5,600 h. Numerical simulations of this thruster using a 2-D axisymmetric, magnetic field-aligned-mesh (MFAM) plasma solver reveal that the process that led to this significant reduction of the erosion was multifaceted. It is found that when the channel receded from its early-in-life geometry to its steady-state configuration several changes in the near-wall plasma and sheath were induced by the magnetic field that, collectively, constituted an effective shielding of the walls from any significant ion bombardment. Because all such changes in the behavior of the ionized gas near the eroding surfaces were caused by the topology of the magnetic field there, we term this process "magnetic shielding."

life modeling↗
NASA NTRS2017

CFD Simulations of the Supersonic Inflatable Aerodynamic Decelerator (SIAD) Ballistic Range Tests

A series of ballistic range tests were performed on a scaled model of the Supersonic Flight Demonstration Test (SFDT) intended to test the Supersonic Inflatable Aerodynamic Decelerator (SIAD) geometry. The purpose of these experiments were to provide aerodynamic coefficients of the vehicle to aid in mission and vehicle design. The experimental data spans the moderate Mach number range, $3.8-2.0$, with a total angle of attack ($alpha_T$) range, $10o-20o$. These conditions are intended to span the Mach-$alpha$ space for the majority of the SFDT experiment. In an effort to validate the predictive capabilities of Computational Fluid Dynamics (CFD) for free-flight aerodynamic behavior, numerical simulations of the ballistic range experiment are performed using the unstructured finite volume Navier-Stokes solver, US3D. Comparisons to raw vehicle attitude, and post-processed aerodynamic coefficients are made between simulated results and experimental data. The resulting comparisons for both raw model attitude and derived aerodynamic coefficients show good agreement with experimental results. Additionally, near body pressure field values for each trajectory simulated are investigated. Extracted surface and wake pressure data gives further insights into dynamic flow coupling leading to a potential mechanism for dynamic instability.

CFD modeling↗
NASA NTRSDate not supplied

Validation of Actuator Disk, Actuator Line and Sliding Mesh Methods within the LAVA Solver

In this study the implementation of actuator-disk, actuator-line and sliding-mesh methodologies in the Launch Ascent and Vehicle Aerodynamics (LAVA) solver is described and validated against several test-cases. The different models are validated against available numerical as well as experimental data. Both steady and unsteady Reynolds-Averaged Navier-Stokes (RANS) simulations using the Spallart Allmaras (SA) turbulence model are performed for several different configurations representative of aeroscience applications. The first part of the paper is focused on the verification and validation of the implemented propulsor models. The first validation case is a theoretical rotor in hover, compared with the 1D analytical solution derived from momentum theory. The second validation is the Rotor-Airframe Interaction Model of Georgia Institute of Technology (GIT), representing the application of actuator disks to top-mounted rotorcraft vehicles or unmanned aerial vehicles (UAVs). The third case is a representative configuration for tip-mounted rotor-craft vehicles such as NASA’s X57 airplane. The second part of the paper shows the implemented models applied to realistic engineering configurations such as NASA’s X57 airplane and the R4 Advanced Ducted Propellor (ADT). A modification to the actuator line method in order to represent the blade geometry more closely is proposed and comparisons with simulations modeling the blade utilizing a sliding-mesh approach are made.

ARMD↗
DOE OSTI2019

Electromagnetics Simulations with Vector-valued Finite Elements in MOOSE

Vector-valued finite elements have become an essential part of modern electromagnetic (EM) numerical simulation. This talk will describe the development and initial validation of a MOOSE-based EM solver application, ELK (Electromagnetics Library for Kinetics and fluids), that utilizes the recently added vector-valued finite element system. The primary motivation for the development of ELK was to study EM wave propagation in radio-frequency (rf) driven systems, such as those encountered in the study of antenna structures, waveguides, and plasma physics. In simulating these scenarios with traditional scalar-valued finite elements, boundary conditions on tangential field components can sometimes be difficult to define, and field discontinuities encountered with certain geometric or structural features (e.g., abrupt corners and inhomogeneous materials) can often lead to non-physical results. Vector-valued elements can largely mitigate and even eliminate these difficulties. ELK verification and validation efforts thus far have focused on standard wave propagation benchmarks for antennas and waveguides, and the impact of vector-valued elements on the implementation and results of these simulations will be described. Progress using ELK for plasma physics simulation, where inhomogeneous simulation domains are commonplace, will also be discussed.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS2015

Numerical Computation of a Continuous-thrust State Transition Matrix Incorporating Accurate Hardware and Ephemeris Models

A significant body of work exists showing that providing a nonlinear programming (NLP) solver with expressions for the problem constraint gradient substantially increases the speed of program execution and can also improve the robustness of convergence, especially for local optimizers. Calculation of these derivatives is often accomplished through the computation of spacecraft's state transition matrix (STM). If the two-body gravitational model is employed as is often done in the context of preliminary design, closed form expressions for these derivatives may be provided. If a high fidelity dynamics model, that might include perturbing forces such as the gravitational effect from multiple third bodies and solar radiation pressure is used then these STM's must be computed numerically. We present a method for the power hardward model and a full ephemeris model. An adaptive-step embedded eight order Dormand-Prince numerical integrator is discussed and a method for the computation of the time of flight derivatives in this framework is presented. The use of these numerically calculated derivatieves offer a substantial improvement over finite differencing in the context of a global optimizer. Specifically the inclusion of these STM's into the low thrust missiondesign tool chain in use at NASA Goddard Spaceflight Center allows for an increased preliminary mission design cadence.

Low Thrust↗
NASA NTRS2012

Numerical Simulation of Rocket Exhaust Interaction with Lunar Soil

This technology development originated from the need to assess the debris threat resulting from soil material erosion induced by landing spacecraft rocket plume impingement on extraterrestrial planetary surfaces. The impact of soil debris was observed to be highly detrimental during NASA s Apollo lunar missions and will pose a threat for any future landings on the Moon, Mars, and other exploration targets. The innovation developed under this program provides a simulation tool that combines modeling of the diverse disciplines of rocket plume impingement gas dynamics, granular soil material liberation, and soil debris particle kinetics into one unified simulation system. The Unified Flow Solver (UFS) developed by CFDRC enabled the efficient, seamless simulation of mixed continuum and rarefied rocket plume flow utilizing a novel direct numerical simulation technique of the Boltzmann gas dynamics equation. The characteristics of the soil granular material response and modeling of the erosion and liberation processes were enabled through novel first principle-based granular mechanics models developed by the University of Florida specifically for the highly irregularly shaped and cohesive lunar regolith material. These tools were integrated into a unique simulation system that accounts for all relevant physics aspects: (1) Modeling of spacecraft rocket plume impingement flow under lunar vacuum environment resulting in a mixed continuum and rarefied flow; (2) Modeling of lunar soil characteristics to capture soil-specific effects of particle size and shape composition, soil layer cohesion and granular flow physics; and (3) Accurate tracking of soil-borne debris particles beginning with aerodynamically driven motion inside the plume to purely ballistic motion in lunar far field conditions. In the earlier project phase of this innovation, the capabilities of the UFS for mixed continuum and rarefied flow situations were validated and demonstrated for lunar lander rocket plume flow impingement under lunar vacuum conditions. Applications and improvements to the granular flow simulation tools contributed by the University of Florida were tested against Earth environment experimental results. Requirements for developing, validating, and demonstrating this solution environment were clearly identified, and an effective second phase execution plan was devised. In this phase, the physics models were refined and fully integrated into a production-oriented simulation tool set. Three-dimensional simulations of Apollo Lunar Excursion Module (LEM) and Altair landers (including full-scale lander geometry) established the practical applicability of the UFS simulation approach and its advanced performance level for large-scale realistic problems.

Liever, Peter↗
NASA NTRS2007

The Attenuation of a Detonation Wave by an Aircraft Engine Axial Turbine Stage

A Constant Volume Combustion Cycle Engine concept consisting of a Pulse Detonation Combustor (PDC) followed by a conventional axial turbine was simulated numerically to determine the attenuation and reflection of a notional PDC pulse by the turbine. The multi-stage, time-accurate, turbomachinery solver TURBO was used to perform the calculation. The solution domain consisted of one notional detonation tube coupled to 5 vane passages and 8 rotor passages representing 1/8th of the annulus. The detonation tube was implemented as an initial value problem with the thermodynamic state of the tube contents, when the detonation wave is about to exit, provided by a 1D code. Pressure time history data from the numerical simulation was compared to experimental data from a similar configuration to verify that the simulation is giving reasonable results. Analysis of the pressure data showed a spectrally averaged attenuation of about 15 dB across the turbine stage. An evaluation of turbine performance is also presented.

VanZante, Dale↗
DOE OSTI2025

Score-Based Physics-Informed Neural Networks for High-Dimensional Fokker–Planck Equations

The Fokker-Planck (FP) equation is a foundational partial differential equation (PDE) in stochastic processes involving Brownian motions. However, the curse of dimensionality (CoD) poses a formidable challenge when dealing with high-dimensional FP equations. Although Monte Carlo simulation and (vanilla) Physics-Informed Neural Networks (PINNs) have shown the potential to tackle CoD, both methods exhibit significant numerical errors in high dimensions when dealing with the probability density function (PDF) associated with Brownian motion. The point-wise PDF values tend to decrease exponentially as dimensionality increases, surpassing the precision of numerical simulations and resulting in substantial errors. In addition, due to its massive sampling, Monte Carlo fails to offer fast sampling. Modeling the logarithm likelihood (LL) via vanilla PINNs transforms the FP equation into a notoriously difficult Hamilton-Jacobi-Bellman (HJB) equation, which is impractical for PINN learning, whose error grows rapidly with dimension. To this end, we propose a novel approach utilizing a score-based solver to fit the score function in stochastic differential equations (SDEs). The score function, defined as the gradient of the LL, plays a fundamental role in inferring LL and PDF and enables fast SDE sampling, offering an effective means to overcome the CoD. Three fitting methods, Score Matching (SM), Sliced Score Matching (SSM), and Score-PINN, are introduced, each contributing unique advantages in computational complexity, accuracy, and generality. The proposed score-based SDE solver operates in two stages: first, employing score matching or Score-PINN to acquire the score function; and second, solving the LL via an ordinary differential equation (ODE) using the obtained score function. Comparative evaluations across these methods showcase varying trade-offs. The proposed methodology is evaluated across diverse SDEs, including anisotropic Ornstein-Uhlenbeck processes, geometric Brownian motion, and Brownian motion with varying eigenspace. We also test various distributions, including Gaussian, Log-normal, Laplace, and Cauchy distributions. The numerical results demonstrate the score-based SDE solver’s stability, speed, and performance across different experimental settings, solidifying its potential as a solution to CoD for high-dimensional FP equations.

97 MATHEMATICS AND COMPUTING↗
NASA NTRSDate not supplied

Improved Benchmarking of Cohesive Elements in Abaqus Standard for Predicting Disbond and Delamination in Composite Structures

Traditional approaches for aircraft certification require the assumption of an initial flaw condition, either represented as barely visible impact damage (BVID) or through inclusion of a Teflon insert to serve as surrogate damage. Based on the initial composite damage state, the structure must be shown to demonstrate structural durability and damage tolerance (DaDT) according to the following criteria: a. Damage displays no detrimental growth under cyclic loading b. The structure is able to sustain design limit load (DLL) Currently, the only available manner for validating structural performance is through test. Since damage can occur over a wide variety of areas within a structure, this approach has proven to be increasingly expensive and time consuming for composite airframes and acreage structure within the design-test-certification building block. A further complicating factor is the requirement to accurately capture the most critical damage morphologies as a starting condition. To understand the severity of the damage, it is either required to experimentally determine the most critical areas at tremendous expense or rely on legacy data of similar structural testing, which limits design space expansion. A preferred solution is to use advanced analysis to provide improved understanding of load margins for critical locations based on a wide variety of potential starting damage conditions. The standard industry approach for DaDT certification adheres to the use of the traditional virtual crack closure technique (VCCT) method. VCCT is generally a preferred method because it conforms to the current certification principles of damage from a known flaw, and when used correctly, can be effective at predicting delamination propagation under static and cyclic loading. The VCCT method requires the inclusion of an initial flaw in the finite element (FE) model requiring a-priori knowledge of the flaw location. This in turn requires a plethora of analysis cases to be examined to cover a reasonable span of potential damage states. Additionally, the VCCT approach requires node-to-node connectivity rendering it incompatible with the best practices and approaches for using continuum damage mechanics (CDM) based progressive damage and failure analysis (PDFA) tools within a typical FE solver. Alternatives to VCCT have emerged in the form of cohesive elements which utilize the cohesive zone model (CZM). Unlike VCCT which models linear elastic fracture mechanics, cohesive elements couples continuum and fracture based responses through the use of bilinear traction separation laws. These laws are defined based on a penalty stiffness, a cohesive strength, and a strain energy release rate. The approach can be mesh regularized with native cohesive elements within many FE solvers such as Abaqus and LS-DYNA. In Phase I of the NASA Advanced Composites Consortium (ACC) post-buckled stiffened panel with BVID, Strength and Life [1], the performance of cohesive elements were benchmarked in comparison to VCCT and LEFM solutions and showed good agreement using Abaqus explicit [2]. To realize savings on current and future programs, it is still necessary to close technical gaps related to the use of cohesive elements with Abaqus Standard. Within a program environment, standard finite element analysis is the preferred analytical capability for quasi-static loading as it eliminates uncertainty due to oscillatory behavior commonly seen with explicit analysis. This oscillatory behavior creates difficulties in writing margins of safety based on the analysis. The use of negative tangent stiffness material models complicates convergence which typically requires the use of numerical controls such as viscous damping to overcome. To date, there has not been a comprehensive study on how to establish best practices for cohesive element convergence for predictive capability within the Abaqus implicit solver. In pursuit of these goals, under the NASA ACC program, several numerical benchmark problems were proposed including pure mode I (double cantilevered beam – DCB), pure mode II (end notch flexure – ENF), and symmetric/unsymmetric evolving mixed mode (single leg bend – SLB). This paper focuses on the use of cohesive elements to model the delamination through the use of CZM. Specifically, finite element models for the DCB, ENF, symmetric SLB, and unsymmetric SLB, are developed and various solution controls for convergence are studied to develop a best practice. Once the best practice has been developed, the predictive capability of the objective CZM model is used to analyze the hat pull-off strength of a standard hat stiffened configuration under various loading conditions.

Abaqus↗
DOE OSTI2014

Scalable Nonlinear Solvers for Fully Implicit Coupled Nuclear Fuel Modeling. Final Report

The focus of the project is on the development and customization of some highly scalable domain decomposition based preconditioning techniques for the numerical solution of nonlinear, coupled systems of partial differential equations (PDEs) arising from nuclear fuel simulations. These high-order PDEs represent multiple interacting physical fields (for example, heat conduction, oxygen transport, solid deformation), each is modeled by a certain type of Cahn-Hilliard and/or Allen-Cahn equations. Most existing approaches involve a careful splitting of the fields and the use of field-by-field iterations to obtain a solution of the coupled problem. Such approaches have many advantages such as ease of implementation since only single field solvers are needed, but also exhibit disadvantages. For example, certain nonlinear interactions between the fields may not be fully captured, and for unsteady problems, stable time integration schemes are difficult to design. In addition, when implemented on large scale parallel computers, the sequential nature of the field-by-field iterations substantially reduces the parallel efficiency. To overcome the disadvantages, fully coupled approaches have been investigated in order to obtain full physics simulations.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
NASA NTRS2004

Numerical Study of Flow Past a Circular Cylinder Using RANS, Hybrid RANS/LES and PANS Formulations

Two multiscale type turbulence models are implemented in the PAB3D solver. The models are based on modifying the Reynolds Averaged Navier-Stokes (RANS) equations. The first scheme is a hybrid RANS/LES model utilizing the two-equation (k(sub epsilon)) model with a RANS/LES transition function dependent on grid spacing and the computed turbulence length scale. The second scheme is a modified version of the partially averaged Navier-Stokes (PANS) model, where the unresolved kinetic energy parameter (f(sub k)) is allowed to vary as a function of grid spacing and the turbulence length scale. Solutions from these models are compared to RANS results and experimental data for a stationary and rotating cylinder. The parameter f(sub k) varies between zero and one and has the characteristic to be equal to one in the viscous sub layer, and when the RANS turbulent viscosity becomes smaller than the LES viscosity. The formulation, usage methodology, and validation example are presented to demonstrate the enhancement of PAB3D's time-accurate and turbulence modeling capabilities. The models are compared to RANS results and experimental data for turbulent separated flows (TS) and laminar separated flows (LS) around stationary and rotating cylinders. For a stationary cylinder, the TS case is accurately simulated using the general two-equation k(sub epsilon) turbulence model (eddy-viscosity model). PAB3D accurately predicts the drag coefficient (CD), lift coefficient (CL) and the Strouhal number (St). The LS case was a challenge for the RANS computation with an eddy-viscosity turbulence model. The RANS/LES and PANS performed well and showed marked improvements over the RANS solution. The modified PANS model was the most accurate. For the rotating cylinder, the spin ratio varied from zero to one, and the PANS results were in good agreement with published experimental data. RANS/LES and PANS capture both temporal and spatial fluctuations and produce large-scale structures that do not occur in the RANS simulation. The current results show promise for the capability of PANS in simulating unsteady and complex flow phenomena.

Elmiligui, Alaa A.↗
NASA NTRS2005

Numerical Study of High-Temperature Jet Flow Using RANS/LES and PANS Formulations

Two multi-scale-type turbulence models are implemented in the PAB3D solver. The models are based on modifying the Reynolds Averaged Navier-Stokes (RANS) equations. The first scheme is a hybrid RANS/LES model utilizing the two-equation (k(epsilon)) model with a RANS/LES transition function dependent on grid spacing and the computed turbulence length scale. The second scheme is a modified version of the Partially Averaged Navier-Stokes (PANS) model, where the unresolved kinetic energy parameter (f(sub k)) is allowed to vary as a function of grid spacing and the turbulence length scale. This parameter is estimated based on a novel two-stage procedure to efficiently estimate the level of scale resolution possible for a given flow on a given grid for Partial Averaged Navier-Stokes (PANS). It has been found that the prescribed scale resolution can play a major role in obtaining accurate flow solutions. The parameter f(sub k) varies between zero and one and equal to one in the viscous sub layer, and when the RANS turbulent viscosity becomes smaller than the LES viscosity. The formulation, usage methodology, and validation examples are presented to demonstrate the enhancement of PAB3D's time-accurate and turbulence modeling capabilities. The accurate simulations of flow and turbulent quantities will provide valuable tool for accurate jet noise predictions. Solutions from these models are compared to RANS results and experimental data for high-temperature jet flows. The current results show promise for the capability of hybrid RANS/LES and PANS in simulating such flow phenomena.

Abdol-Hamid, Khaled S.↗
DOE OSTI2022

A BOUT++ extension for full annular tokamak edge MHD and turbulence simulations

For tokamak edge plasma simulation, a plasma simulation framework BOUT++ employs a dual coordinate system to simulate moderate-n and high-n plasma instability with reasonable computational cost, where n is the toroidal mode number. This coordinate system however limits the computational domain to the toroidal wedge (full torus divided into N parts in the toroidal direction) for computational efficiency and the use of flute-ordering approximation in the field solver calculating the flow potential from the vorticity which may not be valid for low-n modes. Improving numerical treatment of low-n modes is however indispensable to address simulations of low-n current-driven edge localized mode (ELM), ELM control by resonant magnetic perturbations (RMPs), edge turbulence with RMPs and so on. In this work, BOUT++ is extended to simulate the interplay between $n=0$, low-n and high-n plasma components in a full annular tokamak edge domain through hybrid modeling of the flow potential and the vorticity. Low-n modes of flow potential are calculated in an orthogonal flux surface coordinate and high-n modes in the dual coordinate system separately in Fourier space. Finally, the proposed scheme can capture an interplay between $n=1$ global modes and high-n turbulence during pedestal collapse in a full annular torus domain with a circular cross section.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2018

Regularized Finite Element Formulations for Shear Band Instabilities in Metals

In my research I have addressed several fundamental issues of dynamic fracture: 1. established the role of heat conduction as a natural regularizing mechanism in shearband problems (i.e. using the correct physics for this problem), 2. developed a unified model to incorporate fracture through a modified phase field formulation, 3. developed stability analysis tools to determine the onset of localization, 4. developed discretization methods and solvers to solve this problem (on parallel computers) that lead to mesh insensitive results and fast convergence. For such difficult problems, the physics and numerics are strongly coupled. To accurately model the physics, one must develop reliable numerics but reliable numerics depends on correct description of the physics. To this end, we have already made significant progress in the field showing that a fully-coupled implicit solution with heat conduction is the key to obtain reliable results for shear bands (see the next sections for more details). This research fits well within the mission of DOE ASCR program on all aspects: from governing equations to discretization to solvers to parallel computers to applications problems. It also relies on the computational tools developed through the SciDAC institutes FASTMath and SDAV, such as PETSc (and SLEPc), ParaView, Trilinos and Masquite.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2001

High-Performance Java Codes for Computational Fluid Dynamics

The computational science community is reluctant to write large-scale computationally -intensive applications in Java due to concerns over Java's poor performance, despite the claimed software engineering advantages of its object-oriented features. Naive Java implementations of numerical algorithms can perform poorly compared to corresponding Fortran or C implementations. To achieve high performance, Java applications must be designed with good performance as a primary goal. This paper presents the object-oriented design and implementation of two real-world applications from the field of Computational Fluid Dynamics (CFD): a finite-volume fluid flow solver (LAURA, from NASA Langley Research Center), and an unstructured mesh adaptation algorithm (2D_TAG, from NASA Ames Research Center). This work builds on our previous experience with the design of high-performance numerical libraries in Java. We examine the performance of the applications using the currently available Java infrastructure and show that the Java version of the flow solver LAURA performs almost within a factor of 2 of the original procedural version. Our Java version of the mesh adaptation algorithm 2D_TAG performs within a factor of 1.5 of its original procedural version on certain platforms. Our results demonstrate that object-oriented software design principles are not necessarily inimical to high performance.

Riley, Christopher↗