SEARCH · Search NASA
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.
This dataset contains simulation data of a twelve turbine wind farm simulation performed with AMR-Wind. The simulation is documented in Kuhn, M. B., Henry de Frahan, M. T., Mohan, P., Deskos, G., Churchfield, M., Cheung, L., ... & Sprague, M. (2025). AMR‐Wind: A Performance‐Portable, High‐Fidelity Flow Solver for Wind Farm Simulations. Wind Energy, 28(5), e70010 (https://doi.org/10.1002/we.70010).
The present work investigates the modeling of preexascale input/output (I/O) workloads of Adaptive Mesh Refinement (AMR) simulations through a simple proxy application. We collect data from the AMReX Castro framework running on the Summit supercomputer for a wide range of scales and mesh partitions for the hydrodynamic Sedov case as a baseline to provide sufficient coverage to the formulated proxy model. The non-linear analysis data production rates are quantified as a function of a set of input parameters such as output frequency, grid size, number of levels, and the Courant-Friedrichs-Lewy (CFL) condition number for each rank, mesh level and simulation time step. Linear regression is then applied to formulate a simple analytical model which allows to translate AMReX inputs into MACSio proxy I/O application parameters, resulting in a simple “kernel” approximation for data production at each time step. Results show that MACSio can simulate actual AMReX nonlinear “static” I/O workloads to a certain degree of confidence on the Summit supercomputer using the present methodology. The goal is to provide an initial level of understanding of AMR I/O workloads via lightweight proxy applications models to facilitate autotune data management strategies in anticipation of exascale systems.
One of the significant causes of residual losses in superconducting radio-frequency (SRF) cavities is trapped magnetic flux. The flux trapping mechanism depends on many factors that include cool-down conditions, surface preparation techniques, and ambient magnetic field orientation. Suitable diagnostic tools are not yet available to quantitatively correlate such factors’ effect on the flux trapping mechanism. A magnetic field scanning system (MFSS) consisting of AMR sensors, fluxgate magnetometers, or Hall probes is recently commissioned to scan the local magnetic field of trapped vortices around 1.3 GHz single-cell SRF cavities. In this contribution, we will present results from sensitivity calibration and the first tests of AMR sensors in the MFSS.
General relativistic magnetohydrodynamic (GRMHD) simulations have revolutionized our understanding of black hole accretion. Here, we present a GPU-accelerated GRMHD code H-AMR with multifaceted optimizations that, collectively, accelerate computation by 2–5 orders of magnitude for a wide range of applications. First, it introduces a spherical grid with 3D adaptive mesh refinement that operates in each of the three dimensions independently. This allows us to circumvent the Courant condition near the polar singularity, which otherwise cripples high-resolution computational performance. Second, we demonstrate that local adaptive time stepping on a logarithmic spherical-polar grid accelerates computation by a factor of ≲10 compared to traditional hierarchical time-stepping approaches. Jointly, these unique features lead to an effective speed of ~10 9 zone cycles per second per node on 5400 NVIDIA V100 GPUs (i.e., 900 nodes of the OLCF Summit supercomputer). We illustrate H-AMR's computational performance by presenting the first GRMHD simulation of a tilted thin accretion disk threaded by a toroidal magnetic field around a rapidly spinning black hole. With an effective resolution of 13,440 × 4608 × 8092 cells and a total of ≲22 billion cells and ~0.65 × 10 8 time steps, it is among the largest astrophysical simulations ever performed. We find that frame dragging by the black hole tears up the disk into two independently precessing subdisks. The innermost subdisk rotation axis intermittently aligns with the black hole spin, demonstrating for the first time that such long-sought alignment is possible in the absence of large-scale poloidal magnetic fields.
Parallel adaptive mesh refinement (AMR) is an important numerical technique that leads to the efficient solution of many physical and engineering problems. In this paper, we describe how AMR programing can be performed in an object-oreinted way using the modern aspects of Fortran 90 combined with the parallelization features of OpenMP.
The PYRAMID parallel unstructured adaptive mesh refinement (AMR) library has been coupled with the GeoFEST geophysical finite element simulation tool to support parallel active tectonics simulations. Specifically, we have demonstrated modeling of coseismic and postseismic surface displacement due to a simulated Earthquake for the Landers system of interacting faults in Southern California. The new software demonstrated a 25-times resolution improvement and a 4-times reduction in time to solution over the sequential baseline milestone case. Simulations on workstations using a few tens of thousands of stress displacement finite elements can now be expanded to multiple millions of elements with greater than 98% scaled efficiency on various parallel platforms over many hundreds of processors. Our most recent work has demonstrated that we can dynamically adapt the computational grid as stress grows on a fault. In this paper, we will describe the major issues and challenges associated with coupling these two programs to create GeoFEST(P). Performance and visualization results will also be described.
A plasma transport model has been implemented in an Eulerian AMR radiation-hydrodynamics code, xRage, which includes plasma viscosity in the momentum tensor, viscous dissipation in the energy equations, and binary species mixing with consistent species mass and energy fluxes driven by concentration gradients, ion and electron baro-diffusion terms and temperature gradient forces. The physics basis, computational issues, numeric options, and results from several test problems are discussed. The transport coefficients are found to be relatively insensitive to the kinetic correction factors when the concentrations are expressed with the molar fractions and the ion mass differences are large. The contributions to flow dynamics from plasma viscosity and mass diffusion were found to increase significantly as scale lengths decrease in an inertial confinement fusion relevant Kelvin-Helmholtz instability mix layer. The mixing scale lengths in the test case are on the order of 100 μm and smaller for viscous effects to appear and 10 μm or less for significant ion species diffusion, evident over durations on the order of nanoseconds. The temperature gradient driven mass flux is seen to deplete a high Z tracer ion at the ion shock front. The plasma transport model provides the generation of the atomic mix per unit of interfacial area between two species with no free parameters. The evolution of the total atomic mix then depends also on an accurate resolution or estimate of the interfacial area between the species mixing by plasma transport. High resolution simulations or a more Lagrangian-like treatment of species interfaces may be required to distinguish plasma transport and numerical diffusion in an Eulerian computation of complex and dynamically evolving mix regions.
PeleLMeX simulates chemically reacting low Mach number flows with block-structured adaptive mesh refinement (AMR). The code is built upon the AMReX library, which provides the underlying data structures and tools to manage and operate on them across massively parallel computing architectures. PeleLMeX algorithmic features are inherited from its predecessor PeleLM but key improvements allow representation of more complex physical processes. Together with its compressible flow counterpart PeleC, the thermo-chemistry library PelePhysics and the multi-physics library PeleMP, it forms the Pele suite of open-source reactive flow simulation codes.
The Ranchero Magnetic Flux Compression Generator (FCG) has been used to create current pulses in the 10-100 MA range for driving both “static” low inductance (0.5 nH) loads1 for generator demonstration purposes and high inductance (10-20 nH) imploding liner loads2 for ultimate use in physics experiments at very high energy density. Simulations of the standard Ranchero generator have recently shown that it had a design issue that could lead to flux trapping in the generator, and a non- robust predictability in its use in high energy density experiments. A re-examination of the design concept for the standard Ranchero generator, prompted by the possible appearance of an aneurism at the output glide plane, has led to a new generation of Ranchero generators designated the RancheroS (for swooped). This generator has removed the problematic output glide plane and replaced it with a region of constantly increasing diameter in the output end of the FCG cavity in which the armature is driven outward under the influence of an additional HE load not present in the original Ranchero. The resultant RancheroS generator, to be tested in LA43S-L13, probably in early FY17, has a significantly increased initial inductance and may be able to drive a somewhat higher load inductance than the standard Ranchero. This report will use the Eulerian AMR code Roxane to study the ability of the new design to drive static loads, with a goal of providing a database corresponding to the load inductances for which the generator might be used and the anticipated peak currents such loads might produce in physics experiments. Such a database, combined with a simple analytic model of an ideal generator, where d(LI)/dt = 0, and supplemented by earlier estimates of losses in actual use of the standard Ranchero, scaled to estimate the increase in losses due to the longer current carrying perimeter in the RancheroS, can then be used to bound the expectations for the current drive one may apply to any load assembly in future experiments.
This document will be shared with LBNL in April 2019. It contains the initial conditions of 4 explosion problems that will test the ability of the 3-Phase AMR code to cope with various cases of 3- phase flows.
A collaborative team from Carnegie Mellon Univ. (CMU), North Carolina State Univ. (NCSU) and Metglas, South Carolina have studied new high speed motors (HSMs) with high-power density for traction motor applications. These are enabled by hybrid designs, including Flux Switching with Permanent Magnets (FSWPM) motors, exploiting permanent magnets without heavy rare earths (RE-lean) and high induction/high resistivity soft magnetic materials that allow for high switching frequencies needed to increase power densities. Team members include Michael E. McHenry, Prof. Materials Science & Eng., CMU, with > 30 years experience in magnetic materials development; Subashish Bhattacharya, Prof. Electrical Eng. and Freedom Center Director at NCSU with > 30 years experience in development of power electronic components and systems and Eric Theisen, Director of Research at Metglas, the only US located supplier of AMR and MANC materials. The team offers novel axial motor architectures exploiting soft magnetic materials (SMMs) that switch with low loss at high frequencies and heavy rare earth free permanent magnets that address materials criticality issues, supply chain risks, and high costs for traction motors. Axial-flux permanent magnet motors (APFM), offer efficiency improvements reducing rotor losses and also significantly higher power density. Axial-flux construction requires less core material, high torque-to-weight ratio. Since AFPM machines have thin magnets, they are smaller than radial flux motors making them attractive in space-limited applications. Noise and vibration are less and planar air gaps are easily adjusted. Flexibility in air-gap direction allows many topologies.
The potential advantages of the Aeromaneuvering Orbit-to-Orbit Shuttle (AMOOS) over the all-propulsive Orbit Transfer Vehicle (OTV) are shown. In particular, the kit concept studies and the dual fueled AMOOS studies show its versatility and option potential over the all-propulsive vehicle. All of this potential of AMOOS and the Aeromaneuvering Recovery System (AMRS) depends upon the ability to control the trajectory during atmospheric flight and so use an ablative TPS. In turn, this TPS must be light weight, which can be attained by spraying a lightweight ablator directly onto the load bearing skin.
We describe the development of a structured adaptive mesh algorithm (AMR) for the Connection Machine-2 (CM-2). We develop a data layout scheme that preserves locality even for communication between fine and coarse grids. On 8K of a 32K machine we achieve performance slightly less than 1 CPU of the Cray Y-MP. We apply our algorithm to an inviscid compressible flow problem.
The increasing on-chip parallelism has some substantial implications for HPC applications. Currently, hybrid programming models (typically MPI+OpenMP) are employed for mapping software to the hardware in order to leverage the hardware?s architectural features. In this paper, we present an approach that automatically introduces thread level parallelism into Chombo, a parallel adaptive mesh refinement framework for finite difference type PDE solvers. In Chombo, core algorithms are specified in the ChomboFortran, a macro language extension to F77 that is part of the Chombo framework. This domain-specific language forms an already used target language for an automatic migration of the large number of existing algorithms into a hybrid MPI+OpenMP implementation. It also provides access to the auto-tuning methodology that enables tuning certain aspects of an algorithm to hardware characteristics. Performance measurements are presented for a few of the most relevant kernels with respect to a specific application benchmark using this technique as well as benchmark results for the entire application. The kernel benchmarks show that, using auto-tuning, up to a factor of 11 in performance was gained with 4 threads with respect to the serial reference implementation.
The hydrodynamic evolution of the common-envelope (CE) phase of a low-mass binary composed of a 1.05 M{sub Sun} red giant and a 0.6 M{sub Sun} companion has been followed for five orbits of the system using a high-resolution method in three spatial dimensions. During the rapid inspiral phase, the interaction of the companion with the red giant's extended atmosphere causes about 25% of the CE to be ejected from the system, with mass continuing to be lost at the end of the simulation at a rate {approx}2 M{sub Sun} yr{sup -1}. In the process the resulting loss of angular momentum and energy reduces the orbital separation by a factor of seven. After this inspiral phase the eccentricity of the orbit rapidly decreases with time. The gravitational drag dominates hydrodynamic drag at all times in the evolution, and the commonly used Bondi-Hoyle-Lyttleton prescription for estimating the accretion rate onto the companion significantly overestimates the true rate. On scales comparable to the orbital separation, the gas flow in the orbital plane in the vicinity of the two cores is subsonic with the gas nearly corotating with the red giant core and circulating about the red giant companion. On larger scales, 90% of the outflow is contained within 30 Degree-Sign of the orbital plane, and the spiral shocks in this material leave an imprint on the density and velocity structure. Of the energy released by the inspiral of the cores, only about 25% goes toward ejection of the envelope.
Magnetic fields are important contributors to the dynamics of collapsing molecular cloud cores, and can have a major effect on whether collapse results in a single protostar or fragmentation into a binary or multiple protostar system. New models are presented of the collapse of magnetic cloud cores using the adaptive mesh refinement code Enzo2.0. The code was used to calculate the ideal magnetohydrodynamics (MHD) of initially spherical, uniform density, and rotation clouds with density perturbations, i.e., the Boss and Bodenheimer standard isothermal test case for three-dimensional (3D) hydrodynamics codes. After first verifying that Enzo reproduces the binary fragmentation expected for the non-magnetic test case, a large set of models was computed with varied initial magnetic field strengths and directions with respect to the cloud core axis of rotation (parallel or perpendicular), density perturbation amplitudes, and equations of state. Three significantly different outcomes resulted: (1) contraction without sustained collapse, forming a denser cloud core; (2) collapse to form a single protostar with significant spiral arms; and (3) collapse and fragmentation into binary or multiple protostar systems, with multiple spiral arms. Comparisons are also made with previous MHD calculations of similar clouds with a barotropic equations of state. These results for the collapse of initially uniform density spheres illustrate the central importance of both magnetic field direction and field strength for determining the outcome of dynamic protostellar collapse.
We present the results of a large suite of three-dimensional models of the collapse of magnetic molecular cloud cores using the adaptive mesh refinement code Enzo2.2 in the ideal magnetohydrodynamics approximation. The cloud cores are initially either prolate or oblate, centrally condensed clouds with masses of 1.73 or 2.73 M {sub ☉}, respectively. The radial density profiles are Gaussian, with central densities 20 times higher than boundary densities. A barotropic equation of state is used to represent the transition from low density isothermal phases, to high density optically thick phases. The initial magnetic field strength ranges from 6.3 to 100 μG, corresponding to clouds that are strongly to marginally supercritical, respectively, in terms of the mass to magnetic flux ratio. The magnetic field is initially uniform and aligned with the clouds' rotation axes, with initial ratios of rotational to gravitational energy ranging from 10{sup –4} to 0.1. Two significantly different outcomes for collapse result: (1) formation of single protostars with spiral arms, and (2) fragmentation into multiple protostar systems. The transition between these two outcomes depends primarily on the initial magnetic field strength, with fragmentation occurring for mass to flux ratios greater than about 14 times the critical ratio for prolate clouds. Oblate clouds typically fragment into several times more clumps than prolate clouds. Multiple, rather than binary, system formation is the general rule in either case, suggesting that binary stars are primarily the result of the orbital dissolution of multiple protostar systems.
We have implemented non-ideal magnetohydrodynamics (MHD) effects in the adaptive mesh refinement code RAMSES, namely, ambipolar diffusion and Ohmic dissipation, as additional source terms in the ideal MHD equations. We describe in details how we have discretized these terms using the adaptive Cartesian mesh, and how the time step is diminished with respect to the ideal case, in order to perform a stable time integration. We have performed a large suite of test runs, featuring the Barenblatt diffusion test, the Ohmic diffusion test, the C-shock test, and the Alfven wave test. For the latter, we have performed a careful truncation error analysis to estimate the magnitude of the numerical diffusion induced by our Godunov scheme, allowing us to estimate the spatial resolution that is required to address non-ideal MHD effects reliably. We show that our scheme is second-order accurate, and is therefore ideally suited to study non-ideal MHD effects in the context of star formation and molecular cloud dynamics.