Search NASA⌕ Search

SEARCH · Search NASA

Results for “distributed parallelization”

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 1,585 records · Page 88

Fabrication and Tests of a Resistive-Type Superconducting Fault Current Limiter Module Based on Coated Conductors

The resistivetype superconducting fault current limiter (R-SFCL) is simple in structure, near zero impedance at normal state and can be quickly triggered by the fault current. Thus, it is a possible solution to limit the fault current in the power grid. In this paper, we report the fabrication and test for an R-SFCL module The R-SFCL module consists of seven bifilar coils connected in parallel, which were wound with two parallel 2G tapes. A 5mm spacer was co-wound between adjacent turns to act as the highvoltage insulation. In the tests, we test the current distribution among the bifilar coils, and then simulated short-circuit tests were conducted on the module to obtain its performance under fault current surge.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Scale-dependent spatial variabilities of hydrological exchange flows and transit time in a large regulated river

Hydrological exchange flows (HEF) across the river-aquifer interface and the associated residence time of river water in the aquifer have important implications for contaminant plume migration and biogeochemical processes in the river corridor. HEFs and residence time are influenced by both subsurface physical features and hydrologic forcing related to the transport process, which can exhibit complex spatial and temporal variations. In this study, we used a massively parallel subsurface flow model and a particle-tracking model to study the influences of different control factors on spatial variability of HEFs and residence time distributions (RTD) in the Hanford Reach of the Columbia River in Washington State. A total number of 100M particles were randomly injected in time and space and then tracked in a model domain that covers a 51-km 2 area (15.1M model cells). We used hourly river stages and groundwater levels to drive the model to provide dynamic velocity fields for the particle tracking in the simulation period that was longer than 2 years. The groundwater flow simulation and particle-tracking results provide the first comprehensive assessment of the spatial distribution of HEFs and residence time in large complex river corridors. Overall, our results show that the aquifer hydrogeological structure has the strongest correlation with the extent and magnitude of exchange flux. The residence time exhibits complex patterns that are impacted by all the river geomorphologic, hydrodynamic, and hydrogeologic factors and are strongly correlated with the downwelling ratio of exchange flux. The new insights gained through this study can be used to support the development of reduced-order models of HEFs and RTDs for large complex river systems.

54 ENVIRONMENTAL SCIENCES↗

Phase transformation of Ca{sub 4}[Al{sub 6}O{sub 12}]SO{sub 4} and its disordered crystal structure at 1073 K

The phase transformation of Ca{sub 4}[Al{sub 6}O{sub 12}]SO{sub 4} and the crystal structure of its high-temperature phase were investigated by differential thermal analysis, temperature-dependent Raman spectroscopy and high-temperature X-ray powder diffraction (CuKα{sub 1}). We determined the starting temperature of the orthorhombic-to-cubic transformation during heating (=711 K) and that of the reverse transformation during cooling (=742 K). The thermal hysteresis was negative (=−31 K), suggesting the thermoelasticity of the transformation. The space group of the high temperature phase is I4{sup ¯}3m with the unit-cell dimensions of a=0.92426(2) nm and V=0.78955(2) nm{sup 3} (Z=2) at 1073 K. The initial structural model was derived by the direct methods and further refined by the Rietveld method. The final structural model showed the orientational disordering of SO{sub 4} tetrahedra. The maximum-entropy method-based pattern fitting method was used to confirm the validity of the split-atom model, in which conventional structure bias caused by assuming intensity partitioning was minimized. At around the transformation temperature during heating, the vibrational spectra, corresponding to the Raman-active SO{sub 4} internal stretching mode, showed the continuous and gradual change in the slope of full width at half maximum versus temperature curve. This strongly suggests that the orthorhombic-to-cubic phase transformation would be principally accompanied by the statistical disordering in orientation of the SO{sub 4} tetrahedra, without distinct dynamical reorientation. - Graphical abstract: (Left) Three-dimensional electron-density distributions of the SO{sub 4} tetrahedron with the split-atom model, and (right) a bird's eye view of electron densities on the plane parallel to (111). - Highlights: • Crystal structure of Ca{sub 4}[Al{sub 6}O{sub 12}]SO{sub 4} at 1073 K is determined by powder XRD. • The atom arrangements are represented by the split-atom model. • The MPF method is used to confirm the validity of the model. • The phase transition is accompanied by orientational disordering of SO{sub 4} tetrahedra.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Episodic Earthquake Swarms in the Mineral Mountains, Utah Driven by the Roosevelt Hydrothermal System

Over 1,000 earthquakes (-2.0 < M < 2.0), identified using a matched-filter method, occurred in the Mineral Mountains, Utah between 2016 and 2019. The enhanced catalog is complete down to M -0.9 and contains roughly 15 times more events than originally cataloged. Earthquake relocation of ~800 earthquakes shows that activity is concentrated in a <2 km long E-W striking narrow zone, ~4 km east of the Roosevelt hydrothermal system. Two fault orientations, both N-S and E-W parallel to the Opal Mound and Mag Lee faults, respectively, are observed after computing composite focal mechanisms of highly similar earthquakes. Looking solely at the temporal distribution of the seismicity, we identify 15 periods of swarm-like activity, with two major clusters occurring in December 2016, recorded by three stations, and in October 2019 recorded by eight stations. The October 2019 swarm, the best recorded sequence in the area, provides evidence for the underlying triggering mechanism. We show that a complex mechanism of fluid diffusion and aseismic slip is responsible for the swarm evolution with migration velocities reaching 10 km/day. We hypothesize that these episodic swarms in the Mineral Mountains are primarily driven by migrating fluids that originate within the Roosevelt hydrothermal system.

58 GEOSCIENCES↗

Learning local equivariant representations for large-scale atomistic dynamics

Abstract A simultaneously accurate and computationally efficient parametrization of the potential energy surface of molecules and materials is a long-standing goal in the natural sciences. While atom-centered message passing neural networks (MPNNs) have shown remarkable accuracy, their information propagation has limited the accessible length-scales. Local methods, conversely, scale to large simulations but have suffered from inferior accuracy. This work introduces Allegro, a strictly local equivariant deep neural network interatomic potential architecture that simultaneously exhibits excellent accuracy and scalability. Allegro represents a many-body potential using iterated tensor products of learned equivariant representations without atom-centered message passing. Allegro obtains improvements over state-of-the-art methods on QM9 and revMD17. A single tensor product layer outperforms existing deep MPNNs and transformers on QM9. Furthermore, Allegro displays remarkable generalization to out-of-distribution data. Molecular simulations using Allegro recover structural and kinetic properties of an amorphous electrolyte in excellent agreement with ab-initio simulations. Finally, we demonstrate parallelization with a simulation of 100 million atoms.

74 ATOMIC AND MOLECULAR PHYSICS↗

Final Stage in the Design of a Boron Neutron Capture Therapy facility at CEADEN, Cuba

A neutron beam simulation study is carried out to determine the most suitable neutron energy for treatment of shallow and deep-seated brain tumors in the context of Boron Neutron Capture Therapy (BNCT). Two figures-of-merit, the therapeutic gain and the neutron fluence are utilized as beam assessment parameters. An irradiation cavity is used instead of a parallel beam port for the therapy. Calculations are performed using the MCNP5 code. After the optimization of our beam-shaper a study of the dose distribution in the head, neck, tyroids, lungs and upper and middle spine had been made. The therapeutic gain is increased while the current required for one hour treatment is decreased in comparison with the trading prototypes of NG used for BNCT.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

The effect of charge separation on nonlinear electrostatic waves in a magnetized dusty plasma with two-temperature ions

In view of the observations of parallel (to Earth's magnetic field) spiky electric field structures by the FAST satellite, a theoretical study is conducted using a dusty plasma model comprising Boltzmann distributed hot and cool ions, Boltzmann electrons and a negatively charged cold dust fluid to investigate the existence of similar low frequency nonlinear electrostatic waves in a dusty plasma which could have a similar appearance as the observed waveforms. Charge separation effects are incorporated into our model by the inclusion of Poisson's equation as opposed to assuming quasineutrality. The system of nonlinear equations is then numerically solved. The resulting electric field structure is examined as a function of various plasma parameters such as Mach number, driving electric field amplitude, bulk dust drift speed, particle densities and particle temperatures.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Kinetic simulation of capacitively coupled plasmas driven by trapezoidal asymmetric voltage pulses

A kinetic Particle-In-Cell simulation with Monte Carlo Collisions was performed of a geometrically symmetric capacitively coupled, parallel-plate discharge in argon, driven by trapezoidal asymmetric voltage pulses with a period of 200 ns. The discharge was electrically asymmetric, making the ion energy distributions at the two electrodes different from one another. The fraction of the period (α), during which the voltage was kept at a constant (top-flat) positive value, was a critical control parameter. For the parameter range investigated, as α increased, the mean ion energy on the grounded electrode increased and the ions became more directional, whereas the opposite was found for the ions striking the powered electrode. The absolute value of the DC self-bias voltage decreased as α increased. Plasma instabilities, promoted by local double layers and electric field reversals during the time of the positive voltage excursion, were characterized by electron plasma waves launched from the sheath edge.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

New compact hohlraum configuration research at the 1.7 MA Z-pinch generator

A new compact Z-pinch x-ray hohlraum design with parallel-driven x-ray sources was experimentally demonstrated in a full configuration with a central target and tailored shine shields (to provide a symmetric temperature distribution on the target) at the 1.7 MA Zebra generator. This presentation reports on the joint success of two independent lines of research. One of these was the development of new sources – planar wire arrays (PWAs). PWAs turned out to be a prolific radiator. Another success was the drastic improvement in energy efficiency of pulsed-power systems, such as the Load Current Multiplier (LCM). The Zebra/LCM generator almost doubled the plasma load current to 1.7 MA. The two above-mentioned innovative approaches were used in combination to produce a new compact hohlraum design for ICF, as jointly proposed by SNL and UNR. Good agreement between simulated and measured radiation temperature of the central target is shown. Experimental comparison of PWAs with planar foil liners (PFL) - another viable alternative to wire array loads at multi-MA generators show promising data. Results of research at the University of Nevada Reno allowed for the study of hohlraum coupling physics at University-scale generators. The advantages of new hohlraum design applications for multi-MA facilities with W or Au double PWAs or PFL x-ray sources are discussed.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Effect of thermionic cathode heating current self-magnetic field on gaseous plasma generator characteristics

The performance capabilities of the PINK, a plasma generator with a thermionic cathode mounted in the cavity of a hollow cathode, depending for its operation on a non-self-sustained low-pressure gas discharge have been investigated. It has been shown that when a single-filament tungsten cathode 2 mm in diameter is used and the peak filament current is equal to or higher than 100 A, the self-magnetic field of the filament current significantly affects the discharge current and voltage waveforms. This effect is due to changes in the time and space distributions of the emission current density from the hot cathode. When the electron mean free path is close to the characteristic dimensions of the thermionic cathode, the synthesized plasma density distribution is nonuniform and the cathode is etched nonuniformly. The cathode lifetime in this case is 8–12 h. Using a cathode consisting of several parallel-connected tungsten filaments ∼0.8 mm in diameter moderates the effect of the self-magnetic field of the filament current and nearly doubles the cathode lifetime. The use of this type of cathode together with a discharge igniting electrode reduces the minimum operating pressure in the plasma generator to about one third of that required for the generator operation with a single-filament cathode (to 0.04 Pa)

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

SCATTER-DOMINATED INTERPLANETARY TRANSPORT OF SOLAR ENERGETIC PARTICLES IN LARGE GRADUAL EVENTS AND THE FORMATION OF DOUBLE POWER-LAW DIFFERENTIAL FLUENCE SPECTRA OF GROUND-LEVEL EVENTS DURING SOLAR CYCLE 23

The effects of scatter-dominated interplanetary transport on the spectral properties of the differential fluence of large gradual solar energetic particle (SEP) events are investigated analytically. The model assumes for simplicity radial constant solar wind and radial magnetic field. The radial diffusion coefficient is calculated with quasilinear theory by assuming a spectrum of Alfvén waves propagating parallel to the magnetic field. Cross-field transport is neglected. The model takes into consideration several essential features of gradual event transport: nearly isotropic ion distributions, adiabatic deceleration in a divergent solar wind, and particle radial scattering mean free paths increasing with energy. Assuming an impulsive and spherically symmetric injection of SEPs with a power-law spectrum near the Sun, the predicted differential fluence spectrum exhibits at 1 AU three distinctive power laws for different energy domains. The model naturally reproduces the spectral features of the double power-law proton differential fluence spectra that tend to be observed in extremely large SEP events. We select nine western ground-level events (GLEs) out of the 16 GLEs during Solar Cycle 23 and fit the observed double power-law spectra to the analytical predictions. The compression ratio of the accelerating shock wave, the power-law index of the ambient wave intensity, and the proton radial scattering mean free path are determined for the nine GLEs. The derived parameters are generally in agreement with the characteristic values expected for large gradual SEP events.

79 ASTRONOMY AND ASTROPHYSICS↗

Large parallel cosmic string simulations: New results on loop production

Using a new parallel computing technique, we have run the largest cosmic string simulations ever performed. Our results confirm the existence of a long transient period where a nonscaling distribution of small loops is produced at lengths depending on the initial correlation scale. As time passes, this initial population gives way to the true scaling regime, where loops of size approximately equal to one-twentieth the horizon distance become a significant component. We observe similar behavior in matter and radiation eras, as well as in flat space. In the matter era, the scaling population of large loops becomes the dominant component; we expect this to eventually happen in the other eras as well.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Population Mixing in Asymmetric Magnetic Reconnection with a Guide Field

We investigate how population mixing leads to structured electron distribution functions in asymmetric guide-field magnetic reconnection based on particle-in-cell simulations. The change of magnetic connectivity patches populations from different inflow regions to form multicomponent distributions in the exhaust, illustrating the direct consequence of the breaking and rejoining of magnetic flux tubes. Finite Larmor radius (FLR) effects of electrons accelerated by the perpendicular electric fields result in crescent-type nongyrotropic distributions. A new type of nongyrotropy is found to be caused by the combined effects of the FLR and velocity dispersion of electrons accelerated by the parallel electric field. The patching together of populations and the effects of acceleration and the FLR form the first steps of mixing in the exhaust and separatrix regions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

iez: Resource Contention Aware Load Balancing for Large-Scale Parallel File Systems

Parallel I/O performance is crucial to sustaining scientific applications on large-scale High-Performance Computing (HPC) systems. However, I/O load imbalance in the underlying distributed and shared storage systems can significantly reduce overall application performance. There are two conflicting challenges to mitigate this load imbalance: (i) optimizing systemwide data placement to maximize the bandwidth advantages of distributed storage servers, i.e., allocating I/O resources efficiently across applications and job runs; and (ii) optimizing client-centric data movement to minimize I/O load request latency between clients and servers, i.e., allocating I/O resources efficiently in service to a single application and job run. Moreover, existing approaches that require application changes limit wide-spread adoption in commercial or proprietary deployments. We propose iez, an “end-to-end control plane” where clients transparently and adaptively write to a set of selected I/O servers to achieve balanced data placement. Our control plane leverages realtime load information for distributed storage server global data placement while our design model leverages trace-based optimization techniques to minimize I/O load request latency between clients and servers. We evaluate our proposed system on an experimental cluster for two common use cases: synthetic I/O benchmark IOR for large sequential writes and a scientific application I/O kernel, HACC-I/O. Results show read and write performance improvements of up to 34% and 32%, respectively, compared to the state of the art.

Wadhwa, Bharti↗

Costa - Introduction to 2015 Annual Report

In parallel with Sandia National Laboratories having two major locations (NM and CA), along with a number of smaller facilities across the nation, so too is the distribution of scientific, engineering and computing resources. As a part of Sandia’s Institutional Computing Program, CA site-based Sandia computer scientists and engineers have been providing mission and research staff with local CA resident expertise on computing options while also focusing on two growing high performance computing research problems. The first is how to increase system resilience to failure, as machines grow larger, more complex and heterogeneous. The second is how to ensure that computer hardware and configurations are optimized for specialized data analytical mission needs within the overall Sandia computing environment, including the HPC subenvironment. All of these activities support the larger Sandia effort in accelerating development and integration of high performance computing into national security missions. Sandia continues to both promote national R&D objectives, including the recent Presidential Executive Order establishing the National Strategic Computing Initiative and work to ensure that the full range of computing services and capabilities are available for all mission responsibilities, from national security to energy to homeland defense.

97 MATHEMATICS AND COMPUTING↗

Evaluation of Deep Subsurface Resistivity Imaging for Hydrofracture Monitoring

This report describes the results of the first of its kind monitoring of a hydrofracture operation with electromagnetic measurements. The researchers teamed with oil and gas producer Encana Corporation to design and execute a borehole to surface monitoring of three fracture stages at a well pad in central Colorado. The field project consisted of an equipment upgrade, a survey design and modeling phase, several weeks of data collection, and data processing and interpretation. Existing Depth to Surface Resistivity (DSR) instrumentation was upgraded to allow for continuous high precision recording from downhole sources. The full system can now collect data continuously for up to 72 hours, which is sufficient to measure data for 10 frac stages. Next we used numerical modeling and frac treatment data supplied by Encana to design a field survey to detect EM signal from induced fractures. Prior to modeling we developed a novel technique for using well casing as an antenna for a downhole source. Modeling shows that 1) a measurable response for an induced fracture could be achieved if the facture fluid was of high salinity 2) an optimum fracture response is created when the primary source field is parallel to the well casing but perpendicular to the fracture direction. In mid-July, 2014 we installed an array of more than 100 surface sensors, distributed above the treatment wells and extending for approximately 1 km north and 750 m eastward. We applied a 0.6 Hz square wave signal to a downhole current electrode located in a horizontal well 200 m offset from the treatment well with a return electrode on the surface. The data were transmitted to a recording trailer via Wi-Fi where we monitored receiver and transmitter channels continuously in a 72-hour period which covered 7 frac stages, three of which were high salinity. Although the background conditions were very noisy we were able to extract a clear signal from the high salinity stages. Initial data interpretation attempts consisting of trying to directly invert collected data using a simple background starting model did not produce reasonable models. We next used a simulated data set to develop a constrained inversion workflow that places the downhole fracture in the correct location. Finally, we used a combination of forward and inverse models to fit the collected data to a model that incorporated 3 frac stages. This project provided a first-of-its-kind application of EM technology to map an induced fracture. The results demonstrated a clear anomaly during the operation that can be fit to a reasonable model of an induced fracture.

42 ENGINEERING↗

Monte Carlo Simulations Of The Dose Distributions From Carbon Microbeams Used In An Experimental Radiation Therapy Method

Recent upgrades of the MCNPX Monte Carlo code include transport of heavy ions. We employed the new code to simulate the energy and dose distributions produced by carbon beams in rabbit's head in and around a brain tumor. The work was within our experimental technique of interlaced carbon microbeams, which uses two 90 deg. arrays of parallel, thin planes of carbon beams (microbeams) interlacing to produce a solid beam at the target. A similar version of the method was earlier developed with synchrotron-generated x-ray microbeams. We first simulated the Bragg peak in high density polyethylene and other materials, where we could compare the calculated carbon energy deposition to the measured data produced at the NASA Space Radiation Laboratory (NSRL) at Brookhaven National Laboratory (BNL). The results showed that new MCNPX code gives a reasonable account of the carbon beam's dose up to {approx}200 MeV/nucleon beam energy. At higher energies, which were not relevant to our project, the model failed to reproduce the Bragg-peak's extent of increasing nuclear breakup tail. In our model calculations we determined the dose distribution along the beam path, including the angular straggling of the microbeams, and used the data for determining the optimal values of beam spacing in the array for producing adequate beam interlacing at the target. We also determined, for the purpose of Bragg-peak spreading at the target, the relative beam intensities of the consecutive exposures with stepwise lower beam energies, and simulated the resulting dose distribution in the spread out Bragg-peak. The details of the simulation methods used and the results obtained are presented.

62 RADIOLOGY AND NUCLEAR MEDICINE↗