Search NASA⌕ Search

SEARCH · Search NASA

Results for “distributed algorithm”

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 1,729 records · Page 96

ED-cPSD: Fast Phase-Size Distribution via Sequential Erosion-Dilation

The Erosion-Dilation continuous Phase-Size Distribution, ED-cPSD, is an application for calculating continuous pore and particle-size distribution from digital reconstructions and/or image-based structural data. It is based on the erosion-dilation continuous phase-size distribution method. A continuous size distribution is a measure of the probability density of finding a particle or pore of a certain size. These distributions are of interest in any field of study involving porous media, including but not limited to electrochemistry, petroleum engineering, geology, and food science. The algorithm behind the software provides a computationally efficient way to calculate phase-size distributions for large domains. For a 3D battery electrode reconstruction with 1.3 x 10 8 voxels, the particle size distribution is derived in under 2 min on a desktop, while also retaining flexibility and computational efficiency for HPC-scale multi-threading. The software can handle structures with over 10 9 voxels. The algorithm is roughly 280 times faster than a previous version on the same task.

Characterization↗

Understanding performance variability in standard and pipelined parallel Krylov solvers

In this work, we collect data from runs of Krylov subspace methods and pipelined Krylov algorithms in an effort to understand and model the impact of machine noise and other sources of variability on performance. We find large variability of Krylov iterations between compute nodes for standard methods that is reduced in pipelined algorithms, directly supporting conjecture, as well as large variation between statistical distributions of runtimes across iterations. Based on these results, we improve upon a previously introduced nondeterministic performance model by allowing iterations to fluctuate over time. We present our data from runs of various Krylov algorithms across multiple platforms as well as our updated non-stationary model that provides good agreement with observations. We also suggest how it can be used as a predictive tool.

97 MATHEMATICS AND COMPUTING↗

The Suomi National Polar-Orbiting Partnership (SNPP): Continuing NASA Research and Applications

The Suomi National Polar-orbiting Partnership (SNPP) satellite was successfully launched into a polar orbit on October 28, 2011 carrying 5 remote sensing instruments designed to provide data to improve weather forecasts and to increase understanding of long-term climate change. SNPP provides operational continuity of satellite-based observations for NOAA's Polar-orbiting Operational Environmental Satellites (POES) and continues the long-term record of climate quality observations established by NASA's Earth Observing System (EOS) satellites. In the 2003 to 2011 pre-launch timeframe, NASA's SNPP Science Team assessed the adequacy of the operational Raw Data Records (RDRs), Sensor Data Records (SDRs), and Environmental Data Records (EDRs) from the SNPP instruments for use in NASA Earth Science research, examined the operational algorithms used to produce those data records, and proposed a path forward for the production of climate quality products from SNPP. In order to perform these tasks, a distributed data system, the NASA Science Data Segment (SDS), ingested RDRs, SDRs, and EDRs from the NOAA Archive and Distribution and Interface Data Processing Segments, ADS and IDPS, respectively. The SDS also obtained operational algorithms for evaluation purposes from the NOAA Government Resource for Algorithm Verification, Independent Testing and Evaluation (GRAVITE). Within the NASA SDS, five Product Evaluation and Test Elements (PEATEs) received, ingested, and stored data and performed NASA's data processing, evaluation, and analysis activities. The distributed nature of this data distribution system was established by physically housing each PEATE within one of five Climate Analysis Research Systems (CARS) located at either at a NASA or a university institution. The CARS were organized around 5 key EDRs directly in support of the following NASA Earth Science focus areas: atmospheric sounding, ocean, land, ozone, and atmospheric composition products. The PEATES provided the system level interface with members of the NASA SNPP Science Team and other science investigators within each CARS. A sixth Earth Radiation Budget CARS was established at NASA Langley Research Center (NASA LaRC) to support instrument performance, data evaluation, and analysis for the SNPP Clouds and the Earth's Radiant Budget Energy System (CERES) instrument. Following the 2011 launch of SNPP, spacecraft commissioning, and instrument activation, the NASA SNPP Science Team evaluated the operational RDRs, SDRs, and EDRs produced by the NOAA ADS and IDPS. A key part in that evaluation was the NASA Science Team's independent processing of operational RDRs and SDRs to EDRs using the latest NASA science algorithms. The NASA science evaluation was completed in the December 2012 to April 2014 timeframe with the release of a series of NASA Science Team Discipline Reports. In summary, these reports indicated that the RDRs produced by the SNPP instruments were of sufficiently high quality to be used to create data products suitable for NASA Earth System science and applications. However, the quality of the SDRs and EDRs were found to vary greatly when considering suitability for NASA science. The need for improvements in operational algorithms, adoption of different algorithmic approaches, greater monitoring of on-orbit instrument calibration, greater attention to data product validation, and data reprocessing were prominent findings in the reports. In response to these findings, NASA, in late 2013, directed the NASA SNPP Science Team to use SNPP instrument data to develop data products of sufficiently high quality to enable the continuation of EOS time series data records and to develop innovative, practical applications of SNPP data. This direction necessitated a transition of the SDS data system from its pre-launch assessment mode to one of full data processing and production. To do this, the PEATES, which served as NASA's data product testing environment during the prelaunch and early on-orbit periods, were transitioned to Science Investigator-led Processing Systems (SIPS). The distributed data architecture was maintained in this new system by locating the SIPS at the same institutions at which the CARS and PEATES were located. The SIPS acquire raw SNPP instrument Level 0 (i.e. RDR) data over the full SNPP mission from the NOAA ADS and IDPS through the NASA SDS Data Distribution and Depository Element (SD3E). The SIPS process those data into NASA Level 1, Level 2, and global, gridded Level 3 standard products using peer-reviewed algorithms provided by members of the NASA Science Team. The SIPS work with the NASA SNPP Science Team in obtaining enhanced, refined, or alternate real-time algorithms to support the capabilities of the Direct Readout Laboratory (DRL). All data products, algorithm source codes, coefficients, and auxiliary data used in product generation are archived in an assigned NASA Distributed Active Archive Center (DAAC).

remote sensing↗

Results on the neutron energy distribution measurements at the RECH-1 Chilean nuclear reactor

Neutron activations experiments has been perform at the RECH-1 Chilean Nuclear Reactor to measure its neutron flux energy distribution. Samples of pure elements was activated to obtain the saturation activities for each reaction. Using - ray spectroscopy we identify and measure the activity of the reaction product nuclei, obtaining the saturation activities of 20 reactions. GEANT4 and MCNP was used to compute the self shielding factor to correct the cross section for each element. With the Expectation-Maximization algorithm (EM) we were able to unfold the neutron flux energy distribution at dry tube position, near the RECH-1 core. In this work, we present the unfolding results using the EM algorithm.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

A simple algorithm for beam profile diagnostics using a thermographic camera

A new algorithm for digital image processing apparatuses is developed to evaluate profiles of high-intensity DC beams from temperature images of irradiated thin foils. Numerical analyses are performed to examine the reliability of the algorithm. To simulate the temperature images acquired by a thermographic camera, temperature distributions are numerically calculated for 20 MeV proton beams with different parameters. Noise in the temperature images which is added by the camera sensor is also simulated to account for its effect. Using the algorithm, beam profiles are evaluated from the simulated temperature images and compared with exact solutions. We find that niobium is an appropriate material for the thin foil used in the diagnostic system. We also confirm that the algorithm is adaptable over a wide beam current range of 0.11–214 μA, even when employing a general-purpose thermographic camera with rather high noise (ΔT{sub NETD} ≃ 0.3 K; NETD: noise equivalent temperature difference)

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Search for collective phenomena in hadron interactions

New results of the search for collective phenomena have been obtained and analyzed in the present report. The experimental studies are carried out on U-70 accelerator of IHEP in Protvino. It is suggested that these phenomena can be discovered at the energy range of 50-70 GeV in the extreme multiplicity region since the high-density matter can form in this very region. The collective behavior of secondary particles is considered to manifest itself in the Bose-Einstein condensation of pions, Vavilov-Cherenkov gluon radiation, excess of soft-photon yield, and other unique phenomena. The perceptible peak in the angular distribution has been revealed. It was interpreted as the gluon radiation and so the parton matter refraction index was determined. The new software was designed for the track reconstruction based on Kalman Filter technique. This algorithm allows one to estimate more precisely the track parameters (especially momentum). The search for Bose-Einstein condensation can be continued by using the selected events with the multiplicity of more than eight charged particles. The gluon dominance model predictions have shown good agreement with the multiplicity distribution at high multiplicity and confirmed the guark-gluon medium formation under these conditions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

SU-E-T-205: Improving Quality Assurance of HDR Brachytherapy: Verifying Agreement Between Planned and Delivered Dose Distributions Using DICOM RTDose and Advanced Film Dosimetry

Purpose: HDR brachytherapy is undergoing significant development, and quality assurance (QA) checks must keep pace. Current recommendations do not adequately verify delivered against planned dose distributions: This is particularly relevant for new treatment planning system (TPS) calculation algorithms (non TG-43 based), and an era of significant patient-specific plan optimisation. Full system checks are desirable in modern QA recommendations, complementary to device-centric individual tests. We present a QA system incorporating TPS calculation, dose distribution export, HDR unit performance, and dose distribution measurement. Such an approach, more common in external beam radiotherapy, has not previously been reported in the literature for brachytherapy. Methods: Our QA method was tested at 24 UK brachytherapy centres. As a novel approach, we used the TPS DICOM RTDose file export to compare planned dose distribution with that measured using Gafchromic EBT3 films placed around clinical brachytherapy treatment applicators. Gamma analysis was used to compare the dose distributions. Dose difference and distance to agreement were determined at prescription Point A. Accurate film dosimetry was achieved using a glass compression plate at scanning to ensure physically-flat films, simultaneous scanning of known dose films with measurement films, and triple-channel dosimetric analysis. Results: The mean gamma pass rate of RTDose compared to film-measured dose distributions was 98.1% at 3%(local), 2 mm criteria. The mean dose difference, measured to planned, at Point A was -0.5% for plastic treatment applicators and -2.4% for metal applicators, due to shielding not accounted for in TPS. The mean distance to agreement was 0.6 mm. Conclusion: It is recommended to develop brachytherapy QA to include full-system verification of agreement between planned and delivered dose distributions. This is a novel approach for HDR brachytherapy QA. A methodology using advanced film dosimetry and gamma comparison to DICOM RTDose files has been demonstrated as suitable to fulfil this need.

07 ISOTOPE AND RADIATION SOURCES↗

Neural Networks and Graph Algorithms with Next-Generation Processors

The use of graphical processors for distributed computation revolutionized the field of high performance scientific computing. As the Moore's Law era of computing draws to a close, the development of non-Von Neumann systems: neuromorphic processing units, and quantum annealers; again are redefining new territory for computational methods. While these technologies are still in their nascent stages, we discuss their potential to advance computing in two domains: machine learning, and solving constraint satisfaction problems. Each of these processors utilize fundamentally different theoretical models of computation. This raises questions about how to best use them in the design and implementation of applications. While many processors are being developed with a specific domain target, the ubiquity of spin-glass models and neural networks provides an avenue for multi-functional applications. This provides hints at the future infrastructure needed to integrate many next-generation processing units into conventional high-performance computing systems.

Hamilton, Kathleen↗

Estimating and Calibrating DER Model Parameters Using Levenberg–Marquardt Algorithm in Renewable Rich Power Grid

The proliferation of inverter-based distributed energy resources (IBDERs) has increased the number of control variables and dynamic interactions, leading to new grid control challenges. For stability analysis and designing appropriate protection controls, it is important that IBDER models are accurate. This paper focuses on the accurate estimation and parameter calibration of DER_A, a recently proposed aggregated IBDER model. In particular, we focus on the parameters of the reactive power–voltage regulation module. We formulate the problem of parameter tuning as a non-linear least square minimization problem and solve it using the Levenberg–Marquardt (LM) method. The LM method is primarily chosen due to its flexibility in adaptively selecting between the steepest descent and Gauss–Newton methods through a damping parameter. The LM approach is used to minimize the error between the actual measurements and the estimated response of the model. Further, the computational challenges posed by the numerical calculation of the Jacobian are tackled using a quasi-Newton root-finding approach. The proposed method is validated on a real feeder model in the northeastern part of the United States. The feeder is modeled in OpenDSS and the measurements thus obtained are fed to the DER_A model for calibration. The simulation results indicate that our approach is able to successfully calibrate the relevant model parameters quickly and with high accuracy, with a total sum of square error of 3.57 × 10 –7 .

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

On compactness of admissible parameter sets: Convergence and stability in inverse problems for distributed parameter systems

A series of numerical examples is reported and several algorithms compared for estimation of coefficients in differential equation models. Unconstrained, constrained and Tikhonov regularization methods are tested for their behavior with regard to both convergence (of approximation methods for the states and parameters) and stability (continuity of the estimates with respect to perturbations in the data or observed states).

Banks, H. T.↗

Proof-of-concept of a reinforcement learning framework for wind farm energy capture maximization in time-varying wind

Here, we present a proof-of-concept distributed reinforcement learning framework for wind farm energy capture maximization. The algorithm we propose uses Q-Learning in a wake-delayed wind farm environment and considers time-varying, though not yet fully turbulent, wind inflow conditions. These algorithm modifications are used to create the Gradient Approximation with Reinforcement Learning and Incremental Comparison (GARLIC) framework for optimizing wind farm energy capture in time-varying conditions, which is then compared to the FLOw Redirection and Induction in Steady State (FLORIS) static lookup table wind farm controller baseline.

17 WIND ENERGY↗

Feature Detection and Curve Fitting Using Fast Walsh Transforms for Shock Tracking: Applications

Walsh functions form an orthonormal basis set consisting of square waves. Square waves make the system well suited for detecting and representing functions with discontinuities. Given a uniform distribution of 2p cells on a one-dimensional element, it has been proven that the inner product of the Walsh Root function for group p with every polynomial of degree < or = (p - 1) across the element is identically zero. It has also been proven that the magnitude and location of a discontinuous jump, as represented by a Heaviside function, are explicitly identified by its Fast Walsh Transform (FWT) coefficients. These two proofs enable an algorithm that quickly provides a Weighted Least Squares fit to distributions across the element that include a discontinuity. The detection of a discontinuity enables analytic relations to locally describe its evolution and provide increased accuracy. Time accurate examples are provided for advection, Burgers equation, and Riemann problems (diaphragm burst) in closed tubes and de Laval nozzles. New algorithms to detect up to two C0 and/or C1 discontinuities within a single element are developed for application to the Riemann problem, in which a contact discontinuity and shock wave form after the diaphragm bursts.

Gnoffo, Peter A.↗

SU-C-BRC-07: Parametrized GPU Accelerated Electron Monte Carlo Second Check

Purpose: I am presenting a parameterized 3D GPU accelerated electron Monte Carlo second check program. Method: I wrote the 3D grid dose calculation algorithm in CUDA and utilized an NVIDIA GeForce GTX 780 Ti to run all of the calculations. The electron path beyond the distal end of the cone is governed by four parameters: the amplitude of scattering (AMP), the mean and width of a Gaussian energy distribution (E and α), and the percentage of photons. In my code, I adjusted all parameters until the calculated PDD and profile fit the measured 10×10 open beam data within 1%/1mm. I then wrote a user interface for reading the DICOM treatment plan and images in Python. In order to verify the algorithm, I calculated 3D dose distributions on a variety of phantoms and geometries, and compared them with the Eclipse eMC calculations. I also calculated several patient specific dose distributions, including a nose and an ear. Finally, I compared my algorithm’s computation times to Eclipse’s. Results: The calculated MU for all of the investigated geometries agree with the TPS within the TG-114 action level of 5%. The MU for the nose was < 0.5 % different while the MU for the ear at 105 SSD was ∼2 %. Calculation times for a 12MeV 10×10 open beam ranged from 1 second for a 2.5 mm grid resolution with ∼15 million particles to 33 seconds on a 1 mm grid with ∼460 million particles. Eclipse calculation runtimes distributed over 10 FAS workers were 9 seconds to 15 minutes respectively. Conclusion: The GPU accelerated second check allows quick MU verification while accounting for patient specific geometry and heterogeneity.

60 APPLIED LIFE SCIENCES↗

Experiments on active isolation using distributed PVDF error sensors

A control system based on a two-channel narrow-band LMS algorithm is used to isolate periodic vibration at low frequencies on a structure composed of a rigid top plate mounted on a flexible receiving plate. The control performance of distributed PVDF error sensors and accelerometer point sensors is compared. For both sensors, high levels of global reduction, up to 32 dB, have been obtained. It is found that, by driving the PVDF strip output voltage to zero, the controller may force the structure to vibrate so that the integration of the strain under the length of the PVDF strip is zero. This ability of the PVDF sensors to act as spatial filters is especially relevant in active control of sound radiation. It is concluded that the PVDF sensors are flexible, nonfragile, and inexpensive and can be used as strain sensors for active control applications of vibration isolation and sound radiation.

Lefebvre, S.↗

Streamline Integration Using MPI-Hybrid Parallelism on a Large Multicore Architecture

Streamline computation in a very large vector field data set represents a significant challenge due to the nonlocal and data-dependent nature of streamline integration. In this paper, we conduct a study of the performance characteristics of hybrid parallel programming and execution as applied to streamline integration on a large, multicore platform. With multicore processors now prevalent in clusters and supercomputers, there is a need to understand the impact of these hybrid systems in order to make the best implementation choice. We use two MPI-based distribution approaches based on established parallelization paradigms, parallelize over seeds and parallelize over blocks, and present a novel MPI-hybrid algorithm for each approach to compute streamlines. Our findings indicate that the work sharing between cores in the proposed MPI-hybrid parallel implementation results in much improved performance and consumes less communication and I/O bandwidth than a traditional, nonhybrid distributed implementation.

97 MATHEMATICS AND COMPUTING↗

Application of Laser Imaging for Bio/geophysical Studies

SPARTA, Inc. has developed a low-cost, portable laser imager that, among other applications, can be used in bio/geophysical applications. In the application to be discussed here, the system was utilized as an imaging system for background features in a forested locale. The SPARTA mini-ladar system was used at the International Paper Northern Experimental Forest near Howland, Maine to assist in a project designed to study the thermal and radiometric phenomenology at forest edges. The imager was used to obtain data from three complex sites, a 'seed' orchard, a forest edge, and a building. The goal of the study was to demonstrate the usefulness of the laser imager as a tool to obtain geometric and internal structure data about complex 3-D objects in a natural background. The data from these images have been analyzed to obtain information about the distributions of the objects in a scene. A range detection algorithm has been used to identify individual objects in a laser image and an edge detection algorithm then applied to highlight the outlines of discrete objects. An example of an image processed in such a manner is shown. Described here are the results from the study. In addition, results are presented outlining how the laser imaging system could be used to obtain other important information about bio/geophysical systems, such as the distribution of woody material in forests.

Hummel, J. R.↗

Retrieval of the sea spray aerosol mode from submicron particle size distributions and supermicron scattering during LASIC

Abstract. Improved quantification of sea spray aerosol concentration and size is important for determining aerosol effects on clouds and the climate, though attempts to accurately capture the size distribution of the sea spray mode remain limited by the availability of supermicron size distributions. In this work, we introduce a new approach to retrieving lognormal mode fit parameters for a sea spray aerosol mode by combining submicron size distributions with supermicron scattering measurements using a Mie inversion. Submicron size distributions were measured by an ultra-high-sensitivity aerosol spectrometer (UHSAS), and supermicron scattering was taken as the difference between <10 µm and <1 µm three-wavelength integrating nephelometer measurements (NEPH). This UHSAS-NEPH method was applied during background marine periods of the Department of Energy Atmospheric Radiation Measurement Layered Atlantic Smoke Interactions with Clouds (LASIC) campaign on Ascension Island (November 2016–May 2017), when the contribution of sea spray aerosol was expected to represent a large fraction of the aerosol mass and total scattering. Lognormal sea spray modal parameters were retrieved from comparisons between nephelometer measurements and a lookup table of Mie theory-simulated scattering coefficients for low-error solutions that minimized the 0.4–1 µm residual in the UHSAS size distribution. We evaluated the UHSAS-NEPH method with a set of clean marine measurements in the North Atlantic that included supermicron size and chemical measurements, showing that measured supermicron size distributions are needed to constrain the sea spray number concentration but that mass concentration was reasonably characterized using supermicron scattering. For LASIC, the UHSAS-NEPH method retrieved sea spray mode properties for approximately 88 % of the background marine times when the scattering variability and total particle concentration were low (<± 5 Mm−1 and <400 cm−3, respectively), with mass mean diameter ranging from 0.6 to 1.9 µm (1.47 ± 0.17 µm), modal width ranging from 1.1 to 3.97 (2.4±0.3), and mass concentration ranging from 0.18 to 23.0 µg m−3 (8.37. ± 4.1 µg m−3). The measured nephelometer scattering at three wavelengths was found to constrain the mode width marginally at the largest particle sizes in the absence of additional size and chemical measurements for defining parameters for the Mie solutions. Comparing UHSAS-NEPH retrievals to those of a fitting algorithm applied only to the submicron UHSAS number size distribution showed that correlations between retrieved mass concentration and the available mass-based sea spray tracers (coarse scattering, wind speed, and chloride) are low when supermicron measurements are not considered. This work demonstrates the added value of supermicron scattering measurements for retrieving reasonable sea spray mass concentrations, providing the best-available observationally constrained estimate of the sea spray mode properties when supermicron size distribution measurements are not available.

54 ENVIRONMENTAL SCIENCES↗