Search NASA⌕ Search

SEARCH · Search NASA

Results for “compilation”

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 1,765 records · Page 98

Contribution of citizen science towards international biodiversity monitoring

To meet collective obligations towards biodiversity conservation and monitoring, it is essential that the world's governments and non-governmental organisations as well as the research community tap all possible sources of data and information, including new, fast-growing sources such as citizen science (CS), in which volunteers participate in some or all aspects of environmental assessments. Through compilation of a database on CS and community-based monitoring (CBM, a subset of CS) programs, we assess where contributions from CS and CBM are significant and where opportunities for growth exist. We use the Essential Biodiversity Variable framework to describe the range of biodiversity data needed to track progress towards global biodiversity targets, and we assess strengths and gaps in geographical and taxonomic coverage. Our results show that existing CS and CBM data particularly provide large-scale data on species distribution and population abundance, species traits such as phenology, and ecosystem function variables such as primary and secondary productivity. Only birds, Lepidoptera and plants are monitored at scale. Most CS schemes are found in Europe, North America, South Africa, India, and Australia. We then explore what can be learned from successful CS/CBM programs that would facilitate the scaling up of current efforts, how existing strengths in data coverage can be better exploited, and the strategies that could maximise the synergies between CS/CBM and other approaches for monitoring biodiversity, in particular from remote sensing. More and better targeted funding will be needed, if CS/CBM programs are to contribute further to international biodiversity monitoring.

59 BASIC BIOLOGICAL SCIENCES↗

Biomass supply chain equipment for renewable fuels production: A review

The production of renewable fuels is a critical component of global strategies to reduce greenhouse gas (GHG) emissions. Moreover, the collection of raw materials for its production can provide added benefits such as reduction of wildfire risk, additional income for farmers, and decreased disposal costs. Although there is substantial literature on design and modeling of supply chains, the authors were unable to find a single reference with the information needed for the selection and cost estimation of each type of equipment involved in the supply chain. Therefore, the goal of this research is to gather information necessary for the construction and utilization of models that might drive the identification of a feasible supply chain to produce renewable fuels at a commercial scale. The primary objectives are to 1) understand the supply chain of critical feedstocks for renewable fuels production; 2) identify the equipment commercially available for collection and adequation of feedstock; and 3) consolidate information regarding equipment cost, energy consumption, and efficiency, as well as feedstock storage and transportation systems. This paper provides a compilation for five feedstock types studied for sustainable aviation fuel production: 1) agricultural residues and grasses, 2) forest residues, 3) urban wood waste, 4) oilseeds, 5) fats, oils & greases. All the technologies involved from the field to the gate of the preprocessing or conversion unit were reviewed. The information on fats, oils & greases supply chains and equipment purposely designed for forest thinning and pruning was very limited.

: Feedstock, Collection and Adequation, Renewable ↗

Italian prototype building models for urban scale building performance simulation

Urban building energy modeling (UBEM) seeks to evaluate strategies to optimize building energy use at urban scale to support a city's building energy goals. Prototype building models are usually developed to represent typical urban building characteristics of a specific use type, construction year, and climate zone, as detailed characteristics of individual buildings at urban scale are difficult to obtain. This study investigated the Italian building stock, developing 46 building prototypes, based on construction year, for residential and office buildings. The study included 16 single-family buildings, 16 multi-family buildings, and 14 office buildings. Building envelope properties and heating, ventilation, and air conditioning system characteristics were defined according to existing building energy codes and standards for climatic zone E, which covers about half the Italian municipalities. Novel contributions of this study include (1) detailed specifications of prototype building energy models for Italian residential and office buildings that can be adopted by UBEM tools, and (2) a dataset in GeoJSON format of Italian urban buildings compiled from diverse data sources and national standards. Finally, the developed prototype building specifications, the building dataset, and the workflow can be applied to create other building prototypes and to support Italian national building energy efficiency and environmental goals.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Analytical models of hydrogen transport in graphite

The importance of graphite-hydrogen chemical reactions to fusion, fission, and hydrogen storage applications, combined with the rapidly evolving knowledge on the underlying mechanisms, has led to the development of multiple models to describe hydrogen transport in graphite. Significant differences exist among these models, resulting from discrepancies in the modeling assumptions, intended degree of fidelity, and conditions of applicability. This paper attempts at reconciling these apparent differences by providing a comprehensive description of the constitutive equations governing hydrogen transport in graphite at high-temperature, identifying outstanding gaps in knowledge, illustrating how these different models approach them, and proposing alternative analytical formulations grounded on experimental results from hydrogen-graphite studies. Governing equations, closing relations, and simplifying assumptions are discussed for hydrogen transport at the inter-granular and intra-granular level, accompanied by compiled experimental data and illustrated energy diagrams associated to the proposed transport mechanisms. Analytical formulations are provided to reproduce competing hypotheses on the mechanisms, supporting the development of a range of computational models that can enable resolution of outstanding knowledge gaps through comparative testing against experimental data.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Fast estimation of reaction rates in spherical and non-spherical porous catalysts

Here we present a methodology for modeling multi-step reaction rates in porous catalyst particles for use in CFD-DEM and two fluid models. Single-step effectiveness factors based on a Thiele modulus, while useful, cannot accurately capture the cascading reaction systems common in high temperature vapor-phase chemical reactors like fluidized catalytic cracking units and catalytic biomass fast pyrolysis systems. Instead, multi-step effectiveness vectors derived from steady-state solutions to the governing reaction-diffusion equations are needed. Solutions for various catalyst shapes are presented, including spheres, cylinders, and prisms. Computational challenges inherent in repeated evaluation of reaction rates with diffusion limitations are discussed, and an efficient implementation based on pre-computed lookup tables is proposed and demonstrated on a simulation of a fluidized bed reactor. Open-source code is provided for the compilation of reaction rate tables for use in ODE, DEM, and two-fluid models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Predicting concrete compressive strength using hybrid ensembling of surrogate machine learning models

This study aims to implement a hybrid ensemble surrogate machine learning technique in predicting the compressive strength (CS) of concrete, an important parameter used for durability design and service life prediction of concrete structures in civil engineering projects. For this purpose, an experimental database consisting of 1030 records has been compiled from the machine learning repository of the University of California, Irvine. The database was used to train and validate four conventional machine learning (CML) models, namely Artificial Neural Network (ANN), Linear and Non-Linear Multivariate Adaptive Regression Splines (MARS-L and MARS-C), Gaussian Process Regression (GPR), and Minimax Probability Machine Regression (MPMR). Subsequently, the predicted outputs of CML models were combined and trained using ANN to construct the Hybrid Ensemble Model (HENSM). It is observed that the proposed HENSM produces higher predictive accuracy compared to the CML models used in the present study. The predictive performance of all models for CS prediction was compared using the testing dataset and it is found that the HENSM model attained the highest predictive accuracy in both phases. Based on the experimental results, the newly constructed HENSM model is very potential to be a new alternative in handling the overfitting issues of CML models and hence, can be used to predict the concrete CS, including the design of less polluting and more sustainable concrete constructions.

36 MATERIALS SCIENCE↗

The multiple depleted mantle components in the Hawaiian-Emperor chain

Oceanic island basalts are targeted for geochemical study because they provide a direct window into mantle composition and a wealth of information on the dynamics and timescales associated with Earth mixing. Previous studies mainly focused on the shield volcanic stage of oceanic islands and the more fusible, enriched mantle components that are easily distinguished in those basalts. Mantle depleted compositions are typically more difficult to resolve unless large amounts of this material participated in mantle melting (e.g., mid-ocean ridges), or unique processes allow for their compositions to be erupted undiluted, such as very small degrees of melting of a source with minimal fusible enriched components (e.g., rejuvenated basalts) or as xenoliths (e.g., abyssal peridotites). Mantle depleted components, defined here as material with low time-integrated Rb/Sr (low 87 Sr/ 86 Sr) and high time-integrated Sm/Nd and Lu/Hf ratios (high 143 Nd/ 144 Nd and 176 Hf/ 177 Hf) relative to primitive mantle, derive from a potentially very large volume reservoir (up to 80% of the mantle), and therefore need adequate characterization in order estimate the composition of the Earth and mantle-derived melts. Here, this review focuses on mantle depleted compositions in oceanic island basalts using the Hawaiian-Emperor chain as a case study. The Hawaiian-Emperor chain is the ~6000 km long geological record of the deeply sourced Hawaiian mantle plume, active for >81 Myr. Hawaiian volcanism evolves through four volcanic stages as a volcano traverses the Hawaiian plume: alkalic preshield, tholeiitic shield (80–90% volcano volume), alkalic postshield (~1%), and silica undersaturated rejuvenated (<0.1%). We report Pb-Sr-Nd-Hf isotope compositions and trace element concentrations of three rejuvenated Northwest Hawaiian Ridge basalts and compare them to an exhaustive compiled dataset of basalts from the Hawaiian Islands to the Emperor Seamounts. The Northwest Hawaiian Ridge (NWHR) includes 51 volcanoes spanning ~42 m.y. between the bend in the Hawaiian-Emperor chain and the Hawaiian Islands where there is no high-precision isotopic data published on the rejuvenated-stage over ~47% of the chain. NWHR and Hawaiian Island rejuvenated basalts are geochemically similar, indicating a consistent source for rejuvenated volcanism over ~12.5 million years. In contrast, shield-stage basalts from the oldest Emperor Seamounts are more depleted in isotopic composition (i.e., higher 176 Hf/ 177 Hf, and 143 Nd/ 144 Nd with lower 87 Sr/ 86 Sr and 208 Pb*/ 206 Pb*) and trace element concentrations (i.e., much lower concentrations of highly incompatible elements) than all other depleted Hawaiian basalts younger than the bend, including NWHR rejuvenated basalts. The strongly depleted source for the oldest Emperor Seamounts (>70 Ma) was likely related to interaction with the Kula-Pacific-Izanagi mid-ocean ridge spreading system active near the Hawaiian plume in the Late Cretaceous. In contrast, the incompatible trace element ratios of NWHR rejuvenated basalts require a distinct source in the Hawaiian mantle plume that was imprinted by ancient (>1 Ga) partial melting, likely ancient recycled oceanic lithosphere. This review of the geochemistry of Hawaiian depleted components documents the need for the sampling of multiple distinctive depleted compositions, each preferentially melted during specific periods of Hawaiian plume activity. This suggests that the composition of depleted components can evolve during the lifetime of the mantle plume, as observed for enriched components in the Hawaiian mantle plume. Changes in the composition of depleted components are dominantly controlled by the upper mantle tectonic configurations at the time of eruption (i.e., proximity to a mid-ocean ridge), as this effect overwhelms the signal imparted by potentially sampling different lower mantle components through time.

07 ISOTOPE AND RADIATION SOURCES↗

Optimized thermodynamic properties of REE aqueous species (REE 3+ and REEOH 2+ ) and experimental database for modeling the solubility of REE phosphate minerals (monazite, xenotime, and rhabdophane) from 25 to 300 °C

Rare earth elements (REE) are critical elements found in monazite, xenotime, and hydrated REE phosphates which typically form in hydrothermal mineral deposits. Accurate predictions of the solubility of these REE phosphates and the speciation of REE in aqueous fluids are both key to understanding the controls on the transport, fractionation, and deposition of REE in natural systems. Previous monazite and xenotime solubility experiments indicate the presence of large discrepancies between experimentally derived solubility constants versus calculated solubilities by combining different data sources for the thermodynamic properties of minerals and aqueous species at hydrothermal conditions. In this study, these discrepancies were resolved by using the program GEMSFITS to optimize the standard partial molal Gibbs energy of formation (Δ f G° 298 ) of REE aqueous species (REE 3+ and REE hydroxyl complexes) at 298.15 K and 1 bar while keeping the thermodynamic properties fixed for the REE phosphates. A comprehensive experimental database was compiled using solubility data available between 25 and 300 °C. The latter permits conducting thermodynamic parameter optimization of Δ f G° 298 for REE aqueous species. Optimal matching of the rhabdophane solubility data between 25 and 100 °C requires modifying the Δ f G° 298 values of REE 3+ by 1–6 kJ/mol, whereas matching of the monazite solubility data between 100 and 300 °C requires modifying the Δ f G° 298 values of both REE 3+ and REEOH 2+ by ~15–31 kJ/mol and ~2–10 kJ/mol, respectively. For xenotime, adjustments of Δ f G° 298 values by 1–26 kJ/mol are only necessary for the REE 3+ species. The optimizations indicate that the solubility of monazite in acidic solutions is controlled by the light (L)REE 3+ species at <150 °C and the LREEOH 2+ species at >150 °C, whereas the solubility of xenotime is controlled by the heavy (H)REE 3+ species between 25 and 300 °C. Based on the optimization results, we conclude that the revised Helgeson-Kirkham-Flowers equation of state does not reliably predict the thermodynamic properties of REE 3+ , REEOH 2+ , and likely other REE hydroxyl species at hydrothermal conditions. We therefore provide an experimental database (ThermoExp_REE) as a basic framework for future updates, extensions with other ligands, and optimizations as new experimental REE data become available. As a result, the optimized thermodynamic properties of aqueous species and minerals are available open access to accurately predict the solubility of REE phosphates in fluid-rock systems.

58 GEOSCIENCES↗

Plastic additives in the ocean: Use of a comprehensive dataset for meta-analysis and method development

In excess of 13,000 chemicals are added to plastics (‘additives’) to improve performance, durability, and production of plastic products. They are categorized into numerous chemical classes including flame retardants, light stabilizers, antioxidants, and plasticizers. While research on plastic additives in the marine environment has increased over the past decade, there is a lack of methodological standardization. To direct future measurement of plastic additives, we compiled a first-of-its-kind dataset of literature assessing plastic additives in marine environments, delineated by sample type (plastic debris, seawater, sediment, biota). Using this dataset, we performed a meta-analysis to summarize the state of the science. Currently, our dataset includes 217 publications published between 1978 and May 2023. The majority of publications analyzed plastic additives in biota collected from Europe and Asia. Analyses concentrated on plasticizers, brominated flame retardants, and bisphenols. Common sample preparation techniques included Solvent - Agitation extraction for plastic, sediment, and biota samples, and Solid Phase Extraction for seawater samples with dichloromethane and solvent mixtures including dichloromethane as the organic extraction solvent. Finally, most analyses were performed utilizing gas chromatography/mass spectrometry. There are a variety of data gaps illuminated by this meta-analysis, most notably the small number of compounds that have been targeted for detection compared to the large number of additives used in plastic production. The provided dataset facilitates future investigation of trends in plastic additive concentration data in the marine environment (allowing for comparison to toxicity thresholds) and acts as a starting point for optimizing and harmonizing plastic additive analytical methods.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Atomization energies from coupled-cluster calculations augmented with explicitly-correlated perturbation theory.

The atomization energies of the 105 molecules in the test set of Bakowies [D. Bakowies, J. Chem. Phys. 127 (2007) 084105] have been computed with an estimated standard deviation (from the values compiled in the Active Thermochemical Tables) of {+-}0.1 kJ/mol per valence electron in the molecule. Equilibrium geometries and harmonic vibrational frequencies were calculated at the all-electron CCSD(T)/cc-pCVTZ level, that is, at the level of coupled-cluster theory with singles, doubles and non-iterative triples in a correlation-consistent polarized core-valence triple-zeta basis. Single-point energy calculations were performed at the all-electron CCSD(T) level in a correlation-consistent polarized core-valence quadruple-zeta basis (cc-pCVQZ), and several corrections were added: (i) a correction for the basis-set truncation error, obtained from second-order perturbation theory using Slater-type geminals (MP2-F12 theory), (ii) a correction for the effect of anharmonicity on the zero-point vibrational energy, (iii) a relativistic correction, (iv) a correction for the difference between the full CCSDT model (coupled-cluster theory with singles, doubles and triples) and the CCSD(T) approximation, and (v) a correction for connected quadruple excitations obtained from CCSDT(Q) calculations. The correction for the basis-set truncation error was obtained from MP2-F12 calculations by scaling the MP2 basis-set truncation error by an empirically optimized 'interference factor' of f{sub int} = 0.78. The reference values from the Active Thermochemical Tables for 73 molecules in the test set, the equilibrium geometries, the harmonic vibrational frequencies, and all of the energy corrections represent valuable data for performance assessments of additivity schemes that will be developed in the future, in which the basis-set truncation error will be calculated at the level of coupled-cluster theory using Slater-type geminals (CC-F12 theory). Such a scheme will be free of empirical corrections and scaling factors.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Flow-flame interactions causing acoustically coupled heat release fluctuations in a thermo-acoustically unstable gas turbine model combustor

A detailed analysis of the flow-flame interactions associated with acoustically coupled heat-release rate fluctuations was performed for a 10 kW, CH{sub 4}/air, swirl stabilized flame in a gas turbine model combustor exhibiting self-excited thermo-acoustic oscillations at 308 Hz. High-speed stereoscopic particle image velocimetry, OH planar laser induced fluorescence, and OH* chemiluminescence measurements were performed at a sustained repetition rate of 5 kHz, which was sufficient to resolve the relevant combustor dynamics. Using spatio-temporal proper orthogonal decomposition, it was found that the flow-field contained several simultaneous periodic motions: the reactant flux into the combustion chamber periodically oscillated at the thermo-acoustic frequency (308 Hz), a helical precessing vortex core (PVC) circumscribed the burner nozzle at 515 Hz, and the PVC underwent axial contraction and extension at the thermo-acoustic frequency. The global heat release rate fluctuated at the thermo-acoustic frequency, while the heat release centroid circumscribed the combustor at the difference between the thermo-acoustic and PVC frequencies. Hence, the three-dimensional location of the heat release fluctuations depended on the interaction of the PVC with the flame surface. This motivated the compilation of doubly phase resolved statistics based on the phase of both the acoustic and PVC cycles, which showed highly repeatable periodic flow-flame configurations. These include flames stabilized between the inflow and inner recirculation zone, large-scale flame wrap-up by the PVC, radial deflection of the inflow by the PVC, and combustion in the outer recirculation zones. Large oscillations in the flame surface area were observed at the thermo-accoustic frequency that significantly affected the total heat-release oscillations. By filtering the instantaneous reaction layers at different scales, the importance of the various flow-flame interactions affecting the flame area was determined. The greatest contributor was large-scale elongation of the reaction layers associated with the fluctuating reactant flow rate, which accounted for approximately 50% of the fluctuations. The remaining 50% was distributed between fine scale stochastic corrugation and large-scale corrugation due to the PVC. (author)

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Theoretical estimates of flammability bounds for thin condensed fuel diffusion flames in microgravity using detailed models of chemistry and radiation

Recently, U-shape flammability maps have been constructed showing minimal oxygen vs. flame strain for opposed flame spread in micro-gravity by Olson and Ferkul. The U-shape defines the limiting flammability bounds from radiative extinction and flame blow-off. Here, the minimum of the U corresponds to the minimum possible oxidizer concentration where burning can occur, and is an important quantity of interest for fire safety. While high strain extinction bounds have been well analyzed, low strain radiative extinction has not. To estimate low strain extinction, in this study an analytical theory is developed based on thin flame theory coupled with a heat and mass transfer model for solid fuels. A reaction progress variable based on the Damköhler number is adapted in the theory to account for incomplete combustion at high strain rates and enable the capturing of the full flammability map. The analytical model is compared to a one dimensional numerical model w/ detailed chemical kinetics and coupled radiation heat transfer in planar and spherical geometries. The flammability maps are then qualitatively compared to experimental extinguishment data compiled by Olson and Ferkul for cylindrical rods of PMMA showing similar trends. The results show the newly developed analytics capture the radiative extinction bound compared to the numerical model and qualitatively agrees with microgravity data.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗

Rapid multiphase-field model development using a modular free energy based approach with automatic differentiation in MOOSE/MARMOT

In this paper, we present a novel rapid phase-field model development capability using a modular free energy based approach in the finite element based MOOSE framework. The phase-field equations are implemented in a general form as separate code form the model specific data such as the thermodynamic free energy density and mobility functions. This isolates researchers from complicated code and allows them to focus on implementing new material models. In addition, using a native symbolic differentiation algorithm, all required derivatives of the free energy density are automatically determined, removing sources for potential human error, and guaranteeing optimally preconditioned equation systems through fully populated Jacobian matrices. Free energy terms contributing to a phase-field model are abstracted into self contained objects that can be combined at will at simulation run time. Combining multiple chemical and mechanical free energy contributions allows the construction of coupled phase-field, mechanics, and multiphase models. Through just-in-time compilation, we greatly reduce the computational time required to evaluate the differentiated expressions. Lastly, we demonstrate the new capability on a variety of example simulations.

97 MATHEMATICS AND COMPUTING↗

9Cr steel visualization and predictive modeling

The goals of this work were to develop tools to visualize and characterize materials information, to explore new alloy compositions and/or processing, to better predict tensile strength and other mechanical properties. The 9–12 wt% Cr (9Cr) steel test data were compiled from several sources. About 3000 records with 30 predictor variables represent 82 unique steel alloy compositions, variations of thermo-mechanical processing steps and temperatures (homogenization, normalization and tempering cycles), test conditions and outcomes. Detailed data processing steps such as visualization and exploratory analysis, including univariate and bivariate analysis, different clustering techniques used to segment the data, statistical post-hoc analysis to verify significance of the findings, feature engineering to identify predictors of importance, and predictive modeling with cross-validation were performed. The outcome of analysis at each step was reviewed in the context of the domain knowledge of this class of steel, to see if there were underlying physical mechanisms that explain statistical relationships. Data analytics techniques and their parameters were fine-tuned to facilitate interpretation of the results as aligned with the insights from domain experts. Here, the ensemble predictive modeling using Random Forest regressor and post-hoc means comparison corroborated domain knowledge on the role of Co which is known to increase strength of steel alloys through solid-solution strengthening and to affect diffusion of alloying elements and precipitates. Alloys with Co content of 0.7–8 wt% had significantly higher mean strength than the ones without (while having Cr locked within 10–11 wt% range in either group and keeping other compositional element ratios fixed). End products of the computational techniques exploration are presented as the tools that can be used in iterative workflow of materials development and testing.

36 MATERIALS SCIENCE↗

Deep-freeze graph training for latent learning

Scientific and engineering advances are primarily driven by multi-tier conceptual constructs and conditional theoretical frameworks. The theories allow predictions of hypothetical system responses, given a set of approximate conditions (ranges of applicability) imposed on latent parameters that cannot be measured directly. Learning to estimate the latent variables (Latent Learning) helps to pinpoint the anticipated range-edge anomalies and improves the confidence in interpretation, interpolation and extrapolation of limited experimental data. Due to high dimensionality and extreme non-linearity of the materials science problems, very large datasets are typically required for conventional data-driven model development. The vital experimental data collection, particularly on microstructural phases, is very challenging, which makes it difficult to compile a high-quality database. Incorporation of the domain knowledge into the computational graph structure, initialization and optimization processes presents a viable mechanism for developing accurate models, with limited datasets. Furthermore, this study successfully utilized the approach to build the Deep Freeze Graph (DeepFreG) by mapping known causality relationships and by digitizing empirical domain knowledge for Latent Learning (LL), with specific applications in materials science.

36 MATERIALS SCIENCE↗

Overcoming small minirhizotron datasets using transfer learning

Minirhizotron technology is widely used to study root growth and development. Yet, standard approaches for tracing roots in minirhiztron imagery is extremely tedious and time consuming. Machine learning approaches can help to automate this task. However, lack of enough annotated training data is a major limitation for the application of machine learning methods. Transfer learning is a useful technique to help with training when available datasets are limited. In this paper, we investigated the effect of pre-trained features from the massives-cale, irrelevant ImageNet dataset and a relatively moderate-scale, but relevant peanut root dataset on switchgrass root imagery segmentation applications. We compiled two minirhizotron image datasets to accomplish this study: one with 17,550 peanut root images and another with 28 switchgrass root images. Both datasets were paired with manually labeled ground truth masks. Deep neural networks based on the U-net architecture were used with different pre-trained features as initialization for automated, precise pixel-wise root segmentation in minirhizotron imagery. We observed that features pre-trained on a closely related but relatively moderate size dataset like our peanut dataset were more effective than features pre-trained on the large but unrelated ImageNet dataset. Here, we achieved high quality segmentation on peanut root dataset with 99.04% accuracy at the pixel-level and overcame errors in human-labeled ground truth masks. By applying transfer learning technique on limited switchgrass dataset with features pre-trained on peanut dataset, we obtained 99% segmentation accuracy in switchgrass imagery using only 21 images for training (fine tuning). Furthermore, the peanut pre-trained features can help the model converge faster and have much more stable performance.

59 BASIC BIOLOGICAL SCIENCES↗

Implementing Molecular Dynamics for Hybrid High Performance Computers - 1. Short Range Forces

The use of accelerators such as general-purpose graphics processing units (GPGPUs) have become popular in scientific computing applications due to their low cost, impressive floating-point capabilities, high memory bandwidth, and low electrical power requirements. Hybrid high performance computers, machines with more than one type of floating-point processor, are now becoming more prevalent due to these advantages. In this work, we discuss several important issues in porting a large molecular dynamics code for use on parallel hybrid machines - 1) choosing a hybrid parallel decomposition that works on central processing units (CPUs) with distributed memory and accelerator cores with shared memory, 2) minimizing the amount of code that must be ported for efficient acceleration, 3) utilizing the available processing power from both many-core CPUs and accelerators, and 4) choosing a programming model for acceleration. We present our solution to each of these issues for short-range force calculation in the molecular dynamics package LAMMPS. We describe algorithms for efficient short range force calculation on hybrid high performance machines. We describe a new approach for dynamic load balancing of work between CPU and accelerator cores. We describe the Geryon library that allows a single code to compile with both CUDA and OpenCL for use on a variety of accelerators. Finally, we present results on a parallel test cluster containing 32 Fermi GPGPUs and 180 CPU cores.

43 PARTICLE ACCELERATORS↗

Implementing Molecular Dynamics on Hybrid High Performance Computers - Particle-Particle Particle-Mesh

The use of accelerators such as graphics processing units (GPUs) has become popular in scientific computing applications due to their low cost, impressive floating-point capabilities, high memory bandwidth, and low electrical power requirements. Hybrid high-performance computers, machines with nodes containing more than one type of floating-point processor (e.g. CPU and GPU), are now becoming more prevalent due to these advantages. In this paper, we present a continuation of previous work implementing algorithms for using accelerators into the LAMMPS molecular dynamics software for distributed memory parallel hybrid machines. In our previous work, we focused on acceleration for short-range models with an approach intended to harness the processing power of both the accelerator and (multi-core) CPUs. To augment the existing implementations, we present an efficient implementation of long-range electrostatic force calculation for molecular dynamics. Specifically, we present an implementation of the particle-particle particle-mesh method based on the work by Harvey and De Fabritiis. We present benchmark results on the Keeneland InfiniBand GPU cluster. We provide a performance comparison of the same kernels compiled with both CUDA and OpenCL. We discuss limitations to parallel efficiency and future directions for improving performance on hybrid or heterogeneous computers.

97 MATHEMATICS AND COMPUTING↗