Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “SCALE MODELS”

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 523 records · Page 29

Modeling scale-dependent bias on the baryonic acoustic scale with the statistics of peaks of Gaussian random fields

Models of galaxy and halo clustering commonly assume that the tracers can be treated as a continuous field locally biased with respect to the underlying mass distribution. In the peak model pioneered by Bardeen et al.[Astrophys. J. 304, 15 (1986)], one considers instead density maxima of the initial, Gaussian mass density field as an approximation to the formation site of virialized objects. In this paper, the peak model is extended in two ways to improve its predictive accuracy. First, we derive the two-point correlation function of initial density peaks up to second order and demonstrate that a peak-background split approach can be applied to obtain the k-independent and k-dependent peak bias factors at all orders. Second, we explore the gravitational evolution of the peak correlation function within the Zel'dovich approximation. We show that the local (Lagrangian) bias approach emerges as a special case of the peak model, in which all bias parameters are scale independent and there is no statistical velocity bias. We apply our formulas to study how the Lagrangian peak biasing, the diffusion due to large scale flows, and the mode coupling due to nonlocal interactions affect the scale dependence of bias from small separations up to the baryon acoustic oscillation (BAO) scale. For 2{sigma} density peaks collapsing at z=0.3, our model predicts a {approx}5% residual scale-dependent bias around the acoustic scale that arises mostly from first order Lagrangian peak biasing (as opposed to second order gravity mode coupling). We also search for a scale dependence of bias in the large scale autocorrelation of massive halos extracted from a very large N-body simulation provided by the MICE Collaboration. For halos with mass M > or approx. 10{sup 14}M{sub {center_dot}/}h, our measurements demonstrate a scale-dependent bias across the BAO feature which is very well reproduced by a prediction based on the peak model.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A unified gas-kinetic scheme for continuum and rarefied flows IV: Full Boltzmann and model equations

Fluid dynamic equations are valid in their respective modeling scales, such as the particle mean free path scale of the Boltzmann equation and the hydrodynamic scale of the Navier–Stokes (NS) equations. With a variation of the modeling scales, theoretically there should have a continuous spectrum of fluid dynamic equations. Even though the Boltzmann equation is claimed to be valid in all scales, many Boltzmann solvers, including direct simulation Monte Carlo method, require the cell resolution to the order of particle mean free path scale. Therefore, they are still single scale methods. In order to study multiscale flow evolution efficiently, the dynamics in the computational fluid has to be changed with the scales. A direct modeling of flow physics with a changeable scale may become an appropriate approach. The unified gas-kinetic scheme (UGKS) is a direct modeling method in the mesh size scale, and its underlying flow physics depends on the resolution of the cell size relative to the particle mean free path. The cell size of UGKS is not limited by the particle mean free path. With the variation of the ratio between the numerical cell size and local particle mean free path, the UGKS recovers the flow dynamics from the particle transport and collision in the kinetic scale to the wave propagation in the hydrodynamic scale. The previous UGKS is mostly constructed from the evolution solution of kinetic model equations. Even though the UGKS is very accurate and effective in the low transition and continuum flow regimes with the time step being much larger than the particle mean free time, it still has space to develop more accurate flow solver in the region, where the time step is comparable with the local particle mean free time. In such a scale, there is dynamic difference from the full Boltzmann collision term and the model equations. This work is about the further development of the UGKS with the implementation of the full Boltzmann collision term in the region where it is needed. The central ingredient of the UGKS is the coupled treatment of particle transport and collision in the flux evaluation across a cell interface, where a continuous flow dynamics from kinetic to hydrodynamic scales is modeled. The newly developed UGKS has the asymptotic preserving (AP) property of recovering the NS solutions in the continuum flow regime, and the full Boltzmann solution in the rarefied regime. In the mostly unexplored transition regime, the UGKS itself provides a valuable tool for the non-equilibrium flow study. The mathematical properties of the scheme, such as stability, accuracy, and the asymptotic preserving, will be analyzed in this paper as well.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A micro-scale cutting model for UD CFRP composites with thermo-mechanical coupling

Cutting a unidirectional carbon fiber-reinforced polymer (UD CFRP) structure is the basic unit for CFRP machining, which is a complex thermal-mechanically coupled process. To reveal the deformation mechanism and predict cutting force in UD CFRP micro cutting, a micro-scale fracture model for UD CFRP cutting with thermal-mechanical coupling is demonstrated in this paper, which captures the failure modes for fibers, matrix and the interface based on a micro-level RVE using a relatively simple damage based fracture method. The thermal-mechanical coupling model at the micro scale is developed on the basis of the plastic energy dissipation and frictional heating during cutting. Failure models for the fiber, matrix and interface region are applied depending on the material properties of each of these three phases. Numerical simulations based on the above model with different fiber orientations were performed to predict the deformation and forces of different components in UD CFRP. Cutting experiments with the same fiber orientations as considered in the simulations were carried out to validate the force and deformation results. The predicted force and deformation patterns match well with evidence from our experiments. In general, the cutting force is larger than the thrust force regardless of fiber orientation. The cutting force reaches a maximum as the fiber orientation approaches 90 , but thrust forces do not vary substantially across cases. When the fiber orientation is acute, the deformation of fibers is much smaller than when the cutting angle is obtuse. Surface roughness follows the same trend with cutting angle as fiber deformation.

42 ENGINEERING↗

Genome-Scale Metabolic Model for the Green Alga Chlorella vulgaris UTEX 395 Accurately Predicts Phenotypes under Autotrophic, Heterotrophic, and Mixotrophic Growth Conditions

The green microalgae Chlorella vulgaris has been widely recognized as a promising candidate for biofuel production due to its ability to store high lipid content and its natural metabolic versatility. Compartmentalized genome-scale metabolic models constructed from genome sequences enable quantitative insight into the transport and metabolism of compounds within a target organism. These metabolic models have long been utilized to generate optimized design strategies for an improved production process. Here, we describe the reconstruction, validation, and application of a genome-scale metabolic model for C. vulgaris UTEX 395, iCZ843. The reconstruction represents the most comprehensive model for any eukaryotic photosynthetic organism to date, based on the genome size and number of genes in the reconstruction. The highly curated model accurately predicts phenotypes under photoautotrophic, heterotrophic, and mixotrophic conditions. The model was validated against experimental data and lays the foundation for model-driven strain design and medium alteration to improve yield. Calculated flux distributions under different trophic conditions show that a number of key pathways are affected by nitrogen starvation conditions, including central carbon metabolism and amino acid, nucleotide, and pigment biosynthetic pathways. Moreover, model prediction of growth rates under various medium compositions and subsequent experimental validation showed an increased growth rate with the addition of tryptophan and methionine.

09 BIOMASS FUELS↗

Large-scale functional models of visual cortex for remote sensing

Neuroscience has revealed many properties of neurons and of the functional organization of visual cortex that are believed to be essential to human vision, but are missing in standard artificial neural networks. Equally important may be the sheer scale of visual cortex requiring {approx}1 petaflop of computation. In a year, the retina delivers {approx}1 petapixel to the brain, leading to massively large opportunities for learning at many levels of the cortical system. We describe work at Los Alamos National Laboratory (LANL) to develop large-scale functional models of visual cortex on LANL's Roadrunner petaflop supercomputer. An initial run of a simple region VI code achieved 1.144 petaflops during trials at the IBM facility in Poughkeepsie, NY (June 2008). Here, we present criteria for assessing when a set of learned local representations is 'complete' along with general criteria for assessing computer vision models based on their projected scaling behavior. Finally, we extend one class of biologically-inspired learning models to problems of remote sensing imagery.

97 MATHEMATICS AND COMPUTING↗

Representativity error scaling of models of the high temperature test facility

Error scaling is a critical step toward the validation of modeling capabilities for high-temperature gas-cooled reactors (HTGRs). Extensive effort is being made to bring HTGRs into the validation basis of numerous thermal-hydraulics codes. Here, this paper demonstrates how systems-level codes can be leveraged to perform error scaling analyses between an experimental facility and a plant-to-be facility. Specifically, we focus on two conduction cooldown experiments from the High Temperature Test Facility (HTTF) and the General Atomics 350 MW th modular high-temperature gas-cooled reactor (MHTGR-350). The error scaling methodology employed in this study is representativity, which in the context of this work was used to quantify how well experiments captured the physics of the plant facility by comparing sensitivity vectors between the two facilities. In addition, two different RELAP5–3D models of the HTTF were included in the comparison to determine if modeling methodologies notably impact the representativity results. The key figures of merit are the maximum block temperature and the coolant outlet temperature. The time-dependent maximum block temperature had a low representativity of below 0.1 for both models across both transients. The coolant outlet temperature had a higher representativity of around 0.6 for both models during the pressurized conduction cooldown, but it was below 0.2 for the depressurized conduction cooldown. Overall, the HTTF experiments were not representative of the transients in the MHTGR-350. However, these results can play a significant role in informing future potential experiments for HTGR systems that iterate on what the HTTF accomplished.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗

Computation of condition-dependent proteome allocation reveals variability in the macro and micro nutrient requirements for growth

Sustaining a robust metabolic network requires a balanced and fully functioning proteome. In addition to amino acids, many enzymes require cofactors (coenzymes and engrafted prosthetic groups) to function properly. Extensively validated resource allocation models, such as genome-scale models of metabolism and gene expression (ME-models), have the ability to compute an optimal proteome composition underlying a metabolic phenotype, including the provision of all required cofactors. Here we apply the ME-model for Escherichia coli K-12 MG1655 to computationally examine how environmental conditions change the proteome and its accompanying cofactor usage. We found that: (1) The cofactor requirements computed by the ME-model mostly agree with the standard biomass objective function used in models of metabolism alone (M-models); (2) ME-model computations reveal non-intuitive variability in cofactor use under different growth conditions; (3) An analysis of ME-model predicted protein use in aerobic and anaerobic conditions suggests an enrichment in the use of peroxyl scavenging acids in the proteins used to sustain aerobic growth; (4) The ME-model could describe how limitation in key protein components affect the metabolic state of E . coli . Genome-scale models have thus reached a level of sophistication where they reveal intricate properties of functional proteomes and how they support different E . coli lifestyles.

59 BASIC BIOLOGICAL SCIENCES↗

High-Temperature Fe-Based Fischer–Tropsch Synthesis: Experimentally Validated Kinetic Models Implemented at Pellet and Reactor Scales

Pellet- and reactor-scale models for Fischer–Tropsch synthesis (FTS) with a Fe–K/silica catalyst were developed to investigate the sensitivity of the hydrocarbon products and carbon dioxide selectivity to process conditions and feed composition at high temperature (350–400 °C), moderate pressure (1–10 bar), and a range of H 2 /CO ratios (3–1). The major objective of this paper is to develop, validate, and evaluate a high-temperature FTS model that is then used to assess the feasibility of process integration with syngas production. Since there is limited kinetic data available, in literature at these conditions, bench-scale reactor tests were conducted to obtain operational data for parameter fitting of kinetic expressions used in the model. This resulting kinetic model demonstrated agreement with the experimental data with an R 2 of 0.97 to the testing data set and, thus, was feasible to apply at pellet and reactor scales. Here, multiple pellet sizes were modeled to detail the role of transport limitations as the sphere’s diameter approached and exceeded 1 mm. Application of the reactor model indicated that hydrocarbon selectivity depended strongly on temperature, whereas the ratio of olefin to paraffin products decreased with increasing temperature, pressure, and H 2 /CO ratio. Product selectivity was not sensitive to the conversion of carbon monoxide. Furthermore, the roles of the pressure and H 2 /CO ratio were closely coupled. At a H 2 /CO ratio of 3, only slight variations in selectivity occurred over a pressure range of 1–20 bar, whereas at a ratio of 1, selectivity could vary by as much as 30% over the same pressure range. At pressures below 5 bar and temperatures above 350 °C, minimal selectivity to heavy hydrocarbons (C 12+ ) is obtained, and selectivity to midrange products (C 5–11 ) rapidly declined as pressure dropped below 5 bar, which indicated that an operational pressure of at least 5 bar is needed to achieve reasonable yields in this temperature range. These results, while tentative, provide guidelines for further experimentation and evaluation of integrated FTS processes.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Molecular Dynamics Simulations of Anion Exclusion in Clay Interlayer Nanopores

© 2016, Clays and Clay Minerals. All rights reserved. The aqueous chemistry of water films confined between clay mineral surfaces remains an important unknown in predictions of radioelement migration from radioactive waste repositories. This issue is particularly important in the case of long-lived anionic radioisotopes ( 129 I - , 99 TcO - 4 , 36 Cl - ) which interact with clay minerals primarily by anion exclusion. For example, models of ion migration in clayey media do not agree as to whether anions are completely or partially excluded from clay interlayer nanopores. In the present study, this key issue was addressed for Cl - using MD simulations for a range of nanopore widths (6 to 15 Å) overlapping the range of average pore widths that exists in engineered clay barriers. The MD simulation results were compared with the predictions of a thermodynamic model (Donnan Equilibrium model) and two pore-scale models based on the Poisson-Boltzmann equation under the assumption that interlayer water behaves as bulk liquid water. The simulations confirmed that anion exclusion from clay interlayers is greater than predicted by the pore-scale models, particularly at the smallest pore size examined. This greater anion exclusion stems from Cl - being more weakly solvated in nano-confined water than it is in bulk liquid water. Anion exclusion predictions based on the Poisson-Boltzmann equation were consistent with the MD simulation results, however, if the predictions included an ion closest approach distance to the clay mineral surface on the order of 2.0 ± 0.8 Å. These findings suggest that clay interlayers approach a state of complete anion exclusion (hence, ideal semi-permeable membrane properties) at a pore width of 4.2 ± 1.5 Å.

Tournassat, Christophe↗

Computational Modeling of Transport Limitations in Li-Air Batteries

In this paper we investigate transport limitations in the electrodes of lithium-air batteries through computational modeling. We use meso-scale models to consider the effects of dendrites on the current and potential at the anode surface, and to investigate the effects of reaction and transport parameters on the formation of precipitates in the cathode. The formation of dendrites on the anode surface during cycling reduces the transport of ions and can lead to short circuits in the cell. Growth of precipitates in the cathode reduces the specific capacity of the cell due to surface passivation and pore clogging. Both of these degradation mechanisms depend on meso-scale phenomena, such as the pore-scale reactive transport in the cathode. To understand the effects of the meso-scale transport and precipitation on the performance and lifetime of Li-air batteries, meso-scale modeling is needed that is able to resolve the electrodes and their microstructures.

Li-Air Batteries, Smoothed Particle Hydrodynamics,↗

A Multi-scale Mathematical Model of Growth and Coalescence of Bubbles Beneath the Anode in an Aluminum Reduction Cell

Modeling of bubble shapes is challenging because of the wide range of length scales in aluminum reduction cells. A 3D multi-scale mathematical model was developed to understand the nucleation, growth and coalescence of bubbles beneath the anode and to investigate the transition of bubbles from the micro- to macro level. The motion of micro-bubbles is examined using the discrete bubble model (DBM) within a Lagrangian reference. An algorithm for the transition from discrete micro-bubbles to large bubbles, which are fully resolved by the volume of fluid (VOF) approach, is achieved using a user-defined function. The two-way coupling between discrete bubbles and continuous fluids is achieved by inter-phase momentum exchange. The model involves three kinds of bubble coalescence: micro-bubble coalescence is taken into account by the DBM; large bubbles swallowing up micro-bubbles are solved by the discrete-continuum transition model; the coalescence between large bubbles is handled by the VOF method. Numerical results show that the coverage and thickness of bubbles agree well with the experimental data in the literature. Bubble thickness stops increasing when the bubble elongates with the anode bottom, which is 4.0 to 4.5 mm. Meanwhile, two asymmetrical thick heads can be observed under the anode. Bubble release frequency increases with increasing current density.

36 MATERIALS SCIENCE↗

OPER: Optimality-Guided Embedding Table Parallelization for Large-scale Recommendation Model

With the sharp increasing volume of user data, Deep Learning Recommendation Model (DLRM) becomes an indispensable infrastructure in large technology companies. However, large-scale DLRM on the multi-GPU platform is still inefficient due to unbalanced workload partitioning and intensive inter-GPU communication. To this end, we propose OPER, an OPtimality guided Embedding table placement for large-scale Recommendation model training and inference. OPER explores the potential of mitigating remote memory access latency in DLRM through fine-grained embedding table placement. Specifically, OPER proposes a theoretical modeling that builds up the relationship between EMT placement and the embedding communication latency in both training and inference. OPER proves the NP hardness of finding the optimal embedding table placement and proposes a heuristic algorithm that yields near optimal placement. OPER implements a SHMEM-based embedding table training system and a unified embedding index mapping to support fine-grained embedding table sharding and placement. Comprehensive experiments reveal that OPER achieves on average 3.4× and 5.1× speedup on training and inference respectively over state-of-the-art DLRM frameworks.

Wang, Zheng↗

New Efficient Sparse Space Time Algorithms for Superparameterization on Mesoscales

Superparameterization (SP) is a large-scale modeling system with explicit representation of small-scale and mesoscale processes provided by a cloud-resolving model (CRM) embedded in each column of a large-scale model. New efficient sparse space-time algorithms based on the original idea of SP are presented. The large-scale dynamics are unchanged, but the small-scale model is solved in a reduced spatially periodic domain to save the computation cost following a similar idea applied by one of the authors for aquaplanet simulations. In addition, the time interval of integration of the small-scale model is reduced systematically for the same purpose, which results in a different coupling mechanism between the small- and large-scale models. The new algorithms have been applied to a stringent two-dimensional test suite involving moist convection interacting with shear with regimes ranging from strong free and forced squall lines to dying scattered convection as the shear strength varies. The numerical results are compared with the CRM and original SP. It is shown here that for all of the regimes of propagation and dying scattered convection, the large-scale variables such as horizontal velocity and specific humidity are captured in a statistically accurate way (pattern correlations above 0.75) based on space-time reduction of the small-scale models by a factor of 1/3; thus, the new efficient algorithms for SP result in a gain of roughly a factor of 10 in efficiency while retaining a statistical accuracy on the large-scale variables. Even the models with 1/6 reduction in space-time with a gain of 36 in efficiency are able to distinguish between propagating squall lines and dying scattered convection with a pattern correlation above 0.6 for horizontal velocity and specific humidity. These encouraging results suggest the possibility of using these efficient new algorithms for limited-area mesoscale ensemble forecasting.

99 GENERAL AND MISCELLANEOUS↗

Effluent Management Facility Stack Monitor Qualification: Assessment with Surrogate Stacks

Sampling and monitoring at the Waste Treatment and Immobilization Plant (WTP) Effluent Management Facility (EMF) exhaust is required to meet the regulatory criteria that governs stacks that may exceed the 0.1-millirem per year threshold limit given in Title 40 of the Code of Federal Regulations, Part 61, National Emissions Standards for Hazardous Air Pollutants, Subpart H, National Emission Standard for Emissions of Radionuclides other than Radon from Department of Energy Facilities. As a result, the flow conditions and probe location must be within certain bounds as prescribed by the American National Standards Institute/Health Physics Society (ANSI/HPS) standard N13.1-1999. Compliance with the standard may be demonstrated by tests on the stack itself, or with a surrogate stack (i.e., a scale model or other geometrically similar stack). Pacific Northwest National Laboratory (PNNL) has performed scale model stack tests for 11 of the 20 emission points that are planned for the WTP. Additionally, PNNL has full scale stack results from Idaho National Laboratory and PNNL facility stacks to use as references. One or more WTP scale model stack results were expected to be appropriate as a surrogate stack to the EMF. This report describes the EMF stack monitoring location as well as an assessment of existing surrogate stacks used to qualify the EMF stack sampling location.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗

Development of a Genome-Scale Metabolic Model of Clostridium thermocellum and Its Applications for Integration of Multi-Omics Datasets and Computational Strain Design

Solving environmental and social challenges such as climate change requires a shift from our current non-renewable manufacturing model to a sustainable bioeconomy. To lower carbon emissions in the production of fuels and chemicals, plant biomass feedstocks can replace petroleum using microorganisms as biocatalysts. The anaerobic thermophile Clostridium thermocellum is a promising bacterium for bioconversion due to its capability to efficiently degrade lignocellulosic biomass. However, the complex metabolism of C. thermocellum is not fully understood, hindering metabolic engineering to achieve high titers, rates, and yields of targeted molecules. In this study, we developed an updated genome-scale metabolic model of C. thermocellum that accounts for recent metabolic findings, has improved prediction accuracy, and is standard-conformant to ensure easy reproducibility. We illustrated two applications of the developed model. We first formulated a multi-omics integration protocol and used it to understand redox metabolism and potential bottlenecks in biofuel (e.g., ethanol) production in C. thermocellum. Second, we used the metabolic model to design modular cells for efficient production of alcohols and esters with broad applications as flavors, fragrances, solvents, and fuels. The proposed designs not only feature intuitive push-and-pull metabolic engineering strategies, but also present novel manipulations around important central metabolic branch-points. We anticipate the developed genome-scale metabolic model will provide a useful tool for system analysis of C. thermocellum metabolism to fundamentally understand its physiology and guide metabolic engineering strategies to rapidly generate modular production strains for effective biosynthesis of biofuels and biochemicals from lignocellulosic biomass.

59 BASIC BIOLOGICAL SCIENCES↗