Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “reasonable accuracy”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Practical Understanding of Cancer Model Identifiability in Clinical Applications

Mathematical models are a core component in the foundation of cancer theory and have been developed as clinical tools in precision medicine. Modeling studies for clinical applications often assume an individual’s characteristics can be represented as parameters in a model and are used to explain, predict, and optimize treatment outcomes. However, this approach relies on the identifiability of the underlying mathematical models. In this study, we build on the framework of an observing-system simulation experiment to study the identifiability of several models of cancer growth, focusing on the prognostic parameters of each model. Our results demonstrate that the frequency of data collection, the types of data, such as cancer proxy, and the accuracy of measurements all play crucial roles in determining the identifiability of the model. We also found that highly accurate data can allow for reasonably accurate estimates of some parameters, which may be the key to achieving model identifiability in practice. As more complex models required more data for identification, our results support the idea of using models with a clear mechanism that tracks disease progression in clinical settings. For such a model, the subset of model parameters associated with disease progression naturally minimizes the required data for model identifiability.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Accelerated Modeling of Lithium Diffusion in Solid State Electrolytes using Artificial Neural Networks

Abstract Previous efforts to understand structure‐function relationships in high ionic conductivity materials for solid state batteries have predominantly relied on density functional theory (DFT‐) based ab initio molecular dynamics (MD). Such simulations, however, are computationally demanding and cannot be reasonably applied to large systems containing more than a hundred atoms. Here, an artificial neural network (ANN) is trained to accelerate the calculation of high accuracy atomic forces and energies used during such MD simulations. After carefully training a robust ANN for four and five element systems, nearly identical lithium ion diffusivities are obtained for Li 10 GeP 2 S 12 (LGPS) when benchmarking the ANN‐MD results with DFT‐MD. Applying the ANN‐MD approach, the effect of chlorine doping on the lithium diffusivity is calculated in an LGPS‐like structure and it is found that a dopant concentration of 1.3% maximizes ionic conductivity. The optimal concentration balances the competing consequences of effective atomic radii and dielectric constants on lithium diffusion and agrees with the experimental composition. Performing simulations at the resolution necessary to model experimentally relevant and optimal concentrations would be infeasible with traditional DFT‐MD. Systems that require a large number of simulated atoms can be studied more efficiently while maintaining high accuracy with the proposed ANN‐MD framework.

Rao, Karun K.↗
DOE OSTI2022

Very-High Dynamic Range, 10,000 Frames/Second Pixel Array Detector for Electron Microscopy

Precision and accuracy of quantitative scanning transmission electron microscopy (STEM) methods such as ptychography, and the mapping of electric, magnetic, and strain fields depend on the dose. Reasonable acquisition time requires high beam current and the ability to quantitatively detect both large and minute changes in signal. A new hybrid pixel array detector (PAD), the second-generation Electron Microscope Pixel Array Detector (EMPAD-G2), addresses this challenge by advancing the technology of a previous generation PAD, the EMPAD. The EMPAD-G2 images continuously at a frame-rates up to 10 kHz with a dynamic range that spans from low-noise detection of single electrons to electron beam currents exceeding 180 pA per pixel, even at electron energies of 300 keV. The EMPAD-G2 enables rapid collection of high-quality STEM data that simultaneously contain full diffraction information from unsaturated bright-field disks to usable Kikuchi bands and higher-order Laue zones. Test results from 80 to 300 keV are presented, as are first experimental results demonstrating ptychographic reconstructions, strain and polarization maps. We introduce a new information metric, the maximum usable imaging speed (MUIS), to identify when a detector becomes electron-starved, saturated or its pixel count is mismatched with the beam current.

36 MATERIALS SCIENCE↗
DOE OSTI2026

SALSAA: a statistical approach to line shapes from an average atom

Ion-Stark line broadening is a key density diagnostic for hot dense plasmas relevant to inertial fusion and astrophysics. It is caused by interactions of a radiating ion with nearby perturbing ions, whose electric microfields lead to changes in bound-bound transition energies. Ion-Stark broadening becomes increasingly difficult to compute for complex, many-electron ions with myriad transitions. In this paper, we propose a simplified approach to ion-Stark broadening based on self-consistent ion distributions and electronic structure from an average-atom model. We find that this approach reproduces the line shape predictions of one traditional method for high-n K-shell emission lines in aluminum ions with accuracy sufficient for density diagnostics in thermal plasmas with equal ion and electron temperatures. We expect that this approach can be extended to provide a reasonable picture of ion-Stark broadening in many-electron ions, enabling rapid calculations of line profiles in complex spectra.

average-atom↗
DOE OSTI2026

ORCHID: Orchestrated Retrieval-Augmented Classification of High-Risk Property with Intelligent Decision-Making

High-Risk Property (HRP) classification is critical at U.S. Department of Energy (DOE) sites, where inventories include sensitive and often dual-use equipment. Compliance must track evolving rules designated by various export control policies to make transparent and auditable decisions. Traditional expert-only workflows are time-consuming, backlog-prone, and struggle to keep pace with shifting regulatory boundaries. We propose ORCHID, a modular agentic framework for HRP classification that pairs retrieval-augmented generation (RAG) with human oversight to produce policy based outputs that can be audited. Small cooperating agents—retrieval, description refiner, classifier, validator, and feedback logger—coordinate via agent-to-agent messaging and invoke tools through the Model Context Protocol (MCP) for model-agnostic on-premise operation. The interface follows an "Item to Evidence to Decision" loop with step-by-step reasoning, on-policy citations, and append-only audit bundles (run-cards, prompts, evidence). In preliminary tests on real HRP cases, ORCHID improves accuracy and traceability over a non-agentic baseline while deferring uncertain items to Subject Matter Experts (SMEs). The demonstration shows single item submission, grounded citations, SME feedback capture, and exportable audit artifacts—illustrating a practical path to trustworthy LLM assistance in sensitive DOE compliance workflows.

Das, Sanjay [ORNL] (ORCID:0009000542591915)↗
DOE OSTI2020

Metallic Fuel Benchmark Simulations Based on the X430 Experiments

To gauge the accuracy of current BISON metallic fuel performance models and identify models that require further development, Oak Ridge National Laboratory has developed benchmark problems based on two pins of the X430 experiment. These benchmarks recreate the experimental conditions of the X430 series of experiments with as much accuracy as possible given the available data. The benchmarks were implemented in the BISON nuclear fuel performance code and the simulation results were compared with the experimental results to quantify the simulation accuracy. Findings in this work were consistent with previous benchmark problems. The temperature profiles, burnup, and fission gas behavior were all as accurate as can reasonably be expected. The cladding radial growth was accurate enough to not affect the results of other measurements. The fuel axial growth was accurate for U-Zr fuel but not for U-Pu-Zr fuel, which appears to be caused by a lack of U-Pu-Zr data. Swelling correlations have not been developed for U-Pu-Zr fuel because there were not enough data on this fuel type, so BISON used the U-Zr correlations. Finally, recommendations were made for near-term BISON development.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2023

Improved Fuel Particle and Compact Matrix Fission Product Release Prediction with Improved Mass Transfer Model

This study was motivated by the need to incorporate more physics-based mass transfer models in BISON in order to improve fuel particle and compact matrix fission product release prediction. This document details the integration of new modeling capabilities in BISON, including (1) development of sorption mass transfer models; improvement of thermal contact model by adding neon to the gas gap inventory, (3) implementation of PCEA and IG-110 graphite thermal models, (4) validation using Advanced Gas Reactor (AGR)-3/4 compact and capsule rings data. These new capabilities have been shown to enforce the desired physics with satisfactory accuracy. BISON’s predictions of fission product release of AGR-3/4 compacts compare favorably with PARFUME and the concentration profiles across capsule rings show reasonable trend to the experiment measurement.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2022

On the Rapid Calculation of Binding Affinities for Antigen and Antibody Design and Affinity Maturation Simulations

The accurate and efficient calculation of protein-protein binding affinities is an essential component in antibody and antigen design and optimization, and in computer modeling of antibody affinity maturation. Such calculations remain challenging despite advances in computer hardware and algorithms, primarily because proteins are flexible molecules, and thus, require explicit or implicit incorporation of multiple conformational states into the computational procedure. The astronomical size of the amino acid sequence space further compounds the challenge by requiring predictions to be computed within a short time so that many sequence variants can be tested. In this study, we compare three classes of methods for antibody/antigen (Ab/Ag) binding affinity calculations: (i) a method that relies on the physical separation of the Ab/Ag complex in equilibrium molecular dynamics (MD) simulations, (ii) a collection of 18 scoring functions that act on an ensemble of structures created using homology modeling software, and (iii) methods based on the molecular mechanics-generalized Born surface area (MM-GBSA) energy decomposition, in which the individual contributions of the energy terms are scaled to optimize agreement with the experiment. When applied to a set of 49 antibody mutations in two Ab/HIV gp120 complexes, all of the methods are found to have modest accuracy, with the highest Pearson correlations reaching about 0.6. In particular, the most computationally intensive method, i.e., MD simulation, did not outperform several scoring functions. The optimized energy decomposition methods provided marginally higher accuracy, but at the expense of requiring experimental data for parametrization. Within each method class, we examined the effect of the number of independent computational replicates, i.e., modeled structures or reinitialized MD simulations, on the prediction accuracy. We suggest using about ten modeled structures for scoring methods, and about five simulation replicates for MD simulations as a rule of thumb for obtaining reasonable convergence. We anticipate that our study will be a useful resource for practitioners working to incorporate binding affinity calculations within their protein design and optimization process.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Introducing the GeoRePORT Resource Size Tool: Reporting on Geothermal Resource Size Estimations Using the Geothermal Resource Portfolio Optimization and Reporting Technique (GeoRePORT): Preprint

The Geothermal Resource Portfolio Optimization and Reporting Technique (GeoRePORT) was developed with funding from the U.S. Department of Energy Geothermal Technologies Office to assist in identifying and pursuing long-term investment strategies through the development of a resource reporting protocol. The assessment protocols used in GeoRePORT allow for comparison of project attributes across locations and geological settings to understand the feasibility of geothermal development. This work introduces the Resource Size Tool, a new feature within the GeoRePORT package that compiles two independent methods for estimating geothermal resource size in terms of energy capacity in MW. Energy production potential for twenty-three case studies was estimated with the Resource Size Tool in order to 1) generate a reasonable range of resource size estimates for a particular geothermal field; 2) illustrate the advantages and limitations of each methodology (such as data input requirements, estimate accuracy and precision, and the appropriate circumstances of use); and 3) test the ability of the resource size tool to provide useful and accurate information for geothermal stakeholders. The tool employs two methods widely used in the geothermal industry: (1) USGS Volumetric and (2) Power Density. Results from our case studies show general overlap between these two methods in terms of resource size estimates; however, they also reveal key differences between the two approaches that should be considered when using such estimates to drive development. First, the two methods rely on different input parameters and therefore one method may be more appropriate and/or accurate for a given project than the other. Second, the Power Density method was found to generate wider ranges of resource size predictions, more consistently aligning with actual power production of the field but with larger scales of error; whereas the USGS Volumetric method predicts narrower ranges but tends to overestimate when compared to current MW production. Future work will refine variables used in the methods with input data from other sections of GeoRePORT and modify uncertainty levels based on the particular datasets used for a given project.

geological↗
DOE OSTI2021

Validation of the EDGES Low-band Antenna Beam Model

The response of the antenna is a source of uncertainty in measurements with the Experiment to Detect the Global Epoch of Reionization Signature (EDGES). We aim to validate the electromagnetic beam model of the low-band (50–100 MHz) dipole antenna with comparisons between models and against data. We find that simulations of a simplified model of the antenna over an infinite perfectly conducting ground plane are, with one exception, robust to changes in the numerical electromagnetic solver code or algorithm. For simulations of the antenna with the actual finite ground plane and realistic soil properties, we find that two out of three numerical solvers agree well. Applying our analysis pipeline to a simulated drift-scan observation from an early EDGES low-band instrument that had a 10 m × 10 m ground plane, we find residual levels after fitting and removing a five-term foreground model from the simulated data binned in local sidereal time (LST) average about 250 mK with ±40 mK variation between numerical solvers. A similar analysis of the primary 30 m × 30 m sawtooth ground plane reduced the LST-averaged residuals to about 90 mK with ±10 mK between the two viable solvers. More broadly we show that larger ground planes generally perform better than smaller ground planes. Simulated data have a power that is within 4% of real observations, a limitation of net accuracy of the sky and beam models. We observe that residual spectral structures after foreground model fits match qualitatively between simulated data and observations, suggesting that the frequency dependence of the beam is reasonably represented by the models. We find that a soil conductivity of 0.02 S m{sup -1} and relative permittivity of 3.5 yield good agreement between simulated spectra and observations. This is consistent with the soil properties reported by Sutinjo et al. for the Murchison Radio-astronomy Observatory, where EDGES is located.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2021

Computation of the Thermal Expansion Coefficient of Graphene with Gaussian Approximation Potentials

Direct experimental measurement of thermal expansion coefficient without substrate effects is a challenging task for two-dimensional (2D) materials, and its accurate estimation with large-scale ab initio molecular dynamics is computationally very expensive. Machine learning-based interatomic potentials trained with ab initio data have been successfully used in molecular dynamics simulations to decrease the computational cost without compromising the accuracy. In this work, we investigated using Gaussian approximation potentials to reproduce the density functional theory-level accuracy for graphene within both lattice dynamical and molecular dynamical methods, and to extend their applicability to larger length and time scales. Two such potentials are considered, GAP17 and GAP20. GAP17, which was trained with pristine graphene structures, is found to give closer results to density functional theory calculations at different scales. Further vibrational and structural analyses verify that the same conclusions can be deduced with density functional theory level in terms of the reasoning of the thermal expansion behavior, and the negative thermal expansion behavior is associated with long-range out-of-plane phonon vibrations. Thus, it is argued that the enabled larger system sizes by machine learning potentials may even enhance the accuracy compared to small-size-limited ab initio molecular dynamics.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Quantum-Chemically Informed Machine Learning: Prediction of Energies of Organic Molecules with 10 to 14 Non-hydrogen Atoms

High-fidelity quantum-chemical calculations can provide accurate predictions of molecular energies, but their high computational costs limit their utility, especially for larger molecules. We have shown in previous work that machine learning models trained on high-level quantum-chemical calculations (G4MP2) for organic molecules with one to nine non-hydrogen atoms can provide accurate predictions for other molecules of comparable size at much lower costs. Here we demonstrate that such models can also be used to effectively predict energies of molecules larger than those in the training set. To implement this strategy, we first established a set of 191 molecules with 10–14 non-hydrogen atoms having reliable experimental enthalpies of formation. We then assessed the accuracy of computed G4MP2 enthalpies of formation for these 191 molecules. The error in the G4MP2 results was somewhat larger than that for smaller molecules, and the reason for this increase is discussed. Two density functional methods, B3LYP and ωB97X-D, were also used on this set of molecules, with ωB97X-D found to perform better than B3LYP at predicting energies. The G4MP2 energies for the 191 molecules were then predicted using these two functionals with two machine learning methods, the FCHL-Δ and SchNet-Δ models, with the learning done on calculated energies of the one to nine non-hydrogen atom molecules. The better-performing model, FCHL-Δ, gave atomization energies of the 191 organic molecules with 10–14 non-hydrogen atoms within 0.4 kcal/mol of their G4MP2 energies. Thus, this work demonstrates that quantum-chemically informed machine learning can be used to successfully predict the energies of large organic molecules whose size is beyond that in the training set.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Small-𝑥 gluon GPD constrained from deeply virtual 𝐽/𝜓 production and gluon PDF through universal-moment parametrization

We phenomenologically constrain the small-𝑥 and small-𝜉 gluon generalized parton distributions (GPDs) with the deeply virtual 𝐽/𝜓 production (DV⁢𝐽/𝜓⁢P) in the framework of GPDs through universal moment parameterization (GUMP). We use a hybrid cross-section formula combining collinear factorization to the next-to-leading order (NLO) accuracy of the strong coupling 𝛼 𝑠 , with corrections from nonrelativistic QCD to account for the power corrections due to the heavy 𝐽/𝜓 mass. We reach reasonable fit to the measured differential cross sections of DV⁢𝐽/𝜓⁢P by H1 at the Hadron-Electron Ring Accelerator as well as forward gluon parton distribution functions (PDFs) from JAM22 global analysis. We find that both NLO and nonrelativistic corrections are significant for heavy vector meson productions. Of course, the gluon GPD we obtain still contains considerable freedom in need of inputs from other constraints, particularly in the distribution-amplitude-like region.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Demystifying the Resilience of Large Language Models: An End-to-End Perspective

Deep neural networks are known to be resilient to random bit-wise faults in their parameters. However, this resilience has primarily been established through evaluations of classification models. The extent to which this claim holds for large-language models remains underexplored. In this work, we conduct an extensive measurement study on the impact of random bitwise faults in commercial-scale language models. We perform an in-depth analysis of the resulting generation outputs. We first expose that these language models are not truly resilient to random bit-flips. While aggregate metrics such as accuracy may suggest resilience, an in-depth inspection of the generated outputs shows significant degradation in text quality. Our analysis also shows that tasks requiring more complex reasoning suffer more from performance and quality degradation. Moreover, we extend our analysis to models with augmented reasoning capabilities, such as Chain-of-Thought or Mixture of Experts architectures, and characterize their failure scenarios under random bit-flips.

Sun, Yu↗
DOE OSTI2024

Smooth Dispersion Is Physically Appropriate: Assessing and Amending the D4 Dispersion Model

The addition of dispersion corrections to density functionals is essential for accurate energy and geometry predictions. Among them, the D4 scheme is popular due to its low computational cost and high accuracy. However, due to its design, the D4 correction can occasionally lead to anomalies, such as unphysical curvature and bumps in the potential energy surface. We find these anomalies are common in the D4 model, although observable consequences are rarer than in the D3 model for reasons we explain. Nevertheless, we uncover instances of unphysical local minima and stationary points with the D4 scheme and propose two solutions that yield smoother dispersion energy as a function of nuclear position. One is trivial to implement, based on a smoother reparametrization of Gaussian weighting (D4S) to find the effective coordination number. The other replaces Gaussian weighting with soft linear interpolation (D4SL). These new approaches usually remove artificial extremum points, while maintaining accuracy.

Tkachenko, Nikolay V↗
DOE OSTI2022

Neural network based fast prediction of β N limits in HL-2M

Artificial neural networks (NNs) are trained, based on the numerical database, to predict the no-wall and ideal-wall β N limits, due to onset of the n = 1 ( n is the toroidal mode number) ideal external kink instability, for the HL-2M tokamak. The database is constructed by toroidal computations utilizing both the equilibrium code CHEASE (Lütjens et al 1992 Comput. Phys. Commun. 69 287) and the stability code MARS-F (Liu et al 2000 Phys. Plasmas 7 3681). The stability results show that (1) the plasma elongation generally enhances both β N limits, for either positive or negative triangularity plasmas; (2) the effect is more pronounced for positive triangularity plasmas; (3) the computed no-wall β N limit linearly scales with the plasma internal inductance, with the proportionality coefficient ranging between 1 and 5 for HL-2M; (4) the no-wall limit substantially decreases with increasing pressure peaking factor. Furthermore, both the NN model and the convolutional neural network (CNN) model are trained and tested, producing consistent results. The trained NNs predict both the no-wall and ideal-wall limits with as high as 95% accuracy, compared to those directly computed by the stability code. Additional test cases, produced by the Tokamak Simulation Code (Jardin et al 1993 Nucl. Fusion 33 371), also show reasonable performance of the trained NNs, with the relative error being within 10%. The constructed database provides effective references for the future HL-2M operations. The trained NNs can be used as a real-time monitor for disruption prevention in the HL-2M experiments, or serve as part of the integrated modeling tools for ideal kink stability analysis.

Physics↗
DOE OSTI2023

Optimization to Generate Equations of State for Hydrogen Production

On a high level, the larger project in question, HydroGEN, aims to develop software used for finding equations of state (EOS) to optimize catalyst configuration for H 2 production through water splitting. In particular, this summer project focused on solving the nonlinear equations used in fitting the equations. This problem involved using Python to solve a linear system with nonlinear constraints. In order for this to be achieved, Pyomo was used to build a model and the solver Ipopt, interior point optimizer, was used. Pyomo is a Python-based language developed at Sandia; it is an optimization modeling language. Rather than solving the entire problem at once, a toy problem was created, simplifying the problem down to the most important focus. This problem had a known solution, comparable to the calculated solution to assess accuracy and as progress was made towards finding solutions, complexity was gradually added to the problem. After building and solving the toy problem, it was found that it gave reasonably accurate solutions, better compared to the two existing solvers previously used with this project in terms of functionality. The solver is now ready for implementation into the project’s main software.

08 HYDROGEN↗
DOE OSTI2021

An Overview of the Usefulness of Machine Learning Techniques on Network Packet Data

Understanding the health and behavior of a computer network allows for better network efficiency and security. We present an overview of various machine learning techniques for classifying network packet data via packet metadata. While some classical machine learning approaches achieve reasonable results, the most accurate classification can be achieved with deep learning. On the four data sets studied herein, a basic deep learning model achieved at or near 100\% classification accuracy. We also propose a method for determining variable importance as a means for potential transfer learning applications to classifying yet unseen network packet data.

97 MATHEMATICS AND COMPUTING↗