SEARCH · Engineering Papers

Results for “Task Performance and Analysis”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,081 records · Page 60

All sources NASA NTRS DOE OSTI

DOE OSTI2020

Model reduction for hypersonic aerodynamics via conservative LSPG projection and hyper-reduction

High-speed aerospace engineering applications rely heavily on computational fluid dynamics (CFD) models for design and analysis due to the expense and difficulty of flight tests and experiments. This reliance on CFD models necessitates performing accurate and reliable uncertainty quantification (UQ) of the CFD models. However, it is very computationally expensive to run CFD for hypersonic flows due to the fine grid resolution required to capture the strong shocks and large gradients that are typically present. Furthermore, UQ approaches are “many-query” problems requiring many runs with a wide range of input parameters. One way to enable computationally expensive models to be used in such many-query problems is to employ projection-based reduced-order models (ROMs) in lieu of the (high-fidelity) full-order model. In particular, the least-squares Petrov–Galerkin (LSPG) ROM (equipped with hyper-reduction) has demonstrated the ability to significantly reduce simulation costs while retaining high levels of accuracy on a range of problems including subsonic CFD applications. This allows computationally inexpensive LSPG ROM simulations to replace the full-order model simulations in UQ studies, which makes this many-query task tractable, even for large-scale CFD models. This work presents the first application of LSPG to a hypersonic CFD application. In particular, we present results for LSPG ROMs of the HIFiRE-1 in a three-dimensional, turbulent Mach 7.1 flow, showcasing the ability of the ROM to significantly reduce computational costs while maintaining high levels of accuracy in computed quantities of interest.

Fluid Dynamics↗

DOE OSTI2023

Towards Lightweight Data Integration Using Multi-Workflow Provenance and Data Observability

Modern large-scale scientific discovery requires multidisciplinary collaboration across diverse computing facilities, including High Performance Computing (HPC) machines and the Edge-to-Cloud continuum. Integrated data analysis plays a crucial role in scientific discovery, especially in the current AI era, by enabling Responsible AI development, FAIR, Reproducibility, and User Steering. However, the heterogeneous nature of science poses challenges such as dealing with multiple supporting tools, cross-facility environments, and efficient HPC execution. Building on data observability, adapter system design, and provenance, we propose MIDA: an approach for lightweight runtime Multi-workflow Integrated Data Analysis. MIDA defines data observability strategies and adaptability methods for various parallel systems and machine learning tools. With observability, it intercepts the dataflows in the background without requiring instrumentation while integrating domain, provenance, and telemetry data at runtime into a unified database ready for user steering queries. We conduct experiments showing end-to-end multi-workflow analysis integrating data from Dask and MLFlow in a real distributed deep learning use case for materials science that runs on multiple environments with up to 276 GPUs in parallel. We show near-zero overhead running up to 100,000 tasks on 1,680 CPU cores on the Summit supercomputer.

Santos Souza, Renan↗

DOE OSTI2016

CFD and X-ray Analysis of Gaseous Direct Injection from an Outward Opening Injector

Using natural gas in an internal combustion engine (ICE) is emerging as a promising way to improve thermal efficiency and reduce exhaust emissions. In the development of such engine platforms, computational fluid dynamics (CFD) plays a fundamental role in the optimization of geometries and operating parameters. One of the most relevant issues in the simulation of direct injection (DI) gaseous processes is the accurate prediction of the gas jet evolution. The simulation of the injection process for a gaseous fuel does not require complex modeling, nevertheless properly describing high-pressure gas jets remains a challenging task. At the exit of the nozzle, the injected gas is under-expanded, the flow becomes supersonic and shocks occur due to compressibility effects. These phenomena lead to challenging computational requirements resulting from high grid resolution and low computational time-steps. In this paper, the simulation of gaseous injection from an outward opening injector was carried out with the commercial CFD software CONVERGE. Argon was injected into a quiescent chamber filled with Nitrogen. Two free-jet cases with injection pressure of 15 and 10 bar, and ambient pressure of 1 and 2 bar respectively, were tested. For the 15 bar injection pressure and 1 bar ambient pressure, a jet impinging on a barrier case was performed as well. The numerical results were validated against experimental mass measurements using X-ray radiography. The results, presented in terms of 2D maps and transverse sections of equivalent path length of argon, showed good agreement between modeling and experiments in terms of jet penetration, shape and gas mass distribution. These emphasized the significant impact of mesh size on the results for the simulation of this type of injector. Due to the small geometrical dimensions a fine mesh is required in order to deliver an accurate description of the mixture formation. Numerical results were discussed with a particular emphasis on the application to engine simulations, where a trade-off between accuracy and computational time is a necessity.

Bartolucci, Lorenzo↗

DOE OSTI2021

CST R9120-B 30x60 Media Pretreatment and Digestion Standard Development to Support Tank Closure Cesium Removal 1A (TCCR-1A) Operations

In support of continued operations of Savannah River Remediation’s (SRR) ion exchange process to remove radioactive cesium from Savannah River Site (SRS) dissolved saltcake, referred to as the Tank Closure Cesium Removal 1A process (TCCR-1A; containing the second set of at-tank ion exchange columns), SRNL was tasked with: 1) validating the field protocol planned for pretreating the crystalline silicotitanate (CST) ion exchange media to be used in the columns, 2) preparing CST media for use in in-tank batch contact testing, and 3) developing a digestion standard to verify complete CST dissolution prior to characterization following each in-tank batch contact test. A procedure (SRNL L29 Manual, ITS-0229) was developed previously in support of the TCCR Demonstration (which contained the first set of ion exchange columns) based on evaluations of a different CST production lot with a larger particle size range to document the protocol for pretreating CST utilizing conditions similar to what will be performed in the field.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2023

State of the Art of Molecular Visualization in Immersive Virtual Environments

Abstract Visualization plays a crucial role in molecular and structural biology. It has been successfully applied to a variety of tasks, including structural analysis and interactive drug design. While some of the challenges in this area can be overcome with more advanced visualization and interaction techniques, others are challenging primarily due to the limitations of the hardware devices used to interact with the visualized content. Consequently, visualization researchers are increasingly trying to take advantage of new technologies to facilitate the work of domain scientists. Some typical problems associated with classic 2D interfaces, such as regular desktop computers, are a lack of natural spatial understanding and interaction, and a limited field of view. These problems could be solved by immersive virtual environments and corresponding hardware, such as virtual reality head‐mounted displays. Thus, researchers are investigating the potential of immersive virtual environments in the field of molecular visualization. There is already a body of work ranging from educational approaches to protein visualization to applications for collaborative drug design. This review focuses on molecular visualization in immersive virtual environments as a whole, aiming to cover this area comprehensively. We divide the existing papers into different groups based on their application areas, and types of tasks performed. Furthermore, we also include a list of available software tools. We conclude the report with a discussion of potential future research on molecular visualization in immersive environments.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

FORCE Update 2024

The Framework for Optimization of Resources and Economics (FORCE) tool suite is the U.S. Department of Energy’s Nuclear Integrated Energy Systems (IES) Program flagship tool suite for technoeconomic IES analysis of IES. This tool suite is useful for analysis designed to evaluate and improve the technoeconomics of energy production systems, particularly for systems including nuclear technology. In this report, we document the development activity for the FORCE tool suite to extend its capabilities as performed during fiscal year 2024. In addition to reliability and accessibility, capability is one of the three standards guiding the development of the FORCE tool suite and the software codes that are its constituent parts. Extending the capabilities of the FORCE tool suite allows analysis both within the IES program as well as industry, university, and laboratory partners to perform analysis with more accuracy, insight, and impactful narrative. Four areas of capability development were the focus of activity this year: economic parameter uncertainty quantification, multiresolution analysis, components-to-optimization workflow automation, and statespace construction workflows for real-time optimal control. In economic parameter uncertainty quantification, the ability of HERON to capture risk due to scenarios (weather and energy demand uncertainty) was expanded to also include uncertainties in financial parameters such as capital cost or operation and maintenance costs. By including these sources of uncertainty, which are sometimes very large compared with scenario uncertainty, HERON is better able to capture the risk posed by investment in various IES technology. Because of this, analysts can also consider the reduction in risks that can be realized by choice of some technologies. In multiresolution analysis, development activity extended on work completed previously. In fiscal year 2023, methods for decomposing time series signals, such as demand, solar and wind availability, and price profiles, were analyzed and down-selected to those most effective at splitting signals into different resolutions. These resolutions allow considering the influence of different energy demand and supply behaviors across different time scales. For example, energy demand might be divided into seasonal, weekly, and hourly profiles. In fiscal year 2024, this preliminary work was extended and implemented within the Risk Analysis Virtual Environment (RAVEN) risk and uncertainty analysis platform, which is used throughout the FORCE framework. This development of the “multi-resolution time series analysis” (MR-TSA) module in RAVEN allows training synthetic history generators on complex time series. These synthetic history generators can then be used in HERON for generating scenarios that represent possible market and weather scenarios that can be analyzed on different time scales. We envision completing this work in the future, implementing multiresolution dispatch optimization strategies that can make the most beneficial use of these stratified time histories. In components-to-optimization workflow development, workflows for translating user inputs of components into algorithms for algebraic optimization were selected and implemented. Similar algorithms within the Holistic Energy Resource Optimization Network (HERON) were separated from the main code base of HERON and gathered with the components-to-optimization workflows in the new Dispatch Optimization Variable Engine (DOVE) software library. This modularization allows FORCE users to analyze dispatch optimization and energy system duty cycles independently of HERON, which previously was a burdensome task. Additionally, these dispatch optimization algorithms, set up in an independent library, can now be used across all software applications within FORCE, especially including the real-time optimal control software Optimization of Real-time Capacity Allocation (ORCA). Allowing FORCE software to share dispatch optimization algorithms within a single library allows for improved software maintenance and reliability. In statespace characterization workflow development, alternative workflows for optimizing dispatch with additional technical accuracy was the focus, particularly to improve the real-time optimization decision making in ORCA. Using algorithms and workflows initially developed for the Feasible Actuator Range Modifier (FARM), workflows for determining the statespace representation of IES were identified and demonstrated. The resulting dispatch optimization required a more robust optimization algorithm than that originally used in HERON (and moved to DOVE), which required adding an alternate workflow to DOVE that can more accurately match the behavior of physical systems using a partial differential equation representation. In conclusion, capability developments in the FORCE tool suite in fiscal year 2024 have improved the ability of the FORCE tool suite to perform

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2016

Alternative uses of a megavolt tandem accelerator for few-keV studies with ion-source SIMS monitoring

We increase the versatility of a tandem electrostatic accelerator by implementing simple modifications to the standard operation procedure. While keeping its ability to deliver MeV ion beams, we show that the experimental setup can (i) provide good quality ion beams in the few-keV energy range and (ii) be used to study ion-beam surface modification with simultaneous secondary ion mass spectrometry. This latter task is accomplished without using any chamber connected to the accelerator exit. We perform mass spectrometry of the few-keV anions produced in the ion source by measuring their neutral counterparts at the accelerator exit with energies up to 1.7 MeV. With an additional modification, a high-current few-keV regime is obtained, using the ion source as an irradiation chamber and the accelerator itself only as a mass spectrometer. As an example of application, we prepare a sample for the study of ion-beam assisted dewetting of a thin Au film on a Si substrate.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2023

Process-Oriented Diagnostics: Principles, Practice, Community Development, and Common Standards

Process-oriented diagnostics (PODs) aim to provide feedback for model developers through model analysis based on physical hypotheses. However, the step from a diagnostic based on relationships among variables, even when hypothesis driven, to specific guidance for revising model formulation or parameterizations can be substantial. The POD may provide more information than a purely performance-based metric, but a gap between POD principles and providing actionable information for specific model revisions can remain. Furthermore, in coordinating diagnostics development, there is a trade-off between freedom for the developer, aiming to capture innovation, and near-term utility to the modeling center. Best practices that allow for the former, while conforming to specifications that aid the latter, are important for community diagnostics development that leads to tangible model improvements. Promising directions to close the gap between principles and practice include the interaction of PODs with perturbed physics experiments and with more quantitative process models as well as the inclusion of personnel from modeling centers in diagnostics development groups for immediate feedback during climate model revisions. Examples are provided, along with best-practice recommendations, based on practical experience from the NOAA Model Diagnostics Task Force (MDTF). Common standards for metrics and diagnostics that have arisen from a collaboration between the MDTF and the Department of Energy’s Coordinated Model Evaluation Capability are advocated as a means of uniting community diagnostics efforts.

58 GEOSCIENCES↗

DOE OSTI2023

Requirements Description of GAMSOR and GAMSRC

This report presents the modeling and simulation capabilities of Argonne National Laboratory’s GAMSOR and GAMSRC code that is used in present reactor analysis activities. These capabilities will then be used to establish the set of verification tasks necessary to verify GAMSOR for use on commercial projects. A similar approach was taken for the PERSENT, REBUS and DIF3D software packages. The GAMSOR program is a modified version of DIF3D used to obtain the coupled neutron gamma flux distributions for a given reactor system. GAMSOR was first developed in the 1980s and its work has progressed in parallel with DIF3D throughout its various development phases. While GAMSOR is compatible with DIF3D-FD and DIF3D-Nodal, DIF3D-VARIANT is most commonly used today and will be the focus of the present verification effort. As covered in [1], GAMSOR was built to provide more accurate heating distributions by explicitly solving for both the gamma and neutron heating distributions. Without this consideration, gamma heat is normally assumed to be deposited locally at the site of gamma production. This is acceptable for modeling some phenomena because the total energy production is preserved, but it can lead to inaccurate approximations of fluence-dependent behaviors such as fuel performance, swelling, cladding embrittlement, radiation induced creep, and power peaking. The outputs from GAMSOR are contained in multiple intermediate and final output files: FIXSRC, VARSRC, GTFLUX, GHFLUX, PWDINT, NPDINT, and GPDINT. DIF3D normally produces a PWDINT file containing power density results, with the assumption that gamma rays are absorbed where they are emitted (i.e., energy deposition is local with no modeled transport). With GAMSOR, neutron power density (NPDINT) and gamma power density (GPDINT) are both produced, with their summed power densities replacing the results in PWDINT for total heating (i.e., recoverable) power density.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2017

A new APDS approach - alarm processing based on alarm prioritization - 201

We take for granted that the advances in technology bring improvements to the conditions in which humans perform activities associated to them. Is that the case for digital alarm systems in nuclear power plants? The answer is not that simple. A review of the related literature, shows that still, the single most significant problem with alarm systems, is the large number of alarm messages presented to the operator at once. This is an issue that affects analog I and C control rooms with conventional annunciators and gets aggravated in digital I and C based control rooms. Digital alarm systems present the alarms in the form of lists in computer screens. This kind of presentation forces the operator to scroll or navigate pages to review the active alarms. Also, the alarm lists do not allow the operator to apply the pattern recognition practices acquired during training when using with fixed position alarm systems. An Alarm Processing and Diagnostic System (APDS) facilitates the alarm monitoring task by applying alarm reduction techniques based on alarm priorities. It allows the operators to focus on the most important alarms during plant upsets. A new generation of APDS should be based in an improved alarm analysis methodology. To develop such methodology, Tecnatom reviewed applicable standards and our own alarm analysis experiences. This paper documents the process and conclusions of the methodology development. (authors)

42 ENGINEERING↗

DOE OSTI2017

Three-Dimensional Dosimetric Validation of a Magnetic Resonance Guided Intensity Modulated Radiation Therapy System

Purpose: To validate the dosimetric accuracy of a commercially available magnetic resonance guided intensity modulated radiation therapy (MRgIMRT) system using a hybrid approach: 3-dimensional (3D) measurements and Monte Carlo calculations. Methods and Materials: We used PRESAGE radiochromic plastic dosimeters with remote optical computed tomography readout to perform 3D high-resolution measurements, following a novel remote dosimetry protocol. We followed the intensity modulated radiation therapy commissioning recommendations of American Association of Physicists in Medicine Task Group 119, adapted to incorporate 3D data. Preliminary tests (“AP” and “3D-Bands”) were delivered to 9.5-cm usable diameter cylindrical PRESAGE dosimeters to validate the treatment planning system (TPS) for nonmodulated deliveries; assess the sensitivity, uniformity, and rotational symmetry of the PRESAGE dosimeters; and test the robustness of the remote dosimetry protocol. Following this, 4 clinical MRgIMRT plans (“MultiTarget,” “Prostate,” “Head/Neck,” and “C-Shape”) were measured using 13-cm usable diameter PRESAGE dosimeters. For all plans, 3D-γ (3% or 3 mm global, 10% threshold) passing rates were calculated and 3D-γ maps were examined. Point doses were measured with an IBA-CC01 ionization chamber for validation of absolute dose. Finally, by use of an in-house-developed, GPU-accelerated Monte Carlo algorithm (gPENELOPE), we independently calculated dose for all 6 Task Group 119 plans and compared against the TPS. Results: For PRESAGE measurements, 3D-γ analysis yielded passing rates of 98.7%, 99.2%, 98.5%, 98.0%, 99.2%, and 90.7% for AP, 3D-Bands, MultiTarget, Prostate, Head/Neck, and C-Shape, respectively. Ion chamber measurements were within an average of 0.5% (±1.1%) from the TPS dose. Monte Carlo calculations demonstrated good agreement with the TPS, with a mean 3D-γ passing rate of 98.5% ± 1.9% using a stricter 2%/2-mm criterion. Conclusions: We have validated the dosimetric accuracy of a commercial MRgIMRT system using high-resolution 3D techniques. We have demonstrated for the first time that hybrid 3D remote dosimetry is a comprehensive and feasible approach to commissioning MRgIMRT. This may provide better sensitivity in error detection compared with standard 2-dimensional measurements and could be used when implementing complex new magnetic resonance guided radiation therapy technologies.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2020

Status Update on the High Precision Isotopic Measurements on High Burnup LWR Fuel in 2020

The US Department of Energy (DOE) Office of Nuclear Energy (NE) is currently investigating the feasibility of directly disposing dual-purpose (storage and transportation) canisters (DPCs) in a spent nuclear fuel (SNF) repository. Criticality during the repository performance period (10,000 years or more) is one of the major concerns related to direct disposal of DPCs, specifically as the system undergoes degradation in the repository environment and timeframe. Oak Ridge National Laboratory (ORNL) is developing an as-loaded criticality analysis methodology using full (actinides + fission products) burnup credit that exploits the inherent criticality margin associated with actual canister-specific loading configuration. Burnup credit criticality analysis requires validation of the depletion/decay codes used to generate the burned isotopic inventory of an assembly by comparing the code-predicted inventory with experimentally determined isotopic data. Currently, isotopic measurement data for boiling water reactors (BWR) SNF are limited, and additional measurements will be highly beneficial for BWR burnup credit analysis, which is essential to demonstrate disposability of BWR DPCs. Moreover, new pressurized water reactor (PWR) samples of isotopic measurements will expand the PWR sample population and consequently will greatly improve sample statistics. In turn, this will reduce uncertainty in the computational determination of the isotopic composition of commercial SNF and eliminate additional penalties currently used for lack of data. As part of the isotopic analysis task, eight diverse, high burnup (HBU) samples from several PWR rods are being dissolved and isotopically analyzed to provide high-quality measurement data for PWR SNF to reducing the uncertainties associated with PWR isotopic depletion validation. The isotopics of interest focus on nuclides important to burnup credit, shielding, and decay heat in PWR SNF. Additionally, eight high burnup BWR fuel samples from the Limerick nuclear power plant have been identified for measurement to support reducing uncertainties associated with BWR isotopic depletion validation.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2023

Evaluation of Minimum Detectable Activities for Stack Sample Analyses

The Rad-NESHAP program takes samples at major point sources to track how much of select nuclides are being emitted out of the stack. Samples are taken at the site where the radiological operations are being performed. The samples are then sent to analytical laboratories for analysis to determine what nuclides exist in the sample. However, due to the sensitivity of the instruments, false positives are often reported. To reduce the number of false positives, the results reported from the analytical labs are compared to set detection limits. Anything below the set detection limit is considered part of the background while anything above the limit is considered a positive hit. Therefore, the detection limit must be set low enough that any false negatives do not result in a significant dose missed at the end of the year but high enough that there are not enough false positives to skew the annual dose being reported. All detection limits listed in this section for the analyses below were designed to meet the stated equivalent emissions rate and subsequent off-site doses. These minimum detectable activities (MDAs) were developed in the 1990s and documented in the memo cited for each analysis below. These limits and memos are all currently referenced in the Quality Assurance Project Plan (QAPP) for the Rad-NESHAP Compliance Task, which is being updated to a Project Implementation Plan (PIP) concurrently with this document. As part of this update process, the original MDAs defined in this section needed to be verified that they were still conservative and meet the off-site dose levels stated.

2019 MDA↗

DOE OSTI2026

Regularization via f -Divergence: An Application to Multi-Oxide Spectroscopic Analysis

In this paper, we explore the application of convolutional neural networks (CNNs) for predicting the chemical composition of complex geologic samples in a simulated Martian atmospheric environment. Specifically, we aim to characterize oxide weight percentages (wt.%) of rock samples analyzed by remote Laser-Induced Breakdown Spectroscopy (LIBS), framing the problem as a multi-target regression task . Neural networks trained on LIBS spectra are prone to overfitting due to high spectral complexity, limited labeled data, and measurement noise. While regularization is critical for improving generalization, common methods (e.g., ℓ 2 regularization) impose constraints not directly tied to data distribution properties. We propose a novel regularization method based on a specific ƒ-divergence induced by a graph-based estimator, designed to constrain the distributional discrepancy between predictions and targets. This regularizer serves a dual purpose: (a) mitigating overfitting by enforcing a constraint on the distributional difference between predictions and noisy targets, and (b) acting as an auxiliary loss that penalizes large divergences. To enable backpropagation, we develop a differentiable approximation of this particular ƒ-divergence, making the method feasible for neural networks. Experiments on ChemCam and SuperCam LIBS calibration spectra show that mathematical equation-divergence regularization outperforms or matches standard regularization methods (ℓ 1 , ℓ 2 , dropout) and the classical baseline, partial least squares (PLS). Combining ƒ-divergence regularization with standard regularization yields further performance gains, indicating that distributional regularization is useful in this context giving a promising direction for robust model training in planetary science applications. Source code is publicly available at Klein and Li (2025), https://doi.org/10.11578/dc.20250530.7.

58 GEOSCIENCES↗

DOE OSTI2024

Solving the Orszag–Tang vortex magnetohydrodynamics problem with physics-constrained convolutional neural networks

We study the 2D Orszag–Tang vortex magnetohydrodynamics (MHD) problem through the use of physics-constrained convolutional neural networks (PCNNs) for forecasting the density, ρ, and the magnetic field, B, as well as the prediction of B given the velocity field v of the fluid. In addition to translation equivariance from the convolutional architecture, other physics constraints were embedded: absence of magnetic monopoles, non-negativity of ρ, use of only relevant variables, and the periodic boundary conditions of the problem. The use of only relevant variables and the hard constraint of non-negative ρ were found to facilitate learning greatly. The divergenceless condition ∇·B=0 was implemented as a hard constraint up to machine precision through the use of a magnetic potential to define B=∇×A. Residual networks and data augmentation were also used to improve performance. This allowed for some of the residual models to function as surrogate models and provide reasonably accurate simulations. For the prediction task, the PCNNs were evaluated against a physics-informed neural network, which had the ideal MHD induction equation as a soft constraint. Several models were able to generate highly accurate fields, which are visually almost indistinguishable and have low mean squared error. Only methods with built-in hard constraints produced physical fields with ∇·B=0. The use of PCNNs for MHD has the potential to produce physically consistent real-time simulations to serve as virtual diagnostics in cases where inferences must be made with limited observables.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2024

Pretraining Billion-Scale Geospatial Foundational Models on Frontier

As AI workloads increase in scope, generalization capability becomes challenging for small task-specific models and their demand for large amounts of labeled training samples increases. On the contrary, Foundation Models (FMs) are trained with internet-scale unlabeled data via self-supervised learning and have been shown to adapt to various tasks with minimal fine-tuning. Although large FMs have demonstrated significant impact in natural language processing and computer vision, efforts toward FMs for geospatial applications have been restricted to smaller size models, as pretraining larger models requires very large computing resources equipped with state-of-the-art hardware accelerators. Current satellite constellations collect 100+TBs of data a day, resulting in images that are billions of pixels and multimodal in nature. Such geospatial data poses unique challenges opening up new opportunities to develop FMs. We investigate billion scale FMs and HPC training profiles for geospatial applications by pretraining on publicly available data. We studied from end-to-end the performance and impact in the solution by scaling the model size. Our larger 3B parameter size model achieves up to 30% improvement in top1 scene classification accuracy when comparing a 100M parameter model. Moreover, we detail performance experiments on the Frontier supercomputer, America's first exascale system, where we study different model and data parallel approaches using PyTorch's Fully Sharded Data Parallel library. Specifically, we study variants of the Vision Transformer architecture (ViT), conducting performance analysis for ViT models with size up to 15B parameters. By discussing throughput and performance bottlenecks under different parallelism configurations, we offer insights on how to leverage such leadership-class HPC resources when developing large models for geospatial imagery applications.

Tsaris, Aristeidis (aris)↗

DOE OSTI2024

UHT-CAMANCHE: Ultra-High Temperature Ceramic Additively Manufactured Compact Heat Exchangers

The conceptual basis for this project is the convergence of advanced ultra-high temperature ceramic materials and additive manufacturing technologies to produce compact ceramic heat exchangers with complex internal flow path geometries. Task areas were broadly divided into materials and manufacturing development, heat exchanger design, component testing, and techno-economic analysis. Technical challenges included the design and commissioning of new test facilities, improving feature resolution and deposition rate of ceramic additive manufacturing techniques, establishing process-structure-property relationships in additively manufactured ultra-high temperature ceramics, and assessing high temperature materials compatibility in CO 2 environments. The primary candidate material evaluated in this work is a composite comprising zirconium diboride (ZrB2) with 30 vol. % silicon carbide (SiC) which was selected based on its desirable combination of high temperature mechanical properties, high thermal and electrical conductivities, and oxidation resistance. High solids loaded ZrB2-SiC pastes suitable for extrusion-based additive manufacturing were developed for the first time as part of this work. Materials compatibility studies indicate this material oxidizes in CO 2 to form a protective borosilicate scale which transforms to pure silica above 1000°C. Parts made by additive manufacturing displayed enlarged grain sizes produced by pressureless sintering as compared to hot-press sintering. Increases in microstructural coarseness have outsized effect on oxidation performance up to 1400°C due to incomplete oxidation of coarse large diameter SiC particles resulting in lower amounts of silica that apparently inhibit protective scale formation. Additive manufacturing as a forming technique did not appear to significantly affect thermal conductivity, hardness, or elastic modulus, though flexural strength was reduced by half or more as compared to traditionally hot-pressed materials. This effect was attributed to the presence of strength-limiting flaws (ca. 40 microns in size) originating from extrudate inhomogeneities that could potentially be eliminated with further process improvements. Attempts to attain economies of scale for production of multi-kilowatt scale heat exchangers by ceramic additive manufacturing proved difficult. Lack of automation and a modest extrusion rate while retaining fine feature resolution made the overall process labor intensive and limited experimental throughput. A number of full-scale components were taken through post-process heat treatments including drying, binder burnout, and sintering; however, none survived without significant flaws or cracks. Therefore, no operational data from a newly installed heat exchanger test loop were able to be obtained during the performance period. Continued research and development is recommended to improve economic feasibility of ceramic additive manufacturing by standardizing the use of advanced sensors, artificial intelligence, and automation tools to reduce associated labor costs and accelerate production rates. The materials and manufacturing techniques demonstrated in this work are likely to find applications in defense and energy applications.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2026

Soil Water Retention and Hydraulic Conductivity Data and Model at Pump House in East River Watershed, Colorado 2019-2024

This data package includes soil water retention and hydraulic conductivity data and model fitting results from measurements of ex-situ soil samples and in-situ soil sensors near Pump House at Mount Crested Butte in the East River Watershed. Soil water retention curves (SWRC) characterize soil water content as a function of soil water potential. SWRC depends on soil texture and pore structure and can be used to describe the constraints on biogeochemical processes in terms of soil water availability. In this data package, the sample identification follows the format ER-X-Y, where ER refers to East River, X is the location identifier, and Y is the depth identifier at the same X (shallow Y=1). Specifically, ER-PHS, ER-LMC, ER-LMF, and ER-SMN are associated with ecohydrology sites under the East-Taylor Watershed Community Observatory Sites directory, and ER-RBTn (upslope n=1) are sampling transects during the 2019 Rootball Campaign. The sample and location information can be found in metadata.csv. Sampling and Measurements Each sample falls into one of the three sampling methods – (1) intact cores, (2) repacked samples, or (3) soil sensors – and one of the two measurement methods – (a) laboratory or (b) in-situ. Both intact cores and repacked samples were measured using the laboratory methods, which include measurements of soil water potential (HYPROP & WP4C, METER), saturated (KSAT, METER) and unsaturated hydraulic conductivity (HYPROP). The in-situ method uses a pair of co-located soil sensors to measure volumetric water content (TEROS12, METER) and soil water potential (TEROS21, METER), and the hydraulic conductivity was not measured. In comparison, the laboratory methods progress from full saturation to dry conditions, and the in-situ method includes both dry-to-wet and wet-to-dry cycles. The sampling and measurement methods for each sample can be found in metadata.csv, and more information about the measurements is detailed in the Methods section below. Models Retention and hydraulic conductivity data were fitted with four van-Genuchten-type models (specified by “model_name” column in the files): (1) traditional constrained van Genuchten model (“vG_constrained”), (2) traditional unconstrained van Genuchten model (“vG_unconstrained”), (3) PDI-variant of the constrained van Genuchten model (“vG_constrained_PDI”), and (4) PDI-variant of the unconstrained van Genuchten model (“vG_unconstrained_PDI”). The difference between the constrained (1: n) and the unconstrained (2: n, m) van Genuchten models is the number of pore-size distribution parameters in the model equations, giving the unconstrained model more degrees of freedom when fitting the data. Between the traditional and the PDI-variant models, model fitting differs the most at the dry end of the measurements. The traditional models allow infinite suction at the residual water content (water content does not drop below residual water content), and the PDI-variant models enforce a soil water potential value of pF=6.8 (~ -630 MPa) at oven-dryness (water content reaches 0). The inclusion of the van-Genuchten-type models is due to their common application. If other retention models are required, users can access the data in data.csv for further data fitting. More information about the models can be found in the Methods section below. Fitting Tasks The model fitting can be categorized into three levels of tasks (specified by “fitting_task” column in the files). Level 1 (“fit_retention”) only includes retention data fitting (the only level available for the in-situ method). Level 2 (“fit_retention_conductivity”) includes both retention and hydraulic conductivity data fitting, and the saturated hydraulic conductivity (Ks, a parameter of the hydraulic conductivity functions) is fixed by the measurements from KSAT. Level 3 (“fit_retention_conductivity_Ks”) also includes both retention and hydraulic conductivity data fitting, but Ks is a fitted parameter without the constraints from KSAT measurements. Among the same retention models (e.g. vG_constrained models of the same sample), level 1 should produce the best retention data fitting. Level 2 should have the highest misfit of the retention and hydraulic conductivity data, because the retention and hydraulic conductivity functions share common model parameters, and the unsaturated hydraulic conductivity (HYPROP) data fitting is subject to Ks measured independently by KSAT. Level 3 should have mid-level misfits of the retention and hydraulic conductivity data. While level 3 fits the hydraulic conductivity data better than level 2, the fitted Ks value might be unreasonable due to the lack of constraints at the wet end of the measurements. General recommendation when using this data package: (1) Choice of sampling methods: Intact cores and in-situ soil sensors could be prioritized because these sampling methods are less destructive. While the repacked samples were packed to the target bulk density (estimated post-sampling, when sample volume was known), these samples had altered pore structures. Nevertheless, intact cores might suffer from sample gaps that would lead to overestimation of Ks (sample gaps can be inferred from the “soil_sample_volume” column in metadata.csv when the value is < 249). In-situ method also has higher uncertainty in characterizing the wet end of the SWRC because of sensor limitations and the difficulty in reaching full saturation under natural conditions. (2) Choice of fitting tasks: When only retention data is needed, level 1 (“fit_retention”) should be prioritized. When both retention and hydraulic conductivity data are needed, level 2 (“fit_retention_conductivity”) could be prioritized. (3) Choice of models: This could depend on what the downstream models call for. If no specific model is required, model misfit could be used as a ranking criterion. Model misfit values in terms of RMSE can be found in model_parameters.csv. The following files are included in this data package: (1) metadata.csv – This file includes the general information of each sample, including location (description, geocoordinates, elevation), sampling and measurements details (method, depth, time or period, volume, instruments), and soil physical properties (bulk density, saturated hydraulic conductivity, only applicable to physical soil samples). (2) data.csv – This file includes soil water potential, volumetric water content, and unsaturated hydraulic conductivity data of each sample. Column “instrument” specifies the instrument (HYPROP, WP4C, or TEROS) used to perform the measurements. (3) model_fit.csv – This file includes soil water potential, volumetric water content, and unsaturated hydraulic conductivity fitted from the four models and three fitting tasks. Column “model_name” specifies the retention model used, and “fitting_task” specifies the level of data fitting. Missing values indicate that the variable does not apply to that fitting task. (4) model_parameters.csv – This file includes the fitted model parameters, model misfits, and conventional water content thresholds (field capacity and wilting point) from the four models and three fitting tasks. Column “model_name” specifies the retention model used, and “fitting_task” specifies the level of data fitting. Missing values indicate that the parameter does not apply to that model and/or that fitting task. (5) data_Ks.csv – This file includes the saturated hydraulic conductivity measurements from KSAT. (6) /figure/*.png – This folder includes three quick visualizations of the data, retention model fitting results and misfits, and hydraulic conductivity model fitting results, misfits, and parameters. The model fitting results are separated by samples and fitting tasks and colored by models. Zoom-in required. (7) /hyprop/*.bdhx – This folder includes proprietary hyprop files that require the free Labros SoilView-Analysis (METER) to open. Users can explore data fitting using other retention models (i.e. Brooks-Corey, Fredlund-Xing, Kosugi, bimodal models). Be aware that Ks value is pre-entered under “Fitting tab, Conductivity functions parameters” for level 2 fitting. If the value is lost, please refer to metadata.csv under “Ks” column. (8) Six file-level metadata that summarize file, header, column, and variable information of all files. This work was supported by the Watershed Function Science Focus Area at Lawrence Berkeley National Laboratory funded by the US Department of Energy, Office of Science, Biological and Environmental Research under Contract No. DE-AC02-05CH11231.

EARTH SCIENCE > LAND SURFACE > SOILS↗