Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “performance objectives”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

FY24 Development of Improved Grout Waste Forms for Alternative Low Activity Waste Treatment

Since the WTP Low Activity Waste (LAW) Vitrification Facility was not designed to process the entire inventory of Hanford LAW, up to half of the retrieved Hanford LAW will require supplemental immobilization. Immobilizing LAW in a cementitious waste form known as Cast Stone has been investigated as a possible candidate supplemental immobilization technology. In FY21, Washington River Protection Solutions, LLC (WRPS) tasked Atkins and the Vitreous State Laboratory (VSL) of The Catholic University of America (CUA) to perform testing to evaluate methods for reducing the release of COCs, particularly nitrate, 99Tc, and 129I, from cementitious waste forms made from aqueous LAW derived from Hanford Tank Waste. FY22 work built on the FY21 results and further developed formulations while targeting higher waste loadings. The objective of this work was to perform laboratory-scale testing to further refine the most promising formulation(s) that were identified in the FY23 work. The goal of the refinement was to further reduce the release rates for 99Tc, Cr, 129I, and nitrate while maintaining workability of the fresh grout, and to increase waste loading.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2025

LAMP Emittance Budget, Rev. 1

This report summarizes the performance limits of the LANSCE Coupled-Cavity Linac (CCL). These results are captured or summarized directly from the references cited. This report was written in support of the LANSCE Modernization Project (LAMP). This brief report summarizes the emittance budget for the LANSCE Modernization Project (LAMP). While the project Key Performance Parameters (KPPs) specify threshold and objective requirements for charge delivered to each experimental area, no upper limits on beam emittances are specified. To maintain low losses in the high-energy section of the LANSCE linac and hands-on maintenance, some upper limits on beam emittance need to be specified for the new LAMP front-end performance. The scope of the LAMP project replaces the injector section and drift-tube linac (DTL) up to 100 MeV of the existing LANSCE linac. This new design replacement will be integrated with the remaining coupled cavity linac (CCL) which makes up most of the accelerator at LANSCE and accelerates the beam to a final energy of 800 MeV. The initial approach that has been used to define an emittance budget for the new LAMP front end is based on recent and historical measured beam emittances at 100 MeV for the three beam types accelerated at LANSCE: H+ (protons for isotope production), LBEG (H- beam for delivery to proton radiography and to the Lujan neutron spallation target, and MPEG (H- beam for delivery to the Weapons Neutron Research facility). The emittance budget (upper limit) for each beam type has been selected to maintain the losses in the CCL to a level approximately equivalent to those observed in present operations to first order. However, the goal of the LAMP project is to improve the quality of the beams injected into the CCL, if possible, thus allowing for higher average current operation while also lowering beam losses and activation at high beam energies. The table below summarizes the beam measurements evaluated and used to establish a conservative emittance budget for LAMP based on known historical beam losses and activation. However, based on estimates of the CCL admittance and the Isotope Production Facility (IPF) beamline acceptance, a more relaxed transverse emittance upper limit of 0.095 π-cm-mrad, rms, normalized may be acceptable at 100 MeV while still meeting the LAMP performance requirements for charge delivery to each LANSCE experimental area and maintaining hands-on maintenance. This upper limit is supported by a recent analysis of operational data. Additionally, the present conceptual LAMP front-end design meets this requirement.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

Advanced Fuels Campaign Execution Plan

The Advanced Fuels Campaign (AFC) Execution Plan outlines the strategy, mission, scope, near-term and long-term goals, structure, and organization associated with nuclear fuels and materials research, development, and demonstration activities within the Department of Energy’s (DOE) Nuclear Fuel Cycle and Supply Chain (NFCSC) program. NFCSC has been given responsibility to identify and mature advanced fuel technologies for the DOE using a science-based approach, focused on developing a fundamental understanding of nuclear fuels and materials to drive development of integrated nuclear fuel and materials technology. This science-based approach combines theory, experiments, and multiscale modeling and simulation to achieve a predictive understanding of relevant behaviors ranging from fuel fabrication processes (and their resulting fuel microstructures) through fuel/cladding performance under irradiation (in contrast to more empirical, observation-based approaches frequently used in fuel performance modeling and fuel qualification). The traditional scope of AFC includes the evaluation and development of multiple fuel forms to support two fuel cycle options: once-through and full recycle. The word “fuel” is used generically to include conventional fuels, transmutation targets, and any associated cladding or duct materials. The once-through fuel cycle addresses advanced light water reactor fuels with enhanced performance, extended burnup, and reduced waste generation. In fiscal year (FY) 2012, AFC’s scope expanded to include research, development, and demonstration (RD&D) for light water reactor (LWR) fuels with enhanced accident tolerance. Fuel fabrication activities include the development of innovative methods to enhance process efficiencies, reduce waste, and improve control over as-fabricated fuel microstructural properties to achieve desired in-reactor performance. Using modern modeling and simulation approaches, the objective is to predict fresh fuel properties given the feedstock characteristics and fabrication process parameters. The performance-related activities include small-scale, in-reactor, and out-of-reactor phenomenological testing (distinct from, but synergistic with, integral prototypic testing) and extensive, quantitative characterization (focusing on characterization of fuel and cladding materials at the scale of microstructure) both before and after testing. Larger-scale, prototypic experiments are conducted in concert with phenomenological testing to drive a Fuel Development and Qualification program, incorporating a fundamental understanding of fuel behavior performance characteristics. Then, using the tools developed under the productive science-based approach, fuels will be optimized to meet specific performance requirements, thereby minimizing the need to repeatedly perform large-scale, integral experiments over a wide parametric range as a means of experimental exploration. Two significant initiatives are underway within AFC. First, a gap analysis completed in early FY 2019 identified critical irradiation testing needs that are lacking within the national light water reactor (LWR) fuels testbed since the shutdown of the Halden Reactor in 2018. The identified gaps are for instrumented, prototypic testing of LWR fuels, especially under boiling water reactor conditions, ramp conditions, and conditions leading to fuel failure; these needs exist for supporting current LWR fuels and their possible extension to higher burnups, but are especially urgent relative to near-term development and qualification of accident-tolerant fuels. Recommendations that resulted from the Halden Gap Analysis focused on enhancements at Advanced Test Reactor (ATR) and Transient Reactor Test Facility (TREAT) to fill gaps in testing capabilities relative to these needs. Second, a concerted effort to develop and demonstrate a systematic approach to accelerating the development, testing, and qualification of new fuel systems has been initiated. This is highlighted by a test strategy that combines the considerable advances in multiscale, mechanistic fuel modeling of recent years with a MiniFuel separate effects test program in the High Flux Isotope Reactor (HFIR) and a Fission Accelerated Steady-state Testing (FAST) semi-integral accelerated test program in ATR. This approach is being tested/demonstrated using the metallic fuel system, but if successful it is expected to be extensible to multiple fuel types and diverse applications. This document includes an overview of the NFCSC program, a definition of science-based development of nuclear fuels, near-term goals for Advanced LWR fuels (ALFs), and longer-term goals for Advanced Reactor Fuels (ARFs) RD&D. This includes the activities that will be conducted to achieve success toward the grand challenge, as well as the goals and milestones to be achieved over the next few decades of research and development. Long-term goals are based on the DOE Office of Nuclear Energ

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

High Temperature High Vacuum Mechanical Property Assessment of Zirconium Nuclear Fuel Cladding

This report presents the mechanical characterization of a specific Zry-4 cladding batch serving as the foundation for a diverse range of fuel performance research at Oak Ridge National Laboratory (ORNL). This effort supports research needs for the U.S. Department of Energy (DOE), particularly regarding evaluating accident tolerant fuel (ATF) cladding coating concepts, expanding understanding of cladding response to loss-of-coolant accidents (LOCA) transients, refining post-critical heat flux (CHF) limits (t@T), and upcoming irradiation campaigns. The central objective was to define the baseline performance of the substrate Zircaloy-4 (Zry-4) material leveraged across ORNL Advanced Fuel Campaign (AFC) efforts through controlled high-temperature vacuum tensile testing. This work begins to address gaps in existing models where implementation based on nominal heat-treatment labels, such as stress relief annealed (SRA), often fail to capture the interplay of recovery, recrystallization, and grain growth. To quantify this, data was benchmarked against the Pacific Northwest National Laboratory (PNNL) stress strain model to determine where this material falls in comparison to assumed values for materials in the same heat treatment regime. Analysis of the tensile data revealed that this specific SRA batch exhibits a transitional microstructural state best described by an effective cold-work (CW) parameter of 0.09, diverging from the previous estimation of 0.5 for SRA materials. Additionally, comparative testing of Cr coated specimens demonstrated no distinct difference in axial strength relative to the bare substrate. This suggests that the strengthening benefits of Cr coatings observed in burst scenarios are driven by residual stress mechanisms acting solely in the hoop direction.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

Triboelectric Nanogenerator Repeatability and Reproducibility Study

The development of triboelectric nanogenerators (TENGs) has largely focused on optimizing output performance, often at the expense of other critical research considerations such as the development of reliable technical procedures. In particular, the reliability of reported results—specifically repeatability and reproducibility—remains underexplored and is frequently limited to brief discussion within available literature. Without rigorous validation through repeatability and reproducibility studies, the credibility and broader applicability of reported findings remain uncertain. This study addresses this gap by systematically evaluating the repeatability and reproducibility of TENG performance data. Five polymer materials—Kapton, polyethylene (PE), polyethylene terephthalate (PET), polytetrafluoroethylene (PTFE), and polyvinylidene fluoride (PVDF)—were investigated across all pairwise combinations of 25 total combinations for the reproducibility study and three selected pairs of the 25 samples were selected for the repeatability study. For each TENG pairing, we analyzed the methodology, experimental procedures, and resulting performance data to quantify consistency and reliability. The objective of this work is to assess the validity of the collected dataset and determine whether the observed performance trends are consistent for use in future TENG design and optimization studies. Establishing reliable and reproducible data is essential for advancing the development of high-output TENG systems and ensuring their dependable implementation in practical applications.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Optimization of simulated high-field side lower hybrid current drive coupling using machine learning predictions of scrape-off layer density

Lower hybrid current drive (LHCD) is a potential source of non-inductive off-axis current drive (CD) for tokamaks. Although LHCD has been successfully deployed on a number of tokamaks, it is highly sensitive to the scrape-off layer (SOL) conditions local to the LHCD launcher. Large gaps between the launcher and plasma core, SOL turbulence, or edge density perturbations due to edge-localized modes can hamper CD or cause large reflected power. These coupling issues in part motivated the installation of an LHCD launcher on the high-field side (HFS) of DIII-D. On the HFS, the SOL is less turbulent and more controllable compared to the low-field side. This quiescence may result in more predictable edge conditions and thus a more predictable CD. Here, in this work, HFS SOL reflectometry measurements are predicted from global plasma parameters using machine learning models. The SOL predictions coupled with the full-wave simulation of the LHCD launcher allow for the prediction of reflected power, directivity, and arcing risk before the discharge. Launcher performance is then optimized using multi-objective Bayesian optimization, finding the shot parameters that result in an optimal SOL density that maximizes CD while minimizing the risk of arcing. The predictions and optimizations of LHCD performance are then accelerated using a surrogate model of the full-wave LHCD simulation.

Bayesian optimization↗
DOE OSTI2025

Implementation of Genetic Algorithms to Optimize Metal–Organic Frameworks for CO 2 Capture

Metal-organic frameworks (MOFs) are promising materials for CO 2 capture with the potential to use less energy than current industrial CO 2 capture methods. MOFs are highly versatile sorbents, and there is an almost unlimited number of MOFs that could be synthesized. In this work, we used a genetic algorithm (GA) and grand canonical Monte Carlo (GCMC) simulations to efficiently search for high-performing MOFs for CO 2 capture. We analyzed the effects of important GA parameters, including the mutation probability, the number of MOFs per generation and the number of GA generations, on the GA performance. Here, we performed GCMC simulations on-the-fly during the GA procedure to determine the performance of proposed MOFs and optimized their structures using multiple objective functions across different topologies. The GA was able to determine top-performing MOFs balancing CO 2 selectivity versus working capacity and reduced the cost of molecular simulations by a factor of 25 versus brute-force screening of an entire database of structures.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Quantum Imaging with X-rays

Quantum imaging encompasses a broad range of methods that exploit the quantum properties of light to capture information about an object. One such approach involves using a two-photon quantum state, where only one photon interacts with the object being imaged while its entangled partner carries spatial or temporal information. To implement this technique, it is necessary to generate specific quantum states of light and detect photons at the single-photon level. While this method has been successfully demonstrated in the visible electromagnetic spectrum, extending it to X-rays has faced significant challenges due to the difficulties in producing a sufficient rate of X-ray photon pairs and detecting them with adequate resolution. Here, we demonstrate record high rates of correlated X-ray photon pairs produced via a spontaneous parametric down-conversion process and we employ these photons to perform quantum correlation imaging of several objects, including a biological sample (E. cardamomum seedpod). Notably, we report an unprecedented detection rate of about 6,300 pairs per hour and the observation of energy anti-correlation for the X-ray photon pairs. We also present a detailed analysis of the properties of the down-converted X-ray photons, as well as a comprehensive study of the correlation imaging formation, including a study of distortions and corrections. These results mark a substantial advancement in X-ray quantum imaging, expanding the possibilities of X-ray quantum optical technologies, and illustrating the pathway towards enhancing biological imaging with reduced radiation doses.

Gofron, Kaz [ORNL] (ORCID:0000000314415736)↗
DOE OSTI2025

Sparse-Data Deep Learning Strategies for Radiographic Non-Destructive Testing

Radiography is an imaging technique used in a variety of applications, such as medical diagnosis, airport security, and nondestructive testing. We present a deep learning system for extracting information from radiographic images. We perform various prediction tasks using our system, including material classification and regression on the dimensions of a given object that is being radiographed. Our system is designed to address the sparse-data issue for radiographic nondestructive testing applications. It uses a radiographic simulation tool for synthetic data augmentation, and it uses transfer learning with a pre-trained convolutional neural network model. Using this system, our preliminary results indicate that the object geometry regression task saw an improvement of 70% in the R-squared value when using a multi-regime model. In addition, we increase the performance of the object material classification tasks by utilizing data from different imaging systems. In particular, using neutron imaging improved the material classification accuracy by 20% when compared to x-ray imaging.

convolutional neural networks↗
DOE OSTI2025

A mathematical approach to using the forgetting curve to evaluate experience and training factors in human reliability analysis

Traditional human reliability analysis (HRA) methods have difficulty dealing with the dynamic nature of factors such as time and rely on static and expert-judgment-based assessments of performance-shaping factors (PSFs) across limited levels. In this study, we introduce a mathematical approach for dynamically evaluating the experience and training PSF. Our proposed method integrates the psychological concept of the “forgetting curve” to evaluate how PSFs are impacted by the number of trainings and the time elapsed since training. To confirm the validity of the model, we provide experimental data fitted by identifying the quantitative relationship between training and human performance. This research enables dynamic and objective assessments, thus reducing reliance on subjective expert judgment and improving the accuracy of HRA.

99 - GENERAL AND MISCELLANEOUS↗
DOE OSTI2025

A new metrics framework for quantifying and intercomparing atmospheric rivers in observations, reanalyses, and climate models

We present a new atmospheric river (AR) analysis and benchmarking tool, namely Atmospheric River Metrics Package (ARMP). It includes a suite of new AR metrics that are designed for quick analysis of AR characteristics via statistics in gridded climate datasets such as model output and reanalysis. This package can be used for climate model evaluation in comparison with reanalysis and observational products. Integrated metrics such as mean bias and spatial pattern correlation are efficient for diagnosing systematic AR biases in climate models. For example, the package identifies the fact that, in CMIP5 and CMIP6 (Coupled Model Intercomparison Project Phases 5 and 6) models, AR tracks in the South Atlantic are positioned farther poleward compared to ERA5 reanalysis, while in the South Pacific, tracks are generally biased towards the Equator. For the landfalling AR peak season, we find that most climate models simulate a completely opposite seasonal cycle over western Africa. This tool can also be used for identifying and characterizing structural differences among different AR detectors (ARDTs). For example, ARs detected with the Mundhenk algorithm exhibit systematically larger size, width, and length compared to the TempestExtremes (TE) method. The AR metrics developed from this work can be routinely applied for model benchmarking and during the development cycle to trace performance evolution across model versions or generations and set objective targets for the improvement of models. They can also be used by operational centers to perform near-real-time climate and extreme event impact assessments as part of their forecast cycle.

58 GEOSCIENCES↗
DOE OSTI2025

BULKI-Store v0.3.2

BULKI-Store is a distributed object storage system optimized for high-performance computing environments. Built with a Rust core and Python bindings, it efficiently manages scientific and machine learning datasets across HPC clusters. The system employs a client-server architecture with MPI integration, enabling seamless scaling on supercomputers like Perlmutter. BULKI-Store's object-oriented approach provides intuitive data organization with rich metadata support, contrasting with traditional file-based solutions. Key optimizations include selective checkpoint loading, unified checkpoint files, and object chunking for large data transfers. For machine learning workloads, BULKI-Store offers advantages through fine-grained access patterns, dynamic data sharing between training instances, and reduced memory pressure. Memory management features include strategic Python GC calls, minimized data copies, and batch processing capabilities. The system leverages Rayon's thread pool for asynchronous data prefetching and supports multiple CPU architectures (ARM64, x86, AMD, RISC-V). By combining performance optimizations with developer-friendly APIs, BULKI-Store addresses the complex data management challenges of modern HPC applications while maintaining compatibility across heterogeneous computing environments.

Zhang, Wei [Lawrence Berkeley National Laboratory ↗
DOE OSTI2026

Spaced out: An economic framework to explore the impacts of PV panel spacing on large-scale farming in Colorado

CONTEXT Agrivoltaic systems co-locate solar technologies with agricultural operations on an integrated plot of land and potentially provide benefits to both energy and agricultural systems. To date, large-scale (>5-MW) agrivoltaic projects in the United States have been limited to grazing and ecovoltaic applications, raising questions about the impact and scalability of agrivoltaic crop systems. Many agrivoltaic designs raise the height of the solar panels to accommodate agricultural practices while keeping energy density high. However, raising the panels results in increased photovoltaic (PV) development costs, which often are higher than the economic returns of crop production underneath the panels. This leads to unfavorable project economics and the need for other agrivoltaic solutions than raising panels. OBJECTIVE To explore other solutions, we perform an initial feasibility analysis for an agrivoltaic solution that can integrate with large-scale farming practices by increasing the row spacing in between panels. Increased PV row spacing is a low-cost approach for scaling agrivoltaics to accommodate crop production and this spacing can be tailored to required crop equipment for different regions. Increasing row spacing will reduce the power density (PV installed per acre), but in areas that are not land limited, these agrivoltaic designs could be economically feasible. Our analysis establishes a framework for a feasibility analysis for where and with what crops spaced out panel agrivoltaic solutions might be economical. METHODS Using a case study for large-scale agriculture crops in Colorado, we establish a framework for wide-row agrivoltaic economic feasibility analysis. We utilized the System Advisor Model to calculate technoeconomic metrics to compare different row spacing solutions and capture tradeoffs of these system designs. RESULTS AND CONCLUSIONS We find that, in some circumstances, wider row agrivoltaic solutions that allow for continued mechanized crop production can provide economic benefits over a traditional utility-scale PV system. For most crops examined in this analysis, roughly $\$$200/acre in agricultural profit justified spacing out the panels to at least 31.7 ft. to accommodate agrivoltaic configurations versus PV only configurations. Additionally, opportunities for increased agricultural revenue with agrivoltaic systems allow PV project economics to tolerate a larger range of CAPEX variability while remaining economically viable relative to the PV only configurations. SIGNIFICANCE This framework can be adapted for a wide variety of crops and regions and allows for examination of economically favorable sites for future agrivoltaic systems that utilize different configuration and expand opportunities for agrivoltaics.

14 SOLAR ENERGY↗
DOE OSTI2024

A B-spline based gradient-enhanced micropolar implicit material point method for large localized inelastic deformations

The quasi-brittle response of cohesive-frictional materials in numerical simulations is commonly represented by softening plasticity or continuum damage models, either individually or in combination. However, classical models, particularly when coupled with non-associated plasticity, often suffer from ill-posedness and a lack of objectivity in numerical simulations. Moreover, the performance of the finite element method significantly degrades in simulations involving finite strains when mesh distortion reaches excessive levels. This represents a challenge for modeling cohesive-frictional materials, given their tendency to experience strongly localized deformations, such as those occurring during shear band dominated failure. Hence, accurate modeling of the response of cohesive-frictional solids is a demanding task. To address these challenges, we present an extension of the material point method (MPM) for the unified gradient-enhanced micropolar continuum, aiming at the analysis of finite localized inelastic deformations in cohesive-frictional materials. The generalized gradient-enhanced micropolar continuum formulation is employed to tackle challenges related to localization and softening material behavior, while the MPM addresses issues arising from excessive deformations. The method utilizes a B-spline formulation for the rigid background mesh to mitigate the well-known cell crossing errors of the MPM. To demonstrate the performance of the method, 2D and 3D numerical studies on localized failure in sandstone in plane strain compression and triaxial extension tests are presented. A comparison with finite element results confirms the suitability of the formulation. Moreover, an efficient numerical implementation of the formulation is presented, and it is demonstrated that the additional MPM specific overhead is negligible.

B-spline↗
DOE OSTI2026

Online task-space motion control for positioner-coordinated multi-robot manufacturing systems

Incorporating multiple robotic manipulators into large-scale manufacturing systems enhances production efficiency and expands manufacturing capabilities beyond those of single-robot systems. Workpiece positioners in robotic manufacturing have demonstrated significant benefits for process optimization, but coordination strategies for multi-robot systems with shared positioners have received limited attention. This work presents a task-space coordinated trajectory-tracking control framework for multi-robot manufacturing systems, in which robots coordinate their motions within a shared, dynamic workpiece positioning frame. A workpiece positioner actively adjusts the pose of the manufactured component to enable greater operational concurrency and improve overall production efficiency. The proposed motion-coordination scheme employs a distributed and scalable architecture, supporting coordination across heterogeneous multi-robot systems. Two optimization methodologies are introduced to manage kinematic redundancies and maintain continuous, near-optimal operation throughout the manufacturing process. The first strategy exploits a task-space dimensionality reduction to achieve locally optimal configurations by leveraging symmetry-axis rotations of the tool. The second strategy utilizes the workpiece positioner to drive the coordinated robots toward stable and kinematically favorable configurations. For both optimization strategies, multiple objectives are defined to improve key performance metrics, including manipulability, configuration consistency, proximity to mechanical limits, and motion efficiency. Addressing a key limitation of existing coordination approaches, the framework is designed around online setpoint modification, allowing coordinated robots to respond effectively to in-situ process feedback. The proposed control framework is validated using the Robot Operating System (ROS) middleware on a combination of physical and simulated multi-robot system hardware.

Arbogast, Alex [ORNL] (ORCID:0000000154740723)↗
DOE OSTI2026

Transforming jet flavour tagging at ATLAS

Jet flavour tagging enables the identification of jets originating from heavy-flavour quarks in proton–proton collisions at the Large Hadron Collider, playing a critical role in its physics programmes. This paper presents GN2, a transformer-based flavour tagging algorithm deployed by the ATLAS Collaboration that represents a different methodology compared to previous approaches. Designed to classify jets based on the flavour of their constituent particles, GN2 processes low-level tracking information in an end-to-end architecture and incorporates physics-informed auxiliary training objectives to enhance both interpretability and performance. Its performance is validated in both simulation and collision data. The measured c-jet (light-jet) rejection in data is improved by a factor of 3.5 (1.8) for a 70% b-jet tagging efficiency, compared to the previous algorithm. GN2 provides substantial benefits for physics analyses involving heavy-flavour jets, such as measurements of Higgs boson pair production and the couplings of bottom and charm quarks to the Higgs boson, and demonstrates the impact of advanced machine learning methods in experimental particle physics.

Characterization and analytical techniques↗
DOE OSTI2025

Analyzing and Exploring Training Recipes for Large-Scale Transformer-Based Weather Prediction

Abstract The rapid rise of deep learning (DL) in numerical weather prediction (NWP) has led to a proliferation of models which forecast atmospheric variables with comparable or superior skill than traditional physics-based NWP. However, among these leading DL models, there is a wide variance in both the training settings and architecture used. Further, the lack of thorough ablation studies makes it hard to discern which components are most critical to success. In this work, we show that it is possible to attain high forecast skill even with relatively off-the-shelf architectures, simple training procedures, and moderate compute budgets. Specifically, we train a minimally modified Swin Transformer V2 (SwinV2) on ERA5 data and find that it attains superior skill in terms of mean-square errors of deterministic forecasts when compared against the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS). Almost all DL–NWP systems share a core set of hyperparameters and design decisions. To aid and expedite future DL–NWP research, we present an in-depth, systematic exploration of different loss functions, model sizes and depths, patch sizes, and multistep training objectives. We also examine the model performance with metrics beyond the typical accuracy (ACC) and RMSE and investigate how the performance scales with model size. Through our open-source code, scoring pipelines, and models, we share our findings on key aspects of the training pipeline. These ablations reduce the necessity for expensive hyperparameter tuning and lower the barrier to entry for future DL–NWP research. Significance Statement This study investigates the potential of using large-scale transformer-based models for weather prediction, showing that it is possible to achieve high forecast accuracy with simpler, off-the-shelf architectures. By training a minimally modified SwinV2 transformer on ERA5 data, we show that the model achieves competitive forecast skill in terms of mean-square error for key variables, outperforming the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS) at all lead times. Our findings suggest that effective training strategies, such as multistep fine-tuning and channel-weighted losses, significantly enhance the model’s performance. However, we also highlight that these improvements come with trade-offs in other areas, such as ensemble spread and high-frequency spatial detail. This work highlights the promise of deep learning in improving weather forecasts, which could lead to better preparedness and response to weather events, ultimately benefiting society by providing more reliable weather predictions.

Willard, Jared D. [Lawrence Berkeley National Labo↗
DOE OSTI2024

Testing- and Model- Based Optimization of Coal-fired Primary Heater Design for Indirect Supercritical CO 2 Power Cycles (Final Scientific and Technical Report)

The overall objective of this project was to perform the R&D necessary to mitigate the risk associated with the design of a primary heat exchanger for a solid-fired combustion system coupled with an indirect-fired closed-loop Brayton Cycle utilizing supercritical CO 2 . The key technological hurdle was the coupling of a solid-fuel firing system with the primary heater, which poses a singular challenge, which is the management of burner performance and operational conditions in a way to manage heat exchanger tube metal temperatures and temperature ramp rates in the absence of fluid phase change on the inside of the tubes. We designed and built the first ever pseudo power system employing a simple recuperated supercritical CO 2 closed-loop Brayton Cycle coupled to a solid-fuel fired system. Advanced coupled CFD and process modeling were used to design the primary heat exchanger (PHX), which consisted of both radiative and convective sections, to limit tube metal temperatures resulting from the heat release profile of the solid fuel flame near the radiative tubes. The heat exchanger was designed to produce finished CO 2 temperatures of 600 °C a pressure of 20.7 MPa and CO 2 flow of 5.5 kg/s. The constructed PHX was capable of 1.2 MWth heat uptake. During design of the PHX, the modeling showed that most variables influencing flame shape (burner stoichiometric ratio and register velocities and swirl) were not suitable to manage heat flux to the metal surfaces. This is because they substantially increased adiabatic flame temperature through the influence of localized stoichiometric ratio. Excess air and firing rate were the two most powerful variables that could be used to control tube surface temperatures. The coupled system was operated for a total of 407 hours, with the longest continuous run of 248 hours. For 62% of the operational time, the unit was unmanned and in automatic control. The fuels used for the testing included natural gas, two Utah Bituminous coals, woody biomass, and bagasse. During the testing we were able to verify the 1.2 MWth heat uptake and we operated at a finished CO 2 temperature of 607 °C and a pressure of 20.3 MPa simultaneously. The real-time corrosion rate of the Super 304H tube CO 2 surface in the region of the radiative section of the PHX were measured, at an approximate temperature of 550 °C. The two key variables related to corrosion rate are the pressure and flow rate of the CO 2 . A technoeconomic analysis was performed at a scale of 120 MWE. The updated analysis showed that the efficiency of an sCO 2 power producing plant will be related to the pressure drop of the PHX.

01 COAL, LIGNITE, AND PEAT↗