Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Learning with errors”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

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

At least 1,423 records · Page 79
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Physics-informed graph neural networks for predicting cetane number with systematic data quality analysis

Designing alternative fuels for advanced compression ignition engines necessitates a predictive model for cetane number (CN). In this study, the physics-informed graph neural networks are introduced for a reliable CN prediction by considering molecular features pertinent to the physical properties of molecules that affect CN. The reliability of measured data is another key factor to consider for improving the predictive model. Various experimental instruments for measuring CN exist, including standard and non-standard methods. In this regard, a systematic data quality analysis was carried out for the total 630 CNs collected from literature and new measurements in this study using Advanced Fuel Ignition Delay Analyzer (AFIDA). The results from this data curation process were reflected in the model by imposing lower sample weights on the data coming from less reliable measurement techniques. This approach effectively maximized the prediction accuracy while incorporating data from all available sources. Using the sample weights decreased the mean absolute error (MAE) up to 0.8 CN units. The accuracy was also improved by introducing the CN-related physical properties (the number of hydrogen bond donors and acceptors); the test set MAE is 5.74 and 7.01 for the model with and without such properties, respectively. Investigating molecular structural effects on CN was also carried out to gain chemical insights into factors used to design new fuel candidates. The dimensionality reduction analysis of feature vectors showed a clear clustering in terms of functional groups and CN and the structural effect derived from the model was consistent with the physicochemical insights. Finally, this physics-informed model and data curation would be helpful for accurate CN prediction and inform rational fuel design.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

The Global LAnd Surface Satellite (GLASS) evapotranspiration product Version 5.0: Algorithm development and preliminary validation

An accurate estimation of spatially and temporally continuous global terrestrial evapotranspiration (ET) is essential in the assessment of surface energy, water and carbon cycles. The Global LAnd Surface Satellite (GLASS) ET product Version 4.0 (v4.0) based on the Bayesian model averaging (BMA) method was generated to estimate global terrestrial ET. However, certain uncertainty for the GLASS ET product v4.0 limits its application. In this study, we introduced the deep neural networks (DNN) merging framework to improve terrestrial ET estimation for GLASS ET product Version 5.0 (v5.0) generation by integrating five satellite-derived ET products [Moderate Resolution Imaging Spectroradiometer (MODIS) ET product (MOD16), Shuttleworth–Wallace dual-source ET product (SW), Priestley–Taylor-based ET product (PT-JPL), modified satellite-based Priestley–Taylor ET product (MS-PT) and simple hybrid ET product (SIM)]. We compared the performance of DNN method against other merging methods, including GLASS ET algorithm v4.0 (BMA), the gradient boosting regression tree (GBRT) method and the random forest (RF) method, based on 195 global eddy covariance (EC) flux towers covering observations from 2000 through 2015. Validations indicated that the DNN had the highest accuracy among four merging methods across different land cover types, yielding the highest average determination coefficients (R 2 , 0.62), root-mean-squared-error (RMSE, 24.1 W/m 2 ) and Kling–Gupta efficiency (KGE, 0.77) with a of 99% confidence interval. Compared with GLASS ET algorithm v4.0, the DNN improved on the R 2 by approximately 7% (p < 0.01) and the KGE by 10%. Based on the DNN, we then generated 8-day GLASS ET product v5.0 globally with a 1 km spatial resolution from 2001 to 2015 driven by GLASS vegetation and surface net radiation (R n ) datasets and Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA2) datasets. Finally, this global terrestrial ET product provides a valuable dataset for monitoring regional and global water resources and environmental changes.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Accelerating Collective Communication in Data Parallel Training across Deep Learning Frameworks

This work develops new techniques within Horovod, a generic communication library supporting data parallel training across deep learning frameworks. In particular, we improve the Horovod control plane by implementing a new coordination scheme that takes advantage of the characteristics of the typical data parallel training paradigm, namely the repeated execution of collectives on the gradients of a fixed set of tensors. Using a caching strategy, we execute Horovod’s existing coordinator-worker logic only once during a typical training run, replacing it with a more efficient decentralized orchestration strategy using the cached data and a global intersection of a bitvector for the remaining training duration. Next, we introduce a feature for end users to explicitly group collective operations, enabling finer grained control over the communication buffer sizes. To evaluate our proposed strategies, we conduct experiments on a world-class supercomputer — Summit. We compare our proposals to Horovod’s original design and observe 2x performance improvement at a scale of 6000 GPUs; we also compare them against tf.distribute and torch.DDP and achieve 12% better and comparable performance, respectively, using up to 1536 GPUs; we compare our solution against BytePS in typical HPC settings and achieve about 20% better performance on a scale of 768 GPUs. Finally, we test our strategies on a scientific application (STEMDL) using up to 27,600 GPUs (the entire Summit) and show that we achieve a near-linear scaling of 0.93 with a sustained performance of 1.54 exaflops (with standard error +- 0.02) in FP16 precision.

Romero, Joshua↗
DOE OSTI2026

Crowdsourcing the Frontier: Advancing Hybrid Physics‐ML Climate Simulation via a $\$$50,000 Kaggle Competition

Subgrid machine-learning (machine learning [ML]) parameterizations have the potential to introduce a new generation of climate models that incorporate the effects of higher-resolution physics without incurring the prohibitive computational cost associated with more explicit physics-based simulations. However, important issues, ranging from online instability to inconsistent online performance, have limited their operational use for long-term climate projections. To more rapidly drive progress in solving these issues, domain scientists and ML researchers opened up the offline aspect of this problem to the broader ML and data science community with the release of ClimSim, a NeurIPS Data sets and Benchmarks publication, and an associated Kaggle competition. This paper reports on the downstream results of the Kaggle competition by coupling emulators inspired by the winning teams' architectures to an interactive climate model (including full cloud microphysics, a regime historically prone to online instability) and systematically evaluating their online performance. Our results demonstrate that online stability in the low-resolution real-geography setting is reproducible across multiple diverse architectures, which we consider a key milestone. All tested architectures exhibit strikingly similar offline and online biases, though their responses to architecture-agnostic design choices (e.g., expanding the list of input variables) can differ significantly. Multiple Kaggle-inspired architectures achieve state-of-the-art results on certain metrics such as zonal mean bias patterns and global Root Mean Squared Error, indicating that crowdsourcing the essence of the offline problem is one path to improving online performance in hybrid physics-AI climate simulation.

Environmental sciences↗
DOE OSTI2024

Parallax-corrected VISST-derived pixel-level products from satellite GOES-16

The NASA Langley group led by William Smith produced GOES-16 satellite cloud retrievals over an approximate 10 by 10 degree region over the CACTI field campaign location. These retrievals are described here: https://www.arm.gov/capabilities/vaps/visst and are available for download here . They use algorithms historically called VISST that are now referred to as SatCORPS. More information can be found in Trepte et al. (2019), Minnis et al. (2021), and Yost et al. (2021). If using this dataset, please cite these references, the CACTI VISST dataset DOI found at the download link above, and this dataset’s DOI. The CACTI VISST pixel-level retrievals are on a 2 km spatial grid and available every 15 minutes (every 10 minutes late in the campaign), producing 21,765 files for the entire field campaign between October 2018 and April 2019. They are not corrected for parallax error, which is an offset in the actual geographical location of a cloud above the surface due to the satellite viewing the cloud partly from the side off nadir. This dataset applies a correction for parallax using the location relative to the satellite and the retrieved cloud top height above the surface, which allows the dataset to be geo-located with surface-based observations. The parallax correction for each location depends on the longitude, latitude and cloud top height above ground level (AGL) for that longitude and latitude in the original VISST files. The cloud top height AGL requires first computing the surface elevation at each VISST grid point. Data from the Advanced Spaceborne Thermal Emission and Reflection (ASTER) Global Digital Elevation Map Version 3 at 30-m resolution is projected onto the VISST grid using conservative coarsening (conserving surface elevation) in the xESMF Python package. The surface elevation is then subtracted from the VISST-retrieved cloud top height above mean sea level. These cloud top heights AGL are then combined with longitude and latitude to estimate the latitude and longitude corrections. Due to variability in cloud top height, the parallax shifts produce an irregular grid of values since higher cloud tops are shifted further than lower cloud tops. A ball tree-based neighbor search with Haversine distance is performed using the Python-based scikit-learn library to find the nearest VISST grid point to each parallax correction-shifted point. The data value of the shifted point is then assigned to that VISST grid point. In this manner, the irregular geographical shifts to correct for parallax are projected back to the rectilinear VISST grid. Because relatively higher clouds should obscure lower clouds, the variable values for the highest cloud top are preferentially chosen if two or more values are assigned to a grid point. The parallax correction should be viewed as an improved but still imperfect estimation of the cloud top locations, largely because the cloud top height is an imperfect retrieval. Please see the attached README document for further information. Users are encouraged to contact the authors with any additional questions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

From Data to Discovery: AI's Transformative Role in Thin Film Research

The advancement of thin film technologies is pivotal for progress in numerous fields, including energy, electronics, and quantum computing. However, the traditional trial-and-error approach to materials discovery is inherently slow and inefficient. This presentation will showcase how artificial intelligence (AI) is transforming thin film research by enabling a data-driven paradigm shift. We will highlight our past successes in applying AI to understand radiation damage in thin film oxides, demonstrating how graph analytics can unravel complex material behavior. Additionally, we will provide insights into our current work at the National Renewable Energy Laboratory, where we are leading the charge in autonomous materials science. Backed by a $14M investment in our characterization facility, we are developing AI-guided workflows that seamlessly integrate experimentation and AI-guided decision-making. By harnessing the power of AI, we aim to accelerate the discovery and design of high-performance thin films, propelling innovation across a multitude of industries.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Artificial intelligence driven laser parameter search: Inverse design of photonic surfaces using greedy surrogate-based optimization

Photonic surfaces designed with specific optical characteristics are becoming increasingly crucial for novel energy harvesting and storage systems. The design of these surfaces can be achieved by texturing materials using lasers. The optimal adjustment of laser fabrication parameters to achieve target surface optical properties is an open challenge. Thus, we develop a surrogate-based optimization approach. Our framework employs the Random Forest algorithm to model the forward relationship between the laser fabrication parameters and the resulting optical characteristics. During the optimization process, we use a greedy, prediction-based exploration strategy that iteratively selects batches of laser parameters to be used in experimentation by minimizing the predicted discrepancy between the surrogate model’s outputs and the user-defined target optical characteristics. This strategy allows for efficient identification of optimal fabrication parameters without the need to model the error landscape directly. We demonstrate the efficiency and effectiveness of our approach on two synthetic benchmarks and two specific experimental applications of photonic surface inverse design targets. By calculating the average performance of our algorithm compared to other state of the art optimization methods, we show that our algorithm performs, on average, twice as well across all benchmarks. Additionally, a warm starting inverse design technique for changed target optical characteristics enhances the performance of the introduced approach.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

The Dark Energy Survey: Cosmology Results with ∼1500 New High-redshift Type Ia Supernovae Using the Full 5 yr Data Set

Abstract We present cosmological constraints from the sample of Type Ia supernovae (SNe Ia) discovered and measured during the full 5 yr of the Dark Energy Survey (DES) SN program. In contrast to most previous cosmological samples, in which SNe are classified based on their spectra, we classify the DES SNe using a machine learning algorithm applied to their light curves in four photometric bands. Spectroscopic redshifts are acquired from a dedicated follow-up survey of the host galaxies. After accounting for the likelihood of each SN being an SN Ia, we find 1635 DES SNe in the redshift range 0.10 < z < 1.13 that pass quality selection criteria sufficient to constrain cosmological parameters. This quintuples the number of high-quality z > 0.5 SNe compared to the previous leading compilation of Pantheon+ and results in the tightest cosmological constraints achieved by any SN data set to date. To derive cosmological constraints, we combine the DES SN data with a high-quality external low-redshift sample consisting of 194 SNe Ia spanning 0.025 < z < 0.10. Using SN data alone and including systematic uncertainties, we find Ω M = 0.352 ± 0.017 in flat ΛCDM. SN data alone now require acceleration ( q 0 < 0 in ΛCDM) with over 5 σ confidence. We find ( Ω M , w ) = ( 0.264 − 0.096 + 0.074 , − 0.80 − 0.16 + 0.14 ) in flat w CDM. For flat w 0 w a CDM, we find ( Ω M , w 0 , w a ) = ( 0.495 − 0.043 + 0.033 , − 0.36 − 0.30 + 0.36 , − 8.8 − 4.5 + 3.7 ) , consistent with a constant equation of state to within ∼2 σ . Including Planck cosmic microwave background, Sloan Digital Sky Survey baryon acoustic oscillation, and DES 3 × 2pt data gives (Ω M , w ) = (0.321 ± 0.007, −0.941 ± 0.026). In all cases, dark energy is consistent with a cosmological constant to within ∼2 σ . Systematic errors on cosmological parameters are subdominant compared to statistical errors; these results thus pave the way for future photometrically classified SN analyses.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Investigating lab-scaled offshore wind aerodynamic testing failure and developing solutions for early anomaly detections

As offshore wind systems become more complex, the risk of human error or equipment malfunction increases during experimental testing. This study investigates a lab-scale incident involving a 1 : 50 scale 5 MW wind turbine, where a generator failure led to rotor overspeed and a blade–tower strike. To improve early fault detection, we propose a data-driven method based on multivariate long short-term memory (LSTM) models. High-frequency measurements are projected onto principal components, and anomalies are identified using reconstruction error and its time derivative. Two models are trained on different healthy datasets and tested using single- and multi-principal component (1PC and MPC) variations. Results show that combining both error and error derivative improves detection accuracy. The 1PC model detects faults faster, has a higher recall rate, and achieves a 43 % improvement in anomaly detection accuracy, while the MPC model yields higher precision. This approach provides a simple and effective tool for early anomaly detection in lab-scale experiments, helping to reduce the risk of future failures during the testing of new technologies.

17 WIND ENERGY↗
DOE OSTI2026

SpaceNet 9—Cross-Sensor Alignment of Optical and SAR Imagery

Precise registration of high-resolution synthetic aperture radar (SAR) and optical imagery is necessary for realizing the full potential and benefits of multimodal image analysis. However, two significant challenges presently exist. First, there is a lack of annotated datasets and benchmarks available for high-resolution SAR–optical image registration. Second, an assessment of efficient and reliable image registration methods that can precisely align these modalities is lacking. Here, we present a holistic description of the SpaceNet 9 Challenge and its results. We present a description of the dataset and baseline algorithm along with the results of the challenge, including a description of the winning algorithms. We release the SpaceNet 9 dataset along with open-sourcing the winning algorithms and baseline. The objective of SpaceNet 9 was to compute a dense displacement map that indicates the shift needed to align pixels in an optical image to the pixels in a SAR image. The challenge launched in April 2025 and was active for approximately two months. The top five solutions reduced image alignment error from approximately 34 m to under 13 m for public and private test data, with the best results obtaining a registration error of only 8.5 and 6.7 m on the public testing and private testing dataset, respectively. Usage of pretrained image matching models, robust outlier rejection with RANSAC, and estimating local displacement were common among the top solutions. The results of this challenge provide insight into high-resolution SAR–optical image registration and offer opportunities for future benchmarking in this domain. The baseline algorithm, winning solutions, and datasets are available at https://spacenet.ai/sn9-challenge/.

benchmark datasets↗
DOE OSTI2026

HydraGNN_Predictive_GFM_2026 - Ensemble of predictive graph foundation models for atomistic materials modeling

This release contains data and parameters of HydraGNN-based graph foundation models trained as a result of the work published in the pre-print "Exascale Multi-Task Graph Foundation Models for Imbalanced, Multi-Fidelity Atomistic Data" by M. Lupo Pasini et al. (https://arxiv.org/abs/2604.15380). We jointly train on 16 open first-principles datasets (544+ million structures covering 85+ elements) using a multi-task architecture with per-dataset heads and a scalable ADIOS2/DDStore data pipeline. On Frontier, we execute six large-scale DeepHyper hyperparameter optimization campaigns in FP64 and promote the top-performing message-passing models to sustained 2,048-node training, yielding a PaiNN-based lead model. The version of HydraGNN used to generate the outputs provided in this release is HydraGNN v5.0 (https://github.com/ORNL/HydraGNN/releases/tag/v5.0) The list of datasets used for the training of the graph foundation model is the following: 1) Alexandria [1] 2) ANI1x [2] 3) MPTrj [3] 4) Open Catalyst 2020 (OC20) [4] 5) Open Catalyst 2022 (OC22) [5] 6) Open Catalyst 2025 (OC25) [6] 7) Open Direct ir Capture 2023 (ODAC23) [7] 8) Open Materials 2024 (OMat24) [8] 9) Open Molecules 2025 (OMol25) [9] 10) OMol25-neutral (subset of OMol25 that contains only molecules with zero total charge) 11) OMol25-non-neutral (subset of OMol25 that contains only molecules with non-zero total charge) 12) Open Polymers 2026 (OPoly2026) [10] 13) Nabla2DFT [11] 14) QCML [12] 15) QM7X [reference 13] 16) transition1x [14] Dataset references: [1] J. Schmidt et al., “A dataset of 175k stable and metastable materials calculated with the PBEsol and SCAN functionals,” Scientific Data, vol. 9, p. 64, 2022. [2] J. S. Smith et al., “The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for molecules,” Scientific Data, vol. 7, p. 134, 2020. [Online]. Available: https: //www.nature.com/articles/s41597-020-0473-z [3] A. Jain et al., “Commentary: The Materials Project: A materials genome approach to accelerating materials innovation,” APL Materials, vol. 1, no. 1, p. 011002, 07 2013. [Online]. Available: https://doi.org/10.1063/1.4812323 [4] L. Chanussot et al., “Open catalyst 2020 (oc20) dataset and community challenges,” ACS Catalysis, vol. 11, no. 10, pp. 6059–6072, 2021. [Online]. Available: https://doi.org/10.1021/acscatal.0c04525 [5] K. Tran et al., “Open catalyst 2022 (oc22) dataset and challenges for oxidation electrocatalysts,” ACS Catalysis, vol. 13, no. 5, pp. 3066–3084, 2023. [Online]. Available: https://doi.org/10.1021/acscatal.2c05426 [6] S. J. Sahoo et al., “The open catalyst 2025 (oc25) dataset and models for solid-liquid interfaces,” arXiv preprint arXiv:2509.17862, 2025. [Online]. Available: https://arxiv.org/abs/2509.17862 [7] A. Sriram et al., “The open DAC 2023 dataset and challenges for sorbent discovery in direct air capture,” ACS Central Science, vol. 10, no. 5, pp. 923–941, 2024. [8] L. Barroso-Luque et al., “Open materials 2024 (omat24) inorganic materials dataset and models,” 2024. [Online]. Available: https://arxiv.org/abs/2410.12771 [9] D. S. Levine et al., “The open molecules 2025 (OMol25) dataset, evaluations, and models,” 2025. [Online]. Available: https://arxiv.org/abs/2505.08762 [10] D. S. Levine et al., The open polymers 2026 (OPoly26) dataset and evaluations,” arXiv preprint arXiv:2512.23117, 2025. [Online]. Available: https://arxiv.org/abs/2512.23117 [11] K. Khrabrov et al., “Nabla2dft: A universal quantum chemistry dataset of drug-like molecules and a benchmark for neural network potentials,” in NeurIPS 2024 Datasets and Benchmarks Track, 2024. [Online]. Available: https://openreview.net/forum?id=ElUrNM9U8c [12] S. Ganscha et al., “The QCML dataset, quantum chemistry reference data from 33.5M DFT and 14.7B semi-empirical calculations,” Scientific Data, vol. 12, p. 406, 2025. [13] J. Hoja et al., “QM7-X, a comprehensive dataset of quantum-mechanical properties spanning the chemical space of small organic molecules,” Scientific Data, vol. 8, p. 43, 2021. [Online]. Available: https://www.nature.com/articles/s41597-021-00812-2 [14] M. Schreiner et al., “Transition1x - a dataset for building generalizable reactive machine learning potentials,” Scientific Data, vol. 9, p. 779, 2022. The folder "datasets_ADIOS2_format" contains the set of pre-processed datasets in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used for the development and training of GFMs in this work. The "datasets_ADIOS2_format" directory contains 2 sub-directories, one for the version "v1" of the datasets and one for the version "v2" of the datasets. The version "v1" of the datasets provides values of the total energy as they are extracted from the original data as it was released by the respective institutions. The version "v2" of the datasets provides values of the energy that have been realigned. The realignment was performed by training a linear regression model that predicts the total energy as a function of the chemical composition of the atomistic structure, and then subtract such prediction from the original value of the total energy. Both folders "v1" and "v2" contain 16 sub-directories, each corresponding to an ADIOS2-formatted dataset The folder "DeepHyper-results" contains the configurational files and model's parameters for all the 186 HPO trials that were successfully completed by the scalable hyperparameter optimization (HPO) runs on Frontier. The content of the folder "DeepHyper-results" I structured as follows: 1) task-list.txt: list of mpnn name, jobid, and deephyper task id 2) gfm_${MPNN}_${JOBID}_0.${TASKID}: run directory with checkpoint files 3) gfm_${MPNN}: deephyper summary directory (*.csv) for each specific MPNN type 4) deephyper-experiment-${JOBID}: output and error logs for each job The file "deephyper-sorted.csv" contains the details of each HydraGNN model built and tested by HPO, obtained by merging the (*.csv) filed from each HPO run executed. Out of all the HPO trials, we selected 10 to continue the training of the respective HydraGNN models. Due to limited computational budget available in the LRN070 allocation we could not complete the training till convergence for all these 10 selected models. The folder "models" contains multiple sub-folders, one per each HydraGNN model trained. Each model sub-folder contains the parameters of each HydraGNN model, with multiple checkpoint-restarts. The list of sub-folders are as follows: 1) multidataset_hpo-BEST1-fp64 2) multidataset_hpo-BEST2-fp64 3) multidataset_hpo-BEST3-fp64 4) multidataset_hpo-BEST4-fp64 5) multidataset_hpo-BEST5-fp64 6) multidataset_hpo-BEST6-fp64 7) multidataset_hpo-BEST7-fp64 8) multidataset_hpo-BEST8-fp64 9) multidataset_hpo-BEST9-fp64 10) multidataset_hpo-BEST10-fp64 Within each one of these folders, additional auxiliary log files are provided with descriptions about how the training proceeded. The lead PaiNN-model is contained inside "multidataset_hpo-BEST6-fp64". The file "mlp_branch_weights" contains the parameters of the multi-layer perceptron (MLP) used to reconcile the predictions of the 16 output decoding heads of the HydragNN architectures. The MLP takes in input the chemical composition of the atomistic structure and predicts averaging weights to linearly mix the predictions of each output decoding head toward consolidating them into a single one. The folder "1.1billion-structure-inference" contains 1.1 billion atomistic structures randomly generated. Each structures is associated with energy and forces predicted with the lead-PaiNN model combined with the MLP model for reconciliation of the multi-branch predictions generated by the 16 output decoding heads. The folder "1.1billion-structure-inference" contains 9,300 (*.tar.gz) subdirectories, one per Frontier compute node used to execute the inference at exascale. Once uncompressed, each (*.tar.gz) subdirectory contains an ADIOS2 (*.bp) file container, where each atomistic structure is stored as a PyTorch-Geometric Data object. The file "export_dataset_environment_variables.sh" contains the environment variables that need to be set before running the HydraGNN code to reproduce the results provided in this dataset release. The code that can be used to load the ADIOS2 files, load HydraGNN models, and run inference is available at: https://github.com/ORNL/HydraGNN/releases/tag/v5.0

36 MATERIALS SCIENCE↗
DOE OSTI2020

Blind Modeling Validation Exercises Using the Horizontal Dry Cask Simulator

The U.S. Department of Energy (DOE) established a need to understand the thermal-hydraulic properties of dry storage systems for commercial spent nuclear fuel (SNF) in response to a shift towards the storage of high-burnup (HBU) fuel (> 45 gigawatt days per metric ton of uranium, or GWd/MTU). This shift raises concerns regarding cladding integrity, which faces increased risk at the higher temperatures within spent fuel assemblies present within HBU fuel compared to low-burnup fuel (≤ 45 GWd/MTU). A dry cask simulator (DCS) was built at Sandia National Laboratories (SNL) in Albuquerque, New Mexico to produce validation-quality data that can be used to test the accuracy of the modeling used to predict cladding temperatures. These temperatures are critical to evaluating cladding integrity throughout the storage cycle of commercial spent nuclear fuel. A model validation exercise was previously carried out for the DCS in a vertical configuration. Lessons learned during the previous validation exercise have been applied to a new, blind study using a horizontal dry cask simulator (HDCS). Three modeling institutions – the Nuclear Regulatory Commission (NRC), Pacific Northwest National Laboratory (PNNL), and Empresa Nacional del Uranio, S.A., S.M.E. (ENUSA) – were granted access to the input parameters from the DCS Handbook, SAND2017-13058R, and results from a limited data set from the horizontal BWR dry cask simulator tests reported in the HDCS update report, SAND2019-11688R. With this information, each institution was tasked to calculate peak cladding temperatures and air mass flow rates for ten HDCS test cases. Axial as well as vertical and horizontal transverse temperature profiles were also calculated. These calculations were done using modeling codes (ANSYS/Fluent, STAR-CCM+, or COBRA-SFS), each with their own unique combination of modeling assumptions and boundary conditions. For this validation study, the ten test cases of the horizontal dry cask simulator were defined by three independent variables – fuel assembly decay heat (0.5 kW, 1 kW, 2.5 W, and 5 kW), internal backfill pressure (100 kPa and 800 kPa), and backfill gas (helium and air). The plots provided in Chapter 3 of this report show the axial, vertical, and horizontal temperature profiles obtained from the dry cask simulator experiments in the horizontal configuration and the corresponding models used to describe the thermal-hydraulic behavior of this system. The tables provided in Chapter 3 illustrate the closeness of fit of the model data to the experiment data through root mean square (RMS) calculations of the error in peak cladding temperatures (PCTs), PCT axial locations, axial temperature profiles, vertical and horizontal temperature profiles at two different axial locations, and air mass flow rates for the ten test cases, normalized by the experimental results. The model results are assigned arbitrary model numbers to retain anonymity. Due to the relatively flat axial temperature profiles, small temperature gradients resulted in large deviations of all models’ PCT axial location from the experimental PCT axial location. When the PCT axial location error is excluded in the calculation of the combined RMS of the normalized errors that considers PCT, the temperature profiles, and the air mass flow rates, the model data fits the experimental data to within 5%. When the vault information is excluded, the model data fits the experimental data to within 2.5%. An error analysis was developed further for one model, using the model and experimental uncertainties in each validation parameter to calculate validation uncertainties. The uncertainties for each parameter were used to define quantifiable validation criteria. For this analysis, the model was considered validated for a given comparison metric if the normalized error in that metric divided by the validation uncertainty was less than or equal to 1. When considering the combined RMS of the normalized errors of all metrics divided by their validation uncertainties, the model was found to have satisfied the criterion for model validation.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

Neural Posterior Estimation for Scalable and Accurate Inverse Parameter Inference in Li-Ion Batteries

Diagnosing the internal state of Li-ion batteries is critical for battery research, operation of real-world systems, and prognostic evaluation of remaining lifetime. By using physics-based models to perform probabilistic parameter estimation via Bayesian calibration, diagnostics can account for the uncertainty due to model fitness, data noise, and the observability of any given parameter. However, Bayesian calibration in Li-ion batteries using electrochemical data is computationally intensive even when using a fast surrogate in place of physics-based models, requiring many thousands of model evaluations. A fully amortized alternative is neural posterior estimation (NPE). NPE shifts the computational burden from the parameter estimation step to data generation and model training, reducing the parameter estimation time from minutes to milliseconds, enabling real-time applications. The present work shows that NPE can infer parameters equally or more accurately than Bayesian calibration, even if it leads to higher voltage reconstruction errors. We also demonstrate that the higher computational costs for data generation are tractable even in high-dimensional cases (ranging from 6 to 27 estimated parameters). The NPE method also offers several interpretability advantages over Bayesian calibration, such as local parameter sensitivity to specific regions of the voltage curve. The NPE method is demonstrated using an experimental fast charge dataset, with parameter estimates validated against measurements of loss of lithium inventory and loss of active material. The implementation is made available in a companion repository (https://github.com/NatLabRockies/BatFIT).

25 ENERGY STORAGE↗
DOE OSTI2026

Hourly PM 2.5 Estimates across California from 2018 to 2023

This study presents a new data set of hourly PM 2.5 concentrations across California from 2018 to 2023 at a three-kilometer resolution. This data set was developed by assimilating observations from PurpleAir and the U.S. EPA Air Quality System monitors into wildfire smoke forecasts from the High-Resolution Rapid Refresh Smoke (HRRR-Smoke) model using the Gridpoint Statistical Interpolation (GSI) three-dimensional variational data assimilation framework. Archived forecasts of modeled wildfire smoke PM 2.5 from HRRR-Smoke create the background field for assimilation, which is then corrected using surface observations of total PM 2.5 . The resulting reanalysis from GSI provides an estimate of total PM 2.5 that minimizes error from both the observational and the model data. Validation results indicate strong performance, with monthly R 2 values ranging from 0.73 to 0.91 across the six-year data set, comparable to other PM 2.5 data sets. Case studies are presented for three major fire events, the 2018 Camp Fire, 2019 Kincade Fire, and 2020 Lightning Complex Fires to demonstrate the data set’s fidelity in resolving plume dynamics and local exposure patterns. Root-mean-squared error averaged over each month scales with average PM 2.5 concentrations, resulting in a low error under typical conditions but higher absolute errors during extreme smoke events. This is the first long-term, hourly PM 2.5 data set of its kind for California and enables the generation of subdaily exposure metrics, such as peak hourly concentrations, exceedance durations, and time-of-day exposure peaks. The novelty and strong validation of this data set make it a compelling resource for future studies on the impact and significance of subdaily PM 2.5 exposure.

PM2.5↗
DOE OSTI2020

Effects of cluster expansion on the locations of phase transition boundary as a first step to quantify uncertainty in first principles statistical mechanics framework

Predicting phase diagrams from first principle calculations eliminates the need of tedious experimental trials and errors. Fully automating first principle phase diagram calculations without any sort of human intervention has been a long daunting task and troubling scientific communities for decades. This grand problem remains not fully resolved, largely due to the vastly high-dimensional parameter space associated with density functional theory, cluster expansion, lattice Monte Carlo, and the substantial uncertainty propagating through a set of complex simulations. As a first step to tackle this grand problem, we reported a first demonstration of how sensitive phase boundary locations can be to various cluster expansion fittings and input DFT training data. To the best knowledge of the authors, this study reported the first ever attempt to quantify uncertainty in first principles statistical mechanics framework. In addition, a semi-automated phase transition detection algorithm has been devised in this paper to deal with the associated statistical errors and uncertainties from Monte Carlo method and its predecessor DFT calculations and cluster expansions. This algorithm has been applied in a classical cluster expansion Mg-Cd binary alloy system to detect phase transitions at various locations of phase diagram using different cluster expansions to demonstrate its predictive power and quantify uncertainties. The results suggested that using Chebyshev basis function shifted the transition locations toward the dilute solid solution phase and expanded the phase stability of concentrated ordered phases to wider composition range. The addition of perturbed defect configurations into training data set lowered the order-disorder transition temperature to be away from true transition temperature, suggesting the transition nature is indeed configurational disorder dominated rather than defect assisted. Finally, we found that the weighting has negligible effect on the transition locations, except for the case of using Chebyshev basis function to fit all non-weighted configurations that can be susceptible to Monte Carlo sampling hysteresis.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Systematic feature design for cycle life prediction of lithium-ion batteries during formation

Optimization of the formation step in lithium-ion battery manufacturing is challenging due to limited physical understanding of solid-electrolyte interphase formation and the long testing time (∼100 days) for cells to reach the end of life. We propose a systematic feature-design framework that requires minimal domain knowledge for accurate cycle life prediction during formation. By only using two simple Q (V) features designed from our framework, extracted from formation data without any additional diagnostic cycles, we achieved an average of 9.87% error for cycle life prediction. Here, the physics-based investigation guided by the two designed features shows that the voltage ranges identified by our framework capture the effects of formation temperature and microscopic-particle resistance heterogeneity. By designing highly predictive, robust, and interpretable features, our approach can accelerate industrial battery formation research, leveraging the interplay between data-driven feature design and mechanistic understanding.

25 ENERGY STORAGE↗
DOE OSTI2025

SIDDA: SInkhorn Dynamic Domain Adaptation for image classification with equivariant neural networks

Modern neural networks (NNs) often do not generalize well in the presence of a ‘covariate shift’; that is, in situations where the training and test data distributions differ, but the conditional distribution of classification labels given the data remains unchanged. In such cases, NN generalization can be reduced to a problem of learning more robust, domain-invariant features. Domain adaptation (DA) methods include a broad range of techniques aimed at achieving this; however, these methods have struggled with the need for extensive hyperparameter tuning, which then incurs significant computational costs. In this work, we introduce SInkhorn Dynamic Domain Adaptation (SIDDA), an out-of-the-box DA training algorithm built upon the Sinkhorn divergence, that can achieve effective domain alignment with minimal hyperparameter tuning and computational overhead. We demonstrate the efficacy of our method on multiple simulated and real datasets of varying complexity, including simple shapes, handwritten digits, real astronomical observations, and remote sensing data. These datasets exhibit covariate shifts due to noise, blurring, differences between telescopes, and variations in imaging wavelengths. SIDDA is compatible with a variety of NN architectures, and it works particularly well in improving classification accuracy and model calibration when paired with symmetry-aware equivariant NNs (ENNs). We find that SIDDA consistently enhances the generalization capabilities of NNs, achieving up to a ${\approx}40\%$ improvement in classification accuracy on unlabeled target data, while also providing a more modest performance gain of $\lesssim 1\%$ on labeled source data. We also study the efficacy of DA on ENNs with respect to the varying group orders of the dihedral group DN, and find that the model performance improves as the degree of equivariance increases. Finally, if SIDDA achieves proper domain alignment, it also enhances model calibration on both source and target data, with the most significant gains in the unlabeled target domain—achieving over an order of magnitude improvement in the expected calibration error and Brier score. SIDDA’s versatility across various NN models and datasets, combined with its automated approach to domain alignment, has the potential to significantly advance multi-dataset studies by enabling the development of highly generalizable models.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Accurate numerical simulations of open quantum systems using spectral tensor trains

Decoherence between qubits is a major bottleneck in quantum computations. Decoherence results from intrinsic quantum and thermal fluctuations as well as noise in the external fields that perform the measurement and preparation processes. With prescribed colored noise spectra for intrinsic and extrinsic noise, we present a numerical method, Quantum Accelerated Stochastic Propagator Evaluation (Q-ASPEN), to solve the time-dependent noise-averaged reduced density matrix in the presence of intrinsic and extrinsic noise. Q-ASPEN is arbitrarily accurate and can be applied to provide estimates for the resources needed to error-correct quantum computations. We employ spectral tensor trains, which combine the advantages of tensor networks and pseudospectral methods, as a variational ansatz to the quantum relaxation problem and optimize the ansatz using methods typically used to train neural networks. Here, the spectral tensor trains in Q-ASPEN make accurate calculations with tens of quantum levels feasible. We present benchmarks for Q-ASPEN on the spin-boson model in the presence of intrinsic noise and on a quantum chain of up to 32 sites in the presence of extrinsic noise. In our benchmark, the memory cost of Q-ASPEN scales as a low-order polynomial in the size of the system once the number of system states surpasses the number of basis functions used in the spectral expansion.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗