Search NASA⌕ Search

SEARCH · Search NASA

Results for “Accuracy”

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

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

At least 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Leveraging Large Language Models for Real-World Data Evidence: A Framework for Automated Treatment Extraction and Data Harmonization

Background: The ability to comprehensively collect treatment information from cancer patient medical records would enable studies to evaluate real-world benefits and risks tied to specific treatments. Currently, it is difficult to system- atically collect high-quality treatment information because it is often stored in unstructured text. Manually extracting and standardizing drug and regimen data is time-intensive. Recent advances in large language models (LLMs) offer a potential solution for automated extraction of structured treatment information from clinical text. Objective: This study systematically evaluates the utility of four LLMs from the Llama family for automated extraction of oncology treatment information from clinical text. This information can guide researchers using cancer registry data to provide insights into cancer care and outcomes beyond clinical trials. Methods: Four instruction-tuned Llama models with varying parameter counts (1B, 3B, 8B, and 70B) were evaluated for their ability to extract treatment information from clinical documents. A unified oncology knowledge base integrating seven major public data sources was developed to standardize and normalize extracted entities—a critical step for harmonizing data from diverse sources. Extracted treatment data were compared against expert-annotated ground truth. Model performance was assessed using accuracy metrics (Precision, Recall, F1-Score) and opera- tional feasibility metrics, including processing speed and structural compliance of the output. Results: A strong positive correlation was observed between model size and extraction accuracy. F1-score improved from 0.609 for the 1B model to 0.710 (3B), 0.807 (8B), and 0.828 (70B). While larger models demonstrated superior accuracy and compliance, they incurred higher computational costs. The modest performance difference between 8B and 70B suggests diminishing returns with increasing model size. Conclusions: LLMs represent a viable technology for automating oncology treatment extraction. The 8B-parameter model emerged as a highly effective option, balancing high accuracy and computational efficiency. Selecting an appropriate LLM for deployment in cancer registries involves a trade-off between desired accuracy and available operational resources. Harmonizing extracted entities with the oncology knowledge base facilitates standardized integration into common data models, enhancing data quality for real-world evidence analyses.

artificial intelligence↗
DOE OSTI2025

Utilization of Data Augmentation Techniques in Automated Inspection Systems for Defect Detection in Metals With Limited Data

Accurate identification of defects on metal surfaces is of great interest to many industry sectors, such as the automotive and aerospace industries. In contrast to conventional manual inspection techniques, recent automated inspection systems employ deep learning models trained to detect defects rapidly and precisely. The development of these models often requires a substantial image dataset to acquire adequate knowledge of defect features and enhance their predictive accuracy. When data is limited, augmentation techniques are often used to improve the precision and accuracy of defect detection systems. This study examined the prediction performance of two object detection models, namely Faster Region‐based Convolutional Neural Network (Faster R‐CNN) and You Only Look Once version 8 (YOLOv8), to identify dent defects in limited images of cast iron cylinder head surfaces. The original image set contains 46 images with 563 dents. To overcome limited data availability, common image augmentation techniques along with a copy‐paste method were applied. Results show that standard augmentation improved YOLOv8 accuracy by 8.00% and average precision (AP) by 3.00%. On the other hand, the copy‐paste technique achieved a 20.00% increase in accuracy and a 1% increase in AP with just 200 synthetic dents. Furthermore, these results provide support for using the copy‐paste augmentation strategy to enhance defect detection performance, with a limited dataset, contributing to more accurate defect identification in remanufacturing processes.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Parallel-in-Time Solution of Allen-Cahn Equations by Integrating Operator Learning into the Parareal Method

While recent advances in deep learning have shown promising efficiency gains in solving time-dependent partial differential equations (PDEs), matching the accuracy of conventional numerical solvers still remains a challenge. One strategy to improve the accuracy of deep learning-based solutions for time-dependent PDEs is to use the learned model as the coarse propagator in the Parareal method and a traditional numerical method as the fine solver. However, successful integration of deep learning into the Parareal method requires consistency between the coarse and fine solvers, particularly for PDEs exhibiting rapid changes such as sharp transitions. Here, to ensure this consistency, we propose using convolutional neural networks (CNNs) to learn the fully discrete time-stepping operator defined by the same numerical scheme employed as the fine solver. We demonstrate the effectiveness of the proposed method in solving the classical and mass-conservative Allen–Cahn (AC) equations. Through iterative updates in the Parareal algorithm, our approach achieves a significant computational speedup compared to traditional fine solvers while converging to high-accuracy solutions. Our results highlight that the proposed hybrid Parareal algorithm effectively accelerates simulations, particularly when implemented on multiple GPUs, and converges to the desired accuracy in only a few iterations. Another advantage of our method is that the CNN model is trained on trajectory-based data generated from random initial conditions, such that the trained model can be used to solve the AC equations with various initial conditions without retraining. This work demonstrates the potential of integrating neural network methods into parallel-in-time frameworks for efficient and accurate simulations of time-dependent PDEs.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

An Internal Digital Image Correlation Technique for High-Strain Rate Dynamic Experiments

Full-field, quantitative visualization techniques, such as digital image correlation (DIC), have unlocked vast opportunities for experimental mechanics. However, DIC has traditionally been a surface measurement technique, and has not been extended to perform measurements on the interior of specimens for dynamic, full-scale laboratory experiments. This limitation restricts the scope of physics which can be investigated through DIC measurements, especially in the context of heterogeneous materials. The focus of this study is to develop a method for performing internal DIC measurements in dynamic experiments. The aim is to demonstrate its feasibility and accuracy across a range of stresses (up to 650 MPa), strain rates (10 3 - 10 6 s -1 ), and high-strain rate loading conditions (e.g., ramped and shock wave loading). Internal DIC is developed based on the concept of applying a speckle pattern at an inner-plane of a transparent specimen. The high-speed imaging configuration is coupled to the traditional dynamic experimental setups, and is focused on the internal speckle pattern. During the experiment, while the sample deforms dynamically, in-plane, two-dimensional deformations are measured via correlation of the internal speckle pattern. In this study, the viability and accuracy of the internal DIC technique is demonstrated for split-Hopkinson (Kolsky) pressure bar (SHPB) and plate impact experiments. The internal DIC experimental technique is successfully demonstrated in both the SHPB and plate impact experiments. In the SHPB setting, the accuracy of the technique is excellent throughout the deformation regime, with measurement noise of approximately 0.2% strain. In the case of plate impact experiments, the technique performs well, with error and measurement noise of 1% strain. The internal DIC technique has been developed and demonstrated to work well for full-scale dynamic high-strain rate and shock laboratory experiments, and the accuracy is quantified. Here, the technique can aid in investigating the physics and mechanics of the dynamic behavior of materials, including local deformation fields around dynamically loaded material heterogeneities.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Integration and validation of some modules for modelling of high-speed chemically reactive flows in two-phase gas-droplet mixtures

Three modules are integrated into the built-in OpenFOAM rhoCentralFoam solver towards accurate and efficient modelling of high-speed chemically reactive flows in two-phase gas-droplet mixtures within the OpenFOAM 10.0 framework. The first module is the mixture-averaged diffusion model. The second module is the built-in OpenFOAM Lagrangian solver coupled with optimised droplet drag coefficient and convective heat transfer coefficient sub-models. The last module is a sparse stiff chemistry solver based on dynamic adaptive hybrid integration (AHI-S). The optimised droplet sub-models are first verified in correct implementation for subsequent simulations in this work. Further, they show good accuracy against experimental and analytical data in the modelling of ammonia droplet acceleration and cooling in the flowing and/or low-temperature air. The accuracy and efficiency gains related to the mixture-averaged diffusion model and the AHI-S chemistry solver are examined by simulating 1-D detonation propagation in ammonia droplet-free/laden ammoniaoxygen mixtures. Numerical results of detonation propagation speed, gaseous temperature, density, and species distributions around the induction zone show good agreement with experimental data and analytical solutions. Compared to the built-in OpenFOAM diffusion model, the mixture-averaged diffusion model provides different numerical predictions of pulsating instabilities in detonation propagation. It shows better accuracy in depicting the detonation structure within the droplet-free section attributed to improved multi-component diffusion modelling. Compared to the built-in OpenFOAM solver EulerImplicit (backward Euler), the AHI-S chemistry solver reduces the computational cost by around 50%. It achieves satisfactory accuracy in calculating detonation propagation speed within the droplet-free section with the optimal efficiency when the safety factor, β, equals 0.5.

42 ENGINEERING↗
DOE OSTI2025

A cross-dimensional analysis of data-driven short-term load forecasting methods with large-scale smart meter data

Electricity load forecasting is essential to utility operation and power grid stability. A wide spectrum of data-driven methods, ranging from linear regression models to more recent deep learning models have been adopted to forecast electric load over the years. However, there still lacks a holistic evaluation of the applicability of conventional statistical and machine learning based algorithms with respect to different temporal and spatial scopes, computational requirements, and sensitivity of model-tuning. Enabled by a large-scale electricity load profile dataset of over 40,000 residential customers in a utility region, we conducted a cross-dimensional analysis of data-driven load forecasting methods. Three regression-based and seven deep learning algorithms with different model configurations were evaluated in terms of their overall and peak load prediction accuracy, and training burdens, across spatial aggregation levels ranging from the transformer, feeder, substation, to neighborhood. We found, first, the load forecasting accuracy is constrained by a predictability boundary, influenced by the forecasting horizon and spatial aggregation level. Specifically, RandomForest, XGBoost, TFT, TSMixer, and TiDE models achieved less than 10 % prediction error for up to 96-h ahead forecasting for district, substation, and feeder levels, while other models struggle at long-horizon predictions; Second, for winter and summer peak load dates, most models were able to predict the peak demand timing within ± 1 h, but the prediction percentage error varied by models, with TFT and TiDE models being the top performers; Third, models with similar prediction accuracy can differ in training burden by an order of magnitude. Therefore, choosing model configurations that balance prediction performance and computational resource is an important practical consideration for large-scale deployment of the machine learning based load forecasting. The outcome of this study can guide researchers and practitioners to choose the proper load forecasting algorithms based on their problem scope, required accuracy, and available resources. The predictability boundary can serve as a benchmark for electricity load forecasting problems with new algorithms and datasets.

Li, Han↗
DOE OSTI2025

Computationally efficient models for aqueous organic redox flow batteries

The rising usage of intermittent energy has garnered the need for large scale energy storage systems. Redox flow batteries (RFB) based energy storage system shows promising potential. Numerical simulations and machine learning approaches have been widely used to study RFB performance. The development of autonomous material discovery framework and digital twin of energy storage system usually needs to query cell performance through fast response models. In this study, two computationally efficient models are introduced: a physics-based analytical flow battery model (EZBattery), and a machine learning operator model (Deep Operator Network, denoted by DeepONet). Both models can provide cell performance near instantly, and prediction accuracy was systematically examined on an application of evaluating the performances of a 780 cm 2 aqueous organic redox flow battery (AORFB), using potential anolyte candidates in dihydroxyphenazine (DHP)-based family of organic materials. A validated computationally expansive 3-dimensional multi-physics finite element model by COMSOL was used as the ground truth and provided the training data set for the DeepONet. 1280 samples were generated with 10 properties to mimic the different possible anolyte candidates, and the cell performances were evaluated under 10 different combined operating conditions. The accuracy comparisons for the two computationally efficient models show that both models can provide comparable accuracy in predicting cell charging/discharging voltage curves. DeepONet can provide slightly higher overall accuracy than EZBattery with faster calculation speed, but highly relies on the training dataset. EZBattery does not need a training dataset and can provide interpretable physics-based explanations of the results, while being more flexible to adjust to adapt any different cell designs, flow battery architectures, and electrolyte materials.

Analytical model↗
DOE OSTI2025

Nonunitary Variational Quantum Eigensolver with the Localized Active Space Method and Cost Mitigation

Accurately describing strongly correlated systems with affordable quantum resources remains a central challenge for quantum chemistry applications on near and intermediate term quantum computers. The localized active space self-consistent field (LASSCF) approximates the complete active space self-consistent field (CASSCF) by generating active space-based wave functions within specific fragments while treating interfragment correlation with mean-field approach, hence is computationally less expensive. Hardware-efficient ansatzes (HEA) offer affordable and shallower circuits, yet they often fail to capture the necessary correlation. Previously, Jastrow-factor-inspired nonunitary qubit operators were proposed to use with HEA for variational quantum eigensolver (VQE) calculations (so-called nuVQE), as they do not increase circuit depths and recover correlation beyond the mean-field level for Hartree–Fock initial states. Here, in this study, we explore running nuVQE with LASSCF as the initial state. The method, named LAS-nuVQE, is shown to recover interfragment correlations, reach chemical accuracy with a small number of gates (<70) in both H 4 and square cyclobutadiene (C 4 H 4 ), and produces more accurate energetics than its HEA counterparts at all circuit depths. To further address the inherent symmetry-breaking in HEA, we implemented spin-constrained LAS-nuVQE to extend the capabilities of HEA further and show spin-pure results for square cyclobutadiene. We also mitigate the increased measurement overhead of nuVQE via Pauli grouping and shot-frugal sampling, reducing measurement costs by up to 2 orders of magnitude compared to ungrouped operator, and show that one can achieve better accuracy with a small number of shots (10 3–4 ) per one expectation value calculation compared to noiseless simulations with one or two orders of magnitude more shots. Finally, wall clock time estimates show that, with our measurement mitigation protocols, nuVQE becomes a cheaper and more accurate alternative than vanilla VQE with HEA. Taken together, these developments illustrate a practical pathway toward performing multireference chemical simulations with accuracy and affordable resources on today’s quantum hardware, achieving both accuracy and affordability in challenging correlated systems.

Wang, Qiaohong [Univ. of Chicago, IL (United State↗
DOE OSTI2025

Deep Learning for Subsurface Flow: A Comparative Study of U‐Net, Fourier Neural Operators, and Transformers in Underground Hydrogen Storage

Subsurface flow research is essential for the sustainable management of natural resources and the environment. Deep learning (DL) has significantly advanced this field by developing efficient and accurate surrogate models to replace computationally expensive physics‐based simulations. These surrogate models are commonly used to predict the spatiotemporal evolution of state variables, such as gas saturation and reservoir pressure, in heterogeneous geological formations. Despite the various DL models applied to this task, there is a lack of studies systematically comparing their performance. This absence of comparative analysis leads to somewhat arbitrary DL model selection in subsurface flow research, resulting in suboptimal performance and potentially inaccurate predictions. To bridge this gap, we conduct a systematic comparison study of three popular DL architectures—U‐Net, Fourier Neural Operators (FNO), and Segmentation Transformer (SETR)—in surrogate modeling of underground hydrogen storage (UHS). We focus on UHS due to its promise of enhancing clean energy resilience and its cyclic operational conditions that represent common scenarios in various subsurface applications. We evaluate the models based on accuracy, training cost, and inference speed. The comparison shows that U‐Net achieves the highest accuracy, followed by SETR and FNO. Despite its lower accuracy, FNO has the highest inference speed. SETR offers competitive accuracy with the least training memory usage, demonstrating the potential of transformers in learning subsurface flow. Our results provide guidance for selecting DL models for surrogate modeling in a wide range of subsurface flow problems.

42 ENGINEERING↗
DOE OSTI2025

The design space of E(3)-equivariant atom-centred interatomic potentials

Abstract Molecular dynamics simulation is an important tool in computational materials science and chemistry, and in the past decade it has been revolutionized by machine learning. This rapid progress in machine learning interatomic potentials has produced a number of new architectures in just the past few years. Particularly notable among these are the atomic cluster expansion, which unified many of the earlier ideas around atom-density-based descriptors, and Neural Equivariant Interatomic Potentials (NequIP), a message-passing neural network with equivariant features that exhibited state-of-the-art accuracy at the time. Here we construct a mathematical framework that unifies these models: atomic cluster expansion is extended and recast as one layer of a multi-layer architecture, while the linearized version of NequIP is understood as a particular sparsification of a much larger polynomial model. Our framework also provides a practical tool for systematically probing different choices in this unified design space. An ablation study of NequIP, via a set of experiments looking at in- and out-of-domain accuracy and smooth extrapolation very far from the training data, sheds some light on which design choices are critical to achieving high accuracy. A much-simplified version of NequIP, which we call BOTnet (for body-ordered tensor network), has an interpretable architecture and maintains its accuracy on benchmark datasets.

Computer Science↗
DOE OSTI2024

Performance of high-order Godunov-type methods in simulations of astrophysical low Mach number flows

High-order Godunov methods for gas dynamics have become a standard tool for simulating different classes of astrophysical flows. Their accuracy is mostly determined by the spatial interpolant used to reconstruct the pair of Riemann states at cell interfaces and by the Riemann solver that computes the interface fluxes. In most Godunov-type methods, these two steps can be treated independently, so that many different schemes can in principle be built from the same numerical framework. Because astrophysical simulations often test out the limits of what is feasible with the computational resources available, it is essential to find the scheme that produces the numerical solution with the desired accuracy at the lowest computational cost. However, establishing the best combination of numerical options in a Godunov-type method to be used for simulating a complex hydrodynamic problem is a nontrivial task. In fact, formally more accurate schemes do not always outperform simpler and more diffusive methods, especially if sharp gradients are present in the flow. For this work, we used our fully compressible Seven-League Hydro (SLH) code to test the accuracy of six reconstruction methods and three approximate Riemann solvers on two- and three-dimensional (2D and 3D) problems involving subsonic flows only. We considered Mach numbers in the range from 10 −3 to 10 −1 , which are characteristic of many stellar and geophysical flows. In particular, we considered a well-posed, 2D, Kelvin–Helmholtz instability problem and a 3D turbulent convection zone that excites internal gravity waves in an overlying stable layer. Although the different combinations of numerical methods converge to the same solution with increasing grid resolution for most of the quantities analyzed here, we find that (i) there is a spread of almost four orders of magnitude in computational cost per fixed accuracy between the methods tested in this study, with the most performant method being a combination of a low-dissipation Riemann solver and a sextic reconstruction scheme; (ii) the low-dissipation solver always outperforms conventional Riemann solvers on a fixed grid when the reconstruction scheme is kept the same; (iii) in simulations of turbulent flows, increasing the order of spatial reconstruction reduces the characteristic dissipation length scale achieved on a given grid even if the overall scheme is only second order accurate; (iv) reconstruction methods based on slope-limiting techniques tend to generate artificial, high-frequency acoustic waves during the evolution of the flow; and (v) unlimited reconstruction methods introduce oscillations in the thermal stratification near the convective boundary, where the entropy gradient is steep.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

Automated ICRF heating surrogate modeling via machine learning

This work introduces automated machine learning workflows that address critical bottlenecks in surrogate model development for Ion Cyclotron Range of Frequencies (ICRF) heating applications. The automated framework includes data analysis tools that transform raw datasets into actionable insights in seconds, replacing weeks of manual exploratory effort and ensuring consistent, reproducible dataset characterization. By integrating advanced hyperparameter optimization (HPO) methods including Bayesian optimization via BoTorch and Tree-structured Parzen Estimators (TPE), the framework significantly reduces model development time from weeks to hours, decreasing computational cost and required expertise, while enabling high-accuracy surrogate models. Compared to traditional hyperparameter scanning (HPS) techniques such as methodical, randomized, and grid searches, HPO methods achieve superior convergence and predictive performance, even when compared to already well-tuned reference models. On NSTX High Harmonic Fast Wave (HHFW) heating datasets, both Random Forest Regressor (RFR) and neural network surrogates demonstrate improved accuracy, achieving R 2 values beyond 0.97 and 0.98, respectively. The results show that while HPO gains are modest for robust architectures like RFR, they become essential for more sensitive models such as neural networks, highlighting the trade-offs across optimization strategies. Through automated workflows that eliminate manual hyperparameter tuning and require minimal ML expertise, this work enables widespread adoption of high-fidelity surrogate models across the fusion community for real-time plasma control, uncertainty quantification, rapid experimental scenario development, and integrated system optimization.

Sanchez-Villar, Alvaro [Princeton Plasma Physics L↗
DOE OSTI2024

Electronic structure simulations in the cloud computing environment

The transformative impact of modern computational paradigms and technologies, such as high-performance computing, quantum computing, and cloud computing, has opened up profound new opportunities for scientific simulations. Scalable computational chemistry is one beneficiary of this technological progress. The main focus of this paper is on the performance of various quantum chemical formulations, ranging from low-order methods to high-accuracy approaches, implemented in different computational chemistry packages, such as NWChem, NWChemEx, SPEC, ExaChem, and FLOSIC codes on the Azure Quantum Element (AQE) Microsoft cloud services. We pay particular attention to the intricate workflows for performing composite chemistry simulations, associated data curation, and mechanisms for accuracy assessment, as defined by the enabling cloud Computational Chemistry as a Service (CCaaS). Our focus also extends to Arrows' automated workflow for high throughput simulations. Finally, we provide a perspective on the role of cloud computing in supporting the mission of leadership computational facilities (LCFs).

computational chemistry, electronic structure, Clo↗
DOE OSTI2024

Machine learning enables identification of an alternative yeast galactose utilization pathway

How genomic differences contribute to phenotypic differences is a major question in biology. The recently characterized genomes, isolation environments, and qualitative patterns of growth on 122 sources and conditions of 1,154 strains from 1,049 fungal species (nearly all known) in the yeast subphylum Saccharomycotina provide a powerful, yet complex, dataset for addressing this question. We used a random forest algorithm trained on these genomic, metabolic, and environmental data to predict growth on several carbon sources with high accuracy. Known structural genes involved in assimilation of these sources and presence/absence patterns of growth in other sources were important features contributing to prediction accuracy. By further examining growth on galactose, we found that it can be predicted with high accuracy from either genomic (92.2%) or growth data (82.6%) but not from isolation environment data (65.6%). Prediction accuracy was even higher (93.3%) when we combined genomic and growth data. After the GALactose utilization genes, the most important feature for predicting growth on galactose was growth on galactitol, raising the hypothesis that several species in two orders, Serinales and Pichiales (containing the emerging pathogen Candida auris and the genus Ogataea, respectively), have an alternative galactose utilization pathway because they lack the GAL genes. Growth and biochemical assays confirmed that several of these species utilize galactose through an alternative oxidoreductive D-galactose pathway, rather than the canonical GAL pathway. Machine learning approaches are powerful for investigating the evolution of the yeast genotype–phenotype map, and their application will uncover novel biology, even in well-studied traits.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Towards modelling AR Sco: generalized particle dynamics and strong radiation-reaction regimes

Numerical simulations of relativistic plasmas have become more feasible, popular, and crucial for various astrophysical sources with the availability of computational resources. The necessity for high-accuracy particle dynamics is especially highlighted in pulsar modelling due to the extreme associated electromagnetic fields and particle Lorentz factors. Including the radiation-reaction force in the particle dynamics adds even more complexity to the problem, but is crucial for such extreme astrophysical sources. We have also realized the need for such modelling concerning magnetic mirroring and particle injection models proposed for AR Sco, the first white dwarf pulsar. This paper demonstrates the benefits of using higher-order explicit numerical integrators with adaptive time-step methods to solve the full particle dynamics with radiation-reaction forces included. We show that for standard test scenarios, namely various combinations of uniform E- and B-fields and a static dipole B-field, the schemes we use are equivalent to and in extreme field cases outperform standard symplectic integrators in accuracy. We show that the higher-order schemes have massive computational time improvements due to the adaptive time-steps we implement, especially in non-uniform field scenarios and included radiation reaction where the particle gyro-radius rapidly changes. When balancing accuracy and computational time, we identified the adaptive Dormand–Prince eighth-order scheme to be ideal for our use cases. The schemes we use maintain accuracy and stability in describing the particle dynamics and we indicate how a charged particle enters radiation-reaction equilibrium and conforms to the analytical Aristotelian Electrodynamics expectations.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

A Near-Real-Time Model for Predicting Electricity Disruptions in Texas During Winter Storms

There has been an increase in extreme weather events, posing a threat to power grid systems, potentially influenced by factors such as population growth, changes in ecosystems, land cover, and land use in the service area, as well as the growth of certain vegetation types. This research seeks to develop a predictive model to mitigate potential damages caused by future winter storms. This research utilizes the Light Gradient Boosting Machine (LightGBM), incorporating the number of power outages experienced at the county level, geographic details, weather information, and lagged outage and lagged weather data. The developed models were broadly divided into two groups, with six models in each group - one group without optimization and another with optimization, totaling 12 trained models. For model optimization, Bayesian optimization was employed using Root Mean Squared Error (RMSE) as the objective function. In results, when comparing Group 2 (the optimized group) with Group 1 (the non-optimized group), it was found that optimization did not always lead to a reduction in RMSE and Mean Absolute Error (MAE). However, in terms of Mean Directional Accuracy (MDA), while all results in Group 1 were below the baseline accuracy of 0.33, all results in Group 2 exceeded 0.33, with some cases showing an increase of more than three times the baseline. The results indicated that, in the optimized model group, Population and Pressure were the most influential factors when using current weather data and geographical information. When using lagged data, lagged recorded outages and lagged Pressure emerged as the most significant factors. Among the 12 developed models, the L-1-2-O model showed the lowest RMSE and MAE, as well as the highest accuracy, with values of 390.62 households and 168.13 households, respectively. To normalize the RMSE and MAE values, each metric was divided by the average number of households among the counties in Texas. For the L-1-2-O model, the scaled RMSE was 0.88% and the scaled MAE was 0.38%. In terms of MDA, which indicates the accuracy of the prediction direction, the L-1-O model achieved the highest score of 0.41. Although this study focused on Texas, which suffered the greatest impact from the winter storms in 2021, with additional validation, the methodology used in this research could be applied to other regions.

Lee, Jangjae [Texas A & M Univ., College Station, ↗
DOE OSTI2026

Real-Time GPU-Accelerated OFDR With an Integrated Auxiliary Interferometer

A GPU-accelerated optical frequency domain reflectometry (OFDR) system with an improved integrated auxiliary interferometer is proposed. Unlike conventional approaches that require separate auxiliary interferometers and multiple detection channels, the proposed OFDR system embeds this functionality directly into the signal via an intentional beat component. This enables self-calibration of laser nonlinearity while maintaining a cost-effective hardware configuration. Building on this simplified configuration, the system leverages GPU acceleration with an NVIDIA RTX 4070 Ti to achieve real-time performance, delivering high-throughput signal processing for continuous OFDR interrogation. The signal processing pipeline comprises signal capture, resampling for nonlinearity compensation, and frequency shift computation, all optimized for parallel execution. Hardware benchmarking demonstrates substantial acceleration over CPU implementations, achieving up to a 45× speedup for resampling and frequency shift computations and enabling processing latencies below 30 ms. Thermal response validation is conducted under two complementary scenarios: localized heating using a water bath and cryogenic-temperature conditions using liquid nitrogen. Under localized heating, the system achieves an accuracy of 0.249 °C with a thermal sensitivity of 5.971 GHz/°C, while cryogenic-temperature validation demonstrates a frequency shift response with a sensitivity of 2.383 GHz/°C and an accuracy of 2.04 °C. The high acceleration of the proposed GPU-accelerated OFDR system and its accuracy are achieved by exploiting CUDA-based stride indexing, enabling efficient parallel segmentation and processing of large datasets without additional memory copies. The benchmarking results confirm the robustness, accuracy, and deployability of the proposed OFDR system across a wide temperature range, establishing it as a practical platform for real-time distributed fiber sensing in structurally dynamic environments.

Harb, Salah [Lawrence Berkeley National Laboratory↗
DOE OSTI2026

Interpretable Models for Workflow Differentiation in High-Performance Scientific Networks

Scientific workflows in high-performance networks spawn hundreds of interdependent flows that must be managed collectively—yet existing network classifiers treat each flow in isolation, leading to fragmented QoS decisions and missed interflow patterns. We present a novel traffic classification solution that operates at the workflow level, distinguishing entire filetransfer operations from streaming analytics by capturing how concurrent flows interact and burst together. We introduce a workflow identification window (WIW) that ingests raw packet headers from parallel flows into unified tensors, preserving the spatial-temporal patterns that differentiate scientific workflows. This approach achieves 98.7% accuracy using CNN, LSTM, and hybrid architectures, while maintaining 84% accuracy on production traffic collected a week later—demonstrating robustness to temporal drift. By integrating SHAP and GradCAM explainability, we reveal that early-packet timing patterns and cross-flow correlations drive classification decisions, providing operators with interpretable insights. Our system enables coherent workflow-level QoS enforcement and dynamic bandwidth allocation in scientific networks, eliminating manual per-flow configuration while maintaining classification latency at millisecond level.

Giannakou, Anna [LBL, Berkeley]↗