Search NASA⌕ Search

SEARCH · Search NASA

Results for “training data”

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 181 records · Page 10
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Improving Reliability of Large Language Models for Nuclear Power Plant Diagnostics Technical Presentation

Large Language Models (LLMs) struggle out of the box when answering factually about detailed questions, especially in domains that are sparsely represented in their training data. This causes hallucinations and reduces reliability making it difficult for them to be used in practice. This work shows that using RAG techniques can improve factual accuracy and reliability, allowing for the application of LLMs in specialized areas, even when those areas that aren’t extensively covered in their initial training.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2024

Spread spectrum time domain reflectometry (SSTDR) and frequency domain reflectometry (FDR) cable inspection using machine learning

Cables are initially qualified for nuclear power plant use for 40 years. As plants extend their operating license to 60 and 80 years, justification for continued cable use must shift to a condition-based approach since it is cost prohibitive to completely replace cables that are likely still capable of performing their design function. The Pacific Northwest National Laboratory (PNNL) Accelerated and Real Time Experimental Nodal Analysis (ARENA) cable motor test bed was used to test the response of a commercial spread spectrum time domain reflectometry (SSTDR) system, a laboratory instrument software-controlled SSTDR, and a vector network analyzer-based frequency domain reflectometry (FDR) system to various cable anomalies. The three instrument systems were able to interrogate cables over a range of frequency bandwidths that can be helpful for human data analysis. Data were subjected to supervised and unsupervised machine learning (ML) analyses to distinguish normal undamaged cable responses from anomalous cable responses. Both supervised and unsupervised ML approaches produced encouraging results with an undamaged/anomalous prediction accuracy from 0.69% to 0.87%. Recommendations for further development and field implementation include increased and more balanced sample sets particularly including more training data.

SSTDR, FDR, Reflectometry, Machine Learning, ARENA↗
DOE OSTI2024

Predicting U.S. federal fleet electric vehicle charging patterns using internal combustion engine vehicle fueling transaction statistics

Utilizing fueling transactions from internal combustion engine vehicles (ICEVs), the authors estimated how frequently midday public charging would be required for U.S. federal fleet battery electric vehicles (BEVs). Fueling transaction summary statistics are more widely available than trip-level telematics data, making this methodology more accessible and transferable to other researchers and fleet managers considering BEV replacements. For example, readers can easily apply a linear model using only the count of back-to-back fueling events at gas stations over 57 straight-line miles apart to predict days exceeding range. This linear regression predicted binned days exceeding 250 miles at 80% accuracy on a hold-out test set from the same fleet as the training data and 66 % accuracy on a new fleet displaying different driving behaviors. The authors additionally provide linear equations for days exceeding 200 and 300 miles as alternative range estimates to account for differences in BEV range and temperature impacts. Beyond the single-feature linear models which readers can apply, the authors tuned and trained other machine learning models on a variety of fueling transaction statistics including consecutive transaction distances, transaction distance from garage, estimated miles traveled from fuel economy and fuel quantity, and transaction periodicity. Utilizing a subset of 1678 light-duty federal fleet vehicles which contained daily vehicle miles traveled (VMT) in addition to fueling statistics, the authors determined which fueling transaction statistics were most relevant in predicting driving days exceeding 250 miles (an approximation of BEV rated driving range). In support of the U.S. federal fleet transition to zero-emission vehicles (ZEVs), the authors used these statistics and machine learning models to predict the frequency of BEV midday charging. After training models on the subset with VMT, the authors predicted days exceeding rated range for 112,902 light-duty vehicles operating in similar circumstances in the federal fleet using a Support Vector Regressor (SVR). In conclusion, they then used the projections as part of the ZEV Planning and Charging (ZPAC) tool to identify optimal candidates for BEVs for the federal fleet. An anonymized version of ZPAC is included in the supplementary materials.

25 ENERGY STORAGE↗
DOE OSTI2025

32 examples of LLM applications in materials science and chemistry: towards automation, assistants, agents, and accelerated scientific discovery

Abstract Large language models (LLMs) are reshaping many aspects of materials science and chemistry research, enabling advances in molecular property prediction, materials design, scientific automation, knowledge extraction, and more. Recent developments demonstrate that the latest class of models are able to integrate structured and unstructured data, assist in hypothesis generation, and streamline research workflows. To explore the frontier of LLM capabilities across the research lifecycle, we review applications of LLMs through 32 total projects developed during the second annual LLM hackathon for applications in materials science and chemistry, a global hybrid event. These projects spanned seven key research areas: (1) molecular and material property prediction, (2) molecular and material design, (3) automation and novel interfaces, (4) scientific communication and education, (5) research data management and automation, (6) hypothesis generation and evaluation, and (7) knowledge extraction and reasoning from the scientific literature. Collectively, these applications illustrate how LLMs serve as versatile predictive models, platforms for rapid prototyping of domain-specific tools, and much more. In particular, improvements in both open source and proprietary LLM performance through the addition of reasoning, additional training data, and new techniques have expanded effectiveness, particularly in low-data environments and interdisciplinary research. As LLMs continue to improve, their integration into scientific workflows presents both new opportunities and new challenges, requiring ongoing exploration, continued refinement, and further research to address reliability, interpretability, and reproducibility.

Computer Science↗
DOE OSTI2025

Temporally-consistent koopman autoencoders for forecasting dynamical systems

Absence of sufficiently high-quality data often poses a key challenge in data-driven modeling of high-dimensional spatio-temporal dynamical systems. Koopman Autoencoders (KAEs) harness the expressivity of deep neural networks (DNNs), the dimension reduction capabilities of autoencoders, and the spectral properties of the Koopman operator to learn a reduced-order feature space with simpler, linear dynamics. However, the effectiveness of KAEs is hindered by limited and noisy training datasets, leading to poor generalizability. To address this, we introduce the Temporally-Consistent Koopman Autoencoder (tcKAE), designed to generate accurate long-term predictions even with limited and noisy training data. This is achieved through a consistency regularization term that enforces prediction coherence across different time steps, thus enhancing the robustness and generalizability of tcKAE over existing models. We provide analytical justification for this approach based on Koopman spectral theory and empirically demonstrate tcKAE’s superior performance over state-of-the-art KAE models across a variety of test cases, including simple pendulum oscillations, kinetic plasma, and fluid flow data.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Enhancing Drinking Water Quality Modeling: Leveraging Physics Informed Neural Networks for Learning with Imperfect Reaction Models and Partial Data

Chemical kinetics models, typically formulated as systems of ordinary or partial differential equations, are valuable tools for simulating drinking water quality. However, these models often face inaccuracies due to discrepancies between the laboratory and the real-world conditions, as well as limitations in experimental analytical methods, hindering the accurate representation of the true underlying chemical mechanisms. In this study, we propose a Physics Informed Neural Network (PINN), using the eXtreme Theory of Functional Connections, to improve the prediction of chemical concentrations over time. The PINN method accounts for imperfect chemical models and incorporates partial data to improve predictions. Focusing on reactions describing water disinfection residual and disinfectant byproduct formation, which are crucial for public health and regulatory compliance, we demonstrate that the PINN model is able to accurately predict the concentrations of chemical species across various pH values. Notably, the model extends its accuracy to predict concentrations of chemical species not originally included in its training data. The developed method can be extended to a variety of chemical systems, offering a wide array of potential applications.

13 HYDRO ENERGY↗
DOE OSTI2025

Non-intrusive reduced-order modeling for dynamical systems with spatially localized features

This work presents a non-intrusive reduced-order modeling framework for dynamical systems with spatially localized features characterized by slow singular value decay. The proposed approach builds upon two existing methodologies for reduced and full-order non-intrusive modeling, namely Operator Inference (OpInf) and sparse Full-Order Model (sFOM) inference. We decompose the domain into two complementary subdomains that exhibit fast and slow singular value decay. The dynamics of the subdomain exhibiting slow singular value decay are learned with sFOM while the dynamics with intrinsically low dimensionality on the complementary subdomain are learned with OpInf. The resulting, coupled OpInf-sFOM formulation leverages the computational efficiency of OpInf and the high resolution of sFOM, and thus enables fast non-intrusive predictions for conditions beyond those sampled in the training data set. A novel regularization technique with a closed-form solution based on the Gershgorin disk theorem is introduced to promote stable sFOM and OpInf models. We also provide a data-driven indicator for subdomain selection and ensure solution smoothness over the interface via a post-processing interpolation step. We evaluate the efficiency of the approach in terms of offline and online speedup through a quantitative, parametric computational cost analysis. We demonstrate the coupled OpInf-sFOM formulation for two test cases: a one-dimensional Burgers’ model for which accurate predictions beyond the span of the training snapshots are presented, and a two-dimensional parametric model for the Pine Island Glacier ice thickness dynamics, for which the OpInf-sFOM model achieves an average prediction error on the order of 1% with an online speedup factor of approximately 8$\times$ compared to the numerical simulation.

42 ENGINEERING↗
DOE OSTI2025

Data-driven emulation of modal aerosol microphysics via neural operator-based modeling

The complexity and the small characteristic scales of aerosol microphysical processes pose a big challenge for accurate and efficient Earth system simulations at regional and global scales. In this work, we construct and evaluate a surrogate model: the aerosol deep operator network (ADON), a physics-inspired dual-net architecture for emulating the aerosol microphysics parameterization suite in the version 2 of the Energy Earth System Model (E3SMv2). The current version of the surrogate model is trained on a dataset comprising 9.8 million samples obtained from a global E3SMv2 simulation with the horizontal resolution of about one degree under cloud-free conditions. Incorporating domain spatial and temporal coordinates, as well as principle components extracted from training data, the dual-net surrogate model effectively captures the intricate representations of aerosol and the relationship with atmospheric state variables, achieving an R-squared score over $$95.7\%$$ for all the lognormal aerosol modes in the extrapolated regime. The validated model provides feature importance of input variables and their impact on the predictive capacity of the surrogate model in relation to the E3SM. The computational cost of online inference time deployed on CPUs and GPUs with lower precisions highlights ADON’s efficiency and potential in robust predictive modeling for large-scale Earth system computations.

Bai, Zhe↗
DOE OSTI2025

CO2 Plume Imaging with Accelerated Deep Learning-based Data Assimilation Considering Multiple Realizations: Application to the Illinois Basin-Decatur Carbon Sequestration Project

We propose a fast and efficient deep learning workflow for near real-time data assimilation, forecasting and visualization of CO2 plume evolution in saline aquifer and demonstrate its application at a field site. Unlike the previous work, this study incorporates the impact of spatial heterogeneity using multiple realizations. In the proposed workflow, a neural network model utilizes available monitoring data such as downhole pressure measurements as input and predicts the propagating pressure ‘front’ using the diffusive time of flight (DTOF) map which is considered as representative reservoir image of the flow field. The DTOF is the arrival time of pressure front propagation, which can be computed by the Fast Marching Method rapidly without flow simulations. Reservoir model calibration can be implemented by selecting the training data samples that describe the predicted DTOF map based on observed data. The power and efficacy of our workflow is demonstrated by application to the Illinois Basin-Decatur Project.

CO2 plume imaging↗
DOE OSTI2025

A Gigaparsec-scale Hydrodynamic Volume Reconstructed with Deep Learning

The next generation of spectroscopic surveys will map the large-scale structure of the Universe at high redshifts (2 ≤ z ≤ 5) using millions of quasar spectra, enabling major advances in constraining both the standard cosmological model and its extensions. Robust cosmological analyses of these data sets require numerical simulations that both cover gigaparsec volumes and resolve features on ∼10 kpc scales and smaller. However, running such large-volume, high-resolution hydrodynamic simulations is computationally prohibitive. We present a generative deep learning model that enhances a low-resolution, gigaparsec-scale (960 h −1 Mpc) hydrodynamic simulation using a smaller (80 h −1 Mpc) high-resolution input hydrodynamic simulation as training data. The resulting enhanced simulation reproduces the line-of-sight power spectrum to within ∼10% and the three-dimensional power spectrum at the ∼20% level at intermediate to small scales (k ≲ 2 h Mpc −1 ). Our method shows strong promise for producing realistic simulations for cosmological analyses with current surveys such as the Dark Energy Spectroscopic Instrument and upcoming next-generation experiments, but further improvements are needed to accurately recover the large-scale modes. We publicly release the enhanced hydrodynamic simulation, along with a halo catalog from a companion N-body dark matter simulation to support the calibration of data analysis pipelines for these large-scale surveys.

Convolutional neural networks↗
DOE OSTI2026

Neural entropy-stable conservative flux form neural networks for learning hyperbolic conservation laws

We propose a neural entropy-stable conservative flux form neural network (NESCFN) for learning hyperbolic conservation laws and their associated entropy functions directly from solution trajectories, without requiring any predefined numerical discretization. While recent neural network architectures have successfully integrated classical numerical principles into learned models, most rely on prior knowledge of the governing equations or assume a fixed discretization. Our approach removes this dependency by embedding entropy-stable design principles into the learning process itself, enabling the discovery of physically consistent dynamics in a fully data-driven setting. By jointly learning both the flux function and a corresponding entropy, NESCFN promotes conservation and entropy dissipation, which is critical for long-term stability and fidelity in the system of hyperbolic conservation laws. Furthermore, numerical results demonstrate that the method achieves stability and conservation over extended time horizons and accurately captures shock propagation speeds, even without oracle access to future-time solution profiles in the training data.

Conservative flux form↗
DOE OSTI2026

Accelerating Traction Motor Optimization Design with AI Surrogate Models

The advancement of artificial intelligence systems enables the use of data-driven physics-based surrogate models to explore design spaces rapidly and deeply for engineering projects. This work presents a surrogate model workflow that accelerates electric traction motor design optimization by replacing finite element analysis (FEA) with an artificial neural network (ANN) and using this model in a genetic algorithm for design optimization. A baseline interior permanent-magnet motor is parameterized and sampled to generate FEA-labeled training data, after which a feed-forward ANN predicts key outputs (e.g., loss components and weight). The validated surrogate enables genetic-algorithm optimization and deep search over the design space without new FEA runs, producing Pareto-optimal trade-offs between weight and losses and set of optimized designs for rapid downselection of manufacturable motor designs.

Ribeiro, Pedro [ORNL] (ORCID:0009000921026641)↗
DOE OSTI2024

Circumventing data imbalance in magnetic ground state data for magnetic moment predictions

Abstract Magnetic materials play a crucial role in the transition to more sustainable forms of energy and electric vehicles. There is an anticipated shortage in magnetic materials in the future, and as a result there is an urgent need to discover and design new magnetic materials. Computational magnetic material design using density functional theory is daunting because of the challenge in identifying magnetic ground states from a combinatorially large set of possibilities. Machine learning offers a path forward by enabling efficient surrogate models that can more readily enumerate these states, but there is a dearth of training data available, and what is available tends to be imbalanced with too much non-magnetic data. In this work we show that the discrete and previously tackled data imbalance that exists at the level of the magnetic ordering leads to an imbalanced continuous distribution with many zeros when the data is unraveled at the atomic magnetic moment level, which subsequently leads to models with low accuracy for magnetic properties. We mitigate this by using a two-part model framework. Our scheme is able to classify atoms into magnetic and non-magnetic with an F1 score and Matthew’s correlation coefficient (MCC) of ~91% and then to provide an implicit embedding representation that maps directly onto the magnitude of the magnetic moment with a mean absolute error of 0.1 μ B . Beyond screening for new magnetic materials, we demonstrate an additional practical use case of our scheme: the provision of good initial guesses for magnetic moments in first-principles electronic relaxations. Such initialization is shown to lead to faster convergence to configurations that lie closer to the ground state.

Computer Science↗
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 OSTI2026

PRIME: An evaluation framework for protein representation inference and generalization in viral mutation space

Background Protein language models (PLMs) have revolutionized protein fitness prediction, yet their application to rapidly evolving viral pathogens is often confounded by extreme sequence homology. This homology leads to “data leakage” in standard random validation splits, yielding inflated performance metrics that fail to translate into real-world biosurveillance utility. Results We present Protein Representation Inference for Mutation Evaluation (PRIME), a framework that integrates domain-specific fine-tuning with a rigorous position-stratified validation protocol to evaluate viral threats. Using a dataset of 347,432 SARS-CoV-2 receptor binding domain (RBD) sequences, we demonstrate that while random training data split yields deceptive R 2 values (> 0.90), they fail to generalize to novel mutational sites. By benchmarking models up to 650 M parameters, we show that domain-specific fine-tuning of the ESM-C 600 M model with correctly stratified data provides an initial demonstration of predictive signal for binding affinity and expression at unseen mutational sites of binding affinity and expression on unseen sites (R 2 ~0.23), a significant advancement over base foundation models which exhibit no predictive power (R 2 <0). PRIME’s embedding-based clustering identified 3.03% of bat coronavirus sequences as candidates for further experimental prioritization based on their functional similarity to human-infective strains in embedding space, offering a perspective complementary to traditional phylogenetic methods. Conclusion PRIME establishes a new benchmark for the application of PLMs in pathogen surveillance. Our findings demonstrate that state-of-the-art models and fine-tuning, when paired with stratified validation, provide biologically meaningful insights into pathogen evolution and zoonotic risk.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Constraining Galaxy-Halo connection using machine learning

We investigate the potential of machine learning (ML) methods to model small-scale galaxy clustering for constraining Halo Occupation Distribution (HOD) parameters. Our analysis reveals that while many ML algorithms report good statistical fits, they often yield likelihood contours that are significantly biased in both mean values and variances relative to the true model parameters. This highlights the importance of careful data processing and algorithm selection in ML applications for galaxy clustering, as even seemingly robust methods can lead to biased results if not applied correctly. ML tools offer a promising approach to exploring the HOD parameter space with significantly reduced computational costs compared to traditional brute-force methods if their robustness is established. Using our ANN-based pipeline, we successfully recreate some standard results from recent literature. Properly restricting the HOD parameter space, transforming the training data, and carefully selecting ML algorithms are essential for achieving unbiased and robust predictions. Among the methods tested, artificial neural networks (ANNs) outperform random forests (RF) and ridge regression in predicting clustering statistics, when the HOD prior space is appropriately restricted. We demonstrate these findings using the projected two-point correlation function (w p (r p )), angular multipoles of the correlation function (ξ ℓ (r)), and the void probability function (VPF) of Luminous Red Galaxies from Dark Energy Spectroscopic Instrument mocks. Our results show that while combining w p (r p ) and VPF improves parameter constraints, adding the multipoles ξ 0 , ξ 2 , and ξ 4 to w p (r p ) does not significantly improve the constraints.

cosmology↗
DOE OSTI2025

Aided Active Learning (AAL) for Enhanced Critical Heat Flux Prediction

Accurate prediction of critical heat flux (CHF) is crucial for the safe and efficient operation of nuclear reactors. Traditional CHF modeling methods often require extensive experimental data, which are hard to obtain. This study introduces the Aided Active Learning (AAL) framework, which strategically minimizes data requirements without sacrificing model accuracy. Unlike conventional Active Learning (AL), AAL introduces an additional step of randomly selecting a subset from the sample pool before applying the query strategy. To evaluate the performance of AAL, two query strategies—uncertainty-based sampling and error-reduction sampling—were evaluated across the following models: random forest (RF), feedforward neural network (FNN), and variational feedforward neural network (vFNN). The proposed framework demonstrated that AAL effectively reduces the number of training samples needed to achieve comparable predictive accuracy. For the RF model, AL required only 710 samples to achieve an R2 score of 0.98, as compared to the 4,785 samples needed by random sampling. Similarly, the FNN model achieved the same R2 score with just 355 samples when using AL, a significant improvement over the 825 samples required by random sampling. In case of uncertainty-based sampling strategy, vFNN attained an R2 of 0.98 with 3,420 samples, reducing the sample requirement by 47% relative to the 6,440 samples needed for random sampling. Its performance suggests that larger training data are required to fully leverage its uncertainty quantification capabilities.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗