Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “model weights”

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 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Image processing pipeline for AI-driven nanoparticle megalibrary characterization

Recent innovations have made it possible to produce megalibraries, millions of structurally and compositionally distinct nanoparticles on a chip. These megalibraries yield vast volumes of data that are impossible to analyze manually, necessitating the development of automated tools. In previous work, we created a binary classification machine learning model to select quality nanoparticle images for downstream analysis. In this work, we show that adding a custom image processing step before training can produce significantly higher-performing models in a fraction of the time and make them more robust to different image noise levels and microscope acquisition settings. The image processing pipeline proposed here effectively cleans raw nanoparticle images, enhances key features, and allows us to use much lower resolution images and simpler neural network model architectures. These features result in higher performance and significant cost savings. Experiments demonstrate superior performance relative to baseline, including an 18.2% improvement in recall and a 13.1% increase in accuracy. Given the high cost of downstream analysis, it is critical to minimize false positives, and our best-performing model reaches a precision of 95.9% and a weighted F-score of 95.1% on an unseen test set. Additionally, model training time is reduced from hours to less than a minute. We also show that, using this custom image processing pipeline, model performance is significantly improved at lower pixel resolutions compared to downsizing alone. We expect that adopting this pipeline for AI-driven automated nanoparticle characterization will allow researchers to rapidly and accurately analyze much greater volumes of data, thereby accelerating materials discovery.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2026

Personal and environmental predictors of polycyclic aromatic hydrocarbon exposure identified through repeated silicone wristband sampling

This study integrates quantitative data on personal exposure to polycyclic aromatic hydrocarbons (PAHs) in 162 silicone wristbands with demographics, behavioral information, and housing characteristics to explore contributions to residential exposure in a superfund-adjacent community over the course of a year. Forty-six residents completed questionnaires and wore silicone wristbands as personal passive samplers for seven consecutive days on up to four separate occasions in alternating months between November 2022 and June 2023. It was hypothesized that individual behaviors and housing characteristics are sources of dependence and correlation between personal PAH exposures. 50 PAHs were detected at least once, 17 of which were alkylated PAHs. Exposure to PAHs of similar molecular weight was often correlated, notably between naphthalenes (2-rings) and higher molecular weight PAHs (3 or more rings). Generalized linear mixed models identified flooring type, participant age, and sampling month as important predictors of increased PAH exposure, and flooring type, and use of wood stoves or heavy machinery as predictors of increased naphthalene exposure relative to higher molecular weight PAHs. Individual chemical models based on concentration data and detection frequencies corroborated these findings across multiple PAHs. We demonstrate that personal exposure is not static and the degree of variability in personal exposure is individual. Hence, identification of influential exposure factors through repeated measures of chemical exposure and characterization of variability in personal exposure as performed in this study, is important in the development of exposure mitigation strategies.

Bonner, Emily↗
DOE OSTI2024

Accuracy optimized neural networks do not effectively model optic flow tuning in brain area MSTd

Accuracy-optimized convolutional neural networks (CNNs) have emerged as highly effective models at predicting neural responses in brain areas along the primate ventral stream, but it is largely unknown whether they effectively model neurons in the complementary primate dorsal stream. We explored how well CNNs model the optic flow tuning properties of neurons in dorsal area MSTd and we compared our results with the Non-Negative Matrix Factorization (NNMF) model, which successfully models many tuning properties of MSTd neurons. To better understand the role of computational properties in the NNMF model that give rise to optic flow tuning that resembles that of MSTd neurons, we created additional CNN model variants that implement key NNMF constraints – non-negative weights and sparse coding of optic flow. While the CNNs and NNMF models both accurately estimate the observer's self-motion from purely translational or rotational optic flow, NNMF and the CNNs with nonnegative weights yield substantially less accurate estimates than the other CNNs when tested on more complex optic flow that combines observer translation and rotation. Despite its poor accuracy, NNMF gives rise to tuning properties that align more closely with those observed in primate MSTd than any of the accuracy-optimized CNNs. This work offers a step toward a deeper understanding of the computational properties and constraints that describe the optic flow tuning of primate area MSTd.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Weighted FFT estimators for 1D and 3D correlations of the Lyman- α forest

Correlations in the Lyman-α (Lyα) forest, both as a function of line of sight separation (1D) and 3D separation, provide a unique window to the distribution of matter at redshifts not accessible by current galaxy surveys. While optimal quadratic estimators have been used to measure 1D correlations, they are computationally expensive and difficult to extend to 3D analyses. On the other hand, estimators based on the Fast Fourier Transform (FFT) are significantly faster, but are affected by missing data in the spectra (masked pixels) and so far have not used pixel weights to reduce the uncertainties in the measurement. In this publication we describe how to compute the window matrix that enables forward-modelling the impact of masked pixels and weights on the FFT-based estimators. Here, we use Gaussian and hydrodynamical simulations with artificially masked pixels to validate the method on the measurement of 1D correlations. Finally, we show that the formalism can be extended to model the impact on 3D correlations, in particular on the cross-spectrum, the correlation of 1D Fourier modes as a function of transverse separation. This work will enable more precise clustering measurements with the Lyα forest dataset recently collected by the Dark Energy Spectroscopic Instrument (DESI).

Lokken, Martine [Univ. Autonoma de Barcelona (Spai↗
DOE OSTI2025

FM4NPP: A Scaling Foundation Model for Nuclear and Particle Physics

Large language models have revolutionized artificial intelligence by enabling large, generalizable models trained through self-supervision. This paradigm has inspired the development of scientific foundation models (FMs). However, applying this capability to experimental particle physics is challenging due to the sparse, spatially distributed nature of detector data, which differs dramatically from natural language. This work addresses if an FM for particle physics can scale and generalize across diverse tasks. We introduce a new dataset with more than 11 million particle collision events and a suite of downstream tasks and labeled data for evaluation. We propose a novel self-supervised training method for detector data and demonstrate its neural scalability with models that feature up to 188 million parameters. With frozen weights and task-specific adapters, this FM consistently outperforms baseline models across all downstream tasks. The performance also exhibits robust data-efficient adaptation. Further analysis reveals that the representations extracted by the FM are task-agnostic but can be specialized via a single linear mapping for different downstream tasks.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2026

ObstacleSense: Low-Power Neuromorphic Vision for Corridor Obstacle Awareness in Low-Level ADAS

The automotive industry’s pursuit of Level 5 autonomy is constrained by substantial perception-compute power requirements, often reaching 1, 000 + watts in full autonomy stacks. Reducing this energy burden requires rethinking perception not only at the high-end autonomy level, but also at the foundational Advanced Driver Assistance Systems (ADAS) level where low-power, safety-critical sensing can have broad impact. Neuromorphic vision provides a promising starting point: HD Dynamic Vision Sensors (DVS) can operate below 100 mW at the sensor level by reporting only asynchronous brightness changes. However, low-power sensing alone is insufficient if downstream perception reintroduces dense, energy-intensive computation. In particular, many event-driven object-detection pipelines still rely on CNN backbones, while purely spiking alternatives often trade away accuracy or ignore deployment constraints. We introduce ObstacleSense, a highly compact, CNN-free hybrid ANN–SNN framework for Level 0–1 forward-corridor obstacle awareness. Instead of performing full-scene object detection with a convolutional feature backbone, ObstacleSense targets the safety-critical question of whether the ego corridor is occupied and how far the nearest obstacle is. The architecture combines polarity-conditioned event encoding, lightweight temporal spiking dynamics, axial spatial mixing, and coarse-to-fine range estimation within a regular fixed-grid compute pattern. This design avoids the dense CNN backbone commonly used in event-based detection while maintaining a small state footprint suitable for eventual small-FPGA deployment. Before hardware mapping, we evaluate the software implementation using a model-side power proxy derived from MACs, weight and activation traffic, and spiking state updates under shared FP16 assumptions. On simulated CARLA event corpora, the deployment-oriented model achieves 0.9464 objectness F1, 0.9978 grid-level mAP, and 0.8987 m distance Mean Absolute Error at an estimated 1.92 mW proxy cost, while maintaining performance on unseen generalization test sequences.

Johnson-Scott, Zac [ORNL]↗
DOE OSTI2025

Contributions of vegetation heterogeneity within tower footprint to CO 2 flux estimations through graph neural network modeling

Net ecosystem exchange of CO 2 (Fc) measured directly by eddy covariance towers is based on various assumptions, including large, flat and homogenous land cover type. In reality, often a tower site is not large enough for flux measurements, and landscapes consist of patches of different land cover types within the flux footprint. In addition, some portions of fluxes are contributed by different cover types when a footprint exceeds the size of the target ecosystem. The contributions of non-dominant patches to Fc are often ignored. Here, in this study, we propose a novel integrated modeling framework that combines random forest (RF) and XGBoost with a residual correction module based on a deep graph convolutional network (DeeperGCN) to simulate Fc for seven flux measurement sites in southwest Michigan. High-resolution remote sensing vegetation indices, soil properties, meteorological variables, and footprint-weighted spatial features were used as model inputs at three spatial resolutions (10 m, 20 m, 30 m), and their importance in predicting Fc with DeeperGCN was assessed. We found that residual correction using DeeperGCN significantly improved prediction accuracy, with the R 2 increasing from 0.9098 to 0.9479 for RF and from 0.9235 to 0.9433 for XGBoost. At site level, the maximum improvement in R 2 reached 0.1617. Paired t-tests confirmed that these improvements were statistically significant (p < 0.05). Among all predictors, leaf area index and incoming shortwave radiation emerged as the dominant drivers of spatial residual variation, followed by precipitation, relative humidity, and selected vegetation indices. The 20 m resolution yielded the best balance between model performance and computational efficiency. In conclusion, our modeling framework effectively captures both spatial heterogeneity and nonlinear interactions, offering a robust solution for spatially explicit flux modeling in structurally diverse ecosystems beyond the study sites.

footprint model↗
DOE OSTI2026

Flexible AI Models for Grid Resilience

The rapid growth in size and complexity of artificial intelligence (AI) and machine learning (ML) models has led to increased energy demands, posing a threat to the reliability of the existing power grid. This project addresses the challenge of highly intermittent and energy-intensive inference workloads by (1) developing fidelity-adaptive neural networks capable of dynamic response to grid conditions and (2) integrating these networks with power flow simulations to assess their impact on power grid reliability. We will explore both top-down and bottom-up approaches to create hierarchies of submodels that provide a controlled trade-off between power draw and prediction accuracy. The top-down method utilizes NN pruning to reduce a flagship model into progressively smaller, energy-efficient variants. The bottom-up approach employs geometrically principled weight setting strategies to construct depth-efficient models from the ground up. A real-time hardware-in-the-loop (HIL) platform will be developed to simulate a scaled AC power grid, integrating live AI workload power draw and enabling dynamic model switching in response to grid feedback. This work will provide a novel framework for evaluating the impact of flexible AI/ML workloads on grid performance and establish new methodologies for energy-aware computing in data centers. The outcomes will demonstrate that adaptive AI/ML can play a critical role in improving grid stability while advancing NREL's leadership in energy-efficient computing research.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Revealing the Crystalline Architecture of Semicrystalline Ion Exchange Membranes for the Design of Conductive and Durable Alkaline Anion Exchange Membranes

Alkaline anion exchange membrane (AAEM) fuel cells offer a cost-effective alternative to proton exchange membrane (PEM) fuel cells by eliminating the need for expensive precious metal catalysts. In both PEMs and AAEMs, semicrystalline polymers are a common choice, as the crystalline domains can act as mechanical reinforcements that limit swelling and promote mechanical durability in the material. However, spatially resolved characterization of crystalline organization in ion exchange membranes beyond ensemble-averaged X-ray scattering is underrepresented, likely in part due to ionization damage limitations in soft materials. Here, in this study, we resolve the nanometer-size crystallites in semicrystalline ion exchange membranes by applying cryogenic four-dimensional scanning transmission electron microscopy (cryo-4D-STEM) along with data-processing algorithms designed to optimize signals at a low dose to minimize radiation damage. We investigate the effects of synthesis components, including molecular weight and thermal treatment, on a model system of AAEMs in comparison to Nafion, the most commonly used and commercially successful PEM today. We find that excess water uptake in polymer membranes, a property directly associated with weak mechanical durability and with possible negative impacts on ion conductivity, can be reduced by over 30% by varying the polymer's crystalline morphology through changes in synthesis parameters such as molecular weight and thermal history. Our results indicate that this improvement is correlated with smaller crystalline domains with a more homogeneous distribution. More broadly, these results demonstrate how the crystalline architecture of polymer membranes can be tuned through their chemistry and thermal treatment in order to improve their conductivity and durability for commercial fuel cell performance.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Crack Identification and Characterization in Deformed Nb3Sn Rutherford Cable Stacks Using Machine Learning

An investigation of instance segmentation of cracks in Nb3Sn 4-stack 40-strand Rutherford cables using machine learning is presented. Three samples were uniaxially and biaxially loaded before metallographic inspections were performed. The Mask R-CNN model was used in the Detectron2 framework with pre-trained weights but fine-tuned to detect and segment cracks. The model detected cracks with bounding box and mask average precisions (AP) of 42.8 and 27.9, respectively, and was used for instance segmentation of all cracks in the three samples. More cracks were found in the sample pre-loaded along the z-axis (i.e., along the cable length). Pre-loading along the x-axis (i.e., on the cables edges) reduced the number of cracks and changed the crack orientation distribution, away from being highly aligned with the y-axis (i.e., normal to the cables broad faces), i.e., the direction with the highest applied load. Fine-tuning of the Segment Anything Model (SAM) was also studied but performed poorly without human-provided prompts. However, the zero-shot capability of SAM showed high promises to accelerate the image annotation process for applications beyond this study.

Croteau, Jean-Francois↗
DOE OSTI2025

PHASE: Personalized Head-based Automatic Simulation for Electromagnetic properties in 7T MRI

Accurate and individualized human head models are becoming increasingly important for electromagnetic (EM) simulations. These simulations depend on precise anatomical representations to realistically model electric and magnetic field distributions, particularly when evaluating Specific Absorption Rate (SAR) within safety guidelines. State of the art simulations use the Virtual Population due to limited public resources and the impracticality of manually annotating patient data at scale. Here, this paper introduces Personalized Head-based Automatic Simulation for EM properties (PHASE), an automated open-source toolbox that generates high-resolution, patient-specific head models for EM simulations using paired T1-weighted (T1w) magnetic resonance imaging (MRI) and computed tomography (CT) scans with 14 tissue labels. To evaluate the performance of PHASE models, we conduct semi-automated segmentation and EM simulations on 15 real human patients, serving as the gold standard reference. The PHASE model achieved comparable global SAR and localized SAR averaged over 10 grams of tissue (SAR-10g), demonstrating its potential as a promising tool for generating large-scale human model datasets in the future. The code and models of PHASE toolbox have been made publicly available: https://github.com/hrlblab/PHASE.

Deep learning↗
DOE OSTI2026

Natural Language Processing to Inform Agent-Based Modeling: With Application to Modeling Adoption of Medium-Duty Electric Vehicles

Agent-based socio-technical modeling of medium- and heavy-duty (MDHD) electric vehicle (EV) adoption has the potential to provide analysis, prediction, and gui. This paper describes new applications of text analysis developed through machine learning (ML) to build and understand relevant topics and their saliency in the published discourse on adoption of MDHD EVs. This work contributes to the state of the art in topic mining models by defining a new metric of topic ranking (START) that quantifies the importance of predefined topics within the corpus using weighted results for predefined topics from two topic modeling approaches: Latent Dirichlet Allocation (LDA) and BERTopic. The START metric is then demonstrated in practice to model how academia and industry view the EV adoption process based on the respective texts published by these groups. Results show that academic literature places more emphasis on categories of interests such as norms/attitudes and adopter knowledge, while trade journals tend to emphasize long-term cost more than academia. The two bodies of literature agree on the importance of policy and incentives in MDHD EV adoption. Together these results illustrate the potential to use ML-based text analysis to populate the characteristics of agent-based socio-technical models.

Electric vehicle adoption, fleet electrification, ↗
DOE OSTI2025

Analytical simulation of meander morphology from equilibrium to long-term evolution: Impacts of channel geometry and vegetation-induced coarsening

This study introduces an innovative approach to modeling meandering river morphology, integrating and investigating the effects of geometric characteristics and vegetation-induced channel coarsening. The developed comprehensive framework combines several advanced techniques: Genetic Programming for refining the scour factor of transverse bed slope, a Leaf Area Index (LAI)-enhanced analytical model for quantifying vegetative flow resistance, and an upstream-weighted moving average method for efficient approximation of the convolution integral in meander migration calculations. The model is validated against both an idealized Kinoshita meander and a natural bend of the Tumen River (China) in equilibrium, demonstrating its robustness across diverse scales and conditions. The model's ability to simulate the long-term evolution, including cutoff events, provides valuable insight for river management strategies. The current findings demonstrate that channel geometry, particularly width-to-depth ratio, plays a dominant role in meander evolution, with wider channels prone to more complex and rapid morphological changes. Vegetation effects are most pronounced in channels with moderate width-to-depth ratios, where they can significantly influence migration rates and bed topography. A combination of channel widening and deepening, coupled with strategic vegetation management, can effectively enhance navigability while maintaining channel stability in the studied Tumen River reach. Sensitivity analyses highlight the complex interplay between hydraulic conditions, sediment characteristics, and vegetation in shaping river morphology. This research advances understanding of the multifaceted nature of meandering river systems and offers practical tools for informed decision-making in river engineering and environmental management, particularly in the context of climate change and increasing anthropogenic pressures on fluvial ecosystems.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Spike-and-Slab Shrinkage Priors for Structurally Sparse Bayesian Neural Networks

Network complexity and computational efficiency have become increasingly significant aspects of deep learning. Sparse deep learning addresses these challenges by recovering a sparse representation of the underlying target function by reducing heavily overparameterized deep neural networks. Specifically, deep neural architectures compressed via structured sparsity (e.g., node sparsity) provide low-latency inference, higher data throughput, and reduced energy consumption. In this article, we explore two well-established shrinkage techniques, Lasso and Horseshoe, for model compression in Bayesian neural networks (BNNs). To this end, we propose structurally sparse BNNs, which systematically prune excessive nodes with the following: 1) spike-and-slab group Lasso (SS-GL) and 2) SS group Horseshoe (SS-GHS) priors, and develop computationally tractable variational inference, including continuous relaxation of Bernoulli variables. We establish the contraction rates of the variational posterior of our proposed models as a function of the network topology, layerwise node cardinalities, and bounds on the network weights. Furthermore, we empirically demonstrate the competitive performance of our models compared with the baseline models in prediction accuracy, model compression, and inference latency.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Vehicle-Cycle Inventory for Type C School Buses & Intra-City Transit Buses

This report documents the new inventory incorporated into the Research and Development version of Greenhouse gases, Regulated Emissions, and Energy use in Technologies (R&D GREET) 2025 model for the vehicle cycle of Type C school buses and intracity transit buses. The transportation sector contributes significantly to the United States’ energy consumption and resultant emissions (EPA, 2025a). However, public transit plays an important role in mitigating these impacts because it consumes a relatively low amount of energy per passenger (Congressional Budget Office, 2022). Public transit is widely used in the United States; more than 500,000 school buses (EPA, 2025b) and ~75,000 service buses (American Public Transportation Association, 2025) operate in the nation. These are primarily internal combustion engine vehicles (ICEVs) powered by diesel. Original equipment manufacturers (OEMs) are making efforts to electrify U.S. bus fleets by using batteries as a propulsion system to replace internal combustion engines. Electrification can reduce tailpipe emissions, such as particulate matter (with a diameter ≤10 µm [PM 10 ] and with a diameter ≤2.5 µm [PM 2.5 ]) and nitrogen oxides (Jonas et al., 2025; Martinez and Samaras, 2024; EPA, 2025b; Wayne et al., 2009). Hence, any energy and emission impact analysis of public transit must consider both conventional ICEVs and upcoming electric vehicle (EV) options for the school and transit buses that dominate this landscape. To understand the detailed environmental impact profiles of ICEV and EV school and transit buses, it is necessary to conduct a thorough analysis covering both vehicle manufacturing and vehicle use stages. The current literature lacks a detailed vehicle-cycle inventory for school and transit buses, which makes this kind of comparison difficult. To overcome this gap, we developed a comprehensive vehicle-cycle model for school and transit buses in Argonne’s R&D GREET 2025 model. The model is flexible in handling user inputs for key assumptions, such as component weights and material compositions, upstream energy sources for material processing, and vehicle operating parameters, to understand their impacts on energy use and emissions for both school and transit buses. This report is organized as follows: Section 2 provides details on the modeling approach and vehicle specifications (weights and composition of different vehicle components, and vehicle operating parameters) for both school and transit buses. Section 3 provides details on vehicle assembly, disposal, and recycling (ADR) approaches for the two buses. Section 4 includes details about their incorporation into the R&D GREET model.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

Resin Testing and Modeling for Optimal Composite Processing

Polymer composites have properties such as high strength and stiffness, low weight, good thermal and chemical stability, as well as impact and abrasion resistance that make them ideal for high-performance applications. The chemistries of these materials are continuously improving, so determining their properties is vital for successfully producing them and achieving the desired results. Multiple methods can be employed to monitor characteristics such as heat flow, weight, dimension, and modulus as a function of time and temperature. By analyzing this information, models can be developed to predict outcomes of parameters not tested for. In one application, materials proposed for wet filament winding and the production of high pressure vessels can be analyzed to verify they will have the necessary low viscosity for good fiber wetting and long pot life for the extended handling inherent to this process. Such data about a prospective system provides valuable information on how that material could ultimately be processed to yield the desired part.

36 MATERIALS SCIENCE↗
DOE OSTI2025

SineKAN: Kolmogorov-Arnold Networks using sinusoidal activation functions

Recent work has established an alternative to traditional multi-layer perceptron neural networks in the form of Kolmogorov-Arnold Networks (KAN). The general KAN framework uses learnable activation functions on the edges of the computational graph followed by summation on nodes. The learnable edge activation functions in the original implementation are basis spline functions (B-Spline). Here, we present a model in which learnable grids of B-Spline activation functions are replaced by grids of re-weighted sine functions (SineKAN). We evaluate numerical performance of our model on a benchmark vision task. We show that our model can perform better than or comparable to B-Spline KAN models and an alternative KAN implementation based on periodic cosine and sine functions representing a Fourier Series. Further, we show that SineKAN has numerical accuracy that could scale comparably to dense neural networks (DNNs). Compared to the two baseline KAN models, SineKAN achieves a substantial speed increase at all hidden layer sizes, batch sizes, and depths. Current advantage of DNNs due to hardware and software optimizations are discussed along with theoretical scaling. Additionally, properties of SineKAN compared to other KAN implementations and current limitations are also discussed.

Reinhardt, Eric↗