SEARCH · Engineering Papers

Results for “Generative models”

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.

SourceResearch connectionFrom yearThrough year

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

At least 1,297 records · Page 72

All sources NASA NTRS DOE OSTI

DOE OSTI2026

Emerging Flexible Designs for Geospatial Multimodal Foundation Models

Foundation models are rapidly transforming Earth observation by enabling scalable pretraining across diverse unlabeled geospatial modalities. However, their architectural diversity—ranging from encoder-only to encoder-decoder and masked autoencoding paradigms—makes it challenging to assess performance trade-offs in a consistent manner. In this work, we present an apples-to-apples comparison of leading FM architectures designed for geospatial multimodal reasoning, with a particular focus on flexibility across varied spectral band configurations. We standardize pretraining using identical self-supervised learning objectives and training datasets, and evaluate all models under consistent parameterization on the GEOBench benchmark across classification and segmentation tasks. Our results offer new insights into the design trade-offs between model flexibility, modality alignment, and downstream task performance. By highlighting architectural strengths and limitations under controlled conditions, this study provides practical guidance for building next-generation geospatial foundation models capable of robust multimodal reasoning.

Ambrozio Dias, Philipe [ORNL] (ORCID:0000000194277↗

DOE OSTI2020

Analysis of THz generation by multicolor laser pulses with various frequency ratios

Here, terahertz (THz) generation in a gas medium is simulated by quantum calculations with multicolor intense laser pulses of various frequency ratios. By correlating THz radiation with the above-threshold-ionization photoelectron spectrum, we confirm that our previously proposed free-free transition model for the generation of THz radiation is also applicable to multicolor pulses. THz photon emissions can proceed through transitions between continuous states with similar energies that have opposite parities. This mechanism predicts that THz waves can be efficiently generated at special two-color laser frequency ratios when the multiphoton mixing condition is satisfied. Applying this model to multicolor fields (sawtooth wave shape), we provide a quantum-mechanical interpretation for the reason for THz radiation enhancement. A scheme combining multiple lasers to raise THz radiation satisfying multiphoton mixing conditions is also proposed.

74 ATOMIC AND MOLECULAR PHYSICS↗

DOE OSTI2020

Modelling the Milky Way – I. Method and first results fitting the thick disc and halo with DES-Y3 data

ABSTRACT We present a technique to fit the stellar components of the Galaxy by comparing Hess Diagrams (HDs) generated from trilegal models to real data. We apply this technique, which we call mwfitting, to photometric data from the first 3 yr of the Dark Energy Survey (DES). After removing regions containing known resolved stellar systems such as globular clusters, dwarf galaxies, nearby galaxies, the Large Magellanic Cloud, and the Sagittarius Stream, our main sample spans a total area of ∼2300 deg2. We further explore a smaller subset (∼1300 deg2) that excludes all regions with known stellar streams and stellar overdensities. Validation tests on synthetic data possessing similar properties to the DES data show that the method is able to recover input parameters with a precision better than 3 per cent. We fit the DES data with an exponential thick disc model and an oblate double power-law halo model. We find that the best-fitting thick disc model has radial and vertical scale heights of 2.67 ± 0.09 kpc and 925 ± 40 pc, respectively. The stellar halo is fit with a broken power-law density profile with an oblateness of 0.75 ± 0.01, an inner index of 1.82 ± 0.08, an outer index of 4.14 ± 0.05, and a break at 18.52 ± 0.27 kpc from the Galactic centre. Several previously discovered stellar overdensities are recovered in the residual stellar density map, showing the reliability of mwfitting in determining the Galactic components. Simulations made with the best-fitting parameters are a promising way to predict Milky Way star counts for surveys such as the LSST and Euclid.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2022

Investigation of Remanufacturing Methods Utilizing Mechanically-Generated and Gas-Atomized Feedstock Production for Directed Energy Deposition Processing Through Specific Energy Consumption Modeling

The goal of this work was to assess the energy consumption of a novel remanufacturing method that utilizes mechanically-generated feedstock for directed energy deposition. Gas-atomization is the state-of-the-art production process for directed energy deposition feedstock, and for this work was incorporated into a remanufacturing process path to provide a comparison method. Specific energy consumption models of remanufacturing utilizing both of these proposed process paths were developed and applied to a case study to investigate energy saving opportunities for future manufacturing paradigms. The energy modelling was done in three stages. First, the mechanically-generated feedstock production energy consumption model was generated from experimental observation. Second, the gas-atomized feedstock production energy consumption model was generated from a combination of experimental observation, reported estimates from the manufacturer, and data found in the literature. Lastly, the energy consumption model of directed energy deposition was derived from experimental observation and compares favorably with reported estimates in the literature. With the models, the specific energy consumption in the two process paths were compared and their application was demonstrated by estimating the energy consumption to remanufacture a bracket. The two feedstock production methods had similar specific energy consumptions. The specific energy consumption of the directed energy deposition process was the greatest component in the respective remanufacturing paths by an order of magnitude; increasing deposition rate is the most important factor for lowering the overall specific energy consumption. The analyzed remanufacturing technologies were estimated to consume less energy than replacement when repairing up to approximately 15% of the original part’s mass.

42 ENGINEERING↗

DOE OSTI2022

The Virtual Test Bed (VTB) repository: a library of multiphysics reference reactor models using NEAMS tools

With the next generation of nuclear reactors under development, modeling and simulation (MS) tools are being developed by the Nuclear Energy Advanced Modeling and Simulation (NEAMS) program in order to support their design, licensing, and future operation. Mirroring the physical test beds currently under construction (i.e., EBR-II and ZPPR), the Virtual Test Bed (VTB) was launched by the National Reactor Innovation Center (NRIC) in collaboration with NEAMS to support the advanced reactor community. This collaborative effort, which involves multiple teams at both Idaho National Laboratory and Argonne National Laboratory aims to use NEAMS tools to model a wide range of reactor designs. Those models are automatically tested to ensure their continued functionality as the tools are further developed. Examples are extensively documented, each acting as a tutorial for applying the relevant NEAMS tools to that reactor design. Currently, five advanced reactor types (with a total of eight specific design variants) are simulated by a variety of different models. These models range from steady-state, core multiphysics simulations to integrated plant analysis during loss-of flow transients. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

DOE OSTI2022

The Virtual Test Bed (VTB) Repository: A Library of Multiphysics Reference Reactor Models using NEAMS Tools

With the next generation of nuclear reactors under development, modeling and simulation (M&S) tools are being developed by the Nuclear Energy Advanced Modeling and Simulation (NEAMS) program in order to support their design, licensing, and future operation. Mirroring the physical test beds currently under construction (i.e., EBR-II and ZPPR), the Virtual Test Bed (VTB) was launched by the National Reactor Innovation Center (NRIC) in collaboration with NEAMS to support the advanced reactor community. This collaborative effort, which involves multiple teams at both Idaho National Laboratory and Argonne National Laboratory aims to use NEAMS tools to model a wide range of reactor designs. Those models are automatically tested to ensure their continued functionality as the tools are further developed. Examples are extensively documented, each acting as a tutorial for applying the relevant NEAMS tools to that reactor design. Currently, five advanced reactor types (with a total of eight specific design variants) are simulated by a variety of different models. These models range from steady-state, core multiphysics simulations to integrated plant analysis during loss-of-flow transients.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2020

Model-Free Dynamic Voltage Control of a Synchronous Generator-Based Microgrid

The main goal of this paper is to present a new dynamic voltage stability mechanism, based on model-free control (MFC), for effective control and coordination of synchronous generator (SG)-based reactive power resources in a microgrid setting. MFC has shown successful operation in various domains, and this paper presents its first use in the voltage stability of a power system. It is utilized as an online controller to achieve the dynamic voltage stability of a microgrid system under different disturbances and fault conditions. A 21-bus microgrid system fed by SG-based distributed energy resources (DERs) is considered as a use case study. The overall dynamic voltage stability of the microgrid system is investigated using time-domain dynamic simulations during emergency, hazard, and disaster events. Simulation results show that there are significant improvements and enhancements on the dynamic load bus voltage profiles of the microgrid system by the effective model-free control and coordination of the reactive power reserves of the SG-based DERs.

Hatipoglu, Kenan↗

DOE OSTI2025

A matheuristic for design and dispatch of a utility-connected distributed energy system

Modeling distributed power generation systems often requires complicated mathematical expressions that present challenges for commercial optimization solvers. Here, this paper presents a matheuristic to solve a mixed-integer optimization model that informs decisions regarding the design and dispatch of a utility-connected microgrid. We deploy a genetic algorithm to search the system design space and a linear program to solve the economic dispatch problem. The model is a component of a web tool that requires solutions within a few minutes. Our method yields objective function values within 5% of an exogenously produced optimal in fewer than 30 seconds for 90% of our test cases compared to only 10% of our test cases by a traditional optimization solver in the same amount of time.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2023

Graph-based Compact Modeling (GCM) of CMOS transistors for efficient parameter extraction: A machine learning approach

Parameter extraction of compact transistor models is an expensive process, heavily relying on engineering knowledge and experience. To automate such a process, we propose a novel approach, Graph-based Compact Model (GCM), that integrates physical modeling and data-driven learning. GCM utilizes Graph Neural Networks (GNNs) to establish the model structure, while retaining the physicality in compact models. Here, we implement our GCM in Verilog-A to support circuit simulations. As demonstrated with an academic 7 nm FinFET PDK, the new approach automatically generates a GCM model within a minute, and achieves excellent accuracy and efficiency in SPICE.

42 ENGINEERING↗

DOE OSTI2022

Communication System Modeling in Transactive Systems

Transactive energy systems almost always rely on communication systems for proper operation but most analysis of transactive systems do not model the communication system. Often these analysis are performed by those without a communication system modeling or simulation background and the difficulty of implementing such models in the analysis environment is prohibitive. Without this model, an understanding of the communication system requirements to successfully implement a transactive energy system can not be comprehended. This report details a capability of auto-generating communication system models from an electrical distribution system model, discusses the need for such models, discusses the method by which these models were developed in this project, and demonstrates the impact on the performance of a load management system when using such models. This capability has been incorporated into Pacific Northwest National Laboratory’s (PNNL) Transactive Energy Simulation Platform (TESP).

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2021

Developing Practical Models of Complex Salts for Molten Salt Reactors

Molten salt reactors (MSRs) utilize salts as coolant or as the fuel and coolant together with fissile isotopes dissolved in the salt. It is necessary to therefore understand the behavior of the salts to effectively design, operate, and regulate such reactors, and thus there is a need for thermodynamic models for the salt systems. Molten salts, however, are difficult to represent as they exhibit short-range order that is dependent on both composition and temperature. A widely useful approach is the modified quasichemical model in the quadruplet approximation that provides for consideration of first- and second-nearest-neighbor coordination and interactions. Its use in the CALPHAD approach to system modeling requires fitting parameters using standard thermodynamic data such as phase equilibria, heat capacity, and others. A shortcoming of the model is its inability to directly vary coordination numbers with composition or temperature. Another issue is the difficulty in fitting model parameters using regression methods without already having very good initial values. The proposed paper will discuss these issues and note some practical methods for the effective generation of useful models.

Besmann, Theodore M. (ORCID:0000000155980550)↗

DOE OSTI2020

Integrating Light Curve and Atmospheric Modeling of Transiting Exoplanets

Spectral retrieval techniques are currently our best tool to interpret the observed exoplanet atmospheric data. Said techniques retrieve the optimal atmospheric components and parameters by identifying the best fit to an observed transmission/emission spectrum. Over the past decade, our understanding of remote worlds in our galaxy has flourished thanks to the use of increasingly sophisticated spectral retrieval techniques and the collective effort of the community working on exoplanet atmospheric models. A new generation of instruments in space and from the ground is expected to deliver higher quality data in the next decade; it is therefore paramount to upgrade current models and improve their reliability, their completeness, and the numerical speed with which they can be run. In this paper, we address the issue of reliability of the results provided by retrieval models in the presence of systematics of unknown origin. More specifically, we demonstrate that if we fit directly individual light curves at different wavelengths (L-retrieval), instead of fitting transit or eclipse depths, as it is currently done (S-retrieval), the said methodology is more sensitive against astrophysical and instrumental noise. This new approach is tested, in particular, when discrepant simulated observations from Hubble Space Telescope/Wide Field Camera 3 and Spitzer/IRAC are combined. We find that while S-retrievals converge to an incorrect solution without any warning, L-retrievals are able to flag potential discrepancies between the data sets.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Li-ion battery design through microstructural optimization using generative AI

Lithium-ion batteries are used across various applications, necessitating tailored cell designs to enhance performance. Optimizing electrode manufacturing parameters is a key route to achieving this, as these parameters directly influence the microstructure and performance of the cells. However, linking process parameters to performance is complex, and experimental or modeling campaigns are often slow and expensive. This study introduces a fast computational optimization framework for electrode manufacturing parameters. A generative model, trained on a small dataset of microstructural images associated with different manufacturing parameters, efficiently generates representative microstructures for new parameters. This model is integrated into a Bayesian optimization loop that includes microstructure generation, characterization, and simulation, aiming to find optimal manufacturing parameters for a particular application. Significant improvement in the energy density of a 4680 cell is achieved through bespoke cell design, highlighting the importance of cell-scale normalization. The framework’s modularity allows its application to various advanced materials manufacturing scenarios.

batteries↗

DOE OSTI2025

Monitoring and modeling hydrologic conditions in Ukraine for hydropower generation

Study region: The Dnieper and Dniester Rivers of Ukraine. Study focus: The ongoing conflict in Ukraine has caused disruptions to electricity generation, of which hydroelectric sources contribute approximately 9 % to the country’s needs. With the takeover of the Zaporizhzhia nuclear power plant by enemy forces, the loss of the Kakhovka hydroelectric dam, and the future impacts of the conflict on electricity generation unclear, it may be valuable for the Ukrainian government to better understand how it could leverage hydroelectric power sources in the near future. Unfortunately, measurements of river discharge throughout Ukraine ceased data collection in the late 1980’s to early 1990’s. To address this data gap, we developed a protocol that combined satellite-based time-series measurements of river width at seven locations throughout Ukraine from 2013 to 2023 with reanalysis data, climate-model predictions, and hydrologic models to both provide a means of monitoring a proxy for near-real-time discharge and also predict near-term (i.e., 2023–2030) hydrologic patterns for the region. New hydrological insights for the region: We ran new algorithms on 144 WorldView-2 and WorldView-3 satellite images to map rivers and extract width, one of which was validated against river gauge data located along the same river but in a neighboring country. Hydrologic models using two climate scenarios found minimal change in annual discharge at all sites, but magnitude and timing of peak discharge showed a moderate trend. The results suggest that hydropower is underutilized in Ukraine.

13 HYDRO ENERGY↗

DOE OSTI2021

Distributed Solar and Storage Adoption Modeling

The National Renewable Energy Laboratory (NREL) is analyzing the rapidly increasing role of energy storage in the electrical grid through 2050 through its Storage Futures Study. In one phase of the study, NREL used the laboratory's Distributed Generation Market (dGen) model to examine the various future distributed storage capacity adoption scenarios, results, and implications.

agent-based modeling↗

DOE OSTI2024

ESM data downscaling: a comparison of super-resolution deep learning models

Abstract Climate projections at fine spatial resolutions are required to conduct accurate risk assessment for critical infrastructure and design adaptation planning. Generating these projections using advanced Earth system models (ESM) requires significant computational resources. To address this issue, various statistical downscaling techniques have been introduced to generate fine-resolution data from coarse-resolution simulations. In this study, we evaluate and compare five deep learning-based downscaling techniques, namely, super-resolution convolutional neural networks, fast super-resolution convolutional neural network ESM, efficient sub-pixel convolutional neural network, enhanced deep residual network (EDRN), and super-resolution generative adversarial network (SRGAN). These techniques are applied to a dataset generated by the Energy Exascale Earth System Model (E3SM), focusing on key surface variables such as surface temperature, shortwave heat flux, and longwave heat flux. Models are trained and validated using paired fine-resolution (0.25 $$^{\circ }$$ ∘ ) and coarse-resolution (1 $$^{\circ }$$ ∘ ) monthly data obtained from a 9-year simulation. Next, blind testing is performed using monthly data obtained from two different years outside of the training and validation set. To evaluate the efficiency of each technique, different statistical metrics are used, including mean squared error (MSE), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and learned perceptual image patch similarity (LPIPS). The results show that EDRN outperforms other algorithms in terms of PSNR, SSIM, and MSE, but struggles to capture fine-scale features in the data. In contrast, SRGAN, a generative model that uses perceptual loss, excels in capturing fine details at boundaries and internal structures, resulting in lower LPIPS than other methods.

Pawar, Nikhil M. (ORCID:0000000211613289)↗

DOE OSTI2024

Assessment of fine-tuned large language models for real-world chemistry and material science applications

The current generation of large language models (LLMs) has limited chemical knowledge. Recently, it has been shown that these LLMs can learn and predict chemical properties through fine-tuning. Using natural language to train machine learning models opens doors to a wider chemical audience, as field-specific featurization techniques can be omitted. In this work, we explore the potential and limitations of this approach. We studied the performance of fine-tuning three open-source LLMs (GPT-J-6B, Llama-3.1-8B, and Mistral-7B) for a range of different chemical questions. We benchmark their performances against “traditional” machine learning models and find that, in most cases, the fine-tuning approach is superior for a simple classification problem. Depending on the size of the dataset and the type of questions, we also successfully address more sophisticated problems. The most important conclusions of this work are that, for all datasets considered, their conversion into an LLM fine-tuning training set is straightforward and that fine-tuning with even relatively small datasets leads to predictive models. These results suggest that the systematic use of LLMs to guide experiments and simulations will be a powerful technique in any research study, significantly reducing unnecessary experiments or computations.

Van Herck, Joren↗

DOE OSTI2021

Modeling laser-driven ion acceleration with deep learning

Developments in machine learning promise to ameliorate some of the challenges of modeling complex physical systems through neural-network-based surrogate models. High-intensity, short-pulse lasers can be used to accelerate ions to mega-electronvolt energies, but to model such interactions requires computationally expensive techniques such as particle-in-cell simulations. Multilayer neural networks allow one to take a relatively sparse ensemble of simulations and generate a surrogate model that can be used to rapidly search the parameter space of interest. In this study, we created an ensemble of over 1,000 simulations modeling laser-driven ion acceleration and developed a surrogate to study the resulting parameter space. A neural-network-based approach allows for rapid feature discovery not possible for traditional parameter scans given the computational cost. A notable observation made during this study was the dependence of ion energy on the pre-plasma gradient length scale. While this methodology harbors great promise for ion acceleration, it has ready application to all topics in which large-scale parameter scans are restricted by significant computational cost or relatively large, but sparse, domains.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗