Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “deep 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.

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

At least 595 records · Page 33
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Electric Vehicles at Scale (EVs@Scale) Laboratory Consortium Deep-Dive Technical Meetings: High Power Charging (HPC) Summary Report

Electric vehicle (EV) adoption will change the composition of EV charging load to higher-power charging as more medium- (MD) and heavy-duty (HD) applications are electrified, and as all vocations, including light-duty (LD) vehicles, are capable of faster charging. These shifts provide the opportunity for high-power charging (HPC) and facility equipment to evolve and improve efficiency, cost, and space. High-Power Electric Vehicle Charging Hub Integration Platform (eCHIP) project designs and develops a high-power, interoperable charging experimental platform to research, develop, and demonstrate the integration approaches and technology solutions. The project addresses (1) interconnection and management of a grid-tied inverter; (2) development of a DC distribution system that is responsible for system energy management, interoperability, and DC protection; (3) modular DC/DC conversion for vehicle charging; (4) EV charging interface and DC/DC integration; and (5) smart charge control and vehicle-to-edge (vehicle-to-building [V2B], vehicle-to-everything [V2X]) capability. This summary presentation is the first technical progress output of the project. It provides insight for the first deep-dive technical meeting outputs in terms of research presentations and also includes summary of the discussions occurred in the follow-up breakout sessions. The summary presentation covers three technical areas: (1) HPC: State of the art power architectures and the design of the power electronics, (2) Modeling, energy management, and power control in the HPC station, and (3) Next Generation Profiles for high power charging characterization.

ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2021

Assessment of alternative radionuclides for use in a radioisotope thermoelectric generator

238 Pu is the current fuel source for MMRTGs. This is due to favorable properties, such as high decay heat and low gamma emissions. 238 Pu supplies are dwindling rapidly, necessitating the review of potential replacement isotopes for deep space missions. This work considers nine potential replacements, Ac-227, Am-241, Cf-252, Cm-244, Cs-137, Po-210, Sr-90, Th-228, and Th-170. We have assessed performance of alternatives to 238 Pu by modeling the specific decay-heat and gamma spectra of several potential radioisotopes in SCALE-ORIGEN, and comparing these with 238 Pu. An analysis of heat output, heat stability, safety, availability and required mass, baselined against 238 Pu demonstrates that of those radioisotopes studied, 241 Am has properties that make it most conducive for 238 Pu replacement.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2025

An efficient surrogate model of secondary electron formation and evolution

This work extends the adjoint-deep learning framework for runaway electron (RE) evolution, developed by McDevitt et al. [Phys. Plasmas 32, 042503 (2025)], to account for large-angle collisions. By incorporating large-angle collisions, the framework allows the avalanche of REs to be captured, an essential component of RE dynamics. This extension is accomplished by using a Rosenbluth–Putvinski approximation to estimate the distribution of secondary electrons generated by large-angle collisions. By evolving both the primary and multiple generations of secondary electrons, the present formulation can capture both the detailed temporal evolution of a RE population beginning from an arbitrary initial momentum space distribution, along with providing approximations to the saturated growth and decay rates of the RE population. Predictions of the adjoint-deep learning framework are verified against a traditional RE solver, with good agreement present across a broad range of parameters.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

Static versioning in the polyhedral model

An approach is presented to enhancing the optimization process in a polyhedral compiler by introducing compile-time versioning, i.e., the production of several versions of optimized code under varying assumptions on its run-time parameters. We illustrate this process by enabling versioning in the polyhedral processor placement pass. We propose an efficient code generation method and validate that versioning can be useful in a polyhedral compiler by performing benchmarking on a small set of deep learning layers defined for dynamically-sized tensors.

Meister, Benoit J.↗
DOE OSTI2026

Hadron-Argon Cross-Section Measurements at Protodune

The Deep Underground Neutrino Experiment (DUNE) is a next-generation experiment designed to measure neutrino oscillations with unprecedented precision, specifically seeking to identify CP violation in the lepton sector. DUNE utilizes the Liquid Argon Time Projection Chamber (LArTPC) as its core detecting technology, offering high-resolution imaging to reconstruct neutrino interactions on argon. A primary challenge in this reconstruction is the modeling of hadronic final-state interactions (FSI), where hadrons produced at the initial interaction vertex undergo further scattering before escaping the nucleus. In this thesis, I utilize the ProtoDUNE-SP detector and its 1~GeV/$c$ hadron beam at the CERN Neutrino Platform to present the first measurements of total inelastic $\pi^+$-argon and proton-argon cross sections in an energy regime critical to DUNE. I detail the development and validation of the ``slicing method'', which utilizes a kiloton-scale LArTPC to extract hadron-argon cross sections. These measurements provide an indispensable benchmark for informing DUNE's FSI modeling, and our results show no significant tension with current model predictions. Additionally, I study the impact of FSI modeling on DUNE’s sensitivity to oscillation parameters. This study demonstrates that variations in FSI models may be degenerate with oscillation features, highlighting the necessity of using direct experimental data to refine interaction models. Finally, I provide an outlook on related studies that will further contribute to this effort, supporting the ambitious physics goals of upcoming neutrino oscillation programs including DUNE.

Yin-Rui, Liu [U. Chicago (main)]↗
DOE OSTI2025

Reduced-order modeling for efficient cross section library development in high-temperature gas reactor pebble-bed depletion analysis

Accurate modeling of running-in and equilibrium conditions in pebble-bed reactors (PBRs) requires precise microscopic multigroup neutron cross sections. In Griffin, deterministic neutronics calculations rely on multivariate interpolation over large cross section libraries, resulting in significant memory usage and performance bottlenecks. This work, together with a companion paper on Griffin integration, explores reduced-order models (ROMs) to replace interpolation with lightweight surrogates. Several ROM techniques are benchmarked, with deep neural networks (DNNs) demonstrating superior memory efficiency, scalability, and predictive accuracy. A total of 295 DNNs were trained to build a comprehensive isotope library, integrated into Griffin through a custom LibTorch interface for depletion analysis. Initial results demonstrate that DNN-based ROMs drastically reduce memory demands while preserving accuracy, enabling finer tabulations and additional state variables without overhead. In conclusion, the framework also supports online cross section generation and real-time DNN updates through transfer learning, improving fidelity by capturing self-shielding and evolving nuclide compositions during burnup.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗
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↗
DOE OSTI2024

Machine Learning‐Assisted Microearthquake Location Workflow for Monitoring the Newberry Enhanced Geothermal System

Abstract Enhanced geothermal systems (EGS) offer a sustainable energy source but face challenges in accurately locating microearthquakes induced during reservoir stimulation. Locating these microearthquakes provides reliable feedback on the stimulation progress. Current deep learning methods for locating earthquakes require extensive data sets for training, which is problematic as detected microearthquakes are often limited. To address the scarcity of training data, we propose a practical workflow using probabilistic multilayer perceptron (PMLP) which predicts microearthquake locations from cross‐correlation time lags in waveforms. Utilizing a 3D velocity model of Newberry site derived from ambient noise interferometry, we generate numerous synthetic microearthquakes and 3D acoustic waveforms for PMLP training. Accurate synthetic tests prompt us to apply the trained network to the 2012 and 2014 stimulation field waveforms. To enhance the accuracy of source localization, we carefully handpick the P‐arrival times. Predictions on the 2012 stimulation data set show major microseismic activity at depths of 0.5–1.2 km, correlating with a known casing leakage scenario. In the 2014 data set, the majority of predictions concentrate at 2.0–2.9 km depths, consistent with results obtained from conventional physics‐based inversion, and align with the presence of natural fractures from 2.0 to 2.7 km. We validate our findings by comparing the synthetic and field picks, demonstrating a satisfactory match for the first arrivals. By combining the benefits of quick inference speeds and accurate location predictions, we demonstrate the feasibility of using realistic synthetic data set to locate microseismicity for EGS monitoring.

15 GEOTHERMAL ENERGY↗
DOE OSTI2020

Decarbonizing US passenger vehicle transport under electrification and automation uncertainty has a travel budget

Abstract The transportation sector is at the beginning of a transition represented by electrification, shared mobility, and automation, which could lead to either increases or decreases in total travel and energy use. Understanding the factors enabling deep decarbonization of the passenger vehicle sector is essential for planning the required infrastructure investments and technology adoption policies. We examine the requirements for meeting carbon reduction targets of 80% and higher for passenger vehicle transport in the United States (US) by midcentury under uncertainty. We model the changes needed in vehicle electrification, electricity carbon intensity, and travel demand. Since growth in fleet penetration of electric vehicles (EVs) is constrained by fleet stock turnover, we estimate the EV penetration rates needed to meet climate targets. We find for a base case level of passenger vehicle travel, midcentury deep decarbonization of US passenger transport is conditional on reducing the electricity generation carbon intensity to close to zero along with electrification of about 67% or 84% of vehicle travel to meet decarbonization targets of 80% or 90%, respectively. Higher electricity generation carbon intensity and degraded EV fuel economy due to automation would require higher levels of fleet electrification and/or further constrain the total vehicle travel allowable. Transportation deep decarbonization not only depends on electricity decarbonization, but also has a total travel budget, representing a maximum total vehicle travel threshold that still enables meeting a midcentury climate target. This makes encouraging ride sharing, reducing total vehicle travel, and increasing fuel economy in both human-driven and future automated vehicles increasingly important to deep decarbonization.

Alarfaj, Abdullah F. (ORCID:0000000157360662)↗
DOE OSTI2020

NASA’s first ground-based Galactic Cosmic Ray Simulator: Enabling a new era in space radiobiology research

With exciting new NASA plans for a sustainable return to the moon, astronauts will once again leave Earth’s protective magnetosphere only to endure higher levels of radiation from galactic cosmic radiation (GCR) and the possibility of a large solar particle event (SPE). Gateway, lunar landers, and surface habitats will be designed to protect crew against SPEs with vehicle optimization, storm shelter concepts, and/or active dosimetry; however, the ever penetrating GCR will continue to pose the most significant health risks especially as lunar missions increase in duration and as NASA sets its aspirations on Mars. The primary risks of concern include carcinogenesis, central nervous system (CNS) effects resulting in potential in-mission cognitive or behavioral impairment and/or late neurological disorders, degenerative tissue effects including circulatory and heart disease, as well as potential immune system decrements impacting multiple aspects of crew health. Characterization and mitigation of these risks requires a significant reduction in the large biological uncertainties of chronic (low-dose rate) heavy-ion exposures and the validation of countermeasures in a relevant space environment. Historically, most research on understanding space radiation-induced health risks has been performed using acute exposures of monoenergetic single-ion beams. However, the space radiation environment consists of a wide variety of ion species over a broad energy range. Using the fast beam switching and controls systems technology recently developed at the NASA Space Radiation Laboratory (NSRL) at Brookhaven National Laboratory, a new era in radiobiological research is possible. NASA has developed the “GCR Simulator” to generate a spectrum of ion beams that approximates the primary and secondary GCR field experienced at human organ locations within a deep-space vehicle. The majority of the dose is delivered from protons (approximately 65%–75%) and helium ions (approximately 10%–20%) with heavier ions (Z ≥ 3) contributing the remainder. The GCR simulator exposes state-of-the art cellular and animal model systems to 33 sequential beams including 4 proton energies plus degrader, 4 helium energies plus degrader, and the 5 heavy ions of C, O, Si, Ti, and Fe. A polyethylene degrader system is used with the 100 MeV/n H and He beams to provide a nearly continuous distribution of low-energy particles. A 500 mGy exposure, delivering doses from each of the 33 beams, requires approximately 75 minutes. To more closely simulate the low-dose rates found in space, sequential field exposures can be divided into daily fractions over 2 to 6 weeks, with individual beam fractions as low as 0.1 to 0.2 mGy. In the large beam configuration (60 × 60 cm 2 ), 54 special housing cages can accommodate 2 to 3 mice each for an approximately 75 min duration or 15 individually housed rats. On June 15, 2018, the NSRL made a significant achievement by completing the first operational run using the new GCR simulator. This paper discusses NASA’s innovative technology solution for a ground-based GCR simulator at the NSRL to accelerate our understanding and mitigation of health risks faced by astronauts. Ultimately, the GCR simulator will require validation across multiple radiogenic risks, endpoints, doses, and dose rates.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Deep learning-assisted modeling for χ (2) nonlinear optics

Modeling second-order (χ(2)) nonlinear optical processes remains computationally expensive due to the need to resolve fast field oscillations and simulate wave propagation using methods such as the split-step Fourier method (SSFM). This can become a bottleneck in real-time applications, such as high-repetition-rate laser systems requiring rapid feedback and control. We present a long short-term memory-based surrogate model trained on SSFM simulations generated from a start-to-end model of the photocathode drive laser at SLAC National Accelerator Laboratory’s Linac Coherent Light Source II. The model achieves over 250× speedup while maintaining high fidelity, enabling future real-time optimization and laying the foundation for data-integrated modeling frameworks and digital twins of laser systems.

Accelerator Physics (physics.acc-ph)↗
DOE OSTI2022

Understanding and Leveraging the I/O Patterns of Emerging Machine Learning Analytics

The scientific community is currently experiencing unprecedented amounts of data generated by cutting-edge science facilities. Soon facilities will be producing up to 1 PB/s which will force scientist to use more autonomous techniques to learn from the data. The adoption of machine learning methods, like deep learning techniques, in large-scale workflows comes with a shift in the workflow’s computational and I/O patterns. These changes often include iterative processes and model architecture searches, in which datasets are analyzed multiple times in different formats with different model configurations in order to find accurate, reliable and efficient learning models. This shift in behavior brings changes in I/O patterns at the application level as well at the system level. These changes also bring new challenges for the HPC I/O teams, since these patterns contain more complex I/O workloads. In this paper we discuss the I/O patterns experienced by emerging analytical codes that rely on machine learning algorithms and highlight the challenges in designing efficient I/O transfers for such workflows. We comment on how to leverage the data access patterns in order to fetch in a more efficient way the required input data in the format and order given by the needs of the application and how to optimize the data path between collaborative processes. We will motivate our work and show performance gains with a study case of medical applications.

Gainaru, Ana↗
DOE OSTI2026

Machine learning for seismic low-frequency extrapolation

The cycle-skipping problem that plagues full waveform inversion (FWI) can be at least partially mitigated if low frequencies (which encode the kinematics of wave propagation in seismic data) are recorded. However, seismic sources and receivers are band-limited, so seismic data does not generally include signals down to 0 Hz. To improve our ability to solve the seismic inverse problem, one can synthesize this missing low-frequency (LF) content from the recorded high-frequency (HF) data using machine learning (ML) models. Deep learning models such as convolutional neural networks (CNNs) demonstrate impressive ability to perform low frequency extrapolation. However, such models require powerful hardware (GPU machines) and careful training. We assess the extrapolation capabilities of three different ML models that do not require GPU machines, namely, random forest, Gaussian process regression and gradient boosting, on both synthetic and real data. Experimental results on two synthetic data sets (generated from a low velocity lens embedded in a homogeneous medium, and the Marmousi model) demonstrate that FWI applied to the extrapolated data consistently improves inversion accuracy relative to FWI applied to the original data sets that do not contain low frequencies. Application of low-frequency extrapolation to real data from the Northwest Shelf of Australia demonstrates that tree-based ML models such as gradient boosting can outperform CNNs in terms of both accuracy and computational cost on non-GPU architectures.

58 GEOSCIENCES↗
DOE OSTI2024

The expressivity of classical and quantum neural networks on entanglement entropy

Abstract Analytically continuing the von Neumann entropy from Rényi entropies is a challenging task in quantum field theory. While then-th Rényi entropy can be computed using the replica method in the path integral representation of quantum field theory, the analytic continuation can only be achieved for some simple systems on a case-by-case basis. In this work, we propose a general framework to tackle this problem using classical and quantum neural networks with supervised learning. We begin by studying several examples with known von Neumann entropy, where the input data is generated by representing$${\text {Tr}}\rho _A^n$$ Tr ρ A n with a generating function. We adopt KerasTuner to determine the optimal network architecture and hyperparameters with limited data. In addition, we frame a similar problem in terms of quantum machine learning models, where the expressivity of the quantum models for the entanglement entropy as a partial Fourier series is established. Our proposed methods can accurately predict the von Neumann and Rényi entropies numerically, highlighting the potential of deep learning techniques for solving problems in quantum information theory.

Physics↗
DOE OSTI2024

Assessing and benchmarking the fidelity of posterior inference methods for astrophysics data analysis

In this era of large and complex astronomical survey data, interpreting, validating, and comparing inference techniques becomes increasingly difficult. This is particularly critical for emerging inference methods like Simulation-Based Inference (SBI), which offer significant speedup potential and posterior modeling flexibility, especially when deep learning is incorporated. We present a study to assess and compare the performance and uncertainty prediction capability of Bayesian inference algorithms – from traditional MCMC sampling of analytic functions to deep learning-enabled SBI. We focus on testing the capacity of hierarchical inference modeling in those scenarios. Before we extend this study to cosmology, we first use astrophysical simulation data to ensure interpretability. We demonstrate a probabilistic programming implementation of hierarchical and non-hierarchical Bayesian inference using simulations derived from the DeepBench software library, a benchmarking tool developed by our group that generates simple and controllable astrophysical objects from first principles. This study will enable astronomers and physicists to harness the inference potential of these methods with confidence.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Power Production, Inter- and Intra-Array Wake Losses from the U.S. East Coast Offshore Wind Energy Lease Areas

There is an urgent need to develop accurate predictions of power production, wake losses and array–array interactions from multi-GW offshore wind farms in order to enable developments that maximize power benefits, minimize levelized cost of energy and reduce investment uncertainty. New, climatologically representative simulations with the Weather Research and Forecasting (WRF) model are presented and analyzed to address these research needs with a specific focus on offshore wind energy lease areas along the U.S. east coast. These, uniquely detailed, simulations are designed to quantify important sources of wake-loss projection uncertainty. They sample across different wind turbine deployment scenarios and thus span the range of plausible installed capacity densities (ICDs) and also include two wind farm parameterizations (WFPs; Fitch and explicit wake parameterization (EWP)) and consider the precise WRF model release used. System-wide mean capacity factors for ICDs of 3.5 to 6.0 MWkm−2 range from 39 to 45% based on output from Fitch and 50 to 55% from EWP. Wake losses are 27–37% (Fitch) and 11–19% (EWP). The discrepancy in CF and wake losses from the two WFPs derives from two linked effects. First, EWP generates a weaker ‘deep array effect’ within the largest wind farm cluster (area of 3675 km2), though both parameterizations indicate substantial within-array wake losses. If 15 MW wind turbines are deployed at an ICD of 6 MWkm−2 the most heavily waked wind turbines generate an average of only 32–35% of the power of those that experience the freestream (undisturbed) flow. Nevertheless, there is no evidence for saturation of the resource. The wind power density (electrical power generation per unit of surface area) increases with ICD and lies between 2 and 3 Wm−2. Second, EWP also systematically generates smaller whole wind farm wakes. Sampling across all offshore wind energy lease areas and the range of ICD considered, the whole wind farm wake extent for a velocity deficit of 5% is 1.18 to 1.38 times larger in simulations with Fitch. Over three-quarters of the variability in normalized wake extents is attributable to variations in freestream wind speeds, turbulent kinetic energy and boundary layer depth. These dependencies on meteorological parameters allow for the development of computationally efficient emulators of wake extents from Fitch and EWP.

17 WIND ENERGY↗
DOE OSTI2023

Informed Feature Selection for Data Clustering of CSP Plant Production

To make concentrating solar power (CSP) more cost competitive, rigourous optimizations must be run to improve plant design and operations. However, these optimizaitons rely on time consuming annual simulations that solve an electricity dispatch scheduling problem to maximize plant revenue. To reduce the runtime of annual dispatch simulations of CSP plants, a data clustering approach is utilized. This approach assumes that like days of revenue and electricity generation can be identified using weather and price data. Although weather and price are important factors for electricity production, this work investigates how thermal energy storage (TES) inventory at the beginning of a day, denoted as Si, can be used as a supplemental feature to group like days. A framework for creating and training a deep neural network to predict Si is proposed. This model is validated and assessed using eleven sets of testing data that were not used during training. Then, the data clustering approach is performed three seperate times with features of weather and price along with either Si from the neural network, Si from the full annual simulation, or no Si. Ultimately, the results suggest that using Si as an additional clustering feature improves the data clustering simulation accuracy by 1.4%.

Tuman, Matthew J. (ORCID:000900038772051X)↗
DOE OSTI2023

Using Ultrasound Image Augmentation and Ensemble Predictions to Prevent Machine-Learning Model Overfitting

Deep learning predictive models have the potential to simplify and automate medical imaging diagnostics by lowering the skill threshold for image interpretation. However, this requires predictive models that are generalized to handle subject variability as seen clinically. Here, we highlight methods to improve test accuracy of an image classifier model for shrapnel identification using tissue phantom image sets. Using a previously developed image classifier neural network—termed ShrapML—blind test accuracy was less than 70% and was variable depending on the training/test data setup, as determined by a leave one subject out (LOSO) holdout methodology. Introduction of affine transformations for image augmentation or MixUp methodologies to generate additional training sets improved model performance and overall accuracy improved to 75%. Further improvements were made by aggregating predictions across five LOSO holdouts. This was done by bagging confidences or predictions from all LOSOs or the top-3 LOSO confidence models for each image prediction. Top-3 LOSO confidence bagging performed best, with test accuracy improved to greater than 85% accuracy for two different blind tissue phantoms. This was confirmed by gradient-weighted class activation mapping to highlight that the image classifier was tracking shrapnel in the image sets. Overall, data augmentation and ensemble prediction approaches were suitable for creating more generalized predictive models for ultrasound image analysis, a critical step for real-time diagnostic deployment.

60 APPLIED LIFE SCIENCES↗