SEARCH · Engineering Papers

Results for “learning dynamics”

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 1,621 records · Page 90

All sources NASA NTRS DOE OSTI

DOE OSTI2026

Differentiable multiphase flow model for physics-informed machine learning in reservoir pressure management

Accurate subsurface reservoir pressure control is extremely challenging due to geological heterogeneity and multiphase fluid-flow dynamics. Predicting behavior in this setting relies on high-fidelity physics-based simulations that are computationally expensive. Yet, the uncertain, heterogeneous properties that control these flows make it necessary to perform many of these expensive simulations, which is often prohibitive. To address these challenges, we introduce a physics-informed machine learning workflow that couples a fully differentiable multiphase flow simulator, which is implemented in the DPFEHM framework with a convolutional neural network (CNN). The CNN learns to predict fluid extraction rates from heterogeneous permeability fields to enforce pressure limits at critical reservoir locations. By incorporating transient multiphase flow physics into the training process, our method enables more practical and accurate predictions for realistic injection-extraction scenarios compared to previous works. To speed up training, we pretrain the model on single-phase, steady-state simulations and then finetune it on full multiphase scenarios, which dramatically reduces the computational cost. We demonstrate that high-accuracy training can be achieved with fewer than three thousand full-physics multiphase flow simulations – compared to previous estimates requiring up to ten million. This drastic reduction in the number of simulations is achieved by leveraging transfer learning from much less expensive single phase simulations.

25 ENERGY STORAGE↗

DOE OSTI2021

Data-driven Distributed Learning of Multi-agent Systems: A Koopman Operator Approach

Koopman operator theory provides a model-free technique for studying nonlinear dynamical systems purely from data. Since the Koopman operator is infinite-dimensional, researchers have developed several methods that provide a finite-dimensional approximation of the Koopman operator so that it can be applied for practical use cases. One common thing with most of the methods is that their solutions are obtained by solving a centralized minimization problem. In this work, we treat the dynamical system to be a multi-agent system and propose an algorithm to compute the finite-dimensional approximation of the Koopman operator in a distributed manner using the knowledge of the topology of the underlying multi-agent system. The proposed distributed approach is shown to be equivalent to the centralized learning problem and results in a sparse Koopman whose block structure mimics the Laplacian of the multi-agent system. Extensive simulation studies illustrate the proposed framework on the network of oscillators and the IEEE 68 bus system.

Nandanoori, Sai Pushpak↗

DOE OSTI2024

Evaluating a Commercial Dynamic Line Rating Software with the National PMU Dataset

To accelerate the development of data-driven applications for power systems, the Department of Energy (DOE) supported the collection and curation of a synchrophasor dataset spanning two years of observations from transmission utilities across the US. This National PMU Dataset (NPDS) was anonymized and distributed to awardees of a DOE research grant under nondisclosure agreements (NDAs) but has also been retained at PNNL to enable further research. Agreements with data contributors prevent the data from being shared outside the organization. However, establishing a blind research validation methodology is envisioned to maximize the value proposition of the NPDS. In this validation strategy, researchers may share algorithms/software (potentially as executables to protect intellectual property) with PNNL, and PNNL will share feedback about the software’s performance on subsets of the NPDS. Such a blind methodology ensures that sensitive information about critical infrastructure remains protected, but the value of the NPDS can be extended to research beyond PNNL. Through iterative feedback, the algorithms may be tweaked to address real-world artifacts. As the NPDS data is temporally and geographically diverse, it may capture features absent in smaller datasets used during the development of the algorithm under test. This report presents lessons learned from applying the blind validation methodology to LineID™, a synchrophasor-based dynamic line rating software developed by Topolonet Corporation. Improvements made to the software through iterative feedback, limitations of the validation methodology, as well as how the limitations of the NPDS affected the evaluation process are discussed. Observations indicate that the proposed validation methodology can be valuable for evaluating other tools in the future.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

A machine learning study on spinodal clumping in heavy ion collisions

Possible observables of baryon number clustering due to the instabilities occurring at a first order QCD phase transition are discussed. The dynamical formation of baryon clusters at a QCD phase transition can be described by numerical fluid dynamics, augmented with a gradient term and an equation of state with a mechanically unstable region. It is shown that the dynamical description of this phase transition, in nuclear collisions, will lead to the formation of dense baryon clusters at the phase boundary. State-of-the-art machine learning methods find that the coordinate space clumping leaves characteristic imprints on the spatial net density distribution in almost every event. On the other hand the momentum distributions do not show any clear event-by-event features. Lastly, it is shown that the 'third order' cumulant, the skewness, shows a peak at the beam energy where the system, created in the heavy ion collision, reaches the deconfinement phase transition.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2023

Data-driven analysis and prediction of stable phases for high-entropy alloy design

High-entropy alloys (HEAs) represent a promising class of materials with exceptional structural and functional properties. However, their design and optimization pose challenges due to the large composition-phase space coupled with the complex and diverse nature of the phase formation dynamics. In this study, a data-driven approach that utilizes machine learning (ML) techniques to predict HEA phases and their composition-dependent phases is proposed. By employing a comprehensive dataset comprising 5692 experimental records encompassing 50 elements and 11 phase categories, we compare the performance of various ML models. Our analysis identifies the most influential features for accurate phase prediction. Furthermore, the class imbalance is addressed by employing data augmentation methods, raising the number of records to 1500 in each category, and ensuring a balanced representation of phase categories. The results show that XGBoost and Random Forest consistently outperform the other models, achieving 86% accuracy in predicting all phases. Additionally, this work provides an extensive analysis of HEA phase formers, showing the contributions of elements and features to the presence of specific phases. We also examine the impact of including different phases on ML model accuracy and feature significance. Notably, the findings underscore the need for ML model selection based on specific applications and desired predictions, as feature importance varies across models and phases. This study significantly advances the understanding of HEA phase formation, enabling targeted alloy design and fostering progress in the field of materials science.

36 MATERIALS SCIENCE↗

DOE OSTI2025

Latent space mapping: Revolutionizing predictive models for divertor plasma detachment control

The inherent complexity of boundary plasma, characterized by multi-scale and multi-physics challenges, has historically restricted high-fidelity simulations to scientific research due to their intensive computational demands. Consequently, routine applications such as discharge control and scenario development have relied on faster but less accurate empirical methods. This work introduces DivControlNN, a novel machine-learning-based surrogate model designed to address these limitations by enabling quasi-real-time predictions (i.e., ~ 0.2 ms) of boundary and divertor plasma behavior. Trained on over 70,000 2D UEDGE simulations from KSTAR tokamak equilibria, DivControlNN employs latent space mapping to efficiently represent complex divertor plasma states, achieving a computational speed-up of over 10 8 compared to traditional simulations while maintaining a relative error below 20% for key plasma property predictions. During the 2024 KSTAR experimental campaign, a prototype detachment control system powered by DivControlNN successfully demonstrated detachment control on its first attempt, even for a new tungsten divertor configuration and without any fine-tuning. These results highlight the transformative potential of DivControlNN in overcoming diagnostic challenges in future fusion reactors by providing fast, robust, and reliable predictions for advanced integrated control systems.

Artificial neural networks↗

DOE OSTI2022

Multiplicative Shot-Noise: A New Route to Stability of Plastic Networks

Fluctuations of synaptic weights, among many other physical, biological, and ecological quantities, are driven by coincident events of two “parent” processes. Here we propose a multiplicative shot-noise model that can capture the behaviors of a broad range of such natural phenomena, and analytically derive an approximation that accurately predicts its statistics. We apply our results to study the effects of a multiplicative synaptic plasticity rule that was recently extracted from measurements in physiological conditions. Using mean-field theory analysis and network simulations, we investigate how this rule shapes the connectivity and dynamics of recurrent spiking neural networks. The multiplicative plasticity rule is shown to support efficient learning of input stimuli, and it gives a stable, unimodal synaptic-weight distribution with a large fraction of strong synapses. The strong synapses remain stable over long times but do not “run away.” Our results suggest that the multiplicative shot-noise offers a new route to understand the tradeoff between flexibility and stability in neural circuits and other dynamic networks.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2019

MODELING SHOCK AND VIBRATION ON USED NUCLEAR FUEL DURING NORMAL CONDITIONS OF TRANSPORTATION

In 2017, the United States Department of Energy (DOE) collaborated with Spanish and Korean organizations to perform a multimodal transportation test to measure shock and vibration loads imparted to used nuclear fuel (UNF) assemblies. This test used real fuel assembly components containing surrogate fuel mass to approximate the response characteristics of real, irradiated used nuclear fuel. Pacific Northwest National Laboratory was part of the test team and used the data collected during this test to validate numerical models needed to predict the response of real used nuclear fuel in other transportation configurations. This paper summarizes the modeling work and identifies lessons learned related to the modeling and analysis methodology. The modeling includes railcar dynamics using the NUCARS software code and explicit dynamic finite element modeling of used nuclear fuel cladding in LS-DYNA. The NUCARS models were validated against railcar dynamics data collected during captive track testing at the Federal Railroad Administration's Transportation Technology Center in Pueblo, CO. The LS-DYNA models of the fuel cladding were validated against strain gage data collected throughout the test campaign. One of the key results of this work was an assessment of fuel cladding fatigue, and the methods used to calculate fatigue are detailed in this paper. The validated models and analysis methodologies described in this paper will be applied to evaluate future UNF transportation systems.

Klymyshyn, Nicholas A.↗

DOE OSTI2023

Multiobjective Hyperparameter Optimization for Deep Learning Interatomic Potential Training Using NSGA-II

Deep neural network (DNN) potentials are an emerging tool for simulation of dynamical atomistic systems, with the promise of quantum mechanical accuracy at speedups of 10000$\times$. As with other DNN methods, hyperparameters used during training can make a substantial difference in model accuracy, and optimal settings vary with dataset. To enable rapid tuning of hyperparameters for DNN potential training, we developed a scalable multiobjective optimization evolutionary algorithm for supercomputers and tested it on the Summit system at the Oak Ridge Leadership Computing Facility (OLCF). The multiobjective approach is required due to the coupling of two learned values defining the potential: the energy and force. Using a large-scale implementation of the NSGA-II algorithm adapted for training DNN potentials, we discovered several optimal multiobjective combinations, including best choices of activation functions, learning rate scaling scheme, and pairing of the two radial cutoffs used in the three dimensional descriptor function.

Coletti, Mark↗

DOE OSTI2026

ShenCFD

ShenCFD is a fast pseudospectral solver for fluid dynamics written to be maximally Pythonic and maximally useful for machine-learning-based turbulence model discovery.

Saenz, Juan↗

DOE OSTI2017

Power Excursion Mitigation for Flexgrid Defragmentation With Machine Learning

Flexgrid optical networking relies on spectrum defragmentation to groom channel wavelengths and improve spectral efficiency. Various defragmentation methods—hop, make-before-break, and sweep—interact with the power dynamics of erbium-doped fiber amplifiers (EDFA) differently and result in undesired power excursions that exacerbate the post-EDFA power variance. Here we present a machine learning engine that characterizes the channel dependence of power excursions from historical data. We further demonstrate that post-EDFA power variance during hop, make-before-break, and sweep defragmentation methods can be greatly mitigated by the trained machine learning engine with automated and expedited power adjustment predictions.

42 ENGINEERING↗

DOE OSTI2024

Towards a data-driven model of hadronization using normalizing flows

We introduce a model of hadronization based on invertible neural networks that faithfully reproduces a simplified version of the Lund string model for meson hadronization. Additionally, we introduce a new training method for normalizing flows, termed MAGIC, that improves the agreement between simulated and experimental distributions of high-level (macroscopic) observables by adjusting single-emission (microscopic) dynamics. Our results constitute an important step toward realizing a machine-learning based model of hadronization that utilizes experimental data during training. Finally, we demonstrate how a Bayesian extension to this normalizing-flow architecture can be used to provide analysis of statistical and modeling uncertainties on the generated observable distributions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2022

Nucleon and pion PDFs: large-x asymptotics meets functional mimicry

We review strategies to unveil the primordial large- x x structure of the nucleons as well as the pion from hard-scattering experiments. Ideas are presented for learning about the x\to 1 x → 1 limit of nonperturbative QCD dynamics at energy scales of order 1 GeV from collider experiments at much higher scales. The behavior of parton distributions at x\to 1 x → 1 predicted by the quark counting rules and other low-energy theoretical approaches is contrasted with phenomenological PDFs. Polynomial mimicry of PDF parametrizations is one of many factors that influence the apparent power of the (1-x) ( 1 − x ) falloff. We discuss implications of the mimicry for the large- x x falloff of the pion PDFs.

Courtoy, Aurore↗

DOE OSTI2021

Simulation of Electron-Proton Scattering Events by a Feature-Augmented and Transformed Generative Adversarial Network (FAT-GAN)

We apply generative adversarial network (GAN) technology to build an event generator that simulates particle production in electron-proton scattering that is free of theoretical assumptions about underlying particle dynamics. The difficulty of efficiently training a GAN event simulator lies in learning the complicated patterns of the distributions of the particles physical properties. We develop a GAN that selects a set of transformed features from particle momenta that can be generated easily by the generator, and uses these to produce a set of augmented features that improve the sensitivity of the discriminator. The new Feature-Augmented and Transformed GAN (FAT-GAN) is able to faithfully reproduce the distribution of final state electron momenta in inclusive electron scattering, without the need for input derived from domain-based theoretical assumptions. The developed technology can play a significant role in boosting the science of existing and future accelerator facilities, such as the Electron-Ion Collider.

Alanazi, Yasir↗

DOE OSTI2024

Design, Selection, and Evaluation of Reinforcement Learning Single Agents for Ground Target Tracking

Previous approaches for small fixed-wing unmanned air systems that carry strapdown rather than gimbaled cameras achieved satisfactory ground target tracking performance using both standard and deep reinforcement learning algorithms. However, these approaches have significant restrictions and abstractions to the dynamics of the vehicle, such as constant airspeed and constant altitude, because the number of states and actions was necessarily limited. Thus, extensive tuning was required to obtain good tracking performance. The expansion from 4 state–action degrees of freedom to 15 enabled the agent to exploit previous reward functions that produced novel yet undesirable emergent behavior. This paper investigates the causes of and various potential solutions to undesirable emergent behavior in the ground target tracking problem. A combination of changes to the environment, reward structure, action space simplification, command rate, and controller implementation provides insight into obtaining stable tracking results. Consideration is given to reward structure selection and refinement to mitigate undesirable emergent behavior. Results presented in the paper for a simulated environment of a single unmanned air system tracking a randomly moving single ground target show that a soft actor–critic algorithm can produce feasible tracking trajectories without limiting the state space and action space, provided that the environment is properly posed.

Engineering↗

DOE OSTI2024

Regional climate change: consensus, discrepancies, and ways forward

Climate change has emerged across many regions. Some observed regional climate changes, such as amplified Arctic warming and land-sea warming contrasts have been predicted by climate models. However, many other observed regional changes, such as changes in tropical sea surface temperature and monsoon rainfall are not well simulated by climate model ensembles even when taking into account natural internal variability and structural uncertainties in the response of models to anthropogenic radiative forcing. This suggests climate model predictions may not fully reflect what our future will look like. The discrepancies between models and observations are not well understood due to several real and apparent puzzles and limitations such as the “signal-to-noise paradox” and real-world record-shattering extremes falling outside of the possible range predicted by models. Addressing these discrepancies, puzzles and limitations is essential, because understanding and reliably predicting regional climate change is necessary in order to communicate effectively about the underlying drivers of change, provide reliable information to stakeholders, enable societies to adapt, and increase resilience and reduce vulnerability. The challenges of achieving this are greater in the Global South, especially because of the lack of observational data over long time periods and a lack of scientific focus on Global South climate change. To address discrepancies between observations and models, it is important to prioritize resources for understanding regional climate predictions and analyzing where and why models and observations disagree via testing hypotheses of drivers of biases using observations and models. Gaps in understanding can be discovered and filled by exploiting new tools, such as artificial intelligence/machine learning, high-resolution models, new modeling experiments in the model hierarchy, better quantification of forcing, and new observations. Conscious efforts are needed toward creating opportunities that allow regional experts, particularly those from the Global South, to take the lead in regional climate research. This includes co-learning in technical aspects of analyzing simulations and in the physics and dynamics of regional climate change. Finally, improved methods of regional climate communication are needed, which account for the underlying uncertainties, in order to provide reliable and actionable information to stakeholders and the media.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2022

The Evolution of Volatile Memory Forensics

The collection and analysis of volatile memory is a vibrant area of research in the cybersecurity community. The ever-evolving and growing threat landscape is trending towards fileless malware, which avoids traditional detection but can be found by examining a system’s random access memory (RAM). Additionally, volatile memory analysis offers great insight into other malicious vectors. It contains fragments of encrypted files’ contents, as well as lists of running processes, imported modules, and network connections, all of which are difficult or impossible to extract from the file system. For these compelling reasons, recent research efforts have focused on the collection of memory snapshots and methods to analyze them for the presence of malware. However, to the best of our knowledge, no current reviews or surveys exist that systematize the research on both memory acquisition and analysis. We fill that gap with this novel survey by exploring the state-of-the-art tools and techniques for volatile memory acquisition and analysis for malware identification. For memory acquisition methods, we explore the trade-offs many techniques make between snapshot quality, performance overhead, and security. For memory analysis, we examined the traditional forensic methods used, including signature-based methods, dynamic methods performed in a sandbox environment, as well as machine learning-based approaches. We summarize the currently available tools, and suggest areas for more research.

Nyholm, Hannah↗

DOE OSTI2025

Dataset, Code, and Models for Training Deep Learning Potentials for Low Temperature Plasma-Surface Interactions

This repository contains datasets, training scripts, and finished models, and test simulations used in the development of DeepREBO— a machine-learned interatomic potential trained to emulate the REBO2 empirical potential. The data was generated to study deep potential development for simulations of plasma-surface interactions. It uses an active learning framework, starting from a minimal dataset and iteratively expanding it. Included are those generated datasets, the trained models, and simulations used to evaluate the performance of the training process. This resource supports reproducibility and provides a reference framework for training deep potentials in plasma-surface interaction studies.

active learning↗