Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Distributed training”

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 433 records · Page 24
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Data efficiency and extrapolation trends in neural network interatomic potentials

Abstract Recently, key architectural advances have been proposed for neural network interatomic potentials (NNIPs), such as incorporating message-passing networks, equivariance, or many-body expansion terms. Although modern NNIP models exhibit small differences in test accuracy, this metric is still considered the main target when developing new NNIP architectures. In this work, we show how architectural and optimization choices influence the generalization of NNIPs, revealing trends in molecular dynamics (MD) stability, data efficiency, and loss landscapes. Using the 3BPA dataset, we uncover trends in NNIP errors and robustness to noise, showing these metrics are insufficient to predict MD stability in the high-accuracy regime. With a large-scale study on NequIP, MACE, and their optimizers, we show that our metric of loss entropy predicts out-of-distribution error and data efficiency despite being computed only on the training set. This work provides a deep learning justification for probing extrapolation and can inform the development of next-generation NNIPs.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Efficient quantum state tomography with mode-assisted training

Neural networks (NNs) representing quantum states are typically trained using Markov chain Monte Carlo based methods. However, unless specifically designed, such samplers only consist of local moves, making the slow-mixing problem prominent even for extremely simple quantum states. Here, we propose to use mode-assisted training that provides global information via the modes of the NN distribution. Further, applied to quantum state tomography using restricted Boltzmann machines, this method improves the quality of reconstructed quantum states by orders of magnitude. The method is applicable to other types of NNs and may efficiently tackle problems previously unmanageable.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

Probabilistic neural networks for fluid flow surrogate modeling and data recovery

We consider the use of probabilistic neural networks for fluid flow surrogate modeling and data recovery. This framework is constructed by assuming that the target variables are sampled from a Gaussian distribution conditioned on the inputs. Consequently, the overall formulation sets up a procedure to predict the hyperparameters of this distribution which are then used to compute an objective function given training data. We demonstrate that this framework has the ability to provide for prediction confidence intervals based on the assumption of a probabilistic posterior, given an appropriate model architecture and adequate training data. The applicability of the present framework to cases with noisy measurements and limited observations is also assessed. To demonstrate the capabilities of this framework, we consider canonical regression problems of fluid dynamics from the viewpoint of reduced-order modeling and spatial data recovery for four canonical data sets. The examples considered in this study arise from (i) the shallow-water equations, (ii) a two-dimensional cylinder flow, (iii) the wake of a NACA0012 airfoil with a Gurney flap, and (iv) the NOAA sea surface temperature data set. Furthermore, the present results indicate that the probabilistic neural network not only produces a machine-learning-based fluid flow surrogate model but also systematically quantifies the uncertainty therein to assist with model interpretability.

42 ENGINEERING↗
DOE OSTI2022

Skopi: a simulation package for diffractive imaging of noncrystalline biomolecules

X-ray free-electron lasers (XFELs) have the ability to produce ultra-bright femtosecond X-ray pulses for coherent diffraction imaging of biomolecules. While the development of methods and algorithms for macromolecular crystallography is now mature, XFEL experiments involving aerosolized or solvated biomolecular samples offer new challenges in terms of both experimental design and data processing. Skopi is a simulation package that can generate single-hit diffraction images for reconstruction algorithms, multi-hit diffraction images of aggregated particles for training machine learning classifiers using labeled data, diffraction images of randomly distributed particles for fluctuation X-ray scattering algorithms, and diffraction images of reference and target particles for holographic reconstruction algorithms. Skopi is a resource to aid feasibility studies and advance the development of algorithms for noncrystalline experiments at XFEL facilities.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

WaveDenoiser

We developed a robust deep learning model to effectively reduce background noise in the time domain from seismic waveforms. The deep learning model processes a 57-second three-component seismogram to predict and generate a denoised seismogram. The training was conducted using the benchmark STEAD dataset, which comprises globally distributed earthquake signals recorded at local distances ranging from 0 to 350 kilometers.

Chai, Chengping [Oak Ridge National Laboratory (OR↗
DOE OSTI2026

WaveDP

We developed a robust deep learning model designed to effectively reduce background noise and measure signal arrival times from seismic waveforms. This model processes a 57-second, three-component seismogram to predict the probability of Primary (P) and Secondary (S) waves for each timestamp, while also generating a denoised seismogram. Training was conducted using the benchmark STEAD dataset, which includes globally distributed earthquake signals recorded at local distances ranging from 0 to 350 kilometers.

Chai, Chengping [Oak Ridge National Laboratory (OR↗
DOE OSTI2022

Computational microscopy for fast widefield deep-tissue fluorescence imaging using a commercial dual-cannula probe

A solid-glass cannula serves as a micro-endoscope that can deliver excitation light deep inside tissue while also collecting emitted fluorescence. Then, we utilize deep neural networks to reconstruct images from the collected intensity distributions. By using a commercially available dual-cannula probe, and training a separate deep neural network for each cannula, we effectively double the field of view compared to prior work. We demonstrated ex vivo imaging of fluorescent beads and brain slices and in vivo imaging from whole brains. We clearly resolved 4 µm beads, with FOV from each cannula of 0.2 mm (diameter), and produced images from a depth of ∼1.2 mm in the whole brain, currently limited primarily by the labeling. Since no scanning is required, fast widefield fluorescence imaging limited primarily by the brightness of the fluorophores, collection efficiency of our system, and the frame rate of the camera becomes possible.

42 ENGINEERING↗
DOE OSTI2020

Adaptive Machine Learning for Bragg Coherent Diffraction Imaging (BCDI) of 3D Electron Density Maps with Application to La 2-x Ba x CuO 4 (LBCO) High Temperature Superconductor Studies [PowerPoint]

Understanding mesoscale heterogeneity is important for evaluating fatigue and failure of structural materials and for manufacturing processes. Data-driven tools can facilitate in guiding effort investment (measurements and computations) at various stages of the materials development. Our 3D reconstruction approach is to find a set of coefficients which define a bounding surface of the electron density. Data for training the 3D CNN was generated by sampling coefficients from uniform distributions. Test vs. prediction values are shown for the 28 even-valued coefficients for 1000 test structures, with the best and worst performers highlighted. The adaptive part of this work utilized a model independent extremum seeking (ES). The predictions of the CNN were used as the starting point for the ES algorithm. The convergence of the ES algorithm for 3 different structures is illustrated, and the robustness of the adaptive ML approach was demonstrated on experimentally measured 3D crystal from high energy diffraction microscopy. Reconstruction of a non-uniform density volume with different levels of Poisson noise is also shown. Preliminary attempts at reconstructing measured La 2-x Ba x CuO 4 diffraction patterns have so far failed.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
NASA NTRS1978

Summary report of the Turbulence Committee

The effects of atmospheric turbulence are discussed. Aircraft design and aircraft operations are reviewed. Turbulence in terms of intensity and scale in design considerations was examined. Turbulence models were used in the form of discrete gusts, spectral distributions, and probability distributions. Various aspects of the design and operations problems, simulation and training factors of pilots, and weather services and forecasts are reported.

Elderkin, C. E.↗
NASA NTRS1979

When day is done and shadows fall, we miss the airport most of all

Both the effectiveness of pilot training and the safety of flight can be influenced by the distribution of texture in the visual scene, the distance to which the eyes accommodate, and the associated shifts in the apparent size and distance of objects in central and peripheral vision. Studies reviewed and original results presented indicate that these factors are involved in various misjudgments and illusions experienced by pilots: (1) when searching for other airborne traffic or targets, (2) when making approaches to airports over water at night, (3) when breaking out of low clouds on a final approach to a landing by reference to head-up or head-down displays, and (4) when practicing simulated approaches and landings or air-to-surface weapon deliveries by reference to synthetically generated visual systems.

Roscoe, S. N.↗
NASA NTRS1981

Wind energy system time-domain (WEST) analyzers

A portable analyzer which simulates in real time the complex nonlinear dynamics of horizontal axis wind energy systems was constructed. Math models for an aeroelastic rotor featuring nonlinear aerodynamic and inertial terms were implemented with high speed digital controllers and analog calculation. This model was combined with other math models of elastic supports, control systems, a power train and gimballed rotor kinematics. A stroboscopic display system graphically depicting distributed blade loads, motion, and other aerodynamic functions on a cathode ray tube is included. Limited correlation efforts showed good comparison between the results of this analyzer and other sophisticated digital simulations. The digital simulation results were successfully correlated with test data.

Dreier, M. E.↗
NASA NTRS1990

Flight elements subpanel

A summary of findings of the Flight Elements Subpanel is given in viewgraph form. Topics covered include advanced avionics architecture; processors; integrated Global Positioning System and guidance, navigation and control; displays and controls; telemetry; sensors and instrumentation; fault detection and management; electrical power distribution and control; spacecraft power supplies; and in-flight crew training.

Keckler, Claude R.↗
NASA NTRS2017

Beyond Control Centers

For NASA's Resource Prospector (RP) Lunar Rover Mission, we are moving away from a control center concept, to a fully distributed operation utilizing control nodes, with decision support from anywhere via mobile devices. This operations concept will utilize distributed information systems, notifications, mobile data access, and optimized mobile data display for off-console decision support. We see this concept of operations as a step in the evolution of mission operations from a central control center concept to a mission operations anywhere concept. The RP example is part of a trend, in which mission expertise for design, development and operations is distributed across countries and across the globe. Future spacecraft operations will be most cost efficient and flexible by following this distributed expertise, enabling operations from anywhere. For the RP mission we arrived at the decision to utilize a fully distributed operations team, where everyone operates from their home institution, based on evaluating the following factors: the requirement for physical proximity for near-real time command and control decisions; the cost of distributed control nodes vs. a centralized control center; the impact on training and mission preparation of flying the team to a central location. Physical proximity for operational decisions is seldom required, though certain categories of decisions, such as launch abort, or close coordination for mission or safety-critical near-real-time command and control decisions may benefit from co-location. The cost of facilities and operational infrastructure has not been found to be a driving factor for location in our studies. Mission training and preparation benefit from having all operators train and operate from home institutions.

Operations↗
NASA NTRS2018

Estimation of Terrestrial Global Gross Primary Production (GPP) with Satellite Data-Driven Models and Eddy Covariance Flux Data

We estimate global terrestrial gross primary production (GPP) based on models that use satellite data within a simplified light-use efficiency framework that does not rely upon other meteorological inputs. Satellite-based geometry-adjusted reflectances are from the MODerate-resolution Imaging Spectroradiometer (MODIS) and provide information about vegetation structure and chlorophyll content at both high temporal (daily to monthly) and spatial (1 km) resolution. We use satellite-derived solar-induced fluorescence (SIF) to identify regions of high productivity crops and also evaluate the use of downscaled SIF to estimate GPP. We calibrate a set of our satellite-based models with GPP estimates from a subset of distributed eddy covariance flux towers (FLUXNET 2015). The results of the trained models are evaluated using an independent subset of FLUXNET 2015 GPP data. We show that variations in light-use efficiency (LUE) with incident PAR are important and can be easily incorporated into the models. Unlike many LUE-based models, our satellite-based GPP estimates do not use an explicit parameterization of LUE that reduces its value from the potential maximum under limiting conditions such as temperature and water stress. Even without the parameterized downward regulation, our simplified models are shown to perform as well as or better than state-of-the-art satellite data-driven products that incorporate such parameterizations. A significant fraction of both spatial and temporal variability in GPP across plant functional types can be accounted for using our satellite-based models. Our results provide an annual GPP value of 140 Pg C year 1 for 2007 that is within the range of a compilation of observation-based, model, and hybrid results, but is higher than some previous satellite observation-based estimates

CO2↗
DOE OSTI2022

Machine Learning-based Prediction of Departure from Nucleate Boiling Power for the PSBT Benchmark

Machine Learning (ML) has seen an exponential growth in its applications due to its advanced data driven prediction capabilities. The study presents a data-driven approach as a preliminary attempt to predict the power at which departure from nucleate boiling (DNB) occurs in pressurized water reactors (PWRs) by constructing an advanced ML algorithm that takes outlet pressure, inlet temperature and inlet mass flux as the input features. DNB is a critical heat flux (CHF) phenomenon seen in PWRs. The experimental data from the PWR subchannel and bundle tests (PSBT) benchmark is first used to train an artificial neural network (ANN) to predict the DNB power, which produces a root mean square error (RMSE) of 6.89 kW/m when tested on a blind subset of the PSBT data. Since the PSBT dataset is relatively small to train an accurate ANN, a data augmentation methodology based on generative adversarial networks (GANs) is used to expand the training dataset. By assuming that the real data follows a certain distribution, GANs try to learn that underlying distribution to generate similar synthetic data to augment the database and to improve the predictive capabilities of the ANN. The data generated from GANs are validated using 1-nearest neighbor and kernel maximum mean discrepancy. To further ensure data from GAN is similar to PSBT, the data is tested and filtered out using the sub-channel thermal-hydraulic code CTF. The results indicate that with the addition of 120 data points from GAN the RMSE reduces to 4.84 kW/m showing promising results for future developments.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2021

Learning Sequential Distribution System Restoration via Graph-Reinforcement Learning

We report a distribution service restoration algorithm as a fundamental resilient paradigm for system operators provides an optimally coordinated, resilient solution to enhance the restoration performance. The restoration problem is formulated to coordinate distribution generators and controllable switches optimally. A model-based control scheme is usually designed to solve this problem, relying on a precise model and resulting in low scalability. To tackle these limitations, this work proposes a graph-reinforcement learning framework for the restoration problem. We link the power system topology with a graph convolutional network, which captures the complex mechanism of network restoration in power networks and understands the mutual interactions among controllable devices. Latent features over graphical power networks produced by graph convolutional layers are exploited to learn the control policy for network restoration using deep reinforcement learning. The solution scalability is guaranteed by modeling distributed generators as agents in a multi-agent environment and a proper pre-training paradigm. Comparative studies on IEEE 123-node and 8500-node test systems demonstrate the performance of the proposed solution.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

A Data-Driven Framework for Power System Event Type Identification via Safe Semi-Supervised Techniques

Herein this paper investigates the use of phasor measurement unit (PMU) data with deep learning techniques to construct real-time event identification models for transmission networks. Increasing penetration of distributed energy resources represents a great opportunity to achieve decarbonization, as well as challenges in systematic situational awareness. When high-resolution PMU data and sufficient manually recorded event labels are available, the power event identification problem is defined as a statistical classification problem that can be solved by numerous cutting-edge classifiers. However, in real grids, collecting tremendous high-quality event labels is quite expensive. Utilities frequently have a large number of event records without in-depth details (i.e., unlabeled events). To bridge this gap, we propose a novel semi-supervised learning-based method to improve the performance of event classifiers trained with a limited number of labeled events by exploiting the information from massive unlabeled events. In other words, compared to existing data-driven methods, our method requires only a small portion of labeled data to achieve a similar level of accuracy. Meanwhile, this work discusses and addresses the performance degradation caused by class distribution mismatch between the training set and the real applications. Based on the proposed safe learning mechanism, our model does not directly use all unlabeled events during model training, but selectively uses them through a comprehensive evaluation procedure. Numerical studies on a sizable PMU dataset have been used to validate the performance of the proposed method.

42 ENGINEERING↗
DOE OSTI2026

Safe Deep Reinforcement Learning for Active Distribution System Model Predictive Control with EVs and DERs

The temporal and spatial mismatch between PV generation and electric vehicle (EV) charging and discharging may cause voltage violations in active distribution networks. Despite the widespread use of deep reinforcement learning (DRL) in power system optimization and control, it lacks guarantees on constraint satisfaction during both training and deployment. This paper proposes a Lagrangian-based safe DRL approach for model predictive control (MPC) of active distribution systems with large-scale integration of PVs, EVs, and energy storage systems (ESSs). A Transformer-LSTM time-series model is proposed to forecast EV charging demand, which is then formulated as a constraint to ensure charging requirements are met. Using this prediction, a Lagrangian-based safe soft actor-critic (SAC) framework is developed for real-time control in a three-phase unbalanced distribution system, enforcing voltage safety constraints while optimizing the cumulative net reward. By integrating the forecasting model with multi-period constraints, the proposed framework jointly coordinates PV systems, EV charging and discharging, and ESS scheduling within the MPC horizon. Numerical experiments on a modified IEEE 123-bus system with real-world data show that, under a high PV penetration scenario, the proposed method increases the net reward by 30.74% and reduces average voltage violations from 0.0011 p.u. to 0.0002 p.u. compared with standard SAC. Compared with the optimal power flow (OPF) approach, it achieves similar voltage security while yielding lower line losses. It also maintains real-time control capability, reducing operation latency to 53.21 ms per 15-minute control interval. The proposed method remains effective under varying PV/EV penetrations and load conditions.

24 POWER TRANSMISSION AND DISTRIBUTION↗