SEARCH · Search NASA

Results for “task mapping”

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 487 records · Page 27

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Extreme sparsification of physics-augmented neural networks for interpretable model discovery in mechanics

Data-driven constitutive modeling with neural networks has received increased interest in recent years due to its ability to easily incorporate physical and mechanistic constraints and to overcome the challenging and time-consuming task of formulating phenomenological constitutive laws that can accurately capture the observed material response. However, even though neural network-based constitutive laws have been shown to generalize proficiently, the generated representations are not easily interpretable due to their high number of trainable parameters. Sparse regression approaches exist that allow for obtaining interpretable expressions, but the user is tasked with creating a library of model forms which by construction limits their expressiveness to the functional forms provided in the libraries. Here, in this work, we propose to train regularized physics-augmented neural network-based constitutive models utilizing a smoothed version of $L^0$-regularization. This aims to maintain the trustworthiness inherited by the physical constraints, but also enables interpretability which has not been possible thus far on any type of machine learning-based constitutive model where model forms were not assumed a priori but were actually discovered. During the training process, the network simultaneously fits the training data and penalizes the number of active parameters, while also ensuring constitutive constraints such as thermodynamic consistency. We show that the method can reliably obtain interpretable and trustworthy constitutive models for compressible and incompressible hyperelasticity, yield functions, and hardening models for elastoplasticity, using synthetic and experimental data. This work aims to set a new paradigm for interpretable machine learning models in the broad area of solid mechanics where low and limited data is available along with prior knowledge of physical constraints that the learned maps need to obey. This paradigm can potentially be extended to a broader spectrum of scientific exploration.

Data-driven constitutive models↗

DOE OSTI2021

A generalized machine learning workflow to visualize mechanical discontinuity

Accurate detection and mapping of mechanical discontinuity in materials has widespread industrial and research applications. Herein, we developed a generalized machine-learning framework for visualizing single mechanical discontinuity embedded in material of any composition, velocity, density, porosity, and size with limited data. The proposed visualization of discontinuity requires accurate estimations of the length, location, and orientation of the embedded discontinuity by processing multipoint wave-transmission measurements. k-Wave simulator is used to create a large dataset of elastic waveforms recorded during multi-point wave-transmission measurements through materials containing single mechanical discontinuity. k-Wave simulator considers the wave attenuation, dispersion, and mode conversion in wave motion. Discrete wavelet transform (DWT) and statistical feature extraction are essential for data preprocessing prior to the data-driven model development. DWT also minimizes the effect of noise. Using hyper-parameter tuning and cross validation, gradient boosting regression can visualize the mechanical discontinuity with an accuracy of 0.85, in terms of coefficient of determination. A double-layered neural network-based regression has better performance with an accuracy of 0.95. Use of convolutional neural network converts the predictive task from a waveform processing to an image processing problem. Convolutional neural network achieved a generalization performance of 0.91. The proposed generalized workflow requires robust simulation of wave propagation, signal processing, feature engineering, and model evaluation. Sensors closest to the source and those located opposite the source are the most significant for the desired visualization. Notably, the sensors closest to the source capture the non-linear associations, whereas the sensor on the border opposite to the source capture the linear associations between the measured waveforms and the properties of the mechanical discontinuity.

42 ENGINEERING↗

DOE OSTI2025

DeepAndes: A Self-Supervised Vision Foundation Model for Multispectral Remote Sensing Imagery of the Andes

By mapping sites at large scales usingremotely sensed data, archaeologists can generate unique insights into long-term demographic trends, interregional social networks, and human adaptations in the past. Remote sensing surveys complement field-based approaches, and their reach can be especially great when combined with deep learning and computer vision techniques. However, conventional supervised deep learning methods face challenges in annotating fine-grained archaeological features at scale. In addition, while recent vision foundation models have shown remarkable success in learning large-scale remote sensing data with minimal annotations, most off-the-shelf solutions are designed for RGB images rather than multispectral satellite imagery, such as the eight-band data used in our study. In this article, we introduce DeepAndes, a transformer-based vision foundation model trained on three million multispectral satellite images, specifically tailored for Andean archaeology. DeepAndes incorporates a customized DINOv2 self-supervised learning algorithm optimized for eight-band multispectral imagery, marking the first foundation model designed explicitly for the Andes region. We evaluate its image understanding performance through imbalanced image classification, image instance retrieval, and pixel-level semantic segmentation tasks. Our experiments show that DeepAndes achieves superior F1 scores, mean average precision, and Dice scores in few-shot learning scenarios, significantly outperforming models trained from scratch or pretrained on smaller datasets. This underscores the effectiveness of large-scale self-supervised pretraining in archaeological remote sensing.

Guo, Junlin [Vanderbilt Univ., Nashville, TN (Unit↗

DOE OSTI2017

Predictive coarse-graining

We propose a data-driven, coarse-graining formulation in the context of equilibrium statistical mechanics. In contrast to existing techniques which are based on a fine-to-coarse map, we adopt the opposite strategy by prescribing a probabilistic coarse-to-fine map. This corresponds to a directed probabilistic model where the coarse variables play the role of latent generators of the fine scale (all-atom) data. From an information-theoretic perspective, the framework proposed provides an improvement upon the relative entropy method and is capable of quantifying the uncertainty due to the information loss that unavoidably takes place during the coarse-graining process. Furthermore, it can be readily extended to a fully Bayesian model where various sources of uncertainties are reflected in the posterior of the model parameters. The latter can be used to produce not only point estimates of fine-scale reconstructions or macroscopic observables, but more importantly, predictive posterior distributions on these quantities. Predictive posterior distributions reflect the confidence of the model as a function of the amount of data and the level of coarse-graining. The issues of model complexity and model selection are seamlessly addressed by employing a hierarchical prior that favors the discovery of sparse solutions, revealing the most prominent features in the coarse-grained model. A flexible and parallelizable Monte Carlo – Expectation–Maximization (MC-EM) scheme is proposed for carrying out inference and learning tasks. A comparative assessment of the proposed methodology is presented for a lattice spin system and the SPC/E water model.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2015

Managing work flow and logistic at waste processing facilities using the waste compliance and tracking system - 15392

Work flow and logistics at a busy radioactive waste processing complex or permitted treatment, storage, and disposal facility (TSDF) can pose a monumental challenge. An efficient operation requires good communication between engineering, logistics, and waste operators. The Los Alamos National Laboratory (LANL) has implemented an application that facilitates seamless communication of work flow requirements, between engineering, logistics, and waste operations personnel in the field. The Waste Compliance and Tracking System (WCATS), a comprehensive waste management system designed at LANL, provides four key elements to support logistics: (a) the container work path, (b) directed unit operation requests, (c) notification of pending operations, and (d) real-time status tools for office and field workers. The combination of these elements allows everyone to function as a team, with a shared knowledge of what is needed to accomplish facility objectives. The first element is like a road map indicating which unit operations are needed and in what sequence they should be performed; this is referred to as a 'work path'. It automatically routes a waste item or container from operation to operation. For example, a given work path might indicate that waste items require radiological survey, radioassay, and a technical review, in that order. When the work path is assigned to a container, the system enforces the correct order of operations. Directed unit operation 'requests', the second element, fills a gap between pre-planned, formal 'work path' activities and the need for flexibility to direct preparatory operations or deal with unexpected issues, like staging or mining containers in preparation for a shipment or transferring containers from one storage unit to another to clear a path for building maintenance. Logistics engineers prepare these 'requests', and the system monitors the status and completion of the work. The third element, notifications, allows team members to subscribe to and receive notice that an operation is pending, initiated, or complete. This functionality supports a range of communication options, including e-mail, cell phone, pager, or application in-box messaging. Notifications allow waste operations personnel to multi-task, and remain in the loop on the status of facility-wide activities. Finally, the fourth element provides tools that communicate work path pending operations and directed requests to field workers on their mobile computer, allowing them to mine or pick containers from storage units, and leverage the capabilities of a bar-code scanner to avoid selection mistakes. Oversight and summary tools are provided, allowing logistics personnel to monitor work path and request task status, identify work-flow bottlenecks, and maximize the efficiency of waste operations. The Waste Compliance and Tracking System (WCATS) provides an innovative, comprehensive set of tools that allow large-scale radioactive waste processing facilities to manage logistics, maximize efficient use of the workforce, and create a team environment, where everyone from management to the waste handlers have an understanding of the work load and what needs to be accomplished. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2023

classLog: Logistic regression for the classification of genetic sequences

Introduction Sequencing and phylogenetic classification have become a common task in human and animal diagnostic laboratories. It is routine to sequence pathogens to identify genetic variations of diagnostic significance and to use these data in realtime genomic contact tracing and surveillance. Under this paradigm, unprecedented volumes of data are generated that require rapid analysis to provide meaningful inference. Methods We present a machine learning logistic regression pipeline that can assign classifications to genetic sequence data. The pipeline implements an intuitive and customizable approach to developing a trained prediction model that runs in linear time complexity, generating accurate output rapidly, even with incomplete data. Our approach was benchmarked against porcine respiratory and reproductive syndrome virus (PRRSv) and swine H1 influenza A virus (IAV) datasets. Trained classifiers were tested against sequences and simulated datasets that artificially degraded sequence quality at 0, 10, 20, 30, and 40%. Results When applied to a poor-quality sequence data, the classifier achieved between >85% to 95% accuracy for the PRRSv and the swine H1 IAV HA dataset and this increased to near perfect accuracy when using the full dataset. The model also identifies amino acid positions used to determine genetic clade identity through a feature selection ranking within the model. These positions can be mapped onto a maximum-likelihood phylogenetic tree, allowing for the inference of clade defining mutations. Discussion Our approach is implemented as a python package with code available at https://github.com/flu-crew/classLog .

Zeller, Michael A.↗

DOE OSTI2025

Systems Innovation: Modernization & Efficiencies for ESH&Q Reviews

Environmental compliance reviews at INL have traditionally been managed through fragmented systems, relying on multiple spreadsheets and manual processes. This inefficiency led to time-consuming status updates and redundant tasks, such as manually sending reminder emails and transferring data from Excel to the Environmental Review Process (ERP). Initial attempts to streamline these processes using Power Automate and Excel revealed significant limitations, necessitating a more comprehensive solution. To address these immediate inefficiencies, automated workflows were developed using Power Automate. These workflows were designed to send scheduled status update reminders and capture responses through standardized forms, with submitted data flowing directly into centralized Excel trackers. This automation reduced the administrative burden, improved data accuracy, and enabled faster, more consistent reporting. Specifically, email automation achieved a 65% efficiency gain, while data integration saw a 48% improvement, resulting in 91% of project statuses being updated within two months. Despite the improvements brought by Power Automate, the fragmented nature of the review processes persisted. To further enhance efficiency and accuracy, the Integrated Review Tool (IRT) was developed. The IRT aims to centralize review initiation and connect team systems, creating an interconnected data infrastructure that preserves team autonomy while enhancing overall efficiency. This tool automates email reminders, centralizes reviews, and streamlines data integration, significantly improving the accuracy and efficiency of environmental compliance reviews. The design and development of the IRT involved advanced systems methodology, process mapping, project management, and collaboration with subject matter experts. The minimum viable product design is 100% complete, and system development is currently underway, with expected outcomes including a centralized entry point for all ESH&Q reviews, automated routing, real-time tracking and analytics, AI integration, and a user-friendly interface. This project demonstrates the potential of leveraging automation and integrated systems to enhance efficiency, accuracy, and decision-making in environmental reporting and compliance processes at INL.

99 - GENERAL AND MISCELLANEOUS↗

DOE OSTI2025

A formaldehyde and nitrogen dioxide toxics monitor for environmental justice

DOE’s Integrated Field Laboratories (IFLs) require a vast array of measurements to measure urban air quality. Nitrogen dioxide (NO2) and formaldehyde (HCHO) are ubiquitous pollutants with numerous sources, most notably vehicle exhaust. High-spatial density measurements are critical to understanding how certain neighborhoods are more adversely impacted by pollutants. This report describes the results of a Phase I SBIR project to develop and demonstrate a rapid, affordable, low-power monitor to provide high spatial resolution mobile measurements of formaldehyde and nitrogen dioxide at street-level, where people live and walk. The spatial density will be achieved via rapid measurements on a mobile platform. This sensor will quantify spatial disparities in cities of the toxics, HCHO and NO2, as well as carbon monoxide and carbon dioxide for better source identification. These species are detected in the mid infrared by laser absorption spectroscopy, and fiberoptics are used. Tasks accomplished include identification and testing of the appropriate spectroscopic regions; coupling of laser light into a fiberoptic system; development of software to control miniature electronics to scan and control two lasers simultaneously; coupling of the mid infrared fiber into a multipass cell for enhanced pathlength and sensitivity; characterization of sources of instrument noise; and design work for a Phase II prototype.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2016

3D camera assisted fully automated calibration of scanning laser Doppler vibrometers

Scanning laser Doppler vibrometers (LDV) are used to measure full-field vibration shapes of products and structures. In most commercially available scanning laser Doppler vibrometer systems the user manually draws a grid of measurement locations on a 2D camera image of the product. The determination of the correct physical measurement locations can be a time consuming and diffcult task. In this paper we present a new methodology for product testing and quality control that integrates 3D imaging techniques with vibration measurements. This procedure allows to test prototypes in a shorter period because physical measurements locations will be located automatically. The proposed methodology uses a 3D time-of-flight camera to measure the location and orientation of the test-object. The 3D image of the time-of-flight camera is then matched with the 3D-CAD model of the object in which measurement locations are pre-defined. A time of flight camera operates strictly in the near infrared spectrum. To improve the signal to noise ratio in the time-of-flight measurement, a time-of-flight camera uses a band filter. As a result of this filter, the laser spot of most laser vibrometers is invisible in the time-of-flight image. Therefore a 2D RGB-camera is used to find the laser-spot of the vibrometer. The laser spot is matched to the 3D image obtained by the time-of-flight camera. Next an automatic calibration procedure is used to aim the laser at the (pre)defined locations. Another benefit from this methodology is that it incorporates automatic mapping between a CAD model and the vibration measurements. This mapping can be used to visualize measurements directly on a 3D CAD model. Secondly the orientation of the CAD model is known with respect to the laser beam. This information can be used to find the direction of the measured vibration relatively to the surface of the object. With this direction, the vibration measurements can be compared more precisely with numerical experiments.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2025

In situ Detection of Plasma Induced Surface Interaction based on Deep Learning based Visual Diagnostics (Technical Report)

It is characteristic for many plasma devices to undergo plasma-material interaction leading to surface erosion. These processes, often not easily detectable, lead to changes in device performance and lifespan. State-of-the-art lifetime tests and wear experiments require over 1000s hours. A self-consistent model for accurately predicting the erosion's effects is not available. In situ detection of these processes is not a trivial task since the surface variations at the early stages have a micron scale. Such limitations not only restrict testing and prediction capabilities but also slow the development of new thrusters and limit mission duration. To address these challenges, an in-situ diagnostic for real-time erosion assessment has been developed, aiming to expedite lifetime testing and broaden experimental campaigns. Several works were dedicated to real-time and in situ monitoring of material erosion during plasma exposure using laser holography, microscopy, and with telemicroscopes. However, the applicability of these approaches is limited due to complexity, cost and less flexibility as they often require placing diagnostic equipment inside the vacuum chamber. In collaboration with Princeton Collaborative Research Facility (PCRF), Princeton Plasma Physics Laboratory (PPPL), a new diagnostic approach is developed, where geometry modifications to the ceramic channel walls were introduced that would result in accelerated channel erosion. We employed Long-distance microscope (LDM) imagery, combined with Deep-Learning based Shape from focus or depth from focus (DFF or SFF) approach, that provides an accessible and cost-effective solution. LDM employs focus variation techniques to continuously capture multiple images of the target object at distinct focal planes. DFF, an optical focus variation method, generates a 3D topographical surface depth map from a sequence of variably focused images. Combined with the developed diagnostic, this approach offers a controllable means to study erosion under accelerated conditions. In this work, we develop Neural Network-based DFF algorithm applicable for LDM data to quantitatively evaluate plasma induced surface modification from LDM data. Next, we develop Deep Learning-based super-resolution depth map image reconstruction technique to increase the resolution of depth maps obtained from DFF algorithm to improve the accuracy of erosion measurements. Thirdly, we develop several image processing techniques to remove noise and improve the quality of depth map image. Here we report the results of initial tests for this approach. An experimental setup designed and built in PPPL was employed that consists of a 3-cm gridded ion source that produces a neutralized argon beam with energies up to 600 eV. A hexagonal boron nitride (h-BN) ceramic target, designed based on computational predictions, was used. Tests were conducted to reconstruct the complex geometry of the target under the lighting conditions of the operated ion source.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2023

ZeoNet: 3D convolutional neural networks for predicting adsorption in nanoporous zeolites

Zeolites are one of the most widely used materials in the chemical industry due to their nanometer-sized pores that can adsorb and react upon molecules selectively. With hundreds of known framework topologies and hundreds of thousands of computationally predicted structures, the ability to rapidly predict zeolite performance allows researchers to prioritize their efforts on the most promising structures for a given application. Although the accuracy of forcefield-based atomistic simulations has advanced significantly in the past two decades, these simulations can be computationally expensive, especially for long-chain, complex molecules. Here, we present ZeoNet, a representation learning framework using convolutional neural networks (ConvNets) and 3D volumetric representations for predicting adsorption in zeolites. ZeoNet was trained on the task of predicting Henry's constants for adsorption, k H , of n-octadecane in more than 330 000 known and predicted zeolite materials. Employing a 3D grid based on the distances to solvent-accessible surfaces, a volumetric representation that can be generated efficiently, the best-performing ZeoNet achieved a correlation coefficient r 2 = 0.977 and a mean-squared error MSE = 3.8 in ln k H , which corresponds to an error of 9.3 kJ mol -1 in adsorption free energy. In comparison, a model based on hand-designed geometric features has values of r 2 = 0.783 and MSE = 35.7. ZeoNet is also relatively efficient and can process ≈8 structures per second on an Nvidia RTX 2080TI GPU, orders of magnitude faster than forcefield-based simulations. A systematic analysis was conducted to investigate how the choice of ConvNet architectures, the linear dimension (L) and spatial resolution (Δd) of the distance grids, batch size, optimizer, and learning rate impact the model performance. We found that ConvNets based on the ResNet architecture offer the best tradeoff between expressiveness and efficiency. The performance for all models reaches a plateau at L = 30–45 Å and depends less sensitively on grid resolution, with a small benefit around Δd = 0.30–0.45 Å. Finally, saliency maps were visualized to identify which regions of the materials contributed the most to model predictions. It was found, interestingly, that the predictions are driven primarily by the accessible pore volume rather than the region occupied by the framework atoms.

36 MATERIALS SCIENCE↗

DOE OSTI2021

DeepDist: real-value inter-residue distance prediction with deep residual convolutional network

Abstract Background Driven by deep learning, inter-residue contact/distance prediction has been significantly improved and substantially enhanced ab initio protein structure prediction. Currently, most of the distance prediction methods classify inter-residue distances into multiple distance intervals instead of directly predicting real-value distances. The output of the former has to be converted into real-value distances to be used in tertiary structure prediction. Results To explore the potentials of predicting real-value inter-residue distances, we develop a multi-task deep learning distance predictor (DeepDist) based on new residual convolutional network architectures to simultaneously predict real-value inter-residue distances and classify them into multiple distance intervals. Tested on 43 CASP13 hard domains, DeepDist achieves comparable performance in real-value distance prediction and multi-class distance prediction. The average mean square error (MSE) of DeepDist’s real-value distance prediction is 0.896 Å 2 when filtering out the predicted distance ≥ 16 Å, which is lower than 1.003 Å 2 of DeepDist’s multi-class distance prediction. When distance predictions are converted into contact predictions at 8 Å threshold (the standard threshold in the field), the precision of top L/5 and L/2 contact predictions of DeepDist’s multi-class distance prediction is 79.3% and 66.1%, respectively, higher than 78.6% and 64.5% of its real-value distance prediction and the best results in the CASP13 experiment. Conclusions DeepDist can predict inter-residue distances well and improve binary contact prediction over the existing state-of-the-art methods. Moreover, the predicted real-value distances can be directly used to reconstruct protein tertiary structures better than multi-class distance predictions due to the lower MSE. Finally, we demonstrate that predicting the real-value distance map and multi-class distance map at the same time performs better than predicting real-value distances alone.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2020

Savannah River Site H-Canyon Advancing Technologies for Remote Inspections - 20345

In 2017, the DOE Environmental Management Office of Technology Development (DOE-EM TD) sponsored the H-Canyon Advanced Technology Demonstration (ATD) to demonstrate to DOE facilities the value of using new commercial-off-the-shelf (COTS) and near-ready technologies to solve difficult problems and enhance worker safety. The DOE Savannah River Site (SRS) H-Canyon Air Exhaust Tunnel (HCAEX) inspection task was identified as representative of the hazardous, human denied environments which could benefit from advanced technologies. The HCAEX underground concrete tunnel is visually inspected biannually using a camera mounted on a remotely operated vehicle (ROV) designed and built by SRNL. While tunnel images have provided valuable visual information, it is desirable to have a higher order of understanding of the environment to support a more thorough structural integrity (SI) analysis and for long term planning purposes. As part of the ATD, the Concrete Integrated Product Team (CIPT) was formed to identify and evaluate available sensors and methods mature enough to remotely obtain tunnel concrete characterization data of high value and with a high probability of success. The team included SMEs and H-Canyon stakeholders in the field of concrete, nondestructive examination (NDE), structural integrity, sensors and remote systems from SRNL, SRNS, LANL, DOE-SR and the Army Corps of Engineering. The CIPT completed an in-depth identification of customer concrete inspection needs and potential technology solutions. Sensors and methods were evaluated on performance, data usefulness, cost and the feasibility of a successful deployment given the unique tunnel access challenges and environment. Two technologies were identified as promising by the CIPT for near term demonstration and evaluation: Lidar (Light Detection and Ranging) 3-dimensional (3D) mapping and remote robotic deployment of NDE instrumentation. Laser spectroscopy to characterize tunnel surface chemical changes was also of interest, but presently cost prohibitive. This paper will include a discussion of the two efforts underway to evaluate and implement the CIPT recommendations. First, the status of the November 2019 deployment of Lidar at a single location into the tunnel is presented. This initial deployment provided the team a learning curve and lessons learned on the challenges of tunnel deployment to include remote operation and data collection, stabilization of the sensor in high air flow (∼30 mph), ability to achieve a tolerance accuracy of 0.25-inches, and the probability to identify change in tunnel dimensions over time. Secondly, a discussion on the development of the Robotic Arm Concrete Inspection Test Bed capable of deploying NDE instruments to examine custom concrete forms will be presented. Concrete forms simulating the rough concrete surfaces, strength, composition and potential structural defects that can be found at our DOE EM facilities have been designed and built for the test bed. Two state-of-the art concrete NDE instruments have been identified as having potential to work on rough concrete walls, they are being tested and characterized as to their ability to provide desired structural integrity data to include wall thickness and defect identification on the developed test beams. Lastly, lessons learned, and the path forward will be presented. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2021

InversionNet3D: Efficient and Scalable Learning for 3-D Full-Waveform Inversion

Seismic full-waveform inversion (FWI) techniques aim to find a high-resolution subsurface geophysical model provided with waveform data. Some recent effort in data-driven FWI has shown some encouraging results in obtaining 2-D velocity maps. However, due to high computational complexity and large memory consumption, the reconstruction of 3-D high-resolution velocity maps via deep networks is still a great challenge. Here, in this article, we present InversionNet3D (InvNet3D), an efficient and scalable encoder–decoder network for 3-D FWI. The proposed method employs group convolution in the encoder to establish an effective hierarchy for learning information from multiple sources while cutting down unnecessary parameters and operations at the same time. The introduction of invertible layers further reduces the memory consumption of intermediate features during training and, thus, enables the development of deeper networks with more layers and higher capacity as required by different application scenarios. Experiments on the 3-D Kimberlina dataset demonstrate that InvNet3D achieves state-of-the-art reconstruction performance with lower computational cost and lower memory footprint compared to the baseline.

58 GEOSCIENCES↗

DOE OSTI2023

Robotics Plan in Support of DOME Testbed Operations

Various robotic tooling options have been evaluated for the NRIC DOME concept of operations (ConOps). Framatome was contracted to develop a wide-ranging list of commercial off the shelf (COTS) and custom robotic systems to be considered for performing the DOME ConOps functions. Subsequent project tasks from Framatome narrowed down the list and scored the most viable options. The abbreviated list, and associated scoring, has been reviewed and assessed to provide formal recommendations for the robotic ConOps functions of DOME. The overhead telescoping mast & Kraft arm assembly is recommended as the primary system for reactor demobilization and removal. The estimated cost is $\$700$K with a timeline to develop and deliver of about 2.5 years. The mast & Kraft arm would still require an overhead lift system for mobilization and installation. Either the refurbished polar crane or a new gantry crane delivery platform could both serve as the overhead lift and delivery system. The estimated costs for the refurbishment polar crane and new gantry crane systems are, respectively, $\$4.4$M and $\$3$M with about 2.5 years to develop and deliver. A Brokk + Kraft crawler & arm is recommended as a secondary robotic system to support the overhead mast & Kraft arm system. The Brokk + Kraft crawler & arm would cost an estimated $\$726$k and would take about 1.5 years to develop and deliver. The Boston Dynamics SPOT robot is also recommended for the in-service monitoring during experimentation operations. The estimated cost is $\$220$K with a timeline to deliver of 8-14 months. The Elios 3 aerial drone is recommended to perform large area radiation dose rate mapping and visual inspections. The estimated cost is a $\$100$K and will also need about 8-14 months to develop and deliver. A mockup rig is also recommended to be procured an employed to test & verify the robotic capabilities as well as provide needed training for operations personnel. It is estimated that a mockup rig would cost about $\$500$k and would take up to a year to develop and build.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

DOE OSTI2025

The Importance of Being Adaptable: An Exploration of the Power and Limitations of Domain Adaptation for Simulation-Based Inference with Galaxy Clusters

The application of deep machine learning methods in astronomy has exploded in the last decade, with new models showing remarkably improved performance on benchmark tasks. Not nearly enough attention is given to understanding the models' robustness, especially when the test data are systematically different from the training data, or "out of domain." Domain shift poses a significant challenge for simulation-based inference, where models are trained on simulated data but applied to real observational data. In this paper, we explore domain shift and test domain adaptation methods for a specific scientific case: simulation-based inference for estimating galaxy cluster masses from X-ray profiles. We build datasets to mimic simulation-based inference: a training set from the Magneticum simulation, a scatter-augmented training set to capture uncertainties in scaling relations, and a test set derived from the IllustrisTNG simulation. We demonstrate that the Test Set is out of domain in subtle ways that would be difficult to detect without careful analysis. We apply three deep learning methods: a standard neural network (NN), a neural network trained on the scatter-augmented input catalogs, and a Deep Reconstruction-Regression Network (DRRN), a semi-supervised deep model engineered to address domain shift. Although the NN improves results by 17% in the Training Data, it performs 40% worse on the out-of-domain Test Set. Surprisingly, the Scatter-Augmented Neural Network (SANN) performs similarly. While the DRRN is successful in mapping the training and Test Data onto the same latent space, it consistently underperforms compared to a straightforward Yx scaling relation. These results serve as a warning that simulation-based inference must be handled with extreme care, as subtle differences between training simulations and observational data can lead to unforeseen biases creeping into the results.

Ntampaka, Michelle [Baltimore, Space Telescope Sci↗

DOE OSTI2024

Self-Supervised Cloud Classification

Abstract Low-level marine clouds play a pivotal role in Earth’s weather and climate through their interactions with radiation, heat and moisture transport, and the hydrological cycle. These interactions depend on a range of dynamical and microphysical processes that result in a broad diversity of cloud types and spatial structures, and a comprehensive understanding of cloud morphology is critical for continued improvement of our atmospheric modeling and prediction capabilities moving forward. Deep learning has recently accelerated our ability to study clouds using satellite remote sensing, and machine learning classifiers have enabled detailed studies of cloud morphology. A major limitation of deep learning approaches to this problem, however, is the large number of hand-labeled samples that are required for training. This work applies a recently developed self-supervised learning scheme to train a deep convolutional neural network (CNN) to map marine cloud imagery to vector embeddings that capture information about mesoscale cloud morphology and can be used for satellite image classification. The model is evaluated against existing cloud classification datasets and several use cases are demonstrated, including training cloud classifiers with very few labeled samples, interrogation of the CNN’s learned internal feature representations, cross-instrument application, and resilience against sensor calibration drift and changing scene brightness. The self-supervised approach learns meaningful internal representations of cloud structures and achieves comparable classification accuracy to supervised deep learning methods without the expense of creating large hand-annotated training datasets. Significance Statement Marine clouds heavily influence Earth’s weather and climate, and improved understanding of marine clouds is required to improve our atmospheric modeling capabilities and physical understanding of the atmosphere. Recently, deep learning has emerged as a powerful research tool that can be used to identify and study specific marine cloud types in the vast number of images collected by Earth-observing satellites. While powerful, these approaches require hand-labeling of training data, which is prohibitively time intensive. This study evaluates a recently developed self-supervised deep learning method that does not require human-labeled training data for processing images of clouds. We show that the trained algorithm performs competitively with algorithms trained on hand-labeled data for image classification tasks. We also discuss potential downstream uses and demonstrate some exciting features of the approach including application to multiple satellite instruments, resilience against changing image brightness, and its learned internal representations of cloud types. The self-supervised technique removes one of the major hurdles for applying deep learning to very large atmospheric datasets.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

Light Water Sustainability Program: Optimizing Information Automation Using a New Method Based on System-Theoretic Process Analysis

This report describes the interim progress for research supporting the design and optimization of information automation systems for nuclear power plants. Much of the domestic nuclear fleet is currently focused on modernizing technologies and processes, including transitioning toward digitalization in the control room and elsewhere throughout the plant, along with a greater use of automation, artificial intelligence, robotics, and other emerging technologies. While there are significant opportunities to apply these technologies toward greater plant safety, efficiency, and overall cost-effectiveness, optimizing their design and avoiding potential safety and performance risks depends on ensuring that human-performance-related organizational and technical design issues are identified and addressed. This report describes modeling tools and techniques, based on sociotechnical system theory, to support these design goals and their application in the current research effort. The report is intended for senior nuclear energy stakeholders, including regulators, corporate management, and senior plant management. We have developed and employed a method to design an optimized information automation ecosystem (IAE) based on the systems-theoretic constructs underlying sociotechnical systems theory in general and the Systems-Theoretic Accident Modeling and Processes (STAMP) approach in particular. We argue that an IAE can be modeled as an interactive information control system whose behavior can be understood in terms of dynamic control and feedback relationships amongst the system’s technical and organizational components. Up to this point, we have employed a Causal Analysis based on STAMP (CAST) technique to examine a performance- and safety-related incident at an industry partner’s plant that involved the unintentional activation of an emergency diesel generator. This analysis provided insight into the behavior of the plant’s current information control structure within the context of a specific, significant event. Our ongoing analysis is focused on identifying near-term process improvements and longer-term design requirements for an optimized IAE system. The latter analyses will employ a second STAMP-derived technique, System-Theoretic Process Analysis (STPA). STPA is a useful modeling tool for generating and analyzing actual or potential information control structures. Finally, we have begun modeling plantwide organizational relationships and processes. Organizational system modeling will supplement our CAST and STPA findings and provide a basis for mapping out a plantwide information control architecture. CAST analysis findings indicate an important underlying contributor to the incident under investigation, and a significant risk to information automation system performance, was perceived schedule pressure, which exposed weaknesses in interdepartmental coordination between and within responsible plant organizations and challenged the resilience of established plant processes, until a human caused the initiating event. These findings are discussed in terms of their risk to overall system performance and their implications for information automation system resilience and brittleness. We present two preliminary information automation models. The proactive issue resolution model is a test case of an information automation concept with significant near-term potential for application and subsequent reduction in significant plant events. The IAE model is a more general representation of a broader, plantwide information automation system. From our results, we have generated a set of preliminary system-level requirements and safety constraints. These requirements will be further developed over the remainder of our project in collaboration with nuclear industry subject matter experts and specialists in the technical systems under consideration. Additionally, we will continue to pursue the system analyses initiated in the first part of our effort, with a particular emphasis on STPA as the main tool to identify weak or weakening control structures that affect the resilience of organizations and programs. Our intent is to broaden the scope of the analysis from an individual use case to a related set of use cases (e.g., maintenance tasks, compliance tasks) with similar human-system performance challenges. This will enable more generalized findings to refine the Proactive Issue Resolution and IAE models, as well as their system-level requirements and safety constraints. We will use organizational system modeling analyses to supplement STPA findings and model development. We conclude the report with a set of summary recommendations and an initial draft list of system-level requirements and safety constraints for optimized information automation systems.

99 GENERAL AND MISCELLANEOUS↗