Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “generative machine learning”

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 883 records · Page 49

Safety Assurance of Software and Machine Learning Development for Nuclear Instrumentation and Controls

Digital instrumentation and control (DI&C) systems monitor and control parameters in nuclear power plants. Ensuring their safety is a critical part of ensuring overall plant safety. Nuclear power plant licensing generates thousands of safety documents that could be organized more effectively using a safety assurance case (SAC). We conducted a literature survey of SACs and created a SAC framework for DI&C software using Goal Structuring Notation (GSN). This framework focuses on four software development processes: management & assurance, pre-developed software (PDS) qualification, the Software Development Life Cycle (SDLC), and the Machine Learning Development Life Cycle (MLDLC). We organized our framework using a novel level structure that can be applied to other SACs to improve their clarity. Finally, we demonstrate how our framework can be incorporated as part of a SAC for a larger reactor system.

46 - INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AN↗

Optimizing enzymes for plastic upcycling using machine learning design and high throughput experiments

Plastic use is ubiquitous in the modern world, and polyethylene terephthalate (PET) is one of the most abundantly produced plastics (and the most highly produced polyester), with ~65 million metric tons manufactured annually. To the consumer, PET is likely most recognizable as the plastic used to make beverage bottles. Like many plastics, traditional mechanical or chemical means of PET deconstruction and upcycling are costly and inefficient. Because of these challenges, recycled plastic is generally of lower quality and is more expensive to produce than virgin plastic derived from petroleum. Ultimately, this results in most plastic ending up as waste. We view plastic waste as an underutilized resource which, with the development of more efficient and high-quality recycling processes, could (1) generate significant economic value while (2) decreasing petroleum usage and greenhouse gas emissions, as well as (3) minimizing its negative environmental and health impacts. Biocatalytic recycling, or biomanufacturing the basic building blocks of new plastic from plastic waste, is a promising approach to plastic reuse that complements existing recycling technologies. Recently, biological enzymes capable of breaking down PET have garnered significant attention as an attractive means of dealing with the plastic problem. These enzymes are currently undergoing pilot studies for implementation in industrial-scale enzyme-based recycling. However, there are significant limitations to current enzymes, including the need to perform costly pre-processing of the plastic waste before the enzymes are able to work. Further optimization of these enzymes is necessary to make these technologies competitive, and ultimately incentivise industry-wide adoption of this biology-based green recycling technology. n this work we demonstrate a means to design and generate performant biological enzymes, capable of efficiently deconstructing plastic waste. Specifically, we applied recent advances in artificial intelligence, machine learning, and statistical analysis to design new versions and discover natural enzymes capable of breaking down PET. We focused on optimizing key properties that are important for industrial-scale enzymatic recycling such as pH and thermotolerance. Normal testing of enzymatic plastic-deconstruction is extremely labor intensive and so through this work we also developed a robotic-assisted experimental pipeline capable of characterizing thousands of candidate enzymes. The results of this iterative, AI-guided, multi-discipline approach have led to increases in enzymatic breakdown of over 150X over starting enzymes. This work supports the rapidly developing and transformative field of biocatalytic solutions to environmental problems beyond the discovery and predictive understanding of enzymes for polymer recycling, and has wide implications for tackling numerous energy problems such as carbon capture and fixation (e.g., engineering carbon monoxide dehydrogenase and the rubisco-pathway), biomining (e.g., design of lanthanide-binding proteins) and biomanufacturing (e.g., lignin-deconstruction enzymes).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Communication-Avoiding and Memory-Constrained Sparse Matrix-Matrix Multiplication at Extreme Scale

Sparse matrix-matrix multiplication (SpGEMM) is a widely used kernel in various graph, scientific computing and machine learning algorithms. In this paper, we consider SpGEMMs performed on hundreds of thousands of processors generating trillions of nonzeros in the output matrix. Distributed SpGEMM at this extreme scale faces two key challenges: (1) high communication cost and (2) inadequate memory to generate the output. Furthermore, we address these challenges with an integrated communication-avoiding and memory-constrained SpGEMM algorithm that scales to 262,144 cores (more than 1 million hardware threads) and can multiply sparse matrices of any size as long as inputs and a fraction of output fit in the aggregated memory. As we go from 16,384 cores to 262,144 cores on a Cray XC40 supercomputer, the new SpGEMM algorithm runs 10x faster when multiplying large-scale protein-similarity matrices.

97 MATHEMATICS AND COMPUTING↗

Usable Data Abstractions for Next-Generation Scientific Workflows

Data- and computationally-intensive scientific research, such as numerical simulations and inversions or the training of large neural networks in machine learning applications, that are well suited for HPC environments also often require expert insight and evaluation throughout the computation which can be greatly facilitated with the use of interactive computing tools, such as those in the Jupyter ecosystem. HPC workflows and interactive workflows are typically treated as orthogonal, however, the next generation of research will require both. The first challenge we face in this project is thus designing the right level of abstractions to allow interactive capabilities in the JupyterLab environment to allow the working scientist to flexibly explore and query their data at multiple levels, with a minimal amount of customization required of the underlying optimized codes. In addition to these questions regarding the high-level representation of data for interactive use in HPC, we tackled two additional issues that are part of the entire lifecycle of research and that become particularly acute in HPC contexts: how to improve the experience of interfacing with the HPC system's scheduling environment for a scientist focused on exploratory questions, and how can that scientist then best share the results of their work with others in a self-contained, reproducible manner.

97 MATHEMATICS AND COMPUTING↗

Group Structure Machine Learning Proposal

Nuclear data is the linchpin underwriting several fundamental capabilities and mission needs at LANL. New techniques such as machine learning can be brought to bear to solve old problems such as multigroup cross-section accuracy. In neutron transport, generating multigroup cross sections is a complex and arcane task, but a crucial one, as accurate solutions require appropriate cross sections. There are two key challenges when generating multigroup cross sections: (1) choosing an accurate weight function, and (2) choosing appropriate energy boundaries. Often, energy boundaries are chosen using “expert judgment” that is not documented and is difficult to replicate. The long-standing Los Alamos 30-group structure has been in use since at least 1969 and is still in use today. Simplistic attempts over the years since to improve on the 30-group structure have been met with limited success. Machine learning algorithms would enable the selection of appropriate, problem-dependent group boundaries without an inordinate investment of scientist time. We will develop workflows and tools to enable these improved group boundary choices, which will reduce uncertainty and increase predictive capability of neutron-transport applications at LANL.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Machine learning at the edge to improve in-field safeguards inspections

Artificial intelligence (AI) and machine learning (ML) are near-ubiquitous in day-to-day life; from cars with automated driver-assistance, recommender systems, generative content platforms, and large language chatbots. Implementing AI as a tool for international safeguards could significantly decrease the burden on safeguards inspectors and nuclear facility operators. The use of AI would allow inspectors to complete their in-field activities quicker, while identifying patterns and anomalies and freeing inspectors to focus on the uniquely human component of inspections. Sandia National Laboratories has spent the past two and a half years developing on-device machine learning to develop both a digital and robotic assistant. This combined platform, which we term inspecta, has numerous on-device machine learning capabilities that have been demonstrated at the laboratory scale. Here this work describes early successes implementing AI/ML capabilities to reduce the burden of tedious inspector tasks such as seal examination, information recall, note taking, and more.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

A machine-learning-aided data recovery approach for predicting multi-material thermal behaviors in advanced test reactor capsules

Instrumented experiments conducted at test reactors are essential to the deployment of new advanced reactor systems. Designing new experiments and generating data on specific reactor conditions require significant investments in terms of both time and cost. Finite element analysis software can be used to create high-fidelity models of experiment environments in order to support the actual experiments, but computation time remains a concern in terms of applying outcomes to real-time usage of data (e.g., a digital twin [DT]). Here, the present research proposes a machine-learning (ML) aided approach to making temperature and displacement predictions based on the thickness of the outer gas gap on the experimental capsule used for in-pile demonstration of a novel new thermal conductivity probe in the Advanced Test Reactor (ATR). This capsule consisted of U10Zr fuel, a rodlet, sodium, and inner and outer capsules. Gas gaps existed between the fuel and the rodlet, and between the inner and the outer capsule. The learning data pertained to an experimental capsule's radial distributions of temperature and displacement, as obtained based on Abaqus and the physical features. For the first step of ML sequence, the temperature was predicted using three positional parameters. Next, the displacement was predicted using seven additional parameters. Each physical feature was normalized in order to be both nondimensional and standardized. The temperature and displacement predictions showed good agreement with the simulation results in all cases involving interpolation and extrapolation. Furthermore, data similarity enhancement increased the similarity between the training and the target data, thereby increasing the predictive accuracy of the ML models. In certain extrapolation cases involving limited original ML model accuracy, data similarity enhancement and data recovery was able to somewhat improve this accuracy.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Additive Manufacturing – A New Challenge for Automation and Robotics

Additive manufacturing (AM) is a rapidly growing technology descended from the first stereolithography systems. AM describes a variety of material deposition technologies for forming objects in a digital manner layer-by-layer under computer control. Now commonly known as 3D printing, AM quickly branched out into several key directions – material extrusion, sheet lamination, direct energy deposition, vat polymerization, powder bed fusion, binder jetting, and material jetting. The common factor in all AM branches is a foundation in robotics and automation. While most of the mechanical 3D printing structures are based on simple gantry systems, there are Gough-Stewart platforms and, more recently, six or more DoF manipulator-based systems that have been developed. Currently available commercial systems are based on open-loop control with minimum sensing capabilities; the latest systems in development are starting to take advantage of complex feedback loops and layers of advanced sensing and data logging. The Manufacturing Demonstration Facility of Oak Ridge National Laboratory is leading the efforts in applying advanced robotics in the creation of large-scale 3D printers. The recent demonstration of an additively manufactured excavator at the CONEXPO 2017 exhibition in Las Vegas showed that the use of cutting edge robotics and automation is essential for the next generation of additive systems. The future of AM will heavily rely on advanced robotics, machine learning, and the internet of things. This paper summarizes progress in AM; presents the practical aspects, challenges, and lessons learned in developing robotic-based AM systems; and outlines the needs and future directions of robotics for AM.

Nycz, Andrzej↗

Evaluation of normalization strategies for mass spectrometry-based multi-omics datasets

Introduction Data normalization is crucial for multi-omics integration, reducing systematic errors and maximizing the likelihood of discovering true biological variation. Most studies assess normalization for a single omics type or use datasets from separate experiments. Few address time-course data, where normalization might bias temporal differentiation. In this study, we compared common normalization methods and a machine learning approach, Systematical Error Removal using Random Forest (SERRF), using multi-omics datasets generated from the same experiment—even from the same cell lysate. Objectives To develop a straightforward process to assess normalization effects and identify the most robust methods across multi-omics datasets. Methods We analyzed metabolomics, lipidomics, and proteomics datasets from primary human cardiomyocytes and motor neurons exposed to acetylcholine-active compounds over time. Normalization effectiveness was evaluated based on improvement in QC features consistency and observing the change in treatment and time-related variance. Results Probabilistic Quotient Normalization (PQN) and Locally Estimated Scatterplot Smoothing (LOESS) QC were identified as optimal for metabolomics and lipidomics, while PQN, Median, and LOESS normalization excelled for proteomics. These methods consistently enhanced QC feature consistency in metabolomics and lipidomics, and preserved time-related variance or treatment-related variance in proteomics, demonstrating their effectiveness and robustness. SERRF normalization, applied only to metabolomics in this study, outperformed other methods in some datasets but inadvertently masked treatment-related variance in others. Conclusion Our evaluation identified PQN and LoessQC as the top methods for metabolomics and lipidomics, and PQN, Median, and Loess normalization for proteomics, in multi-omics integration in a temporal study.

60 APPLIED LIFE SCIENCES↗

Machine learning surrogates for ion energy–angle distributions in thermal and RF plasma sheaths

Ion energy–angle distributions (IEADs) at material surfaces are a critical input for plasma–material interaction (PMI) studies in fusion devices, yet they are computationally expensive to obtain using particle-in-cell (PIC) simulations. In this work, we develop a machine learning surrogate based on a deep deconvolutional neural network (DDeCNN) trained on large databases generated with the hPIC2 code. The surrogate is capable of reconstructing IEADs from sheath parameters for both thermal and radio-frequency (RF) plasmas, including cases with multiple ion species. Across thousands of test cases, the model achieves high accuracy, with over 97 % of predictions classified as good or average based on standard error metrics (MAE, MSE, L2). Even in the more challenging RF and multi-species regimes, the surrogate reliably captures the multi-peak structure of PIC results. Once trained, the surrogate produces IEADs in milliseconds on a common workstation, yielding speedups of six to seven orders of magnitude compared with running a full PIC simulation. This computational gain enables dense parameter scans and direct coupling of IEAD predictions with PMI and erosion models on whole-device scales in fusion-relevant conditions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Nanoengineering of non-aqueous liquid electrolyte solutions for future lithium metal batteries

Research and development of non-aqueous electrolyte solutions are essential for practical advancement towards the production of high-energy lithium metal batteries (LMBs). An ideal LMB electrolyte solution should enable highly efficient, uniform and prolonged lithium metal plating and stripping, preserve the electrodes’ electro(chemo)mechanical properties and ensure compatibility with all cell components. However, despite extensive research efforts, scientists have yet to achieve an electrolyte design that meets these requirements simultaneously. Here, by examining the nanoengineering aspects of various non-aqueous electrolyte solution designs, we elucidate the understanding of the nanoscale physicochemical and electrochemical processes taking place in LMBs, which are mainly governed by the thermodynamic and kinetic properties of the electrolyte system. We also explore emerging research directions and propose an accelerated, iterative framework that integrates nanoengineering principles with machine learning, high-throughput computation and experimentation to facilitate the development of next-generation non-aqueous electrolyte solutions for practical LMBs.

Weintz, Dominik↗

NENCI-2021. I. A large benchmark database of non-equilibrium non-covalent interactions emphasizing close intermolecular contacts

In this work, we present NENCI-2021, a benchmark database of ~8000 Non-Equilibirum Non-Covalent Interaction energies for a large and diverse selection of intermolecular complexes of biological and chemical relevance. To meet the growing demand for large and high-quality quantum mechanical data in the chemical sciences, NENCI-2021 starts with the 101 molecular dimers in the widely used S66 and S101 databases and extends the scope of these works by (i) including 40 cation–π and anion–π complexes, a fundamentally important class of non-covalent interactions that are found throughout nature and pose a substantial challenge to theory, and (ii) systematically sampling all 141 intermolecular potential energy surfaces (PESs) by simultaneously varying the intermolecular distance and intermolecular angle in each dimer. Designed with an emphasis on close contacts, the complexes in NENCI-2021 were generated by sampling seven intermolecular distances along each PES (ranging from 0.7× to 1.1× the equilibrium separation) and nine intermolecular angles per distance (five for each ion–π complex), yielding an extensive database of 7763 benchmark intermolecular interaction energies (E int ) obtained at the coupled-cluster with singles, doubles, and perturbative triples/complete basis set [CCSD(T)/CBS] level of theory. The E int values in NENCI-2021 span a total of 225.3 kcal/mol, ranging from -38.5 to +186.8 kcal/mol, with a mean (median) E int value of -1.06 kcal/mol (-2.39 kcal/mol). In addition, a wide range of intermolecular atom-pair distances are also present in NENCI-2021, where close intermolecular contacts involving atoms that are located within the so-called van der Waals envelope are prevalent—these interactions, in particular, pose an enormous challenge for molecular modeling and are observed in many important chemical and biological systems. A detailed symmetry-adapted perturbation theory (SAPT)- based energy decomposition analysis also confirms the diverse and comprehensive nature of the intermolecular binding motifs present in NENCI-2021, which now includes a significant number of primarily induction-bound dimers (e.g., cation–π complexes). NENCI-2021 thus spans all regions of the SAPT ternary diagram, thereby warranting a new four-category classification scheme that includes complexes primarily bound by electrostatics (3499), induction (700), dispersion (1372), or mixtures thereof (2192). A critical error analysis performed on a representative set of intermolecular complexes in NENCI-2021 demonstrates that the E int values provided herein have an average error of ±0.1 kcal/mol, even for complexes with strongly repulsive E int values, and maximum errors of ±0.2–0.3 kcal/mol (i.e., ~±1.0 kJ/mol) for the most challenging cases. For these reasons, we expect that NENCI-2021 will play an important role in the testing, training, and development of next-generation classical and polarizable force fields, density functional theory approximations, wavefunction theory methods, and machine learning based intra- and inter-molecular potentials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Finite-element-based simulations of electrodes for CO 2 cascade reduction reactions

The multielectron reduction of CO 2 to liquid fuels could be a path to scalable energy storage, but reaching this goal requires major advances in catalysis and systems engineering. Cascade catalysis, which couples sequential reactions without isolating intermediates, has emerged as a promising route to enhance selectivity and efficiency in CO 2 reduction (CO 2 R). In this review, we examine how finite-element-based simulations of continuum model [finite element method (FEM)] approaches are being used to analyze and guide CO 2 R cascade systems. We first outline the fundamentals of cascade catalysis and recent advances in catalytic materials (metallic, molecular, and hybrid architectures). We then focus on FEM developments at the electrode and device scales, emphasizing how these models capture transport phenomena, local microenvironments, and geometry-dependent effects. To clarify design principles, we present case studies of cascade electrodes organized in systems without and with integrated semiconductors. We further emphasize the integration of FEM with multiscale frameworks (density functional theory, molecular dynamics, kinetic Monte Carlo) and its role in bridging atomic-level insights with device-level performance. Finally, we identify current limitations and future prospects, including improved boundary conditions, coupling with operando experiments, and machine learning-accelerated model development. Together, these insights provide design principles for next-generation CO 2 R cascade systems for efficient solar fuel production.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Towards quantum machine learning with tensor networks

Machine learning is a promising application of quantum computing, but challenges remain for implementation today because near-term devices have a limited number of physical qubits and high error rates. Motivated by the usefulness of tensor networks for machine learning in the classical context, we propose quantum computing approaches to both discriminative and generative learning, with circuits based on tree and matrix product state tensor networks, that could already have benefits with such near-term devices. The result is a unified framework in which classical and quantum computing can benefit from the same theoretical and algorithmic developments, and the same model can be trained classically then transferred to the quantum setting for additional optimization. Tensor network circuits can also provide qubit-efficient schemes in which, depending on the architecture, the number of physical qubits required scales only logarithmically with, or independently of the input or output data sizes. Here, we demonstrate our proposals with numerical experiments, training a discriminative model to perform handwriting recognition using a hybrid quantum-classical optimization procedure that could be carried out on quantum hardware today, and testing the noise resilience of the trained model.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Resolving turbulent magnetohydrodynamics: a hybrid operator-diffusion framework

We present a hybrid machine learning framework that combines physics-informed neural operators (PINOs) with score-based generative diffusion models to simulate the full spatio-temporal evolution of two-dimensional, incompressible, resistive magnetohydrodynamic turbulence across a broad range of Reynolds numbers (Re). The framework leverages the equation-constrained generalization capabilities of PINOs to predict coherent, low-frequency dynamics, while a conditional diffusion model stochastically corrects high-frequency residuals, enabling accurate modeling of fully developed turbulence. Trained on a comprehensive ensemble of high-fidelity simulations with Re ϵ {100, 250, 500, 750, 1000, 3000, 10000}, the approach achieves state-of-the-art accuracy in regimes previously inaccessible to deterministic surrogates. At Re = 1000 and 3000, the model faithfully reconstructs the full spectral energy distributions of both velocity and magnetic fields late into the simulation, capturing non-Gaussian statistics, intermittent structures, and cross-field correlations with high fidelity. At extreme turbulence levels (Re = 10 000), it remains the first surrogate capable of recovering the high-wavenumber evolution of the magnetic field, preserving large-scale morphology and enabling statistically meaningful predictions.

Diffusion-Integrated Neural Operators↗

SVM-Based Synchronized Fault Detection for 100% Renewable Microgrids

Traditional protection schemes face significant challenges when applied to microgrids with high penetrations of renewables with inverter-based resources (IBRs). The proliferation of advanced sensing and communication technologies has generated copious data, offering an opportunity to overcome these limitations using data-driven machine learning approaches. This work proposes a novel approach based on a support vector machine (SVM) for detecting faults within a 100% renewable microgrid. The approach encompasses a systematic offline training stage for the development of a linear SVM-based fault detection algorithm. This process covers offline data collection from the microgrid under study, the extraction of features such as positive- and negative-sequence components and the total harmonic distortion of the voltage and current measurements of the relays, and the design of the linear SVM-based classifier. During the online implementation, however, different classifiers can exhibit asynchronicity in detecting the fault inception at different subcycle-to-cycle period-level delays. To circumvent this asynchronicity issue, a separate algorithm is developed for each relay to estimate the fault inception time as close to the real fault time. The performance of the proposed SVM-based synchronized fault detection method is evaluated using online time-domain simulation studies on a microgrid test system. The results corroborate the reliability of the fault detection scheme when tested under various fault cases (fault types, locations, and impedances) and non-fault cases during both grid-tied and islanded operation modes.

100% microgrid↗

Enhancing Diagnosis through AI-driven Analysis of Reflectance Confocal Microscopy

Reflectance Confocal Microscopy (RCM) is a non-invasive imaging technique used in biomedical research and clinical dermatology. It provides virtual high-resolution images of the skin and superficial tissues, reducing the need for physical biopsies. RCM employs a laser light source to illuminate the tissue, capturing the reflected light to generate detailed images of microscopic structures at various depths. Recent studies explored AI and machine learning, particularly CNNs, for analyzing RCM images. Our study proposes a segmentation strategy based on textural features to identify clinically significant regions, empowering dermatologists in effective image interpretation and boosting diagnostic confidence. This approach promises to advance dermatological diagnosis and treatment.

Yoon, Hong-Jun↗

Dynamic Importance Sampling

Dynamic Importance (DynIm) Sampling is a new approach for importance sampling in high-dimensional space. DynIm has two key characteristics. (1) Importance sampling: The notion of importance of candidate samples is used to guide the sampling, where importance is defined based on the (dis)similarity from previously selected samples (via Euclidean distance metric). (2) Dynamic sampling: The sampling can be performed dynamically, as new candidates are generated and new samples must be selected. DynIm was developed as part of a machine learning based coupling of scales in large multiscale simulations and can be used for a variety of applications.

Bhatia, Harsh↗