Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Training Analysis”

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 1,639 records · Page 91
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Chapter 4: Physically informed deep learning networks for simulating microstructure evolution of 3D polycrystals

As discussed in the previous chapter, high energy diffraction microscopy (HEDM) is used to study the micromechanical evolution of a material during in situ loading. HEDM experiments have been used to verify crystal plasticity (CP) simulations [119, 91, 90, 120], for experimental planning, material design, and to further analyze experimental results. However, Fast Fourier transform-based CP (CP-FFT) or finite element-based CP (CP-FE) methods are often too slow to be used in real-time during an experiment. CP-FFT is faster than CP-FE simulations due to the absence of meshing, but can still take hours to simulate the response of a single volume depending on the size and number of strain steps [127]. Reducing computation time would create a larger exploration space in planning and design, and enable faster analysis of experimental results and real-time feedback during an experiment. This research expands upon previous works to develop a workflow for predicting the full-field evolution of a 3D polycrystal. The workflow is simplified from previous works to predict only orientation and elastic strain tensors (from which stress tensors are calculated). The network is physically informed through loss functions and network architecture for a more robust model. The orientation predictions are informed about the cubic crystal symmetry of the material by incorporating disorientation and misorientation information into the network architecture and loss. The Von Mises stress is used to enforce the correct stress-strain trends in the strain tensor predictions. Additional total strain steps from the elastic and elastoplastic region are included to better capture the stress-strain evolution at smaller total strain steps. Material and hardening parameters are additional inputs into the networks to further inform the network and to study the network’s ability to predict different materials other than those used for training.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Super-resolution and segmentation deep learning for breast cancer histopathology image analysis

Traditionally, a high-performance microscope with a large numerical aperture is required to acquire high-resolution images. However, the images’ size is typically tremendous. Therefore, they are not conveniently managed and transferred across a computer network or stored in a limited computer storage system. As a result, image compression is commonly used to reduce image size resulting in poor image resolution. Here, we demonstrate custom convolution neural networks (CNNs) for both super-resolution image enhancement from low-resolution images and characterization of both cells and nuclei from hematoxylin and eosin (H&E) stained breast cancer histopathological images by using a combination of generator and discriminator networks so-called super-resolution generative adversarial network-based on aggregated residual transformation (SRGAN-ResNeXt) to facilitate cancer diagnosis in low resource settings. The results provide high enhancement in image quality where the peak signal-to-noise ratio and structural similarity of our network results are over 30 dB and 0.93, respectively. The derived performance is superior to the results obtained from both the bicubic interpolation and the well-known SRGAN deep-learning methods. In addition, another custom CNN is used to perform image segmentation from the generated high-resolution breast cancer images derived with our model with an average Intersection over Union of 0.869 and an average dice similarity coefficient of 0.893 for the H&E image segmentation results. Finally, we propose the jointly trained SRGAN-ResNeXt and Inception U-net Models, which applied the weights from the individually trained SRGAN-ResNeXt and inception U-net models as the pre-trained weights for transfer learning. The jointly trained model’s results are progressively improved and promising. We anticipate these custom CNNs can help resolve the inaccessibility of advanced microscopes or whole slide imaging (WSI) systems to acquire high-resolution images from low-performance microscopes located in remote-constraint settings.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Latent Stochastic Differential Equations for Modeling Quasar Variability and Inferring Black Hole Properties

Quasars are bright and unobscured active galactic nuclei (AGN) thought to be powered by the accretion of matter around supermassive black holes at the centers of galaxies. The temporal variability of a quasar’s brightness contains valuable information about its physical properties. The UV/optical variability is thought to be a stochastic process, often represented as a damped random walk described by a stochastic differential equation (SDE). Upcoming wide-field telescopes such as the Rubin Observatory Legacy Survey of Space and Time (LSST) are expected to observe tens of millions of AGN in multiple filters over a ten year period, so there is a need for efficient and automated modeling techniques that can handle the large volume of data. Latent SDEs are machine learning models well suited for modeling quasar variability, as they can explicitly capture the underlying stochastic dynamics. In this work, we adapt latent SDEs to jointly reconstruct multivariate quasar light curves and infer their physical properties such as the black hole mass, inclination angle, and temperature slope. Our model is trained on realistic simulations of LSST ten year quasar light curves, and we demonstrate its ability to reconstruct quasar light curves even in the presence of long seasonal gaps and irregular sampling across different bands, outperforming a multioutput Gaussian process regression baseline. Our method has the potential to provide a deeper understanding of the physical properties of quasars and is applicable to a wide range of other multivariate time series with missing data and irregular sampling.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

A unified large language model–based framework for heterogeneous PV image diagnosis

With advances in imaging technologies, modern photovoltaic (PV) systems generate large volumes of heterogeneous image data, including visible, electroluminescence (EL), and infrared (IR) images. Existing PV image analysis models, particularly deep learning approaches, are typically task-specific and lack cross-modality generalization. To address this limitation, this paper proposes an open-source large language model (LLM)–based unified framework for heterogeneous PV image diagnostics. Through task-aware diagnostic prompting, the framework enables analysis of visible, EL, and IR images within a single pipeline, supporting both zero-shot and few-shot inference and binary and multiclass classification. It is compatible with state-of-the-art multimodal LLMs, including ChatGPT, Gemini, Claude, Qwen, and CLIP. The framework is evaluated on PV module condition classification (clean, soiling, snow, hail, and bird droppings) using visible images, cell crack detection using EL images, and hotspot detection using IR images. GPT-5.1 in few-shot mode achieves the best performance, with classification accuracy exceeding 97.3%. Open-source models such as Qwen and CLIP also deliver competitive results on visible images (around 90% accuracy), though their performance is more limited on EL and IR modalities. On the full ELPV dataset, the framework achieves 83.5% zero-shot accuracy, within 2.8% of the supervised CNN baseline, confirming scalability to larger benchmarks. Practical aspects such as reproducibility, response latency, and confidence estimation are systematically analyzed. The framework operates across PV image modalities without modality- or task-specific training, making it well suited as a rapid pre-screening tool to support downstream detailed diagnostics. A benchmark dataset of diverse labeled PV images is also released.

Li, Baojie↗
DOE OSTI2025

Meeting Global Health Needs via Infectious Disease Forecasting: Development of a Reliable Data-Driven Framework

Infectious diseases (IDs) have a significant detrimental impact on global health. Timely and accurate ID forecasting can result in more informed implementation of control measures and prevention policies. To meet the operational decision-making needs of real-world circumstances, we aimed to build a standardized, reliable, and trustworthy ID forecasting pipeline and visualization dashboard that is generalizable across a wide range of modeling techniques, IDs, and global locations. We forecasted 6 diverse, zoonotic diseases (brucellosis, campylobacteriosis, Middle East respiratory syndrome, Q fever, tick-borne encephalitis, and tularemia) across 4 continents and 8 countries. We included a wide range of statistical, machine learning, and deep learning models (n=9) and trained them on a multitude of features (average n=2326) within the One Health landscape, including demography, landscape, climate, and socioeconomic factors. The pipeline and dashboard were created in consideration of crucial operational metrics—prediction accuracy, computational efficiency, spatiotemporal generalizability, uncertainty quantification, and interpretability—which are essential to strategic data-driven decisions. While no single best model was suitable for all disease, region, and country combinations, our ensemble technique selects the best-performing model for each given scenario to achieve the closest prediction. For new or emerging diseases in a region, the ensemble model can predict how the disease may behave in the new region using a pretrained model from a similar region with a history of that disease. The data visualization dashboard provides a clean interface of important analytical metrics, such as ID temporal patterns, forecasts, prediction uncertainties, and model feature importance across all geographic locations and disease combinations. As the need for real-time, operational ID forecasting capabilities increases, this standardized and automated platform for data collection, analysis, and reporting is a major step forward in enabling evidence-based public health decisions and policies for the prevention and mitigation of future ID outbreaks.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

Randomized Algorithms for Scientific Computing (RASC)

Randomized algorithms have propelled advances in artificial intelligence (AI) and represent a foundational research area in advancing AI for Science. Future advancements in DOE Office of Science priority areas such as climate science, astrophysics, fusion, advanced materials, combustion, and quantum computing all require randomized algorithms for surmounting challenges of complexity, robustness, and scalability. Advances in data collection and numerical simulation have changed the dynamics of scientific research and motivate the need for randomized algorithms. For instance, advances in imaging technologies such as X-ray ptychography, electron microscopy, electron energy loss spectroscopy, or adaptive optics lattice light-sheet microscopy collect hyperspectral imaging and scattering data in terabytes, at breakneck speed enabled by state-of-the-art detectors. The data collection is exceptionally fast compared with its analysis. Likewise, advances in high-performance architectures have made exascale computing a reality and changed the economies of scientific computing in the process. Floating-point operations that create data are essentially free in comparison with data movement. Thus far, most approaches have focused on creating faster hardware. Ironically, this faster hardware has exacerbated the problem by making data still easier to create. Under such an onslaught, scientists often resort to heuristic deterministic sampling schemes (e.g., low-precision arithmetic, sampling every nth element) and sacrifice potentially valuable accuracy. Dramatically better results can be achieved via randomized algorithms, reducing the data size as much as or more than naive deterministic subsampling can achieve, while retaining the high accuracy of computing on the full data set. By randomized algorithms we mean those algorithms that employ some form of randomness in internal algorithmic decisions to accelerate time to solution, increase scalability, or improve reliability. Examples include matrix sketching for solving large-scale least-squares problems (see Figure 1) and stochastic gradient descent for training machine learning models. We are not recommending heuristic methods but rather randomized algorithms that have certificates of correctness and probabilistic guarantees of optimality and near-optimality. Such approaches can be useful beyond acceleration, for example, in understanding how to avoid measure zero worst-case scenarios that plague methods such as QR matrix factorization.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

A Comprehensive Review of Latent Space Dynamics Identification Algorithms for Intrusive and Non-Intrusive Reduced-Order-Modeling

Numerical solvers of partial differential equations (PDEs) have been widely employed for simulating physical systems. However, the computational cost remains a major bottleneck in various scientific and engineering applications, which has motivated the development of reduced-order models (ROMs). Recently, machine-learning-based ROMs have gained significant popularity and are promising for addressing some limitations of traditional ROM methods, especially for advection dominated systems. In this chapter, we focus on a particular framework known as Latent Space Dynamics Identification (LaSDI), which transforms the high-fidelity data, governed by a PDE, to simpler and low-dimensional latent-space data, governed by ordinary differential equations (ODEs). These ODEs can be learned and subsequently interpolated to make ROM predictions. Each building block of LaSDI can be easily modulated depending on the application, which makes the LaSDI framework highly flexible. In particular, we present strategies to enforce the laws of thermodynamics into LaSDI models (tLaSDI), enhance robustness in the presence of noise through the weak form (WLaSDI), select high-fidelity training data efficiently through active learning (gLaSDI, GPLaSDI), and quantify the ROM prediction uncertainty through Gaussian processes (GPLaSDI). We demonstrate the performance of different LaSDI approaches on Burgers equation, a non-linear heat conduction problem, and a plasma physics problem, showing that LaSDI algorithms can achieve relative errors of less than a few percent and up to thousands of times speed-ups.

Computational Engineering, Finance, and Science (c↗
DOE OSTI2023

Controlling Molecular Structure and Spin with Multiconfigurational Quantum Chemistry (Final Technical Report)

For many first-row transition metal complexes, structure-property relationships can be obtained from high-level molecular geometry optimizations and subsequent electronic structure studies. The project funded under this award utilized newly implemented fully internally contracted (FIC) nuclear gradients for extended multi-state (XMS) complete active space second-order multireference perturbation theory (CASPT2) to explore geometry changes in first-row transition metal coordination complexes. At the start of the project, only one full geometry optimization using FIC-CASPT2 analytical gradients had been reported (J. Chem. Theory Comput., 2016, 12 (8), 3781), and many open-questions regarding the performance and achievable accuracy in applying such computations to larger complexes persisted. Key deliverables in the report include the implementation of the numerical Hessian and subsequent vibrational analysis in the BAGEL program package. This includes both full Hessian vibrational analysis (FHVA) and a partial Hessian vibrational analysis (PHVA). The latter of which allows us to reduce the significant cost of computing the vibrational frequencies in a large molecule by focusing on the modes of highest interest. The gradient code also proved useful in the development of an approach to improve the Hubbard-U correction by removing bias in predicting spin-splitting in Fe(II) complexes via plane-wave density functional theory (DFT). Finally, we showed for three families of complexes (spin-crossover (SCO) complexes, metallocorroles complexes, and those with metal-metal bonds) that CASPT2 can result in good molecular geometries and established best practices in undertaking this work. We are now using this approach in collaborative projects with experimental groups, which would not have been possible at the start of this project. A common theme also arose through this work that static and dynamic correlation must be recovered in a balanced way to yield quantitative results. By systematically varying how both effects were included, we were able to make key chemical insights. This work supported the training of two postdoctoral scholars, one graduate student, and one undergraduate student in applying high-level wavefunction based methods to challenging systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Reconstructing porous media using generative flow networks

One area of intense scientific interest for the study of sandstones, carbonates, and shale at the pore scale is the use of limited image and petrophysical data to generate multiple realizations of a rock’s pore structure. Such images aid efforts to quantify uncertainty in petrophysical properties, including porosity–permeability transforms. We develop and evaluate a deep learning-based method to synthesize porous media volumes using a so-called generative flow model trained on x-ray microscope images of rock texture and pore structure. These models are optimized on a log-likelihood objective and they synthesize large and realistic three-dimensional images. Further, we demonstrate the rapid generation of sandstone image volumes that display realism as gauged by quantitative comparison of topological features using Minkowski functionals of porosity, specific surface area, and the Euler–Poincaré characteristic (i.e., pore connectivity). We also evaluate the single-phase permeability using Navier–Stokes and lattice Boltzmann methods and show that transport properties of the generated samples match measured trends.

58 GEOSCIENCES↗
DOE OSTI2022

The Geothermal Artificial Intelligence for geothermal exploration

Exploration of geothermal resources involves analysis and management of a large number of uncertainties, which makes investment and operations decisions challenging. Remote Sensing (RS), Machine Learning (ML) and Artificial Intelligence (AI) have potential in managing the challenges of geothermal exploration. In this paper, we present a methodology that integrates RS, ML and AI to create an initial assessment of geothermal potential, by resorting to known indicators of geothermal areas namely mineral markers, surface temperature, faults and deformation. We demonstrated the implementation of the method in two sites (Brady and Desert Peak geothermal sites) that are close to each other but have different characteristics (Brady having clear surface manifestations and Desert Peak being a blind site). Here, we processed various satellite images and geospatial data for mineral markers, temperature, faults and deformation and then implemented ML methods to obtain pattern of surface manifestation of geothermal sites. We developed an AI that uses patterns from surface manifestations to predict geothermal potential of each pixel. We tested the Geothermal AI using independent data sets obtaining accuracy of 92-95%; also tested the Geothermal AI trained on one site by executing it for the other site to predict the geothermal / non-geothermal delineation, the Geothermal AI performed quite well in prediction with 72-76% accuracy.

15 GEOTHERMAL ENERGY↗
DOE OSTI2023

Securely Aggregated Coded Matrix Inversion

Coded computing is a method for mitigating straggling workers in a centralized computing network, by using erasure-coding techniques. Federated learning is a decentralized model for training data distributed across client devices. In this work we propose approximating the inverse of an aggregated data matrix, where the data is generated by clients; similar to the federated learning paradigm, while also being resilient to stragglers. To do so, we propose a coded computing method based on gradient coding. We modify this method so that the coordinator does not access the local data at any point; while the clients access the aggregated matrix in order to complete their tasks. Here, the network we consider is not centrally administrated, and the communications which take place are secure against potential eavesdroppers.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Feature Extraction for Pipeline Defects Inspection Based Upon Distributed Acoustic Fiber Optic Sensing Data

Fiber-optic distributed acoustic sensing (DAS) is becoming an increasingly important tool for real-time monitoring of energy and civil infrastructure structural health such as pipelines. We present a systematic theoretical study of the potential for DAS to be directly coupled with guided ultrasonic waves typically used in conventional acoustic non-destructive evaluation (NDE) methods for real-time pipeline health monitoring. We are referring to this innovative new NDE technique as ultrasonic guided wave and optical fiber sensor fusion. In the practical application of DAS coupled with guided ultrasonic waves, the structural design of (1) the specific guided waves excited, (2) the physical installation of the acoustic transducers and the fiber optic sensors, and (3) the functional performance specifications (gauge length, sensitivity, Etc.) of fiber optic DAS have an important influence on overall capabilities of the monitoring system. Meanwhile, physics-based analysis of acoustic waves is still a challenge due to the complex nature of the Lamb wave when it propagates, scatters, and disperses in the presence of structural defects. In this work, we simulate carbon steel pipes relevant for oil and gas pipeline applications with diameters of approximately 6-12” and wall thickness of 0.5” as the objects to be monitored. By establishing and implementing these capabilities, we seek to pursue an in-depth study on structural parameter optimization of DAS network, measurement range, and signal processing with an ultimate goal of increasing the sensitivity and efficacy of DAS to defect identification for various modes of corrosion expected in practice. To study the characteristics of scattered acoustic waves and performance of DAS for defect identification, we simulated the response of DAS for multiple pipe structures, defect types, and DAS sensor network configuration using finite element software Ansys, then the properties of signal response are extracted to construct defect-sensitive features. The raw data simulated, and the associated features extracted can ultimately be utilized as annotated training data to benchmark various designs for DAS applications, guided acoustic excitation sources, and learning model parameters to enhance early detection of potentially problematic defects.

Pipeline Defects Inspection, Fiber-optic sensors, ↗
DOE OSTI2024

Rare Earth Elements (REE) and Critical Minerals (CM) in Middle Pennsylvanian-Age Coals and Associated Sediments in the Central Appalachian Basin, Eastern U.S.A.

The Central Appalachian Basin (CAB) of Kentucky, Tennessee, West Virginia and Virginia has a long history of coal mining and oil and gas extraction that has empowered the regional and national economies, the development of infrastructure, and a highly trained energy resources work force. As our societal demands for advanced technologies have rapidly increased in recent years, coal-related materials are viewed as an important new unconventional domestic source of critical minerals (CM) that are required for telecommunications, aerospace and transportation industries, electronics, the transition to low-carbon emissions energy production, and many consumer products. Coal-related materials encompass coal, associated sediments, coal mining waste materials, produced waters, and ash residues from coal-fired power plants. An important objective of the Evolve Central Appalachia Project (Evolve CAPP), sponsored by the U.S. Department of Energy (DOE) National Energy Technology Laboratory (NETL) is to assess the quantity and distribution of CM resources in the CAB region. The rare earth elements (REE) are considered highest priority, although other important CM such as niobium, gallium, and zirconium are known to occur in the Middle Pennsylvanian-age coals and sediments. Working with coal industry partners who provided access to drill cores, coal-related sediments and waste materials, over 600 samples have been collected for laboratory analysis, and over 730 materials have been scanned using portable x-ray fluorescence (pXRF) equipment. The application of pXRF provides the means for real-time semiquantitative analysis of CM content at very close spacings (typically 2-3 inch intervals) along drill core and in-situ channel samples that span the roof, coal seam, and floor rock. The comparison of pXRF geochemical data with laboratory results, geologic data, and downhole spectral gamma logs can provide high resolution input to lithologic and depositional models for future CM resource evaluations. The preliminary findings show that pXRF is capable of accurately measuring low concentrations of many of the CM with a high level of confidence (Ba, Cr, Ga, K, Nb, Rb, Sr, Th, Y), whereas for others (La, Ce, Co, Mn, Nd, Ni, Sc, Ti, V) the detection limits are very high or spectral interferences increase the uncertainty. Notably, the mean abundances of Y (34 ppm), La (86 ppm), Ga (40 ppm), and V (146 ppm) in coal underclays in the CAB region are up to 7X enriched compared with the overlying coal. These values also exceed the reported concentrations in published reference materials for upper continental crustal rocks (Rudnick and Gao, 2003), North American Shale Composite (Gromet et al., 1984), and North American coal (Finkelman, 1993). The pXRF data are in part verified by laboratory results that indicate the mean Y abundance (36 ppm) is highly correlated (R2 = 0.807) with total REE (ΣREE). The correlation is even higher (R2 = 0.957) with heavy REE (ΣHREE). Applying these correlations to the pXRF data for the coal underclays, the mean estimated values for ΣREE+Y and ΣHREE+Y are 270 ppm and 58 ppm, respectively. Although these average values are not considered high, the range of Y measured by pXRF in the coal underclays extended as high as 114 ppm, which would suggest ΣREE+Y equal to 847 ppm. The mean abundances of Zr (192 ppm) and Th (21 ppm) in coal underclays are also enriched compared with the overlying coal and these results likely reflect the presence of resistant detrital heavy minerals such as monazite, xenotime, and zircon in the underclay matrix. Several of the profiled coal seams and associated wall rocks contained thin volcanic ash layers up to 4-5 inches in thickness. The extent to which these ash fall layers provided a source for CM under the paleoenvironmental conditions that resulted in coal deposits in the CAB remains to be fully studied. Continuing investigations in the Evolve CAPP study area will include laboratory determinations of mineralogic and clay compositions, and evaluations of CM geochemical mobility in the coal and coal underclays.

Lassetter, Billy↗
DOE OSTI2026

Validated Reactive Force Field Quantifies MXene Interfacial Properties, Mechanics, and Thermal Transport

MXenes combine rich surface chemistry, mechanical strength, and high conductivity for a multitude of emerging applications. Predictive modeling supports accelerated materials designs and has been limited by the absence of validated and transferable force fields. Here, we introduce an interpretable, reactive INTERFACE force field (IFF and IFF-R) for Ti 3 C 2 T x MXenes that is trained based on chemical knowledge and achieves quantitative agreement with experiments across lattice parameters (<0.5%), density (<0.2%), liquid contact angles, Raman spectra, and the in-plane elastic modulus (∼320 GPa). The models cover surface terminations from hydroxyl (−OH) to fluorine (−F) groups and are extensible to other chemistries. We introduce pH-resolved surface chemistry and identify dopamine adsorption mechanisms at MXene–aqueous interfaces supported by QCM-D and UV–Vis experiments. The data reveal coplanar and perpendicular binding modes and concentration-dependent multilayer assembly. We predict previously inaccessible properties, including termination-dependent cleavage energies, interlayer shear moduli and dynamic shear failure, nanoindentation and brittle fracture, anisotropic in-plane and out-of-plane thermal conductivities, including the role of defects. Agreement with available experimental data is consistently close and exceeds DFT accuracy across the benchmark properties examined. The IFF/IFF-R model is compatible with CHARMM, AMBER, OPLS, and CVFF force fields for simulations of MXenes with diverse surface terminations, electrolyte interfaces, biointerfaces, and polymer composites without additional parameters. Parameter sets, 3D models, and analysis scripts are provided for community use. The validated, reactive, and transferable IFF framework facilitates predictive design of MXene-based films, membranes, sensing interfaces, and composites.

MXene↗
DOE OSTI2022

Computational evolution of high-performing unfused non-fullerene acceptors for organic solar cells

Materials optimization for organic solar cells (OSCs) is a highly active field, with many approaches using empirical experimental synthesis, computational brute force to screen a subset of chemical space, or generative machine learning methods that often require significant training sets. While these methods may find high-performing materials, they can be inefficient and time-consuming. Genetic algorithms (GAs) are an alternative approach, allowing for the “virtual synthesis” of molecules and a prediction of their “fitness” for some property, with new candidates suggested based on good characteristics of previously generated molecules. In this work, a GA is used to discover high-performing unfused non-fullerene acceptors (NFAs) based on an empirical prediction of power conversion efficiency (PCE) and provides design rules for future work. The electron-withdrawing/donating strength, as well as the sequence and symmetry, of those units are examined. The utilization of a GA over a brute-force approach resulted in speedups up to 1.8 × 10 12 . New types of units, not frequently seen in OSCs, are suggested, and in total 5426 NFAs are discovered with the GA. Of these, 1087 NFAs are predicted to have a PCE greater than 18%, which is roughly the current record efficiency. While the symmetry of the sequence showed no correlation with PCE, analysis of the sequence arrangement revealed that higher performance can be achieved with a donor core and acceptor end groups. Future NFA designs should consider this strategy as an alternative to the current A-D-A'-D-A architecture.

14 SOLAR ENERGY↗
DOE OSTI2023

Machine-Learning-aided Approach for Predicting the Thermal Expansion Behaviors in Advanced Test Reactor Capsules (NURETH-20 full paper)

Instrumented experiments at test reactors are essential to deploying new advanced reactor systems. Designing new experiments and generating data on specific conditions require both time and cost investment. A high-fidelity model of the experiment environment can be created using finite element analysis software to support the actual experiments, but computation time is still a concern in applying outcomes to real-time usage (e.g., a digital twin). This research proposes a machine-learning-aided approach to temperature and displacement predictions, based on the thickness of the outer gas gap on the experimental capsule used for the in-pile demonstration of a novel thermal conductivity probe in the Advanced Test Reactor. The capsule consisted of U10Zr fuel, a rodlet, sodium, and inner and outer capsules. There were gas gaps between the fuel and rodlet and between the inner and outer capsule. The learning data consisted of an experimental capsule’s radial distributions of temperature and displacement, as obtained from Abaqus and the physical features. For the first step, temperature was predicted using three positional parameters. Then the displacement was predicted using six different positional parameters. Each physical feature was normalized to be both nondimensional and standardized. The temperature and displacement predictions showed good agreement in all cases involving interpolation and extrapolation. Also, data similarity enhancement increased the similarity between training and target data increasing the predictive accuracy of machine-learning models. In some cases of extrapolation, the accuracy of the machine-learning model showed limited performance, but still data similarity enhancement improved the accuracy.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2022

Predicting Rare Earth Element Potential in Produced and Geothermal Waters of the United States via Emergent Self-Organizing Maps

This work applies emergent self-organizing map (ESOM) techniques, a form of machine learning, in the multidimensional interpretation and prediction of rare earth element (REE) abundance in produced and geothermal waters in the United States. Visualization of the variables in the ESOM trained using the input data shows that each REE, with the exception of Eu, follows the same distribution patterns and that no single parameter appears to control their distribution. Cross-validation, using a random subsample of the starting data and only using major ions, shows that predictions are generally accurate to within an order of magnitude. Using the same approach, an abridged version of the U.S. Geological Survey Produced Waters Database, Version 2.3 (which includes both data from produced and geothermal waters) was mapped to the ESOM and predicted values were generated for samples that contained enough variables to be effectively mapped. Results show that in general, produced and geothermal waters are predicted to be enriched in REEs by an order of magnitude or more relative to seawater, with maximum predicted enrichments in excess of 1000-fold. Cartographic mapping of the resulting predictions indicates that maximum REE concentrations exceed values in seawater across the majority of geologic basins investigated and that REEs are typically spatially co-associated. The factors causing this co-association were not determined from ESOM analysis, but based on the information currently available, REE content in produced and geothermal waters is not directly controlled by lithology, reservoir temperature, or salinity.

Engle, Mark A. (ORCID:0000000152587374)↗
DOE OSTI2021

Adaptive machine learning for time-varying systems: low dimensional latent space tuning

Machine learning (ML) tools such as encoder-decoder convolutional neural networks (CNN) can represent incredibly complex nonlinear functions which map between combinations of images and scalars. For example, CNNs can be used to map combinations of accelerator parameters and images which are 2D projections of the 6D phase space distributions of charged particle beams as they are transported between various particle accelerator locations. Despite their strengths, applying ML to time-varying systems, or systems with shifting distributions, is an open problem, especially for large systems for which collecting new data for re-training is impractical or interrupts operations. Particle accelerators are one example of large time-varying systems for which collecting detailed training data requires lengthy dedicated beam measurements which may no longer be available during regular operations. We present a novel method of adaptive ML for time-varying systems. Our approach is to map very high (N ≈ 100k) dimensional inputs (a combination of scalar parameters and images) into the low dimensional (N ≈ 2) latent space at the output of the encoder section of an encoder-decoder CNN. We then actively tune the low dimensional latent space-based representation of complex system dynamics by the addition of an adaptively tuned feedback vector directly before the decoder sections builds back up to our image-based high-dimensional phase space density representations. This method allows us to learn correlations within and to quickly tune the characteristics of incredibly large parameter space systems and to track their evolution in real time based on feedback without massive new data sets for re-training. We demonstrate that our method can accurately predict and track the phase space of charged particle beams at various locations in a particle accelerator by adaptively adjusting in real-time while the unknown input beam distribution of the accelerator is changing in shape, charge, and offset and while the RF system of the accelerator itself is also changing in an unpredictable way. For FACET-II we demonstrate that such an approach has the potential to use transverse deflecting cavity and energy spread spectrum beam measurements to accurately predict 2D projections of the 6D phase space of the electron beam at the plasma wakefield acceleration interaction point where such diagnostics are unavailable.

47 OTHER INSTRUMENTATION↗