Search NASA⌕ Search

SEARCH · Search NASA

Results for “Sample handling”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

HHL algorithm with mapping function and enhanced sampling for model predictive control in microgrids

Here, this paper presents a refined quantum Harrow Hassidim Lloyd (HHL) algorithm for microgrid control. The first novelty of the developed method is that a mapping shift function enables the original HHL algorithm to handle general linear equations with non-singular and indefinite matrix. Second, a method of Matrix Extension for Amplifying Sampling Probabilities of Intended Solution (ME-ASPI) is proposed to design the reformulated linear algebraic equations, allowing for improved sampling efficiency of the quantum tomography in the refined HHL algorithm. Then, we applied the method to solve the model predictive control (MPC) problem in nonlinear dynamical microgrids. Specifically, with the ME-ASPI method, the refined HHL algorithm can effectively obtain the intended partial optimal control inputs for MPC. The optimization of quadratic programming problem in each time step of MPC is transformed into a linear system problem, which is addressed by the proposed quantum solver through using only partial information, with the time complexity improved from $\mathscr{O}(\mathscr{N}^{2.37286})$ classically to $\mathscr{O}(\mathscr{N}^{2} log \mathscr{N}$ x $p$ log $p)$ in quantum. Numerical examples have validated the effectiveness of the refined HHL algorithm with the proposed mapping function and the ME-ASPI method. By leveraging quantum properties, the proposed method provides a hybrid quantum–classical framework for microgrid control. This generic method can also potentially tackle many other challenges in analyzing and controlling general complex engineered systems.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

Sputter sample preparation for ion beam delivery of radium-223 at ATLAS

A radium-223 ion beam was delivered to an experiment from the electron cyclotron resonance ion source, ECR2, at the Argonne Tandem Linac Accelerator System (ATLAS). Here, the radium-223 material was in a nitrate salt form within a vial, prior to being converted to a usable sputter sample. The sputter sample was produced using a new sample preparation method, where the radium nitrate was dissolved into a solution and pipetted onto pressed aluminum powder. This sample was then allowed to dry, distributing the radium-223 material throughout the sputter sample. Ion source operation using the radium sputter sample is described with the operating parameters listed. The intensity and energy requirements for this ion beam were 1 × 10 6 particles/s and 1.07 GeV, respectively. Because the intensity is relatively low compared to most experiments at ATLAS, previously developed accelerator mass spectrometry methods were used Scott et al. to avoid the need for tuning of the low-intensity beam of interest. Handling of the radium material, as well as loading and unloading of the sputter sample from ECR2, required collaboration with Health Physics. Procedures were used and dry runs were carried out before, during, and after the experiment to ensure the safety of the workers. The processes used and lessons learned are described within.

07 ISOTOPE AND RADIATION SOURCES↗
DOE OSTI2023

BEYONDPLANCK II. CMB mapmaking through Gibbs sampling

We present a Gibbs sampling solution to the mapmaking problem for cosmic microwave background (CMB) measurements that builds on existing destriping methodology. Gibbs sampling breaks the computationally heavy destriping problem into two separate steps: noise filtering and map binning. Considered as two separate steps, both are computationally much cheaper than solving the combined problem. This provides a huge performance benefit as compared to traditional methods and it allows us, for the first time, to bring the destriping baseline length to a single sample. Here, we applied the Gibbs procedure to simulated Planck 30 GHz data. We find that gaps in the time-ordered data are handled efficiently by filling them in with simulated noise as part of the Gibbs process. The Gibbs procedure yields a chain of map samples, from which we are able to compute the posterior mean as a best-estimate map. The variation in the chain provides information on the correlated residual noise, without the need to construct a full noise covariance matrix. However, if only a single maximum-likelihood frequency map estimate is required, we find that traditional conjugate gradient solvers converge much faster than a Gibbs sampler in terms of the total number of iterations. The conceptual advantages of the Gibbs sampling approach lies in statistically well-defined error propagation and systematic error correction. This methodology thus forms the conceptual basis for the mapmaking algorithm employed in the BEYONDPLANCK framework, which implements the first end-to-end Bayesian analysis pipeline for CMB observations.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Efficient distributed continual learning for steering experiments in real-time

Deep learning has emerged as a powerful method for extracting valuable information from large volumes of data. However, when new training data arrives continuously (i.e., is not fully available from the beginning), incremental training suffers from catastrophic forgetting (i.e., new patterns are reinforced at the expense of previously acquired knowledge). Training from scratch each time new training data becomes available would result in extremely long training times and massive data accumulation. Rehearsal-based continual learning has shown promise for addressing the catastrophic forgetting challenge, but research to date has not addressed performance and scalability. To fill this gap, we propose an approach based on a distributed rehearsal buffer that efficiently complements data-parallel training on multiple GPUs to achieve high accuracy, short runtime, and scalability. It leverages a set of buffers (local to each GPU) and uses several asynchronous techniques for updating these local buffers in an embarrassingly parallel fashion, all while handling the communication overheads necessary to augment input minibatches using unbiased, global sampling. We further propose a generalization of rehearsal buffers to support both classification and generative learning tasks, as well as more advanced rehearsal strategies (notably Dark Experience Replay, leveraging knowledge distillation). We illustrate this approach with a real-life HPC streaming application from the domain of ptychographic image reconstruction. Furthermore, we run extensive experiments on up to 128 GPUs of the ThetaGPU supercomputer to compare our approach with baselines representative of training-from-scratch (the upper bound in terms of accuracy) and incremental training (the lower bound). Results show that rehearsal-based continual learning achieves a top-5 validation accuracy close to the upper bound, while simultaneously exhibiting a runtime close to the lower bound.

Asynchronous data management↗
DOE OSTI2020

Deposition of sulfate aerosols with positive Δ 33 S in the Neoarchean

Anomalous sulfur isotope compositions present in Archean rocks have been intensely scrutinized over the last 20 years because they record key aspects of Earth's atmospheric composition prior to the appearance of free molecular oxygen ca. 2.3 billion years ago. These isotopic compositions can be described as mass anomalous fractionations (MAF) and are produced in the atmosphere as UV light interacts with SO 2 molecules. Most interpretations suggest that atmospheric processes generate a reduced S-phase with a positive ( 33 S-enriched) MAF signature, as measured in pyrites, and an oxidized S-phase with a negative anomaly, as measured in bedded barite deposits. However, recent data for carbonate-associated sulfate (CAS) — a direct proxy for the isotopic composition of sulfur from seawater sulfate — in Neoarchean rocks showed no such negative values, but rather the opposite. To understand if the positive MAF anomalies we measured in Neoarchean CAS reflect secondary processes (diagenetic, metamorphic, handling) instead of original signals of Archean seawater sulfate, we collected additional sample suites with various degrees of preservation and metamorphic alteration across the Campbellrand-Malmani platform in South Africa. Results illustrate that within this comprehensive suite, less-altered samples all contain positive MAF values while secondary processes tend to either remove CAS from the sample and/or decrease the 33 S-enrichment. This positive MAF signal in sulfate is therefore reasonably interpreted as a primary depositional origin, and implies that the assumption that sulfate always carries a negative MAF anomaly throughout the Archean rock record needs to be reconsidered. Our CAS observations suggest that future experiments and calculations should also consider atmospheric and/or sulfur cycling processes that can produce oxidized sulfur with a positive MAF signature.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Interpreting T-History Data on Supercooling and Phase Change Enthalpy of a Salt-Hydrogel Complex for Building Thermal Management

Phase change materials (PCMs) are promising materials for storing and discharging latent heat in buildings to regulate the thermal environment. Measurement of salt hydrate PCM thermal performance is often challenging because of supercooling and phase segregation. Though differential scanning calorimetry (DSC) has been widely used to measure thermal properties of various phase change materials, DSC uses milligrams of sample and is unrealistic for predicting large-scale materials performance. The T-history method [1], in contrast, handles higher volumes but the analysis typically requires a lumped capacitance assumption. Here, we report and analyze T-History data on a novel salt-hydrogel complex with high viscosity, that renders low Biot number sample preparation extremely challenging. We re-examine the standard T-history method and compared against DSC data in this complex material. We report details of an experimental setup, built with necessary additions to a controlled environmental chamber. We carefully examine the validity of the lumped capacitance assumption and suggest a modified analysis using a computational approach for reliable data. Experimental parameters such as ramp rate and temperature range of the environmental chamber are considered via computational modeling to understand the effect of ramp rate on supercooling [2]. Finally, we report thermal cycling experiments on the salt-hydrogel complex. Our modified approach to measuring supercooling and enthalpy of fusion at large scales is important in better understanding the performance of phase-change materials at scale for building thermal storage. References: [1] Marin, Jose M, Belen Zalba, Luisa F Cabeza, and Harald Mehling. “Determination of Enthalpy Temperature Curves of Phase Change Materials with the Temperature-History Method: Improvement to Temperature Dependent Properties.” Measurement Science and Technology 14, no. 2 (February 1, 2003): 184–89. https://doi.org/10.1088/0957-0233/14/2/305. [2] Safari, A., R. Saidur, F.A. Sulaiman, Yan Xu, and Joe Dong. “A Review on Supercooling of Phase Change Materials in Thermal Energy Storage Systems.” Renewable and Sustainable Energy Reviews 70 (April 2017): 905–19. https://doi.org/10.1016/j.rser.2016.11.272

thermodynamics↗
DOE OSTI2026

The Art of Automation: Translating Electron Microscopy Workflows Into Automated Processes

Acquiring data using a scanning transmission electron microscope (STEM) is a complex, multi-step process. The intricacy of the process depends on the type of sample, composition of the material, desired results of the experiment, resolution requirement and other experimental factors. Each experiment presents unique complications, such as sample drift and contamination, that the microscopist must consider when acquiring data. All these challenges are handled fluidly and expertly by experienced microscopists, but to reach new levels of innovation in material development, including greater reproducibility, throughput, and precision, the automation of these workflows is essential. The initial phase of this work involved translating intuition-based workflows into discrete, programmable steps. Some common key stages in STEM workflows are the initial tuning, scanning the sample for areas of interest, and then acquiring the data. Each stage can be broken further into specific parameter adjustments, such as aberration correction and dwell time optimization, depending on the experiment. When deconstructing various experiments each step was assessed for automation feasibility based on the amount of real time operator decisions. There are steps that lend themselves to automation more readily than others, such as course focusing and sample screening, but there is potential for full automation of all stages with time. As an initial step, an automated montage routine was developed, allowing for the efficient acquisition of large portions of the sample without requiring continuous intervention from the operator. The automation of this small process of the procedure demonstrates the value of this capability. A major challenge in automation arises from discrepancies between commanded, reported and actual stage movements. Using systematic tests, stage movement was quantified. This error can be corrected algorithmically for more accurate workflows in the future. Expanding automation capabilities would result in larger, more efficient data acquisition which allows for more robust statistical analysis. Additionally, this work lays the groundwork for a closed loop system where machine learning algorithms would intake automatically acquired data and make real time decisions. By progressively automating this instrument, this work establishes the foundation for fully automated experimentation in transmission electron microscopy.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Stochastic projective splitting

Here, we present a new, stochastic variant of the projective splitting (PS) family of algorithms for inclusion problems involving the sum of any finite number of maximal monotone operators. This new variant uses a stochastic oracle to evaluate one of the operators, which is assumed to be Lipschitz continuous, and (deterministic) resolvents to process the remaining operators. Our proposal is the first version of PS with such stochastic capabilities. We envision the primary application being machine learning (ML) problems, with the method’s stochastic features facilitating “mini-batch” sampling of datasets. Since it uses a monotone operator formulation, the method can handle not only Lipschitz-smooth loss minimization, but also min–max and noncooperative game formulations, with better convergence properties than the gradient descent-ascent methods commonly applied in such settings. The proposed method can handle any number of constraints and nonsmooth regularizers via projection and proximal operators. We prove almost-sure convergence of the iterates to a solution and a convergence rate result for the expected residual, and close with numerical experiments on a distributionally robust sparse logistic regression problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Status Report on Development of a Cask to Enable Pulsed Neutron Characterization of Irradiated Fuel

We present the design of a shielding cask that allows pulsed neutron characterization of irradiated fuel rodlets prior to their destructive examination in hot cells. The goal is to provide complementary and informative measurements that will inform subsequent hot cell PIE examinations by identifying typical and atypical regions with respect to microstructure, tomographic data, or isotope densities. Time-of-flight neutron diffraction has the potential to offer efficient, non-destructive and non-contact microstructural characterization of irradiated fuel pins with spatial resolution of 1 mm3 to 1 cm3 while for energy-resolved neutron imaging (and by extension tomography) a resolution of 100 μm was demonstrated. The potential results include crystallinity vs. amorphous volumes and microstructural information such as phase compositions and textures from the diffraction data as well as distances (e.g. pellet to cladding), cracks, and isotopic distributions of minor actinides, fission products as well as fission gas partial pressures e.g. in the plenum from energy resolved neutron imaging. The capability will identify regions of atypical behavior and provide cost effective bulk characterization of entire samples prior to destructive evaluation in hot cells. This effort is a collaboration between LANL, INL, ORNL, and UC Berkeley. A key facet of the initiative is the ability to cost effectively ship irradiated fuels from INL to the pulsed spallation neutron source at LANSCE. The irradiated fuel would be loaded in the custom designed cask (designated RaMHaM) at INL hot cells which then could be shipped in a BRR Type B shipment container between INL and LANL. No hot cell is required after the loading of the sample at INL in the pathway laid out in this report, greatly simplifying handling at LANL.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2021

Progress Report on Graphite-Salt Interaction Studies (FY21)

This report summarizes the activities performed in FY21 to investigate the interactions between nuclear graphite and molten salts, such as FLiNaK. First, building on last year’s achievements, the team improved the procedure for measurements on pressurized salt intrusion in nuclear graphite and performed new experiments at variable pressures with specimens of different sizes and shapes. Additionally, the team actively provided suggestions and comments for ASTM D8091, Standard Guide for Impregnation of Graphite with Molten Salts , which came up for periodic revision in 2021. Second, the team expanded its capabilities to further characterize salt-impregnated graphite by procuring and installing a laser-induced breakdown spectroscopy (LIBS) instrument capable of producing a 3D chemical composition of a surface and subsurface region of specimens. A methodology for accurately analyzing information was developed, and a manuscript was submitted for publication. The LIBS instrument sample cell is sealed in controlled inert atmosphere, which is an advantage for using humidity-sensitive samples and Be-contaminated samples. This aligns with the team’s effort to expand the capabilities of handling FLiBe and Becontaining materials in two newly installed glovebox units. Third, the team performed preliminary tribology tests for pebble graphite wear in contact with stainless steel in dry state and in molten salt at high temperature. These tests are needed to establish the baseline for the wear of pebble graphite by friction against metallic walls. The friction and wear of pebbles in the fluoride salt-cooled hightemperature reactor are expected to generate dust, which is a concern for other components’ properties and safe reactor operation.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2025

Paper or Plastic? Multiscale Material Handling Properties of Two Model Municipal Solid Waste Streams

Purpose: Municipal Solid Waste (MSW) is a potentially valuable sustainable feedstock for fuel and chemical production due to its carbon-rich content and low cost. This study aims to assess the material handling properties of paperand plastic-rich MSW feedstocks to mitigate equipment failure and processing downtime. Methods: The material handling properties of crumbled MSW feedstocks were measured using apowder rheometer with mass flow hopper calculations to assess handling performance. Inverse gas chromatography was use to measure the surface energy differences between feedstocks. Electron microscopy and Raman spectroscopy was used to evaluate microscale features that may contribute to material handling differences. Results: Plastic and paper rich feedstocks crumbled to a nominal 2 mm particle size were observed to have similar flow and handling characteristics with reasonable hopper outlets. 2 mm plastic rich crumbles, with their higher bulk density, exhibited superior flow performance. By contrast, 4 mm material required significantly larger hopper outlets, indicating poor flowability. Paper rich and 4 mm plastic rich samples displayed broad particle size distributions, which contributed to particle interlocking, jamming, and other flow issues. Electron microscopy revealed that plastic rich samples were significantly smoother, enhancing their flowability compared to the rougher, paper rich materials. Conclusions: This study establishes critical material handling baselines for processing MSW as a viable feedstock for fuel and chemical production. The findings highlight the importance of optimizing particle size and feedstock composition to improve flowability and handling performance.

09 BIOMASS FUELS↗
DOE OSTI2021

Functional capacities of microbial communities to carry out large scale geochemical processes are maintained during ex situ anaerobic incubation

Mechanisms controlling CO 2 and CH 4 production in wetlands are central to understanding carbon cycling and greenhouse gas exchange. However, the volatility of these respiration products complicates quantifying their rates of production in the field. Attempts to circumvent the challenges through closed system incubations, from which gases cannot escape, have been used to investigate bulk in situ geochemistry. Efforts towards mapping mechanistic linkages between geochemistry and microbiology have raised concern regarding sampling and incubation-induced perturbations. Microorganisms are impacted by oxygen exposure, increased temperatures and accumulation of metabolic products during handling, storage, and incubation. We probed the extent of these perturbations, and their influence on incubation results, using high-resolution geochemical and microbial gene-based community profiling of anaerobically incubated material from three wetland habitats across a permafrost peatland. We compared the original field samples to the material anaerobically incubated over 50 days. Bulk geochemistry and phylum-level microbiota in incubations largely reflected field observations, but divergence between field and incubations occurred in both geochemistry and lineage-level microbial composition when examined at closer resolution. Despite the changes in representative lineages over time, inferred metabolic function with regards to carbon cycling largely reproduced field results suggesting functional consistency. Habitat differences among the source materials remained the largest driver of variation in geochemical and microbial differences among the samples in both incubations and field results. While incubations may have limited usefulness for identifying specific mechanisms, they remain a viable tool for probing bulk-scale questions related to anaerobic C cycling, including CO 2 and CH 4 dynamics.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Large Scale Caching and Streaming of Training Data for Online Deep Learning

The training of deep neural network models on large data remains a difficult problem, despite progress towards scalable techniques. In particular, there is a mismatch between the random but predetermined order in which AI flows select training samples and the streaming I/O patterns for which traditional HPC data storage (e.g., parallel file systems) are designed. In addition, as more data are obtained, it is feasible neither simply to train learning models incrementally, due to catastrophic forgetting (i.e., bias towards new samples), nor to train frequently from scratch, due to prohibitive time and/or resource constraints. In this paper, we study data management techniques that combine caching and streaming with rehearsal support in order to enable efficient access to training samples in both offline training and continual learning. We revisit state-of-art streaming approaches based on data pipelines that transparently handle prefetching, caching, shuffling, and data augmentation, and discuss the challenges and opportunities that arise when combining these methods with data-parallel training techniques. We also report on preliminary experiments that evaluate the I/O overheads involved in accessing the training samples from a parallel file system (PFS) under several concurrency scenarios, highlighting the impact of the PFS on the design of the data pipelines.

data pipelines↗
DOE OSTI2020

Scalable molecular dynamics on CPU and GPU architectures with NAMD

NAMD is a molecular dynamics program designed for high-performance simulations of very large biological objects on CPU- and GPU-based architectures. NAMD offers scalable performance on petascale parallel supercomputers consisting of hundreds of thousands of cores, as well as on inexpensive commodity clusters commonly found in academic environments. It is written in C++ and leans on Charm++ parallel objects for optimal performance on low-latency architectures. NAMD is a versatile, multipurpose code that gathers state-of-the-art algorithms to carry out simulations in apt thermodynamic ensembles, using the widely popular CHARMM, AMBER, OPLS, and GROMOS biomolecular force fields. Here, we review the main features of NAMD that allow both equilibrium and enhanced-sampling molecular dynamics simulations with numerical efficiency. We describe the underlying concepts utilized by NAMD and their implementation, most notably for handling long-range electrostatics; controlling the temperature, pressure, and pH; applying external potentials on tailored grids; leveraging massively parallel resources in multiple-copy simulations; and hybrid quantum-mechanical/molecular-mechanical descriptions. We detail the variety of options offered by NAMD for enhanced-sampling simulations aimed at determining free-energy differences of either alchemical or geometrical transformations and outline their applicability to specific problems. Last, we discuss the roadmap for the development of NAMD and our current efforts toward achieving optimal performance on GPU-based architectures, for pushing back the limitations that have prevented biologically realistic billion-atom objects to be fruitfully simulated, and for making large-scale simulations less expensive and easier to set up, run, and analyze. NAMD is distributed free of charge with its source code at www.ks.uiuc.edu.

high-performance computing↗
DOE OSTI2022

Analysis of sparse recovery for Legendre expansions using envelope bound

We provide novel sufficient conditions for the uniform recovery of sparse Legendre expansions using ℓ 1 minimization, where the sampling points are drawn according to orthogonalization (uniform) measure. So far, conditions of the form m ≳ Θ 2 s x log factors have been relied on to determine the minimum number of samples m that guarantees successful reconstruction of s-sparse vectors when the measurement matrix is associated to an orthonormal system. However, in case of sparse Legendre expansions, the uniform bound Θ of Legendre systems is so high that these conditions are unable to provide meaningful guarantees. Here, in this paper, we present an analysis which employs the envelop bound of all Legendre polynomials instead, and prove a new recovery guarantee for s-sparse Legendre expansions, m ≳ Θs 2 x log factors, which is independent of Θ. Arguably, this is the first recovery condition established for orthonormal systems without assuming the uniform boundedness of the sampling matrix. The key ingredient of our analysis is an extension of chaining arguments, recently developed in Bourgain and Chkifa et al., to handle the envelope bound. Furthermore, our recovery condition is proved via restricted eigenvalue property, a less demanding replacement of restricted isometry property which is perfectly suited to the considered scenario. Along the way, we derive simple criteria to detect good sample sets. Our numerical tests show that sets of uniformly sampled points that meet these criteria will perform better recovery on average.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Device for residue handling minimization with vacuum-assisted separations

A device for collection of residue from separation columns. A manifold provides for multiple sample ports that are fluidly connected to a vacuum port by at least one channel. Residue from the separation columns is pulled through the sample ports, into the channel, and then to the vacuum port. From the vacuum port the residue can be moved, under the force of the vacuum, directly to a drain or to a collection vessel connected with a drain.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

D2NO: Efficient handling of heterogeneous input function spaces with distributed deep neural operators

Neural operators have been applied in various scientific fields, such as solving parametric partial differential equations, dynamical systems with control, and inverse problems. However, challenges arise when dealing with input functions that exhibit heterogeneous properties, requiring multiple sensors to handle functions with minimal regularity. To address this issue, discretization-invariant neural operators have been used, allowing the sampling of diverse input functions with different sensor locations. However, existing frameworks still require an equal number of sensors for all functions. We propose a novel distributed approach to further relax the discretization requirements and solve the heterogeneous dataset challenges. Our method involves partitioning the input function space and processing individual input functions using independent and separate neural networks. A centralized neural network is used to handle shared information across all output functions. This distributed methodology reduces the number of gradient descent back-propagation steps, improving efficiency while maintaining accuracy. Here, we demonstrate that the corresponding neural network is a universal approximator of continuous nonlinear operators and present three numerical examples to validate its performance.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Uncertainty-aware particle segmentation for electron microscopy at varied length scales

Electron microscopy is indispensable for examining the morphology and composition of solid materials at the sub-micron scale. To study the powder samples that are widely used in materials development, scanning electron microscopes (SEMs) are increasingly used at the laboratory scale to generate large datasets with hundreds of images. Parsing these images to identify distinct particles and determine their morphology requires careful analysis, and automating this process remains challenging. In this work, we enhance the Mask R-CNN architecture to develop a method for automated segmentation of particles in SEM images. We address several challenges inherent to measurements, such as image blur and particle agglomeration. Moreover, our method accounts for prediction uncertainty when such issues prevent accurate segmentation of a particle. Recognizing that disparate length scales are often present in large datasets, we use this framework to create two models that are separately trained to handle images obtained at low or high magnification. By testing these models on a variety of inorganic samples, our approach to particle segmentation surpasses an established automated segmentation method and yields comparable results to the predictions of three domain experts, revealing comparable accuracy while requiring a fraction of the time. These findings highlight the potential of deep learning in advancing autonomous workflows for materials characterization.

36 MATERIALS SCIENCE↗