Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Point Optimization”

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
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

FCC feasibility studies: Impact of tracker- and calorimeter-detector performance on jet flavor identification and Higgs physics analyses

The ambitious physics program planned for the Future Circular Collider electron-positron phase imposes stringent constraints on detector performance. This study systematically investigates how different detector configurations impact jet flavor identification and their effects on high-profile physics analyses. Using Higgs boson coupling measurements and searches for invisible Higgs decays as benchmarks, we evaluate the sensitivity of these analyses to variations in tracker and calorimeter detector properties. We examine modifications to single-point resolution, material budget, silicon layer placement, and particle identification capabilities, quantifying their effects on flavor-tagging performance. Additionally, we present the first comprehensive study of Higgs-to-invisible decay detection using full detector simulation, providing insights for optimizing detector designs at lepton colliders.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Hiperclust

This software leverages transfer learning to analyze atom probe tomography (APT) data. It is trained on synthetic data and then applies this knowledge to predict the optimal number of clusters for a given APT dataset. Initially, the software used preliminary clustering to estimate the general structure of the data. Based on this, it provides suggestions for key parameters like minimum cluster size and minimum number of points. These parameters are critical for algorithms like HDBSCAN, ensuring accurate cluster formation without the need for trial-and-error testing. The software runs on High-Performance computing (HPC) systems, enabling fast, scalable analysis of large APT datasets, ultimately saving time and improving the reliability of clustering outcomes.

Tang, Yalei [Idaho National Laboratory (INL), Idah↗
DOE OSTI2024

Impact of Biofouling on Point Absorber Wave Energy Converter Performance and Control

Biofouling is a well-documented problem in naval engineering, but little is known about its effect on wave energy converter (WEC) performance. In this study, the software WEC-Sim is used to simulate the performance of a point absorber WEC that has been biofouled by “hard” species (e.g., mussels, barnacles) to varying degrees. Specifically, biofouling is assumed to change the nonlinear drag forces acting on the WEC, which have quantifiable effects on key performance characteristics such as optimal damping conditions, power, peak displacement, and peak velocity. The results of this analysis are then used to discuss strategies for WEC control as it relates to biofouling. Furthermore, the results show that average power production can decrease by as much as 15% with heavy biofouling and require an adjustment of the optimal control law by up to 20%.

Biofouling↗
DOE OSTI2020

Taming the fixed-node error in diffusion Monte Carlo via range separation

By combining density-functional theory (DFT) and wave function theory via the range separation (RS) of the interelectronic Coulomb operator, we obtain accurate fixed-node diffusion Monte Carlo (FN-DMC) energies with compact multi-determinant trial wave functions. In particular, we combine here short-range exchange-correlation functionals with a flavor of selected configuration interaction known as configuration interaction using a perturbative selection made iteratively (CIPSI), a scheme that we label RS-DFT-CIPSI. One of the take-home messages of the present study is that RS-DFT-CIPSI trial wave functions yield lower fixed-node energies with more compact multi-determinant expansions than CIPSI, especially for small basis sets. Indeed, as the CIPSI component of RS-DFT-CIPSI is relieved from describing the short-range part of the correlation hole around the electron-electron coalescence points, the number of determinants in the trial wave function required to reach a given accuracy is significantly reduced as compared to a conventional CIPSI calculation. Importantly, by performing various numerical experiments, we evidence that the RS-DFT scheme essentially plays the role of a simple Jastrow factor by mimicking short-range correlation effects, hence avoiding the burden of performing a stochastic optimization. Considering the 55 atomization energies of the Gaussian-1 benchmark set of molecules, we show that using a fixed value of mu = 0.5 bohr(-1) provides effective error cancellations as well as compact trial wave functions, making the present method a good candidate for the accurate description of large chemical systems.

Scemama, Anthony↗
DOE OSTI2025

Hydrogen Bond Benchmark: Focal‐Point Analysis and Assessment of DFT Functionals

We performed a hierarchical, convergent ab initio benchmark study and systematically analyzed the performance of density functional approximations for describing hydrogen bonds in small neutral, cationic, and anionic complexes, as well as in larger systems involving amide, urea, deltamide, and squaramide moieties. Focal point analyses (FPA), extrapolating to the ab initio limit, were carried out using correlated wave function methods up to CCSDT(Q) for the small complexes and CCSD(T) for the larger systems, together with correlation-consistent Gaussian basis sets up to the complete basis set limit. Optimized geometries and vibrational frequencies were obtained at the CCSD(T) level. The resulting FPA hydrogen-bond energies converge within a few tenths of a kcal mol −1 . These reference data were used to evaluate 60 density functionals (including 12 dispersion-corrected), spanning the local-density approximation (LDA), generalized gradient approximations (GGAs), meta-GGAs, hybrids, meta-hybrids, double-hybrids, and range-separated hybrids. Overall, the meta-hybrid M06-2X provides the best performance for both hydrogen bond energies and geometries, while the dispersion-corrected GGAs BLYP-D3(BJ) and BLYP-D4 also yield accurate hydrogen-bond data and can serve as cost-effective options for studying large and complex systems.

coupled cluster theory↗
DOE OSTI2025

Development of an interatomic potential for the Ta–Li system

A new interatomic potential for the Ta–Li system is introduced to facilitate the study of phase stability, mechanical properties and non-equilibrium dynamics after Li implantation in Ta. Here, this potential is based on a generalization of the embedded atom method (GEAM) and includes contributions from embedding energy, explicit two- and three-body interactions, and nonlocal many-body interaction terms. The parameters of the potential are optimized using energies and atomic forces for a wide range of configurations obtained from ab initio density functional theory (DFT) calculations. The potential is rigorously validated across a range of physical properties, including elastic constants, equations of state, phonon dispersion curves, point defect properties, and melting temperatures for different compositions. Although our potential is trained on a small dataset, its accuracy is comparable to that of available machine learning potentials for Li and Ta. Our simulations show that at temperatures below 500 K, Li atoms in Ta–Li alloys form clusters separated by Ta-rich domains, and we find no evidence of ordered phase formation. For Li concentrations below a few percent, Li atoms preferentially segregate to surfaces and grain boundaries. However, in alloys containing more than ~10% Li, the accumulation of Li in symmetric-tilt grain boundaries can lead to one of the following effects: formation of amorphous-like regions, changes in grain boundary structural units, or lateral movement of the grain boundary.

GEAM potential↗
DOE OSTI2025

Rational Electrolyte Design for Elevated-Temperature and Thermally Stable Lithium-Ion Batteries with Nickel-Rich Cathodes

As the energy density of lithium-ion batteries (LIBs) increases, the shortened cycle life and the increased safety hazard of LIBs are drawing increasing concerns. To address such challenges, a series of localized high-concentration electrolytes (LHCEs) based on a solvating-solvent mixture of tetramethylene sulfone and trimethyl phosphate and a high flash-point diluent 1H,1H,5H-octafluoropentyl 1,1,2,2-tetrafluoroethyl ether were designed. The LHCEs exhibited non-flammability and greatly suppressed heat release at high temperatures, which would potentially improve the safety performance of the LIBs. Moreover, the optimal LHCE achieved capacity retentions of 87.1% and 81.7% in graphite||LiNi 0.8 Mn 0.1 Co 0.1 O 2 cells after 500 cycles at 25 °C and 45 °C, respectively, which were significantly better than the conventional electrolyte, whose capacity retentions were only 75.2% and 38.5% under the same condition. Mechanistic studies revealed that the LHCE not only formed a more robust solid electrolyte interphase, but also exhibited improved anodic stability, compared with the conventional electrolyte. Further, this work sheds light in rational electrolyte design for high energy density LIBs with high battery performance and low safety concerns.

25 ENERGY STORAGE↗
DOE OSTI2025

Hepatocellular Metabolic Profile: Understanding Post-Thawing Metabolic Shift in Primary Hepatocytes In Vitro

Primary human hepatocytes (PHHs) are widely used as in vitro models for liver function and drug metabolism studies, yet their metabolic stability post-thawing remains an open question. To better characterize early metabolic changes, we conducted a time-course experiment using liquid chromatography-tandem mass spectrometry (LC-MS/MS) to analyze metabolic shifts in PHHs cultured in suspension. Unexposed and exposed (acetaminophen-treated) samples were evaluated, and TITAN analysis was applied to determine the time point of maximal metabolic change at both individual metabolite and global metabolic profile levels. Our results indicate that the majority of metabolic shifts occur within the first five hours post-thawing. In the early culture time points, substantial metabolic overlap was observed between unexposed and exposed cells, suggesting a conserved biological response likely related to cellular recovery. However, at later time points, metabolite profiles diverged, with acetaminophen treatment-specific metabolic changes emerging, potentially reflecting differences in homeostatic restoration versus hepatotoxic responses. Our study highlights the importance of considering early post-thawing metabolic dynamics in experimental design and offers insights for optimizing hepatocyte culture protocols to better replicate in vivo physiological conditions.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Rapid Screening of Liquid Metal Wetting for a Materials Compatibility Library

Wetting behavior of molten metals on solid substrates is a critical phenomenon influencing numerous industrial applications, including welding, anti-corrosion coatings, and metal additive manufacturing (AM). In particular, molten metal jetting (MMJ), an emerging AM technology, requires that the molten metal remain pinned at the nozzle exit. Thus, each new metal requires a specific nozzle material to ensure consistent droplet ejection and deposition, making it important to rapidly identify the appropriate wetting combinations. However, traditional measurements of wetting angles require expensive equipment and only allow one combination of materials to be investigated at a time which can be time consuming. This work introduces a rapid screening method based on sessile droplet experiments to evaluate wetting profiles across multiple metal–substrate combinations simultaneously. This study investigates the wetting interactions of molten Al alloy (Al4008), Cu, and Sn on various ceramic and metal substrates to identify optimal material combinations for MMJ nozzle designs. Results demonstrate that Al4008 achieves wetting on ceramic substrates such as AlN, TiO 2 , and SiC, with varying mechanisms including chemical reactions and weak surface interactions. Additionally, theoretical predictions regarding miscibility gaps and melting point differences were verified for Cu and Sn on refractory metals like Mo and W. Findings from this study contribute to the establishment of a materials compatibility library, enabling the selection of wetting/non-wetting combinations for stable MMJ operation. This resource not only advances MMJ technologies but also provides valuable insights for broader applications such as welding, coating, and printed electronics.

Materials science↗
DOE OSTI2023

Improving the Quality of Geothermal Data Through Data Standards and Pipelines Within the Geothermal Data Repository: Preprint

For machine learning outputs to be applicable to real world problems, high quality data are needed to ensure high quality results. With the more recent emphasis on machine learning in geothermal, there is an increasing need for greater focus on the quality of the data available for use in these projects. For example, Geothermal Operational Optimization Using Machine Learning (GOOML) utilized large quantities of geothermal power plant operational data to inform power plant operational configurations to maximize power generation. High quality datasets result from dependable sensors or devices collecting data, high frequency of measurements, sufficient data points, adequate metadata, reliable storage of data, and sufficient data curation. Another component that contributes to high quality data is reusability, which can be enhanced through data standardization. Data Standardization creates consistency in formatting and contents of like datasets, lessening preprocessing requirements and ensuring adequate information provided by a given dataset. The Geothermal Data Repository (GDR) aims to help improve data quality through automated data standardization for high-value datasets through the implementation of data pipelines alongside reliable and accessible long-term storage for datasets. As such, the GDR has decided to shift away from recommending the use of Excel-based content models and towards the implementation of automated data pipelines. This takes the burden of data standardization off the user and project team and will increase the availability of standardized geothermal data available through the GDR. A set of recommendations, or a data standard for each data type will exist with each data pipeline in order to advise data collection for maximum usability for future research. This paper serves to describe the GDR's proposed transition towards data standardization through automated data pipelines, to discuss the need for and value of such a shift, and to call for suggestions from the community regarding the most useful data standards and pipelines.

data↗
DOE OSTI2026

Dark Energy Survey Year 6 Results: MagLim++ Lens Sample Selection and Measurements of Galaxy Clustering

Galaxy clustering is a sensitive probe of the expansion history and growth of structure of the universe, and key degeneracies can be broken by combining these data with measurements of cosmic shear and galaxy-galaxy lensing (a so-called 3$\times$2pt analysis). The largest and least biased statistical samples of galaxies for use in clustering analyses can be collected photometrically through large imaging surveys. However, selecting clean photometric subsamples for cosmology are crucial for avoiding contamination that can bias cosmological constraints. Here we present the MagLim++ galaxy sample, selected to optimize for cosmological constraining power and incorporating an array of novel quality cuts to identify and remove residual contamination. This sample comes from the full six years of observations from the Dark Energy Survey. We present measurements of the two-point angular clustering ($w(θ)$) of 9,186,205 galaxies distributed over 4031 sq. degrees and in six tomographic redshift bins centered at $\bar{z}\approx$ [0.31, 0.44, 0.62, 0.78, 0.90, 1.01]. These measurements are used as part of the 3$\times$2pt and other DES Y6 legacy cosmological analyses in companion works. We describe the battery of null tests and mitigation schemes implemented to address observational, astrophysical, and methodological systematics in the analysis. The resulting $w(θ)$ measurements have a S/N = 149 (90.2 for linear scales only), which we use to place galaxy-clustering-only constraints on the matter density of the Universe, $Ω_m=0.311^{+0.023}_{-0.035}$, and amplitude of galaxy clustering in each redshift bin, $b_iσ_8=[1.16^{+0.04}_{-0.06},\ 1.40^{+0.04}_{-0.06},\ 1.57^{+0.04}_{-0.06},\ 1.59^{+0.04}_{-0.05},\ 1.50^{+0.04}_{-0.05},\ 1.74^{+0.06}_{-0.08}]$.

Weaverdyck, N. [BCCP, Berkeley; UC, Berkeley; LBL,↗
DOE OSTI2024

Degradation of Poly- and Perfluoroalkyl Substances (PFAS) in Water via High Power, Energy-Efficient Electron Beam Accelerator

The goal of the 2-year workplan was to see if electron beam (EB) could be used to break down a sub-set of the larger chemical family of per and polyfluoroalkylated substances (PFAS) in an energy efficient and economical manner when compared to conventional water treatment technologies. Year one (Y1) work focused on sample EB treatment work in the Fermi National Accelerator Laboratory’s (FNALs) Accelerator Applications Demonstration and Development (A2D2) EB accelerator. While there are reportedly thousands of types of PFAS, for the point of most of the work herein, a small subset was examined, typically perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA). PFOA and PFOS are two of the most well studied PFAS and are studied for baseline evaluations and are considered most useful. The work from Y1 provided information about the optimal operating parameters and additives to use when treating PFOS and PFOA via EB. The data were then used to see where in a water treatment system an EB accelerator would be best suited to treat PFAS. A conventional water treatment technology, GAC, was then compared to e-beam treatment technology with respect to energy and costs for treatment. In year two (Y2), several conventional e-beam accelerator designs, and FNAL’s developmental compact SRF accelerator design, were evaluated for their suitability in PFAS treatment, from an energy efficiency and cost standpoint. Several EB parameters were evaluated and optimized for the removal of PFOA and PFOS from water at normal pressure and temperature, measured as total PFAS removal. Under the optimized test conditions both PFOA showed complete destruction to inorganic fluoride, and PFOS to inorganic fluoride and sulfate, with mass balance. The effect on PFAS removal relative to solution pH, total EB dose, EB dose rate, dissolved oxygen concentration (DO), temperature, and initial PFAS concentration were evaluated. In general, PFOA was easier to destroy than PFOS. Degradation products, typically observed under less-than-optimal EB conditions, provided insight to degradation mechanisms. Products were identified to rule out possible deleterious biproduct formation. The water radiolysis radical reaction kinetics with PFOS and PFOA were not dependent on the initial concentration over 5-orders of magnitude from 2 μg/L to 20 mg/L. This is thought to be because there was an overabundance of the reactive water radiolysis radicals relative to PFAS molecules and largely attributed to aqueous electrons. The reaction rates appeared to be diffusion limited. Testing at higher concentrations (100-200 mg/L) showed a decrease in removal efficiency, suggesting alternative kinetics, possibly second order rates, at higher concentrations. In all, we successfully defined a set of optimal EB parameters to treat PFOA and PFOS at concentrations of 20 mg/L in water with destruction efficiencies near 100%. We further tested the optimized EB parameters with other types of PFAS, including shorter and longer fluorocarbon chain homologs of PFOA and PFOS, and PFAS with alternative functional groups such as sulfonamides. Based on our results EB can be optimized as an effective destructive technology for removing PFAS from water. The conditions optimized for PFOA and PFOS were less effective with ultra-short fluorocarbon compounds like TFMS, PFES, PFPS and PFBS, and likely require re-optimization of parameters to them. In all, it was determined that from a cost and energy efficiency standpoint, EB would be best applied to waste streams with relatively high concentrations of PFOS and PFOA and is not as cost effective as GAC treatment for removing low concentrations of PFAS from water. Higher concentrations of PFAS can be found in the wastewater of conventional treatment processes such as RO and IE and therefore EB may be used to supplement such treatment technologies. Some real-world IE regeneration wash water and RO reject water containing higher concentrations of PFAS and obtained from pilot scale industrial wastewater treatment system at a fluorochemical manufacturing facility, showed that EB could remove PFAS from such types of wastewaters. The IE regenerant wash water appeared to be the most efficient of the two types of wastewaters tested. However, some further optimization of the EB parameters for the specific PFAS types present in those wastewaters may be required. Also, the effects of co-present TOC and mineral salts should be considered during such optimization efforts. From the experimental Y1 results it was seen that the aqueous electron drives degradation of the PFAS. In a hypothetical water treatment skid using EB for PFAS destruction the parameters of the system should be optimized to promote aqueous electron production. Before EB treatment, the PFAS should be preconcentrated when possible, the pH should be raised to pH 10 or higher to enhance aqueous electron production, and the water should be nitrogen purged to remove dissolved oxygen to minimize aqueous electron scavenging. An excel spreadsheet was created that calculates optimal conditions based on inlet PFAS concentration and desired outlet concentration, by optimizing the accelerator power, dose rate, water treatment rate, pH and dissolved oxygen levels to reach the desired endpoint. Given this information on accelerator operating conditions five different EB accelerator systems were compared. One EB system was a continuous-wave, linear superconducting accelerator being designed at Fermilab. Three other EB systems (IMPELA at 5% and 25% duty factor and the ILU-14) were normal conducting pulsed linear accelerators. The fifth system was an IBA Rhodotron which is a normal conducting, circular, continuous-wave accelerator. The accelerator efficiency (% of the incoming power that is used in water treatment) was the dominating factor in accelerator choice. The radio frequency (RF) power supply and the accelerator design (superconducting versus warm technology) drive the accelerator efficiency. The IBA Rhodotron was seen to be the most energy efficient commercially available technology with a wall-plug (total) power efficiency of 43% at 400 kW. The Fermilab design, with a prototype for a different application currently being fabricated, was the most energy efficient at 55% when driven by a Klystron RF power supply and as high as 77% when powered by a magnetron. As the Fermilab design was the most energy efficient by approximately 10-30%, further design work was done on the accelerator and beam delivery system specific to the destruction of PFAS in water. The Fermilab design is unique from industrial accelerators in that is superconducting. Superconducting technology allows for the acceleration of electrons without losses. The accelerator must be cooled to below the point where it is superconducting and is operated around 4 degrees Kelvin. The bulk of the design work for the accelerator is on making the accelerator as energy efficient as possible so that it does not require liquid helium and can be cooled with conduction cooling via cryocoolers. Final design work resulted in an EB accelerator that would operate at minimally 200 kW and 10 MeV. Prototype construction would cost $\$ $7.8 million dollars when driven by a Klystron power supply. A second version of the same accelerator would cost $\$ $5.5 million dollars when driven by a magnetron that is still under development. The commercially available 300 kW IBA Rhodotron cost was estimated at approximately $\$ $9 million. While it is hard to directly compare, an operational GAC system used by 3M for groundwater treatment capital cost (2022 dollars) was estimated to cost $\$ $3.3 million. While the capital expense of the EB accelerator systems was higher than GAC, the accelerator EB treatment would result in destruction of the PFAS and not just sequestration of PFAS to form a new waste stream that requires further treatment or disposal. The operating cost to destroy the PFAS via 400 kw EB system was less than $\$ $1000/kg of PFAS destroyed when treating at a 20 mg/L PFAS concentration, compared to GAC with operating costs that calculated at $\$ $27,530 per kg of PFAS sequestered when treating 100 μg/L PFOA and PFOS combined concentration.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Modeling and Characterization of Surface Discharges in Insulating Material for Spacers: Electrode Shape, Discharge Mode, and Revision of the Creepage Concept

In the design of MV AC and DC spacers, the predominant factors are surface and interface conditions. Design is generally carried out on specifications and standards which are based on long-term experience and lab testing. However, the diffusion of power electronics with a trend to increase electric field, switching frequency, and rise time to achieve higher power density calls for an innovative, global approach to optimized insulation system design. A new methodology, based on field simulation, discharge modeling, and partial discharge inception measurements, called the three-leg approach, can form the basis to optimize insulation design for any type of supply voltage waveform. This paper focuses on the influence of the type of electrode on the inception and phenomenology of surface discharges and, as a consequence, on the interpretation of the results used for application of the three-leg approach. It is demonstrated that a typical electrode system used for insulating material testing can generate both gas and surface discharges at the triple point, when the electrodes have a smooth profile that is used to avoid corona or flashover. Hence, testing partial discharge may not provide a straightforward indication of the surface discharge inception and, thus, be partially misleading for insulation design. Another takeover is that such analysis must benefit from PD testing tools endowed with analytics able to provide automatic identification of the type of defect generating PD, i.e., internal, surface, and corona, since design and remedy actions can be taken, and adequate insulating materials developed, only knowing the type of source generating PD. Hence, testing partial discharge may not provide a straightforward indication of surface discharge inception and, thus, be partially misleading for insulation design. In addition to the importance of the three-leg approach to favor reliable and optimized design of insulation systems, there is a clear need to have a PD testing tool endowed with analytics. It should preferably be able to provide automatic identification of the type of defect generating PD, i.e., internal, surface, and corona.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Convex Optimization for Nonequilibrium Steady States on a Hybrid Quantum Processor

Finding the transient and steady state properties of open quantum systems is a central problem in various fields of quantum technologies. Here, in this work, we present a quantum-assisted algorithm to determine the steady states of open system dynamics. By reformulating the problem of finding the fixed point of Lindblad dynamics as a feasibility semidefinite program, we bypass several well-known issues with variational quantum approaches to solving for steady states. We demonstrate that our hybrid approach allows us to estimate the steady states of higher dimensional open quantum systems and discuss how our method can find multiple steady states for systems with symmetries.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Partial Least Squares, Experimental Design, and Near-Infrared Spectrophotometry for the Remote Quantification of Nitric Acid Concentration and Temperature

Near-infrared spectrophotometry and partial least squares regression (PLSR) were evaluated to create a pleasantly simple yet effective approach for measuring HNO3 concentration with varying temperature levels. A training set, which covered HNO3 concentrations (0.1–8 M) and temperature (10–40 °C), was selected using a D-optimal design to minimize the number of samples required in the calibration set for PLSR analysis. The top D-optimal-selected PLSR models had root mean squared error of prediction values of 1.4% for HNO 3 and 4.0% for temperature. The PLSR models built from spectra collected on static samples were validated against flow tests including HNO 3 concentration and temperature gradients to test abnormal conditions (e.g., bubbles) and the model performance between sample points in the factor space. Based on cross-validation and prediction modeling statistics, the designed near-infrared absorption approach can provide remote, quantitative analysis of HNO 3 concentration and temperature for production-oriented applications in facilities where laser safety challenges would inhibit the implementation of other optical techniques (e.g., Raman spectroscopy) and in which space, time, and/or resources are constrained. The experimental design approach effectively minimized the number of samples in the training set and maintained or improved PLSR model performance, which makes the described chemometric approach more amenable to nuclear field applications.

07 ISOTOPE AND RADIATION SOURCES↗
DOE OSTI2022

PemNet: A Transfer Learning-Based Modeling Approach of High-Temperature Polymer Electrolyte Membrane Electrochemical Systems

Widespread adoption of high-temperature electrochemical systems such as polymer electrolyte membrane fuel cells (HT-PEMFCs) requires models and computational tools for accurate optimization and guiding new materials for enhancing fuel cell performance and durability. Furthermore, while robust and better suited for extrapolation, knowledge-based modeling has limitations as it is time-consuming and requires information about the system that is not always available (e.g., material properties and interfacial behavior between different materials). Data-driven modeling, on the other hand, is easier to implement but often necessitates large datasets that could be difficult to obtain. In this contribution, knowledge-based modeling and data-driven modeling are combined by implementing a few-shot learning (FSL) approach. A knowledge-based model originally developed for a HT-PEMFCs was used to generate simulated data (887,735 points) and used to pretrain a neural network source model tuned via a genetic algorithm-based AutoML. Then, experimental datasets from HT-PEMFCs with different materials and operating conditions (~50 points each) were used to train six target models via FSL. Models for the unseen data reached high accuracies in all cases (rRMSE < 10%).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2021

Discovery of a weak topological insulating state and van Hove singularity in triclinic RhBi2

Abstract Time reversal symmetric (TRS) invariant topological insulators (TIs) fullfil a paradigmatic role in the field of topological materials, standing at the origin of its development. Apart from TRS protected strong TIs, it was realized early on that more confounding weak topological insulators (WTI) exist. WTIs depend on translational symmetry and exhibit topological surface states only in certain directions making it significantly more difficult to match the experimental success of strong TIs. We here report on the discovery of a WTI state in RhBi 2 that belongs to the optimal space group P $$\bar{1}$$ 1 ¯ , which is the only space group where symmetry indicated eigenvalues enumerate all possible invariants due to absence of additional constraining crystalline symmetries. Our ARPES, DFT calculations, and effective model reveal topological surface states with saddle points that are located in the vicinity of a Dirac point resulting in a van Hove singularity (VHS) along the (100) direction close to the Fermi energy ( E F ). Due to the combination of exotic features, this material offers great potential as a material platform for novel quantum effects.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Towards High-Performance AI4NP Applications on Modern GPU Platforms

The evolution of modern heterogeneous accelerators, such as GPUs, has significantly advanced the landscape of artificial intelligence (AI). There is a notable surge to adopt AI within the nuclear physics domain (AI4NP). While most AI4NP studies focus on feasibility analysis, our attention is directed towards evaluating their performance on contemporary GPUs that integrate tensor cores. We first benchmark the throughput of hyperparameterized multi-layer perceptron (MLP) models. We then examine the performance of an AI4NP application: Hydra. We assess the performance gain and accuracy loss caused by the tensor cores for low-precision floating-point operations. Our experiments encompass the PyTorch and TensorFlow Keras frameworks on NVIDIA’s T4 and A100 GPUs. We explore the behavior of different GPU hardware platforms and AI software tools. This study can be a valuable resource for guiding the performance optimization of larger-scale deployments of AI4NP applications.

Mei, Xinxin↗