Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “dl”

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 343 records · Page 19
All sourcesNASA NTRSDOE OSTI
DOE OSTI2015

STOCK: Structure mapper and online coarse-graining kit for molecular simulations

We present a web toolkit STructure mapper and Online Coarse-graining Kit for setting up coarse-grained molecular simulations. The kit consists of two tools: structure mapping and Boltzmann inversion tools. The aim of the first tool is to define a molecular mapping from high, e.g. all-atom, to low, i.e. coarse-grained, resolution. Using a graphical user interface it generates input files, which are compatible with standard coarse-graining packages, e.g. VOTCA and DL_CGMAP. Our second tool generates effective potentials for coarse-grained simulations preserving the structural properties, e.g. radial distribution functions, of the underlying higher resolution model. The required distribution functions can be provided by any simulation package. Simulations are performed on a local machine and only the distributions are uploaded to the server. The applicability of the toolkit is validated by mapping atomistic pentane and polyalanine molecules to a coarse-grained representation. Effective potentials are derived for systems of TIP3P (transferable intermolecular potential 3 point) water molecules and salt solution. The presented coarse-graining web toolkit is available at http://stock.cmm.ki.si.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2022

Fast and Accurate Predictions of Total Energy for Solid Solution Alloys with Graph Convolutional Neural Networks

We use graph convolutional neural networks (GCNNs) to produce fast and accurate predictions of the total energy of solid solution binary alloys. GCNNs allow us to abstract the lattice structure of a solid material as a graph, whereby atoms are modeled as nodes and metallic bonds as edges. This representation naturally incorporates information about the structure of the material, thereby eliminating the need for computationally expensive data pre-processing which would be required with standard neural network (NN) approaches. We train GCNNs on ab-initio density functional theory (DFT) for copper-gold (CuAu) and iron-platinum (FePt) data that has been generated by running the LSMS-3 code, which implements a locally self-consistent multiple scattering method, on OLCF supercomputers Titan and Summit. GCNN outperforms the ab-initio DFT simulation by orders of magnitude in terms of computational time to produce the estimate of the total energy for a given atomic configuration of the lattice structure. We compare the predictive performance of GCNN models against a standard NN such as dense feedforward multi-layer perceptron (MLP) by using the root-mean-squared errors to quantify the predictive quality of the deep learning (DL) models. We find that the attainable accuracy of GCNNs is at least an order of magnitude better than that of the MLP.

Lupo Pasini, Massimiliano↗
DOE OSTI2022

Secure Collaborative Environment for Seamless Sharing of Scientific Knowledge

In a secure collaborative environment, tera-bytes of data generated from powerful scientific instruments are used to train secure machine learning (ML) models on exascale computing systems, which are then securely shared with internal or external collaborators as cloud-based services. Devising such a secure platform is necessary for seamless scientific knowledge sharing without compromising individual, or institute-level, intellectual property and privacy details. By enabling new computing opportunities with sensitive data, we envision a secure collaborative environment that will play a significant role in accelerating scientific discovery. Several recent technological advancements have made it possible to realize these capabilities. In this paper, we present our efforts at ORNL toward developing a secure computation platform. We present a use case where scientific data generated from complex instruments, like those at the Spallation Neutron Source (SNS), are used to train a differential privacy enabled deep learning (DL) network on Summit, which is then hosted as a secure multi-party computation (MPC) service on ORNL’s Compute and Data Environment for Science (CADES) cloud computing platform for third-party inference. In this feasibility study, we discuss the challenges involved, elaborate on leveraged technologies, analyze relevant performance results and present the future vision of our work to establish secure collaboration capabilities within and outside of ORNL.

Yoginath, Srikanth↗
DOE OSTI2025

Testing CP properties of the Higgs boson coupling to τ leptons with heterogeneous graphs

In this paper we explore the possibility of utilizing Deep Learning in measuring the CP properties of the coupling of Higgs boson to τ leptons at the High Luminosity Large Hadron Collider. We employ three Deep Learning (DL) networks, Multi-Layer Perceptron (MLP), Graph Convolution Network (GCN), and Graph Transformer Network (GTN) to enhance signal-to-background separation. The angle between τ lepton decay planes at the detector level is CP-sensitive observables, and we develop Heterogeneous Graphs that integrate diverse node and edge structures to incorporate the CP-sensitive observable efficiently. Using simplified detector simulations we estimate the reconstruction accuracy of the angle between τ lepton planes at the detector level, considering hadronic τ decay modes and standard model backgrounds. With $\sqrt{s}$ = 14 TeV and $\mathcal{L}$ = 100 fb -1 , MLP excludes CP mixing angles above 20° at 68% confidence level (CL), while GCN and GTN achieve exclusions at 90% CL and 95% CL, respectively. The networks also achieve a 3σ significance in excluding a pure CP-odd state.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2019

Defatted microalgae ( Nannochloropsis sp.) from biorefinery as a potential feed protein source to replace fishmeal in European sea bass diets

Here, the present work focuses on the use of defatted biomass of the microalga Nannochloropsis sp. from the biodiesel industry, as a partial substitute of fish meal (FM) in diets for European sea bass. The effects of increasing inclusion levels of microalgal meal on growth performance, body composition, nutrient utilization, gut morphology, and innate immunity were evaluated after 93 days. A reference alga-free diet was the control (CTRL) diet, and the three experimental diets contained 5 (MA5), 10 (MA10), and 15% (MA15) of the microalgal meal. The microalga-rich diets were supplemented with DL-methionine to assure sea bass dietary requirement. Overall, nutrient apparent digestibilities (ADCs) of the diets were not altered by the microalgal inclusion, but energy ADC was highest in fish fed the CTRL diet (90% vs 88%). At the end of the trial, fish growth performance was similar among dietary treatments (DGI of 1.0), but fish fed MA10 had a significantly higher feed conversion ratio than those fed CTRL and MA5. Final whole body composition and nutrient gain of fish fed the different diets were similar. No significant differences were detected in gut morphology among treatments. Innate immune parameters (lysozyme, alternative complement pathway—ACH50, and peroxidase) were examined, and ACH50 of the fish fed MA15 was significantly lower than those fed MA10, suggesting a dose-dependent stimulation of the innate immune response. The present results indicate that defatted microalgal meal can replace fishmeal in European sea bass diets—at inclusion levels of up to 15%—contributing to a circular economy approach.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2017

Synthesis and Photophysical Study of New Green Fluorescent TPA Based Poly(azomethine)s

A series of poly(azomethine)s (PAMs) were synthesized from N{sup 1}-(4-aminophenyl)-N{sup 1}-(4-phenoxyphenyl)benzene-1,4-diamine (DA) and various dialdehydes to investigate the influence of structure of polymer chain and triphenylamine-based phenoxy pendant group on the optoelectronic properties. The structural characterization of the resulting poly(azomethine)s was carried out by solubility test, gel permeation chromatography, viscosity measurement, fourier transform infrared (FTIR) spectral and CHN elemental analysis. The photophysical and electrochemical properties of the materials were scrutinized by UV–vis, photoluminescence, time correlation photon counting spectral analysis (TCSP) and cyclic voltammetry. The thermal stability of the poly(azomethine)s was assessed by differential scanning calorimetry and thermogravimetric analysis found to be stable upto 300 °C. These polymers exhibit moderate inherent viscosity range from 0.99 to 1.15 g dL{sup − 1} and appreciable organosolubility. The presence of triphenylamine and azomethine (CH = N) linkage in our synthesized materials rendered them fluorescent, emitting green light upon excitation at 375 nm with quantum efficiencies of 3.9–8.5%. The pendant phenoxy group at para-position in new poly(azomethine)s has also lowered the onset oxidation potentials and elevated the HOMO levels. Additionally, the presence of conjugation increases the fluorescence time of the excited state in conjugated polymers which was found in the range 9.22–11.17 ns, sufficient to be use in future optoelectronic applications.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2012

Dose-Escalated Intensity-Modulated Radiotherapy Is Feasible and May Improve Locoregional Control and Laryngeal Preservation in Laryngo-Hypopharyngeal Cancers

Purpose: To determine the safety and outcomes of induction chemotherapy followed by dose-escalated intensity-modulated radiotherapy (IMRT) with concomitant chemotherapy in locally advanced squamous cell cancer of the larynx and hypopharynx (LA-SCCL/H). Methods and Materials: A sequential cohort Phase I/II trial design was used to evaluate moderate acceleration and dose escalation. Patients with LA-SCCL/H received IMRT at two dose levels (DL): DL1, 63 Gy/28 fractions (Fx) to planning target volume 1 (PTV1) and 51.8 Gy/28 Fx to PTV2; DL2, 67.2 Gy/28 Fx and 56 Gy/28 Fx to PTV1 and PTV2, respectively. Patients received induction cisplatin/5-fluorouracil and concomitant cisplatin. Acute and late toxicities and tumor control rates were recorded. Results: Between September 2002 and January 2008, 60 patients (29 DL1, 31 DL2) with Stage III (41% DL1, 52% DL2) and Stage IV (52% DL1, 48% DL2) disease were recruited. Median (range) follow-up for DL1 was 51.2 (12.1-77.3) months and for DL2 was 36.2 (4.2-63.3) months. Acute Grade 3 (G3) dysphagia was higher in DL2 (87% DL2 vs. 59% DL1), but other toxicities were equivalent. One patient in DL1 required dilatation of a pharyngeal stricture (G3 dysphagia). In DL2, 2 patients developed benign pharyngeal strictures at 1 year. One underwent a laryngo-pharyngectomy and the other a dilatation. No other G3/G4 toxicities were reported. Overall complete response was 79% (DL1) and 84% (DL2). Two-year locoregional progression-free survival rates were 64.2% (95% confidence interval, 43.5-78.9%) in DL1 and 78.4% (58.1-89.7%) in DL2. Two-year laryngeal preservation rates were 88.7% (68.5-96.3%) in DL1 and 96.4% (77.7-99.5%) in DL2. Conclusions: At a mean follow-up of 36 months, dose-escalated chemotherapy-IMRT at DL2 has so far been safe to deliver. In this study, DL2 delivered high rates of locoregional control, progression-free survival, and organ preservation and has been selected as the experimental arm in a Cancer Research UK Phase III study.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2021

Using Auto-Segmentation to Reduce Contouring and Dose Inconsistency in Clinical Trials: The Simulated Impact on RTOG 0617

Contouring inconsistencies are known but understudied in clinical radiation therapy trials. We applied auto-contouring to the Radiation Therapy Oncology Group (RTOG) 0617 dose escalation trial data. We hypothesized that the trial heart doses were higher than reported due to inconsistent and insufficient heart segmentation. We tested our hypothesis by comparing doses between deep-learning (DL) segmented hearts and trial hearts.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2014

A First-Passage Kinetic Monte Carlo method for reaction–drift–diffusion processes

Stochastic reaction–diffusion models are now a popular tool for studying physical systems in which both the explicit diffusion of molecules and noise in the chemical reaction process play important roles. The Smoluchowski diffusion-limited reaction model (SDLR) is one of several that have been used to study biological systems. Exact realizations of the underlying stochastic processes described by the SDLR model can be generated by the recently proposed First-Passage Kinetic Monte Carlo (FPKMC) method. This exactness relies on sampling analytical solutions to one and two-body diffusion equations in simplified protective domains. In this work we extend the FPKMC to allow for drift arising from fixed, background potentials. As the corresponding Fokker–Planck equations that describe the motion of each molecule can no longer be solved analytically, we develop a hybrid method that discretizes the protective domains. The discretization is chosen so that the drift–diffusion of each molecule within its protective domain is approximated by a continuous-time random walk on a lattice. New lattices are defined dynamically as the protective domains are updated, hence we will refer to our method as Dynamic Lattice FPKMC or DL-FPKMC. We focus primarily on the one-dimensional case in this manuscript, and demonstrate the numerical convergence and accuracy of our method in this case for both smooth and discontinuous potentials. We also present applications of our method, which illustrate the impact of drift on reaction kinetics.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2026

Evaluating lightweight unsupervised online IDS for masquerade attacks in CAN

Vehicular controller area networks (CANs) are susceptible to masquerade attacks by malicious adversaries. In masquerade attacks, adversaries silence a targeted ID and then send malicious frames with forged content at the expected timing of benign frames. As masquerade attacks could seriously harm vehicle functionality and are the stealthiest attacks to detect in CAN, recent work has devoted attention to compare frameworks for detecting masquerade attacks in CAN. However, most existing works report offline evaluations using CAN logs already collected using simulations that do not comply with the domain’s real-time constraints. Here we contribute to advance the state of the art by presenting a comparative evaluation of four different non-deep learning (DL)-based unsupervised online intrusion detection systems (IDS) for masquerade attacks in CAN. Our approach differs from existing comparative evaluations in that we analyze the effect of controlling streaming data conditions in a sliding window setting. In doing so, we use realistic masquerade attacks being replayed from the ROAD dataset. We show that although evaluated IDS are not effective at detecting every attack type, the method that relies on detecting changes in the hierarchical structure of clusters of time series produces the best results at the expense of higher computational overhead. We discuss limitations, open challenges, and how the evaluated methods can be used for practical unsupervised online CAN IDS for masquerade attacks.

Anomaly detection↗
DOE OSTI2025

Impact of organic acids on extraction of rare earth elements: Mechanisms and optimization

Organic acids are increasingly recognized as an environmentally friendly and efficient selective leaching agent of critical minerals. However, their impact on subsequent extraction processes remains unclear. This study investigates the extraction behavior of rare earth elements (REEs), representative of critical minerals, in the presence of various organic acids, including citric acid, maleic acid, malonic acid, DL-malic acid, L(+)-tartaric acid, and L-ascorbic acid. The results show that organic acids slightly reduce the extraction of REEs but greatly increase the extraction of aluminum (Al³⁺), making it harder to separate REEs from other elements. To avoid this, it is best to keep organic acid concentrations as low as possible in practical applications. However, by optimizing extraction conditions, these negative effects can be minimized. Specifically, adjusting the contact time between solutions allows for efficient REE extraction while limiting unwanted aluminum extraction. Among the acids tested, malonic acid and L(+)-tartaric acid were found to be the most effective for selectively extracting REEs. Mechanistically, organic acids are likely to form complexes with REEs and D2EHPA during extraction, except for Y (III), offering practical guidelines for optimizing REE recovery. Beyond extraction, this study also highlights a way to recover organic acids by precipitating REEs from stripping solutions using oxalic acid, which adds both environmental and economic benefits. Furthermore, these findings provide useful insights for optimizing REE recovery and offer a reference for extracting other critical minerals from solutions containing organic acids.

D2EHPA↗
DOE OSTI2016

Quantitative characterization of material surface — Application to Ni + Mo electrolytic composite coatings

A new approach to numerical analysis of maps of material surface has been proposed and discussed in detail. It was concluded that the roughness factor RF and the root mean square roughness S{sub q} show a saturation effect with increasing size of the analysed maps what allows determining the optimal map dimension representative of the examined material. A quantitative method of determining predominant direction of the surface texture based on the power spectral density function is also proposed and discussed. The elaborated method was applied in surface analysis of Ni + Mo composite coatings. It was shown that co-deposition of molybdenum particles in nickel matrix leads to an increase in surface roughness. In addition, a decrease in size of the embedded Mo particles in Ni matrix causes an increase of both the surface roughness and the surface texture. It was also stated that the relation between the roughness factor and the double layer capacitance C{sub dl} of the studied coatings is linear and allows determining the double layer capacitance of the smooth nickel electrode. - Highlights: •Optimization of the procedure for the scanning of the material surface •Quantitative determination of the surface roughness and texture intensity •Proposition of the parameter describing privileged direction of the surface texture •Determination of the double layer capacitance of the smooth electrode.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Engineering of increased L-Threonine production in bacteria by combinatorial cloning and machine learning

The goal of this study is to develop a general strategy for bacterial engineering using an integrated synthetic biology and machine learning (ML) approach. This strategy was developed in the context of increasing L-threonine production in Escherichia coli ATCC 21277. A set of 16 genes was initially selected based on metabolic pathway relevance to threonine biosynthesis and used for combinatorial cloning to construct a set of 385 strains to generate training data (i.e., a range of L-threonine titers linked to each of the specific gene combinations). Hybrid (regression/classification) deep learning (DL) models were developed and used to predict additional gene combinations in subsequent rounds of combinatorial cloning for increased L-threonine production based on the training data. As a result, E. coli strains built after just three rounds of iterative combinatorial cloning and model prediction generated higher L-threonine titers (from 2.7 g/L to 8.4 g/L) than those of patented L-threonine strains being used as controls (4-5 g/L). Interesting combinations of genes in L-threonine production included deletions of the tdh, metL, dapA, and dhaM genes as well as overexpression of the pntAB, ppc, and aspC genes. Mechanistic analysis of the metabolic system constraints for the best performing constructs offers ways to improve the models by adjusting weights for specific gene combinations. Graph theory analysis of pairwise gene modifications and corresponding levels of L-threonine production also suggests additional rules that can be incorporated into future ML models.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2018

Electroluminescence pulse shape and electron diffusion in liquid argon measured in a dual-phase TPC

Here, we report the measurement of the longitudinal diffusion constant in liquid argon with the DarkSide-50 dual-phase time projection chamber. The measurement is performed at drift electric fields of 100 V/cm, 150 V/cm, and 200 V/cm using high statistics $^{39}$Ar decays from atmospheric argon. We derive an expression to describe the pulse shape of the electroluminescence signal (S2) in dual-phase TPCs. The derived S2 pulse shape is fit to events from the uppermost portion of the TPC in order to characterize the radial dependence of the signal. The results are provided as inputs to the measurement of the longitudinal diffusion constant DL, which we find to be (4.12 $\pm$ 0.04) cm$^2$/s for a selection of 140keV electron recoil events in 200V/cm drift field and 2.8kV/cm extraction field. To study the systematics of our measurement we examine datasets of varying event energy, field strength, and detector volume yielding a weighted average value for the diffusion constant of (4.09 $\pm$ 0.09) cm$^2$ /s. The measured longitudinal diffusion constant is observed to have an energy dependence, and within the studied energy range the result is systematically lower than other results in the literature.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2017

Comparative H-mode density limit studies in JET and AUG

Identification of the mechanisms for the H-mode density limit in machines with fully metallic walls, and their scaling to future devices is essential to find for these machines the optimal operational boundaries with the highest attainable density and confinement. Systematic investigations of H-mode density limit plasmas in experiments with deuterium external gas fuelling have been performed on machines with fully metallic walls, JET and AUG and results have been compared with one another. Basically, the operation phases are identical for both tokamaks: the stable H-mode phase, degrading H-mode phase, breakdown of the H-mode with energy confinement deterioration usually accompanied by a dithering cycling phase, followed by the l-mode phase. The observed H-mode density limit on both machines is found close to the Greenwald limit (n/n GW = 0.8–1.1 in the observed magnetic configurations). The similar behavior of the radiation on both tokamaks demonstrates that the density limit (DL) is neither related to additional energy losses from the confined region by radiation, nor to an inward collapse of the hot discharge core induced by overcooling of the plasma periphery by radiation. It was observed on both machines that detachment, as well as the X-point MARFE itself, does not trigger a transition in the confinement regime and thus does not present a limit on the plasma density. It is the plasma confinement, most likely determined by edge parameters, which is ultimately responsible for the transition from H- to l-mode. The measured Greenwald fractions are found to be consistent with the predictions from different theoretical models, based on MHD instability theory in the near-SOL.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Precision Dijet Acoplanarity Tomography of the Chromo Structure of Perfect QCD Fluids

Dijet acoplanarity is dominated by vacuum (Sudakov) pQCD radiation even in Pb+Pb collisions, but future higher precision measurements of the tails of the acoplanarity distributions can help to resolve separately the medium opacity, χ=L/λ and the color screening scale μ 2 from the path averaged BDMS saturation scale Q$^2_s$[χ,μ]=∫dL$\hat{q}$(E,T(L))∝μ 2 L/λ that is already well constrained by nuclear modification factor data on R AA (p T ). We compare Gaussian (BDMS) and non-Gaussian (GLV) models of medium broadening of vacuum (Sudakov) induced acoplanarity distributions in A+A. With few percent accuracy on the ratio of A+A to p+p distributions, experiments can easily identify non-Gaussian Landau and Rutherford tails due to multiple collisions. However, we find that sub-percent precision will be required to constrain χ and μ separately from Q s .

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2022

Deep inverse photonic design: A tutorial

Deep learning (DL) has been used to design deep neural networks (DNNs) which have recently been applied to solving inverse problems in artificial electromagnetic materials (AEMs). Although inverse problems are often ill-posed, and therefore are difficult to solve, deep inverse models (DIMs) have achieved impressive results often surpassing capabilities possible with other approaches. We overview the process of deep inverse learning applied to AEM problems, including the building of data sets, design of a forward model, and comparison of inverse approaches including limitations. We conclude by detailing some important outstanding issues of deep inverse design of AEMs, and present an outlook of this exciting field.

42 ENGINEERING↗