Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Feature Engineering”

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 469 records · Page 26

Modular cell-free expression plasmids to accelerate biological design in cells

Industrial biotechnology aims to produce high-value products from renewable resources. This can be challenging because model microorganisms—organisms that are easy to use like Escherichia coli—often lack the machinery required to utilize desired feedstocks like lignocellulosic biomass or syngas. Non-model organisms, such as Clostridium, are industrially proven and have desirable metabolic features but have several hurdles to mainstream use. Namely, these species grow more slowly than conventional laboratory microbes, and genetic tools for engineering them are far less prevalent. To address these hurdles for accelerating cellular design, cell-free synthetic biology has matured as an approach for characterizing non-model organisms and rapidly testing metabolic pathways in vitro. Unfortunately, cell-free systems can require specialized DNA architectures with minimal regulation that are not compatible with cellular expression. In this work, we develop a modular vector system that allows for T7 expression of desired enzymes for cell-free expression and direct Golden Gate assembly into Clostridium expression vectors. Utilizing the Joint Genome Institute’s DNA Synthesis Community Science Program, we designed and synthesized these plasmids and genes required for our projects allowing us to shuttle DNA easily between our in vitro and in vivo experiments. We next validated that these vectors were sufficient for cell-free expression of functional enzymes, performing on par with the previous state-of-the-art. Lastly, we demonstrated automated six-part DNA assemblies for Clostridium autoethanogenum expression with efficiencies ranging from 68% to 90%. We anticipate this system of plasmids will enable a framework for facile testing of biosynthetic pathways in vitro and in vivo by shortening development cycles.

59 BASIC BIOLOGICAL SCIENCES↗

Transforming Energy Through Computational Excellence: A View From NREL

At the National Renewable Energy Laboratory (NREL)—a U.S. Department of Energy laboratory—computational science, high-performance computing, applied mathematics, advanced computer science, visualization, and data play a pivotal role in advancing energy abundance, affordability, security, and reliability. From fundamental scientifc discovery to systems engineering and analysis, NREL researchers tackle market-relevant challenges to develop solutions for an independent energy system that is reliable, resilient and secure. Collaborative partnerships with industry, government, and academia ensure that our research remains cutting edge, impactful, applicable, and aligned with real-world energy needs. This special issue of Computing in Science & Engineering highlights exemplary NREL projects where computational tools and methodologies drive discovery and accelerate innovation in scalable and integrated energy systems. The featured articles explore the role of computational modeling, high-performance computing, generative AI, and adaptive computing in advancing independent energy solutions, optimizing sustainability research, and enhancing decision-making for energy solutions using a broad mix of energy technologies. Here, these contributions demonstrate how NREL’s computational research bridges the gap between theoretical advancements and practical implementation, emphasizing interdisciplinary collaboration and a commitment to innovation, with a focus on translating computational excellence into real-world impact, thus accelerate progress toward national energy goals. By showcasing cutting-edge research at the intersection of computational science and energy systems, this issue aims to inspire and inform researchers, practitioners, and policymakers dedicated to shaping a more reliable energy future.

97 MATHEMATICS AND COMPUTING↗

Machine learning methods for weather forecasting

SAND2025-14466O This repository contains code for developing, training, and evaluating machine learning models for weather and climate forecasting, including forecast skill assessment, feature importance analysis, and reproducible workflows for model comparison. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Holthuijzen, Maike [Sandia National Lab. (SNL-CA),↗

High-density SSiC 3D-printed Lattices for Compact HTHP Aero-engine Recuperators

The proposed project utilizes advanced ceramic-based 3D-printing technology to develop light, low-cost, ultra-compact, high-temperature high-pressure (HTHP) heat exchangers (HX) for the harsh oxidizing environment of aircraft gas-turbine engines. Typical 3D printing technologies (powder bed or lithographic techniques) cannot produce large high-density, monolithic ceramic material with intricate internal features and channels required for high-temperature, high-pressure recuperators. Furthermore, typical 3D printing technologies are not yet economical for large high-density components. A high-density material is required for hermeticity, high-temperature strength, and oxidation, corrosion, and wear resistance. The team has invented an extrusion-based additive manufacturing technology (referred to as robocasting) and developed processing techniques to tailor composition and rheological mix properties for rapid 3D printing of high-density alumina and SiC parts. The goal of the proposed project is to use robocasting technology to demonstrate commercializable high-density 3D-printed ceramic HX modules with high durability and quality, thereby reducing the risk of thermo-mechanical failure under extreme environments. In addition, the proposed project improves the quality of the ceramic 3D-printing process over a large printed area and HTHP operating conditions. Ceramic 3D printing technology also provides a means to create advanced topologies that are unobtainable with conventional manufacturing.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Leadership at Interfaces: Capitalizing on the "Gaps" in Science, Technology and Education [Slides]

Leadership in the scientific enterprise is most commonly exemplified by those who commandeer major discoveries and critical breakthroughs that expand boundaries, naturally underpinning a competitive spirit among peers in the field. However, many other leadership opportunities exist at the interfaces of such boundaries - as well as at the interfaces of scientific disciplines, classes of professions and across generations - and may enable alternative guiding philosophies that are intrinsically less prone to competition and saturation, and perhaps appreciation. This presentation features a personal reflection of scientific and leadership philosophies that were intentionally manifested at, and by, irreplaceable interfaces of human creativity. Reaction engineering, an increasingly vanishing discipline within chemical engineering, is emphasized as a manifestation of interfacial leadership and serves as a vehicle to trace pivotal career moments in both the speaker's scientific discoveries and in his relational values within human enterprises.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Enhanced Radiation Damage Tolerance of Amorphous Interphase and Grain Boundary Complexions in Cu-Ta

Amorphous interfacial complexions are particularly resistant to radiation damage and have been primarily studied in alloys with good glass-forming ability, yet recent reports suggest that these features can form even in immiscible alloys such as Cu-Ta under irradiation. In this work, the mechanisms of damage production and annihilation due to primary knock-on atom collisions are investigated for amorphous interphase and grain boundaries in a Cu-Ta alloy using atomistic simulations. Amorphous complexions, in particular amorphous interphase complexions that separate Cu and Ta grains, result in less residual defect damage than their ordered counterparts. Stemming from the nanophase chemical separation in this alloy, the amorphous complexions exhibit a highly heterogeneous distribution of atomic excess volume, as compared to a good glass former like Cu-Zr. Complexion thickness, a tunable structural descriptor, plays a vital role in damage resistance. Thicker interfacial films are more damage-tolerant because they alter the defect production rate due to differences in intrinsic displacement threshold energies during the collision cascade. Overall, the findings of this work highlight the importance of interfacial engineering in enhancing the properties of materials operating in radiation-prone environments and the promise of amorphous complexions as particularly radiation damage-tolerant microstructural features.

36 MATERIALS SCIENCE↗

Biomimicry in Clean Energy Futures: Workshop Report

On July 25, 2023, PNNL hosted a workshop to explore the potential for biomimicry – looking to nature and biology for solutions to human engineering challenges – to support clean energy futures and associated research. Participants represented a range of technical expertise ranging from ecological modeling to cyber-security. Workshop supporting material featured a literature review of biomimetic approaches in all renewable energy technologies and inspired from all kingdoms of life. Participants noted clear contrasts between natural design strategies and clean energy deployment and the electric grid. They also agreed that biomimicry offered a rich area of research potential but lacked definitive benefits or results in commercialization phases.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Geometric learning for computational mechanics Part II: Graph embedding for interpretable multiscale plasticity

The history-dependent behaviors of classical plasticity models are often driven by internal variables evolved according to phenomenological laws. The difficulty to interpret how these internal variables represent a history of deformation, the lack of direct measurement of these internal variables for calibration and validation, and the weak physical underpinning of those phenomenological laws have long been criticized as barriers to creating realistic models. In this work, geometric machine learning on graph data (e.g. finite element solutions) is used as a means to establish a connection between nonlinear dimensional reduction techniques and plasticity models. Geometric learning-based encoding on graphs allows the embedding of rich time-history data onto a low-dimensional Euclidean space such that the evolution of plastic deformation can be predicted in the embedded feature space. Finally, a corresponding decoder can then convert these low-dimensional internal variables back into a weighted graph such that the dominating topological features of plastic deformation can be observed and analyzed.

42 ENGINEERING↗

Degrees of Rate Control at Non(pseudo)steady-State Conditions

The degrees of rate control enable identification of rate-controlling transition states and species during (pseudo-)steady-state reaction. We demonstrate a method for identifying rate-controlling transition states and species at non-(pseudo-)steady-state conditions through modification of the definitions for degrees of rate control and sensitivities by superimposing the condition that time constants of the analytical rate function are kept constant when taking the partial derivatives of rate with respect to the rate constants/free energies. In conclusion, the utility of transient degrees of rate control is demonstrated for catalytic reaction networks to rationalize features of the transient rate including extrema and discontinuities.

42 ENGINEERING↗

Connectivity-informed drainage network generation using deep convolution generative adversarial networks

Abstract Stochastic network modeling is often limited by high computational costs to generate a large number of networks enough for meaningful statistical evaluation. In this study, Deep Convolutional Generative Adversarial Networks (DCGANs) were applied to quickly reproduce drainage networks from the already generated network samples without repetitive long modeling of the stochastic network model, Gibb’s model. In particular, we developed a novel connectivity-informed method that converts the drainage network images to the directional information of flow on each node of the drainage network, and then transforms it into multiple binary layers where the connectivity constraints between nodes in the drainage network are stored. DCGANs trained with three different types of training samples were compared; (1) original drainage network images, (2) their corresponding directional information only, and (3) the connectivity-informed directional information. A comparison of generated images demonstrated that the novel connectivity-informed method outperformed the other two methods by training DCGANs more effectively and better reproducing accurate drainage networks due to its compact representation of the network complexity and connectivity. This work highlights that DCGANs can be applicable for high contrast images common in earth and material sciences where the network, fractures, and other high contrast features are important.

42 ENGINEERING↗

AutoAtlas: Neural Network for 3D Unsupervised Partitioning and Representation Learning

Here we present a novel neural network architecture called AutoAtlas for fully unsupervised partitioning and representation learning of 3D brain Magnetic Resonance Imaging (MRI) volumes. AutoAtlas consists of two neural network components: one neural network to perform multi-label partitioning based on local texture in the volume, and a second neural network to compress the information contained within each partition. We train both of these components simultaneously by optimizing a loss function that is designed to promote accurate reconstruction of each partition, while encouraging spatially smooth and contiguous partitioning, and discouraging relatively small partitions. We show that the partitions adapt to the subject specific structural variations of brain tissue while consistently appearing at similar spatial locations across subjects. AutoAtlas also produces very low dimensional features that represent local texture of each partition. We demonstrate prediction of metadata associated with each subject using the derived feature representations and compare the results to prediction using features derived from FreeSurfer anatomical parcellation. Since our features are intrinsically linked to distinct partitions, we can then map values of interest, such as partition-specific feature importance scores onto the brain for visualization.

42 ENGINEERING↗

A Projected Network Model of Online Disinformation Cascades

Within the past half-decade, it has become overwhelmingly clear that suppressing the spread of deliberate false and misleading information is of the utmost importance for protecting democratic institutions. Disinformation has been found to come from both foreign and domestic actors, but the effects from either can be disastrous. From the simple encouragement of unwarranted distrust to conspiracy theories promoting violence, the results of disinformation have put the functionality of American democracy under direct threat. Present scientific challenges posed by this problem include detecting disinformation, quantifying its potential impact, and preventing its amplification. We present a model on which we can experiment with possible strategies toward the third challenge: the prevention of amplification. This is a social contagion network model, which is decomposed into layers to represent physical, ''offline'', interactions as well as virtual interactions on a social media platform. Along with the topological modifications to the standard contagion model, we use state-transition rules designed specifically for disinformation, and distinguish between contagious and non-contagious infected nodes. We use this framework to explore the effect of grassroots social movements on the size of disinformation cascades by simulating these cascades in scenarios where a proportion of the agents remove themselves from the social platform. We also test the efficacy of strategies that could be implemented at the administrative level by the online platform to minimize such spread. These top-down strategies include banning agents who disseminate false information, or providing corrective information to individuals exposed to false information to decrease their probability of believing it. We find an abrupt transition to smaller cascades when a critical number of random agents are removed from the platform, as well as steady decreases in the size of cascades with increasingly more convincing corrective information. Finally, we compare simulated cascades on this framework with real cascades of disinformation recorded on Whatsapp surrounding the 2019 Indian election. We find a set of hyperparameter values that produces a distribution of cascades matching the scaling exponent of the distribution of actual cascades recorded in the dataset. We acknowledge the available future directions for improving the performance of the framework and validation methods, as well as ways to extend the model to capture additional features of social contagion.

42 ENGINEERING↗

3D Printing at Los Alamos

At Los Alamos National Laboratory (LANL), their additive manufacturing division is seeking to answer the question whether it is possible to 3D print on a 5th axis. The end goal is to achieve true continuous fiber 3D printing. My goal on this project was to learn about additive manufacturing, and to learn how to operate the experimental 3D printer. Each print that was done demonstrated a different pattern that would change the strength and integrity of the print. Overall prints with minimal strand separation and featured weave-like patterns proved to have the best integrity.

42 ENGINEERING↗

Spark-ignited direct-injection engine combustion systems

A direct-injection stratified charge internal combustion engine includes a combustion cylinder to receive an air-fuel mixture, and an air intake port to inlet air into the combustion cylinder. The direct-injection engine also includes a fuel injector configured to deliver fuel within the cylinder in a spray pattern substantially aligned to a cylinder central axis to create the air-fuel mixture. A spark igniter is located within a path of the spray pattern to ignite combustion of the air-fuel mixture. The direct-injection engine further includes a movable piston defining a lower boundary of the combustion cylinder to contain the combustion of the air-fuel mixture. The piston is configured to include a bowl portion having local geometric features located on an intake port side of the combustion cylinder to redirect fluid flow towards a vortex in fluid communication with a combustion location near the cylinder central axis.

42 ENGINEERING↗

Investigation of the effects of passive pre-chamber nozzle pattern and ignition system on engine performance and emissions

The impact of passive pre-chamber (PC) internal volume, nozzle hole pattern (i.e. with and without a central axial nozzle), and PC igniter plug type on performance and emissions was investigated in an optically accessible, single-cylinder, gasoline research engine. The two PC igniter plugs investigated were a conventional inductive coil spark plug and a nanosecond repetitively pulsed (NRP) plasma discharge system previously demonstrated to accelerate early flame propagation. The baseline PC design featured a funnel shaped internal volume with a PC tip that contained six radial nozzles and a larger central axial nozzle. Two additional PC tip geometries were evaluated where either the baseline internal volume was increased or the axial nozzle was removed and the radial nozzle diameters were increased. A sweep of charge equivalence ratios ( ϕ) from nearly stoichiometric to the lean limit was performed for a fixed engine speed (1300 revolutions per minute), and engine load (3.5 bar gross indicated mean effective pressure). Time-resolved PC and main chamber (MC) pressure data as well as MC emissions data were collected to analyze engine performance and emissions characteristics. Combustion in the MC was further investigated using high-speed excited methylidyne radical (CH*) chemiluminescence imaging. Collected results highlighted that while all PC tips and ignition systems exhibited similar performance and emissions down to ϕ = 0.8, relevant differences in thermal efficiency and emissions for leaner charge mixtures were observed, with the results highly dependent on the nozzle pattern and ignition system. Major deviations were correlated to preferential de-pressurization of the PC through the axial nozzle for lean conditions that was not observed for mixture conditions closer to stoichiometric. Results show that a combination of radial and axial nozzle patterns in the PC extended lean-stability limits at the low-load condition evaluated. Further benefits were observed with the use of NRP ignition systems due to faster combustion within the PC volume provided that the volume was sufficiently large.

Di Sabatino, Francesco↗

Let’s Unleash the Network Judgment: A Self-Supervised Approach for Cloud Image Analysis

Accurate cloud type identification and coverage analysis are crucial in understanding the Earth’s radiative budget. Traditional computer vision methods rely on low-level visual features of clouds for estimating cloud coverage or sky conditions. Several handcrafted approaches have been proposed; however, scope for improvement still exists. Newer deep neural networks (DNNs) have demonstrated superior performance for cloud segmentation and categorization. These methods, however, need expert engineering intervention in the preprocessing steps—in the traditional methods—or human assistance in assigning cloud or clear sky labels to a pixel for training DNNs. Such human mediation imposes considerable time and labor costs. We present the application of a new self-supervised learning approach to autonomously extract relevant features from sky images captured by ground-based cameras, for the classification and segmentation of clouds. We evaluate a joint embedding architecture that uses self-knowledge distillation plus regularization. We use two datasets to demonstrate the network’s ability to classify and segment sky images—one with ~85,000 images collected from our ground-based camera and another with 400 labeled images from the WSISEG database. We find that this approach can discriminate full-sky images based on cloud coverage, diurnal variation, and cloud base height. Additionally, it semantically segments the cloud areas without labels. The approach shows competitive performance in all tested tasks, suggesting a new alternative for cloud characterization.

54 ENVIRONMENTAL SCIENCES↗

Engineering of flat bands and Dirac bands in two-dimensional covalent organic frameworks (COFs): relationships among molecular orbital symmetry, lattice symmetry, and electronic-structure characteristics

Two-dimensional covalent organic frameworks (2D-COFs), also referred to as 2D polymer networks, display unusual electronic-structure characteristics, which can significantly enrich and broaden the fields of electronics and spintronics. In this Focus article, the objective of this report is to lay the groundwork for the conceptual description of the fundamental relationships among the COF electronic structures, the symmetries of their 2D lattices, and the frontier molecular orbitals (MOs) of their core and linker components. We focus on monolayers of hexagonal COFs and use tight-binding model analyses to highlight the critical role of the frontier-MO symmetry, in addition to lattice symmetry, in determining the nature of the electronic bands near the Fermi level. We rationalize the intriguing feature that, when the core unit has degenerate highest occupied MOs [or lowest unoccupied MOs], the COF highest valence band [or lowest conduction band] is flat but degenerate with a dispersive band at a high-symmetry point of the Brillouin zone; the consequences of having such band characteristics are briefly described. Multi-layer and bulk 2D COFs are found to maintain the salient features of the monolayer electronic structures albeit with a reduced bandgap due to the interlayer coupling. This Focus article is thus meant to provide an effective framework for the engineering of flat and Dirac bands in 2D polymer networks.

36 MATERIALS SCIENCE↗

High frequency lithium niobate film-thickness-mode optomechanical resonator

High-frequency optomechanical resonators are in demand as transduction devices to bridge microwave and optical fields. Thin-film lithium niobate is a promising platform for implementing high-frequency optomechanics for its low optical loss and strong piezoelectric coefficients. However, its strong piezoelectricity is also known to introduce excess phonon loss. Here, we present lithium niobate optomechanical resonators with film-thickness-mode mechanical resonances up to 5.2 GHz, reaching the operating frequency regime of superconducting qubits. By engineering the mechanical anchor to minimize the phonon loss, we achieve a high quality factor up to 12 500 at cryogenic temperatures and, hence, a frequency-quality factor product of 6.6 × 10 13 . Our system also features interference between piezo-optomechanical and electro-optic modulation. A theoretical model is derived to analyze these two effects and their interference.

42 ENGINEERING↗