SEARCH · Engineering Papers

Results for “feature selection”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,297 records · Page 72

All sources NASA NTRS DOE OSTI

NASA NTRSDate not supplied

Fluid System Stability Analysis Techniques

Fluid systems, or networks, consist of multiple components that work together to achieve some desired thermofluid state. For any generic application, this fluid state can be a combination of the fluid pressure, flow rate, enthalpy, or species concentration. Fluid system components, such as pumps and valves, are often governed by nonlinear differential equations, resulting in complex component-to-component interactions. System-level fluid network stability occurs when the flow through the system can maintain a steady-state solution in the presence of small perturbations, which depends on these component interactions. System instability, however, can go undetected until issues arise during integrated system testing. This presentation explores a method for system designers to think of the fluid network as an assembly of components, each with their own thermofluid surfaces of partial stability, called nullclines. The intersections of all nullclines yields system-level solutions, called equilibrium points. When designers define operating points, they are tuning system parameters so that these equilibrium points move to the desired location in the thermofluid state plane. However, linearization theory shows us that the dynamic behavior around these points can be unstable. The local stability of these equilibrium points can be assessed analytically with eigen-analysis, or numerically by propagating state-plane samples to construct a phase portrait. Investigating a phase portrait can help designers gain a qualitative understanding of a system’s dynamic performance. This understanding can then help inform requirement definitions, component selection, and operational procedures. This presentation includes an example of the phase portrait technique on a system featuring a centrifugal pump and a back-pressure regulator (BPR). Numerical modeling of this system suggests that equilibrium points on the left-hand side of the pump curve are dynamically unstable.

Fluid System↗

NASA NTRSDate not supplied

Bingo: A Customizable Framework for Symbolic Regression with Genetic Programming

In this paper, we introduce Bingo, a flexible and customizable yet performant Python framework for symbolic regression with genetic programming. Bingo maintains a modular code structure for simple abstraction and easily swappable components. Fitness functions, selection methods, and constant optimization methods allow for easy problem-specific customization. Bingo also maintains several features for increased efficiency such as parallelism, equation simplification, and a C++ backend. We compare Bingo’s performance to other genetic programming for symbolic regression (GPSR) methods to show that it is both competitive and flexible.

machine learning↗

NASA NTRSDate not supplied

Bingo: A Customizable Framework for Symbolic Regression with Genetic Programming

David Randall↗

NASA NTRSDate not supplied

Developmental Flight Instrumentation: Review of Space Shuttle, Ares I-X, and Artemis I

Ascent vehicles in the developmental stages of the program are instrumented with Developmental Flight Instrumentation (DFI) sensors. These sensors establish a link between a vehicle and engineers on the ground to communicate conditions experienced during the ascent. These data are then compared to pre-flight predictions used in the design process. The aerodynamic, acoustic, thermal, and structural data are either telemetered to ground stations during the ascent or stored on the vehicle for post-flight recovery and archived at the Huntsville Operations Support Center (HOSC). Following NASA's Artemis I Space Launch System (SLS) launch on November 16, 2020, data from three separate programs are available at the HOSC: Space Shuttle Program (Space Transport System (STS)), Constellation Program (Ares I-X), and Artemis Program (SLS). Availability of these data presents a unique opportunity to examine DFI data from three distinct vehicles and analyze the broad impact of the DFI data on the understanding of transonic aerodynamics. Classical spectrogram and Empirical Mode Decomposition techniques were used to present data in aerodynamically analogous regions on each vehicle. On the SLS and Ares I-X, a region downstream of the Launch Abort System motors was chosen. Comparing SLS and the STS, a region downstream of booster Froward Attach Hardware was selected as analogous flow region. Some other regions of interest were also identified. Although similarities in flow features on three vehicles were identified, some challenges in the comparison were also encountered, especially due to poor temporal and spatial resolution of Shuttle measurements.

Space Shuttle↗

NASA NTRSDate not supplied

Developmental Flight Instrumentation: Review of Space Shuttle, Ares I-X, and Artemis I

Space Shuttle↗

NASA NTRSDate not supplied

Cloud and Precipitation Analyses using Merged Datasets from Two Airborne Microwave Radiometers Covering 10–183 GHz

Microwave radiometers provide valuable insight into the structure and characteristics of clouds and precipitation. In NASA’s airborne remote-sensing arsenal, the Advanced Microwave Precipitation Radiometer (AMPR) and the Conical Scanning Millimeter-wave Imaging Radiometer (CoSMIR) have been used extensively in field campaigns throughout the world. AMPR operates with four channels between 10.7 and 85.5 GHz, while CoSMIR operates with nine channels ranging from 50.3 to 183.31 GHz. Although these datasets provide key information when used separately, the combination of these radiometers covers virtually the full range of frequencies used by the Global Precipitation Measurement (GPM) Microwave Imager (GMI), enabling suborbital observations to compare with GPM spaceborne measurements. This presentation will detail the merger of AMPR and CoSMIR data during two NASA airborne field campaigns: the Integrated Precipitation and Hydrology Experiment (IPHEx) and the Olympic Mountains Experiment / Radar Definition Experiment (OLYMPEX/RADEX). Other field campaigns, such as the Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS), may be included as well. Using the merged brightness temperature dataset, features of precipitating and non-precipitating clouds containing liquid and/or ice hydrometeors will be discussed from selected flight segments. Geophysical retrievals derived from these brightness temperatures using a one-dimensional variational (1DVAR) technique and/or multi-linear regression equations will also be employed in these analyses. Observed transitions between precipitating and non-precipitating systems will be explored in greater detail. Dropsonde data will be used to provide environmental contexts throughout each flight, and additional observations (e.g., from airborne and/or land-based radar) will be incorporated to supplement the radiometer-based results. The broader implications of these results and pathways for future work will also be discussed.

Corey G. Amiot↗

DOE OSTI2023

Chemical looping conversion of CH4/CO2 to syngas on 5wt.%Ni@Ce0.6Zr0.4O2 catalyst: Impact of dynamic accumulation of surface carbon and oxygen vacancies

Syngas, a combination of carbon monoxide (CO) and hydrogen (H2), is a precursor to many chemicals and fuels, contributing to the billion-dollar global hydrocarbon industry. Chemical looping reforming (CLR) of the greenhouse gases methane (CH4) and carbon dioxide (CO2) allows for energy-efficient production of syngas. 5wt.% nickel (Ni) on ceria-zirconia (5wt%Ni@Ce0.6Zr0.4O2) mixed metal oxide catalyst was investigated here to explore pathways for enhanced syngas production on sustainable earth-abundant transition metal supported catalysts. The role of reduction-oxidation (redox) state of the catalyst, and carbon formation on the catalyst surface during chemical looping is explored to drive superior reaction kinetics, conversions, selectivity, and syngas ratios (H2/CO). The bulk and surface structure of the catalyst, along with carbon deposition features were characterized by electron microscopy, X-Ray diffraction, and ex situ Raman spectroscopy. The dynamic evolution of catalysts under CLR reaction conditions and the intrinsic reaction mechanisms were probed with in situ Raman spectroscopy, in situ Fourier transform spectroscopy, dynamic oxygen storage capacity (DOSC), and the Temporal Analysis of Products (TAP) reactor studies. Intrinsic kinetics of syngas production via CLR was correlated to the redox state of catalyst and the participation of nickel-catalyzed multiwalled carbon nanotube (CNT) growth, allowing enhanced CLR reaction performance.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

NASA NTRS1995

GHRS observations of cool, low-gravity star. 2: Flow and turbulent velocities in the outer atmosphere of gamma CRU CIS (M3.4 III)

The Goddard High Resoulution Spectrograph (GHRS) on the Hubble Space Telescope (HST) has been used to obtain medium (R = 20,000) and high (R = 85,000) resoultion UV spectra of chromosphere emission features for the M3.4 III star gamma Cru. Small Science Aperture (SSA) G270M and Echelle-B spectra of selected regions in the 2300-2850 A range were obtained to determine the kinematics of the chromosphere using lines of C2), Fe2, Co2, Si1/2), Ni2, Mn2, and Mg2. Profiles of C2) (UV 0.01) lines and fluorescently excited lines of low optical depth indicate average turbulent velocities (Doppler FWHM) of 30.2 +/- 1.3 and 28.8 +/- 1.3 km/s, respectively. The fluorescent emission lines (mean RV = 21.3 +/- 0.9 km/s) and the wings of the emission components of Fe2 lines (mean RV = 22.8 +/- 0.4 km/s) are approximately at rest relative to the radial velocity of the star (21 km/s), while the C2) lines show a modest inflow (mean RV = 23.1 +/- 0.9 km/s). The more opaque lines of Fe2 and Mg2 exhibit complex profiles resulting from line formation in an optically thick, extended expanding atmosphere. The emission wings of these lines are broadened by multiple scattering, and they are centered near the photospheric radial velocity. Closer to line center, these strong lines show a strong blueshifted self-absorption feature (already seen in IUE data), indicative of formation in an expanding chromosphere, and a previously unseen dip in the profiles on the red side of line center. The absorption components, when extracted using simple Gaussian fits, show strong correlations with the relative optical depths of the lines. The derived absorption flow velocities converge to the photospheric velocity as one examines spectra features formed deeper in the atmosphere. The blueward abosrption velocity increases in magnitude from about 7 to 14 km/s with increasing line optical depth - the strong absorptions directly map the acceleration of the outflowing stellar wind, while the interpretation of the weaker redshifted abosrptions is more ambiguous, indicating either an inflow of material or formation in an extended, spherically expanding outflow. The Mg2 and Fe2 profiles, taken together, imply that the wind speed decreases between the atmospheric layers where the Mg2 and Fe2 self-absorption components are formed. Interstellar absorptions are seen in the resonance lines of Mg2 (UV 1) and Fe2 (UV1) with zero-volt lower levels, at about -3 km/s, consistent with models of the interstellar medium in the direction of gamma Cru. Finally, we have detected the Mg2 'satellite lines' seen in solar spectra obtained above the limb. In gamma Cru these lines are probably fluorescently excited by H Ly beta.

Carpenter, Kenneth G.↗

DOE OSTI2026

Phase Picking Beyond Local Distances: Where Waveform Filtering Still Matters for Deep Learning Models

Waveform filtering is a standard step in traditional seismic phase picking but often receives little attention in deep learning workflows, where models are typically trained on raw or minimally processed waveforms. Although this strategy performs well for local events, we show that performance can degrade substantially at regional distances. To address this limitation, we introduce two ways to incorporate multiband-filtered waveforms into deep learning phase pickers. The stacking approach concatenates filtered inputs along the channel dimension, while the branching approach processes each frequency band through a dedicated network branch before feature fusion. Both approaches can substantially improve performance across epicentral distances of 0° to 20°, but their effectiveness depends strongly on the selected frequency bands. Tests with multiple filter banks show that filter-bank design should be treated as part of model optimization rather than as a fixed preprocessing choice. Grad-CAM analysis of the branching model indicates that band importance varies among waveform samples and across training realizations, with only a weak overall preference for the 0.25 to 0.5 Hz band. These results show that no single filter band is consistently optimal and demonstrate that explicit feature engineering remains valuable for robust deep learning-based seismic phase picking.

58 GEOSCIENCES↗

NASA NTRSDate not supplied

Reversion to the mean of the Solar Wind Input to the Magnetosphere

Earth’s magnetosphere responds to solar wind input over time scales ranging from minutes to days. On minute times scales, researchers seek to identify discontinuities in the solar wind that may trigger transient magnetospheric phenomena such as flux transfer events or substorm onsets. On hour time scales, researchers seek to determine the solar wind phenomena that drive the growth phase of geomagnetic substorms and the storage of magnetic flux within the magnetotail. On longer time scales, researchers seek to determine the solar wind features than propel geomagnetic storms through their initial, main, and recovery phases. Researchers must, perforce, rely upon spacecraft located near the L1 libration point for observations of the solar wind input into the magnetosphere over periods ranging from hours to days. Some of the hazards of doing this are well known. Features seen far upstream at locations away from the Sun-Earth line may never arrive at Earth due to structure in the solar wind transverse to the Sun-Earth line. Alternatively, they may arrive at unexpected times due to errors in calculating their orientation and velocity. Other hazards when predicting the magnetospheric response to varying solar wind conditions are less well understood. When observations of large solar wind drivers at spacecraft far upstream and off the Sun-Earth line are used to select events for study, the ‘reversion to the mean’ effect associated with spatial structure and errors in arrival timing for solar wind features at Earth demands that the magnetosphere will actually be subjected to and respond to solar wind drivers with much lower strengths. Consequently, the magnetospheric response must saturate for the highest solar wind drivers.

David Sibeck↗

DOE OSTI2020

Properties of protein unfolded states suggest broad selection for expanded conformational ensembles

Much attention is being paid to conformational biases in the ensembles of intrinsically disordered proteins. However, it is currently unknown whether or how conformational biases within the disordered ensembles of foldable proteins affect function in vivo. Recently, we demonstrated that water can be a good solvent for unfolded polypeptide chains, even those with a hydrophobic and charged sequence composition typical of folded proteins. These results run counter to the generally accepted model that protein folding begins with hydrophobicity-driven chain collapse. Here we investigate what other features, beyond amino acid composition, govern chain collapse. We found that local clustering of hydrophobic and/or charged residues leads to significant collapse of the unfolded ensemble of pertactin, a secreted autotransporter virulence protein from Bordetella pertussis , as measured by small angle X-ray scattering (SAXS). Sequence patterns that lead to collapse also correlate with increased intermolecular polypeptide chain association and aggregation. Crucially, sequence patterns that support an expanded conformational ensemble enhance pertactin secretion to the bacterial cell surface. Similar sequence pattern features are enriched across the large and diverse family of autotransporter virulence proteins, suggesting sequence patterns that favor an expanded conformational ensemble are under selection for efficient autotransporter protein secretion, a necessary prerequisite for virulence. More broadly, we found that sequence patterns that lead to more expanded conformational ensembles are enriched across water-soluble proteins in general, suggesting protein sequences are under selection to regulate collapse and minimize protein aggregation, in addition to their roles in stabilizing folded protein structures.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2021

Protein model accuracy estimation empowered by deep learning and inter-residue distance prediction in CASP14

Abstract The inter-residue contact prediction and deep learning showed the promise to improve the estimation of protein model accuracy (EMA) in the 13th Critical Assessment of Protein Structure Prediction (CASP13). To further leverage the improved inter-residue distance predictions to enhance EMA, during the 2020 CASP14 experiment, we integrated several new inter-residue distance features with the existing model quality assessment features in several deep learning methods to predict the quality of protein structural models. According to the evaluation of performance in selecting the best model from the models of CASP14 targets, our three multi-model predictors of estimating model accuracy (MULTICOM-CONSTRUCT, MULTICOM-AI, and MULTICOM-CLUSTER) achieve the averaged loss of 0.073, 0.079, and 0.081, respectively, in terms of the global distance test score (GDT-TS). The three methods are ranked first, second, and third out of all 68 CASP14 predictors. MULTICOM-DEEP, the single-model predictor of estimating model accuracy (EMA), is ranked within top 10 among all the single-model EMA methods according to GDT-TS score loss. The results demonstrate that inter-residue distance features are valuable inputs for deep learning to predict the quality of protein structural models. However, larger training datasets and better ways of leveraging inter-residue distance information are needed to fully explore its potentials.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2023

An automated and portable method for selecting an optimal GPU frequency

Power consumption poses a significant challenge in current and emerging graphics processing unit (GPU) enabled high-performance computing systems. In modern GPUs, dynamic voltage frequency scaling (DVFS) appears to be a reliable control to regulate power consumption and performance. However, the DVFS design space is large - hence, brute-force approaches are infeasible to select the optimal frequency. Furthermore, no single frequency can be universally optimal for applications with varying computational intensities. Thus, the application's complexity and the availability of a wide range of frequency settings are a challenge in selecting the optimal frequency configuration for a given GPU workload. To that end, this paper proposes a systematic approach that consists of three steps. The feature characterization study identifies the fine-grain GPU utilization metrics that influence the power consumption and execution time of a given workload. To understand the performance, power, and energy consumption behaviors of a workload across GPU's DVFS design space, we derived analytical power and performance models using the identified fine-grain features. Here, it is shown that the same set of GPU utilization metrics can estimate both the power consumption and execution time while being agnostic of changes to frequency and input sizes. Applying a power control with the single objective of reducing power may cause performance degradation, leading to more energy consumption. A multi-objective approach is proposed to select the optimal GPU DVFS configuration for a workload that reduces power consumption with negligible degradation in performance. The evaluation was conducted using SPEC ACCEL benchmarks and three real applications - NAMD LAMMPS, and LSTM on NVIDIA GV100, GA100, and AMD MI210 GPUs. On average, real applications showed 29.6% energy savings with a performance loss of 5.2% on GA100 and 22.6% energy savings with a performance loss of 4.7% on GV100. Moreover, the proposed models are portable to real applications, GPU architectures, and vendors, and require metric collection at only the default frequency rather than all supported DVFS configurations. Additionally, we conducted a comparison between our models and the GPU assembly instructions (PTX)-based static models. The results revealed a significant reduction in the average error rates, with a decrease from 19.7% to 3.1% for power models and from 29.4% to 5.2% for performance models.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1991

Emission-line and continuum properties of 92 bright QSOs - Luminosity dependence and differences between radio-selected and optically selected samples

The emission line and continuum properties of new spectra of 92 QSOs are investigated. A detailed deconvolution of the semiforbidden C III 1909 A blend reveals that the secondary feature is almost certainly Al III 1859 A. A study of the Mg II 2800 A emission-line profiles suggests that the broad 'shoulders' often observed are probably intrinsic to the Mg II emission line. It is shown that the QSOs discovered by slitless spectroscopic methods do not differ significantly from those discovered by UV excess surveys. Significant correlations with luminosity are found for W(Mg II), W(semiforbidden C III), I(C III)/I(Mg II), and the continuum ratio fv(1909 A)/fv(2800 A). All of the significant correlations are substantially stronger for a radio-selected subsample compared to a radio-quiet subsample. The C IV and semiforbidden C III emission lines each have a mean blueshift relative to Mg II of 450 km/s while semiforbidden C IV has a mean blueshift relative to semiforbidden C III of 200 km/s.

Steidel, Charles C.↗

DOE OSTI2023

Ethanol Upgrading to n -Butanol Using Transition-Metal-Incorporated Poly(triazine)imide Frameworks

The upgrading of ethanol to n-butanol was performed using a molecular catalyst integrated into a carbon nitride support, one of the first examples of a supported molecular catalyst performing the Guerbet process. Initial studies using crystalline poly(triazine)imide (PTI) with lithium or transition-metal cations imbedded in the support together with a base as the catalyst system did not produce any significant amounts of n-butanol. However, when using the catalyst material formed by treatment of PTI-LiCl with [(Cp*)IrCl 2 ] 2 (Cp* = pentamethylcyclopentadienyl) along with sodium hydroxide, a 59% selectivity for butanol (13% yield) was obtained at 145 °C. This PTI-(Cp*)Ir material exhibited distinct UV–vis absorption features and powder X-ray diffractions which differ from those of the parent PTI-LiCl and [(Cp*)IrCl 2 ] 2 . Here, the PTI-(Cp*)Ir material was found to have a metal loading of 27% iridium per empirical unit of the framework. Along with the formation of n-butanol from the Guerbet reaction, the presence of higher chain alcohols was also observed.

Guerbet↗

DOE OSTI2023

Model selection and signal extraction using Gaussian Process regression

We present a novel computational approach for extracting localized signals from smooth background distributions. We focus on datasets that can be naturally presented as binned integer counts, demonstrating our procedure on the CERN open dataset with the Higgs boson signature, from the ATLAS collaboration at the Large Hadron Collider. Our approach is based on Gaussian Process (GP) regression — a powerful and flexible machine learning technique which has allowed us to model the background without specifying its functional form explicitly and separately measure the background and signal contributions in a robust and reproducible manner. Unlike functional fits, our GP-regression-based approach does not need to be constantly updated as more data becomes available. We discuss how to select the GP kernel type, considering trade-offs between kernel complexity and its ability to capture the features of the background distribution. We show that our GP framework can be used to detect the Higgs boson resonance in the data with more statistical significance than a polynomial fit specifically tailored to the dataset. Finally, we use Markov Chain Monte Carlo (MCMC) sampling to confirm the statistical significance of the extracted Higgs signature.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2021

Multi-scale Coupled Processes Modeling of Fractures as Porous, Interfacial and Granular Systems from Rock Images with the Numerical Manifold Method

Abstract The greatest challenges of rigorously modeling coupled hydro-mechanical processes in fractured rocks at different scales are associated with computational geometry. In addition, selections of continuous or discontinuous models, physical laws, and coupling priorities at different scales based on different geometric features determine the applicability of a numerical model for a certain type of problem. In this study, we present our multi-scale modeling capabilities that have been developed based on the numerical manifold method for analyzing coupled hydro-mechanical processes in fractured rocks. Based on their geometric features, the fractures are modeled as continua—finite-thickness porous zones, and discontinua—discontinuous interfaces and microscale asperities and granular systems. Different governing equations, physical laws, coupling priorities, and approaches for addressing fracture intersections and shearing are then applied to describe these. We applied these models to simulate coupled processes in fractured rocks using realistic geometry obtained from rock images at different scales. We first calculated shearing of a single fracture with different models and demonstrated the impacts of asperities on shearing. We then applied the continuous and discontinuous models to simulate a network of rough fractures, demonstrating that contact dynamics contribute significantly to the geometric, multi-physical evolution of systems where rough fractures are not mineral filled. For a discrete fracture network, our coupled processes modeling demonstrates that shearing of the discrete fractures can have a major impact on stress and pore pressure distribution. Lastly, we applied the discontinuous granular model to simulate evolution of a complex granular system with a deformation band, demonstrating that the deformation band can dominate contact dynamics, the structural and the stress evolution of the granular system.

58 GEOSCIENCES↗

DOE OSTI2020

Dual-atom Ag 2 /graphene catalyst for efficient electroreduction of CO 2 to CO

Electrochemical reduction of CO 2 into value-added carbon compounds offers a promising strategy to mitigate global warming, but present challenges for chemistry due to the poor selectivity and stability of electrocatalysts. In this work, we report a dual-atom Ag 2 /graphene catalyst featuring well-defined AgN 3 -AgN 3 active site for CO 2 electrochemical reduction. This dual-atom catalyst can drive CO 2 reduction reaction at a potential as high as -0.25 V, and exhibit excellent CO Faradic efficiency up to 93.4 % with a current density of 11.87 mA cm -2 at -0.7 V and long-term stability, far surpassing the single-atom Ag 1 /graphene and the traditional silver nanoparticle catalysts. Finally, DFT calculations reveal that the dual-atom Ag site lowers the barrier for the formation of *COOH by stabilizing the *CO 2 through the concomitant interactions with the C and an O atom of CO 2 , resulting in excellent catalytic performance.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗