SEARCH · Engineering Papers

Results for “Feature importance”

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 127 records · Page 7

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Thermodynamically Optimized Machine-Learned Reaction Coordinates for Hydrophobic Ligand Dissociation

Ligand unbinding is mediated by its free energy change, which has intertwined contributions from both energy and entropy. It is important, but not easy, to quantify their individual contributions to the free energy profile. We model hydrophobic ligand unbinding for two systems, a methane particle and a C 60 fullerene, both unbinding from hydrophobic pockets in all-atom water. Using a modified deep learning framework, we learn a thermodynamically optimized reaction coordinate to describe the hydrophobic ligand dissociation for both systems. Interpretation of these reaction coordinates reveals the roles of entropic and enthalpic forces as the ligand and pocket sizes change. In both cases, we observe that the free-energy barrier to unbinding is dominated by entropy considerations. Furthermore, the process of methane unbinding is driven by methane solvation, while fullerene unbinding is driven first by pocket wetting and then fullerene wetting. For both solutes, the direct importance of the distance from the binding pocket to the learned reaction coordinate is present, but low. Furthermore, our framework and subsequent feature important analysis thus give useful thermodynamic insight into hydrophobic ligand dissociation problems that are otherwise difficult to glean.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS2008

Algorithms for Spectral Decomposition with Applications to Optical Plume Anomaly Detection

The analysis of spectral signals for features that represent physical phenomenon is ubiquitous in the science and engineering communities. There are two main approaches that can be taken to extract relevant features from these high-dimensional data streams. The first set of approaches relies on extracting features using a physics-based paradigm where the underlying physical mechanism that generates the spectra is used to infer the most important features in the data stream. We focus on a complementary methodology that uses a data-driven technique that is informed by the underlying physics but also has the ability to adapt to unmodeled system attributes and dynamics. We discuss the following four algorithms: Spectral Decomposition Algorithm (SDA), Non-Negative Matrix Factorization (NMF), Independent Component Analysis (ICA) and Principal Components Analysis (PCA) and compare their performance on a spectral emulator which we use to generate artificial data with known statistical properties. This spectral emulator mimics the real-world phenomena arising from the plume of the space shuttle main engine and can be used to validate the results that arise from various spectral decomposition algorithms and is very useful for situations where real-world systems have very low probabilities of fault or failure. Our results indicate that methods like SDA and NMF provide a straightforward way of incorporating prior physical knowledge while NMF with a tuning mechanism can give superior performance on some tests. We demonstrate these algorithms to detect potential system-health issues on data from a spectral emulator with tunable health parameters.

Srivastava, Askok N.↗

DOE OSTI2025

Predicting Initial Trans-Membrane Pressure for Optimized Operations in UF Unit Using Random Forest

With the growing scarcity of freshwater, innovative process design mechanisms like Reverse Osmosis (RO) are increasingly gaining attention among water treatment utilities to address the rising demand. Ensuring reliable water production necessitates efficient resource utilization, minimizing downtime in (ultra-filtration) UF systems. Recent advancements in machine learning (ML) have enabled the development of accurate data-driven models for Model Predictive Control (MPC), often requiring minimal prior knowledge of underlying physical processes. In this study, we present predictive regression models based on Random Forest (RF) and Auto-Regressive (AR) approaches to forecast the initial Trans-Membrane Pressure (TMP) for each filtration cycle in data generated by Direct Potable Reuse (DPR) systems. The proposed RF-based model demonstrates superior performance compared to baseline methods, including historical mean, Last Observation Carried Forward (LOCF), and naïve AR models, across various forecasting horizons in terms of root mean square error (RMSE) metric. To evaluate how different classes of process variables contribute to TMP dynamics over time, we examine the feature importance of independent covariates across multiple forecast horizons. This analysis provides insight into the temporal relevance of operational and sensor-derived features, guiding control and monitoring strategies. Additionally, the impact of hyperparameter tuning on TMP prediction performance is studied for both direct and recursive RF modelling approaches across increasing forecast horizons. Accurate prediction of initial TMP is critical for optimizing RO operations, as it enables the development of robust modelling frameworks by accurately estimating membrane fouling trends, thereby enhancing process efficiency and long-term reliability. The demonstrated efficacy of the RF-based approach highlights its potential as a tool for real-time decision-making in water treatment systems, paving the way for advanced process optimization and sustainable water resource management.

Mukherjee, Subrata [ORNL] (ORCID:0000000309930338)↗

DOE OSTI2023

Future Atmospheric Rivers and Impacts on Precipitation: Overview of the ARTMIP Tier 2 High–Resolution Global Warming Experiment

Atmospheric rivers (ARs) are long, narrow synoptic scale weather features important for Earth’s hydrological cycle typically transporting water vapor poleward, delivering precipitation important for local climates. Understanding ARs in a warming climate is problematic because the AR response to climate change is tied to how the feature is defined. The Atmospheric River Tracking Method Intercomparison Project (ARTMIP) provides insights into this problem by comparing 16 atmospheric river detection tools (ARDTs) to a common data set consisting of high resolution climate change simulations from a global atmospheric general circulation model. ARDTs mostly show increases in frequency and intensity, but the scale of the response is largely dependent on algorithmic criteria. Across ARDTs, bulk characteristics suggest intensity and spatial footprint are inversely correlated, and most focus regions experience increases in precipitation volume coming from extreme ARs. The spread of the AR precipitation response under climate change is large and dependent on ARDT selection.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2023

Future Atmospheric Rivers and Impacts on Precipitation: Overview of the ARTMIP Tier2 High-Resolution Global Warming Experiment

Atmospheric rivers (ARs) are long, narrow synoptic scale weather features important for Earth’s hydrological cycle typically transporting water vapor poleward, delivering precipitation important for local climates. Understanding ARs in a warming climate is problematic because the AR response to climate change is tied to how the feature is defined. The Atmospheric River Tracking Method Intercomparison Project (ARTMIP) provides insights into this problem by comparing 16 atmospheric river detection tools (ARDTs) to a common dataset consisting of high resolution climate change simulations from a global atmospheric general circulation model. ARDTs mostly show increases in frequency and intensity, but the scale of the response is largely dependent on algorithmic criteria. Across ARDTs, bulk characteristics suggest intensity and spatial footprint are inversely correlated, and most focus regions experience increases in precipitation volume coming from extreme ARs. The spread of the AR precipitation response under climate change is large and dependent on ARDT selection.

Atmospheric Rivers↗

NASA NTRS1973

QA4, a language for artificial intelligence.

Introduction of a language for problem solving and specifically robot planning, program verification, and synthesis and theorem proving. This language, called question-answerer 4 (QA4), embodies many features that have been found useful for constructing problem solvers but have to be programmed explicitly by the user of a conventional language. The most important features of QA4 are described, and examples are provided for most of the material introduced. Language features include backtracking, parallel processing, pattern matching, set manipulation, and pattern-triggered function activation. The language is most convenient for use in an interactive way and has extensive trace and edit facilities.

Derksen, J. A. C.↗

DOE OSTI2024

Insights into Prismatic Loop Formation in Irradiated Fe–Cr Alloys from Hypothesis-Driven Active Learning and Causal Analysis

Neutron and electron irradiation experimental studies conducted on body-centered cubic Fe and Fe–Cr alloys have established two prismatic dislocation loop populations, which have Burgers vectors of either a/2$\langle$111$\rangle$ or a$\langle$100$\rangle$. Here, the loop formation depends on factors such as dose (D), dose rate (D rt ), temperature (T), chromium content (Cr%), and other alloying elements. Hence, it is important to understand how irradiation-induced dislocation loops evolve conditional upon the loop characteristics, such as loop density (DD), average loop size d̅, and irradiation parameters (D, D rt , T, and irradiation type), which is still an active area of research. To understand these complex structure–property relationships, machine learning (ML) is employed in a three-step approach. This includes imputing missing data with a k-nearest neighbor, generating functionalized features, and assessing feature importance with random forest classification and regression. Physics-based features are incorporated in a hypothesis-driven active learning scheme to overcome data unavailability challenges. Insights obtained from ML models (i) to categorize dislocation loop types, show the highest correlation with d̅; (ii) Log(DD), obtained through mathematical formulations involving D, Cr%, d̅, and T (e.g., Log(DD) ~ D + exp(-Cr%) + 1/d̅ and log(DD) ~ D + exp(-Cr%) + 1/T). Hypothesis-driven active learning is able to predict Log(DD) in which the experimental date is not known. Causal models verify cause–effect relationships for dislocation loop classification and irradiation factors in FeCr alloys.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

NASA NTRS2013

A Small Mission Featuring an Imaging X-ray Polarimeter with High Sensitivity

We present a detailed description of a small mission capable of obtaining high precision and meaningful measurement of the X‐ray polarization of a variety of different classes of cosmic X‐ray sources. Compared to other ideas that have been suggested this experiment has demonstrated in the laboratory a number of extremely important features relevant to the ultimate selection of such a mission by a funding agency. The most important of these questions are: 1) Have you demonstrated the sensitivity to a polarized beam at the energies of interest (i.e. the energies which represent the majority (not the minority) of detected photons from the X‐ray source of interest? 2) Have you demonstrated that the device's sensitivity to an unpolarized beam is really negligible and/or quantified the impact of any systematic effects upon actual measurements? We present our answers to these questions backed up by laboratory measurements and give an overview of the mission.

Weisskopf, Martin C.↗

DOE OSTI2025

Machine Learning-Assisted Recovery of Delicate Kinetic Information from Transient Reactor Experiments

Identifying active sites and their roles in chemical reaction steps remains a vital challenge in heterogeneous catalysis. Transient experiments offer a unique way to probe active sites and distinguish subtle kinetic features. Although physics-based analysis methods may be well-developed, they can be highly susceptible to experimental noise, and smoothing methods may erase or even distort important features; a smooth curve is not always the best curve. We demonstrate a new workflow for the direct interpretation of intrinsic kinetic information from exit flux curves measured in transient reactor experiments. This workflow contains three artificial neural networks (ANNs), including a noise reducer, a concentration predictor, and a rate predictor to analyze experimental data, followed by the virtual TAP (VTAP) physics-based reactor model and density functional theory (DFT) calculations of adsorption energies on specific sites. We use this workflow to analyze the data from experiments titrating Pt/Al 2 O 3 and Pt/SiO 2 catalysts with carbon monoxide (CO) in the temporal analysis of products (TAP) reactor. Our workflow separates the time-evolving chemical reaction and mass transfer information contained in the TAP pulse response. The existence of strong- and weak-binding sites on the Pt/Al 2 O 3 catalyst is observed in the catalyst titration experiment in the transient reactor. The structures of the strong- and weak-binding sites are then identified by using DFT calculations. We find that the Pt/SiO 2 catalyst has only strong-binding sites, which aligns with the inactive support effect of SiO 2 . We demonstrate how machine learning methods provide unique insights with high-resolution data analysis that cannot be achieved by using state-of-the-art physics-based methods.

Adsorption↗

NASA NTRS1992

A neuro-fuzzy architecture for real-time applications

Neural networks and fuzzy expert systems perform the same task of functional mapping using entirely different approaches. Each approach has certain unique features. The ability to learn specific input-output mappings from large input/output data possibly corrupted by noise and the ability to adapt or continue learning are some important features of neural networks. Fuzzy expert systems are known for their ability to deal with fuzzy information and incomplete/imprecise data in a structured, logical way. Since both of these techniques implement the same task (that of functional mapping--we regard 'inferencing' as one specific category under this class), a fusion of the two concepts that retains their unique features while overcoming their individual drawbacks will have excellent applications in the real world. In this paper, we arrive at a new architecture by fusing the two concepts. The architecture has the trainability/adaptibility (based on input/output observations) property of the neural networks and the architectural features that are unique to fuzzy expert systems. It also does not require specific information such as fuzzy rules, defuzzification procedure used, etc., though any such information can be integrated into the architecture. We show that this architecture can provide better performance than is possible from a single two or three layer feedforward neural network. Further, we show that this new architecture can be used as an efficient vehicle for hardware implementation of complex fuzzy expert systems for real-time applications. A numerical example is provided to show the potential of this approach.

Ramamoorthy, P. A.↗

DOE OSTI2024

Improved Subseasonal Forecasting of Extreme Polar Vortices Using Machine Learning

Our research was focused on forecasting the position and shape of the winter stratospheric polar vortex at a subseasonal timescale of 15 days in advance. To achieve this, we employed both statistical and neural network machine learning techniques. The analysis was performed on 42 winter seasons of reanalysis data provided by NASA giving us a total of 6,342 days of data. The state of the polar vortex for determined by using geometric moments to calculate the centroid latitude and the aspect ratio of an ellipse fit onto the vortex. Timeseries for thirty additional precursors were calculated to help improve the predictive capabilities of the algorithm. Feature importance of these precursors was performed using random forest to measure the predictive importance and the ideal number of precursors. Then, using the precursors identified as important, various statistical methods were tested for predictive accuracy with random forest and nearest neighbor performing the best. An echo state network, a type of recurrent neural network that features sparsely connected hidden layer and a reduced number of trainable parameters that allows for rapid training and testing, was also implemented for the forecasting problem. Hyperparameter tuning was performed for each methods using a subset of the training data. The algorithms were trained and tuned on the first 41 years of data, then tested for accuracy on the final year. In general, the centroid latitude of the polar vortex proved easier to predict than the aspect ratio across all algorithms. Random forest outperformed other statistical forecasting algorithms overall but struggled to predict extreme values. Forecasting from echo state network suggested a strong predictive capability past 15 days, but further work is required to fully realize the potential of recurrent neural network approaches.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2024

Jet classification using high-level features from anatomy of top jets

Recent advancements in deep learning models have significantly enhanced jet classification performance by analyzing low-level features (LLFs). However, this approach often leads to less interpretable models, emphasizing the need to understand the decision-making process and to identify the high-level features (HLFs) crucial for explaining jet classification. To address this, we consider the top jet tagging problems and introduce an analysis model (AM) that analyzes selected HLFs designed to capture important features of top jets. Our AM mainly consists of the following three modules: a relation network analyzing two-point energy correlations, mathematical morphology and Minkowski functionals for generalizing jet constituent multiplicities, and a recursive neural network analyzing subjet constituent multiplicity to enhance sensitivity to subjet color charges. We demonstrate that our AM achieves performance comparable to the Particle Transformer (ParT) while requiring fewer computational resources in a comparison of top jet tagging using jets simulated at the hadronic calorimeter angular resolution scale. Furthermore, as a more constrained architecture than ParT, the AM exhibits smaller training uncertainties because of the bias-variance tradeoff. We also compare the information content of AM and ParT by decorrelating the features already learned by AM. Lastly, we briefly comment on the results of AM with finer angular resolution inputs.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2022

Forecasting induced seismicity in Oklahoma using machine learning methods

Oklahoma earthquakes in the past decade have been mostly associated with wastewater injection. Here we use a machine learning technique—the Random Forest to forecast induced seismicity rate in Oklahoma based on injection-related parameters. We split the data into training (2011.01–2015.05) and test (2015.06–2020.12) periods. The model forecasts seismicity rate during the test period based on input features, including operational parameters (injection rate and pressure), geological information (depth to basement), and modeled pore pressure and poroelastic stress. The results show overall good match with observed seismicity rate (adjusted R 2 of 0.75). The model shows that pore pressure rate and poroelastic stressing rates are the two most important features in forecasting. The absolute values of pore pressure and poroelastic stress, and the injection rate itself, are less important than the stressing rates. These findings further emphasize that temporal changes of stressing rates would lead to significant changes in seismicity rates.

58 GEOSCIENCES↗

DOE OSTI2023

Quantum dynamical effects of vibrational strong coupling in chemical reactivity

Abstract Recent experiments suggest that ground state chemical reactivity can be modified when placing molecular systems inside infrared cavities where molecular vibrations are strongly coupled to electromagnetic radiation. This phenomenon lacks a firm theoretical explanation. Here, we employ an exact quantum dynamics approach to investigate a model of cavity-modified chemical reactions in the condensed phase. The model contains the coupling of the reaction coordinate to a generic solvent, cavity coupling to either the reaction coordinate or a non-reactive mode, and the coupling of the cavity to lossy modes. Thus, many of the most important features needed for realistic modeling of the cavity modification of chemical reactions are included. We find that when a molecule is coupled to an optical cavity it is essential to treat the problem quantum mechanically to obtain a quantitative account of alterations to reactivity. We find sizable and sharp changes in the rate constant that are associated with quantum mechanical state splittings and resonances. The features that emerge from our simulations are closer to those observed in experiments than are previous calculations, even for realistically small values of coupling and cavity loss. This work highlights the importance of a fully quantum treatment of vibrational polariton chemistry.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

Variability of near-fault seismic risk to reinforced concrete buildings based on high-resolution physics-based ground motion simulations

Broadband physics-based simulated earthquake ground motions are utilized to characterize the regional-scale seismic risk to modern reinforced concrete (RC) structures. A highly dense dataset of ground motions covering a 100-km urn:x-wiley:00988847:media:eqe3413:eqe3413-math-0001 40-km domain was generated using kinematic fault rupture models with varying rupture characteristics to represent shallow crustal earthquakes and resolved up to frequencies of 5 Hz. Over 40,000 nonlinear response history simulations of short- and mid-rise RC special moment frame buildings were conducted using simulation models that are capable of representing nonlinear behavior and component deterioration effects. The spatial variability of structural risk within a single earthquake scenario and between different rupture scenarios is examined, and the regions of strongest directivity effects and highest structural demands are identified. The structural demands may vary by factor of up to 8.0 at very short distances from the fault, and the large dispersion in the demands decreases significantly beyond a distance of 15 km. The interstory drift and member rotation demands are substantially impacted by important features of the geological structure and the characteristics of the rupture scenarios, particularly the presence of localized high-slip regions. The frequency characteristics of the structures are found to play an important role in determining the effects of near-fault ground motions on the structural response and expected damage. The results of this study suggest that the simulated ground motions, particularly those generated using the hybrid rupture approach, may offer reasonable structural risk estimates for low-frequency structures and conservative estimates for high-frequency structures.

58 GEOSCIENCES↗

NASA NTRS1993

Impact basins on Venus and some interplanetary comparisons

Impact is one of the many processes that have shaped the surface of Venus. The largest impact craters, basins, are important features affecting the evolution of the terrestrial planets. Because Venus has an atmosphere, a gravity similar to Earth's, and a surface target with a high geothermal gradient, venusian basins provide an important comparative set of data to test our ideas about basin-forming impacts and their geological effects on the evolution of the crusts of the terrestrial planets.

Spudis, Paul D.↗

DOE OSTI2022

Machine Learning in Infectious Disease for Risk Factor Identification and Hypothesis Generation: Proof of Concept Using Invasive Candidiasis

Machine learning (ML) models can handle large data sets without assuming underlying relationships and can be useful for evaluating disease characteristics, yet they are more commonly used for predicting individual disease risk than for identifying factors at the population level. We offer a proof of concept applying random forest (RF) algorithms to Candida-positive hospital encounters in an electronic health record database of patients in the United States. Candida-positive encounters were extracted from the Cerner HealthFacts database; invasive infections were laboratory-positive sterile site Candida infections. Features included demographics, admission source, care setting, physician specialty, diagnostic and procedure codes, and medications received before the first positive Candida culture. We used RF to assess risk factors for 3 outcomes: any invasive candidiasis (IC) vs non-IC, within-species IC vs non-IC (eg, invasive C. glabrata vs noninvasive C. glabrata), and between-species IC (eg, invasive C. glabrata vs all other IC). Fourteen of 169 (8%) variables were consistently identified as important features in the ML models. When evaluating within-species IC, for example, invasive C. glabrata vs non-invasive C. glabrata, we identified known features like central venous catheters, intensive care unit stay, and gastrointestinal operations. In contrast, important variables for invasive C. glabrata vs all other IC included renal disease and medications like diabetes therapeutics, cholesterol medications, and antiarrhythmics. Known and novel risk factors for IC were identified using ML, demonstrating the hypothesis-generating utility of this approach for infectious disease conditions about which less is known, specifically at the species level or for rarer diseases.

60 APPLIED LIFE SCIENCES↗

NASA NTRS2008

SiC/SiC Composites: The Effect of Fiber Type and Fiber Architecture on Mechanical Properties

Woven SiC/SiC composites represent a broad family of composites with a broad range of properties which are of interest for many energy-based and aero-based applications. Two important features of SiC/SiC composites which one must consider are the reinforcing fibers themselves and the fiber-architecture they are formed into. The range of choices for these two features can result in a wide range of elastic, mechanical, thermal, and electrical properties. In this presentation, it will be demonstrated how the effect of fiber-type and fiber architecture effects the important property of "matrix cracking stress" for slurry-cast melt-infiltrated SiC matrix composites, which is often considered to be a critical design parameter for this system of composites.

Morscher, Gregory N.↗