Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “model interpretability”

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 901 records · Page 50
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Evolution of a Geological Model for Co-Producing Electricity at the Blackburn Oil Field, Nevada: Preprint

Oil fields characterized by significant proportions of hot water in the reservoir offer a unique opportunity for geothermal energy co-production and utilization. Development of an accurate geological model is essential in understanding the subsurface. However, this is a challenge at the preliminary stages of resource assessment when data on the subsurface parameters and subterranean fault network is scarce. Our focus in this paper is to explore a methodology for developing an accurate conceptual geological model of the complexly faulted reservoir at the Blackburn Field, Nevada, to create a numerical simulation that will eventually be used to assess the potential of co-producing electricity. There are four steps to the evolution of this model: (1) collection of public data and research on the geological setting of the system; (2) digitization of seismic time-interpretation and cross-sections to create the most representative model of the subsurface; and (3) creation of a natural-state numerical model to simulate the thermal gradient through the subsurface. Utilizing publicly available data of the geological system and Leapfrog conceptual modeling software, we refined a preliminary model to be more representative of the subsurface geology of an active geothermal exploration stage project. The conceptual geological model developed has been imported into VOLSUNG reservoir modeling software to create a natural-state model, further characterizing the Blackburn Field reservoir's thermal gradient. In the future, new seismic data will be utilized to further evolve the conceptual geological model for numerical simulations.

Basin and Range province↗
DOE OSTI2022

A fully Lagrangian, non-parametric bias model for dark matter halos

Here, we present a non-parametric Lagrangian biasing model and fit the ratio of the halo and mass densities at the field level using the mass-weighted halo field in the AbacusSummit simulations at z=0.5. Unlike the perturbative halo bias model that has been widely used in interpreting the observed large-scale structure traced by galaxies, we find a non-negative halo-to-mass ratio that increases monotonically with the linear overdensity δ 1 in the initial Lagrangian space. The bias expansion, however, does not guarantee non-negativity of the halo counts, and may lead to rising halo number counts at negative overdensities. The shape of the halo-to-mass ratio is unlikely to be described by a polynomial function of δ 1 and other quantities. Especially for massive halos with 6×10 12 h -1 M ⊙ , the halo-to-mass ratio starts soaring up at δ 1 >0, substantially different from the predictions of the bias expansion. We show that for the halo masses we consider (M>3×10 11 h -1 M ⊙ ) a non-parametric halo-to-mass ratio as a function of δ 1 and its local derivative ∇^2δ 1 can recover the halo power spectra to sub-percent level accuracy for wavenumbers k=0.01-0.1 h Mpc -1 given a proper smoothing scale to filter the initial density field, even though we do not fit the power spectrum directly. However, there is mild dependence of the recovery of the halo power spectrum on the smoothing scale and other input parameters. At k<0.01 h Mpc -1 and for massive halos with M>6x10 12 h -1 M ⊙ , our non-parametric model leads to a few percent overestimation of the halo power spectrum, indicating the need for larger or multiple smoothing scales. The halo-to-mass ratios obtained qualitatively agree with intuitions from extended Press-Schechter theory. We compare our framework to the bias expansion and discuss possible extensions.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Interpreting Ionic Conductivity for Polymer Electrolyte Fuel Cell Catalyst Layers with Electrochemical Impedance Spectroscopy and Transmission Line Modeling

A cathode catalyst layer containing optimally distributed ionomer is critical to reduce the platinum loading and increase its utilization in polymer electrolyte fuel cells. Here, electrochemical impedance spectroscopy (EIS) was used to measure effective ionic conductivity of pseudo catalyst layers (PCLs) at a relative humidity (RH) range of 50%-120%. These results are compared to previous work using the hydrogen pump (HP) method. EIS effective ionic conductivity results reported here are higher than those from the HP because in the HP set-up ionic pathways must be effectively connected through the PCL to be counted, whereas in the EIS measurement, ionomer segments that are in contact with the membrane but are not effectively connected all the way through the PCL can be detected. Double layer capacitances and effective ionic conductivities of Pt/C catalyst layers with various supports and ionomer to carbon (I/C) ratios were studied. High surface area carbon support resulted in a lower effective ionic conductivity compared to the graphitized carbon support due to worse ionomer dispersion. Effective ionic conductivities of Pt/C layers were compared to that of PCLs. On average, effective ionic conductivities of Pt/C layers were higher than PCLs because of possible carbon agglomeratio n within the PCLs.

25 ENERGY STORAGE↗
DOE OSTI2026

Commutative Algebra Modeling in Materials Science – A Case Study on Metal–Organic Frameworks (MOFs)

Metal-organic frameworks (MOFs) are a class of important crystalline and highly porous materials whose hierarchical geometry and chemistry hinder interpretable predictions in materials properties. Commutative algebra is a branch of abstract algebra that has been rarely applied in data and material sciences. We introduce the first ever commutative algebra modeling and prediction in materials science. Specifically, category-specific commutative algebra (CSCA) is proposed as a new framework for MOF representation and learning. It integrates element-based categorization with multiscale algebraic invariants to encode both local coordination motifs and global network organization of MOFs. These algebraically consistent, chemically aware representations enable compact, interpretable, and data efficient modeling of MOF properties such as Henry’s constants and uptake capacities for common gases. Compared to traditional geometric and graph-based approaches, CSCA achieves comparable or superior predictive accuracy while substantially improving interpretability and stability across data sets. By aligning commutative algebra with the chemical hierarchy, the CSCA establishes a rigorous and generalizable paradigm for understanding structure and property relationships in porous materials and provides a nonlinear algebra-based framework for data-driven material discovery.

Khaemba, Caleb S.↗
DOE OSTI2022

Elucidating the temperature dependence of TRIP in Q&P steels using synchrotron X-Ray diffraction, constituent phase properties, and strain-based kinetics models

Understanding the deformation-induced martensitic transformation (DIMT) is critical for interpreting the structure-property relationships that govern the performance of transformation-induced plasticity (TRIP) assisted steels. However, modern TRIP-assisted steels often exhibit DIMT kinetics that are not easily captured by existing empirical models based on bulk tensile strain. We address this challenge by combined bulk uniaxial tensile tests and in-situ high energy synchrotron X-ray diffraction, which resolved the phase volume fractions, stress-strain response, and microstructure evolution of each constituent phase. A modification of the Olson-Cohen model is implemented, which describes the martensitic transformation kinetics as a function of the estimated partitioned strain in austenite, rather than the bulk tensile strain. This DIMT kinetic model is used as a framework to clarify the root cause of an insufficiently understood toughness trough reported for TRIP-assisted steels during deformation at elevated temperatures. Here, the importance of the temperature-dependent toughness is discussed, based on the opportunity to modify deformation processes to tailor the DIMT kinetics and mechanical properties during forming and in service.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Characterization of contaminants in the Lyman-alpha forest auto-correlation with DESI

Baryon Acoustic Oscillations can be measured with sub-percent precision above redshift two with the Lyman-α (Lyα) forest auto-correlation and its cross-correlation with quasar positions. This is one of the key goals of the Dark Energy Spectroscopic Instrument (DESI) which started its main survey in May 2021. We present in this paper a study of the contaminants to the Lyα forest which are mainly caused by correlated signals introduced by the spectroscopic data processing pipeline as well as astrophysical contaminants due to foreground absorption in the intergalactic medium. Notably, an excess signal caused by the sky background subtraction noise is present in the Lyα auto-correlation in the first line-of-sight separation bin. We use synthetic data to isolate this contribution, we also characterize the effect of spectro-photometric calibration noise, and propose a simple model to account for both effects in the analysis of the Lyα forest. We then measure the auto-correlation of the quasar flux transmission fraction of low redshift quasars, where there is no Lyα forest absorption but only its contaminants. We demonstrate that we can interpret the data with a two-component model: data processing noise and triply ionized Silicon and Carbon auto-correlations. This result can be used to improve the modeling of the Lyα auto-correlation function measured with DESI.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Effect of Anoxic Iron Corrosion on WIPP Brine Geochemistry FY23 Final Report (U)

A 280-day study was completed to evaluate the effect of zero-valent iron (Fe 0 ) on the Waste Isolation Pilot Plant (WIPP) brine geochemistry under anticipated reducing conditions. Hydrogen (H 2 ) gas is expected to be present in the repository after closure due to the anoxic corrosion of a vast quantity of iron contained in the waste forms disposed at WIPP; therefore, a background argon atmosphere containing H 2 was chosen for this study. WIPP groundwater brine pH and E h will impact the mobility and fate of plutonium within the repository. Modeling and laboratory results for Castile WIPP brine indicate that equilibrium fa values relative to the standard hydrogen electrode (SHE) are 40 mV more reducing (i.e., more negative) than those for Salado WIPP brine (-480 mV vs. -440 mV, respectively) because of the higher pH of the Castile brine (pH 9 .3 for Castile vs. pH 8.8 for Salado). The E h and pH data were corrected for the effects of high ionic strength. The experimental results for both brines are consistent with thermodynamic predictions using OLI Systems' Mixed Solvent Electrolyte chemical equilibrium model. The measured and corrected pH and E h data from this study are provided in Table ES-I and Table ES-2, respectively. The experimental study, with four test conditions in triplicate, was performed in a dual glovebox with a nominally 3 vol.% H 2 in argon atmosphere (target H 2 range: 3 ± I vol.%). Simulants containing MgO only ( experimental control) and MgO+Fe 0 (WIPP base case) were prepared for both the Salado and Castile brines. MgO was included in all simulants to account for the use of bulk magnesium oxide in the WIPP repository. Fe 0 was included in some simulants to incorporate the effects of the anoxic corrosion of iron and in-situ hydrogen generation in the study. The brine compositions were developed by Sandia National Laboratory (SNL; Xiong, 2008) and have been used in previous WIPP evaluations. The test method (agitation, etc.) is partially based on ASTM D3987-12. Twelve rounds of periodic measurements of pH and E h were performed over the course of the study. Chemical analysis results for liquids and solids (ICP-MS, ICP-ES, IC Anion, TIC, SEM-EDX) are consistent with the pH, E h , and thermodynamic modeling results. This study included the following conditions that deviate from anticipated post-closure conditions following brine intrusion, but were selected to facilitate bench-scale testing to validate modeling of pH and E h for the post-closure WIP P repository: an anoxic glove box atmosphere containing ≤ 4 vol. % H 2 vs. substantially higher H 2 gas concentrations assumed in the WIPP Performance Assessment (PA); a significantly higher liquid-to-solid test ratio compared to the much lower phase ratio anticipated in the WIP P repository; agitation of the simulant bottles to maximize mass transfer; and finally the use of Fe 0 reagents having a much greater surface area than expected in the WIP P repository. Non-representative conditions were chosen for various reasons such as: to provide bounding conservative results, to provide a margin of safety for testing, or to facilitate simulant sub-sampling and analysis. In a parallel effort, aqueous electrolyte thermodynamic models were developed for the synthetic Salado and Castile brines to inform the experimental design, facilitate laboratory data interpretation, and allow extension of evaluations beyond the parameters tested. Thermodynamic modeling simulations including the MgO and Fe 0 additives that are directly relevant to the experimental measurements (e.g., pH calibration curve, ORP corrections) are included in this report. The measured fa of the simulants was close to the OLI model predictions for both brines and was largely controlled by the background H 2 partial pressure in the vapor phase as well as H 2 generated in situ in the aqueous phase by the Fe 0 corrosion. The H 2 gas-phase concentration tested and thermodynamically evaluated was much lower than is assumed in the WIPP PA; however, H 2 (g) concentrations significantly below this level are still predicted to result in very reducing conditions. In conclusion: • The experimental results are consistent with thermodynamic model predictions for fa, pH, and the effects of high ionic strength. • Evidence to date suggests that the H2 concentration in the glovebox atmosphere ultimately determined the final E h values of the simulants and resulted in highly reducing conditions. As a result, little difference was observed between the control simulants containing only MgO and the WIPP base-case simulants that contained MgO and Fe 0 . • This test methodology is recommended for future studies evaluating WIPP repository conditions. The methodology includes: (1) background H 2 in argon with agitation ( or could alternatively include in-situ-generated H 2 in sealed bottles); (2) carefully measured and corrected ORP data ( with much effort focused on allowing the probes to fully stabilize); and (3) ionic-strength-corrected pH data. Other best practices, such as simulant sparging/handling, ORP probe replacement, etc., should also be considered. • The coupling of experimental studies and thermodynamic modeling is also highly recommended because these methods inform and direct one another leading to greater confidence in and understanding of the results.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2024

Interpreting Transformers for Jet Tagging

Machine learning (ML) algorithms, particularly attention-based transformer models, have become indispensable for analyzing the vast data generated by particle physics experiments like ATLAS and CMS at the CERN LHC. Particle Transformer (ParT), a state-of-the-art model, leverages particle-level attention to improve jet-tagging tasks, which are critical for identifying particles resulting from proton collisions. This study focuses on interpreting ParT by analyzing attention heat maps and particle-pair correlations on the $\eta$-$\phi$ plane, revealing a binary attention pattern where each particle attends to at most one other particle. At the same time, we observe that ParT shows varying focus on important particles and subjets depending on decay, indicating that the model learns traditional jet substructure observables. These insights enhance our understanding of the model's internal workings and learning process, offering potential avenues for improving the efficiency of transformer architectures in future high-energy physics applications.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

First search for an anomalous excess of charged-current $\nu_e$ interactions without visible pions using the full MicroBooNE dataset

This public note presents an investigation of low-energy electron-neutrino events in the Fermilab Booster Neutrino Beam by the MicroBooNE experiment. This search is motivated by the excess of low energy electromagnetic events observed by the MiniBooNE experiment and, more broadly, by the landscape of neutrino anomalies observed at short baselines. This is the first measurement to use all of the data collected by the MicroBooNE experiment, corresponding to $1.1\times 10^{21}$ protons on target. Two exclusive samples of electron neutrinos without visible pions are used, one with visible protons and one without any visible protons. MicroBooNE data is compared to two empirical models of the MiniBooNE low energy excess, one obtained by enhancing the electron-neutrino content as a function of the neutrino energy, and one representing the excess for the first time as a function of the kinematics of shower energy and angle. This measurement excludes an electron-like interpretation of the MiniBooNE excess based on these models at $\geq 99\%$ confidence level in all kinematic variables.

43 PARTICLE ACCELERATORS↗
DOE OSTI2021

Unified model of short- and long-term HIV viral rebound for clinical trial planning

Antiretroviral therapy (ART) effectively controls HIV infection, suppressing HIV viral loads. Typically suspension of therapy is rapidly followed by rebound of viral loads to high, pre-therapy levels. Indeed, a recent study showed that approximately 90% of treatment interruption study participants show viral rebound within at most a few months of therapy suspension, but the remaining 10%, showed viral rebound some months, or years, after ART suspension. Some may even never rebound. We investigate and compare branching process models aimed at gaining insight into these viral dynamics. Specifically, we provide a theory that explains both short- and long-term viral rebounds, and post-treatment control, via a multitype branching process with time-inhomogeneous rates, validated with data from Li et al. We discuss the associated biological interpretation and implications of our best-fit model. To test the effectiveness of an experimental intervention in delaying or preventing rebound, the standard practice is to suspend therapy and monitor the study participants for rebound. We close with a discussion of an important application of our modelling in the design of such clinical trials.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Machine learning enables interpretable discovery of innovative polymers for gas separation membranes

Polymer membranes perform innumerable separations with far-reaching environmental implications. Despite decades of research, design of new membrane materials remains a largely Edisonian process. To address this shortcoming, we demonstrate a generalizable, accurate machine learning (ML) implementation for the discovery of innovative polymers with ideal performance. Specifically, multitask ML models are trained on experimental data to link polymer chemistry to gas permeabilities of He, H 2 , O 2 , N 2 , CO 2 , and CH 4 . We interpret the ML models and extract valuable insights into the contributions of different chemical moieties to permeability and selectivity. We then screen over 9 million hypothetical polymers and identify thousands that lie well above current performance upper bounds, including hundreds of never-before-seen ultrapermeable polymer membranes with O 2 and CO 2 permeability greater than 10 4 and 10 5 Barrers, respectively. High-fidelity molecular dynamics simulations confirm the ML-predicted gas permeabilities of the promising candidates, which suggests that many can be translated to reality.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Comparison between Alcator C-Mod ICRF experiments and 3D full wave simulations

Reliable modeling of ion cyclotron range of frequencies (ICRF) antenna performance is essential for interpreting present experiments and guiding the design of future reactors. In this work, a 3D model of the Alcator C-Mod field-aligned antenna is implemented in the Petra-M finite-element framework [S. Shiraiwa et al 2023 Nucl. Fusion 63 026024] and benchmarked against experimental results. Four experimental cases are examined. First, the simulated rectified sheath potentials on the antenna limiters are compared with measurements from a power tapering experiment. Second, proof-of-principle far-field sheath simulations are performed. In scenarios with low single-pass absorption, simulations predict enhanced sheath potentials on a distant poloidal limiter in the far field of the antenna, consistent with experimental observations. Third, the simulated antenna loading during edge localized modes follows experimental trends and appears to be dominated by the density gradient at the pedestal. Finally, Petra-M predicts the unintended excitation of high-$k$ ∥ modes during monopole phasing operation, in agreement with experimental evidence of poor wave coupling and accessibility to the plasma core. Overall, reasonable agreement is found between the simulations and experiments. At the same time, areas of imperfect agreement are identified. These provide important guidance on the limits of the current state-of-the-art modeling, which should be kept in mind when using it as a predictive tool for future reactors.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2024

Explaining Neural Spike Activity for Simulated Bio-plausible Network through Deep Sequence Learning

With significant improvements in large-scale simulations of brain models, there is a growing need to develop tools for rapid analysis and interpreting the simulation results. In this work, we explore the potential of sequential deep learning models to understand and explain the network dynamics among the neurons extracted from a large-scale neural simulation in STACS (Simulation Tool for Asynchronous Cortical Stream). Our method employs a representative neuroscience model that abstracts the cortical dynamics with a reservoir of randomly connected spiking neurons with a low stable spike firing rate throughout the simulation duration. We subsequently analyze the spike dynamics of the simulated spiking neural network through an autoencoder model and an attention-based mechanism.

Kulkarni, Shruti↗
DOE OSTI2024

Diffusion model approach to simulating electron-proton scattering events

Generative artificial intelligence is a fast-growing area of research offering various avenues for exploration in high-energy nuclear physics. In this work, we explore the use of generative models for simulating electron-proton collisions relevant to experiments like the Continuous Electron Beam Accelerator Facility and the future Electron-Ion Collider (EIC). These experiments play a critical role in advancing our understanding of nucleons and nuclei in terms of quark and gluon degrees of freedom. The use of generative models for simulating collider events faces several challenges such as the sparsity of the data, the presence of global or eventwide constraints, and steeply falling particle distributions. In this work, we focus on the implementation of diffusion models for the simulation of electron-proton scattering events at EIC energies. Our results demonstrate that diffusion models can reproduce relevant observables such as momentum distributions and correlations of particles, momentum sum rules, and the leading electron kinematics, all of which are of particular interest in electron-proton collisions. Although the sampling process is relatively slow compared to other machine-learning architectures, we find diffusion models can generate high-quality samples. We foresee various applications of our work including inference for nuclear structure, interpretable generative machine learning, and searches of physics beyond the Standard Model. Published by the American Physical Society 2024

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Search for bosons of an extended Higgs sector in b quark final states in proton-proton collisions at $\sqrt{\textrm{s}}$ = 13 TeV

A search for beyond-the-standard-model neutral Higgs bosons decaying to a pair of bottom quarks, and produced in association with at least one additional bottom quark, is performed with the CMS detector. The data were recorded in proton-proton collisions at a centre-of-mass energy of 13 TeV at the CERN LHC and correspond to an integrated luminosity of 36.7–126.9 fb −1 , depending on the probed mass range. No signal above the standard model background expectation is observed. Upper limits on the production cross section times branching fraction are set for Higgs bosons in the mass range of 125–1800 GeV. The results are interpreted in benchmark scenarios of the minimal supersymmetric standard model, as well as suitable classes of two-Higgs-doublet models.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2022

A fourth-order phase-field fracture model: Formulation and numerical solution using a continuous/discontinuous Galerkin method

Modeling crack initiation and propagation in brittle materials is of great importance to be able to predict sudden loss of load-carrying capacity and prevent catastrophic failure under severe dynamic loading conditions. Second-order phase-field fracture models have gained wide adoption given their ability to capture the formation of complex fracture patterns, e.g. via crack merging and branching, and their suitability for implementation within the context of the conventional finite element method. Higher-order phase-field models have also been proposed to increase the regularity of the exact solution and thus increase the spatial convergence rate of its numerical approximation. However, they require special numerical techniques to enforce the necessary continuity of the phase field solution. In this paper, we derive a fourth-order phase-field model of fracture in two independent ways; namely, from Hamilton’s principle and from a higher-order micromechanics-based approach. The latter approach is novel, and provides a physical interpretation of the higher-order terms in the model. In addition, we propose a continuous/discontinuous Galerkin (C/DG) method for use in computing the approximate phase-field solution. This method employs Lagrange polynomial shape functions to guarantee -continuity of the solution at inter-element boundaries, and enforces the required regularity with the aid of additional variational and interior penalty terms in the weak form. Finally, the phase-field equation is coupled with the momentum balance equation to model dynamic fracture problems in hyper-elastic materials. Two benchmark problems are presented to compare the numerical behavior of the C/DG method with mixed finite element methods.

42 ENGINEERING↗
DOE OSTI2023

Modelling populations of kilonovae

Abstract The 2017 detection of a kilonova coincident with gravitational-wave emission has identified neutron star mergers as the major source of the heaviest elements and dramatically constrained alternative theories of gravity. Observing a population of such sources has the potential to transform cosmology, nuclear physics, and astrophysics. However, with only one confident multi-messenger detection currently available, modelling the diversity of signals expected from such a population requires improved theoretical understanding. In particular, models that are quick to evaluate and are calibrated with more detailed multi-physics simulations are needed to design observational strategies for kilonovae detection and to obtain rapid-response interpretations of new observations. We use grey-opacity models to construct populations of kilonovae, spanning ejecta parameters predicted by numerical simulations. Our modelling focuses on wavelengths relevant for upcoming optical surveys, such as the Rubin Observatory Legacy Survey of Space and Time (LSST). In these simulations, we implement heating rates that are based on nuclear reaction network calculations. We create a Gaussian-process emulator for kilonova grey opacities, calibrated with detailed radiative transfer simulations. Using recent fits to numerical relativity simulations, we predict how the ejecta parameters from binary neutron star (BNS) mergers shape the population of kilonovae, accounting for the viewing-angle dependence. Our simulated population of BNS mergers produce peak i-band absolute magnitudes of −20 ≤ Mi ≤ −11. A comparison with detailed radiative transfer calculations indicates that further improvements are needed to accurately reproduce spectral shapes over the full light curve evolution.

79 ASTRONOMY AND ASTROPHYSICS↗