Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “model inference”

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 613 records · Page 34
All sourcesNASA NTRSDOE OSTI
DOE OSTI2017

Impact of Trucking Network Flow on Preferred Biorefinery Locations in the Southern United States

The impact of the trucking transportation network flow was modeled for the southern United States. The study addresses a gap in existing research by applying a Bayesian logistic regression and Geographic Information System (GIS) geospatial analysis to predict biorefinery site locations. A one-way trucking cost assuming a 128.8 km (80-mile) haul distance was estimated by the Biomass Site Assessment model. The “median family income,” “timberland annual growth-to-removal ratio,” and “transportation delays” were significant in determining mill location. Transportation delays that directly impacted the costs of trucking are presented. Here, a logistic model with Bayesian inference was used to identify preferred site locations, and locations not preferential for a mill location. The model predicted that higher probability locations for smaller biomass mills (feedstock capacity, the size of sawmills) were in southern Alabama, southern Georgia, southeast Mississippi, southern Virginia, western Louisiana, western Arkansas, and eastern Texas. The higher probability locations for large capacity mills (feedstock capacity, the size for pulp and paper mills) were in southeastern Alabama, southern Georgia, central North Carolina, and the Mississippi Delta regions.

09 BIOMASS FUELS↗
NASA NTRS1988

A simulation engine: Combining an expert system with a simulation language

Expert systems were applied in a number of ways to the field of simulation. One of these is the application of an expert system to drive a simulation, by making run-time decisions which effect the simulation. This approach was successful for a number of specific simulation models. The simulation engine extends this capability by supporting this type of interaction in a general purpose setting. A general purpose simulation language is provided for building models, and an inference engine is provided for knowledge processing. This combination results in a mechanism which allows general purpose models to be simulated in concert with interactive knowledge bases.

Spiegel, James R.↗
DOE OSTI2025

Data-driven reduced-order models for port-Hamiltonian systems with operator inference

Hamiltonian operator inference has been developed in Sharma et al. (2022) to learn structure-preserving reduced-order models (ROMs) for Hamiltonian systems. The method constructs a low-dimensional model using only data and knowledge of the functional form of the Hamiltonian. The resulting ROMs preserve the intrinsic structure of the system, ensuring that the mechanical and physical properties of the system are maintained. In this work, we extend this approach to port-Hamiltonian systems, which generalize Hamiltonian systems by including energy dissipation, external input, and output. Based on snapshots of the system’s state and output, together with the information about the functional form of the Hamiltonian, reduced operators are inferred through optimization and are then used to construct data-driven ROMs. To further alleviate the complexity of evaluating nonlinear terms in the ROMs, a hyper-reduction method via discrete empirical interpolation is applied. Accordingly, we derive error estimates for the ROM approximations of the state and output. Lastly, we demonstrate the structure preservation, as well as the accuracy of the proposed port-Hamiltonian operator inference framework, through numerical experiments on a linear mass–spring-damper problem and a nonlinear Toda lattice problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Conin

SAND2025-07645O Conin is a Python library that supports constrained analysis of probabilistic graphical models (PGMs). It enables constrained inference and learning for hidden Markov models, Bayesian networks, dynamic Bayesian networks, and Markov networks. Conin interfaces with the pgmpy library to specify general probabilistic graphical models with a variety of optimization solvers to support learning and inference. 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.

Hart, William [Sandia National Lab. (SNL-CA), Live↗
DOE OSTI2012

CONSTRAINING SATELLITE GALAXY STELLAR MASS LOSS AND PREDICTING INTRAHALO LIGHT. I. FRAMEWORK AND RESULTS AT LOW REDSHIFT

We introduce a new technique that uses galaxy clustering to constrain how satellite galaxies lose stellar mass and contribute to the diffuse 'intrahalo light' (IHL). We implement two models that relate satellite galaxy stellar mass loss to the detailed knowledge of subhalo dark matter mass loss. Model 1 assumes that the fractional stellar mass loss of a galaxy, from the time of merging into a larger halo until the final redshift, is proportional to the fractional amount of dark matter mass loss of the subhalo it lives in. Model 2 accounts for a delay in the time that stellar mass is lost due to the fact that the galaxy resides deep in the potential well of the subhalo and the subhalo may experience dark matter mass loss for some time before the galaxy is affected. We use these models to predict the stellar masses of a population of galaxies and we use abundance matching to predict the clustering of several r-band luminosity threshold samples from the Sloan Digital Sky Survey. Abundance matching assuming no stellar mass loss (akin to abundance matching at the time of subhalo infall) overestimates the correlation function on small scales ({approx}< 1 Mpc), while allowing too much stellar mass loss leads to an underestimate of small-scale clustering. For each luminosity threshold sample, we are thus able to constrain the amount of stellar mass loss required to match the observed clustering. We find that satellite galaxy stellar mass loss is strongly luminosity dependent, with less luminous satellite galaxies experiencing substantially more efficient stellar mass loss than luminous satellites. With constrained stellar mass loss models, we can infer the amount of stellar mass that is deposited into the IHL. We find that both of our model predictions for the mean amount of IHL as a function of halo mass are consistent with current observational measurements. However, our two models predict a different amount of scatter in the IHL from halo to halo, with Model 2 being favored by observations. This demonstrates that a comparison to IHL measurements provides independent verification of our stellar mass loss models, as well as additional constraining power.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

OPER: Optimality-Guided Embedding Table Parallelization for Large-scale Recommendation Model

With the sharp increasing volume of user data, Deep Learning Recommendation Model (DLRM) becomes an indispensable infrastructure in large technology companies. However, large-scale DLRM on the multi-GPU platform is still inefficient due to unbalanced workload partitioning and intensive inter-GPU communication. To this end, we propose OPER, an OPtimality guided Embedding table placement for large-scale Recommendation model training and inference. OPER explores the potential of mitigating remote memory access latency in DLRM through fine-grained embedding table placement. Specifically, OPER proposes a theoretical modeling that builds up the relationship between EMT placement and the embedding communication latency in both training and inference. OPER proves the NP hardness of finding the optimal embedding table placement and proposes a heuristic algorithm that yields near optimal placement. OPER implements a SHMEM-based embedding table training system and a unified embedding index mapping to support fine-grained embedding table sharding and placement. Comprehensive experiments reveal that OPER achieves on average 3.4× and 5.1× speedup on training and inference respectively over state-of-the-art DLRM frameworks.

Wang, Zheng↗
DOE OSTI2024

A framework for strategic discovery of credible neural network surrogate models under uncertainty

The widespread integration of deep neural networks in developing data-driven surrogate models for high-fidelity simulations of complex physical systems highlights the critical necessity for robust uncertainty quantification techniques and credibility assessment methodologies, ensuring the reliable deployment of surrogate models in consequential decision-making. Here, this study presents the Occam Plausibility Algorithm for surrogate models (OPAL-surrogate), providing a systematic framework to uncover predictive neural network-based surrogate models within the large space of potential models, including various neural network classes and choices of architecture and hyperparameters. The framework is grounded in hierarchical Bayesian inferences and employs model validation tests to evaluate the credibility and prediction reliability of the surrogate models under uncertainty. Leveraging these principles, OPAL-surrogate introduces a systematic and efficient strategy for balancing the trade-off between model complexity, accuracy, and prediction uncertainty. The effectiveness of OPAL-surrogate is demonstrated through two modeling problems, including the deformation of porous materials for building insulation and turbulent combustion flow for ablation of solid fuels within hybrid rocket motors.

42 ENGINEERING↗
NASA NTRS1978

Nitric oxide formation in gas turbine engines: A theoretical and experimental study

A modified Zeldovich kinetic scheme was used to predict nitric oxide formation in the burned gases. Nonuniformities in fuel-air ratio in the primary zone were accounted for by a distribution of fuel-air ratios. This was followed by one or more dilution zones in which a Monte Carlo calculation was employed to follow the mixing and dilution processes. Predictions of NOX emissions were compared with various available experimental data, and satisfactory agreement was achieved. In particular, the model is applied to the NASA swirl-can modular combustor. The operating characteristics of this combustor which can be inferred from the modeling predictions are described. Parametric studies are presented which examine the influence of the modeling parameters on the NOX emission level. A series of flow visualization experiments demonstrates the fuel droplet breakup and turbulent recirculation processes. A tracer experiment quantitatively follows the jets from the swirler as they move downstream and entrain surrounding gases. Techniques were developed for calculating both fuel-air ratio and degree of nonuniformity from measurements of CO2, CO, O2, and hydrocarbons. A burning experiment made use of these techniques to map out the flow field in terms of local equivalence ratio and mixture nonuniformity.

Mikus, T.↗
NASA NTRS2015

Evolution of H2O, CO, and CO2 Production in Comet C/2009 P1 Garradd During the 2011-2012 Apparition

We present analysis of high spectral resolution NIR spectra of CO and H2O in Comet C/2009 P1 (Garradd) taken during its 2011-2012 apparition with the CSHELL instrument on NASA's Infrared Telescope Facility (IRTF). We also present analysis of observations of atomic oxygen in Comet Garradd obtained with the ARCES echelle spectrometer mounted on the ARC 3.5-m telescope at Apache Point Observatory and the Tull Coude spectrograph on the Harlan J. Smith 2.7-m telescope at McDonald Observatory. The observations of atomic oxygen serve as a proxy for H2O and CO2. We confirm the high CO abundance in Comet Garradd and the asymmetry in the CO/H2O ratio with respect to perihelion reported by previous studies. From the oxygen observations, we infer that the CO2/H2O ratio decreased as the comet moved towards the Sun, which is expected based on current sublimation models. We also infer that the CO2/H2O ratio was higher pre-perihelion than post-perihelion. We observe evidence for the icy grain source of H2O reported by several studies pre-perihelion, and argue that this source is significantly less abundant post-perihelion. Since H2O, CO2, and CO are the primary ices in comets, they drive the activity. We use our measurements of these important volatiles in an attempt to explain the evolution of Garradd's activity over the apparition.

Comets↗
DOE OSTI2022

Modeling of transient conduction in building envelope assemblies: A review

As buildings age, retrofits are becoming an increasingly important topic for the ever-growing and aging existing building stock. To compare designs or evaluate in-service building envelopes, thermal modeling is utilized to evaluate the thermal performance of envelope assemblies; however, it can be difficult to model the thermal performance of as-built assemblies due to degradation or missing documentation. To address this issue, inverse modeling can be applied to infer the properties of as-built envelope assemblies. This paper presents a review of the literature and published methods to model and infer the transient conductive performance of building envelopes. This review serves as a survey of existing transient conduction algorithms to evaluate performance, computational speed, and relevance for inverse modeling applications. In addition to the literature review, this work also evaluates the computational performance of the most prevalent transient conduction algorithms against the ASHRAE 1052RP toolkit to assess inverse modeling potential. This methodology serves as a foundation for future research to characterize the transient thermal performance of as-built building envelope assemblies.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2012

Inference of gene regulatory networks from genome-wide knockout fitness data

Motivation: Genome-wide fitness is an emerging type of highthroughput biological data generated for individual organisms by creating libraries of knockouts, subjecting them to broad ranges of environmental conditions, and measuring the resulting clone-specific fitnesses. Since fitness is an organism-scale measure of gene regulatory network behaviour, it may offer certain advantages when insights into such phenotypical and functional features are of primary interest over individual gene expression. Previous works have shown that genome-wide fitness data can be used to uncover novel gene regulatory interactions, when compared with results of more conventional gene expression analysis. Yet, to date, few algorithms have been proposed for systematically using genome-wide mutant fitness data for gene regulatory network inference. Results: In this article, we describe a model and propose an inference algorithm for using fitness data from knockout libraries to identify underlying gene regulatory networks. Unlike most prior methods, the presented approach captures not only structural, but also dynamical and non-linear nature of biomolecular systems involved. A state– space model with non-linear basis is used for dynamically describing gene regulatory networks. Network structure is then elucidated by estimating unknown model parameters. Unscented Kalman filter is used to cope with the non-linearities introduced in the model, which also enables the algorithm to run in on-line mode for practical use. Here, we demonstrate that the algorithm provides satisfying results for both synthetic data as well as empirical measurements of GAL network in yeast Saccharomyces cerevisiae and TyrR–LiuR network in bacteria Shewanella oneidensis.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2017

On the galaxy–halo connection in the EAGLE simulation

Empirical models of galaxy formation require assumptions about the correlations between galaxy and halo properties. These may be calibrated against observations or inferred from physical models such as hydrodynamical simulations. In this Letter, we use the EAGLE simulation to investigate the correlation of galaxy size with halo properties. We motivate this analysis by noting that the common assumption of angular momentum partition between baryons and dark matter in rotationally supported galaxies overpredicts both the spread in the stellar mass–size relation and the anticorrelation of size and velocity residuals, indicating a problem with the galaxy–halo connection it implies. We find the EAGLE galaxy population to perform significantly better on both statistics, and trace this success to the weakness of the correlations of galaxy size with halo mass, concentration and spin at fixed stellar mass. Here by, using these correlations in empirical models will enable fine-grained aspects of galaxy scalings to be matched.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2016

THE IMPACT OF SURFACE TEMPERATURE INHOMOGENEITIES ON QUIESCENT NEUTRON STAR RADIUS MEASUREMENTS

Fitting the thermal X-ray spectra of neutron stars (NSs) in quiescent X-ray binaries can constrain the masses and radii of NSs. The effect of undetected hot spots on the spectrum, and thus on the inferred NS mass and radius, has not yet been explored for appropriate atmospheres and spectra. A hot spot would harden the observed spectrum, so that spectral modeling tends to infer radii that are too small. However, a hot spot may also produce detectable pulsations. We simulated the effects of a hot spot on the pulsed fraction and spectrum of the quiescent NSs X5 and X7 in the globular cluster 47 Tucanae, using appropriate spectra and beaming for hydrogen atmosphere models, incorporating special and general relativistic effects, and sampling a range of system angles. We searched for pulsations in archival Chandra HRC-S observations of X5 and X7, placing 90% confidence upper limits on their pulsed fractions below 16%. We use these pulsation limits to constrain the temperature differential of any hot spots, and to then constrain the effects of possible hot spots on the X-ray spectrum and the inferred radius from spectral fitting. We find that hot spots below our pulsation limit could bias the spectroscopically inferred radius downward by up to 28%. For Cen X-4 (which has deeper published pulsation searches), an undetected hot spot could bias its inferred radius downward by up to 10%. Improving constraints on pulsations from quiescent LMXBs may be essential for progress in constraining their radii.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2009

Determination of Chemical Kinetic Rate Constants of a Model for Carbothermal Processing of Lunar Regolith Simulant Using Methane

We have previously developed a chemical conversion model of the carbothermal processing of lunar regolith using methane to predict the rate of production of carbon monoxide. In this carbothermal process, gaseous methane is pyrolyzed as it flows over the hot surface of a molten zone of lunar regolith and is converted to carbon and hydrogen. Hydrogen is carried away by the exiting stream of gases and carbon is deposited on the melt surface. The deposited carbon mixes with the melt and reacts with the metal oxides in it to produce carbon monoxide that bubbles out of the melt. In our model, we assume that the flux of carbon deposited is equal to the product of the surface reaction rate constant gamma and the concentration of methane adjacent to the melt surface. Similarly, the rate of consumption of carbon per unit volume in the melt is equal to the product of the melt reaction rate constant k and the concentrations of carbon and metal oxide in the melt. In this paper, we describe our effort to determine gamma and k by comparison of the predictions from our model with test data obtained by ORBITEC (Orbital Technologies Corporation). The concentration of methane adjacent to the melt surface is a necessary input to the model. It is inferred from the test data by a mass balance of methane, adopting the usual assumptions of the continuously-stirred-tank-reactor model, whereby the average concentration of a given gaseous species equals its exit concentration. The reaction rates gamma and k have been determined by a non-linear least-squares fit to the test data for the production of carbon monoxide and the fraction of the incoming methane that is converted. The comparison of test data with our model predictions using the determined chemical kinetic rate constants provides a consistent interpretation of the process over the full range of temperatures, pressures, and methane flow rates used in the tests, thereby increasing our confidence to use the model for scale-up purposes.

Balasubramaniam, R↗
DOE OSTI2011

ON THE RELIABILITY OF STELLAR AGES AND AGE SPREADS INFERRED FROM PRE-MAIN-SEQUENCE EVOLUTIONARY MODELS

We revisit the problem of low-mass pre-main-sequence stellar evolution and its observational consequences for where stars fall on the Hertzsprung-Russell diagram (HRD). In contrast to most previous work, our models follow stars as they grow from small masses via accretion, and we perform a systematic study of how the stars' HRD evolution is influenced by their initial radius, by the radiative properties of the accretion flow, and by the accretion history, using both simple idealized accretion histories and histories taken from numerical simulations of star cluster formation. We compare our numerical results to both non-accreting isochrones and to the positions of observed stars in the HRD, with a goal of determining whether both the absolute ages and the age dispersions inferred from non-accreting isochrones are reliable. We show that non-accreting isochrones can sometimes overestimate stellar ages for more massive stars (those with effective temperatures above {approx}3500 K), thereby explaining why non-accreting isochrones often suggest a systematic age difference between more and less massive stars in the same cluster. However, we also find the only way to produce a similar overestimate for the ages of cooler stars is if these stars grow from {approx}0.01 M{sub sun} seed protostars that are an order of magnitude smaller than predicted by current theoretical models, and if the size of the seed protostar correlates systematically with the final stellar mass at the end of accretion. We therefore conclude that, unless both of these conditions are met, inferred ages and age spreads for cool stars are reliable, at least to the extent that the observed bolometric luminosities and temperatures are accurate. Finally, we note that the time dependence of the mass accretion rate has remarkably little effect on low-mass stars' evolution on the HRD, and that such time dependence may be neglected for all stars except those with effective temperatures above {approx}4000 K.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2013

THE KINEMATICS AND IONIZATION OF NUCLEAR GAS CLOUDS IN CENTAURUS A

Neumayer et al. established the existence of a blueshifted cloud in the core of Centaurus A, within a few parsecs of the nucleus and close to the radio jet. We propose that the cloud has been impacted by the jet, and that it is in the foreground of the jet, accounting for its blueshifted emission on the southern side of the nucleus. We consider both shock excitation and photoionization models for the excitation of the cloud. Shock models do not account for the [Si VI] and [Ca VIII] emission line fluxes. However, X-ray observations indicate a source of ionizing photons in the core of Centaurus A; photoionization by the inferred flux incident on the cloud can account for the fluxes in these lines relative to Brackett-{gamma}. The power-law slope of the ionizing continuum matches that inferred from synchrotron models of the X-rays. The logarithm of the ionization parameter is -1.9, typical of that in Seyfert galaxies and consistent with the value proposed for dusty ionized plasmas. The model cloud density depends upon the Lorentz factor of the blazar and the inclination of our line of sight to the jet axis. For acute inclinations, the inferred density is consistent with expected cloud densities. However, for moderate inclinations of the jet to the line of sight, high Lorentz factors imply cloud densities in excess of 10{sup 5} cm{sup -3} and very low filling factors, suggesting that models of the gamma-ray emission should incorporate jet Lorentz factors {approx}< 5.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1997

An Assessment of the Regional Distribution of the Oxygen-Isotope Ratio in Northeastern Canada

A compilation of mean values of the oxygen-isotope ratio relative to standard mean ocean Water for 22 sites representative of conditions in north-eastern Canada is complemented with data on mean annual surface temperature, latitude, surface elevation, and mean annual shortest distance to open ocean denoted by the 10% sea-ice concentration boundary. Stepwise regression analysis is used to develop a multivariate model suitable to infer the distribution of 6 1"0 in an area of complex topography and possibly mixed source of advected water vapor. The best model is produced by a run in the backward mode at the 95% confidence level in which only temperature, latitude and distance to the open ocean remain in the model (the correlation coefficient is 0.915, the adjusted coefficient of determination is 0.809, the root mean square residual is 1.62). This model is similar to the best 6180 predictive model derived elsewhere for Greenland, suggesting a common principal source of advected moisture.

Giovinetto, Mario B.↗
DOE OSTI2025

An attention-based neural ordinary differential equation framework for modeling inelastic processes

To preserve strictly conservative behavior as well as model the variety of dissipative behavior displayed by solid materials, we propose a significant enhancement to the internal state variable-neural ordinary differential equation (ISV-NODE) framework. In this data-driven, physics-constrained modeling framework internal states are inferred rather than prescribed. The ISV-NODE consists of: (a) a stress model dependent on observable deformation and inferred internal state, and (b) a model of the evolution of the internal states. The enhancements to ISV-NODE proposed in this work are multifold: (a) a partially input convex neural network stress potential provides polyconvexity in terms of observed strain while leaving the inferred state unconstrained, and (b) an internal state flow model uses common latent features to inform novel attention-based gating and drives the flow of internal state only in dissipative regimes. We demonstrated that this architecture can accurately model dissipative and conservative behavior across an isotropic, isothermal elastic-viscoelastic-elastoplastic spectrum with three exemplars, while maintaining fundamental principles by design.

97 MATHEMATICS AND COMPUTING↗