Search NASA⌕ Search

SEARCH · Search NASA

Results for “Model discrepancy”

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 523 records · Page 29
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Bridging the Gap between Cosmological Simulations with Graph Neural Networks and Domain Adaptation

Deep learning models have been shown to outperform methods that rely on summary statistics, like the power spectrum, in extracting information from complex cosmological data sets. However, due to differences in the subgrid physics implementation and numerical approximations across different simulation suites, models trained on data from one cosmological simulation show a drop in performance when tested on another. Similarly, models trained on any of the simulations would also likely experience a drop in performance when applied to observational data. Training on data from two different suites of the CAMELS hydrodynamic cosmological simulations, we examine the generalization capabilities of Domain Adaptive Graph Neural Networks (DA-GNNs). By utilizing GNNs, we capitalize on their capacity to capture structured scale-free cosmological information from galaxy distributions. Moreover, by including unsupervised domain adaptation via Maximum Mean Discrepancy (MMD), we enable our models to extract domain-invariant features. We demonstrate that DA-GNN achieves higher accuracy and robustness on cross dataset tasks (up to 28% better relative error and up to almost an order of magnitude better χ 2 ). Using data visualizations, we show the effects of domain adaptation on proper latent space data alignment. This shows that DA-GNNs are a promising method for extracting domain-independent cosmological information, a vital step toward robust deep learning for real cosmic survey data.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Domain Adaptive Graph Neural Networks for Constraining Cosmological Parameters Across Multiple Data Sets

Deep learning models have been shown to outperform methods that rely on summary statistics, like the power spectrum, in extracting information from complex cosmological data sets. However, due to differences in the subgrid physics implementation and numerical approximations across different simulation suites, models trained on data from one cosmological simulation show a drop in performance when tested on another. Similarly, models trained on any of the simulations would also likely experience a drop in performance when applied to observational data. Training on data from two different suites of the CAMELS hydrodynamic cosmological simulations, we examine the generalization capabilities of Domain Adaptive Graph Neural Networks (DA-GNNs). By utilizing GNNs, we capitalize on their capacity to capture structured scale-free cosmological information from galaxy distributions. Moreover, by including unsupervised domain adaptation via Maximum Mean Discrepancy (MMD), we enable our models to extract domain-invariant features. We demonstrate that DA-GNN achieves higher accuracy and robustness on cross-dataset tasks (up to $28\%$ better relative error and up to almost an order of magnitude better $\chi^2$). Using data visualizations, we show the effects of domain adaptation on proper latent space data alignment. This shows that DA-GNNs are a promising method for extracting domain-independent cosmological information, a vital step toward robust deep learning for real cosmic survey data.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

VALIDATION, VERIFICATION, AND CALIBRATION THROUGH A CAUSAL LENS

This paper presents an alternative method based on causal inference to perform validation, verification, and calibration of simulation models. While classical validation and verification approaches focus on the identification of the associations between data elements using statistical and machine learning methods, the novel methods in this paper focus instead on the identification of causal relationships between data elements. Statistical and machine learning-based approaches are strictly data-driven, meaning that they provide quantitative comparison measures between datasets without explicitly considering the hypotheses behind them. This can lead to the erroneous conclusion that, if two data sets are close enough, then the models that generated them are similar. In addition, when experimental and simulated data differ to an extent that fails to meet the acceptance criteria, calibration techniques are used to tweak simulation model parameters to reduce the gap between the two types of data. This produces the false expectation that a simulation model will match reality. The methods presented in this paper move away from these strictly data-driven methods for validation and calibration toward more robust, model-driven methods based on causal inference. Causal inference aims to identify the possible mechanisms that might have generated data. Thus, this analysis targets the prediction of the effects when one (or more) of the identified mechanisms are altered. There are many approaches to identify, quantify and illustrate causal relationships. For the scope of this paper, directed graphs are employed as causal models. If the directed graph lacks cycles it is known as a directed acyclic graph (DAG). A node in such a graph represents an observed data element while a directed edge connecting two nodes represents a causal relationship between two variables. The developed causal methods are designed to extract causal models from simulation models and from experimental data. Causal models capture the causal relationships between data elements (e.g., simulated and experimental data). In this context, validation and verification are performed by comparing causal models. The proposed approach does not only inform system analysts on how a simulation model matches real-world data, but also identifies elements of the simulation model that should be revised when discrepancies between simulation and experimental data are observed. Through these causal methods, analysts have a means to identify the portion of the model equation(s) that are behind an edge connecting two variables. Hence, once the structural differences between causal models have been determined, model calibration can occur by changing only those model parameters that impact the identified causal relationships.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1980

Vibrational kinetics in CO electric discharge lasers - Modeling and experiments

A model of CO laser vibrational kinetics is developed, and predicted vibrational distributions are compared with measurements. The experimental distributions were obtained at various flow locations in a transverse CW discharge in supersonic (M = 3) flow. Good qualitative agreement is obtained in the comparisons, including the prediction of a total inversion at low discharge current densities. The major area of discrepancy is an observed loss in vibrational energy downstream of the discharge which is not predicted by the model. This discrepancy may be due to three-dimensional effects in the experiment which are not included in the model. Possible kinetic effects which may contribute to vibrational energy loss are also examined.

Stanton, A. C.↗
DOE OSTI2023

Secondary Pion Production using the LINAC Beam on Graphite and Tungsten Targets

The Fermilab 400-MeV Linac primary proton beam can produce precision, single-species secondary beams using a production target in the MeV Test Area experimental hall; specifically, pions from 4 up to 120 MeV kinetic energy. (These low energy pions rapidly decay into a muon beam in a secondary collection and transport beamline.) Although graphite is a standard production target material given its ultra-high temperature tolerance and low impact on the post-target primary beam, secondary pion/muon total production cross sections increase approximately as Z1/3 for positive muons and N2/3 for negative muons. Higher mass targets such as tungsten can potentially increase low-energy m+ and m- rates by factors of 3 and 8, respectively. Initial production studies have confirmed higher muon yield from heavy targets (tungsten vs carbon), but also, surprisingly, show significant differences in pion production between modern hadronic models (GENIEhad) and between GEANT and MCNP, state of the art production modeling codes. This project will further investigate the production discrepancies between these two target materials and between different physics models with high-statistics runs using the GEANT-based code, G4Beamline to better quantify the results and compare with production data which will be obtained using the MTA secondary beamline.

Ahmed, Shiza↗
NASA NTRS2003

Coulomb Stress Accumulation along the San Andreas Fault System

Stress accumulation rates along the primary segments of the San Andreas Fault system are computed using a three-dimensional (3-D) elastic half-space model with realistic fault geometry. The model is developed in the Fourier domain by solving for the response of an elastic half-space due to a point vector body force and analytically integrating the force from a locking depth to infinite depth. This approach is then applied to the San Andreas Fault system using published slip rates along 18 major fault strands of the fault zone. GPS-derived horizontal velocity measurements spanning the entire 1700 x 200 km region are then used to solve for apparent locking depth along each primary fault segment. This simple model fits remarkably well (2.43 mm/yr RMS misfit), although some discrepancies occur in the Eastern California Shear Zone. The model also predicts vertical uplift and subsidence rates that are in agreement with independent geologic and geodetic estimates. In addition, shear and normal stresses along the major fault strands are used to compute Coulomb stress accumulation rate. As a result, we find earthquake recurrence intervals along the San Andreas Fault system to be inversely proportional to Coulomb stress accumulation rate, in agreement with typical coseismic stress drops of 1 - 10 MPa. This 3-D deformation model can ultimately be extended to include both time-dependent forcing and viscoelastic response.

Smith, Bridget↗
NASA NTRS2020

Bennu Shape Model Validation Methods and Results

The OSIRIS-REx Independent Shape Modeling Team, staffed by the Jet Propulsion Laboratory (JPL), has used stereophotoclinometry (SPC) to produce a three-dimensional shape model of Bennu. The SPC process is informed by (but is separate from) orbit determination, so it is important to ensure that the resulting shape model is consistent with all available spacecraft tracking data. Specifically, checking the shape model’s consistency with LIDAR measurements can illuminate any discrepancies, because LIDAR measurements are highly correlated with the shape model. This study focuses on the JPL experience with these LIDAR measurements and the greater context of the shape model validation process.

Lauretta, Dante S.↗
NASA NTRS2003

Measurement of Low Amounts of Precipitable Water Vapor Using Ground-Based Millimeterwave Radiometry

Extremely dry conditions characterized by amounts of precipitable water vapor (PWV) as as 1-2 mm commonly occur in high-latitude regions during the winter months. While such atmospheres carry only a few percent of the latent heat energy compared to tropical atmospheres, the effects of low vapor amounts on the polar radiation budget - both directly through modulation of longwave radiation and indirectly through the formation of clouds - are considerable. Accurate measurements of precipitable water vapor (PWV) during such dry conditions are needed to improve polar radiation models for use in understanding and predicting change in the climatically sensitive polar regions. To this end, the strong water vapor absorption at 183.310 GHz provides a unique means of measuring low amounts of PWV. Weighting function analysis, forward model calculations based upon a 7-year radiosonde dataset, and retrieval simulations consistently predict that radiometric measurements made using several millimeter-wavelength (MMW) channels near the 183 GHz line, together with established microwave (MW) measurements at the 22.235 GHz water vapor line and -3 1 GHz atmospheric absorption window can be used to determine within 5% uncertainty the full range of PWV expected in the Arctic. This unique collective capability stands in spite of accuracy limitations stemming from uncertainties due to the sensitivity of the vertical distribution of temperature and water vapor at MMW channels. In this study the potential of MMW radiometry using the 183 GHz line for measuring low amounts of PWV is demonstrated both theoretically and experimentally. The study uses data obtained during March 1999 as part of an experiment conducted at the Department of Energy s Cloud and Radiation Testbed (CART) near Barrow, Alaska. Several radiometers from both NOAA and NASA were deployed during the experiment to provide the first combined MMW and MW ground-based data set during dry arctic conditions. Single-channel retrievals of PWV were performed using the MW and MMW data. Discrepancies in the retrieved values were found to be consistent with differences observed between measured brightness temperatures (TBs) and forward-modeled TBs based on concurrent radiosonde profiles. These discrepancies are greater than can be explained by measurement error alone and are attributed to absorption model uncertainty. We discuss here the measurements, retrieval technique, and line model discrepancies along with difficulties and potential of MMW/MW PWV measurement.

Racette, Paul E.↗
NASA NTRS1980

Estimation of the accuracy of dynamic flight-determined coefficients

This paper discusses means of assessing the accuracy of maximum likelihood parameter estimates obtained from dynamic flight data. The commonly used analytical predictors of accuracy are compared from both statistical and simplified geometric standpoints. Emphasizing practical considerations, such as modeling error, the accuracy predictions are evaluated with real and simulated data. Improved computations of the Cramer-Rao bound to correct large discrepancies caused by colored noise and modeling error are presented. This corrected Cramer-Rao bound is the best available analytical predictor of accuracy. Engineering judgement, aided by such analytical tools, is the final arbiter of accuracy estimation.

Maine, R. E.↗
DOE OSTI2025

Development of an ERT‐Based Framework for Bentonite Buffers Monitoring From Laboratory Tests: 2. Quantitative Moisture Dynamics Estimation Model

Abstract The long‐term containment of high‐level radioactive waste in geological disposal repositories relies on Engineered Barrier Systems (EBS), with bentonite clay emerging as a candidate material due to its unique properties. Understanding moisture dynamics within bentonite buffers is crucial for EBS performance, as it directly influences the material's swelling capacity, thermal and hydraulic conductivity, mechanical properties, and long‐term evolution under complex thermal‐hydrological‐mechanical (THM) processes. This study develops an advanced Electrical Resistivity Tomography (ERT)‐based framework to quantitatively monitor moisture dynamics under THM conditions. Our framework extends the Waxman‐Smits model to incorporate the coupled effects of temperature, water content, fluid chemistry, and mechanical changes on bentonite's electrical properties. Utilizing HotBENT‐Lab data from our companion paper, which includes electrical conductivity, CT density, and thermocouple measurements, this study offers a novel methodological framework bridging different scales of the model. Our results show that the extended model can estimate water content from ERT data, capturing spatial and temporal variations in moisture distribution within bentonite columns. However, the model tends to overestimate water content compared to CT density‐derived measurements. We address this discrepancy by incorporating a simplified swelling effect model, which improves agreement between ERT and CT density‐based water content estimates. We also discuss model limitations, including simplified treatment of swelling and micropore effects, and propose a conceptual framework for transitioning from laboratory to field applications, addressing challenges such as parameter scalability, field validation methods, and integration of diverse data sources. This ERT‐based framework can potentially advance real‐world moisture monitoring of bentonite‐based EBS in nuclear waste repositories. Plain Language Summary Safely containing high‐level radioactive waste depends on barriers made from materials like bentonite clay, which is effective because it swells and seals in the waste. To ensure these barriers work well over time, it's important to understand how moisture moves through the clay. Our study developed a new method using ERT to monitor moisture levels in bentonite under conditions that mimic those in actual storage sites, including changes in temperature, water content, and mechanical stress. This study improved an existing model to better account for how these factors affect the clay, allowing us to create more accurate moisture maps. Initially, the proposed model overestimated the amount of water in the clay, but its accuracy was improved by factoring in how the clay swells when wet. This study also identified some limitations of the model and suggested ways to adapt it for use in real‐world waste storage sites. This new approach could lead to better monitoring and safety checks for nuclear waste storage systems, helping to ensure long‐term containment. Key Points This work develops an ERT‐based framework extending the Waxman‐Smits model to monitor bentonite moisture dynamics during coupled THM processes The extended model accurately estimates water content from Electrical Resistivity Tomography data, incorporating swelling effects to improve precision This work proposes a conceptual framework for transitioning from laboratory to field applications, advancing EBS monitoring in nuclear waste repositories

Chen, Hang↗
NASA NTRS2020

Progressive Failure Analysis Correlation with Notched Composite Laminate Test Data

Testing of small and intermediate laminate panels with notches was performed and Progressive Failure Analyses (PFA) models were developed for aid in test planning and for correlation with the test data. Two progressive damage failure models were included in the study: the commercially available Abaqus built-in damage model and COmplete STress Reduction (COSTR) damage model developed at NASA Langley Research Center (LaRC). The finite elements models used for the analysis were developed using shell elements. The pre-test PFA results obtained from the two damage models were compared to test data. Then a post-test PFA with updated material properties based on additional available material property data was executed with the COSTR damage model and compared to the test data. The panels tested in compression all exhibited less scatter in the failure load with self-similar failure behavior. The panels tested in tension exhibited more scatter in the failure load with a failure mode that included delamination of plies. However, a self-similar crack path was also observed as the overall failure mode for all panels tested. When comparing test data to the pre-test analysis results, both damage models with shell element models were considered adequate for predicting the behavior and failure load of both the small and intermediate panels when tested in compression. However, for the tension load cases, larger differences were observed between the test and analysis results. When comparing test data to the post-test analysis performed using the COSTR damage model, better correlation was observed. However, large discrepancies were still observed for the tension load cases. Consequently, a higher fidelity finite element model including solid elements for sub-laminates and the modeling of cohesive layers between sub-laminate layers is recommended for tension loading of notched composite laminates.

Damage Model,Progressive Failure Analysis,Open Hol↗
DOE OSTI2024

Projecting Large Fires in the Western US With an Interpretable and Accurate Hybrid Machine Learning Method

More frequent and widespread large fires are occurring in the western United States (US), yet reliable methods for predicting these fires, particularly with extended lead times and a high spatial resolution, remain challenging. In this study, we proposed an interpretable and accurate hybrid machine learning (ML) model, that explicitly represented the controls of fuel flammability, fuel availability, and human suppression effects on fires. The model demonstrated notable accuracy with a F 1 -score of 0.846 ± 0.012, surpassing process-driven fire danger indices and four commonly used ML models by up to 40% and 9%, respectively. More importantly, the ML model showed remarkably higher interpretability relative to other ML models. Specifically, by demystifying the “black box” of each ML model using the explainable AI techniques, we identified substantial structural differences across ML fire models, even among those with similar accuracy. The relationships between fires and their drivers, identified by our model, were aligned closer with established fire physical principles. The ML structural discrepancy led to diverse fire predictions and our model predictions exhibited greater consistency with actual fire occurrence. With the highly interpretable and accurate model, we revealed the strong compound effects from multiple climate variables related to evaporative demand, energy release component, temperature, and wind speed, on the dynamics of large fires and megafires in the western US. Our findings highlight the importance of assessing the structural integrity of models in addition to their accuracy. They also underscore the critical need to address the rise in compound climate extremes linked to large wildfires.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Investigating laser ablated plume dynamics of carbon and aluminum targets

Recently acquired high-resolution images of nanosecond laser ablation plumes suggest a strong correlation between the internal plume structure and the type of material being ablated. However, the details of this relation are currently not well understood. In this work, we attempt to explore this correlation using a 2D radiation hydrodynamics model to study the dependence of internal plume structure formation on the ablation material. Here spatio-temporal emission maps and plume expansion velocities from experimental measurements are compared with the model predictions, including synthetic emission maps. The shape and expansion rate of an outer air plume region are found to be in good agreement for both carbon and aluminum, as are the inner material plume dynamics for carbon ablation. The largest disagreement is observed in the case of a polished aluminum target, where the chaotic inner plume features seen in the experimental images are not observed in the model. The possible physical mechanisms responsible for this discrepancy are discussed. This effort constitutes a continued development toward a predictive model of ablation plume dynamics and chemistry for various materials in extreme environments.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS1996

Hydroacoustic forcing function modeling using DNS database

A wall pressure frequency spectrum model (Blake 1971 ) has been evaluated using databases from Direct Numerical Simulations (DNS) of a turbulent boundary layer (Na & Moin 1996). Good agreement is found for moderate to strong adverse pressure gradient flows in the absence of separation. In the separated flow region, the model underpredicts the directly calculated spectra by an order of magnitude. The discrepancy is attributed to the violation of the model assumptions in that part of the flow domain. DNS computed coherence length scales and the normalized wall pressure cross-spectra are compared with experimental data. The DNS results are consistent with experimental observations.

Zawadzki, I.↗
NASA NTRS1996

Stratospheric Chlorine partitioning: Constraints from Shuttle-borne Measurements of [HCl], [ClNO3] and [ClO]

Measured stratospheric mixing ratios of HCl, ClNO3, and ClO from ATMOS and MAS are poorly reproduced by models using recommended kinetic parameters. This discrepancy is not resolved by new rates for the reactions Cl+CH4 and OH+HCl derived from weighted fits to laboratory measurements. A deficit in modeled [HCl] and corresponding overprediction of [ClNO3] and [ClO], which increases with altitude, suggests that production of HCl between 20 and 50 km is much faster than predicted from recommended rates.

Michelsen, H. A.↗
NASA NTRS1983

Path discrepancies between great circle and rhumb line

A mathematical model for a comparative analysis of great circle vs. rhumb line navigation in the continental United States has been developed at the Avionics Engineering Center, Ohio University. A FORTRAN simulation of the model has been implemented on the IBM 370 computer. The simulation predicts pertinent navigation information for the two flight paths. The basis for the project, which is a part of an M.S. thesis, is to provide a data base for computing discrepancies between the two flight paths. This document briefly describes the model and discusses the implications of the results obtained.

Kaul, R.↗
NASA NTRS1975

Effect of the Jovian oblateness on Pioneer 10/11 radio occultations

The oblateness of Jupiter, which was neglected in the standard inversion technique for radio occultation data, has a substantial effect on the refractivity profiles of the neutral Jovian atmosphere derived from the Pioneer 10/11 occultations. The geometry of the Pioneer 10 entry and exit occultations has the effect that oblateness should account for much of the discrepancy of the results with plausible atmospheric models. The effect of the oblateness on Pioneer 11 entry is smaller, where much of the discrepancy could be due to other effects, such as cycle-slip in the spacecraft transponder. Theory indicates that the Pioneer 11 exit occultation should be affected by the oblateness in the opposite sense from the first three occultations; and, indeed, the data which had not been invertible under the spherical assumption produced reasonable results when an oblateness correction was applied.

Hubbard, W. B.↗
DOE OSTI2021

Model correction and updating of a stochastic degradation model for failure prognostics of miter gates

Understanding the degradation of the quoin block is vital for failure prognostics in miter gates. Due to the complicated degradation mechanism, degradation models based on simplifications and assumptions cannot accurately describe the damage evolution. It is observed that small errors in a simplified degradation model can lead to a large discrepancy in the remaining useful life estimation attributed to error accumulation over time. Aiming to address this issue in failure prognostics, this paper presents a dynamic model correction framework for a simplified degradation model using strain measurements. In the proposed framework, a polynomial chaos expansion (PCE) model is employed to compensate the missing physics in a simplified stochastic degradation model. Here, a maximum likelihood estimation method is developed to estimate the uncertain parameters of the simplified physics-based degradation model along with the unknown PCE model parameters using strain measurements as the observables. The updated damage degradation model is then applied to failure prognostics of a miter gate. Results of a case study show that the proposed approach can effectively improve the accuracy of failure prognostics in miter gates.

42 ENGINEERING↗