Search NASA⌕ Search

SEARCH · Search NASA

Results for “prediction error”

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 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

The pp → W(→ lν) + γ process at next-to-next-to-leading order

We present details of the calculation of the pp → W(→ lν)γ process at next-to-next-to-leading order in QCD, calculated using the jettiness slicing method. The calculation is based entirely on analytic amplitudes. Because of the radiation zero, the NLO QCD contribution from the gq channel is as important as the contribution from the Born $ q\overline{q} $ process, disrupting the normal counting of leading and sub-leading contributions. We also assess the importance of electroweak (EW) corrections, including the EW corrections to both the six-parton channel 0 →$ \overline{u} d\nu {e}^{+}\gamma g $ and the five-parton channel 0 →$ \overline{u} d\nu {e}^{+}\gamma $. Previous experimental results have been shown to agree with theoretical predictions, taking into account the large experimental errors. With the advent of run II data from the LHC, the statistical errors on the data will decrease, and will be competitive with the error on theoretical predictions for the first time. We present numerical results for $ \sqrt{s} $ = 7 and 13 TeV. Analytic results for the one-loop six-parton QCD amplitude and the tree-level seven-parton QCD amplitude are presented in appendices.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2021

Modeling protein structures from predicted contacts with modern molecular dynamics potentials: accuracy, sensitivity, and refinement

Protein structure prediction has become increasingly popular and successful in recent years. An essential step for fragment-free, template-free methods is the generation of a final three-dimensional protein model from a set of predicted amino acid contacts that are often described by interresidue pairwise atomic distances. Here we explore the use of modern, open-source molecular dynamics (MD) engines, which have been continually developed over the last three decades with all-atom Hamiltonians to model biomolecular structure and dynamics, to generate accurate protein structures starting from a set of inferred pairwise distances. Additionally, the ability of MD empirical physical potentials to correct inaccuracies in the predicted geometries is tested. We rigorously characterize the effect of modeling parameters on results, the effect of different amounts of error in the predicted distances on the final structures, and test the ability of post-processing analysis to sort the best models out of a set of statistical replicas. We find that with exact distances and with noisy distances, the method can produce excellent structural models, and that the molecular dynamics force field seems to help correct errors in distance predictions, resisting the effects of applied noise.

Davidson, Russ↗
DOE OSTI2021

Uncertainty bounds for multivariate machine learning predictions on high-strain brittle fracture

Simulation of the crack network evolution on high strain rate impact experiments performed in brittle materials is very compute-intensive. The cost increases even more if multiple simulations are needed to account for the randomness in crack length, location, and orientation, which is inherently found in real-world materials. Constructing a machine learning emulator can make the process faster by orders of magnitude. There has been little work, however, on assessing the error associated with their predictions. Estimating these errors is imperative for meaningful overall uncertainty quantification. In this work, we extend the heteroscedastic uncertainty estimates to bound a multiple output machine learning emulator. Overall, we find that the response prediction is accurate within its predicted errors, but with a somewhat conservative estimate of uncertainty.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Effects of natural zeolites on field-scale geologic noble gas transport

Improving predictive models for noble gas transport through natural materials at the field-scale is an essential component of improving US nuclear monitoring capabilities. Several field-scale experiments with a gas transport component have been conducted at the Nevada National Security Site (Non-Proliferation Experiment, Underground Nuclear Explosion Signatures Experiment). However, the models associated with these experiments have not treated zeolite minerals as gas adsorbing phases. This is significant as zeolites are a common alteration mineral with a high abundance at these field sites and are shown here to significantly fractionate noble gases during field-scale transport. This fractionation and associated retardation can complicate gas transport predictions by reducing the signal-to-noise ratio to the detector (e.g. mass spectrometers or radiation detectors) enough to mask the signal or make the data difficult to interpret. Omitting adsorption-related retardation data of noble gases in predictive gas transport models therefore results in systematic errors in model predictions where zeolites are present.Herein is presented noble gas adsorption data collected on zeolitized and non-zeolitized tuff. Experimental results were obtained using a unique piezometric adsorption system designed and built for this study. Data collected were then related to pure-phase mineral analyses conducted on clinoptilolite, mordenite, and quartz. These results quantify the adsorption capacity of materials present in field-scale systems, enabling the modeling of low-permeability rocks as significant sorption reservoirs vital to bulk transport predictions.

58 GEOSCIENCES↗
DOE OSTI2023

Accuracy of hourly energy predictions for demand flexibility applications

Decarbonization goals in the United States electricity sector are increasing the levels of renewable energy generation in the electricity supply system, and are driving increased attention to building electrification, which will increase the magnitude and shift the timing of the electricity system peak. These changes are motivating new approaches to coordinate building electricity demand with low-carbon renewable generation, elevating the importance of demand flexibility (DF) in buildings and the need to quantify the temporal impacts of DF. In this paper, we first characterize the hourly predictive accuracy of six commonly used baseline models in an application context of quantifying building-level load shift. Our analysis revealed insights such as hours of the day (afternoons), periods of the week (weekends), and seasons (summer) that were predicted with more accuracy than other time periods. In addition, the analysis showed tendencies toward overprediction or underprediction of load. Secondly, we provide the first published investigation of baseline erosion from repeated dispatch of building load shifting. We observed that as the baseline period is pushed back further from the prediction day, the distribution of errors across baseline model predictions increases, with notable inflection points near the three-week erosion point for two of the three models.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2024

Comparison of MeshGraphNet Techniques for Subsurface Behavior Prediction during CO2 Sequestration

Carbon sequestration is a vital part of the effort to mitigate anthropogenic climate change. Previously, we have shown that Graph Neural Networks (GNNs) provide the ability to extract meaningful insights during prediction of subsurface behavior in carbon storage projects. However, these models have struggled with long-term prediction accuracy due to error accumulation caused by autoregressive prediction. This research leverages the Illinois Basin – Decatur Project (IBDP) dataset to examine strategies for minimizing loss over time in a MeshGraphNet GNN model to improve reliability of predictions while minimizing inferencing time.

Holcomb, Paul↗
DOE OSTI2022

Extension of the PINN diffusion model to k-eigenvalue problems

This paper extends our recent work on the Physics-Informed Neural Networks (PINN) approach for the fixed source diffusion models and applies it to the diffusion theory based k-eigenvalue problems. To make the PINN equitable for the eigenvalue problems, we introduce a novel integral regularization term to the loss function in the framework, and allow the direct inference of the principal eigenvalue and the associated eigenfunction. The regularization term enforces a pre-defined value on the integration of the model predictions, and this value can be directly related to a physical property of the system. We also introduce an additional learnable parameter to approximate the principal eigenvalue. As a proof of principle, we solve the one-group two-dimensional k-eigenvalue neutron diffusion equation in this work. We then provide two numerical examples to demonstrate the applicability of the PINN approach. In each example, we solve the k-eigenvalue diffusion equation in a multi-region configuration constrained with a set of Robin boundary conditions for generality. We use a FEM solution based on the power-iteration method to verify the results of the PINN solution. The results showed relative percentage error in the predicted eigenvalue of about 0.77% and about 1.2% for example 1 and example 2, respectively. The mean absolute error in the predicted flux for example 1 is ∼ 0.002 and for example 2 is ∼ 0.0024. These results indicate some preliminary successes of the PINN application to k-eigenvalue problems. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2020

Data-driven selection of stiff chemistry ODE solver in operator-splitting schemes

Most computational fluid dynamics simulations of practical combustion applications employ operator-splitting schemes, where chemistry and transport are separated and integrated with distinct numerical methods. The changes in composition due to chemistry are evaluated by solving ordinary differential equations (ODE) in each cell of the computational domain, which typically dominates the computational cost when detailed chemistry is considered. In this work, a data-driven approach for the selection of chemistry ODE solvers in operator-splitting schemes is presented. Neural networks are used to predict the ODE solvers CPU times and errors for a given thermochemical state. This allows the selection of an optimal ODE solver on a cell-by-cell, timestep-by-timestep basis. The models are trained using a wide set of thermochemical states generated through partially-stirred reactors and flames simulations. The methodology is validated by quantifying the prediction errors, the classification accuracy, and the computational speedup. The model predicts the optimal ODE solver for 70 to 95% of the validation cases and decreases the computional cost by a factor of 3 or more. The generalizability of the methodology to different chemical mechanisms and different fuels is assessed and it is shown that the model’s performance is only slightly degraded and its applicability is significantly enhanced if the inputs to the neural networks are restricted to a small set of thermochemical state variables present in most chemical mechanisms. In conclusion, the models are used in an homogeneous reactor case and a multi-dimensional CFD simulation of a diesel spray at high pressure where a speedup of more than 3 is achieved.

42 ENGINEERING↗
DOE OSTI2026

Data for: A hybrid biophysical-machine learning framework for diurnal surface energy flux estimation using proximal sensing

Thermal-based remote sensing of surface energy fluxes has traditionally relied on high spatial resolution satellite data with revisit frequencies on the order of weeks. In this study, we evaluate a biophysics-based analytical surface energy balance model for predicting latent energy (LE) and sensible heat (H) fluxes using proximal sensing observations. The Surface Temperature Initiated Closure (STIC1.2) model has been extensively validated across a wide range of spatial and temporal scales using various satellite-derived thermal datasets. Here we extend this validation by applying STIC at sub-hourly temporal resolution over multiple growing seasons for four distinct agricultural systems. We further develop and evaluate novel STIC variants that incorporate machine learning (ML) techniques to eliminate the need for specific surface energy balance observations, specifically net radiation and soil heat flux, thereby enhancing model applicability in data-sparse settings. The integration of an ML component to estimate surface available energy is shown to have strong predictive performance for both LE (R2 = 0.81-0.94) and H (R2 = 0.46-0.72) across all agricultural systems examined here, demonstrating the potential of hybrid biophysical – machine learning approaches for surface energy balance modeling with minimal data requirements. This study concludes with a novel application of explainable machine learning (exML) to diagnose sources of model error. This exML framework attributes residual prediction errors to both model input variables and environmental drivers not explicitly included in the simulation experiments. This approach provides a new pathway for improving model design and integrating previously overlooked yet influential variables into future model iterations.

Agricultural Sciences↗
DOE OSTI2026

A Hybrid Biophysical‐Machine Learning Framework for Diurnal Surface Energy Flux Estimation Using Proximal Sensing

Thermal infrared-based remote sensing of surface energy fluxes has traditionally relied on high spatial resolution satellite data with revisit frequencies on the order of weeks. In this study, we evaluate a biophysics-based analytical surface energy balance model for predicting latent energy ( LE ) and sensible heat ( H ) fluxes using proximal sensing observations. The Surface Temperature Initiated Closure (STIC1.2) model has been extensively validated across a wide range of spatial and temporal scales using various satellite-derived thermal infrared data sets. Here we extend this validation by applying STIC at sub-hourly temporal resolution over multiple growing seasons for four distinct agricultural systems. We further develop and evaluate novel STIC variants that incorporate machine learning (ML) techniques to eliminate the need for surface energy balance observations, specifically net radiation and soil heat flux, thereby enhancing model applicability in data-sparse settings. The integration of a ML component to estimate surface available energy is shown to have strong predictive performance for both LE (R 2 = 0.81–0.94) and H (R 2 = 0.46–0.72) across all agricultural systems examined here, demonstrating the potential of hybrid biophysical-machine learning approaches for surface energy balance modeling with minimal data requirements. This study concludes with a novel application of explainable machine learning (exML) to diagnose sources of model error. This exML framework attributes residual prediction errors to both model input variables and environmental drivers not explicitly included in the simulation experiments. This approach provides a new pathway for improving model design and integrating previously overlooked yet influential variables into future model iterations.

evapotranspiration↗
DOE OSTI2021

Cartesian message passing neural networks for directional properties: Fast and transferable atomic multipoles

The message passing neural network (MPNN) framework is a promising tool for modeling atomic properties but is, until recently, incompatible with directional properties, such as Cartesian tensors. We propose a modified Cartesian MPNN (CMPNN) suitable for predicting atom-centered multipoles, an essential component of ab initio force fields. The efficacy of this model is demonstrated on a newly developed dataset consisting of 46 623 chemical structures and corresponding high-quality atomic multipoles, which was deposited into the publicly available Molecular Sciences Software Institute QCArchive server. We show that the CMPNN accurately predicts atom-centered charges, dipoles, and quadrupoles and that errors in the predicted atomic multipoles have a negligible effect on multipole–multipole electrostatic energies. The CMPNN is accurate enough to model conformational dependencies of a molecule’s electronic structure. This opens up the possibility of recomputing atomic multipoles on the fly throughout a simulation in which they might exhibit strong conformational dependence.

Glick, Zachary L. (ORCID:0000000309002849)↗
DOE OSTI2025

Learning quantum computers' errors using interpretable neural networks

Learning and reducing the errors and noise in quantum computing systems is necessary for achieving quantum computation’s promise. However, rapid advances in experimental quantum computing are making this task increasingly difficult, because state-of-the-art systems now contain hundreds of qubits and many characterization techniques are hard to apply at this scale. Furthermore, complex kinds of errors in these systems, such as crosstalk and non-Markovian effects, must be understood and decreased, but these errors are challenging to study with most existing methods. In this project, we explored using neural networks for scalable characterization of complex errors in quantum computers. We proposed and demonstrated characterizing a quantum computer’s errors with neural networks that have interpretable parameters corresponding to the rates of different kinds of errors, within a sparse Lindbladian parameterization for errors. To enable scaling to many qubit systems, these networks then predict how these errors combine within quantum circuits and impact their outcomes using an efficient approximations. We demonstrated these networks ability to learn coherent crosstalk errors and context-dependent errors in a simulated 4-qubit system.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

EFCOG Human Performance Improvement Task Group: Task 20-2, Asking Better Error Precursor Questions for Effective Job Planning, Pre-Job Briefs, and Event Investigations

The second principle of human performance states: “error-likely situations are predictable, manageable, and preventable.” An error-likely situation comes into play when task-related factors exceed the capabilities of the individual, creating a mismatch at the point when the individual is “touching” either the physical or the paper plant. The simple presence of adverse conditions cannot be error-likely unless a specific action is to occur within that set of adverse conditions.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2023

Battery Charge Curve Prediction via Feature Extraction and Supervised Machine Learning

Real-time onboard state monitoring and estimation of a battery over its lifetime is indispensable for the safe and durable operation of battery-powered devices. In this study, a methodology to predict the entire constant-current cycling curve with limited input information that can be collected in a short period of time is developed. A total of 10 066 charge curves of LiNiO 2 -based batteries at a constant C-rate are collected. With the combination of a feature extraction step and a multiple linear regression step, the method can accurately predict an entire battery charge curve with an error of < 2% using only 10% of the charge curve as the input information. The method is further validated across other battery chemistries (LiCoO 2 -based) using open-access datasets. The prediction error of the charge curves for the LiCoO 2 -based battery is around 2% with only 5% of the charge curve as the input information, indicating the generalization of the developed methodology for predicting battery cycling curves. The developed method paves the way for fast onboard health status monitoring and estimation for batteries during practical applications.

25 ENERGY STORAGE↗
DOE OSTI2021

Development of a U-19Pu-10Zr fuel performance benchmark case based on the IFR-1 experiment

Metallic nuclear fuels are subject to research and development for use in advanced reactors. Robust, accurate metallic fuel performance models are important for the design, safety analysis, and licensing of these reactors. However, metallic fuel performance models require additional development; they are not as mature as UO 2 fuel performance models. A benchmark case based on the IFR-1 experiment was developed to better gauge the accuracy of existing models, identify models for high-priority development, and potentially quantify any future improvements made by further model development.This work collected publicly available information on the IFR-1 experiment and used it to develop the benchmark case. Fuel behavior during the IFR-1 irradiation was simulated by using the fuel performance code BISON, and the predicted results were compared with postirradiation examination data from the IFR-1 experiment. Furthermore, a sensitivity study and tuning studies were performed as a preliminary investigation into the causes of inaccurate temperature and dimensional change predictions.The benchmark predicted reasonably accurate values for the burnup and fission gas release. There was error in the predicted temperatures, which could be explained by uncertainty in the input parameters and legacy temperatures. BISON over-predicted dimensional changes in the fuel and cladding. The sensitivity study showed that the dimensional changes were most sensitive to the fuel swelling anisotropy and the cladding void swelling model. Future benchmark and model development should focus on cladding swelling behaviors to improve dimensional change predictions.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2020

A Study on Modeled Wind Speed Errors Using the U.S. Department of Energy Buoys

Pacific Northwest National Laboratory (PNNL) operates two AXYS WindSentinel lidar buoys for the U.S. Department of Energy’s Wind Energy Technologies Office. The purpose of these buoys is to collect hub-height winds and supporting meteorological and oceanographic information to facilitate the development of wind energy in the U.S. waters. The first deployment for one buoy was off the coast of Virginia from December 2014 to May 2016, and the first deployment for the other buoy was off the coast of New Jersey from November 2015 until February 2017. This report describes recent analysis of data collected during these first two deployments. Specifically, we compare hub-height wind speed estimates using Monin-Obukhov Similarity Theory (MOST) to the lidar measurements, and examine how those errors are affected by wind direction, atmospheric stability, wind-wave direction differences, and various measures of the wave-state. The comparisons are done using standard similarity functions based on MOST; including the Businger - Dyer, the Beljaars & Holtslag and the Vickers & Mahrt similarity functions. All models produce large errors over the range of atmospheric stabilities that were observed, with the largest errors occurring for stable flows. The Vickers & Mahrt function resulted in the largest overall bias and standard deviation, while Beljaars & Holtslag function gave the smallest bias and standard deviation due to its better performance under stable conditions. The models perform best under unstable conditions, but even in this regime there is a consistent overestimation of the wind speed of between roughly 0 to 1 ms -1 compared to the lidar measurements. We identify specific metocean conditions (i.e. stability and wind and wave directions) at each of the deployment locations that lead to large errors in MOST predictions. Finally, a coupled ocean-atmosphere model framework was investigated to simulate large errors in weather research forecasting (WRF).

17 WIND ENERGY↗
DOE OSTI2024

A deep learning-based workflow for fast prediction of 3D state variables in geological carbon storage: A dimension reduction approach

Deep learning (DL) models are extensively used as surrogate models for high-fidelity simulations of multiphase fluid flow in porous media at large scales, enabling fast forecasts of the spatial–temporal evolution of three-dimensional (3D) state variables in geological carbon storage (GCS). However, training these models in high-dimensional space remains computationally demanding and prone to overfitting because of limited training data. This paper presents a novel workflow to address these challenges by integrating dimension reduction (DR) methods. Here, the proposed workflow employed pre-trained DR models to extract the latent variables of geological models and state variables and utilized the multi-layer perceptron (MLP) for constructing mapping functions between the input and output variables in latent spaces. Subsequently, the pre-trained reconstruction models converted the MLP-predicted latent state variables to their original high-dimensional form. Furthermore, we proposed a novel strategy for the DR and reconstruction of 3D saturation fields to account for the unique data characteristics of sparsity, nonuniformity, and discontinuity. The proposed strategy applied PCA and inverse PCA for 2D average saturation fields and developed a DL-based 3D reconstruction model, leveraging three 2D average saturation fields as input to produce a 3D saturation field as output. The pre-training of DR and reconstruction models and training of MLP models were conducted on 84 Gulf of Mexico (GoM) simulations and evaluated on 12 testing simulations. Each simulation contained 720 monthly time steps, with the first 360 months as the injection period and the rest as the post-injection period. The proposed workflow, incorporating DR and DL models, accurately predicts the normalized 3D pressure fields, achieving mean square error (MSE) of 2.92 × 10 -7 compared to the ground truth obtained from a full-physics simulator. Furthermore, the proposed strategy outperformed PCA and convolutional autoencoder (CAE) models on 3D saturation fields, resulting in minor workflow prediction errors with an MSE of 2.93 × 10 -5 . The results suggest the proposed workflow provides sufficient predictive fidelity across temporal and spatial scales, and enables a speedup of 160 times compared to the full-physics simulator, facilitating improved decision-making and risk assessment for large-scale GCS management in real-time scenarios.

3D reconstruction model↗
DOE OSTI2022

Synthetic aromatic kerosene property prediction improvements with isomer specific characterization via GCxGC and vacuum ultraviolet spectroscopy

This research explores an advanced method of fuel composition determination and builds upon typical hydrocarbon group type analyses performed with two-dimensional gas chromatography (GCxGC). In this study, structural information of individual species within Virent’s Synthetic Aromatic Kerosene (SAK) is identified by vacuum ultraviolet (VUV) spectroscopy. By mass, 71.3% of the components elute within six peaks of the chromatogram, from which 12 unique species are identified through a novel deconvolution method. Overall, the identification of 93.6%m across 26 structural isomers is made by the methods described in this work. With 93.6%m ascribed to specific isomers, the precision of fuel property predictions improves dramatically. For example, the absolute error of the viscosity prediction is reduced by 90% because of this advancement in diagnostic capability, and its 95-percentile confidence interval (precision only) is reduced by 93%. Additionally, the properties of SAK, blended with hydro processed esters fatty acids (HEFA), are demonstrated to have blended properties consistent with conventional jet fuel.

10 SYNTHETIC FUELS↗