SEARCH · Search NASA

Results for “Model Uncertainty”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 649 records · Page 36

All sources NASA NTRS DOE OSTI

NASA NTRS2021

In-Flight Turbulence Emulation for Fixed-Wing Aircraft using Equivalent Multisine Excitations

A procedure is discussed for designing control-equivalent turbulence input (CETI) excitations to emulate a fixed-wing aircraft response to turbulence while flying through calm air. These excitations are added to the actuator commands to produce the same motion of the aircraft that would have resulted from flying in actual turbulence. Although CETI models are usually shaping filters driven by white noise inputs, the equivalent control inputs in this report were constructed as multisine excitations, which enables mimicking the response to turbulence with arbitrary power spectra, within hardware limitations. Practical aspects of the approach are discussed, such as actuator dynamics, model uncertainty, actual in-flight turbulence, and others. The procedure is demonstrated using a flight dynamics simulation model of a subscale generic jet transport aircraft.

control-equivalent turbulence inputs (CETI)↗

DOE OSTI2024

Boundary-Aware Adversarial Learning Domain Adaption and Active Learning for Cross-Sensor Building Extraction

The use of convolutional neural networks (CNNs) for building extraction from remote sensing images has been widely studied and many public datasets have been made available for accelerating development of these CNN models. Yet adapting pretrained models at scale in real-world scenarios remains a challenging task. The main barrier is that certain new labels are still needed to compensate for domain shifting between the labeled data and new images that potentially cover new geographic locations or that are from a different sensor. In this article, we propose to add informatively labeled samples from a new image pool under the paradigm of active learning. To select the most useful samples based on model uncertainty, we first tackle the problem of uncalibrated uncertainty estimation due to distribution shifting by adapting feature extractors with boundary-based adversarial learning. Calibrated uncertainty is used as the query criterion in the active learning process, where the most uncertain samples are selected for annotation and included for model retraining. The proposed workflow was tested with three data pairs in which each workflow represents a scenario often encountered in real-world applications, including adapting pretrained models to new images collected with different sensors or to new geographic areas where appearances and types of buildings are very different. Compared to several baselines, including random sampling, temperature scaling (a well-known uncertainty calibration technique), different query strategies, and active domain adaptation methods, the proposed workflow shows that strategically querying a smaller set of samples for labeling achieves comparable or better building extraction performance. The proposed method reduces the number of labeled samples required to achieve sufficient model accuracy, thus significantly reducing hundreds of person-hours for labeled data creation. In addition, we include a few considerations when deploying this workflow in a GPU cluster that can be easily adapted to achieve operational building extraction model retraining.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Model averaging approaches to data subset selection

Model averaging is a useful and robust method for dealing with model uncertainty in statistical analysis. Often, it is useful to consider data subset selection at the same time, in which model selection criteria are used to compare models across different subsets of the data. Two different criteria have been proposed in the literature for how the data subsets should be weighted. We compare the two criteria closely in a unified treatment based on the Kullback-Leibler divergence and conclude that one of them is subtly flawed and will tend to yield larger uncertainties due to loss of information. Here, analytical and numerical examples are provided.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

NASA NTRS2018

Random Variables with Moment-Matching Staircase Density Functions

This paper proposes a family of random variables for uncertainty modeling. The variables of interest have a bounded support set, and prescribed values for the first four moments. We present the feasibility conditions for the existence of any of such variables, and propose a class of variables that conforms to such constraints. This class is called staircase because the density of its members is a piecewise constant function. Convex optimization is used to calculate their distributions according to several optimality criteria, including maximal entropy and maximal log-likelihood. The flexibility and efficiency of staircases enable modeling phenomena having a possibly skewed and/or multimodal response at a low computational cost. Furthermore, we provide a means to account for the uncertainty in the distribution caused by estimating staircases from data. These ideas are illustrated by generating empirical staircase predictor models. We consider the case in which the predictor matches the sample moments exactly (a setting applicable to large datasets), as well as the case in which the predictor accounts for the sampling error in such moments (a setting applicable to sparse datasets). A predictor model for the dynamics of an aeroelastic airfoil subject to flutter instability is used as an example. The resulting predictor not only describes the system's response accurately, but also enables carrying out a risk analysis for safe flight.

Luis G. Crespo↗

DOE OSTI2019

Comprehensive analysis of anomalous ANITA events disfavors a diffuse tau-neutrino flux origin

Recently, the ANITA collaboration reported on two upward-going extensive air shower events consistent with a primary particle that emerges from the surface of the Antarctic ice sheet. These events may be of ν τ origin, in which the neutrino interacts within the Earth to produce a τ lepton that emerges from the Earth, decays in the atmosphere, and initiates an extensive air shower. In this paper we estimate an upper bound on the ANITA acceptance to a diffuse ν τ flux detected via τ-lepton-induced air showers within the bounds of standard model uncertainties. Furthermore, by comparing this estimate with the acceptance of Pierre Auger Observatory and IceCube and assuming standard model interactions, we conclude that a ν τ origin of these events would imply a neutrino flux at least two orders of magnitude above current bounds.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2024

Projective Integral Updates for High-Dimensional Variational Inference

Variational inference is an approximation framework for Bayesian inference that seeks to improve quantified uncertainty in predictions by optimizing a simplified distribution over parameters to stand in for the full posterior. Capturing model variations that remain consistent with training data enables more robust predictions by reducing parameter sensitivity. This work introduces a fixed-point optimization for variational inference that is applicable when every feasible log density can be expressed as a linear combination of functions from a given basis. In such cases, the optimizer becomes a fixed-point of projective integral updates. When the basis spans univariate quadratics in each parameter, the feasible distributions are Gaussian mean-fields and the projective integral updates yield quasi-Newton variational Bayes (QNVB). Other bases and updates are also possible. Since these updates require high-dimensional integration, this work begins by proposing an efficient quasirandom sequence of quadratures for mean-field distributions. Each iterate of the sequence contains two evaluation points that combine to correctly integrate all univariate quadratic functions and, if the mean-field factors are symmetric, all univariate cubics. More importantly, averaging results over short subsequences achieves periodic exactness on a much larger space of multivariate polynomials of quadratic total degree. The corresponding variational updates require four loss evaluations with standard (not second-order) backpropagation to eliminate error terms from over half of all multivariate quadratic basis functions. Furthermore, this integration technique is motivated by first proposing stochastic blocked mean-field quadratures, which may be useful in other contexts. A PyTorch implementation of QNVB allows for better control over model uncertainty during training than competing methods. Experiments demonstrate superior generalizability for multiple learning problems and architectures.

Gaussian mean-field↗

DOE OSTI2023

Machine Learning for Daily Forecasts of Arctic Sea Ice Motion: An Attribution Assessment of Model Predictive Skill

Physics-based simulations of Arctic sea ice are highly complex, involving transport between different phases, length scales, and time scales. Resultantly, numerical simulations of sea ice dynamics have a high computational cost and model uncertainty. We employ data-driven machine learning (ML) to make predictions of sea ice motion. The ML models are built to predict present-day sea ice velocity given present-day wind velocity and previous-day sea ice concentration and velocity. Models are trained using reanalysis winds and satellite-derived sea ice properties. We compare the predictions of three different models: persistence (PS), linear regression (LR), and a convolutional neural network (CNN). We quantify the spatiotemporal variability of the correlation between observations and the statistical model predictions. Additionally, we analyze model performance in comparison to variability in properties related to ice motion (wind velocity, ice velocity, ice concentration, distance from coast, bathymetric depth) to understand the processes related to decreases in model performance. Results indicate that a CNN makes skillful predictions of daily sea ice velocity with a correlation up to 0.81 between predicted and observed sea ice velocity, while the LR and PS implementations exhibit correlations of 0.78 and 0.69, respectively. The correlation varies spatially and seasonally: lower values occur in shallow coastal regions and during times of minimum sea ice extent. LR parameter analysis indicates that wind velocity plays the largest role in predicting sea ice velocity on 1-day time scales, particularly in the central Arctic. Regions where wind velocity has the largest LR parameter are regions where the CNN has higher predictive skill than the LR.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2014

Variations in Modeled Dengue Transmission over Puerto Rico Using a Climate Driven Dynamic Model

Dengue fever is a mosquito‐borne viral disease reemerging throughout much of the tropical Americas. Dengue virus transmission is explicitly influenced by climate and the environment through its primary vector, Aedes aegypti. Temperature regulates Ae. aegypti development, survival, and replication rates as well as the incubation period of the virus within the mosquito. Precipitation provides water for many of the preferred breeding habitats of the mosquito, including buckets, old tires, and other places water can collect. Because of variations in topography, ocean influences and atmospheric processes, temperature and rainfall patterns vary across Puerto Rico and so do dengue virus transmission rates. Using NASA's TRMM (Tropical Rainfall Measuring Mission) satellite for precipitation input, ground‐based observations for temperature input, and laboratory confirmed dengue cases reported by the Centers for Disease Control and Prevention for parameter calibration, we modeled dengue transmission at the county level across Puerto Rico from 2010‐2013 using a dynamic dengue transmission model that includes interacting vector ecology and epidemiological components. Employing a Monte Carlo approach, we performed ensembles of several thousands of model simulations for each county in order to resolve the model uncertainty arising from using different combinations of parameter values that are not well known. The top 1% of model simulations that best reproduced the reported dengue case data were then analyzed to determine the most important parameters for dengue virus transmission in each county, as well as the relative influence of climate variability on transmission. These results can be used by public health workers to implement dengue control methods that are targeted for specific locations and climate conditions.

Morin, Cory↗

DOE OSTI2016

DESIGN CHARACTERISTICS OF THE IDAHO NATIONAL LABORATORY HIGH-TEMPERATURE GAS-COOLED TEST REACTOR

Uncertainty and sensitivity analysis is an indispensable element of any substantial attempt in reactor simulation validation. The quantification of uncertainties in nuclear engineering has grown more important and the IAEA Coordinated Research Program (CRP) on High-Temperature Gas Cooled Reactor (HTGR) initiated in 2012 aims to investigate the various uncertainty quantification methodologies for this type of reactors. The first phase of the CRP is dedicated to the estimation of cell and lattice model uncertainties due to the neutron cross sections co-variances. Phase II is oriented towards the investigation of propagated uncertainties from the lattice to the coupled neutronics/thermal hydraulics core calculations. Nominal results for the prismatic single block (Ex.I-2a) and super cell models (Ex.I-2c) have been obtained using the SCALE 6.1.3 two-dimensional lattice code NEWT coupled to the TRITON sequence for cross section generation. In this work, the TRITON/NEWT-flux-weighted cross sections obtained for Ex.I-2a and various models of Ex.I-2c is utilized to perform a sensitivity analysis of the MHTGR-350 core power densities and eigenvalues. The core solutions are obtained with the INL coupled code PHISICS/RELAP5-3D, utilizing a fixed-temperature feedback for Ex. II-1a.. It is observed that the core power density does not vary significantly in shape, but the magnitude of these variations increases as the moderator-to-fuel ratio increases in the super cell lattice models.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2020

Total Measurement Uncertainty in Neutron Coincidence Multiplicity Analysis

Neutron multiplicity counting is the most commonly used nondestructive assay technique for determining the plutonium mass within containers of scrap PuO 2 or mixed oxide (MOX). In multiplicity analysis, the 240 Pu eff mass, leakage multiplication, and alpha ratio (the ratio of [α, n]-to-spontaneous fission neutron production) are the three primary unknown sample properties. They must be determined simultaneously. To solve for these three unknowns in a multiplicity assay, three measured values are needed: the singles, doubles, and triples neutron count rates. While the analysis is limited to solving for three unknowns, there are many additional factors that impact the observed count rates and contribute to the measurement uncertainty. In this study we investigate the various uncertainty contributors for the multiplicity analysis through a combination of traditional uncertainty propagation techniques supplemented by Monte Carlo simulations to address the dependences not explicitly expressed by the point source model. Uncertainties arising from counting statistics, calibration parameters, calibration method, nuclear data, and various material characteristics (isotopic abundances, chemical form, density, and impurities) are considered. A Total Measurement Uncertainty (TMU) estimate is then developed from these uncertainty contributors. This study is confined to multiplicity analysis of items commonly encountered in international safeguards applications. That is, the study focused on Pu oxides and MOX materials for the masses ranging up to 4000 grams total Pu. Multiplicity measurements were simulated using MCNP V6 based on the Plutonium Scrap Multiplicity Counter (PSMC), Epithermal Multiplicity Counter (ENMC), Pyrochemical Multiplicity Counter, and Large Epithermal Multiplicity Counter (LEMC) for this study; however, this report focuses on the parameterization of the uncertainties for the PSMC. The performance differences between the PSMC and the other multiplicity counting systems are relatively small, primarily manifesting in the impact on measurement precision so that the evaluation developed for the PSMC can be applied to the other multiplicity counting systems. Finally an analysis tool, the Multiplicity TMU Estimator, was developed from this study to serve as an aid for evaluation of the total measurement uncertainty of multiplicity assay results obtained from the commonly used INCC acquisition and analysis software.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

NASA NTRS1984

Time-dependent model of the Martian atmosphere for use in orbit lifetime and sustenance studies

A time-dependent model of the Martian atmosphere suitable for calculation of long-term aerodynamic effects on low altitude satellites is presented. The atmospheric model is both position dependent, through latitude and longitude effects, and time dependent. The time dependency includes diurnal and seasonal effects, effects of annual motion, long and short term solar activity effects, and periodic dust storm effects. Nine constituent gases are included in the model. Uncertainties in exospheric temperature, turbidity, and turbopause altitude are used to produce bounds on the expected density. A computer model - a Fortran subroutine which, when given the Julian date, Cartesian position of the sun and the spacecraft in aerocentric coordinates, returns the local values of mass density, temperature, scale height, and upper and lower bounds on the mass density is presented.

Culp, R. D.↗

NASA NTRS2010

Loop Shaping Control Design for a Supersonic Propulsion System Model Using Quantitative Feedback Theory (QFT) Specifications and Bounds

This paper covers the propulsion system component modeling and controls development of an integrated mixed compression inlet and turbojet engine that will be used for an overall vehicle Aero-Propulso-Servo-Elastic (APSE) model. Using previously created nonlinear component-level propulsion system models, a linear integrated propulsion system model and loop shaping control design have been developed. The design includes both inlet normal shock position control and jet engine rotor speed control for a potential supersonic commercial transport. A preliminary investigation of the impacts of the aero-elastic effects on the incoming flow field to the propulsion system are discussed, however, the focus here is on developing a methodology for the propulsion controls design that prevents unstart in the inlet and minimizes the thrust oscillation experienced by the vehicle. Quantitative Feedback Theory (QFT) specifications and bounds, and aspects of classical loop shaping are used in the control design process. Model uncertainty is incorporated in the design to address possible error in the system identification mapping of the nonlinear component models into the integrated linear model.

Connolly, Joseph W.↗

DOE OSTI2016

Updates to the Spatial Analysis and Decision Assistance (SADA) Freeware Tool - 16583

SADA is a collaboration between the University of Tennessee and Oak Ridge National Laboratory to present a freeware desktop application integrating environmental risk analytics, spatial modeling, and decision sciences for use at EPA, NRC, DOE, and DoD contaminated sites. The software embeds risk assessment, uncertainty modeling, and downstream decision processes within a spatial context. Updates to Version 6 will provide will include a rich GIS environment and a new 3-D visualization model which permits immersive visualization of contamination, with topology, and the built environment. The risk models presented are developed in conjunction with Superfund to implement all scenarios in their Regional Screening Levels (RSLs) and radionuclide preliminary remediation goals (PRGs). In addition, SADA will be able to ingest high resolution LandScan population data providing insights into the proximity of population to contamination. A brief overview of the progress in provided along with anticipated next steps. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2022

Robust constrained tension control for high-precision roll-to-roll processes

Tension control is critical for maintaining good product quality in most roll-to-roll (R2R) production systems. Previous work has primarily focused on improving the disturbance rejection performance of tension controllers. Here, a robust linear parameter-varying model predictive control (LPV-MPC) scheme is designed to enhance the tension tracking performance of a pilot R2R system for deposition of materials used in flexible thin film applications. The performance of a tension controller may degrade due to disturbances associated with model uncertainties and the slowly-changing dynamics in R2R systems. We introduce a method that separately treats these two sources of disturbance. The controller utilizes an incremental model to eliminate the errors caused by the mismatch between the nominal model and the actual system. A tube-based MPC formulation combined with scheduled parameters adequately updates models and corrects for the time-varying dynamics. Constraints on the rated motor torque are incorporated in the MPC to maintain the controller reliability and avoid machine failures. We illustrate the operation of our control algorithm through simulation of an actual R2R system. The controller outperforms the benchmarks in terms of fast transient response and offset-free tension tracking. Furthermore, it also demonstrates immunity from variations due to parametric uncertainties.

42 ENGINEERING↗

NASA NTRS2017

Data Assimilation and Uncertainties in Early Solar Cycle Predictions

Stochastic nature of solar activity variations together with our limited knowledge of the dynamo mechanism and subsurface dynamics causes uncertainty in predictions of the solar cycle. For improving the physics-‐based predictions we can take advantage of the mathematical data assimilation approach that allows us to take into account both, observational errors and model uncertainties, and provide estimates of the next solar cycle along with prediction uncertainties. In this study we use the Parker's migratory dynamo model together with the equation of magnetic helicity balance, which reproduces main properties of the sunspot cycles and allow us to minimize discrepancies between the observed global activity variations and the model solution. The test simulation runs show that a reliable prediction can be obtained for two phases of preceding solar cycle: 1) if the polar field reversals shortly after the solar maxima (strong toroidal field and weak poloidal field), and 2) during the solar minima (strongest poloidal and weak toroidal fields). The early estimate of Cycle 25 obtained by this method shows that this cycle will start in 2019 – 2020, reach the maximum in 2023 – 2024, and that the mean sunspot number at the maximum will be about 90 (for the v2.0 sunspot number series).

Data Assimilation↗

DOE OSTI2014

Full waveform inversion of solar interior flows

The inference of flows of material in the interior of the Sun is a subject of major interest in helioseismology. Here, we apply techniques of full waveform inversion (FWI) to synthetic data to test flow inversions. In this idealized setup, we do not model seismic realization noise, training the focus entirely on the problem of whether a chosen supergranulation flow model can be seismically recovered. We define the misfit functional as a sum of L {sub 2} norm deviations in travel times between prediction and observation, as measured using short-distance filtered f and p {sub 1} and large-distance unfiltered p modes. FWI allows for the introduction of measurements of choice and iteratively improving the background model, while monitoring the evolution of the misfit in all desired categories. Although the misfit is seen to uniformly reduce in all categories, convergence to the true model is very slow, possibly because it is trapped in a local minimum. The primary source of error is inaccurate depth localization, which, due to density stratification, leads to wrong ratios of horizontal and vertical flow velocities ({sup c}ross talk{sup )}. In the present formulation, the lack of sufficient temporal frequency and spatial resolution makes it difficult to accurately localize flow profiles at depth. We therefore suggest that the most efficient way to discover the global minimum is to perform a probabilistic forward search, involving calculating the misfit associated with a broad range of models (generated, for instance, by a Monte Carlo algorithm) and locating the deepest minimum. Such techniques possess the added advantage of being able to quantify model uncertainty as well as realization noise (data uncertainty).

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2023

Knowledge-Informed Uncertainty-Aware Machine Learning for Time Series Forecasting of Dynamical Engineered Systems

The high complexity and multiscale nature of many engineered systems—such as those in nuclear power plants—make representing and forecasting their dynamic behavior challenging. Physics-based models can be overly complex and computationally intractable, whereas machine learning (ML) tools are often data-hungry and prone to unphysical solutions. This study proposes a knowledge-informed ML-aided hybrid residual modeling approach that offers accurate and efficient time series forecasting for the operation of dynamical engineered systems. Hybrid residual modeling entails a baseline solution from domain knowledge and known physics expressions about the system dynamics integrated with an ML model to capture undiscovered information from the mismatch (i.e., residuals) between true states from measurements and baseline-predicted outputs. This study further quantifies the ML model uncertainty to provide trustworthy solutions. Real-time operational data from thermal-hydraulic flow loops of the cryogenic moderator system in Oak Ridge National Laboratory’s Spallation Neutron Source facility were used to demonstrate the potential of knowledge-informed uncertainty-aware ML in real-world applications. The state variables of the cryogenic helium loop were modeled with (1) first principles–based system identification (sysID), (2) long short-term memory (LSTM) neural network, and (3) hybrid sysID (baseline) + LSTM (residual). The superior predictive capability of the sysID+LSTM model versus stand-alone sysID and LSTM is confirmed by average performance metrics and individual data points across different prediction horizons. By creating a robust representation of the underlying physical system, the widely applicable hybrid residual modeling approach will enable the future development of digital twins for performance prediction, prognostics, and operation control.

Zhao, Xingang↗

DOE OSTI2022

Diffuse Radiation Forcing Constraints on Gross Primary Productivity and Global Terrestrial Evapotranspiration

Abstract The diffuse radiation fertilization effect—the increase in plant productivity in the presence of higher diffuse radiation ( K ↓,d )—is an important yet understudied aspect of atmosphere‐biosphere interactions and can modify the terrestrial carbon, energy, and water budgets. The K ↓,d fertilization effect links the carbon cycle with clouds and aerosols, all of which are large sources of uncertainties for our current understanding of the Earth system and for future climate projections. Here we establish to what extent observational and modeling uncertainty in sunlight's diffuse fraction ( k d ) affects simulated gross primary productivity (GPP) and terrestrial evapotranspiration ( λE ). We find only 48 eddy covariance sites with simultaneous sufficient measurements of K ↓,d with none in the tropical climate zone, making it difficult to constrain this mechanism globally using observations. Using a land modeling framework based on the latest version of the Community Land Model, we find that global GPP ranges from 114 Pg C year −1 when using k d forcing from the Modern‐Era Retrospective analysis for Research and Applications, version 2 reanalysis to a ∼7% higher value of 122 Pg C year −1 when using the Clouds and the Earth's Radiant Energy System satellite product, with especially strong differences apparent over the tropical region (mean increase ∼9%). The differences in λE , although smaller (−0.4%) due to competing changes in shaded and sunlit leaf transpiration, can be greater than regional impacts of individual forcing agents like aerosols. Our results demonstrate the importance of comprehensively and systematically validating the simulated k d by atmosphere modules as well as the response to differences in k d within land modules across Earth System Models.

54 ENVIRONMENTAL SCIENCES↗