Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “error propagation”

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 361 records · Page 20
All sourcesNASA NTRSDOE OSTI
NASA NTRS2007

Space Shuttle Debris Impact Tool Assessment Using the Modern Design of Experiments

Complex computer codes are used to estimate thermal and structural reentry loads on the Shuttle Orbiter induced by ice and foam debris impact during ascent. Such debris can create cavities in the Shuttle Thermal Protection System. The sizes and shapes of these cavities are approximated to accommodate a code limitation that requires simple "shoebox" geometries to describe the cavities -- rectangular areas and planar walls that are at constant angles with respect to vertical. These approximations induce uncertainty in the code results. The Modern Design of Experiments (MDOE) has recently been applied to develop a series of resource-minimal computational experiments designed to generate low-order polynomial graduating functions to approximate the more complex underlying codes. These polynomial functions were then used to propagate cavity geometry errors to estimate the uncertainty they induce in the reentry load calculations performed by the underlying code. This paper describes a methodological study focused on evaluating the application of MDOE to future operational codes in a rapid and low-cost way to assess the effects of cavity geometry uncertainty.

DeLoach, Richard↗
NASA NTRS2007

Antarctic Circumpolar Current Transport Variability during 2003-05 from GRACE

Gravity Recovery and Climate Experiment (GRACE) gravity data spanning January 2003-November 2005 are used as proxies for ocean bottom pressure (BP) averaged over 1 month, spherical Gaussian caps 500 km in radius, and along paths bracketing the Antarctic Circumpolar Current's various fronts. The GRACE BP signals are compared with those derived from the Estimating the Circulation and Climate of the Ocean (ECCO) ocean modeling-assimilation system, and to a non-Boussinesq version of the Regional Ocean Model System (ROMS). The discrepancy found between GRACE and the models is 1.7 cm H2O (1 cm H2O approx. 1 hPa), slightly lower than the 1.9 cmH2O estimated by the authors independently from propagation of GRACE errors. The northern signals are weak and uncorrelated among basins. The southern signals are strong, with a common seasonality. The seasonal cycle GRACE data observed in the Pacific and Indian Ocean sectors of the ACC are consistent, with annual and semiannual amplitudes of 3.6 and 0.6 cmH2O (1.1 and 0.6 cm H2O with ECCO), the average over the full southern path peaks (stronger ACC) in the southern winter, on days of year 197 and 97 for the annual and semiannual components, respectively; the Atlantic Ocean annual peak is 20 days earlier. An approximate conversion factor of 3.1 Sv (Sv equiv 10(exp 6)cu m/s) of barotropic transport variability per cm H2O of BP change is estimated. Wind stress data time series from the Quick Scatterometer (QuikSCAT), averaged monthly, zonally, and over the latitude band 40(deg)- 65(deg)S, are also constructed and subsampled at the same months as with the GRACE data. The annual and semiannual harmonics of the wind stress peak on days 198 and 82, respectively. A decreasing trend over the 3 yr is observed in the three data types.

Gravity Recovery and Climate Experiment (GRACE)↗
NASA NTRS2006

Antarctic Circumpolar Current Transport Variability during 2003-05 from GRACE

Gravity Recovery and Climate Experiment (GRACE) gravity data spanning January 2003 - November 2005 are used as proxies for ocean bottom pressure (BP) averaged over 1 month, spherical Gaussian caps 500 km in radius, and along paths bracketing the Antarctic Circumpolar Current's various fronts. The GRACE BP signals are compared with those derived from the Estimating the Circulation and Climate of the Ocean (ECCO) ocean modeling-assimilation system, and to a non-Boussinesq version of the Regional Ocean Model System (ROMS). The discrepancy found between GRACE and the models is 1.7 cm(sub H2O) (1 cm(sub H2O) similar to 1 hPa), slightly lower than the 1.9 cm(sub H2O) estimated by the authors independently from propagation of GRACE errors. The northern signals are weak and uncorrelated among basins. The southern signals are strong, with a common seasonality. The seasonal cycle GRACE data observed in the Pacific and Indian Ocean sectors of the ACC are consistent, with annual and semiannual amplitudes of 3.6 and 0.6 cm(sub H2O) (1.1 and 0.6 cm(sub H2O) with ECCO), the average over the full southern path peaks (stronger ACC) in the southern winter, on days of year 197 and 97 for the annual and semiannual components, respectively; the Atlantic Ocean annual peak is 20 days earlier. An approximate conversion factor of 3.1 Sv ( Sv equivalent to 10(exp 6) m(exp 3) s(exp -1)) of barotropic transport variability per cm(sub H2O) of BP change is estimated. Wind stress data time series from the Quick Scatterometer (QuikSCAT), averaged monthly, zonally, and over the latitude band 40 de - 65 deg S, are also constructed and subsampled at the same months as with the GRACE data. The annual and semiannual harmonics of the wind stress peak on days 198 and 82, respectively. A decreasing trend over the 3 yr is observed in the three data types.

oceanographic regions↗
NASA NTRS2016

Laboratory Testing and Performance Verification of the CHARIS Integral Field Spectrograph

The Coronagraphic High Angular Resolution Imaging Spectrograph (CHARIS) is an integral field spectrograph (IFS) that has been built for the Subaru telescope. CHARIS has two imaging modes; the high-resolution mode is R82, R69, and R82 in J, H, and K bands respectively while the low-resolution discovery mode uses a second low-resolution prism with R19 spanning 1.15-2.37 microns (J+H+K bands). The discovery mode is meant to augment the low inner working angle of the Subaru Coronagraphic Extreme Adaptive Optics (SCExAO) adaptive optics system, which feeds CHARIS a coronagraphic image. The goal is to detect and characterize brown dwarfs and hot Jovian planets down to contrasts five orders of magnitude dimmer than their parent star at an inner working angle as low as 80 milliarcseconds. CHARIS constrains spectral crosstalk through several key aspects of the optical design. Additionally, the repeatability of alignment of certain optical components is critical to the calibrations required for the data pipeline. Specifically the relative alignment of the lens let array, prism, and detector must be highly stable and repeatable between imaging modes. We report on the measured repeatability and stability of these mechanisms, measurements of spectral crosstalk in the instrument, and the propagation of these errors through the data pipeline. Another key design feature of CHARIS is the prism, which pairs Barium Fluoride with Ohara L-BBH2 high index glass. The dispersion of the prism is significantly more uniform than other glass choices, and the CHARIS prisms represent the first NIR astronomical instrument that uses L-BBH2as the high index material. This material choice was key to the utility of the discovery mode, so significant efforts were put into cryogenic characterization of the material. The final performance of the prism assemblies in their operating environment is described in detail. The spectrograph is going through final alignment, cryogenic cycling, and is being delivered to the Subaru telescope in April 2016. This paper is a report on the laboratory performance of the spectrograph, and its current status in the commissioning process so that observers will better understand the instrument capabilities. We will also discuss the lessons learned during the testing process and their impact on future high-contrast imaging spectrographs for wavefront control.

Coronagraphic High Angular Resolution Imaging Spec↗
NASA NTRS2019

A Communication Channel Density Estimating Generative Adversarial Network

Autoencoder-based communication systems use neural network channel models to backwardly propagate message reconstruction error gradients across an approximation of the physical communication channel. In this work, we develop and test a new generative adversarial network (GAN) architecture for the purpose of training a stochastic channel approximating neural network. In previous research, investigators have focused on additive white Gaussian noise (AWGN) channels and/or simplified Rayleigh fading channels, both of which are linear and have well defined analytic solutions. Given that training a neural network is computationally expensive, channel approximation networks— and more generally the autoencoder systems—should be evaluated in communication environments that are traditionally difficult. To that end, our investigation focuses on channels that contain a combination of non-linear amplifier distortion, pulse shape filtering, intersymbol interference, frequency-dependent group delay, multipath, and non-Gaussian statistics. Each of our models are trained without any prior knowledge of the channel. We show that the trained models have learned to generalize over an arbitrary amplifier drive level and constellation alphabet. We demonstrate the versatility of our GAN architecture by comparing the marginal probability density function of several channel simulations with that of their corresponding neural network approximations

Smith, Aaron↗
NASA NTRS2019

Novel Classifier of Neural Impairment

The proposed research aligns with Ames Core Competency within the area of Intelligent/Adaptive Systems under Human Systems Integration. This will serve as a robust, efficient, and user-friendly human performance assessment tool that could provide immediate neural status of crewmembers in low-orbit and deep space missions. The human operator is a critical element of crew safety and should be evaluated consistently to mitigate potential off-nominal situations propagated from human error. Flagged deficits in oculomotor performance could act as an alert signal for crewmembers to opt out of cognitively demanding tasks or to prevent irreversible compromise in crewmember health.

impairment↗
NASA NTRS2018

Realistic Covariance Generation for the GPM Spacecraft

We present several different methods for generating realistic predictive covariance, including Monte-Carlo simulations and more direct linear methods which require the addition of process noise. The Monte-Carlo simulation starts with an epoch uncertainty sample basis and propagates each trial to a time of interest in the future. The variance-covariance of the state elements as well as other higher order sample statistics can be readily computed from the propagated sample. While this method preserves the nonlinear effects on the propagated uncertainty, it is computationally intensive as a statistically significant sample size must be considered in the propagation process. Moreover, if the epoch covariance is optimistic, this can result in an underestimation of the prediction error. Another method is to propagate a state sensitivity matrix simultaneously with the satellite state, which allows the epoch covariances to be propagated forward in a linear fashion. This method does not preserve the non-linearity of the satellite state uncertainty but is much less computationally intensive. The propagated state covariance is the scaled to represent the realistic level of GPM state uncertainty via a "e-tuning process." The tuning process generates an inflation factor based on the observed error statistics of the predictive satellite trajectories when compared to the definitive ones. Difference tuning strategies are considered and compared via Goodness-of-Fit method testing for the Gaussian properties of the scaled covariance.

prediction error↗
DOE OSTI2023

Deep Learning for Multigroup Cross-Section Representation in Two-Step Core Calculations

Here we investigate using deep learning, a type of machine-learning algorithm employing multiple layers of artificial neurons, for the mathematical representation of multigroup cross sections for use in the Griffin reactor multiphysics code for two-step deterministic neutronics calculations. A three-dimensional fuel element typical of a high-temperature gas reactor as well as a two-dimensional sodium-cooled fast reactor lattice are modeled using the Serpent Monte Carlo code, and multigroup macroscopic cross sections are generated for various state parameters to produce a training data set and a separate validation data set. A fully connected, feedforward neural network is trained using the open-source PyTorch machine-learning framework, and its accuracy is compared against the standard piecewise linear interpolation model. Additionally, we provide in this work a generic technique for propagating the cross-section model errors up to the k eff using sensitivity coefficients with the first-order uncertainty propagation rule. Quantifying the eigenvalue error due to the cross-section regression errors is especially practical for appropriately selecting the mathematical representation of the cross sections. We demonstrate that the artificial neural network model produces lower errors and therefore enables better accuracy relative to the piecewise linear model when the cross sections exhibit nonlinear dependencies; especially when a coarse grid is employed, where the errors can be halved by the artificial neural network. However, for linearly dependent multigroup cross sections as found for the sodium-cooled fast reactor case, a simpler linear regression outperforms deeper networks.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS1988

Optical propagation through a homogeneous turbulent shear flow

Effects of organized turbulent structures on the propagation of an optical beam in a homogeneous shear flow were studied. A passive-scalar field in a computed turbulent shear flow is used to represent index-of-refraction fluctuations, and phase errors induced in a coherent optical beam by turbulent fluctuations are computed. The organized vortical structures produce a scalar distribution with elongated regions of intense fluctuations which have an inclination with respect to the mean flow similar to that of the characteristic hairpin eddies. It is found that r.m.s. phase error is minimized by propagating approximately normal to the inclined vortical structures. Two-point correlations of vorticity and scalar fluctuation suggest that the regions of intense scalar fluctuation are produced primarily by the hairpin eddies.

Truman, C. Randall↗
NASA NTRS1989

Optical propagation through a homogeneous turbulent shear flow

Effects of organized turbulent structures on the propagation of an optical beam in a homogeneous shear flow have been studied. A passive-scalar field in a computed turbulent shear flow is used to represent index-of-refraction fluctuations, and phase errors induced in a coherent optical beam by turbulent fluctuations are computed. The organized vortical structures produce a scalar distribution with elongated regions of intense fluctuations which have an inclination with respect to the mean flow similar to that of the characteristic hairpin eddies. It is found that rms phase error is minimized by propagating approximately normal to the inclined vortical structures. Two-point correlations of vorticity and scalar fluctuation suggest that the regions of intense scalar fluctuation are produced primarily by the hairpin eddies.

Truman, C. Randall↗
NASA NTRS2024

Evaluation of 10‐m Wind Speed From ISD Meteorological Stations and the MERRA‐2 Reanalysis: Impacts on Dust Emission in the Arabian Peninsula

Mineral dust is one of the most important aerosols when studying the radiative balance and climate of the planet. There are different dust emission schemes utilized by the atmospheric modeling communities, many of which disagree on basic output quantities such as mass of dust emitted and distribution of mass among size bins. In this work, we examined mineral dust emission from a leading model scheme, the Goddard Chemistry Aerosol Radiation and Transport (GOCART), as utilized in the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) Reanalysis and compared it to dust emissions calculated using wind measurements from ground based weather stations located in the Arabian Peninsula that are included in the National Oceanic and Atmospheric Administration’s (NOAA) integrated surface database (ISD). An intercomparison of 10-m wind speed is shown for the Arabian Peninsula region, differences of the observed and modeled wind field are quantified, and impacts of differences on dust emissions are calculated. This analysis shows 10-m winds in the ISD were generally lower than MERRA-2 winds, which propagated to dust emissions errors. Our estimate of one of the most significant mass impacts in dust emission is 0.178 Tg/year/grid box with a percent change of over 200% to the recalculated dust emissions from MERRA-2. These differences in wind speed propagated to a difference in dust mass emitted by the use of a static source function which aids in scaling the mass emitted by the availability of dust in each grid. Additionally, the magnitude of these differences varies seasonally.

winds↗
DOE OSTI2025

ATTNChecker: Highly-Optimized Fault Tolerant Attention for Large Language Model Training

Large Language Models (LLMs) have demonstrated remarkable performance in various natural language processing tasks. However, the training of these models is computationally intensive and susceptible to faults, particularly in the attention mechanism, which is a critical component of transformer-based LLMs. In this paper, we investigate the impact of faults on LLM training, focusing on INF, NaN, and near-INF values in the computation results with systematic fault injection experiments. We observe the propagation patterns of these errors, which can trigger non-trainable states in the model and disrupt training, forcing the procedure to load from checkpoints. To mitigate the impact of these faults, we propose ATTNChecker, the first Algorithm-Based Fault Tolerance (ABFT) technique tailored for the attention mechanism in LLMs. ATTNChecker is designed based on fault propagation patterns of LLM and incorporates performance optimization to adapt to both system reliability and model vulnerability while providing lightweight protection for fast LLM training. Evaluations on four LLMs show that ATTNChecker on average incurs on average 7% overhead on training while detecting and correcting all extreme errors. Compared with the state-of-the-art checkpoint/restore approach, ATTNChecker reduces recovery overhead by up to 49×.

Liang, Yuhang [University of Alabama - Birmingham]↗
DOE OSTI2023

Count Every Trip: Finding the Uncertainty in Energy Estimates Made from Inferred Travel Modes

To properly inform transport policy and infrastructure changes, transportation related metrics need both measured values and uncertainties of those values. Travel monitoring smartphone apps can record people's travel behavior, but trip data quality is limited by sensor errors, user labeling rates and the accuracy of inference algorithms used for travel diary creation. We discuss the use of phone app recorded travel diary data to estimate energy consumption, and propose the use of propagation of variance to find error bars for such estimates. We define energy consumption for one trip as trip length times the energy intensity per distance unit of the travel mode used. We characterize trip length errors with relative error and inferred trip mode errors with confusion matrix columns. The resulting variances of each measurement are then propagated to the final calculated energy consumption. We tested our uncertainty methods on a dataset that used phone app data combined with prompted recall, consisting of 92,234 labeled trips for over 500,000 miles. Accounting for uncertainty using expected energy intensities and variance propagation gives a dataset-wide aggregate energy consumption percent error of about 9%, within one standard deviation from the truth. Future work could involve applying similar methods to other travel diary based metrics.

ADVANCED PROPULSION SYSTEMS,MATHEMATICS AND COMPUT↗
DOE OSTI2023

Count Every Trip: Finding the Uncertainty in Energy Estimates Made from Inferred Travel Modes

To properly inform transport policy and infrastructure changes, transportation related metrics need both measured values and uncertainties of those values. Travel monitoring smartphone apps can record people's travel behavior, but trip data quality is limited by sensor errors, user labeling rates and the accuracy of inference algorithms used for travel diary creation. We discuss the use of phone app recorded travel diary data to estimate energy consumption, and propose the use of propagation of variance to find error bars for such estimates. We define energy consumption for one trip as trip length times the energy intensity per distance unit of the travel mode used. We characterize trip length errors with relative error and inferred trip mode errors with confusion matrix columns. The resulting variances of each measurement are then propagated to the final calculated energy consumption. We tested our uncertainty methods on a dataset that used phone app data combined with prompted recall, consisting of 92,234 labeled trips for over 500,000 miles. Accounting for uncertainty using expected energy intensities and variance propagation gives a dataset-wide aggregate energy consumption percent error of about 9%, within one standard deviation from the truth. Future work could involve applying similar methods to other travel diary based metrics.

ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2023

Estimating Travel Energy Consumption Uncertainty Based on Inferred Travel Mode and Sensed Travel Length

To properly inform transport policy and infrastructure changes, transportation related metrics need both measured values and uncertainties of those values. Travel monitoring smartphone apps can record people's travel behavior, but trip data quality is limited by sensor errors, user labeling rates and the accuracy of inference algorithms used for travel diary creation. We discuss the use of phone app recorded travel diary data to estimate energy consumption, and propose the use of propagation of variance to find error bars for such estimates. We define energy consumption for one trip as trip length times the energy intensity per distance unit of the travel mode used. We characterize trip length errors with relative error and inferred trip mode errors with confusion matrix columns. The resulting variances of each measurement are then propagated to the final calculated energy consumption. We tested our uncertainty methods on a dataset that used phone app data combined with prompted recall, consisting of 92,234 labeled trips for over 500,000 miles. Accounting for uncertainty using expected energy intensities and variance propagation gives a dataset-wide aggregate energy consumption percent error of about 8%, within one standard deviation from the truth. Future work could involve applying similar methods to other travel diary based metrics.

ADVANCED PROPULSION SYSTEMS,ENERGY PLANNING, POLIC↗
NASA NTRS2010

Atmospheric Compensation for Uplink Arrays via Radiometry

Uplink arrays for communications applications are gaining increased visibility within the NASA and military community due to the enhanced flexibility and reliability they provide. When compared with the conventional large, single aperture antennas currently comprising the Deep Space Network (DSN), for example, smaller aperture antenna arrays have the benefits of providing fault tolerance (reduced single-point failure), reduced maintenance cost, and enhanced capabilities such as electronic beam-steering and multi-beam operation. However, signal combining of antenna array elements spaced many wavelengths apart becomes problematic due to the inherent instability of earth's turbulent atmosphere, particularly at the frequencies of interest to the DSN (i.e., Ka-band). Degradation in the power combining of the individual elements comprising the array arises due to uncorrelated phase errors introduced as the signals propagate through the troposphere. It is well known that the fundamental source of this error is due to the inhomogeneous distribution of water vapor in the atmosphere [1]. Several techniques have been proposed to circumvent this issue, including the use of phase calibration towers and a moon bounce to generate a feedback loop which would provide a means of intermittent calibration of the system phase errors (thermal drifts, atmosphere) [2,3]. However, these techniques require repositioning of the antenna elements to perform this operation which ultimately results in reduced system availability. And, though they are sufficient for compensating for slow varying phase drifts, they are insufficient to compensate for faster varying phase errors, such as those introduced by the atmosphere. In this paper, preliminary radiometry and interferometry measurements collected by the NASA Glenn Research Center are analyzed and indicate that the use of optimized water vapor radiometers as a feedback system in a communications platform could provide the necessary atmospheric compensation technique to enhance the beamforming of uplink arrays.

Nessel, James A.↗
NASA NTRS1986

Precision of vertical position estimates from very long baseline interferometry

It is found that the precision of VLBI measurements of vertical crustal motions is limited by errors in the modeling of the propagation delay through the earth's neutral atmosphere, and by errors in finding the orientation of a fixed crust coordinate system in the VLBI reference frame. The repeatability of baseline length measurements study indicates a vertical position precision of about 8 cm, averaged over 13 sites and 4.5 years of data, while the repeatability of vertical position estimates for a Richmond, FL site is found to yield a precision of about 7 cm for 42 observations made over an 11-month period. An overall precision of 8 cm for a 24-hour VLBI observing session is obtained, and the effects of earth orientation parameter errors is contingent on the distances between the VLBI sites.

Herring, T. A.↗
DOE OSTI2022

Nuclear data uncertainty propagation applied to the versatile test reactor conceptual design

We report the Versatile Test Reactor (VTR) currently under development is a 300 MWth sodium-cooled fast reactor (SFR) fueled with ternary metal alloy fuel, which aims to accelerate the testing of advanced nuclear fuels, materials, instrumentation, and sensors in high flux environments that are necessary to license the next generation of advanced reactor concepts. To support the VTR design process, uncertainties associated with the nuclear data has been propagated through the reactor core neutronics calculation to global parameters of interest, such as the core multiplication factor, kinetic parameters, and various reactivity feedback coefficients, following the sensitivity based uncertainty propagation approach. By folding the sensitivity coefficients, separately computed by the generalized perturbation theory code PERSENT and Monte Carlo code Serpent 2, with the variance-covariance matrices from COMMARA-2.0, we obtain the reaction-wise, isotope-wise, and overall uncertainties for each response of interest due to nuclear data uncertainty. With Serpent 2, the statistical error of the uncertainty is obtained by propagating the statistical error of the sensitivity coefficients through the same process using a newly developed uncertainty propagation method. From both codes, the overall top uncertainty contributors are found to be the cross section of Fe-56 elastic scattering, Na-23 elastic scattering, and U 238 inelastic scattering. The large contributions of the Fe-56 elastic scattering cross sections to global parameters are due to its relatively large relative uncertainty of 5–10% in nuclear data and the large volume of Fe-containing reflector assemblies in the fairly compact VTR core design. Both codes agreed well for the overall uncertainty estimates of all responses of interest, except the delayed neutron fraction, prompt neutron generation time, and the coolant density feedback coefficient, where Serpent 2 yielded a much larger value than PERSENT due to the large statistical error of sensitivity coefficients. The calculated uncertainties are also compared to those associated with other SFR cores. Another outcome of this study is a variance-covariance matrix of reactivity coefficients, which can be used in the subsequent uncertainty propagation to the system level to investigate the impact of identified uncertainties on system responses in the safety analysis.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗