Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “variational inference”

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

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

At least 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Hydrodynamic and Radiographic Toolbox (HART)

With a multi-lab and university team, we propose to develop new methods for a Hydrodynamic and Radiographic Toolbox (HART) that will enable a fuller and more extensive use of experimental radiographic data towards better characterizing and reducing uncertainties in predictive modeling of weapons performance. We will achieve this by leveraging recent developments in areas of computational imaging, statistical and machine learning, and reduced order modeling of hydrodynamics. In terms of software practices, by partnering with XCP (RISTRA Project) we will conform to recent XCP standards that are inline with modern software practices and standards and ensure compatibility and ease of inter-operability with existing codes. Three new activities under this proposal include (a) the development and use of deep learning-based surrogates to accelerate reconstruction and variational inference of density fields from radiographs of hydrotests, (b) a model-data fusion strategy that couples deep learning-based density reconstructions with fast hydrodynamics simulators to better constrain the reconstruction, and (c) a method for treating asymmetries using techniques adopted from limited view tomography. All three new activities will be based on improved treatment of scatter, noise, beam spot movement, detector blur, and flat fielding in the forward model, and a use of sophisticated priors to aid in the re construction. The improvements to the forward model and improved algorithmic design of the reconstruction when complete will be contained in the iterative reconstruction code SHIVA—a code project that we have recently initiated. The many ways in which machine learning can be used in the reconstruction work will be contained in a code HERMES that has been initiated with DTRA support. For example, the significant levels of acceleration that will likely be achieved by the use of machine learning techniques will permit us (and are required) to quantify uncertainties in density retrievals. Next, the two-way coupling between density reconstruction and model-based simulation of the hydrodynamics will be contained in code EREBUS, and will permit a fuller realization of the potential of the data to constrain the hydrodynamic model and better address issues related to asymmetries in the problem. Finally, we anticipate that the better consistency with physics achieved in our reconstructions will allow them to be used by X-Division more so than today.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗
DOE OSTI2023

A spatially regularized detector for emergent/re-emergent disease outbreaks

Early detection of outbreaks caused by emergent pathogens, using epidemiological surveillance data i.e., daily case counts, is difficult. This is because the data tend to be noisy during the early epoch of the outbreak. In contrast, the spread-rate of the disease tends to be well-behaved, as it depends only on the mixing patterns of the population and the characteristics of the pathogen, neither of which behave erratically in space-time. In this report, we explore whether the spread-rate can be used for epidemiological surveillance, conditional on case count data. Estimating the spread-rate from case count data allows us to exploit exogenous information, e.g., incubation period distributions etc., which can considerably smooth out any erratic temporal behavior. Further, epidemiological dynamics are spatially correlated, and if case counts are available for multiple areal units e.g., counties, these correlations could potentially be used to suppress noise in the early epoch data. These exogenous information and structure are not exploited by conventional syndromic surveillance detectors to extract a well-behaved latent variable for monitoring purposes. The technical challenge lies in the estimation of the spread-rate field jointly over a collection of areal units; further, the spread-rate varies over time. We develop a method based on mean-field variational inference to approximately estimate the spread-rate field, using a Gaussian Random Field Model for spatial regularization. The method is tested on the estimation of spread-rate in the thirty-three counties of New Mexico and detect the arrival of the Fall 2020 COVID-19 wave in September 2020. We find that the method is scalable, but underestimates the uncertainty in the estimated spread-rate field. We detect the arrival of the Fall 2020 wave a week ahead of conventional syndromic surveillance algorithms, but our simplistic detection algorithm, based on simple anomaly detection, suffers from a high false positive rate, similar to conventional detectors.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2017

Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing

In Hezaveh et al. (2017) we showed that deep learning can be used for model parameter estimation and trained convolutional neural networks to determine the parameters of strong gravitational lensing systems. Here we demonstrate a method for obtaining the uncertainties of these parameters. We review the framework of variational inference to obtain approximate posteriors of Bayesian neural networks and apply it to a network trained to estimate the parameters of the Singular Isothermal Ellipsoid plus external shear and total flux magnification. We show that the method can capture the uncertainties due to different levels of noise in the input data, as well as training and architecture-related errors made by the network. To evaluate the accuracy of the resulting uncertainties, we calculate the coverage probabilities of marginalized distributions for each lensing parameter. By tuning a single hyperparameter, the dropout rate, we obtain coverage probabilities approximately equal to the confidence levels for which they were calculated, resulting in accurate and precise uncertainty estimates. Our results suggest that neural networks can be a fast alternative to Monte Carlo Markov Chains for parameter uncertainty estimation in many practical applications, allowing more than seven orders of magnitude improvement in speed.

Perreault Levasseur, Laurence↗
NASA NTRS1977

A global thermospheric model based on mass spectrometer and incoherent scatter data MSIS. II - Composition

Measurements of O, He, and Ar from neutral gas mass spectrometers on four satellites (Ogo 6, San Marco 3, Aeros A, and AEC-C) and inferred oxygen and hydrogen densities from an ion mass spectrometer on AE-C have been combined with a neutral temperature and nitrogen density model to produce a global model of thermospheric composition in terms of inferred variations at 120 km. The data set covers the time period from mid-1969 to mid-1975. The MSIS (mass spectrometer and incoherent scatter data) model is compared with the Ogo 6 model (Hedin et al., 1974). Ar variations at 120 km tend to be in phase with temperature variations and inverse to the He, O, and H variations.

Hedin, A. E.↗
NASA NTRS1985

Carbon and sulfur cycling through geologic time

Mathematical models of the coupled global systems of sedimentary reservoirs and fluxes are used to infer variations in reservoir sizes and rates of sedimentation over periods of hundreds of millions of years. Perhaps most interesting is the coupled sulfide/sulfate carbon/carbonate system that controls global oxygen and carbon dioxide production and consumption is discussed.

Garrels, R. M.↗
NASA NTRS1985

Small-scale convection beneath the transverse ranges, California: Implications for interpretation of gravity anomalies

Tomographic inversion of upper mantle P wave velocity heterogeneities beneath southern California shows two prominent features: an east-west trending curtain of high velocity material (up to 3% fast) in the upper 250 km beneath the Transverse Ranges and a region of low velocity material (up to 4% slow) in the 100 km beneath the Salton Trough. These seismic velocity anomalies were interpreted as due to small scale convection in the mantle. Using this hypothesis and assuming that temperature and density anomalies are linearly related to seismic velocity anomalies through standard coefficients of proportionality, leads to inferred variations of approx. + or - 300 C and approx. + or - 0.03 g/cc.

Humphreys, E. D.↗
NASA NTRS1985

Fractured toughness of Si3N4 measured with short bar chevron-notched specimens

The short bar chevron-notched specimen is used to measure the plane strain fracture toughness of hot pressed Si3N4. Specimen proportions and chevron-notch angle are varied, thereby varying the amount of crack extension to maximum load (upon which K sub IC is based). The measured toughness (4.68 + or - 0.19 MNm to the 3/2 power) is independent of these variations, inferring that the material has a flat crack growth resistance curve.

Salem, J. A.↗
NASA NTRS1987

Fracture toughness of Si3N4 measured with short bar chevron-notched specimens

The short bar chevron-notched specimen is used to measure the plane strain fracture toughness of hot pressed Si3N4. Specimen proportions and chevron-notch angle are varied, thereby varying the amount of crack extension to maximum load (upon which K sub IC is based). The measured toughness (4.68 + or 0.19 MNm to the 3/2 power) is independent of these variations, inferring that the material has a flat crack growth resistance curve.

Salem, Jonathan A.↗
NASA NTRS1992

Coronal abundances in solar active regions measured by the Solar Maximum Mission flat crystal spectrometer

High resolution soft X-ray spectra acquired by the Flat Crystal Spectrometer (FCS) on solar Maximum Mission provide an excellent data base to study the relative abundances of O, Ne, Mg, and Fe in solar active regions. The FCS data show significant variability for all combinations of these elements. The largest variation occurs for Fe:Ne, which shows region to region changes of up to a factor of 7, and frequent factor of 2 variations in day to day samples of a given region. The atomic data and the ionization balance calculations used to interpret the line ratios affect the actual abundance values obtained, but have little effect on the magnitude of the total range of variation inferred. Resonance scattering of Fe XVII could cause a systematic offset in the abundances determined, but cannot be responsbile for the bulk of the observed variability. While abundance variability complicates the derivation of plasma parameters from spectroscopic measurements, it should offer exciting new clues to the processes which form and heat the corona.

Saba, Julia L. R.↗
NASA NTRS2003

The Chemical Composition and Gas-to-Dust Mass Ratio of Nearby Interstellar Matter

We use recent results on interstellar gas toward nearby stars and interstellar by-products within the solar system to select among the equilibrium radiative transfer models of the nearest interstellar material presented in Slavin & Frisch. For the assumption that O/H - 400 parts per million, models 2 and 8 are found to yield good fits to available data on interstellar material inside and outside of the heliosphere, with the exception of the Ne abundance in the pickup ion and anomalous cosmic-ray populations. For these models, the interstellar medium (ISM) at the entry point to the heliosphere has n(H(sup 0)) = 0.202-0.208/cu cm, n(He(sup 0) = 0.0137-0.0152/cu cm, and ionizations X(H) = 0.29-0.30, X(He) = 0.47-0.51. These best models suggest that the chemical composition of the nearby ISM is approx.60%-70% subsolar if S is undepleted. Both H(0) and H(+) need to be included when evaluating abundances of ions found in warm diffuse clouds. Models 2 and 8 yield an H filtration factor of approx.0.46. Gas-to-dust mass ratios for the ISM toward epsilon CMa are R(sub gd) = 178-183 for solar abundances of Holweger or R(sub gd) = 611-657 for an interstellar abundance standard 70% solar. Direct observations of dust grains in the solar system by Ulysses and Galileo yield R(sub gd) appr0x. 115 for models 2 and 8, supporting earlier results (Frisch and coworkers). If the local ISM abundances are subsolar, then gas and dust are decoupled over small spatial scales. The inferred variation in R(sub gd) over parsec length scales is consistent with the fact that the ISM near the Sun is part of a dynamically active cluster of cloudlets flowing away from the Sco-Cen association. Observations toward stars within approx.500 pc show that R(sub gd) correlates with the percentage of the dust mass that is carried by iron, suggesting that an Fe-rich grain core (by mass) remains after grain destruction. Evidently large dust grains (>10(exp -13) g) and small dust grains (<10(exp -13) g) are not well mixed over parsec length spatial scales in the ISM. It also appears that very small C-dominated dust grains have been destroyed in the ISM within several parsecs of the Sun, since C appears to be essentially undepleted. However, if gas-dust coupling breaks down over the cloud lifetime, the missing mass arguments applied here to determine R(sub gd) and dust grain mineralogy are not appropriate.

Frisch, Priscilla C.↗
NASA NTRS2012

Application of GRACE for Monitoring Groundwater in Data Scarce Regions

In the United States, groundwater storage is somewhat well monitored (spatial and temporal data gaps notwithstanding) and abundant data are freely and easily accessible. Outside of the U.S., groundwater often is not monitored systematically and where it is the data are rarely centralized and made available. Since 2002 the Gravity Recovery and Climate Experiment (GRACE) satellite mission has delivered gravity field observations which have been used to infer variations in total terrestrial water storage, including groundwater, at regional to continental scales. Challenges to using GRACE for groundwater monitoring include its relatively coarse spatial and temporal resolutions, its inability to differentiate groundwater from other types of water on and under the land surface, and typical 2-3 month data latency. Data assimilation can be used to overcome these challenges, but uncertainty in the results remains and is difficult to quantify without independent observations. Nevertheless, the results are preferable to the alternative - no data at all- and GRACE has already revealed groundwater variability and trends in regions where only anecdotal evidence existed previously.

Rodell, Matt↗
NASA NTRS1983

Case studies using GOES infrared data and a planetary boundary layer model to infer regional scale variations in soil moisture

Modeled temperature data from a one-dimensional, time-dependent, initial value, planetary boundary layer model for 16 separate model runs with varying initial values of moisture availability are applied, by the use of a regression equation, to longwave infrared GOES satellite data to infer moisture availability over a regional area in the central U.S. This was done for several days during the summers of 1978 and 1980 where a large gradient in the antecedent precipitation index (API) represented the boundary between a drought area and a region of near normal precipitation. Correlations between satellite derived moisture availability and API were found to exist. Errors from the presence of clouds, water vapor and other spatial inhomogeneities made the use of the measurement for anything except the relative degree of moisture availability dubious.

Rose, F. G.↗
NASA NTRS1975

Winter bulge and diurnal variations in hydrogen inferred from AE-C composition measurements

The atomic hydrogen distribution at 250 km during December 1974 solstice was inferred, considering charge exchange equilibrium, from Atmosphere Explorer-C measurements of n(H(+)), n(O(+) and N(O). An empirical model, derived from the observations by least suqare analysis in terms of spherical harmonics, has the following characteristics: (1) n(H) increases by as much as a factor of two between the summer and winter hemispheres, (2) the n(H) diurnal variation is largest at the equator and (3) the diurnal variation is larger in the winter hemisphere than in the summer. Similar analysis of the gas temperature derived from n(N2) measurements reveals that all n(H) and Tg spherical harmonic coefficients are anticorrelated. Both the diurnal and latitudinal (annual) n(H) and Tg amplitudes are in substantial agreement with the zero flux condition, in which exospheric flow dominates the hydrogen distribution. The observed diurnal phase lag of n(H) with respect to Tg is about one hour, agreeing with theory.

Brinton, H. C.↗
NASA NTRS1985

A cool anomaly off northern California - An investigation using IR imagery and in situ data

The OPTOMA (Ocean Prediction Through Observation, Modeling and Analysis) program is developing an ocean descriptive-predictive system for four-dimensional data assimilation. It is presently concerned with the mesoscale variability in the California Current System (CCS). The present paper has the objective to assess the relationship of surface temperature structure to subsurface temperature structure and flow fields. Surface temperature anomalies are related to the mesoscale horizontal advective and subsurface mass fields. The cool anomaly off northern California in summer 1982 is discussed, taking into account the temperature structure along transects, T-S variations and inferred water masses, variation of horizontal temperature patterns with depth, and synoptic information from IR data.

Rienecker, M. M.↗
NASA NTRS1991

Inferring solar structure variations from photometric and helioseismic observations

Present large scale photometric and helioseismic data may be related to spatial and temporal deviations from an otherwise static, spherical solar convection zone (SCZ). New space experiments combine precise photometric, helioseismic, and magnetic observational capabilities (like SOHO and MDI) and will provide data needed to understand the interaction of magnetic fields and global scale circulation in the SCZ. Even simple physical arguments make it clear that the anticipated accuracy of these new experiments virtually guarantees new constraints on models of the SCZ.

Kuhn, J. R.↗
NASA NTRS2006

Rigid Body Rate Inference from Attitude Variation

In this paper we research the extraction of the angular rate vector from attitude information without differentiation, in particular from quaternion measurements. We show that instead of using a Kalman filter of some kind, it is possible to obtain good rate estimates, suitable for spacecraft attitude control loop damping, using simple feedback loops, thereby eliminating the need for recurrent covariance computation performed when a Kalman filter is used. This considerably simplifies the computations required for rate estimation in gyro-less spacecraft. Some interesting qualities of the Kalman filter gain are explored, proven and utilized. We examine two kinds of feedback loops, one with varying gain that is proportional to the well known Q matrix, which is computed using the measured quaternion, and the other type of feedback loop is one with constant coefficients. The latter type includes two kinds; namely, a proportional feedback loop, and a proportional-integral feedback loop. The various schemes are examined through simulations and their performance is compared. It is shown that all schemes are adequate for extracting the angular velocity at an accuracy suitable for control loop damping.

Bar-Itzhack, I. Y.↗