Search NASA⌕ Search

SEARCH · Search NASA

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 919 records · Page 51

Characterizing the subglacial environment of lower Thwaites Glacier using radar modeling

Understanding the spatial heterogeneity beneath Thwaites Glacier, West Antarctica, is vital to projecting its impact on future sea levels. Radar-echo sounding (RES) is commonly used to infer subglacial conditions, but these data can be challenging to interpret. We assess basal heterogeneity across Thwaites Glacier by comparing RES returns to a radar backscattering simulator for over 400 km of RES data. The modeled variations in bed returned power exhibited a strong correlation with actual RES data in 40% of our simulated flight segments, which we consider evidence for a relatively homogeneous glacier bed. Other sites (40%) demonstrated improved fit quality when hydrology or substrate transitions were introduced in the bed material model. The remaining simulated segments (20%) were diagnosed as having more complex basal heterogeneity. The spatial distribution of complex heterogeneity appears to coincide with asymmetric patterns in the RES specularity content, which has been interpreted in previous studies as a signature for channelized hydrology. Conversely, the homogeneous substrate locations coincide with areas of fast-moving ice in western Thwaites. Our simulation method can isolate power variations induced by material heterogeneity vs topography, which is an important limitation of existing RES analysis methods.

Antarctic glaciology↗

Response of the upper atmosphere to variations in the solar soft x-ray irradiance

Terrestrial far ultraviolet (FUV) airglow emissions have been suggested as a means for remote sensing the structure of the upper atmosphere. The energy which leads to the excitation of FUV airglow emissions is solar irradiance at extreme ultraviolet (EUV) and soft x-ray wavelengths. Solar irradiance at these wavelengths is known to be highly variable; studies of nitric oxide (NO) in the lower thermosphere have suggested a variability of more than an order of magnitude in the solar soft x-ray irradiance. To properly interpret the FUV airflow, the magnitude of the solar energy deposition must be known. Previous analyses have used the electron impact excited Lyman-Birge-Hopfield (LBH) bands of N2 to infer the flux of photoelectrons in the atmosphere and thus to infer the magnitude of the solar irradiance. This dissertation presents the first simultaneous measurements of the FUV airglow, the major atmospheric constituent densities, and the solar EUV and soft x-ray irradiances. The measurements were made on three flights of an identical sounding rocket payload at different levels of solar activity. The linear response in brightness of the LBH bands to variations in solar irradiance is demonstrated. In addition to the N2 LBH bands, atomic oxygen lines at 135.6 and 130.4 nm are also studied. Unlike the LBH bands, these emissions undergo radiative transfer effects in the atmosphere. The OI emission at 135.6 nm is found to be well modeled using a radiative transfer calculation and the known excitation processes. Unfortunately, the assumed processes leading to OI 130.4 nm excitation are found to be insufficient to reproduce the observed variability of this emission. Production of NO in the atmosphere is examined; it is shown that a lower than previously reported variability in the solar soft x-ray irradiance is required to explain the variability of NO.

Bailey, Scott Martin↗

Multi-Epoch XMM Observations of NGC4258

The goal of our program is to characterize time variation of X-ray emission from the AGN in NGC4258 and to look for correlations with radio emission (maser line and continuum). Characterization of the X-ray properties includes estimation of absorbing column, flux, and Fe K emission line strength. In the past year we have completed reduction of five epochs of XMM data and 42 epochs of radio spectroscopic and imaging data. We obtained four of the X-ray epochs as part of our own program. The fifth epoch is public and was obtained by a GTO program. In addition to the XMM data, we have retrieved and reduced four epochs of data from the CXO archive. The radio data had been obtained with the VLA, VLBA, and Effelsberg lOOm antenna. Taken together, the X-ray and radio data are a unique time monitoring database. We have found that the Fe line faded over the period mid-1999 to mid-2000 and is now almost undetectable by XMM. In this same period, the radio emission showed no unusual activity. We infer that the line emission arises either in a time varying disk corona much less than 1 pc in size or from a portion of the disk well within the inner edge of the molecular disk responsible for the radio line emission. We have found that variations in X-ray absorbing column, source flux, and radio emission are not correlated, from which we argue that the molecular disk is either not clumpy or that it has a very low areal filling factor since it is believed to cross the line of sight to the central engine at a radius of about 0.3 pc. In the past month, the public CXO data we have analyzed has been published by another group, which claims the provocative discovery of time variable Fe K absorption lines at one epoch. We do not see these features in our reduction of the same data or in our own XMM data. This will be emphasized in our publication.

Greenhill, Lincoln J.↗

Stratospheric effects of solar ultraviolet variations on the solar rotation time scale

A summary is presented of some current work on measurement and interpretation of stratospheric ozone and temperature responses to observed short term solar ultraviolet variations. Although some studies have yielded provisional evidence for a nearly in-phase ozone-solar cycle relationship, they extend at most over only one or two 11 year cycles so the statistical significance of the correlations is not large. Similarly, the relatively short lengths of individual satellite data sets combined with the problem of estimating the effect of changes in instrument sensitivity (drift) during the observing period have complicated attempts to infer long term or solar cycle ozone trends. The solar rotation and active region development time scale provides an alternate time scale for which detailed studies of middle atmospheric ozone and temperature responses to solar ultraviolet variability are currently possible using available satellite data sets. At tropical latitudes where planetary wave amplitudes are relatively small, clear correlative evidence for the existence of middle atmospheric ozone and temperature responses to short term solar ultraviolet variations has been obtained in recent years. These measurements will ultimately allow improved empirical and theoretical calculations of longer term solar induced ozone and temperature variations at low and middle latitudes.

Hood, L. L.↗

Geomechanical simulation of the stress tensor rotation caused by injection of cold water in a deep geothermal reservoir

We present a three-dimensional thermohydromechanical numerical study of the evolution and distribution of the stress tensor within the northwest part of The Geysers geothermal reservoir (in California), including a detailed study of the region around one injection well from 2003 to 2012. Initially, after imposing a normal faulting stress regime, we calculated local changes in the stress regime around injection wells. Our results were compared with previously published studies in which the stress state was inferred from inverting the focal plane mechanism of seismic events. Our main finding is that changes in stress tensor orientation are caused by injection-induced progressive cooling of the reservoir, as well as by the seasonal variations in injection rate. Because of the gravity flow and cooling around a liquid zone formed by the injection, the vertical stress reduction is larger and propagates far below the injection well. At the same time, the horizontal stress increases, mostly because of stress redistribution below and above the cooling area. These two phenomena cause the rotation of the stress tensor and the appearance of a strike-slip regime above, inside, and below the cooling area. The cooling and the associated rotation of the stress regime can play a significant role in the observed long-term deepening of the microseismicity below active injection wells.

58 GEOSCIENCES↗

Learning model combining convolutional deep neural network with a self-attention mechanism for AC optimal power flow

Alternating current optimal power flow (OPF) analysis is critical for efficient and reliable operation of power systems. For large systems or repetitive computations, the traditional methods such as the direct and gradient methods, or non-traditional methods, such as the genetic algorithm and simulating annealing, are time-consuming and unsuitable for real-time computing. The work in this paper proposes a novel framework to obtain the optimal solution of power flow in real-time using a combination of convolutional neural networks and a self-attention mechanism. All parameters of the power networks are rearranged in an image-like shape of a multi-channel image where each channel is a two-dimensional matrix. The proposed approach is adaptive with every input size of power systems as well as frequent variations of network topologies without intervention to the framework core. The encompassment of all power system contexts in which all parameters of internal elements, generation costs, and topology information are included, contributes to the higher accuracy of inference compared to other current machine-learning-based OPF-solving methods. Besides, the proposed framework established on ubiquitous platforms is effortlessly integrated into current infrastructures of power systems, and the great efficiency along with the computation speed may serve as a critical point for practical implications, such as enabling faster decision-making during real-time operations, predicting system contingencies, and remedial actions based on an offline pre-trained model. Furthermore, this supervised learning process is applied to the dataset of four case studies of meshed power systems: the IEEE 5-bus system (IEEE-5), the IEEE 30-bus system (IEEE-30), the IEEE 39-bus system (IEEE-39), and the IEEE 57-bus system (IEEE-57) to prove the efficacy of the proposed method.

42 ENGINEERING↗

Structural and physical properties of layered oxy-arsenides LnRuAsO (Ln=La, Nd, Sm, Gd)

Polycrystalline samples of LaRuAsO, NdRuAsO, SmRuAsO, and GdRuAsO have been synthesized and studied using powder x-ray diffraction, electrical transport, magnetization, and heat capacity measurements. Variations in structural properties across the series reveal a trend toward more ideal tetrahedral coordination around Ru as the size of the rare earth element is reduced. The lattice parameters of these Ru compounds show a more anisotropic response to variation in Ln than their Fe analogs, and significant anisotropy in thermal expansion is also observed. Transport measurements show metallic behavior, and carrier concentrations near 10{sup 21}-10{sup 22} electrons per cm{sup 3} are inferred from simple analysis of Hall effect measurements. Anomalies in resistivity, magnetization, and heat capacity indicate antiferromagnetic ordering of rare earth moments at 5 K for GdRuAsO, 4.5 K for SmRuAsO, and <2K for NdRuAsO. Magnetization measurements on LaRuAsO show no evidence of a magnetic moment on Ru. Observed behaviors are compared to those reported for similar Fe and Ru compounds. - Graphical abstract: Analysis of crystal structure, electrical transport, magnetic susceptibility, and heat capacity of LaRuAsO, NdRuAsO, SmRuAsO, and GdRuAsO. Highlights: Black-Right-Pointing-Pointer Experimental investigation of LaRuAsO, NdRuAsO, SmRuAsO, and GdRuAsO. Black-Right-Pointing-Pointer Anisotropic lattice response to changing Ln radius and temperature. Black-Right-Pointing-Pointer Ru coordination becomes more ideal as Ln radius is reduced. Black-Right-Pointing-Pointer Transport measurements reveal metallic conduction dominated by electrons. Black-Right-Pointing-Pointer Magnetic measurements indicate antiferromagnetic ordering Nd, Sm, and Gd moments.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Structural, dielectric and ferroelectric properties of La and Ni codoped BiFeO{sub 3}

The codoped Bi{sub 1-x}La{sub x}Fe{sub 1-y}Ni{sub y}O{sub 3} (x = 0.0, 0.1; y = 0.0, 0.05) samples are processed by solid-state reaction route in order to study structural, dielectric and ferroelectric properties. X-ray diffraction patterns of La/Ni doped samples have been performed using the rhombohedral (R3c) symmetry. Raman scattering measurement infers nine Raman active phonon modes. The substitution of Ni ion at Fe-site in BiFeO{sub 3} essentially changes the modes position i.e. all the modes are observed to shift at lower wave number. Frequency variation of the dielectric constant (ε′) and dielectric loss (tanδ) has been investigated. It decreases with increasing frequency of the applied alternating field and become constant at high frequencies signifies substantial role of interfacial polarization. The remnant polarization and coercive field for BLFO, BFNO, BLFNO are 0.08, 0.03, 0.64 μC/cm{sup 2} and 0.24, 0.22, 0.25 kV/cm, respectively. La/Ni ion doping in BFO has successfully enhanced the polarization.

36 MATERIALS SCIENCE↗

XMM-NEWTON OBSERVATIONS OF THE DWARF NOVA RU Peg IN QUIESCENCE: PROBE OF THE BOUNDARY LAYER

We present an analysis of X-ray and UV data obtained with the XMM-Newton Observatory of the long-period dwarf nova RU Peg. RU Peg contains a massive white dwarf (WD), possibly the hottest WD in a dwarf nova (DN), it has a low inclination, thus optimally exposing its X-ray emitting boundary layer (BL), and has an excellent trigonometric parallax distance. We modeled the X-ray data using XSPEC assuming a multi-temperature plasma emission model built from the MEKAL code (i.e., CEVMKL). We obtained a maximum temperature of 31.7 keV, based on the European Photon Imaging Camera MOS1, 2 and pn data, indicating that RU Peg has an X-ray spectrum harder than most DNe, except U Gem. This result is consistent with and indirectly confirms the large mass of the WD in RU Peg. The X-ray luminosity we computed corresponds to a BL luminosity for a mass accretion rate of 2 Multiplication-Sign 10{sup -11} M{sub sun} yr{sup -1} (assuming M{sub wd} = 1.3 M{sub sun}), in agreement with the expected quiescent accretion rate. The modeling of the O VIII emission line at 19 A as observed by the Reflection Grating Spectrometer implies a projected stellar rotational velocity v{sub rot}sin i = 695 km s{sup -1}, i.e., the line is emitted from material rotating at {approx}936-1245 km s{sup -1} (i {approx} 34 Degree-Sign -48 Degree-Sign) or about 1/6 of the Keplerian speed; this velocity is much larger than the rotation speed of the WD inferred from the Far Ultraviolet Spectroscopic Explorer spectrum. Cross-correletion analysis yielded an undelayed (time lag {approx} 0) component and a delayed component of 116 {+-} 17 s where the X-ray variations/fluctuations lagged the UV variations. This indicates that the UV fluctuations in the inner disk are propagated into the X-ray emitting region in about 116 s. The undelayed component may be related to irradiation effects.

79 ASTRONOMY AND ASTROPHYSICS↗

Infall-driven protostellar accretion and the solution to the luminosity problem

We investigate the role of mass infall in the formation and evolution of protostars. To avoid ad hoc initial and boundary conditions, we consider the infall resulting self-consistently from modeling the formation of stellar clusters in turbulent molecular clouds. We show that infall rates in turbulent clouds are comparable to accretion rates inferred from protostellar luminosities or measured in pre-main-sequence stars. They should not be neglected in modeling the luminosity of protostars and the evolution of disks, even after the embedded protostellar phase. We find large variations of infall rates from protostar to protostar, and large fluctuations during the evolution of individual protostars. In most cases, the infall rate is initially of order 10{sup –5} M {sub ☉} yr{sup –1}, and may either decay rapidly in the formation of low-mass stars, or remain relatively large when more massive stars are formed. The simulation reproduces well the observed characteristic values and scatter of protostellar luminosities and matches the observed protostellar luminosity function. The luminosity problem is therefore solved once realistic protostellar infall histories are accounted for, with no need for extreme accretion episodes. These results are based on a simulation of randomly driven magnetohydrodynamic turbulence on a scale of 4 pc, including self-gravity, adaptive-mesh refinement to a resolution of 50 AU, and accreting sink particles. The simulation yields a low star formation rate, consistent with the observations, and a mass distribution of sink particles consistent with the observed stellar initial mass function during the whole duration of the simulation, forming nearly 1300 sink particles over 3.2 Myr.

79 ASTRONOMY AND ASTROPHYSICS↗

Indications of flow near maximum compression in layered deuterium-tritium implosions at the National Ignition Facility

Here, an accurate understanding of burn dynamics in implosions of cryogenically layered deuterium (D) and tritium (T) filled capsules, obtained partly through precision diagnosis of these experiments, is essential for assessing the impediments to achieving ignition at the National Ignition Facility. We present measurements of neutrons from such implosions. The apparent ion temperatures T ion are inferred from the variance of the primary neutron spectrum. Consistently higher DT than DD T ion are observed and the difference is seen to increase with increasing apparent DT T ion . The line-of-sight rms variations of both DD and DT T ion are small, ~150eV, indicating an isotropic source. The DD neutron yields are consistently high relative to the DT neutron yields given the observed T ion . Spatial and temporal variations of the DT temperature and density, DD-DT differential attenuation in the surrounding DT fuel, and fluid motion variations contribute to a DT Tion greater than the DD T ion , but are in a one-dimensional model insufficient to explain the data. We hypothesize that in a three-dimensional interpretation, these effects combined could explain the results.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Crystallization of a lunar magma ocean

The results of applying the presently most detailed thermal model to the cooling, crystallization and differentiation of a lunar magma ocean are used to delineate its general features, i.e., those features that are model independent within reasonable variations in ocean characteristics. Two primitive lunar compositions are used to illustrate the effect of different compositions and phase equilibria on the solidification time and mineralogy of magma oceans. The magma ocean thermal model led to some inferences about the thermal evolution of the outer portion of the moon modeled as isolated magma pods rather than as a magma ocean.

Minear, J. W.↗

Cometary kilometric radio waves and plasma waves correlated with ion pick-up effect at Comet Halley

Bow-shock movements at Comet Halley are inferred from the discrete spectra of the cometary kilometric radiation (30-195 kHz); the observed emissions can be interpreted as being generated and propagating from the moving shock. The shock motion is possibly associated with the time variation of the solar wind and cometary outgassing. It is concluded that these plasma wave phenomena are manifestations of ion pick-up processes, which occur even in a remote region 7 million to 10 million km from the cometary nucleus.

Oya, H.↗

Satellite remote sensing of primary production

Leaf structure and function are shown to result in distinctive variations in the absorption and reflection of solar radiation from plant canopies. The leaf properties that determine the radiation-interception characteristics of plant canopies are directly linked to photosynthesis, stomatal resistance and evapotranspiration and can be inferred from measurements of reflected solar energy. The effects of off-nadir viewing and atmospheric constituents, coupled with the need to measure changing surface conditions, emphasize the need for multitemporal measurements of reflected radiation if primary production is to be estimated.

Tucker, C. J.↗

Out-of-ecliptic tests of the inverse correlation between solar wind speed and coronal expansion factor

The paper addresses the question of whether out-of-ecliptic measurements satisfy the inverse correlation between wind speed at 1 AU and flux-tube divergence in the corona, already found from measurements in the ecliptic. Using the in-ecliptic calibration, out-of-ecliptic speeds are derived from coronal expansion factors determined from global observations of photospheric field and their current-free coronal extension. These derived speeds are compared with speeds inferred from interplanetary scintillation measurements during 1972-1988 and with in situ speeds measured by the Pioneer 11 spacecraft at 16 deg N latitude during 1984-1988. These three sets of wind speed show the same overall variation with latitude and time during the sunspot cycle, with higher latitudes having more years of fast wind than lower latitudes and all latitudes having slow wind at sunspot maximum. Although some detailed discrepancies are also present, the overall agreement is comparable to that achieved in the ecliptic plane.

Sheeley, N. R., Jr.↗

The magnetic evolution of AR 6555 which lead to two impulsive, readily compact, X-type flares

We study the evolution of the vector magnetic field and the sunspot motions observed in AR 6555 during 23-26 Mar. 1991. This region displays two locations of large magnetic shear that were also sites of flare activity. The first location produced two large (X-class) flares during the period covered by our observations. The second location had larger magnetic shear than the first, but produced only small (M- and C-class) flares during our observations. We study the evolution of the photospheric magnetic field in relation to the large flares in the first location. These flares occurred around the same included polarity, and have very similar characteristics (soft X-ray light curves, energies, etc.). However, the whole active region has changed substantially in the period between them. We found several characteristics of the region that appear related to the occurrence of these flares. (1) The flares occurred near regions of large magnetic 'shear,' but not at the locations of maximum shear or maximum field. (2) Potential field extrapolations of the observed field suggest that the topology changed, prior to the first of the two flares, in such a way that a null appeared in the coarse magnetic field. (3) This null was located close to both X-class flares, and remained in that location for a few days while the two flares were observed. (4) The flaring region has a pattern of vector field and sunspot motions in which material is 'squeezed' along the polarity inversion line. This pattern is very different from that usually associated with shearing arcades, but it is similar to that suggested previously by Fontenla and Davis. The vertical electric currents, inferred from the transverse field, are consistent with this pattern. (5) A major reconfiguration of the longitudinal field and the vertical electric currents occurred just prior to the first of the two flares. Both changes imply substantial variations of the magnetic structure of the region. On the basis of the available data we suggest that these changes made the flaring possible, and we develop a scenario that can explain the origin of the magnetic free energy that was released in these flares.

Ambastha, A.↗

The Magnetic Evolution of AR 6555 which led to Two Impulsive, Relatively Compact, X-Type Flares

We study the evolution of the vector magnetic field and the sunspot motions observed in AR 6555 during 1991 March 23-26. This region displays two locations of large magnetic shear that were also sites of flare activity. The first location produced two large (X-class) flares during the period covered by our observations. The second location had larger magnetic shear than the first but produced only small (M- and C-class) flares during our observations. We study the evolution of the photospheric magnetic field in relation to the large flares in the first location. These flares occurred around the same included polarity and have very similar characteristics (soft X-ray light curves, energies, etc,). However, the whole active region has changed substantially in the period between them. We found several characteristics of the region that appear related to the occurrence of these flares: (1) The flares occurred near regions of large magnetic 'shear' but not at the locations of maximum shear or maximum field. (2) Potential field extrapolations of the observed field suggest that the topology changed, prior to the first of the two flares, in such a way that a null appeared in the coarse magnetic field. (3) This null was located close to both X-class flares and remained in that location for a few days while the two flares were observed. (4) The flaring region has a pattern of vector field and sunspot motions in which material is 'squeezed' along the polarity inversion line. This pattern is very different from that usually associated with shearing arcades, but it is similar to that suggested previously by Fontenia and Davis. The vertical electric currents, inferred from the transverse field, are consistent with this pattern. (5) A major reconfiguration of the longitudinal field and the vertical electric currents occurred just prior to the first of the two flares. Both changes imply substantial variations of the magnetic structure of the region. On the basis of the available data we suggest that these changes made the flaring possible, and we develop a scenario that can explain the origin of the magnetic free-energy that was released in these flares.

Fontenla, J. M.↗

A Neural Network Aero Design System for Advanced Turbo-Engines

An inverse design method calculates the blade shape that produces a prescribed input pressure distribution. By controlling this input pressure distribution the aerodynamic design objectives can easily be met. Because of the intrinsic relationship between pressure distribution and airfoil physical properties, a Neural Network can be trained to choose the optimal pressure distribution that would meet a set of physical requirements. Neural network systems have been attempted in the context of direct design methods. From properties ascribed to a set of blades the neural network is trained to infer the properties of an 'interpolated' blade shape. The problem is that, especially in transonic regimes where we deal with intrinsically non linear and ill posed problems, small perturbations of the blade shape can produce very large variations of the flow parameters. It is very unlikely that, under these circumstances, a neural network will be able to find the proper solution. The unique situation in the present method is that the neural network can be trained to extract the required input pressure distribution from a database of pressure distributions while the inverse method will still compute the exact blade shape that corresponds to this 'interpolated' input pressure distribution. In other words, the interpolation process is transferred to a smoother problem, namely, finding what pressure distribution would produce the required flow conditions and, once this is done, the inverse method will compute the exact solution for this problem. The use of neural network is, in this context, highly related to the use of proper optimization techniques. The optimization is used essentially as an automation procedure to force the input pressure distributions to achieve the required aero and structural design parameters. A multilayered feed forward network with back-propagation is used to train the system for pattern association and classification.

Sanz, Jose M.↗