Search NASA⌕ Search

SEARCH · Search NASA

Results for “velocity data”

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 901 records · Page 50
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Seismic Waveform Inversion Capability on Resource-Constrained Edge Devices

Seismic full wave inversion (FWI) is a widely used non-linear seismic imaging method used to reconstruct subsurface velocity images, however it is time consuming, has high computational cost and depend heavily on human interaction. Recently, deep learning has accelerated it’s use in several data-driven techniques, however most deep learning techniques suffer from overfitting and stability issues. In this work, we propose an edge computing-based data-driven inversion technique based on supervised deep convolutional neural network to accurately reconstruct the subsurface velocities. Deep learning based data-driven technique depends mostly on bulk data training. In this work, we train our deep convolutional neural network (DCN) (UNet and InversionNet) on the raw seismic data and their corresponding velocity models during the training phase to learn the non-linear mapping between the seismic data and velocity models. The trained network is then used to estimate the velocity models from new input seismic data during the prediction phase. The prediction phase is performed on a resource-constrained edge device such as Raspberry Pi. Raspberry Pi provides real-time and on-device computational power to execute the inference process. In addition, we demonstrate robustness of our models to perform inversion in the presence on noise by performing both noise-aware and no-noise training and feeding the resulting trained models with noise at different signal-to-noise (SNR) ratio values. We make great efforts to achieve very feasible inference times on the Raspberry Pi for both models. Specifically, the inference times per prediction for UNet and InversionNet models on Raspberry Pi were 22 and 4 s respectively whilst inference times for both models on the GPU were 2 and 18 s which are very comparable. Finally, we have designed a user-friendly interactive graphical user interface (GUI) to automate the model execution and inversion process on the Raspberry Pi.

Manu, Daniel (ORCID:0000000154982677)↗
NASA NTRS1973

Turbulence measurements in a compressible boundary layer subjected to a shock-wave-induced adverse pressure gradient.

The rms intensities of fluctuating mass flux and total temperature and their correlation coefficients are given for the case of an adiabatic, Mach 4, axisymmetric shock-wave boundary-layer interaction. Data were obtained upstream, within, and downstream of the interaction by the use of constant temperature hot-wire anemometer. Turbulence spectra and quantitative behavior from oscilloscope traces are shown at selected locations. The measurements indicate that certain frequencies of the turbulence are increased as a result of the interaction and that the mass flux and total temperature fluctuations remain highly correlated over most of the boundary layer throughout the interaction. The present data are also transformed to rms intensities of fluctuating static temperature and velocity and compared with existing data obtained in adiabatic flows.

Rose, W. C.↗
NASA NTRS1984

Crossed hot-wire data acquisition and reduction system

The report describes a system for rapid computerized calibration acquisition, and processing of data from a crossed hot-wire anemometer is described. Advantages of the system are its speed, minimal use of analog electronics, and improved accuracy of the resulting data. Two components of mean velocity and turbulence statistics up to third order are provided by the data reduction. Details of the hardware, calibration procedures, response equations, software, and sample results from measurements in a turbulent plane mixing layer are presented.

Westphal, R. V.↗
NASA NTRS2001

Separating Spatial and Temporal Variations of the Aurora Using Two Nearly Colocated Satellites

This final report describes the efforts accomplished during the grant's period of performance, covering the period of 1 May 1997 to 30 April 2001, of a NASA Supporting Research and Technology Program grant under the Ionospheric, Thermospheric, and Mesospheric Physics component of the Sun-Earth Connections program. We have met and exceeded the goals set forth in the proposed research objectives. Referred publications have appeared in the scientific literature and several others are in the review process. In addition, numerous invited and contributed presentations of these studies were presented at national and international meetings during the performance period. One graduate student completed his PhD and won two AGU Best Student Paper awards based on research funded by this grant. These studies are summarized below. The science goal delineated in the initial proposal was "to systematically explore the temporal and spatial characteristics of the aurora in a way heretofore impossible, using data from two coplanar DMSP spacecraft." We accomplished this goal through a series of related studies. One study used these unique data to establish the role of Ps6 waves in coupling between the magnetosphere and the auroral ionosphere (omega bands) during the recovery phase of a magnetic storm; the published paper demonstrated the causal relationships between geospace processes occurring in different regions and established a simple conceptual model based on the fortuitous constellation of observations. In the second string of papers, we used these data to explore velocity-dispersed ions (VDIS) in and near the cusp, to test region identification models, and to look at space/time structure of auroral precipitation. On the first topic, the unique DMSP data revealed a remarkable double VDIS with a latitudinal overlap. This could only be explained in terms of a unified reconnection geometry that builds on several earlier unrelated models. The paper outlining this discovery has drawn considerable attention from the community and is currently in press - it adds significantly to the debate over whether reconnection is study state versus bursty and patchy versus global. The second paper develops the model further by incorporating the electron signature - these ionospheric particle precipitation signatures reveal the presence of magnetospheric "fossilized" FTEs, demonstrating the power of ionospheric measurements as a remote diagnostic of magnetospheric processes. Finally, the general nature of aurora] stability and coherence and region identification by particle characteristics were fully explored in a final paper. We identify candidate mechanisms controlling coherence time scales and length scales and refine boundary region identification criteria. We also use the dual-DMSP observations to identify the open and closed LLBL region and related its significance to the generalized bursty, multiple x-line model developed in the first paper. All of these topics are chapters of Dr. Boudouridis' recently completed PhD thesis.

Spence, Harlan E.↗
DOE OSTI2020

The Effects of Assumed Source Depth and Shear-Wave Velocity on Moment Tensors Estimated for Small, Contained Chemical Explosions in Granite

ABSTRACT The Source Phenomenology Experiment (SPE-Arizona) included of a series of chemical explosions detonated within a copper mine in Arizona. This study focuses on ground motions from detonations in the copper mine, which are analyzed to assess the uniqueness of the resulting source representation when the source region propagation characteristics have a range of possible models. P-wave velocities are well constrained by refraction data with less constraint of the S-wave velocities. The effects of explosion source depth and VS are assessed with Green’s functions for a range of models in which VP is held constant. Propagation models with a Poisson’s value of 0.25 and a source depth 30–60 m most accurately replicate the data. The explosion was detonated at a centroid depth of 30 m, so trade-offs in depth are demonstrated. The compensated linear vector dipole and explosion components of the Green’s functions convolved with a Mueller–Murphy source function are compared. Both produce significant energy in the 2–12 Hz band, due to surface-wave contributions with no clear depth dependencies above 20 Hz. The range of propagation models is used with the observational data to invert for the frequency-domain moment tensor. Fits to the data from these inversions have cross-correlation values of 0.64, demonstrating effectiveness in replicating the observations with the assumed propagation path effects and resulting source function. Inversions produce horizontal dipoles (Mxx and Myy), roughly half the maximum amplitude of Mzz, consistent with a compensated linear vector dipole source, which is frequency dependent. Denny and Johnson, Mueller–Murphy, Walter and Ford, and the revised Mueller–Murphy source models, parameterized for granite, are compared to the moment tensors. Despite a nonisotropic moment tensor source, the revised Mueller–Murphy isotropic source model best replicates the long-period moments, overshoot, and corner frequency.

Geochemistry & Geophysics↗
NASA NTRS2005

Dynamic Weakening (Extinction) of Simple Hydrocarbon-air Counterflow Diffusion Flames by Oscillatory Inflows

This study of laminar non-premixed HC-air flames used an Oscillatory-input Opposed Jet Burner (OOJB) system developed from a previously well-characterized 7.2-mm Pyrex-nozzle OJB system. Over 600 dynamic Flame Strength (FS) measurements were obtained on unanchored (free-floating) laminar Counterflow Diffusion Flames (CFDF's). Flames were stabilized using plug inflows having steady-plus-sinusoidal axial velocities of varied magnitude, frequency, f, up to 1600 Hz, and phase angle from 0 (most data) to 360 degrees. Dynamic FS is defined as the maximum average air input velocity (U(sub air), at nozzle exit) a CFDF can sustain before strain-induced extinction occurs due to prescribed oscillatory peak-to-peak velocity inputs superimposed on steady inputs. Initially, dynamic flame extinction data were obtained at low f, and were supported by 25-120 Hz Hot-Wire cold-flow velocity data at nozzle exits. Later, expanded extinction data were supported by 4-1600 Hz Probe Microphone (PM) pk/pk P data at nozzle exits. The PM data were first obtained without flows, and later with cold stagnating flows, which better represent speaker-diaphragm dynamics during runs. The PM approach enabled characterizations of Dynamic Flame Weakening (DFW) of CFDF's from 8 to 1600 Hz. DFW was defined as % decrease in FS per Pascal of pk/pk P oscillation, namely, DFW = - 100 d(U(sub air) / U(sub air),0Hz) / d(pkpk P). The linear normalization with respect to acoustic pressure magnitude (and steady state (SS) FS) led to a DFW unaffected by strong internal resonances. For the C2H4/N2-air system, from 8 to 20 Hz, DFW is constant at 8.52 plus or minus 0.20 (% weakening)/Pa. This reflects a quasi-steady flame response to an acoustically induced dU(sub air)/dP. Also, it is surprisingly independent of C2H4/N2 mole fraction due to normalization by SS FS. From 20 to approximately 150 Hz, the C2H4/N2 air-flames weakened progressively less, with an inflection at approximately 70 Hz, and became asymptotically insensitive (DFW approximately 0) at approximately 300 Hz, which continued to 1600 Hz. The DFW of CH4-air flames followed a similar pattern, but showed much greater weakening than C2H4/N2-air flames; i.e., the quasi-steady DFW (8 to approximately 15 Hz) was 44.3 %/Pa, or approximately 5x larger, even though the 0 Hz (SS) FS was only 3.0 x smaller. The quasi-steady DFW's of C3H8-air and C2H6-air were intermediate at 34.8 and 20.9 %Pa, respectively. The DFW profiles of all four fuels, at various frequencies, correlated well but non-linearly with respective SS FS's. Notably, the DFW profile for C3H8 air fell more rapidly in the range greater than 15 to 60 Hz, compared with the 1- and 2-carbon fuels. This may indicate a shift in chemical kinetics, and/or O2 transport to a flame that moved closer to the fuel-side. In conclusion, Dynamic Flame Weakening limits appear significant and unique for each fuel, and correlate closely, but non-linearly, with Steady-State Flame Strengths at any given frequency. For reasons unknown, the dynamic flames didn't weaken more at intermediate frequencies (e.g., at 20-50 Hz) than they did at low frequencies (less than 15 Hz), where quasi-steady weakening appears to dominate. Quasi-steady flame weakening ostensibly represents a transient input strain rate maximum that just exceeds the steady-state strain-rate-limited extinction limit for a few cycles. Clearly, further detailed mechanistic understanding is needed in the fall-off region.

Pellett, G.↗
DOE OSTI2023

High-velocity CP2 stars in the Galactic halo

The second subclass of chemically peculiar stars, the CP2 stars, are early-type stars exhibiting anomalous abundances with variable line strengths possibly also accompanied by photometric variability that typically belong to the Galactic disk. However, a small fraction of these objects were recently found to be located far from the Galactic plane and are thought to belong to the Galactic halo, which is unexpected for this class of objects. Our study investigates the halo membership of the proposed halo CP2 star candidates based on their velocities and Galactic orbits, to determine their points of origin. In addition, we searched for further halo CP2 star candidates using an as-yet-unpublished sample of CP stars. Our analysis relied on the astrometry from the early third Gaia data release and radial velocities based on the spectroscopy from LAMOST and SDSS, which was also employed in spectral classification. The light variability of the confirmed CP2 stars in our sample was analyzed using data from the ZTF and ATLAS surveys. After filtering our initial sample using kinematic and spectroscopic criteria, we identified six CP2 stars with kinematical properties consistent with a halo membership. The orbits of these stars are in agreement with an origin in the Galactic disk where they were probably ejected through dynamical interactions or in the binary supernova scenario, making them the first runaway CP2 stars known.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Calculation of Velocities from Explosive Shot Test Fast-Frame Imagery

A cylinder of the experimental HE is detonated inside of an aquarium and sequential fast-frame images are taken of the resulting shock wave. We wish to use the sequence of images to extract quantitative data regarding the detonation velocity and velocity of the shock wave which can then be used to tune the parameters of the equation of state for the experimental HE material.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
NASA NTRS1996

Grant Proposal for the Continuation of the Voyager Interstellar Mission: LECP Investigation

This proposal documents the plans of the Low Energy Charged Particle (LECP) investigation team for participation in NASA's Voyager Interstellar Mission (VIM) as the Voyager 1 and 2 spacecraft explore the outer reaches of the heliosphere and search for the termination shock and the heliopause. The proposal covers the four year period from 1 January 1997 to 31 December 2000. The LECP instruments on Voyager 1 and 2 measure in situ intensities of charged particles with energies from about 30 keV to 100 MeV for ions, and about 20 keV to greater than 10 MeV for electrons. The instruments provide detailed spectral, angular, and compositional information about the particles. Composition is available for greater than 200 keV/nuc using multi-parameter measurements. Angular information is obtained by a mechanically scanned platform that rotates at various commanded rates. Measurements of low energy ion and electron intensities versus time and spatial location within the heliosphere contain an abundance of information regarding various transport and acceleration processes on both local (approx. 1 hr, approx. 0.01 AU) and global (approx. 11 yrs, approx. 100 AU) scales. The LECP instruments provide unique observations of such dynamical processes, and we anticipate that it will return critical information regarding the boundaries of the heliosphere. Several recent and exciting discoveries based on LECP measurements emphasize the important role that low energy charged particle distributions play in physical processes in the interplanetary medium. Yet, at the same time, these discoveries also underscore the fact that our understanding of processes in the outer heliosphere is, in most cases, incomplete, and in others, only rudimentary at best. Among the discoveries referred to above are the following: (1) Shocks: Examination of greater than 30 keV ion intensities have revealed: (a) a total absence of acceleration beyond only -100-200 keV at a strong transient shock in May 1991 at 35 AU, despite an enhanced level of seed particles; (b) a large transient shock in September 1991 of global scale, with intensities of shock-accelerated ions greater than or equal to 30 keV to approx. 30 MeV showing complex, highly energy-dependent spatial evolution, and small-scale (approx. few gyroradii), often anisotropic, micro-structures; (c) recurrent intensity increases in greater than or equal to 30 keV to -few MeV ions, with structures that, in some cases, show no correlation with the associated corotating shock. (2) Superthermal ion pressure: A global merged interaction region with a leading shock, downstream of which the superthermal ion (greater than or equal to 30 keV to approx. 4 MeV) pressure is comparable to that of the thermal plasma, and the total particle pressure yields a plasma beta of order unity. (3) Pickup ions: Measurements of the C/O ratio within transient structures at 35-45 AU showing the first clear evidence that transient shocks can pre-accelerate interstellar pickup ions from approx. 1 keV/nuc to at least 1 MeV/nuc. (4) Seed particles: Injection of ions for acceleration to high energies at the termination shock is unlikely to be a problem, since interplanetary transient and recurrent shocks are continually accelerating ions, of solar wind or interstellar origin, to highly superthermal energies. (5) Precursor electrons: Ambient solar electrons (greater than or equal to few tens of keV) that exist in the outer heliosphere ca form a broad precursor, several days wide, that is upstream of the termination shock and potentially observable a few months prior to the shock crossing. (6) Solar wind velocity at Voyager 1: We can use LECP ion data to obtain the solar wind velocity at Voyager 1, enabling us to provide critical measurement of the plasma flow as we approach and encounter the termination shock and other regions (necessary due to the partial failure of the Voyager 1 PLS experiment). The work of the LECP investigator team during the VIM will include: (1) Continuing operations with regard to the receipt, processing, verification, cataloging, display, and distribution of the data from the LECP instruments on Voyager 1 and 2, (2) Monitoring the health and performance of the LECP instruments, and evaluating and characterizing the response of the LECP instruments to various energetic particle and plasma environments, (3) Participating in, and supporting Voyager Project planning exercises and other coordinated activities relevant to exploration of the outer heliosphere, (4) Developing analysis techniques and operational procedures suitable for searching for and characterizing the boundaries and unique regions of the outher heliosphere, (5) Continuing the preparation of data sets appropriate for submission to the National Space Sciences Data Center (NSSDC) and, where appropriate, the Planetary Data System (PDS), (6) Maintaining direct Web access to online LECP data through the JHU/APL Voyager LECP home page, (7) Performing scientific evaluations of the Voyager 1 and 2 LECP data sets in conjunction with other data sets and other investigators, with particular focus on the outer regions of the heliosphere, and (8) Publishing the results of these evaluations in the scientific literature and presenting the results in scientific conferences.

Krimigis, Stamatios M.↗
DOE OSTI2020

Orbital Refinement and Stellar Properties for the HD 9446, HD 43691, and HD 179079 Planetary Systems

The Transit Ephemeris Refinement and Monitoring Survey is a project that aims to detect transits of intermediate-long period planets by refining orbital parameters of the known radial velocity planets using additional data from ground-based telescopes, calculating a revised transit ephemeris for the planet, then monitoring the planet host star during the predicted transit window. Here we present the results from three systems that had high probabilities of transiting planets: HD 9446 b and c, HD 43691 b, and HD 179079 b. We provide new radial velocity (RV) measurements that are then used to improve the orbital solution for the known planets. We search the RV data for indications of additional planets in orbit and find that HD 9446 shows a strong linear trend of 4.8σ. Using the newly refined planet orbital solutions, which include a new best-fit solution for the orbital period of HD 9446 c, and an improved transit ephemerides, we found no evidence of transiting planets in the photometry for each system. Transits of HD 9446 b can be ruled out completely and transits HD 9446 c and HD 43691 b can be ruled out for impact parameters up to b = 0.5778 and b = 0.898, respectively, due to gaps in the photometry. A transit of HD 179079 b cannot be ruled out, however, due to the relatively small size of this planet compared to the large star and thus low signal to noise. We determine properties of the three host stars through spectroscopic analysis and find through photometric analysis that HD 9446 exhibits periodic variability.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1986

Observations and a model of gravity-wave variability in the middle atmosphere

A major goal was to determine what portion of the gravity-wave frequency spectrum accounted for the majority of the momentum flux and divergence, as this has important implications for the middle atmosphere response. It was found that approx. 70% of the total flux and divergence was due to wave motions with observed periods less than 1 hour, consistent with expectations based on the shape of the observed gravity-wave spectrum (FrItts, 1984). This dominance of the momentum flux and divergence by high-frequency motions implies a potential for the modulation of those quantities by large-amplitude motions at lower frequencies. A second, striking aspect of the velocity and momentum flux data is its dramatic diurnal variability, particularly at certain levels. This variability is illustrated with the momentum flux, computed in 8-hr blocks. The dominant contributions here are due to waves with periods less than 1 hr. The variability with height and size of the mean square velocity in the west beam and the momentum flux, energed over the 3-day period. A detailed analysis of the various tidal motions present during this data interval was performed, and it was determined that variations in the zontal wind profile imposed by the diurnal tidal motion are probably responsible for the modulation of the gravity-wave amplitudes and momentum fluxes.

Fritts, D. C.↗
NASA NTRS1973

Velocity structure and evolution of the moon

Seismic data from the Apollo Passive Seismic Network stations are analyzed to determine the velocity structure and to infer the composition and physical properties of the lunar interior. Data from artificial impacts (SIBV booster and LM-ascent stage) cover a distance range of 9 to 1750 km. Travel times and amplitudes, as well as theoretical seismograms, are used to derive a velocity model for the outer 150 km of the moon. The P-wave velocity model confirms an earlier report of a lunar crust in the eastern part of Oceanus Procellarum. The crust is about 60 km thick and may consist of two layers in the mare regions. Possible values for the P-wave velocity in the uppermost mantle are between 7.6 and 9.0 km/sec. The 9 km/sec velocity represents either a localized heterogeneous unit, or a thin layer less than about 40 km in thickness. The elastic properties of the deep interior, as inferred from the seismograms of natural events (meteoroid impacts and moonquakes) occurring at great distances, indicate that there is an increase in attenuation and a possible decrease of velocity at depths below about 1000 km.

Toksoz, M. N.↗
DOE OSTI2024

Bayesian parameter estimation and evaluation of the K -ω shear stress transport model for plane impinging jets

Numerical simulations with semi-empirical turbulence models are commonly used to model impinging jets, often used for cooling solid surfaces. In this work, the constants in the k-ω shear stress transport model in ANSYS FLUENT are calibrated to experimental velocity and heat transfer data for a plane turbulent impinging air jet to determine if Kennedy-O'Hagan calibration (Kennedy and O'Hagan 2001 J. R. Stat. Soc. B 63 425–64) can improve predictions of near-surface velocities and surface Nusselt numbers for similar flows. Impinging jets have been proposed to cool the target plates of the divertor in future magnetic fusion energy reactors, where simulations are used to estimate divertor performance. The flat-plate divertor (Wang et al 2009 Fusion Sci. Technol .56 1023–7) uses a plane jet of helium issuing from a B = 0.5 mm slot to cool a surface with radius of curvature of 44 B at a distance 4 B from the slot. Predictions from the calibrated numerical model are compared with independent experimental data at different flow conditions, as well as surface temperature data for a flat plate divertor test section. The contribution of this work is evaluation of the accuracy of a calibrated turbulence model for modest extrapolations in flow geometry and flow conditions for a plane impinging jet.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS2004

Underexpanded Screeching Jets From Circular, Rectangular, and Elliptic Nozzles

The screech frequency and amplitude, the shock spacing, the hydrodynamic-acoustic standing wave spacing, and the convective velocity of large organized structures are measured in the nominal Mach number range of 1.1 less than or = Mj less that or = l0.9 for supersonic, underexpanded jets exhausting from a circular, a rectangular and an elliptic nozzle. This provides a carefully measured data set useful in comparing the importance of various physical parameters in the screech generation process. The hydrodynamic-acoustic standing wave is formed between the potential pressure field of large turbulent structures and the acoustic pressure field of the screech sound. It has been demonstrated earlier that in the currently available screech frequency prediction models replacement of the shock spacing by the standing wave spacing provides an exact expression. In view of this newly found evidence, a comparison is made between the average standing wavelength and the average shock spacing. It is found that there exists a small, yet important, difference, which is dependent on the azimuthal screech mode. For example, in the flapping modes of circular, rectangular, and elliptic jets, the standing wavelength is slightly longer than the shock spacing, while for the helical screech mode in a circular jet the opposite is true. This difference accounts for the departure of the existing models from predicting the exact screech frequency. Another important parameter, necessary in screech prediction, is the convective velocity of the large organized structures. It is demonstrated that the presence of the hydrodynamic-acoustic standing wave, even inside the jet shear layer, becomes a significant source of error in the convective velocity data obtained using the conventional methods. However, a new relationship, using the standing wavelength and screech frequency is shown to provide more accurate results.

Panda, J.↗
NASA NTRS1984

Outer satellite atmospheres: Their nature and planetary interactions

Significant insights regarding the nature and interactions of Io and the planetary magnetosphere were gained through modeling studies of the spatial morphology and brightness of the Io sodium cloud. East-west intensity asymmetries in Region A are consistent with an east-west electric field and the offset of the magnetic and planetary-spin axes. East-west orbital asymmetries and the absolute brightness of Region B suggest a low-velocity (3 km/sec) satellite source of 1 to 2 x 10(26) sodium atoms/sec. The time-varying spatial structure of the sodium directional features in near Region C provides direct evidence for a magnetospheric-wind-driven escape mechanism with a high-velocity (20 km/sec) source of 1 x 10(26) atoms/sec and a flux distribution enhanced at the equator relative to the poles. A model for the Io potassium cloud is presented and analysis of data suggests a low velocity source rate of 5 x 10(24) atoms/sec. To understand the role of Titan and non-Titan sources for H atoms in the Saturn system, the lifetime of hydrogen in the planetary magnetosphere was incorporated into the earlier Titan torus model of Smyth (1981) and its expected impact discussed. A particle trajectory model for cometary hydrogen is presented and applied to the Lyman-alpha distribution of Comet Kohoutek (1973XII).

Smyth, W. H.↗
NASA NTRS2000

GPS Sounding Rocket Development at NASA with Simultaneous Multi-Payload Tracking Application

An inverse differential GPS system has been developed for Sounding Rocket use which includes the flight unit and a ground station capable of extracting GPS data from sounding rocket telemetry, performing a real time differential solution and graphically displaying the rocket's path relative to a predicted trajectory plot. Accuracy has been proven to within less than 10 meters. Postprocessing has increased the precision to within 10 - 20 centimeters. The system has been successfully flown several times and delivered to the Sounding Program Office for routine field use. In addition to providing position, velocity and time GPS data has been used on sounding rockets for vehicle performance analysis, effecting a one hundred fold improvement in data time tagging, and steering an optical tracking device to intercept payloads launched from over the horizon. Precise velocity separation information and timing has been provided to multiple payload systems. Future plans include its use for Range Safety and enabling of interferometric techniques. The technology and software developed also has potential application to small satellite navigation and formation flying.

Bull, Barton↗
NASA NTRS1978

Intermediate-velocity gas in the local interstellar medium

Scans of interstellar ultraviolet absorption lines of N I, N II, and Si III for 17 stars are combined with previously published data for 30 stars. The extremal velocities at which detectable absorption occurs are tabulated, and it is shown that these are correlated for the three species. The data suggest that intermediate-velocity gas (20 to 60 km/s), best known from Na I and Ca II absorption, contains both neutral and ionized hydrogen. Features characteristic of intermediate-velocity isothermal shocks (greater than 60 km/s) are conspicuously rare. The intermediate-velocity gas may be in the form of clouds containing both H I and H II regions or of radiative shocks propagating in the interstellar medium; in the latter case the gas should be detectable in H-alpha emission.

Cowie, L. L.↗
DOE OSTI2020

Learning and meta-learning of stochastic advection–diffusion–reaction systems from sparse measurements

Physics-informed neural networks (PINNs) were recently proposed in [18] as an alternative way to solve partial differential equations (PDEs). A neural network (NN) represents the solution, while a PDE-induced NN is coupled to the solution NN, and all differential operators are treated using automatic differentiation. Here, we first employ the standard PINN and a stochastic version, sPINN, to solve forward and inverse problems governed by a non-linear advection–diffusion–reaction (ADR) equation, assuming we have some sparse measurements of the concentration field at random or pre-selected locations. Subsequently, we attempt to optimise the hyper-parameters of sPINN by using the Bayesian optimisation method (meta-learning) and compare the results with the empirically selected hyper-parameters of sPINN. In particular, for the first part in solving the inverse deterministic ADR, we assume that we only have a few high-fidelity measurements, whereas the rest of the data is of lower fidelity. Hence, the PINN is trained using a composite multi-fidelity network, first introduced in [12], that learns the correlations between the multi-fidelity data and predicts the unknown values of diffusivity, transport velocity and two reaction constants as well as the concentration field. For the stochastic ADR, we employ a Karhunen–Loève (KL) expansion to represent the stochastic diffusivity, and arbitrary polynomial chaos (aPC) to represent the stochastic solution. Correspondingly, we design multiple NNs to represent the mean of the solution and learn each aPC mode separately, whereas we employ a separate NN to represent the mean of diffusivity and another NN to learn all modes of the KL expansion. For the inverse problem, in addition to stochastic diffusivity and concentration fields, we also aim to obtain the (unknown) deterministic values of transport velocity and reaction constants. The available data correspond to 7spatial points for the diffusivity and 20 space–time points for the solution, both sampled 2000 times. We obtain good accuracy for the deterministic parameters of the order of 1–2% and excellent accuracy for the mean and variance of the stochastic fields, better than three digits of accuracy. In the second part, we consider the previous stochastic inverse problem, and we use Bayesian optimisation to find five hyper-parameters of sPINN, namely the width, depth and learning rate of two NNs for learning the modes. Here, we obtain much deeper and wider optimal NNs compared to the manual tuning, leading to even better accuracy, i.e., errors less than 1% for the deterministic values, and about an order of magnitude less for the stochastic fields.

97 MATHEMATICS AND COMPUTING↗