A Sharp Estimate on the Transient Time of Distributed Stochastic Gradient Descent
Not provided.
SEARCH · Search NASA
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.
Not provided.
On-site inspections (OSI) will be an important process to deter and help verify compliance with the Comprehensive Nuclear-Test-Ban Treaty (CTBT). An important tool in narrowing the possible locations to collect evidence of a nuclear test during an on-site inspection may be over-flights of the general area using aerial gamma spectroscopy which can measure the energy and intensity of gamma radiation and help identify areas that may warrant further investigation of areas of high concentrations of radioactivity. This paper will investigate the capabilities of gamma ray detectors that are typically used in aerial searches. Modeling and simulation results of the detector response for radionuclide species for an OSI will be presented for a variety of assumed releases, depositions on the ground, and times after a suspected Treaty violation for typical over flight heights and speeds. This data will provide information on the possible applicability for airborne spectroscopy and the challenges and limitations of this tool for OSI. Of particular interest will be analysis of the data for gross count, regions of interest, and isotope identification types of algorithms and the characteristics of each.
Abstract not provided
Abstract not provided.
Abstract not provided.
The potential for polarimetric Doppler radar measurements to improve predictions of ice microphysical processes within an idealized model–observational framework is examined. In an effort to more rigorously constrain ice growth processes (e.g., vapor deposition) with observations of natural clouds, a novel framework is developed to compare simulated and observed radar measurements, coupling a bulk adaptive-habit model of vapor growth to a polarimetric radar forward model. Bayesian inference on key microphysical model parameters is then used, via a Markov chain Monte Carlo sampler, to estimate the probability distribution of the model parameters. The statistical formalism of this method allows for robust estimates of the optimal parameter values, along with (non-Gaussian) estimates of their uncertainty. To demonstrate this framework, observations from Department of Energy radars in the Arctic during a case of pristine ice precipitation are used to constrain vapor deposition parameters in the adaptive habit model. The resulting parameter probability distributions provide physically plausible changes in ice particle density and aspect ratio during growth. A lack of direct constraint on the number concentration produces a range of possible mean particle sizes, with the mean size inversely correlated to number concentration. Consistency is found between the estimated inherent growth ratio and independent laboratory measurements, increasing confidence in the parameter PDFs and demonstrating the effectiveness of the radar measurements in constraining the parameters. Furthermore, the combined Doppler and polarimetric observations produce the highest-confidence estimates of the parameter PDFs, with the Doppler measurements providing a stronger constraint for this case.
Solar flares produce hard X-ray emission, the photon spectrum of which is often represented by a combination of thermal and power-law distributions. However, the estimates of the number and total energy of non-thermal electrons are sensitive to the determination of the power-law cutoff energy. Here, we revisit an 'above-the-loop' coronal source observed by RHESSI on 2007 December 31 and show that a kappa distribution model can also be used to fit its spectrum. Because the kappa distribution has a Maxwellian-like core in addition to a high-energy power-law tail, the emission measure and temperature of the instantaneous electrons can be derived without assuming the cutoff energy. Moreover, the non-thermal fractions of electron number/energy densities can be uniquely estimated because they are functions of only the power-law index. With the kappa distribution model, we estimated that the total electron density of the coronal source region was {approx}2.4 Multiplication-Sign 10{sup 10} cm{sup -3}. We also estimated without assuming the source volume that a moderate fraction ({approx}20%) of electrons in the source region was non-thermal and carried {approx}52% of the total electron energy. The temperature was 28 MK, and the power-law index {delta} of the electron density distribution was -4.3. These results are compared to the conventional power-law models with and without a thermal core component.
We have developed a borehole methodology to estimate formation thermal conductivity in situ with a spatial resolution of one meter. In parallel with a fiber-optic distributed temperature sensor (DTS), a resistance heater is deployed to create a controlled thermal perturbation. The transient thermal data is inverted to estimate the formation's thermal conductivity. We refer to this instrumentation as a Distributed Thermal Perturbation Sensor (DTPS), given the distributed nature of the DTS measurement technology. The DTPS was deployed in permafrost at the High Lake Project Site (67 degrees 22 minutes N, 110 degrees 50 minutes W), Nunavut, Canada. Based on DTPS data, a thermal conductivity profile was estimated along the length of a wellbore. Using the thermal conductivity profile, the baseline geothermal profile was then inverted to estimate a ground surface temperature history (GSTH) for the High Lake region. The GSTH exhibits a 100-year long warming trend, with a present-day ground surface temperature increase of 3.0 {+-} 0.8 C over the long-term average.
ABSTRACT We present clustering redshift measurements for Dark Energy Survey (DES) lens sample galaxies used in weak gravitational lensing and galaxy clustering studies. To perform these measurements, we cross-correlate with spectroscopic galaxies from the Baryon Acoustic Oscillation Survey (BOSS) and its extension, eBOSS. We validate our methodology in simulations, including a new technique to calibrate systematic errors that result from the galaxy clustering bias, and we find that our method is generally unbiased in calibrating the mean redshift. We apply our method to the data, and estimate the redshift distribution for 11 different photometrically selected bins. We find general agreement between clustering redshift and photometric redshift estimates, with differences on the inferred mean redshift found to be below |Δz| = 0.01 in most of the bins. We also test a method to calibrate a width parameter for redshift distributions, which we found necessary to use for some of our samples. Our typical uncertainties on the mean redshift ranged from 0.003 to 0.008, while our uncertainties on the width ranged from 4 to 9 per cent. We discuss how these results calibrate the photometric redshift distributions used in companion papers for DES Year 3 results.
The need to understand the risks and implications of traffic incidents involving hydrogen fuel cell electric vehicles in tunnels is increasing in importance with higher numbers of these vehicles being deployed. A risk analysis was performed to capture potential scenarios that could occur in the event of a crash and provide a quantitative calculation for the probability of each scenario occurring, with a qualitative categorization of possible consequences. The risk analysis was structured using an event sequence diagram with probability distributions on each event in the tree and random sampling was used to estimate resulting probability distributions for each end-state scenario. Here, the most likely consequence of a crash is no additional hazard from the hydrogen fuel (98.1–99.9% probability) beyond the existing hazards in a vehicle crash, although some factors need additional data and study to validate. These scenarios include minor crashes with no release or ignition of hydrogen. When the hydrogen does ignite, it is most likely a jet flame from the pressure relief device release due to a hydrocarbon fire (0.03–1.8% probability). This work represents a detailed assessment of the state-of-knowledge of the likelihood associated with various vehicle crash scenarios. This is used in an event sequence framework with uncertainty propagation to estimate uncertainty around the probability of each scenario occurring.
In the radiology department of the Mexican National Institute of Neurology and Neurosurgery, a dedicated institute in Mexico City, on average 19.3 computed tomography (CT) examinations are performed daily on hospitalized patients for neurological disease diagnosis, control scans and follow-up imaging. The purpose of this work was to estimate the effective dose received by hospitalized patients who underwent a diagnostic CT scan using typical effective dose values for all CT types and to obtain the estimated effective dose distributions received by surgical and non-surgical patients. Effective patient doses were estimated from values per study type reported in the applications guide provided by the scanner manufacturer. This retrospective study included all hospitalized patients who underwent a diagnostic CT scan between 1 January 2011 and 31 December 2012. A total of 8777 CT scans were performed in this two-year period. Simple brain scan was the CT type performed the most (74.3%) followed by contrasted brain scan (6.1%) and head angiotomography (5.7%). The average number of CT scans per patient was 2.83; the average effective dose per patient was 7.9 mSv; the mean estimated radiation dose was significantly higher for surgical (9.1 mSv) than non-surgical patients (6.0 mSv). Three percent of the patients had 10 or more brain CT scans and exceeded the organ radiation dose threshold set by the International Commission on Radiological Protection for deterministic effects of the eye-lens. Although radiation patient doses from CT scans were in general relatively low, 187 patients received a high effective dose (>20 mSv) and 3% might develop cataract from cumulative doses to the eye lens.
The liquid cloud microphysical properties for stratiform rain (SR) have been investigated in Part 1 of this series. Since convective rain (CR) has characteristics in raindrop size distribution (DSD) and precipitation that are distinct from SR, we investigate the CR properties in this study by using 20 hr of CR samples collected by 17 Automatic Parsivel Units disdrometers during the Midlatitude Continental Convective Clouds Experiment over the Atmospheric Radiation Measurement Southern Great Plains site. A full spectrum of DSD is constructed based on a total of 23 size channels (0.321 to 9.785 mm), and both Gamma and Exponential fitted functions are applied to extract the DSD shape parameters. Compared to SR properties, CR has distinct features including broader size range and narrower exponential slope parameter (λ E ). These results indicate that the assumption of constant exponential intercept parameter ( N 0E ) is inappropriate for CR rain rate estimates. Additionally, the subsetting scheme of the CR DSD spectra also has a strong impact on the Gamma/Exponential functions. Therefore, a new CR DSD parameterization scheme is developed by choosing the appropriate spectra subset with the constraint of rain rate and using the constant λ E instead of constant N 0E . With the input of the radar reflectivity at the lowest observed height, the newly calculated CR rain rates match well with the collocated surface rain gauge measurements (127 Mesonet stations and 17 Automatic Parsivel Units), while the rain rates calculated using traditional Z-R relationship are 3–4 times larger, indicating that constant λ E is a better assumption for CR DSD.
In numerous applications, surrogate models are used as a replacement for accurate parameter-to-observable mappings when solving large-scale inverse problems governed by partial differential equations (PDEs). The surrogate model may be a computationally cheaper alternative to the accurate parameter-to-observable mappings and/or may ignore additional unknowns or sources of uncertainty. The Bayesian approximation error (BAE) approach provides a means to account for the induced uncertainties and approximation errors, i.e. the errors between the accurate parameter-to-observable mapping and the surrogate. The statistics of these errors are, however, in general unknown a priori, and are thus calculated using Monte Carlo sampling. Although the sampling is typically carried out offline, i.e. before considering the data, the process can still represent a computational bottleneck. In this work, we develop a scalable computational approach for reducing the costs associated with the sampling stage of the BAE approach. Specifically, we consider the Taylor expansion of the accurate and surrogate forward models with respect to the uncertain parameter fields either as a control variate for variance reduction or as a means to directly and efficiently approximate the mean and covariance of the approximation errors. We propose efficient methods for evaluating the expressions for the mean and covariance of the Taylor approximations based on linear(-ized) PDE solves. Furthermore, the proposed approach is independent of the dimension of the uncertain parameter, depending instead on the intrinsic dimension of the data, ensuring scalability to high-dimensional problems. The potential benefits of the proposed approach are demonstrated for two high-dimensional inverse problems governed by PDE examples, namely for the estimation of a distributed Robin boundary coefficient in a linear diffusion problem, and for a coefficient estimation problem governed by a nonlinear diffusion problem.
The objectives are: (1) Estimate number and distribution of spring Chinook salmon Oncorhynchus tshawytscha redds and spawners in the John Day River subbasin; and (2) Estimate smolt-to-adult survival rates (SAR) and out-migrant abundance for spring Chinook and summer steelhead O. mykiss and life history characteristics of summer steelhead. The John Day River subbasin supports one of the last remaining intact wild populations of spring Chinook salmon and summer steelhead in the Columbia River Basin. These populations, however, remain depressed relative to historic levels. Between the completion of the life history and natural escapement study in 1984 and the start of this project in 1998, spring Chinook spawning surveys did not provide adequate information to assess age structure, progeny-to-parent production values, smolt-to-adult survival (SAR), or natural spawning escapement. Further, only very limited information is available for steelhead life history, escapement, and productivity measures in the John Day subbasin. Numerous habitat protection and rehabilitation projects to improve salmonid freshwater production and survival have also been implemented in the basin and are in need of effectiveness monitoring. While our monitoring efforts outlined here will not specifically measure the effectiveness of any particular project, they will provide much needed background information for developing context for project-specific effectiveness monitoring efforts. To meet the data needs as index stocks, to assess the long-term effectiveness of habitat projects, and to differentiate freshwater and ocean survival, sufficient annual estimates of spawner escapement, age structure, SAR, egg-to-smolt survival, smolt-per-redd ratio, and freshwater habitat use are essential. We have begun to meet this need through spawning ground surveys initiated for spring Chinook salmon in 1998 and smolt PIT-tagging efforts initiated in 1999. Additional sampling and analyses to meet these goals include an estimate of smolt abundance and SAR rates, and an updated measure of the freshwater distribution of critical life stages. Because Columbia Basin managers have identified the John Day subbasin spring Chinook population as an index population for assessing the effects of alternative future management actions on salmon stocks in the Columbia Basin (Schaller et al. 1999) we continue our ongoing studies. This project is high priority based on the high level of emphasis the NWPPC Fish and Wildlife Program, Subbasin Summaries, NMFS, and the Oregon Plan for Salmon and Watersheds have placed on monitoring and evaluation to provide the real-time data to guide restoration and adaptive management in the region. By implementing the proposed program we have been able to address many of the goals for population status monitoring, such as defining areas currently used by spring Chinook for holding and spawning habitats and determining range expansion or contraction of summer rearing and spawning populations. The BiOp describes these goals as defining population growth rates (adult monitoring), detecting changes in those growth rates or relative abundance in a reasonable time (adult/juvenile monitoring), estimating juvenile abundance and survival rates (juvenile/smolt monitoring), and identifying stage-specific survival (adult-to-smolt, smolt-to-adult).
Approaches to compute or estimate the output probability distributions from the quantum approximate optimization algorithm (QAOA) are needed to assess the likelihood it will obtain a quantum computational advantage. We analyze output from QAOA circuits solving 7200 random MaxCut instances, with $n$ = 14–23 qubits and depth parameter $p$ ≤ 12 and find that the average basis state probabilities follow approximate Boltzmann distributions: The average probabilities scale exponentially with their energy (cut value), with a peak at the optimal solution. Furthermore, we describe the rate of exponential scaling or effective temperature in terms of a series with a leading-order term $T$ ~ $C$ min /$n$ $\sqrt{p}$, with $C$ min the optimal solution energy. Using this scaling, we generate approximate output distributions with up to 38 qubits and find these give accurate accounts of important performance metrics in cases we can simulate exactly.
Purpose: In post-mastectomy radiation therapy (RT), skin dose must be accurately estimated to assess skin reactions such as erythema, desquamation and necrosis. Planning systems cannot always provide accurate dosimetry for target volumes distal to skin. Therefore, in-vivo dosimetry is necessary. A female anthropomorphic phantom was used with optically stimulated luminescence dosimeters (nanoDots) to measure dose to chest wall skin. In addition, EBT2 films was employed to measure dose to left lung and heart in post-mastectomy RT. Methods: Films and nanoDots were calibrated under full buildup conditions at 100cm SAD for 6MV photons. Five pieces of films were placed between slabs of Rando phantom to assess dose to left lung and heart. Two layers of 0.5cm thick bolus were used to cover the whole left chest. Six pairs of nanoDots were placed at medical and lateral aspects on the bolus surface, between the 0.5cm bolus layers, and under the bolus. Three control nanoDots were placed on chest wall to quantify imaging dose. The phantom was CT scanned with all dosimeters in place, and treatment planning was performed with tangential fields (200cGy). All dosimeters were contoured on CT and dose was extracted. NanoDots were read using nanoDot reader and films were scanned using film scanner. The measured and calculated doses were tabulated. Results: Dose to 12 nanoDots were evaluated. Dose variance for surface nanoDots were +3.8%, +2.7%, −5% and −9.8%. Those at lateral positions, with greater beam obliquity had larger variance than the medial positions. A similar trend was observed for other nanoDots (Table1). Point doses from films for heart and the left lung were 112.7cGy and 108.7cGy, with +10.2% and +9.04% deviation from calculated values, respectively. Conclusion: Dosimetry provided by the advanced planning system was verified using NanoDots and films. Both nanoDots and films provided good estimation of dose distribution in post-mastectomy RT.
The US Congress has passed legislation dictating that all government agencies establish a plan and process for improving energy efficiencies at their sites. In response to this legislation, Oak Ridge National Laboratory (ORNL) has recently conducted a pilot study to explore the deployment of a wireless sensor system for a real-time measurement-based energy efficiency optimization. With particular focus on the 12-mile long steam distribution network in our campus, we propose an integrated system-level approach to optimize energy delivery within the steam distribution system. Our approach leverages an integrated wireless sensor and real-time monitoring capability. We make real time state assessment on the steam trap health and steam flow estimate of the distribution system by mounting acoustic sensors on the steam pipes/traps/valves and observing measurements of these sensors with state estimators for system health. Our assessments are based on a spectral-based energy signature scheme that interprets acoustic vibration sensor data to estimate steam flow rates and assess steam traps status. Experimental results show that the energy signature scheme has the potential to identify different steam trap states and it has sufficient sensitivity to estimate flow rate. Moreover, results indicate a nearly quadratic relationship over the test region between the overall energy signature factor and flow rate in the pipe. We are able to present the steam flow and steam trap status, sensor readings, and the assessed alerts as an interactive overlay within a web-based Google Earth geographic platform that enables decision makers to take remedial action. The goal is to achieve significant energy-saving in steam lines by monitoring and acting on leaking steam pipes/traps/valves. We believe our demonstration serves as an instantiation of a platform that extends implementation to include newer modalities to manage water flow, sewage and energy consumption.
Thermal heterogeneities within energy conversion and storage, material processing, nuclear processes, aerospace, and military applications are often inaccessible to characterization by insertion sensors. When sensor deployment is possible, conventional pointwise temperature probes quickly degrade when inserted into harsh environments typical of such processes. We developed spatially-resolved ultrasonic thermometry to noninvasively measure the spatial distributions of thermal properties in such applications, even when sizable thermal gradients are present. Our method divides the path of ultrasonic propagation into segments bound by echogenic features, which create echoes in pulse-echo mode, encoding the information about interior temperature distributions. We use the acquired ultrasonic responses to estimate the internal temperature distributions by solving an inverse problem or concatenating segmental estimates. This work describes the implementation and industrial testing of the developed method at a coal-fired electrical power generation plant. We inserted an echogenically segmented Inconel 625 waveguide into the combustion zone of the utility-scale boiler and continuously acquired ultrasonic data while keeping sensitive components away from the damaging combustion environment. The accuracy of the time-dependent temperature distributions reconstructed from the ultrasonic measurements was comparable to that of thermocouples. The resiliency of ultrasonic thermometry to harsh combustion conditions was far superior to conventional insertion sensors. The measurements obtained during plant operation captured daily steam generation cycles in response to changing customer demand and intermittent contributions of renewable power sources to the power grid. These measurements have revealed new insights into the relationship between the dynamic power generation load and the conditions inside the steam generator. As a result, the successful industrial testing of spatially-resolved ultrasonic thermometry in solids indicates that the developed technology has matured to become an attractive alternative to conventional sensing in solving challenging problems of long-term thermal characterizations in extreme environments.