Search NASA⌕ Search

SEARCH · Search NASA

Results for “log normal distribution”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2014

RE-EXAMINING SUNSPOT TILT ANGLE TO INCLUDE ANTI-HALE STATISTICS

Sunspot groups and bipolar magnetic regions (BMRs) serve as an observational diagnostic of the solar cycle. We use Debrecen Photohelographic Data (DPD) from 1974-2014 that determined sunspot tilt angles from daily white light observations, and data provided by Li and Ulrich that determined sunspot magnetic tilt angle using Mount Wilson magnetograms from 1974-2012. The magnetograms allowed for BMR tilt angles that were anti-Hale in configuration, so tilt values ranged from 0 to 360° rather than the more common ±90°. We explore the visual representation of magnetic tilt angles on a traditional butterfly diagram by plotting the mean area-weighted latitude of umbral activity in each bipolar sunspot group, including tilt information. The large scatter of tilt angles over the course of a single cycle and hemisphere prevents Joy's law from being visually identified in the tilt-butterfly diagram without further binning. The average latitude of anti-Hale regions does not differ from the average latitude of all regions in both hemispheres. The distribution of anti-Hale sunspot tilt angles are broadly distributed between 0 and 360° with a weak preference for east-west alignment 180° from their expected Joy's law angle. The anti-Hale sunspots display a log-normal size distribution similar to that of all sunspots, indicating no preferred size for anti-Hale sunspots. We report that 8.4% ± 0.8% of all bipolar sunspot regions are misclassified as Hale in traditional catalogs. This percentage is slightly higher for groups within 5° of the equator due to the misalignment of the magnetic and heliographic equators.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2016

TESTING THE PROPAGATING FLUCTUATIONS MODEL WITH A LONG, GLOBAL ACCRETION DISK SIMULATION

The broadband variability of many accreting systems displays characteristic structures; log-normal flux distributions, root-mean square (rms)-flux relations, and long inter-band lags. These characteristics are usually interpreted as inward propagating fluctuations of the mass accretion rate in an accretion disk driven by stochasticity of the angular momentum transport mechanism. We present the first analysis of propagating fluctuations in a long-duration, high-resolution, global three-dimensional magnetohydrodynamic (MHD) simulation of a geometrically thin ( h / r ≈ 0.1) accretion disk around a black hole. While the dynamical-timescale turbulent fluctuations in the Maxwell stresses are too rapid to drive radially coherent fluctuations in the accretion rate, we find that the low-frequency quasi-periodic dynamo action introduces low-frequency fluctuations in the Maxwell stresses, which then drive the propagating fluctuations. Examining both the mass accretion rate and emission proxies, we recover log-normality, linear rms-flux relations, and radial coherence that would produce inter-band lags. Hence, we successfully relate and connect the phenomenology of propagating fluctuations to modern MHD accretion disk theory.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2017

Lensing and the Warm-hot Intergalactic Medium

The correlation of weak lensing and Cosmic Microwave Anisotropy (CMB) data traces the pressure distribution of the hot, ionized gas and the underlying matter density field. The measured correlation is dominated by baryons residing in halos. Detecting the contribution from unbound gas by measuring the residual cross-correlation after masking all known halos requires a theoretical understanding of this correlation and its dependence with model parameters. Our model assumes that the gas in filaments is well described by a log-normal probability distribution function, with temperatures 10{sup 5–7} K and overdensities ξ ≤ 100. The lensing-Comptonization cross-correlation is dominated by gas with overdensities in the range ξ ≈ [3–33]; the signal is generated at redshifts z ≤ 1. If only 10% of the measured cross-correlation is due to unbound gas, then the most recent measurements set an upper limit of T-bar {sub e}≲10{sup 6} K on the mean temperature of the intergalactic Medium. The amplitude is proportional to the baryon fraction stored in filaments. The lensing-Comptonization power spectrum peaks at a different scale than the gas in halos, making it possible to distinguish both contributions. To trace the distribution of the low-density and low-temperature plasma on cosmological scales, the effect of halos will have to be subtracted from the data, requiring observations with larger signal-to-noise ratios than are currently available.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2019

Knock Intensity Distribution and a Stochastic Control Framework for Knock Control

One of the main factors limiting the efficiency of spark-ignited (SI) engines is the occurrence of engine knock. In high temperature and high pressure in-cylinder conditions, the fuel–air mixture auto-ignites creating pressure shock waves in the cylinder. Knock can significantly damage the engine and hinder its performance; as such, conservative knock control strategies are generally implemented which avoid such operating conditions at the cost of lower thermal efficiencies. Significant improvements in the performance of conventional knock controllers are possible if the properties of the knock process are better characterized and exploited in knock controller designs. One of the methods undertaken to better characterize knocking instances is to employ a probabilistic approach, in which the likelihood of knock is derived from the statistical distribution of knock intensity (KI). In this paper, it is shown that KI values at a fixed operating point for single fuel and dual fuel engines are accurately described using a mixed lognormal distribution. The fitting accuracy is compared against those for a randomly generated mixed-lognormally distributed dataset, and shown to exceed a 95% accuracy threshold for almost all of the operating points tested. Additionally, this paper discusses a stochastic knock control approach that leverages the mixed lognormal distribution to adjust spark timing based on KI measurements. Overall, this more informed knock control strategy would allow for improvements in engine performance and fuel efficiency by minimizing knock occurrences.

42 ENGINEERING↗
DOE OSTI2019

A probabilistic perspective on thermodynamic parameter uncertainties: Understanding aqueous speciation of mercury

Speciation plays an important role in determining the fate and transport of metals in terrestrial surface and subsurface systems. Equilibrium speciation modeling in aqueous systems relies on thermodynamic constants (log K values) of complexes, which are subject to uncertainties. In this study, using Monte Carlo (MC) simulations with Latin hypercube sampling (LHS) we systematically analyze the propagation of thermodynamic constant uncertainty through speciation modeling of an inorganic mercury-sulfide-chloride-water system. We find that seemingly small variances of the input log K normal distributions can lead to output species concentrations spanning multiple orders of magnitude, with highly skewed probability distributions. When equilibrium with mineral metacinnabar (β-HgS(s)) is neglected, the relative uncertainty of each output species is strongly positively correlated with the skewness of its concentration probability distribution, i.e., the lowest uncertainty occurs when the species concentration probability distribution is the most negatively skewed as its concentration approaches the total element concentration limit. The highest uncertainty in the identity of dominant species is located around species equivalence points. For cases where the mineral equilibrium is included, we derive analytical probability density functions for the log concentrations of all major species in the system. The mineral log K uncertainty is found to be an important contributor to all output concentration uncertainties. The analysis of combined effects of both pH and total sulfide concentration on output concentration uncertainties shows that high concentration uncertainties occur under highly sulfidic alkaline conditions and low uncertainties at low pH and sulfide concentrations. Finally, an analysis as presented here can distinguish between conditions that require a full uncertainty analysis and those for which the classical deterministic speciation modeling suffices.

58 GEOSCIENCES↗
DOE OSTI2021

The Epoch of Giant Planet Migration Planet Search Program. I. Near-infrared Radial Velocity Jitter of Young Sun-like Stars

We present early results from the Epoch of Giant Planet Migration program, a precise radial velocity (RV) survey of more than 100 intermediate-age (∼20–200 Myr) G and K dwarfs with the Habitable Zone Planet Finder spectrograph (HPF) at McDonald Observatory’s Hobby–Eberly Telescope. The goals of this program are to determine the timescale and dominant physical mechanism of giant planet migration interior to the water ice line of Sun-like stars. Here, we summarize results from the first 14 months of this program, with a focus on our custom RV pipeline for HPF, a measurement of the intrinsic near-infrared RV activity of young Solar analogs, and modeling the underlying population-level distribution of stellar jitter. We demonstrate on-sky stability at the sub-2 m s{sup −1} level for the K2 standard HD 3765 using a least-squares matching method to extract precise RVs. Based on a subsample of 29 stars with at least three RV measurements from our program, we find a median rms level of 34 m s{sup −1}. This is nearly a factor of 2 lower than the median rms level in the optical of 60 m s{sup −1} for a comparison sample with similar ages and spectral types as our targets. The observed near-infrared jitter measurements for this subsample are well reproduced with a log-normal parent distribution with μ = 4.15 and σ = 1.02. Finally, by compiling rms values from previous planet search programs, we show that near-infrared jitter for G and K dwarfs generally decays with age in a similar fashion to optical wavelengths, albeit with a shallower slope and lower overall values for ages ≲1 Gyr.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2011

Advancements in the Representation of Cloud-Aerosol Microphysics in the GEOS-5 AGCM

Despite numerous challenges, the physical parameterization of cloud-aerosol interactions in atmospheric GCMs has become a top priority for advancement because of our need to simulate and understand past, current, and future indirect effects of aerosols on clouds. The challenges stem from the involvement of wide range of cloud-scale dynamics and aerosol activation physical processes. Cloud dynamics modulate cloud areal extent and condensate, while aerosol activation depends on aerosol mass load, size distribution, internal mixing state, and nucleating properties, and ultimately determines cloud optical properties via particle sizes. Both macro- and micro-scale processes are obviously important for cloud-radiation interactions. We will present the main features of cloud microphysical properties in the GEOS- 5 Atmospheric GCM (AGCM) as simulated by the McRAS-AC (Microphysics of Clouds with Relaxed Arakawa-Schubert and Aerosol-Cloud interaction) scheme. McRAS-AC uses Fountoukis and Nenes (2005) aerosol activation for liquid clouds, and has an option for either Liu and Penner (2005) or Barahona and Nenes (2008, 2009) aerosol activation for ice clouds. Aerosol loading (on-line or climatological) comes from GOCART, with an assumed log-normal size distribution. Other features of McRAS-AC are level-by-level cloud-scale thermodynamics, and Seifert-Beheng (2001)-type precipitation microphysics, particularly from moist convection. Results from Single-Column Model simulations will be shown to demonstrate how cloud radiative properties, lifetimes, and precipitation are influenced by different parameterization assumptions. Corresponding fields from year-long simulations of the full AGCM will also be presented with geographical distributions of cloud effective particle sizes compared to satellite retrievals. While the primary emphasis will be on current climate, simulation results with perturbed aerosol loadings will also be shown to expose the radiative sensitivity of the microphysical parameterization.

Lee, D.↗
NASA NTRS1993

Size distributions in two porous chondritic micrometeorites

Quantitative size measurements of granular units (GUs), and nm-sized minerals in these units, in two porous chondritic micrometeorites are investigated. The matrix of these micrometeorites consist of loosely packed, 0.1 micron-sized, GUs. These objects were a major component of the solar nebula dust that accreted into protoplanets. The matrix in micrometeorite W7010*A2 has a fractal dimension with a small coefficient that supports efficient sticking of carbon-rich GUs during accretion. The fractal nature of the matrix provides a way to calculate the density using the aggregate size. The resulting very low density for porous chondritic micrometeorites is 0.08-0.14 g/cu cm, which supports the view that they are the solid debris from unconsolidated solar system bodies. Chondritic GUs contain ultrafine olivines, pyroxenes, and sulfides, embedded in hydrocarbons and amorphous carbons. Nanocrystals in the micrometeorites W7010*A2 and U2015*B show log normal size distributions. The high incidence of disk-shaped grains, a changeover from disk-shaped to euhedral grains, the unevolved nature of the size distributions, and multiple populations for grains less than 127 nm in size, are consistent with continuous postaccretion nucleation and growth in amorphous GUs, including coarsening via Ostwald ripening.

Rietmeijer, Frans J. M.↗
NASA NTRSDate not supplied

Impacts of aerosol size-dependent below-cloud scavenging on tropospheric aerosol in the NASA GEOS model

Cloud scavenging is the major sink for a suite of tropospheric aerosols, and thus largely affects the global distribution and lifetime of aerosols as well as the formation of clouds. Simplified model parameterizations of cloud scavenging of aerosols substantially contribute to large uncertainties in the simulated aerosol burdens, spatial distributions, lifetimes, as well as aerosol radiative forcing. Here we evaluate the wet scavenging scheme in the current NASA GEOS model, develop more physically-based parameterizations of below-cloud scavenging, and assess the impacts of the new scavenging parameterizations on the GEOS simulations of aerosols. The current below-cloud scavenging (BCS; washout) scheme uses a first-order removal with a constant scavenging coefficient. We replaced the current scheme with a size-dependent parameterization of Croft et al.(Atmos. Chem. Phys., 2009; hereafter referred to as C09). The C09 BCS parameterizations calculate aerosol below-cloud scavenging coefficients based on a size-dependent collision efficiency between aerosol particles and raindrops (or snow crystals). A look-up table of compiled collision efficiencies provides values for sixty aerosol sizes and nine rain flux rates. Snow-aerosol collision efficiencies are a function of the sixty aerosol sizes only. We assume a log-normal size distribution for aerosols (bulk) in GOCART. The collision efficiencies are then calculated online using a bilinear interpolation from the look-up table for the given aerosol sizes and precipitation flux rates in the model.

aerosols↗
NASA NTRSDate not supplied

Impacts of aerosol size-dependent below-cloud scavenging on tropospheric aerosol in the NASA GEOS model

Wet deposition is the major sink for a suite of tropospheric aerosols, and thus largely affects the global distribution and lifetime of aerosols as well as the formation of clouds in the atmosphere. In this study, we implement the aerosol size-dependent below-cloud scavenging (BCS; washout) parameterization of Croft et al. (Atmos. Chem. Phys., 2009; hereafter referred to as C09) in the GOCART bulk aerosol scheme coupled with the NASA Goddard Earth Observing System (GEOS) model, and evaluate the simulated tropospheric aerosols against surface, airborne, and satellite observations. The current below-cloud scavenging in GOCART employs a first-order removal function with a constant scavenging coefficient that is independent of aerosol sizes and precipitation fluxes. The C09 BCS parameterization instead calculates aerosol below-cloud scavenging coefficients based on a size-dependent collision efficiency between aerosol particles and raindrops (or snow crystals). A look-up table of compiled collision efficiencies provides scavenging coefficients for sixty aerosol sizes and nine rain flux rates. We determine instantaneous coefficients according to the C09 BCS look-up table, precipitation fluxes, and a log-normal size distribution of aerosols. The implemented BCS parameterization tends to result in a large diversity of scavenging coefficient values with higher magnitudes, leading to more efficient scavenging. The model-simulated global mean tropospheric Pb-210 aerosol lifetime decreases from 7.4 days to 5.8 days. The separate consideration of rain and snow conditions in the scavenging coefficient lookup table leads to a bi-mode distribution of scavenging coefficients. These changes also cause increased mass fractions of smaller aerosols compared to the current simulation. We examine the impacts of the C09 BCS parameterization on the simulated aerosol vertical profiles, tropospheric lifetimes, and deposition fluxes through comparisons with NASA aircraft measurements (ATom), surface observations of aerosol concentrations (IMPROVE) and wet deposition fluxes (NADP and APQMP), and MODIS aerosol optical depth (AOD) retrievals.

Aerosols↗
DOE OSTI2020

Visualization of Multifractal Superconductivity in a Two-Dimensional Transition Metal Dichalcogenide in the Weak-Disorder Regime

Eigenstate multifractality is a distinctive feature of noninteracting disordered metals close to a metal–insulator transition, whose properties are expected to extend to superconductivity. While multifractality in three dimensions (3D) only develops near the critical point for specific strong-disorder strengths, multifractality in 2D systems is expected to be observable even for weak disorder. Here we provide evidence for multifractal features in the superconducting state of an intrinsic, weakly disordered single-layer NbSe 2 by means of low-temperature scanning tunneling microscopy/spectroscopy. The superconducting gap, characterized by its width, depth, and coherence peaks’ amplitude, shows a characteristic spatial modulation coincident with the periodicity of the quasiparticle interference pattern. The strong spatial inhomogeneity of the superconducting gap width, proportional to the local order parameter in the weak-disorder regime, follows a log-normal statistical distribution as well as a power-law decay of the two-point correlation function, in agreement with our theoretical model. Furthermore, the experimental singularity spectrum f(α) shows anomalous scaling behavior typical from 2D weakly disordered systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2014

NuSTAR observations of the state transition of millisecond pulsar binary PSR J1023+0038

We report NuSTAR observations of the millisecond pulsar-low-mass X-ray binary (LMXB) transition system PSR J1023+0038 from 2013 June and October, before and after the formation of an accretion disk around the neutron star. Between June 10 and 12, a few days to two weeks before the radio disappearance of the pulsar, the 3-79 keV X-ray spectrum was well fit by a simple power law with a photon index of Γ=1.17{sub −0.07}{sup +0.08} (at 90% confidence) with a 3-79 keV luminosity of 7.4 ± 0.4 × 10{sup 32} erg s{sup –1}. Significant orbital modulation was observed with a modulation fraction of 36% ± 10%. During the October 19-21 observation, the spectrum is described by a softer power law (Γ=1.66{sub −0.05}{sup +0.06}) with an average luminosity of 5.8 ± 0.2 × 10{sup 33} erg s{sup –1} and a peak luminosity of ≈1.2 × 10{sup 34} erg s{sup –1} observed during a flare. No significant orbital modulation was detected. The spectral observations are consistent with previous and current multiwavelength observations and show the hard X-ray power law extending to 79 keV without a spectral break. Sharp-edged, flat-bottomed dips are observed with widths between 30 and 1000 s and ingress and egress timescales of 30-60 s. No change in hardness ratio was observed during the dips. Consecutive dip separations are log-normal in distribution with a typical separation of approximately 400 s. These dips are distinct from dipping activity observed in LMXBs. We compare and contrast these dips to observations of dips and state changes in the similar transition systems PSR J1824–2452I and XSS J1227.0–4859 and discuss possible interpretations based on the transitions in the inner disk.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2014

Most sub-arcsecond companions of Kepler exoplanet candidate host stars are gravitationally bound

Using the known detection limits for high-resolution imaging observations and the statistical properties of true binary and line-of-sight companions, we estimate the binary fraction of Kepler exoplanet host stars. Our speckle imaging programs at the WIYN 3.5 m and Gemini North 8.1 m telescopes have observed over 600 Kepler objects of interest and detected 49 stellar companions within ∼1 arcsec. Assuming binary stars follow a log-normal period distribution for an effective temperature range of 3000-10,000 K, then the model predicts that the vast majority of detected sub-arcsecond companions are long period (P > 50 yr), gravitationally bound companions. In comparing the model predictions to the number of real detections in both observational programs, we conclude that the overall binary fraction of host stars is similar to the 40%-50% rate observed for field stars.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2015

MODELING THE ANOMALY OF SURFACE NUMBER DENSITIES OF GALAXIES ON THE GALACTIC EXTINCTION MAP DUE TO THEIR FIR EMISSION CONTAMINATION

The most widely used Galactic extinction map is constructed assuming that the observed far-infrared (FIR) fluxes come entirely from Galactic dust. According to the earlier suggestion by Yahata et al., we consider how FIR emission of galaxies affects the SFD map. We first compute the surface number density of Sloan Digital Sky Survey (SDSS) DR7 galaxies as a function of the r-band extinction, A {sub r,} {sub SFD}. We confirm that the surface densities of those galaxies positively correlate with A {sub r,} {sub SFD} for A {sub r,} {sub SFD} < 0.1, as first discovered by Yahata et al. for SDSS DR4 galaxies. Next we construct an analytical model to compute the surface density of galaxies, taking into account the contamination of their FIR emission. We adopt a log-normal probability distribution for the ratio of 100 μm and r-band luminosities of each galaxy, y ≡ (νL){sub 100} {sub μm}/(νL) {sub r}. Then we search for the mean and rms values of y that fit the observed anomaly, using the analytical model. The required values to reproduce the anomaly are roughly consistent with those measured from the stacking analysis of SDSS galaxies. Due to the limitation of our statistical modeling, we are not yet able to remove the FIR contamination of galaxies from the extinction map. Nevertheless, the agreement with the model prediction suggests that the FIR emission of galaxies is mainly responsible for the observed anomaly. Whereas the corresponding systematic error in the Galactic extinction map is 0.1-1 mmag, it is directly correlated with galaxy clustering and thus needs to be carefully examined in precision cosmology.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2015

GREEN BANK TELESCOPE AND SWIFT X-RAY TELESCOPE OBSERVATIONS OF THE GALACTIC CENTER RADIO MAGNETAR SGR J1745–2900

We present results from eight months of Green Bank Telescope 8.7 GHz observations and nearly 18 months of Swift X-ray telescope observations of the radio magnetar SGR J1745–2900. We tracked the radio and X-ray flux density, polarization properties, profile evolution, rotation, and single-pulse behavior. We identified two main periods of activity. The first is characterized by approximately 5.5 months of relatively stable evolution in radio flux density, rotation, and profile shape, while in the second these properties varied substantially. Specifically, a third profile component emerged and the radio flux also became more variable. The single pulse properties also changed, most notably with a larger fraction of pulses with pulse widths ∼5–20 ms in the erratic state. Bright single pulses are well described by a log-normal energy distribution at low energies, but with an excess at high energies. The 2–10 keV flux decayed steadily since the initial X-ray outburst, while the radio flux remained stable to within ∼20% during the stable state. A joint pulsar timing analysis of the radio and X-ray data shows a level of timing noise unprecedented in a radio magnetar, though during the time covered by the radio data alone the timing noise was at a level similar to that observed in other radio magnetars. While SGR J1745–2900 is similar to other radio magnetars in many regards, it differs by having experienced a period of relative stability in the radio that now appears to have ended, while the X-ray properties evolved independently.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2017

Globular cluster seeding by primordial black hole population

Primordial black holes (PBHs) that form in the early Universe in the modified Affleck-Dine (AD) mechanism of baryogenesis should have intrinsic log-normal mass distribution of PBHs. We show that the parameters of this distribution adjusted to provide the required spatial density of massive seeds (≥ 10{sup 4} M {sub ⊙}) for early galaxy formation and not violating the dark matter density constraints, predict the existence of the population of intermediate-mass PBHs with a number density of 0∼ 100 Mpc{sup −3}. We argue that the population of intermediate-mass AD PBHs can also seed the formation of globular clusters in galaxies. In this scenario, each globular cluster should host an intermediate-mass black hole with a mass of a few thousand solar masses, and should not obligatorily be immersed in a massive dark matter halo.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Score-Based Physics-Informed Neural Networks for High-Dimensional Fokker–Planck Equations

The Fokker-Planck (FP) equation is a foundational partial differential equation (PDE) in stochastic processes involving Brownian motions. However, the curse of dimensionality (CoD) poses a formidable challenge when dealing with high-dimensional FP equations. Although Monte Carlo simulation and (vanilla) Physics-Informed Neural Networks (PINNs) have shown the potential to tackle CoD, both methods exhibit significant numerical errors in high dimensions when dealing with the probability density function (PDF) associated with Brownian motion. The point-wise PDF values tend to decrease exponentially as dimensionality increases, surpassing the precision of numerical simulations and resulting in substantial errors. In addition, due to its massive sampling, Monte Carlo fails to offer fast sampling. Modeling the logarithm likelihood (LL) via vanilla PINNs transforms the FP equation into a notoriously difficult Hamilton-Jacobi-Bellman (HJB) equation, which is impractical for PINN learning, whose error grows rapidly with dimension. To this end, we propose a novel approach utilizing a score-based solver to fit the score function in stochastic differential equations (SDEs). The score function, defined as the gradient of the LL, plays a fundamental role in inferring LL and PDF and enables fast SDE sampling, offering an effective means to overcome the CoD. Three fitting methods, Score Matching (SM), Sliced Score Matching (SSM), and Score-PINN, are introduced, each contributing unique advantages in computational complexity, accuracy, and generality. The proposed score-based SDE solver operates in two stages: first, employing score matching or Score-PINN to acquire the score function; and second, solving the LL via an ordinary differential equation (ODE) using the obtained score function. Comparative evaluations across these methods showcase varying trade-offs. The proposed methodology is evaluated across diverse SDEs, including anisotropic Ornstein-Uhlenbeck processes, geometric Brownian motion, and Brownian motion with varying eigenspace. We also test various distributions, including Gaussian, Log-normal, Laplace, and Cauchy distributions. The numerical results demonstrate the score-based SDE solver’s stability, speed, and performance across different experimental settings, solidifying its potential as a solution to CoD for high-dimensional FP equations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Anomaly Detection in Liquid Sodium Cold Trap Operation with Multisensory Data Fusion Using Long Short-Term Memory Autoencoder

Sodium-cooled fast reactors (SFR), which use high temperature fluid near ambient pressure as coolant, are one of the most promising types of GEN IV reactors. One of the unique challenges of SFR operation is purification of high temperature liquid sodium with a cold trap to prevent corrosion and obstructing small orifices. We have developed a deep learning long short-term memory (LSTM) autoencoder for continuous monitoring of a cold trap and detection of operational anomaly. Transient data were obtained from the Mechanisms Engineering Test Loop (METL) liquid sodium facility at Argonne National Laboratory. The cold trap purification at METL is monitored with 31 variables, which are sensors measuring fluid temperatures, pressures and flow rates, and controller signals. Loss-of-coolant type anomaly in the cold trap operation was generated by temporarily choking one of the blowers, which resulted in temperature and flow rate spikes. The input layer of the autoencoder consisted of all the variables involved in monitoring the cold trap. The LSTM autoencoder was trained on the data corresponding to cold trap startup and normal operation regime, with the loss function calculated as the mean absolute error (MAE). The loss during training was determined to follow log-normal density distribution. During monitoring, we investigated a performance of the LSTM autoencoder for different loss threshold values, set at a progressively increasing number of standard deviations from the mean. The anomaly signal in the data was gradually attenuated, while preserving the noise of the original time series, so that the signal-to-noise ratio (SNR) averaged across all sensors decreased below unity. Results demonstrate detection of anomalies with sensor-averaged SNR < 1.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗