Search NASA⌕ Search

SEARCH · Search NASA

Results for “Methods:statistical”

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 1,549 records · Page 86
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

Uncertainty quantification tools for multiphase gas-solid flow simulations using MFIX

Computational fluid dynamics (CFD) has been widely studied and used in the scientific community and in the industry. Various models were proposed to solve problems in different areas. However, all models deviate from reality. Uncertainty quantification (UQ) process evaluates the overall uncertainties associated with the prediction of quantities of interest. In particular it studies the propagation of input uncertainties to the outputs of the models so that confidence intervals can be provided for the simulation results. In the present work, a non-intrusive quadrature-based uncertainty quantification (QBUQ) approach is proposed. The probability distribution function (PDF) of the system response can be then reconstructed using extended quadrature method of moments (EQMOM) and extended conditional quadrature method of moments (ECQMOM). The report first explains the theory of QBUQ approach, including methods to generate samples for problems with single or multiple uncertain input parameters, low order statistics, and required number of samples. Then methods for univariate PDF reconstruction (EQMOM) and multivariate PDF reconstruction (ECQMOM) are explained. The implementation of QBUQ approach into the open-source CFD code MFIX is discussed next. At last, QBUQ approach is demonstrated in several applications. The method is first applied to two examples: a developing flow in a channel with uncertain viscosity, and an oblique shock problem with uncertain upstream Mach number. The error in the prediction of the moment response is studied as a function of the number of samples, and the accuracy of the moments required to reconstruct the PDF of the system response is discussed. The QBUQ approach is then demonstrated by considering a bubbling fluidized bed as example application. The mean particle size is assumed to be the uncertain input parameter. The system is simulated with a standard two-fluid model with kinetic theory closures for the particulate phase implemented into MFIX. The effect of uncertainty on the disperse-phase volume fraction, on the phase velocities and on the pressure drop inside the fluidized bed are examined, and the reconstructed PDFs are provided for the three quantities studied. Then the approach is applied to a bubbling fluidized bed with two uncertain parameters, particle-particle and particle-wall restitution coefficients. Contour plots of the mean and standard deviation of solid volume fraction, solid phase velocities and gas pressure are provided. The PDFs of the response are reconstructed using EQMOM with appropriate kernel density functions. The simulation results are compared to experimental data provided by the 2013 NETL small-scale challenge problem. Lastly, the proposed procedure is demonstrated by considering a riser of a circulating fluidized bed as an example application. The mean particle size is considered to be the uncertain input parameter. Contour plots of the mean and standard deviation of solid volume fraction, solid phase velocities, and granular temperature are provided. Mean values and confidence intervals of the quantities of interest are compared to the experiment results. The univariate and bivariate PDF reconstructions of the system response are performed using EQMOM and ECQMOM.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2013

Large-Scale Optimization for Bayesian Inference in Complex Systems

The SAGUARO (Scalable Algorithms for Groundwater Uncertainty Analysis and Robust Optimization) Project focused on the development of scalable numerical algorithms for large-scale Bayesian inversion in complex systems that capitalize on advances in large-scale simulation-based optimization and inversion methods. The project was a collaborative effort among MIT, the University of Texas at Austin, Georgia Institute of Technology, and Sandia National Laboratories. The research was directed in three complementary areas: efficient approximations of the Hessian operator, reductions in complexity of forward simulations via stochastic spectral approximations and model reduction, and employing large-scale optimization concepts to accelerate sampling. The MIT₋Sandia component of the SAGUARO Project addressed the intractability of conventional sampling methods for large-scale statistical inverse problems by devising reduced-order models that are faithful to the full-order model over a wide range of parameter values; sampling then employs the reduced model rather than the full model, resulting in very large computational savings. Results indicate little effect on the computed posterior distribution. On the other hand, in the Texas₋Georgia Tech component of the project, we retain the full-order model, but exploit inverse problem structure (adjoint-based gradients and partial Hessian information of the parameter-to-observation map) to implicitly extract lower dimensional information on the posterior distribution; this greatly speeds up sampling methods, so that fewer sampling points are needed. We can think of these two approaches as "reduce then sample" and "sample then reduce." In fact, these two approaches are complementary, and can be used in conjunction with each other. Moreover, they both exploit deterministic inverse problem structure, in the form of adjoint-based gradient and Hessian information of the underlying parameter-to-observation map, to achieve their speedups.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2013

Nonlinear ion-acoustic structures in a nonextensive electron–positron–ion–dust plasma: Modulational instability and rogue waves

The nonlinear propagation of planar and nonplanar (cylindrical and spherical) ion-acoustic waves in an unmagnetized electron–positron–ion–dust plasma with two-electron temperature distributions is investigated in the context of the nonextensive statistics. Using the reductive perturbation method, a modified nonlinear Schrödinger equation is derived for the potential wave amplitude. The effects of plasma parameters on the modulational instability of ion-acoustic waves are discussed in detail for planar as well as for cylindrical and spherical geometries. In addition, for the planar case, we analyze how the plasma parameters influence the nonlinear structures of the first- and second-order ion-acoustic rogue waves within the modulational instability region. The present results may be helpful in providing a good fit between the theoretical analysis and real applications in future spatial observations and laboratory plasma experiments. -- Highlights: ► Modulational instability of ion-acoustic waves in a new plasma model is discussed. ► Tsallis’s statistics is considered in the model. ► The second-order ion-acoustic rogue wave is studied for the first time.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2018

Extracting Knowledge from Data through Catalysis Informatics

Catalysis informatics is a distinct sub-field that lies at the intersection of cheminformatics and materials informatics, but with unique challenges arising from the dynamic, surface-sensitive, multi-scale nature of heterogeneous catalysis. The ideas behind catalysis informatics can be traced back decades, but the field is only recently emerging due to advances in data infrastructure, statistics, machine learning, and computational methods. In this work we review the field from early works on expert systems and knowledge engines to more recent approaches utilizing machine-learning and uncertainty quantification. The data-information-knowledge hierarchy is introduced and used to classify various developments. The chemical master equation and micro-kinetic models are proposed as a quantitative representation of catalysis knowledge, which can be used to generate explanative and predictive hypotheses for the understanding and discovery of catalytic materials. We discuss future prospects for the field, including improved quantitative coupling of experiment/theory, advanced micro-kinetic models, and the development of open-source software tools. As a result, integration of existing chemical and physical models with emerging statistical and computational tools presents a promising route toward the automated design, discovery, and optimization of heterogeneous catalytic processes.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2015

Quantum renormalizations in anisotropic multisublattice magnets and the modification of magnetic susceptibility under irradiation

The dispersion equation of a strongly anisotropic one-dimensional magnet catena-[Fe{sup II}(ClO{sub 4}){sub 2}(Fe{sup III}(bpca){sub 2})]ClO{sub 4} containing alternating high-spin (HS) (S = 2) and low-spin (LS) (S = 1/2) iron ions is obtained by the diagram technique for Hubbard operators. The analysis of this equation yields six branches in the excitation spectrum of this magnet. It is important that the crystal field for ions with spin S = 2 is described by the Hamiltonian of single-ion easy-plane anisotropy, whose orientation is changed by 90° when passing from one HS iron ion to another. The U(N) transformation technique in the atomic representation is applied to diagonalize a single-ion Hamiltonian with a large number of levels. It is shown that the modulation of the orientation of easy magnetization planes leads to a model of a ferrimagnet with easy-axis anisotropy and to the formation of energy spectrum with a large gap. For HS iron ions, a decrease in the mean value of the spin projection due to quantum fluctuations is calculated. The analysis of the specific features of the spectrum of elementary excitations allows one to establish a correspondence to a generalized Ising model for which the magnetic susceptibility is calculated in a wide range of temperatures by the transfer-matrix method. The introduction of a statistical ensemble that takes into account the presence of chains of different lengths and the presence of iron ions with different spins allows one to describe the experimentally observed modification of the magnetic susceptibility of the magnet under optical irradiation.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2018

Statistical analysis of the chemical attribution signatures of 3-methylfentanyl and its methods of production

Chemical attribution of the origin of an illegal drug is a key component of forensic efforts aimed at combating illicit and clandestine manufacture of drugs and pharmaceuticals. The results of these studies yield detailed information on synthesis byproducts, reagents, and precursors that can be used to identify the method of manufacture. Here in the present work, chemical attribution signatures (CAS) associated with the synthesis of the analgesic 3-methylfentanyl, N-(3-methyl-1-phenethylpiperidin-4-yl)-N-phenylpropanamide, were investigated. Eighteen crude samples from six synthesis methods were generated, the analysis of which was used to identify signatures (i.e. chemical compounds) that were important in the discrimination of synthetic route. These methods were carefully selected to minimize the use of scheduled precursors, complicated laboratory equipment, number of steps, and extreme reaction conditions. Using gas and liquid chromatographies combined with time-of-flight mass spectrometry (GC-QTOF and LC-QTOF) over 160 distinct species were monitored. Lastly, analysis of this combined data set was performed using modern machine learning techniques capable of reducing the size of the data set, prioritizing key chemical attribution signatures, and identifying the method of production for blindly synthesized 3-methylfentanyl materials.

3-methylfentanyl↗
DOE OSTI2020

The completed SDSS-IV extended baryon oscillation spectroscopic survey: pairwise-inverse probability and angular correction for fibre collisions in clustering measurements

ABSTRACT The completed extended Baryon Oscillation Spectroscopic Survey (eBOSS) catalogues contain redshifts of 344 080 quasars at 0.8 < z < 2.2, 174 816 luminous red galaxies between 0.6 < z < 1.0, and 173 736 emission-line galaxies over 0.6 < z < 1.1 in order to constrain the expansion history of the Universe and the growth rate of structure through clustering measurements. Mechanical limitations of the fibre-fed spectrograph on the Sloan telescope prevent two fibres being placed closer than 62 arcsec in a single pass of the instrument. These ‘fibre collisions’ strongly correlate with the intrinsic clustering of targets and can bias measurements of the two-point correlation function resulting in a systematic error on the inferred values of the cosmological parameters. We combine the new techniques of pairwise-inverse probability and the angular upweighting (PIP+ANG) to correct the clustering measurements for the effect of fibre collisions. Using mock catalogues, we show that our corrections provide unbiased measurements, within data precision, of both the projected $\rm {\mathit{ w}_p}\left(\mathit{ r}_p\right)$ and the redshift-space multipole ξ(ℓ = 0, 2, 4)(s) correlation functions down to $0.1\, h^{-1}{\rm Mpc}$, regardless of the tracer type. We apply the corrections to the eBOSS DR16 catalogues. We find that, on scales $s\gtrsim 20\, h^{-1}{\rm Mpc}$ for ξℓ, as used to make baryon acoustic oscillation and large-scale redshift-space distortion measurements, approximate methods such as nearest-neighbour upweighting are sufficiently accurate given the statistical errors of the data. Using the PIP method, for the first time for a spectroscopic program of the Sloan Digital Sky Survey, we are able to successfully access the one-halo term in the clustering measurements down to $\sim 0.1\, h^{-1}{\rm Mpc}$ scales. Our results will therefore allow studies that use the small-scale clustering to strengthen the constraints on both cosmological parameters and the halo occupation distribution models.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

The sensitive surface chemistry of Co-free, Ni-rich layered oxides: identifying experimental conditions that influence characterization results

Recent studies have suggested that Co-free, Ni-rich layered cathodes (e.g., doped LiNiO2) can provide promising battery performance for practical applications. However, these layered cathodes suffer from significant surface instability during various stages of the sample history, which generates inherent challenges for achieving stable battery performance and obtaining statistically representative characterization results. To reliably report the surface chemistry of these materials, delicate controls of stepwise sample preparation are required. In this study, we aim to reveal how the surface chemistry of LiNiO2 based materials changes with various environments, including human exhalation, sample storage, sample preparation, electrochemistry cycling, and surface doping. Our results demonstrate that the surface of these materials is highly reactive and prone to alter at various stages of sample handling and characterization. The sensitive surface could impact the interpretation of the surface chemical and structural information, including surface carbonate formation, transition metal reduction and dissolution, and surface reconstruction. Importantly, the heterogeneity of the surface degradation calls for a consolidation of nanoscale, high-resolution characterization, and ensemble-averaged methods in order to improve statistical representation. Furthermore, the doping chemistry can effectively mitigate the surface degradation and improve overall battery performance due to the enhanced surface oxygen retention. Our study highlights the necessity of strict measurements through complementary characterizations at multiple length scales to eliminate unintentional biased conclusions.

surface chemistry, Co-free Ni-rich cathodes, Istab↗
DOE OSTI2021

FY21 Progress Report: SRNL Analysis of ICCWR LCM and WAMS data for Corrosion and Cracking

The development of algorithms for machine learning and data analysis for the 3013 Surveillance Program is a collaborative effort by the Savannah River National Laboratory (SRNL) and the University of South Carolina (USC). For corrosion detection, Laser Confocal Microscope (LCM) or Wide Area 3D Measurement System (WAMS) data is extracted from large binary files, with software written to convert the data to physical attributes (e.g., height, color and grayscale values; all as functions of a location in a plane projection). A user-friendly Matlab Graphical User Interface (GUI) that reads data from either LCM or WAMS files was developed to integrate input data with software developed for processing and evaluation. The GUI can selectively download binary data, interrogate data attributes, label data, flag significant features, execute Machine Learning (ML) algorithms, output parameters for trained ML algorithms, report ML model accuracy with respect to labeled data, and generate graphical representations for various analyses. Features can be called out by user-specified thresholds, manual labeling or machine learning algorithms when they have been completed. The ability to rapidly label data is important because of the volume of data required for training machine learning algorithms. The GUI has the flexibility to allow addition of improved ML algorithms, methods for data visualization, and statistical computations. Statistical analyses via the GUI include areas of pits within a defined range of pit depths, correlations between Red-Green-Blue (RGB) or grayscale intensity and relative surface height, covariances between values associated with features, and feature histograms. The development of supervised machine learning algorithms, however, has been hindered by a lack of training data. The machine learning algorithms for crack identification are being refined but require improvements to the true positive rate for crack detection. This shortcoming is an artifact of the limited training data currently available, perhaps more so than the structure of the neural networks. At present, the best results are had from a consensus over an ensemble of randomly generated Deep Neural Network (DNN) or Convolutional Neural Network (CNN) algorithms. Although the consensus accuracy method has yielded optimum true positive and true negative rates in excess of 80%, additional validation testing is necessary. In addition to the suite of LCM data that was initially used, and which represents the majority of the work presented in this report, WAMS image data was also reviewed at a preliminary level. The review included a comparison between image resolution and dynamic range for each method. WAMS (ZON file) image data was found to have a pixel pitch of 3.69μm compared to 1 μm for the LCM (vk4 file) data, which implies a lower resolution for the WAMS images. Conversely, the ratio of dynamic range of the WAMS data to the LCM data was approximately 41:20 for height data, suggesting that information from WAMS should more accurately determine the depth of pits. At present, the significance of the greater dynamic range of the WAMS data relative to the LCM data has not yet been evaluated.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2019

Regression for Assessing Contaminant Plume Trends

The Department of Energy’s (DOE) Hanford site has significant dissolved-phase groundwater contaminant plumes, for which remediation activities are ongoing. Assessing concentration trends is critical for making decisions regarding remedy implementation or performance. The SOCRATES single-page web application provides a quality-assured framework for data access and consistent data analytics using standard methods. To provide trend information for remedial decisions, statistical analyses based on ordinary least squares (OLS) are already supported by SOCRATES. The presence of censored data, however, can skew the result of an OLS regression. Other methods, such as the Akritas-Theil-Sen (ATS) or Tobit regressions can be more powerful. These two regression methods are used by Hanford site personnel to analyze groundwater contaminant concentration data, which often include nondetects (censored data). The ATS method uses comparisons of observation-pair ranks and the Tobit method uses a maximum likelihood estimation method. The Tobit method can also accommodate covariates, such as river stage or groundwater level fluctuations. Both ATS and Tobit regressions can be performed with the statistics software R, but to incorporate these methods into SOCRATES, their R code was ported into JavaScript. For quality assurance verification, the new code was tested with hexavalent chromium concentration data from the Hanford 100-K West Area and compared to regression analyses that were previously conducted on this data using R.

100-K Area↗
DOE OSTI2014

On the Molière theory of multiple scattering of charged particles (1947–1948) and its critique in subsequent years

Several “misconceptions” regarding the theory of multiple scattering of fast charged particles in matter developed by Molière in 1947–1948 and its application in the analysis of experimental results are discussed. It is shown that the critics of this theory misinterpreted the Molière method for determining the cross section of particle scattering by atoms with the screening of their nuclear fields by electron shells described by the Thomas-Fermi statistical model. If the original Molière method is applied consistently, the obtained scattering cross section generally agrees with the results of later classical calculations carried out by Lindhard and his collaborators and other authors.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Chemical Understanding of Actinide Separations

This work provides a broad overview of the separation processes used to isolate actinides and the experimentally and computationally determined chemical characteristics that define those separations. The redox chemistry of the actinides plays a pivotal role in both aqueous and pyrochemical processing separations. The near-overlapping energies of the 6d and 5f orbitals in the light actinides allow for facile adjustment of actinide oxidation states, which is used in many established separation methods. In contrast, the stable, generally 3+ oxidation states of the mid- and heavy actinides can make it difficult to separate them from the similarly lanthanides(III). In aqueous separations, the tendency of the actinides to form anionic and neutral aqueous complexes with a variety of complexants (especially soft donors) is used to achieve high separation factors between chemically similar elements in both solid–liquid separations and liquid–liquid extraction. This selectivity can be further tuned through the use of specialized organic or solid-phase ligands. Pyroprocessing separations utilize the unique redox behavior of the actinides to adjust their distribution between a molten salt electrolyte and either a solid electrode or molten metal phase. Atomic-level insights into the mechanisms underlying actinide separation processes, with the ultimate goal of predicting separation behavior, can be provided by electronic structure and statistical mechanical-based calculation methods.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

How electronic dynamics with Pauli exclusion produces Fermi-Dirac statistics

It is important that any dynamics method approaches the correct population distribution at long times. In this paper, we derive a one-body reduced density matrix dynamics for electrons in energetic contact with a bath. We obtain a remarkable equation of motion which shows that in order to reach equilibrium properly, rates of electron transitions depend on the density matrix. Even though the bath drives the electrons towards a Boltzmann distribution, hole blocking factors in our equation of motion cause the electronic populations to relax to a Fermi-Dirac distribution. These factors are an old concept, but we show how they can be derived with a combination of time-dependent perturbation theory and the extended normal ordering of Mukherjee and Kutzelnigg for a general electronic state. The resulting non-equilibrium kinetic equations generalize the usual Redfield theory to many-electron systems, while ensuring that the orbital occupations remain between zero and one. In numerical applications of our equations, we show that relaxation rates of molecules are not constant because of the blocking effect. Other applications to model atomic chains are also presented which highlight the importance of treating both dephasing and relaxation. Finally, we show how the bath localizes the electron density matrix.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2015

An improved numerical method to compute neutron/gamma deexcitation cascades starting from a high spin state

Numerous nuclear processes involve the deexcitation of a compound nucleus through the emission of several neutrons, gamma-rays and/or conversion electrons. The characteristics of such a deexcitation are commonly derived from a total statistical framework often called “Hauser–Feshbach” method. In this work, we highlight a numerical limitation of this kind of method in the case of the deexcitation of a high spin initial state. To circumvent this issue, an improved technique called the Fluctuating Structure Properties (FSP) method is presented. Two FSP algorithms are derived and benchmarked on the calculation of the total radiative width for a thermal neutron capture on 238 U. We compare the standard method with these FSP algorithms for the prediction of particle multiplicities in the deexcitation of a high spin level of 143 Ba. The gamma multiplicity turns out to be very sensitive to the numerical method. The bias between the two techniques can reach 1.5 γγ/cascade. Lastly, the uncertainty of these calculations coming from the lack of knowledge on nuclear structure is estimated via the FSP method.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2017

A Multiscale Bayesian Data Integration Approach for Mapping Air Dose Rates around the Fukushima Daiichi Nuclear Power Plant - 17166

This paper presents a multiscale data integration method to estimate the spatial distribution of air dose rates in the regional scale around the Fukushima Daiichi Nuclear Power Plant. We integrate various types of datasets, such as ground-based walk and car surveys, and airborne surveys, all of which have different scales, resolutions, spatial coverage, and accuracy. This method is based on geo-statistics to represent spatial heterogeneous structures, and also on Bayesian hierarchical models to integrate multiscale, multi-type datasets in a consistent manner. The Bayesian method allows us to quantify the uncertainty in the estimates, and to provide the confidence intervals that are critical for robust decision-making. Although this approach is primarily data-driven, it has great flexibility to include mechanistic models for representing radiation transport or other complex correlations. We demonstrate our approach using three types of datasets collected at the same time over Fukushima City, Japan: (1) coarse-resolution airborne surveys covering the entire area, (2) car surveys along major roads, and (3) walk surveys in multiple neighborhoods. Results show that the method can successfully integrate three types of datasets in a consistent manner, and create an integrated map (including the confidence intervals) of air dose rates over the domain in high resolution. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2023

Equibiaxial flexural strength determination of UO 2 using a ball-on-ring test

To increase nuclear fuel performance and reliability, their mechanical properties require an accurate and statistically relevant assessment. Biaxial flexural strength tests provide an alternative to bend bar techniques for assessing mechanical behavior; namely, the transverse rupture strength (TRS) of ceramic samples. Biaxial test samples require simple geometries and minimal surface preparation, reducing fabrication costs and handling hazards. This study investigated the TRS of polycrystalline UO 2 fuel forms at room temperature using a ball-on-ring test fixture. Here, pellets were fabricated from UO 2 powder using conventional powder processing and sintering techniques. The TRS and Weibull parameters were obtained through Weibull statistics on over 60 UO 2 samples tested under equibiaxial flexure. The larger sample size in this study enabled a more robust Weibull statistical analysis than alternative test methods, which may not capture the stochastic failure of typical ceramics. Furthermore, two different loading ball diameters were employed to assess the impact of contact damage on fracture strength. While Hertzian contact damage was observed with the smaller loading ball, the fracture strength remained unaffected. A fracture analysis of the tested UO 2 samples indicated a mixture of intergranular and transgranular fracture that transitions to transgranular fracture with increasing distance from the fracture origin. The characteristic strength of the combined data sets was determined to be 148 MPa, and the Weibull modulus was determined to be 9.1. The TRS values and Weibull parameters were close to values found in the literature for alternative testing techniques using samples with similar microstructure and density. The findings in this study validate the ball-on-ring method used to obtain the TRS of UO 2 with a sample geometry more representative of nuclear fuels. Additionally, experimental TRS results from this study can be implemented in modeling codes to predict fuel performance, which is critical to fuel burnup extension and advanced nuclear fuel technologies.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2011

Model-Independent Characterization of Charge Diffusion in Thick Fully Depleted CCDs

We present a new method to measure charge diffusion in charge-coupled devices (CCDs). The method is based on a statistical characterization of the shapes of charge clouds produced by low-energy X-rays using known properties of the two-dimensional Gaussian point-spread function (PSF). The algorithm produces reliable upper and lower bounds on the size of the PSF for photons converting near the entrance window of a device. It is optimally suited to the case of thick back-illuminated CCDs where the X-ray absorption length is smaller than the silicon thickness and the diffusion scale is of the same order as the pixel size. The only assumptions are that the charge cloud width is a monotonically increasing function of distance from the conversion point to the buried channel, and that the conversion probability is peaked at the surface. Otherwise, no physical models of carrier transport or knowledge of the electric field profile in the CCD are needed. In suboptimal conditions, the upper bound increases and the lower bound is unaffected, so confidence in the correctness of results is retained. The new method has been benchmarked against Monte Carlo simulations and tested on X-ray images measured on thick high-resistivity prototype CCDs for the Large Synoptic Survey Telescope. In Monte Carlo simulations of noiseless images having the optimal diffusion scale, the upper bound approximated the true PSF within 5%, increasing to 10% in simulations with added noise. Even with severely undersampled or truncated PSFs, the method brackets the true value to within 25%. Our method is accurate and computationally efficient and offers a fast and simple experimental setup.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

Domain Adaptive Graph Neural Networks for Constraining Cosmological Parameters Across Multiple Data Sets

Deep learning models have been shown to outperform methods that rely on summary statistics, like the power spectrum, in extracting information from complex cosmological data sets. However, due to differences in the subgrid physics implementation and numerical approximations across different simulation suites, models trained on data from one cosmological simulation show a drop in performance when tested on another. Similarly, models trained on any of the simulations would also likely experience a drop in performance when applied to observational data. Training on data from two different suites of the CAMELS hydrodynamic cosmological simulations, we examine the generalization capabilities of Domain Adaptive Graph Neural Networks (DA-GNNs). By utilizing GNNs, we capitalize on their capacity to capture structured scale-free cosmological information from galaxy distributions. Moreover, by including unsupervised domain adaptation via Maximum Mean Discrepancy (MMD), we enable our models to extract domain-invariant features. We demonstrate that DA-GNN achieves higher accuracy and robustness on cross-dataset tasks. Using data visualizations, we show the effects of domain adaptation on proper latent space data alignment. This shows that DA-GNNs are a promising method for extracting domain-independent cosmological information, a vital step toward robust deep learning for real cosmic survey data.

79 ASTRONOMY AND ASTROPHYSICS↗