Search NASA⌕ Search

SEARCH · Search NASA

Results for “spectral error”

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 307 records · Page 17
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Introductory guide to backgrounds in XPS spectra and their impact on determining peak intensities

Photoelectron and Auger peaks are central to most of the important uses of x-ray photoelectron spectroscopy (XPS), and thus, they receive the most attention in many types of analysis. Quantitative chemical analysis using XPS requires the assessment of the intensities of the photoemission peaks of the elements detected. Determination of peak intensities requires separation of the photoelectron peak signals from the background on which the peaks rest. For the determination of peak area intensities, the background is subtracted from overall signal intensity. The spectral background is also critical when peak fitting is used to determine intensities of overlapping peaks, and the model of background used in this process can impact the results. In addition to the impact on quantitative analysis, information about the depth distribution of elements in the near surface region can often be obtained by visual inspection of the background and quantified using appropriate modeling. Here, this introductory guide provides some basic information about backgrounds in the XPS analysis, describes the types of background models that are commonly used, suggests some of their strengths and weaknesses, and provides examples of their use and misuse. Although the fundamental nature of some components of the background signals in XPS is not understood, indicating that none of the models in use are fully correct and the area is subject to active research, appropriate good practices have been established for most routine analysis. The guide describes good practices, identifies errors that frequently appear in the literature, and uses examples to demonstrate the impacts of background selections on determinations of peak intensities.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Methods for Determining the Infrared Complex Refractive Indices n (λ) and k (λ) from Organic Solid Powders: Comparison of a Sucrose Single Crystal and a Pressed Pellet

There exists a need to measure optical constants n and k for many materials, especially organic solids in the infrared spectral region, but at present, there are only two established methods for making such measurements: single-angle (SA) reflectance and infrared spectroscopic ellipsometry (IRSE). The former derives the n/k vectors via measurement of the change in the amplitude of light reflected from the sample and then uses the Kramers–Kronig transform, while the latter measures the change in polarization of the reflected light from the sample to derive n/k. Here we make a direct comparison between the n/k vectors derived from SA measurements of a sucrose single crystal and the n/k values obtained after the crystal had been crushed, ground to powder, and pressed into a pellet form. The pellet was measured not only by SA but also via the IRSE method, and we found that pressing specular-quality pellets was critical, especially for the SA data. The n/k values of this typical organic molecule all agree to within 0.04 root-mean-square error, confirming that a pressed pellet can be used instead of a crystal to derive the intrinsic n/k vectors. In conclusion, this result is important for the many organic materials that, unlike sugar, do not have centimeter-sized crystals readily available.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

New Measurements of the Lyα Forest Continuum and Effective Optical Depth with LyCAN and DESI Y1 Data

Abstract We present the Ly α Continuum Analysis Network (LyCAN), a convolutional neural network that predicts the unabsorbed quasar continuum within the rest-frame wavelength range of 1040–1600 Å based on the red side of the Ly α emission line (1216–1600 Å). We developed synthetic spectra based on a Gaussian mixture model representation of nonnegative matrix factorization (NMF) coefficients. These coefficients were derived from high-resolution, low-redshift ( z < 0.2) Hubble Space Telescope/Cosmic Origins Spectrograph (COS) quasar spectra. We supplemented this COS-based synthetic sample with an equal number of DESI Year 5 mock spectra. LyCAN performs extremely well on testing sets, achieving a median error in the forest region of 1.5% on the DESI mock sample, 2.0% on the COS-based synthetic sample, and 4.1% on the original COS spectra. LyCAN outperforms principal component analysis (PCA) and NMF-based prediction methods using the same training set by 40% or more. We predict the intrinsic continua of 83,635 DESI Year 1 spectra in the redshift range of 2.1 ≤ z ≤ 4.2 and perform an absolute measurement of the evolution of the effective optical depth. This is the largest sample employed to measure the optical depth evolution to date. We fit a power law of the form τ ( z ) = τ 0 ( 1 + z ) γ to our measurements and find τ 0 = (2.46 ± 0.14) × 10 −3 and γ = 3.62 ± 0.04. Our results show particular agreement with high-resolution, ground-based observations around z = 2, indicating that LyCAN is able to predict the quasar continuum in the forest region with only spectral information outside the forest.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Rapid monitoring of fermentations: a feasibility study on biological 2,3-butanediol production

2,3-butanediol (2,3-BDO) is an economically important platform chemical that can be produced by the fermentation of sugars using an engineered strain of Zymomonas mobilis . These fermentations require continuous monitoring and modification of fermentation conditions to maximize 2,3-BDO yields and minimize the production of the undesired coproducts glycerol and acetoin. Because of the time required for sampling and off-line chromatographic measurement of fermentation samples, the ability of fermentation scientists to modify fermentation conditions in a timely manner is limited. The goal of this study was to test if near-infrared spectroscopy (NIRS) along with multivariate statistics could reduce the time needed for this analysis and enable real-time monitoring and control of the fermentation. In this work we developed partial least squares (PLS) calibration models to predict the concentrations of glucose, xylose, 2,3-BDO, acetoin, and glycerol in fermentations via NIRS using two different spectrometers and two different spectroscopy modalities. We first evaluated the feasibility of rapid NIRS monitoring through experiments where we measured the signals from each analyte of interest and built NIRS-based PLS models using spectra from synthetic samples containing uncorrelated concentrations of these analytes. All analytes showed unique spectral signatures, and this initial modeling showed that all analytes could be detected simultaneously. We then began work with samples from laboratory fermentation experiments and tested the feasibility of regression model development across two spectral collection modalities (at-line and on-line) and two instruments: a laboratory-grade instrument and a low-cost instrument with a more limited spectral range. All modalities showed promise in the ability to monitor Z. mobilis fermentations of glucose and xylose to 2,3-BDO. The low-cost instrument displayed a lower signal-to-noise ratio than the laboratory-grade instrument, which led to comparatively lower performance overall, but still provided sufficient accuracy to monitor fermentation trends. While the ease of use of on-line monitoring systems was favored as compared to at-line systems due to the lack of sampling required and potential for automated process control, we observed some decrease in performance due to the additional complexity of the sample matrix. We have demonstrated that NIRS combined with multivariate analysis can be used for at-line and on-line monitoring of the concentrations of glucose, xylose, 2,3-BDO, acetoin, and glycerol during Z. mobilis fermentations. The decrease in signal-to-noise ratio when using a low-cost spectrometer led to greater prediction error than the laboratory-grade spectrometer for at-line monitoring. The on-line monitoring modality showed great promise for real time process control via NIRS.

09 BIOMASS FUELS↗
DOE OSTI2022

Spectral evidence of solar neighborhood analogs in CALIFA galaxies

We introduce a novel nonparametric method to find solar neighborhood analogs (SNAs) in extragalactic integral field spectroscopic surveys. The main ansatz is that the physical properties of the solar neighborhood (SN) should be encoded in its optical stellar spectrum. We assume that our best estimate of such a spectrum is the one extracted from the analysis performed by the Code for Stellar properties Heuristic Assignment (CoSHA) from the MaStar stellar library. It follows that finding SNAs in other galaxies consist in matching, in a χ 2 sense, the SN reference spectrum across the optical extent of the observed galaxies. We applied this procedure to a selection of CALIFA galaxies, by requiring a close to face-on projection, relative isolation, and non-active galactic nucleus. We explore how the local and global properties of the SNAs (stellar age, metallicity, dust extinction, mass-to-light ratio, stellar surface mass density, star-formation density, and galactocentric distance) and their corresponding host galaxies (morphological type, total stellar mass, star-formation rate, and effective radius) compared with those of the SN and the Milky Way (MW). We find that SNAs are located preferentially in S(B)a–S(B)c galaxies, in a ring-like structure, which radii seem to scale with the galaxy size. Despite the known sources of systematics and errors, most properties present a considerable agreement with the literature on the SN. We conclude that the solar neighborhood is relatively common in our sample of SNAs. Our results warrant a systematic exploration of correlations among the physical properties of the SNAs and their host galaxies. We reckon that our method should inform current models of the galactic habitable zone in our MW and other galaxies.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

An alternative electrode design for accurate spectral induced polarization measurements in variably saturated porous media

SUMMARY Applications of spectral induced polarization (SIP) require electrodes that maintain hydrologic contact with surrounding soils to capture small electrical responses, often observed as phase shifts in milliradians. For unsaturated soils, electrodes must overcome increased electrical contact impedance due to reduced pore fluid. Traditional designs use a ceramic membrane electrode (CME) with a water reservoir and metal conductor, requiring periodic maintenance to retain electrolytic solution. For field applications where maintenance is impractical, alternative designs are needed. This study evaluated a new electrode design (silica flour electrode, SFE) alongside a CME. SFEs use packed silica flour to store water via capillary forces against a metal conductor. This study examined both designs in three variably saturated soils at soil suctions up to 700 mbar and soil water contents below 1 percent, with SIP measurements across 0.01 to 10 000 Hz frequencies. SFEs match CMEs at high frequencies and perform better at lower frequencies, without requiring ongoing maintenance, making them ideal for field use. In water-only experiments, CMEs produced errors and high noise below 1.5 Hz, whereas SFEs were more accurate. However, CMEs performed better above 300 Hz. In fine sand, SFEs performed better due to the relatively lower contact impedance as compared to CMEs. Both electrode types performed comparably in silty sand and silt loam soils, although CMEs required ongoing maintenance, suggesting potential for long-term reliability issues.

58 GEOSCIENCES↗
DOE OSTI2025

Catalight─An Open-Source Automated Photocatalytic Reactor Package Illustrated through Plasmonic Acetylene Hydrogenation

An open-source and modular Python package, Catalight, is developed and demonstrated to automate (photo)catalysis measurements. (Photo)catalysis experiments require studying several parameters to evaluate performance, including the temperature, gas flow rate and composition, illumination power, and spectral profile. Catalight orchestrates measurements over this complicated parameter space and systematically stores, analyzes, and visualizes the results. To showcase the capabilities of Catalight, we perform an automated apparent activation barrier measurement of acetylene hydrogenation over a plasmonic AuPd catalyst on an Al 2 O 3 support, simultaneously varying laser power, wavelength, and temperature in a multiday experiment controlled by a simple Python script. Our chemical results unexpectedly show an increased activation barrier upon light excitation, contrary to previous findings for other plasmonic reactions and catalysts. We show that the reaction rate order with respect to both acetylene and hydrogen remains unchanged upon illumination, suggesting that molecular surface coverage is not changed by light. By analyzing the inhomogeneity of the laser-induced heating, we attribute these results to a partial photothermal effect combined with a photochemical/hot electron-driven mechanism. In conclusion, our findings highlight the capabilities of a new experiment automation tool; explore the photocatalytic mechanism for an industrially relevant reaction; and identify systematic sources of error in canonical photocatalysis experimental procedures.

Catalysts↗
DOE OSTI2020

NEON AOP foliar trait maps, maps of model uncertainty estimates, and conifer map, East River, CO 2018

This data package contains mapped trait estimates and their uncertainties, and conifer map, for the National Ecological Observatory Network's Airborne Observation Platform survey data acquired over the Upper East River, Colorado in 2018. For full details, please see associated reference. in brief, trait models were developed independently for needle and non-needle leaf species using partial least squares regression (PLSR) using ground data from additional datasets: doi:10.15485/1618130, doi:10.15485/1618132, and doi:10.15485/1631278, merged with extracted spectral data from doi:10.15485/1618131. We separated vegetated pixels into needle and non-needle classes in order to generate a classification map based on the spectral differences between these leaf types (conifer.tif). We trained a deep learning model with custom architecture, detailed in Chadwick et al. In Press. The model performed with 0.998 true positive rate and 0.982 true negative rate, with ‘positives’ being non-needle identification. We then utilized PLSR to generate models of foliar traits for each leaf type. So that we could also map uncertainty in these predictions, we generated ten different models for needle and non-needle leaf species using different testing holdout sets of discrete sites. Each of these models was developed with a 100-fold cross validation procedure that utilized a 70% training set and 30% validation set with each fold, and then assessed based on the 10% of testing sites that were not included in that model’s development. The mean predicted value across the 10 models is used for the trait estimate in each pixel across the study area. The models are applied according to the leaf type designation in the conifer.tif map. The errors are the standard deviation across the 10 different models developed, with high error suggesting instability in model prediction and areas where values may not be reliable for ecological inference. These maps are only applied to areas with a NDVI > 0.5 to exclude non-vegetated areas. Shade masks could be applied to these data (doi:10.15485/1618131), but have not been for this data package. These data are also available on Google Earth Engine: https://code.earthengine.google.com/?asset=users/kdc/ER_NEON

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Leveraging Machine Learning Capabilities for the Characterization of Irradiated Uranium: A Case Study of Analysis Methods for Nuclear Safeguards and Nuclear Forensics

Nondestructively determining the initial enrichment of irradiated uranium is a complex and laborious multivariable problem due to the presence of fission products. This work demonstrates the capabilities of machine learning to analyze gamma-ray spectral data to determine initial enrichment without knowledge of the decay time of the sample. The approach developed is agnostic to the particular scenario and is applicable to a wide variety of applications in nuclear forensics and nuclear safeguards. We irradiated 5 mg uranium standard reference materials at discrete enrichment values ranging from 0.02% to 97% 235 U (weight percent) in UT Austin’s Nuclear Engineering Teaching Laboratory TRIGA Mark II 1.1 MW research reactor, allowed each to decay for 8 hours, and then measured each sample via gamma-ray spectrometry for 50 hours post-irradiation yielding 1,400 individual gamma-ray spectra discretized into 8,192 energy bins. We then trained decision trees models to analyze individual gamma-ray spectra and estimate the associated initial enrichment without knowledge of the time since end of irradiation. We evaluated the performance of the models with a reserved test set not used for training or calibrating the model. A decision tree model constructed with this procedure achieved a mean absolute error in initial enrichment determination of 2.3% (weight percent 235 U). Next, we implemented a principal component analysis pre-processing routine of the gamma-ray spectrometry data to reduce the dimensionality of the dataset from 8,192 channels in the spectrum to 10 principal components while retaining over 99% of the inherent variance in the data. Decision tree models constructed with these data demonstrated decreased mean absolute error in enrichment determination, reduced computational time, and decreased complexity. A single decision tree model constructed with this procedure achieved a mean absolute error in initial enrichment determination of 0.05% (weight percent 235 U). Furthermore, we analyzed these models with learning curves to ensure that overfitting did not occur. The capabilities provided by these models can be naturally extended to other application-focused measurements in the fields of nuclear safeguards, nuclear forensics, and nuclear non-proliferation.

Drescher, Adam↗
DOE OSTI2026

ACRRF High-Bay Dose Calculations using MCNP (Part A)

Analytical tools and models have been developed as a starting point for directly assessing dose in the Annular Core Research Reactor Facility (ACRRF) due to reactor operation. Key results include peak dose along the Central Cavity (CC) centerline (beamline) at the cavity level, dose throughout the High-Bay (HB), and dose on the facility roof for partially-shielded reactor operation where the 4” insert is removed from the CC Shield Plug (SP). Model results in the beamline are benchmarked against measured doses from passive dosimetry evaluations. Personnel total (neutron and gamma) dose in the ACRRF HB is calculated using Monte Carlo N-Particle (MCNP). Various CC and SP configurations are analyzed, including unshielded (no SP) and partially shielded (SP installed but 4” insert removed). Novel application of Variance Reduction (VR) techniques, namely the Surface Source Write (SSW) and Surface Source Read (SSR) capabilities in MCNP, enable impressive resolution (in a Monte Carlo modeling sense) of dose throughout much the facility. The VR techniques reduce stochastic error for challenging tallies, with more advanced techniques explored in the companion to this report (Part B) [1]. Supplementary studies (including a verification analysis) and pedagogic evaluations in Part B involve neutron spectra, angular distributions, and the dose impact of facility characteristics. With the SP 4” insert removed and the Lead-Boron (44”) Bucket (LB–44) in the reactor cavity, Total Effective Dose (TED) within the CC beamline is ≈140 rem per 300 MJ of reactor yield (or 3900 rem per hour at 100% Steady-State (SS) power). With no SP (unshielded) and a Free-Field (FF) cavity, TED within the beamline is ≈610 rem per 300 MJ (or 17000 rem per hour at 100% SS power). Due to the predicted collimation of radiation by the reactor pool (and partial SP, if present), beamline dose is much greater than the scattered radiation field surrounding the cavity and reactor tank. Comparisons are made to beamline dosimetry measurements to validate the model. Model predictions agree reasonably well (⪅10%) with measured quantities of neutron fluence, gamma fluence, and spectral metrics. Away from the beamline, comparisons made to previous dose measurements in the HB agreement within an order of magnitude.

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2020

On negative induced polarization in frequency domain measurements

SUMMARY Induced polarization (IP) has been widely used to non-invasively characterize electrical conduction and polarization in the subsurface resulting from an applied electric field. Earth materials exhibit a lossy capacitance defined by a negative intrinsic phase in frequency-domain IP (FDIP) or a positive intrinsic chargeability in time-domain IP (TDIP). However, error-free positive apparent phase or negative apparent chargeability (i.e. negative IP effects) can occur in IP measurements over heterogeneous media. While negative IP effects in TDIP data sets have been discussed, no studies have addressed this topic in detail for FDIP measurements. We describe theory and numerical modelling to explain the origin of negative IP effects in FDIP measurements. A positive apparent phase may occur when a relatively high polarizability feature falls into negative sensitivity zones of complex resistivity measurements. The polarity of the apparent phase is determined by the distribution of subsurface intrinsic phase and resistivity, with the resistivity impacting the apparent phase polarity via its control on the sensitivity distribution. A physical explanation for the occurrence of positive apparent phase data is provided by an electric circuit model representing a four-electrode measurement. We also show that the apparent phase polarity will be frequency dependent when resistivity changes significantly with frequency (i.e. in the presence of significant IP effects). Consequently, negative IP effects manifest themselves in the shape of apparent phase spectra recorded with multifrequency (spectral IP) data sets. Our results imply that positive apparent phase measurements should be anticipated and should be retained during inversion and interpretation of single frequency and spectral IP data sets.

Geochemistry & Geophysics↗
DOE OSTI2023

Residual estimation for grid modification in wall-modeled large eddy simulation using unstructured high-order methods

Here, the accuracy and computational cost of a large eddy simulation are highly dependent on the computational grid. Building optimal grids manually from a priori knowledge is not feasible in most practical use cases; instead, solution-adaptive strategies can provide a robust and cost-efficient method to generate a grid with the desired accuracy. We adapt the residual estimation algorithm developed by Toosi and Larsson for Discontinuous Galerkin Spectral Elements Methods (DGSEM) to guide the grid-adaptation process. The core of the method is the computation of the estimated modeling residual using the polynomial basis functions used in DGSEM, and the averaging of the estimated residual over each element. The final method is assessed in multiple channel flow test cases and for the transonic flow over an airfoil, in both cases making use of mortar interfaces between elements with hanging nodes. The method is found to be robust and reliable, and to provide solutions on grids with significantly fewer elements at comparable accuracy compared to when using human-generated grids.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Metrics for Intercomparison of Remapping Algorithms (MIRA) protocol applied to Earth system models

Abstract. Strongly coupled nonlinear phenomena such as those described by Earth system models (ESMs) are composed of multiple component models with independent mesh topologies and scalable numerical solvers. A common operation in ESMs is to remap or interpolate component solution fields defined on their computational mesh to another mesh with a different combinatorial structure and decomposition, e.g., from the atmosphere to the ocean, during the temporal integration of the coupled system. Several remapping schemes are currently in use or available for ESMs. However, a unified approach to compare the properties of these different schemes has not been attempted previously. We present a rigorous methodology for the evaluation and intercomparison of remapping methods through an independently implemented suite of metrics that measure the ability of a method to adhere to constraints such as grid independence, monotonicity, global conservation, and local extrema or feature preservation. A comprehensive set of numerical evaluations is conducted based on a progression of scalar fields from idealized and smooth to more general climate data with strong discontinuities and strict bounds. We examine four remapping algorithms with distinct design approaches, namely ESMF Regrid (Hill et al., 2004), TempestRemap (Ullrich and Taylor, 2015), generalized moving least squares (GMLS) (Trask and Kuberry, 2020) with post-processing filters, and WLS-ENOR (Li et al., 2020). By repeated iterative application of the high-order remapping methods to the test fields, we verify the accuracy of each scheme in terms of their observed convergence order for smooth data and determine the bounded error propagation using challenging, realistic field data on both uniform and regionally refined mesh cases. In addition to retaining high-order accuracy under idealized conditions, the methods also demonstrate robust remapping performance when dealing with non-smooth data. There is a failure to maintain monotonicity in the traditional L2-minimization approaches used in ESMF and TempestRemap, in contrast to stable recovery through nonlinear filters used in both meshless GMLS and hybrid mesh-based WLS-ENOR schemes. Local feature preservation analysis indicates that high-order methods perform better than low-order dissipative schemes for all test cases. The behavior of these remappers remains consistent when applied on regionally refined meshes, indicating mesh-invariant implementations. The MIRA intercomparison protocol proposed in this paper and the detailed comparison of the four algorithms demonstrate that the new schemes, namely GMLS and WLS-ENOR, are competitive compared to standard conservative minimization methods requiring computation of mesh intersections. The work presented in this paper provides a foundation that can be extended to include complex field definitions, realistic mesh topologies, and spectral element discretizations, thereby allowing for a more complete analysis of production-ready remapping packages.

58 GEOSCIENCES↗
DOE OSTI2020

An Algorithm to Retrieve Total Precipitable Water Vapor in the Atmosphere from FengYun 3D Medium Resolution Spectral Imager 2 (FY-3D MERSI-2) Data

The atmosphere has substantial effects on optical remote sensing imagery of the Earth&rsquo;s surface from space. These effects come through the functioning of atmospheric particles on the radiometric transfer from the Earth&rsquo;s surface through the atmosphere to the sensor in space. Precipitable water vapor (PWV), CO 2 , ozone, and aerosol in the atmosphere are very important among the particles through their functioning. This study presented an algorithm to retrieve total PWV from the Chinese second-generation polar-orbiting meteorological satellite FengYun 3D Medium Resolution Spectral Imager 2 (FY-3D MERSI-2) data, which have three near-infrared (NIR) water vapor absorbing channels, i.e., channel 16, 17, and 18. The algorithm was improved from the radiance ratio technique initially developed for Moderate-Resolution Imaging Spectroradiometer (MODIS) data. MODTRAN 5 was used to simulate the process of radiant transfer from the ground surfaces to the sensor at various atmospheric conditions for estimation of the coefficients of ratio technique, which was achieved through statistical regression analysis between the simulated radiance and transmittance values for FY-3D MERSI-2 NIR channels. The algorithm was then constructed as a linear combination of the three-water vapor absorbing channels of FY-3D MERSI-2. Measurements from two ground-based reference datasets were used to validate the algorithm: the sun photometer measurements of Aerosol Robotic Network (AERONET) and the microwave radiometer measurements of Energy&rsquo;s Atmospheric Radiation Measurement Program (ARMP). The validation results showed that the algorithm performs very well when compared with the ground-based reference datasets. The estimated PWV values come with root mean square error (RMSE) of 0.28 g/cm 2 for the ARMP and 0.26 g/cm 2 for the AERONET datasets, with bias of 0.072 g/cm 2 and 0.096 g/cm 2 for the two reference datasets, respectively. The accuracy of the proposed algorithm revealed a better consistency with ground-based reference datasets. Thus, the proposed algorithm could be used as an alternative to retrieve PWV from FY-3D MERSI-2 data for various remote sensing applications such as agricultural monitoring, climate change, hydrologic cycle, and so on at various regional and global scales.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Using active learning to improve quasar identification for the DESI spectra processing pipeline

The Dark Energy Spectroscopic Instrument (DESI) survey uses an automatic spectral classification pipeline to classify spectra. QuasarNET is a convolutional neural network used as part of this pipeline originally trained using data from the Baryon Oscillation Spectroscopic Survey (BOSS). In this paper we implement an active learning algorithm to optimally select spectra to use for training a new version of the QuasarNET weights file using only DESI data, with the goal of improving classification accuracy. This active learning algorithm includes a novel outlier rejection step using a Self-Organizing Map to ensure we label spectra representative of the larger quasar sample observed in DESI. We perform two iterations of the active learning pipeline, assembling a final dataset of 5600 labeled spectra, a small subset of the approximately 1.3 million quasar targets in DESI's Data Release 1. When splitting the spectra into training and validation subsets we achieve similar performance to the previously trained weights file in completeness and purity calculated on the validation dataset but do so with less than one tenth of the amount of training data. The new weights also more consistently classify objects in the same way when used on unlabeled data compared to the old weights file. In the process of improving QuasarNET's classification accuracy we discovered a systemic error in QuasarNET's redshift estimation and used our findings to improve our understanding of QuasarNET's redshifts.

Machine learning↗
DOE OSTI2022

Neural-network learning of SPOD latent dynamics

Here, we aim to reconstruct the latent space dynamics of high dimensional, quasi-stationary systems using model order reduction via the spectral proper orthogonal decomposition (SPOD). The proposed method is based on three fundamental steps: in the first, once that the mean flow field has been subtracted from the realizations (also referred to as snapshots), we compress the data from a high-dimensional representation to a lower dimensional one by constructing the SPOD latent space; in the second, we build the time-dependent coefficients by projecting the snapshots containing the fluctuations onto the SPOD basis and we learn their evolution in time with the aid of recurrent neural networks; in the third, we reconstruct the high-dimensional data from the learnt lower -dimensional representation. The proposed method is demonstrated on two different test cases, namely, a compressible jet flow, and a geophysical problem known as the Madden-Julian Oscillation. An extensive comparison between SPOD and the equivalent POD-based counterpart is provided and differences between the two approaches are highlighted. The numerical results suggest that the proposed model is able to provide low rank predictions of complex statistically stationary data and to provide insights into the evolution of phenomena characterized by specific range of frequencies. The comparison between POD and SPOD surrogate strategies highlights the need for further work on the characterization of the interplay of error between data reduction techniques and neural network forecasts.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Application of machine learning to estimate fireball characteristics and their uncertainty from infrared spectral data

Experiments or events involving high explosives (HE) can be monitored remotely by infrared (IR) sensors to gather information about the configuration or materials involved in the device. Researchers at the Air Force Institute of Technology (AFIT) developed a phenomenological model for HE fireball spectra in the IR range that allows for parameters to be extracted from Fourier transform infrared (FTIR) data. This model includes parameters tied to physical characteristics of the fireball: temperature, size, soot, and gas concentrations. Previous works have sought to recover these parameters by the fitting of either whole spectra or select wavenumber bands to this phenomenological model. Difficulties arise due to the complex relationships between the parameters to be fit. Uncertainty quantification of the estimated fireball parameters is also problematic since HE experiments do not have any ground truth information on the parameters. It is suggested that artificial neural network (ANN) based approaches may be well suited to this problem, because of their ability to capture complex and highly nonlinear relationships. As such, this work seeks to explore the efficacy of deep artificial neural networks (DNNs) for this problem of parameter recovery from spectra and to also investigate the uncertainty of recovering the fireball parameters from FTIR data. Networks are designed using the hyperparameter optimization tool Hyperopt and trained/tested on artificial data generated using the phenomenological model developed by AFIT. The results of applying the network to the artificial data set are compared to a physics-based band approach that uses a selected number of bands based on their physical properties. Information on the uncertainty of estimating parameters from remotely sensed experimental data is obtained by treating the accuracy of the DNN model on artificial data as an upper bound and by examining the impact of emissivity due to soot on parameter estimation error; the results for artificial data are likely to be optimistic as compared to recovering parameters from experimental data.

42 ENGINEERING↗
DOE OSTI2024

Fe Kα XANES, Fe Kβ HERFD XANES and EPMA flank method determinations of the oxidation state of Fe in garnet

The ferric to total iron ratios (Fe 3+ /ΣFe) of garnets can be paired with thermodynamic mineral activity models to quantify the oxygen fugacity of garnet-bearing rocks. However, techniques with a high analytical and spatial resolution are necessary to distinguish differences in garnet Fe 3+ /ΣFe ratios at the percent level and to accurately measure garnets that are zoned or contain inclusions. We acquired conventional Fe Kα and high-resolution energy fluorescence detection (HERFD) Fe Kβ X-ray absorption near edge structure (XANES) spectra and electron microprobe flank method analyses on a suite of 27 peridotitic and eclogitic garnets with Fe 3+ /ΣFe ratios previously determined by Mössbauer spectroscopy to evaluate the precision of each technique. We examined variations in the energy and intensity of three XANES spectral features as a function of Fe 3+ /ΣFe ratios: 1) the intensity ratio of two-post edge features (I-ratio; Fe Kα only); 2) the energy of the Fe edge at 90% normalized intensity (E 0.9 ; Fe Kα only) and 3) the pre-edge centroid energy (Fe Kα and HERFD Fe Kβ). In accordance with previous work, we find the energies of garnet pre-edge centroids are relatively insensitive to Fe 3+ /ΣFe ratios. The I-ratios of peridotitic and eclogitic garnets are offset from each other at low Fe 3+ /ΣFe ratios (≤0.13); I-ratio garnet XANES calibrations are composition-specific. The E 0.9 feature is independent of garnet major element composition in spectra that have been corrected for the effects of self-absorption. We produce two Fe Kα garnet XANES calibrations based on variations in the E0.9 feature; one calibration with all garnet reference materials included (Fe 3+ /ΣFe up to 1.0; “all garnet calibration”) and another calibration specific to garnets with low Fe 3+ /ΣFe ratios (“low ferric calibration”). Fe 3+ /ΣFe ratios calculated from the mean of up to 25 flank method measurements on eight garnet reference materials fall within 4% absolute of a one-to-one correlation with Fe 3+ /ΣFe ratios measured by Mössbauer. The standard error of the mean Fe 3+ /ΣFe ratio calculated from flank method approaches the Mössbauer-determined Fe 3+ /ΣFe ratio within estimated error (3%) after three analyses. Flank method precision is enhanced at higher beam current; however, the precision of the flank method does not approach the precision of XANES under any microprobe analytical condition tested here. Garnet reference materials detailed here are available by request to the Smithsonian Institution.

58 GEOSCIENCES↗