Search NASA⌕ Search

SEARCH · Search NASA

Results for “Sample Handling”

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 667 records · Page 37
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

AI-Batt (Autonomous Identification of Battery Life Models) [SWR 21-36]

Autonomous Identification of Battery Life Models (AI-Batt) AI-Batt is a MATLAB code base for developing lifetime models for batteries from accelerated aging data. The code base provides many functions for processing, visualizing, and modeling battery aging data, making the data processing, exploration, and modeling workflow substantially faster. These tools are tailored for working with battery aging data sets, which usually consist of many separate time-series for each cell, with many test conditions and possible replicates at each condition, which makes it difficult to simply process or visualize the data set. Complex modeling tasks, such as cross-validation, sensitivity analysis, and uncertainty quantification have been implemented to enable thorough statistical investigation of model predictions. Additionally, several machine-learning algorithms are implemented to autonomously identify suitable models via symbolic regression. Data processing functions automatically cast data from the struct data type, which is commonly used to store experimental data, but is not an acceptable input for most algorithms, to the table data type, which can be easily used as input to any optimization algorithm. Also, the data can be separated into time-invariant and time-variant data tables, which is helpful for exploring the data set as well as developing separate models for time-variant and time-invariant aging mechanisms. For example, in aging tests with constant temperature, temperature is a time-invariant experimental condition. Visualization tools enable plotting of data, model fits, and model simulations possible with single-line function calls, empowering data exploration of complex data sets with both time-varying and time-invariant trends. Plots can be automatically generated for the whole data set, or separated by data group (groups of test replicates) or individual data series. Data points or data series can be automatically colored by the value of a variable with a variety of color maps, and model predictions can also be colored by the value of a fit statistic. Comparisons between data sets and the predictions/simulations of different models on the same data set can be easily plotted as well. Distributions of parameter values from bootstrap resampling can be plotted to visualize the reliability of parameter estimation, or determine any correlations between parameters. Modeling tools handle the complex task of creating and parsing symbolic equations for modeling battery lifetime. Equations are parsed to grab relevant data variables, parameter values, or specified sub-models for input into optimization, evaluation, or simulation functions. Models can be optimized locally (one set of parameters for each data series), bi-level (some parameters shared across the data set), or globally (single set of parameters for all data). Functions implementing symbolic regression algorithms help users to discover effective model equations, even in poorly sampled, high-dimensional data.

Smith, Kandler [National Renewable Energy Lab. (NR↗
DOE OSTI2026

Filtration Performance of Simulated 200 West Area Waste Feeds

This report describes the scaled experimental system and approach used to examine dead-end filtration performance of representative 200W waste feeds. The scaled system, which was originally designed and assembled to test Tank Side Cesium Removal (TSCR) system performance with higher-than-expected solid loadings in 2021 (Schonewill et al. 2021), was repurposed to conduct the current experiments at ~1/145 of full scale (based on throughput). Six experimental runs were conducted with five different 200W waste feed simulants: three using a DEF module scaled for TSCR and three using a DEF module scaled for the 200W process modules (based on the current design for the Advanced Modular Pretreatment System). Each experiment was run continuously for multiple days with an operating approach prototypic of the full-scale system. Staff performing the experimental runs monitored performance, obtained data from calibrated process instruments, and collected samples for observation and analysis. The measured data are presented with a focus on assessing DEF performance – specifically, the filters’ differential pressure response to the five waste simulants, frequency and efficacy of backwashing, and baseline recovery between experimental runs; data related to ion exchange column performance are also discussed in cases where the opportunity arose. The experimental campaign demonstrated that the DEFs satisfied their primary function of protecting the ion exchange column from solid intrusion for all the representative simulants used. The filters readily handled solids loadings of =500 ppm (and even greater), especially the modules scaled to the 200W process modules. Adjustments to the processing flow rate and reductions in feed temperature were observed to affect the rate of differential pressure increase on the filters, but neither adversely affected the ability of the DEFs to perform their primary function. Backflushing reliably recovered filter performance in all runs, although it did not prevent irreversible fouling for one simulant. The run that exhibited irreversible fouling established that both the quantity and the nature of the solids being filtered need to be considered when projecting filter performance.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
NASA NTRS1987

Some aspects of space propulsion with extraterrestrial resources

Extraterrestrial resources for space processing of chemicals, in general, and propellants, in particular, are explored quantitatively. It is seen that, for several candidate space mission scenarios, space processing of both space resources and earth-carried resources can make decisive differences in the mission success for a given payload. To fix ideas and demonstrate trends, the specific case of water splitting to extract oxygen, discard (or use without storage) the resulting hydrogen, and burn earth-carried noncryogenic liquid fuel(s) in a simple rocket motor, designed for periodic thrusting, is treated in some detail. Experimental hardware is assembled and demonstrated to perform adequately, besides showing compactness of the space-packaged 'capsule' module that is self-contained. Building upon previous studies, the concept of in situ propellant production (ISPP) is reexamined in light of more recent energy and materials technologies. Missions to comets and Mars Sample Return are mentioned as candidate scenarios. The mission duration, reliability-repairability of hardware, resource availability in low earth orbit (LEO), and the thrust requirements are considered in turn. It is seen that space storage of hydrogen for extended durations (5-10 years) involves problems that require detailed studies, besides involving many presently unanswered issues. A study of the energy option in LEO and in deep space is developed in simple terms. The different solar, radioisotope, and nuclear power sources are mentioned. Storage and handling of raw and processed chemicals are considered.

Ramohalli, Kumar↗
NASA NTRS2013

Sample Analysis at Mars Instrument Simulator

The Sample Analysis at Mars Instrument Simulator (SAMSIM) is a numerical model dedicated to plan and validate operations of the Sample Analysis at Mars (SAM) instrument on the surface of Mars. The SAM instrument suite, currently operating on the Mars Science Laboratory (MSL), is an analytical laboratory designed to investigate the chemical and isotopic composition of the atmosphere and volatiles extracted from solid samples. SAMSIM was developed using Matlab and Simulink libraries of MathWorks Inc. to provide MSL mission planners with accurate predictions of the instrument electrical, thermal, mechanical, and fluid responses to scripted commands. This tool is a first example of a multi-purpose, full-scale numerical modeling of a flight instrument with the purpose of supplementing or even eliminating entirely the need for a hardware engineer model during instrument development and operation. SAMSIM simulates the complex interactions that occur between the instrument Command and Data Handling unit (C&DH) and all subsystems during the execution of experiment sequences. A typical SAM experiment takes many hours to complete and involves hundreds of components. During the simulation, the electrical, mechanical, thermal, and gas dynamics states of each hardware component are accurately modeled and propagated within the simulation environment at faster than real time. This allows the simulation, in just a few minutes, of experiment sequences that takes many hours to execute on the real instrument. The SAMSIM model is divided into five distinct but interacting modules: software, mechanical, thermal, gas flow, and electrical modules. The software module simulates the instrument C&DH by executing a customized version of the instrument flight software in a Matlab environment. The inputs and outputs to this synthetic C&DH are mapped to virtual sensors and command lines that mimic in their structure and connectivity the layout of the instrument harnesses. This module executes, and thus validates, complex command scripts prior to their up-linking to the SAM instrument. As an output, this module generates synthetic data and message logs at a rate that is similar to the actual instrument.

Benna, Mehdi↗
NASA NTRS1997

Optimization of 2005 Sample Return for Mars Exopaleontology

If we wish to retain the search for past/present life as an objective in the program, exobiological milestones need to be addressed at each opportunity. At this stage of Mars exploration, these milestones are complimentary to those of other disciplines. Achieving the goals of the exobiological community will require a long-term view of the problem and a phased plan of exploration. Decision trees are a realistic way to assess present needs and to benefit from future missions. Timely progress in achieving milestones identified will ensure we are moving ahead at each opportunity in meeting what the Agency (and public!) regards as an essential part of the program. It is clear that life-related issues must be handled carefully as they are presented to NASA and the public not to create an imbalance, or more importantly, misunderstanding about where we are in the exploration plan.

Farmer, Jack D.↗
NASA NTRS2000

Treatability Study of Pentaborane(9)

Procedures for the safe destruction of liquid pentaborane(9), B5H9, by solvolysis were investigated. The objective of the study was to establish the optimum conditions for a pilot plant operation that would use water, or alcohol, or water-alcohol mixtures as the solvolysis reagent Small amounts of B5H9 sprayed from a syringe will not necessarily enflame, nor will a small pool on a spot plate. Therefore, a procedure was developed to reproducibly demonstrate the flammability of B5H9 In these tests every sample of neat B5H9 ignited and burned with a very sooty flame till the sample was consumed. The spontaneous self-ignition of B5H9 was quenched by the addition of small concentrations of the ethers THF (tetrahydrofuran) or DME (1,2-dimethoxy ethane). It was found that 10% (volume) of either provided total quenching with a large margin of safety. When these stabilized solutions were exposed to air, they decomposed and evaporated leaving a residue that was identified by NMR analysis as boric acid. Most of the laboratory solvolysis experiments used the 90% B5H9, 10% THF solution. This mixture was safer to handle and its solvolysis reactivity was virtually identical to that of 100% B5H9. Reaction rates were analyzed by measurement of hydrogen evolved during the solvolysis reactions. In terms of the minimum overall complete reaction time, the data indicate that 50150 alcohol/water is the optimum solvolysis reagent. This reaction produced a mixture of boric acid, B(OH)3, and triethoxyborane, B(OEt)3 [Et = C2H5], and mixed exchange derivatives thereof.

McDonald, Joseph K.↗
NASA NTRS2022

Compositional Measurements of Saturn's Upper Atmosphere and Rings From Cassini INMS: An Extended Analysis of Measurements From Cassini's Grand Finale Orbits

The Cassini spacecraft's final orbits sampled Saturn's atmosphere and returned surprisingly complex mass spectra from the Ion and Neutral Mass Spectrometer (INMS). Signal returned from the instrument included native Saturn species, as expected, as well as a significant amount of signal attributed to vaporized ices and higher mass organics believed to be owing into Saturn's atmosphere from the rings. In this paper, we present an in-depth compositional analysis of the mass spectra returned from Cassini's last few orbits. We use a mass spectral deconvolution algorithm designed specifically to handle the complexities involved with unit resolution spaceflight mass spectrometry data to determine the relative abundance of species detected in the observations. We calculate the downward external flux and mass deposition rates of ring volatile species into Saturn's atmosphere and conclude that during these observations ring material was being deposited into Saturn's equatorial region at a rate on the order of 10(exp 4) kg/s.

Saturn↗
NASA NTRS2022

Compositional Measurements of Saturn's Upper Atmosphere and Rings From Cassini INMS: An Extended Analysis of Measurements From Cassini's Grand Finale Orbits

The Cassini spacecraft's final orbits sampled Saturn's atmosphere and returned surprisingly complex mass spectra from the Ion and Neutral Mass Spectrometer (INMS). Signal returned from the instrument included native Saturn species, as expected, as well as a significant amount of signal attributed to vaporized ices and higher mass organics believed to be flowing into Saturn's atmosphere from the rings. In this paper, we present an in-depth compositional analysis of the mass spectra returned from Cassini's last few orbits. We use a mass spectral deconvolution algorithm designed specifically to handle the complexities involved with unit resolution spaceflight mass spectrometry data to determine the relative abundance of species detected in the observations. We calculate the downward external flux and mass deposition rates of ring volatile species into Saturn's atmosphere and conclude that during these observations ring material was being deposited into Saturn's equatorial region at a rate on the order of 10(exp 4) kg/s.

Saturn↗
NASA NTRS2000

Adiabatic Compression Sensitivity of Liquid Fuels and Monopropellants

Liquid rocket propellants can be sensitive to rapid compression. Such liquids may undergo decomposition and their handling may be accompanied with risk. Decomposition produces small gas bubbles in the liquid, which upon rapid compression may cause catastrophic explosions. The rapid compression can result from mechanical shocks applied on the tank containing the liquid or from rapid closure of the valves installed on the lines. It is desirable to determine the conditions that may promote explosive reactions. At Air Force Research Laboratory (AFRL), we constructed an apparatus and established a safe procedure for estimating the sensitivity of propellant materials towards mechanical shocks (Adiabatic Compression Tester). A sample is placed on a stainless steel U-tube, held isothermally at a temperature between 20 and 150 C then exposed to an abrupt mechanical shock of nitrogen gas at a pressure between 6.9 and 20.7 MPa (1000 to 3000 psi). The apparatus is computer interfaced and is driven with LABTECH NOTEBOOK-pro (registered) Software. In this presentation, the design of the apparatus is shown, the operating procedure is outlined, and the safety issues are addressed. The results obtained on different energetic materials are presented.

Ismail, Ismail M. K.↗
NASA NTRSDate not supplied

Reconstructing Subgrid Cloud Variability Guided by CloudSat/CALIPSO Observations

Predicting realistic cloud subgrid variability remains a challenge for Global Climate Models (GCMs) even though it can play an important role for proper representation of processes pertaining to cloud microphysics, precipitation, and radiation, but also for comparisons with satellite observations which are of much higher resolution than model grids. The diagnostic approach of subgrid cloud variability is commonly handled by subcolumn cloud generators. For a specific GCM, one ideally wants the subgrid variability used for comparisons with satellite observations to be created by the same generator and with the same rules as the one used for model integration. With this in mind, we have embarked in an effort to test and improve cloud subcolumn generators appropriate for GCMs. For this purpose, we use cloud (hydrometeor) products from active observations by the CloudSat radar (CPR) and the CALIPSO lidar (CALIOP). Cloud products from active sensors while suffering significant sampling and coverage drawbacks have the advantage of resolving both horizontal and vertical variability. The main question is: given a profile of cloud condensate mean and variance, can we create a subgrid cloud field that is statistically similar to the observed subgrid cloud field? By “statistically”, we suggest that we do not aspire to reproduce “well” each individual subgrid cloud field of a GCM-scale region, but that our generator performs well for a large ensemble of cases. We simulate radar, passive imager, and radiation flux fields from the observed 2D cloud fields and create one-point statistics; we use the profiles of cloud condensate mean and variance as input to the generator to create subgrid cloud fields; we compare the statistics; we adjust the rules of the generator to create the best possible agreement between the radar, imager and radiation field statistics. In this process, the active observations have actually a dual role: they provide the actual subgrid cloud field which can be reduced to a profile of mean and variance used by the subcolumn generator, but they also provide the rules needed by the generator, such as measures (e.g, decorrelation length) of the vertical overlap of cloud fraction and of the condensate horizontal variability. This presentation will show our progress using Cloudsat and CALIPSO products in this dual fashion.

cloud↗
NASA NTRS2021

BUMPER: A Tool for Analyzing Spacecraft Micrometeoroid and Orbital Debris Risk

“Bumper” is a computer program for analyzing spacecraft micrometeoroid and orbital debris (MMOD) risk. Bumper was developed in the late-1980s and has been continuously maintained and used since. The user base has grown from a few government entities now include numerous commercial entities as well. The National Aeronautics and Space Administration (NASA) Johnson Space Center (JSC) Hypervelocity Impact Technology (HVIT) group is responsible for all aspects of the Bumper software. Bumper has been used to characterize MMOD risk on many spacecraft. All of the International Space Station (ISS) modules, visiting vehicles and numerous external components and systems have been analyzed. Bumper was used to analyze the Space Shuttle, Orion, and many space probes, telescopes and satellites. Bumper is also being used to analyze future spacecraft such as the Deep Space Gateway (DSG) and Mars Sample Return (MSR) missions. The Bumper Configuration Control Board (CCB) ensures that all changes to the code are approved, reviewed, and documented. The current Bumper version – “Bumper 3” – is a Fortran executable that utilizes a 64-bit architecture. Bumper has numerous features that make it a powerful tool for analyzing spacecraft MMOD risk. Bumper uses the latest orbital debris and meteoroidenvironment models. Bumper also has a large library of ballistic limit “damage” equations available that can be used for a wide variety of MMOD shielding configurations. Bumper can also handle large spacecraft finite elementmodels (FEMs) and conducts checks of the model. This paper introduces Bumper and the MMOD risk analysis process using a simplified cube-shaped spacecraft model

hypervelocity↗
NASA NTRS2024

PACE Technical Report Series, Volume 12: The PACE Level 1C data format

NASA's Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission will make global ocean color and atmospheric measurements to provide extended data records of ocean ecology and global biogeochemistry, along with polarimetric measurements for advanced observations of aerosols, clouds and the ocean. PACE will contain three instruments: the primary Ocean Color Instrument (OCI), and two multi-angle polarimeters (MAPs). The latter instruments are contributed under a ‘Do-No-Harm’ (to the rest of the PACE mission) principle, and the PACE Science Data Processing System (SDPS) is only required to produce Level-1b (geolocated radiances with calibration applied) data without performance requirements. However, there is a strong desire to produce data in a format that merges the disparate spatial resolutions, viewing geometry and sampling nature of the three instruments. Our terminology for this format is Level 1c (L1C). This format will be an input to Level 2 algorithms produced from standalone MAP instrument observations, or from algorithms employing multi-sensor fusion. Creating the L1C format has several components. This includes choice of projection method, the means by which multi-angle views are properly incorporated into that projection (‘aggregation’) the means to represent wavelength and light polarization state, the selection of data to be included within the L1C file, and the handling of ancillary data either required for L1C file generation or needed in that format for L2 processing.

PACE↗
NASA NTRS1985

Physical, anthropometrical, and body composition characteristics of workers at Kennedy Space Center

At the Kennedy Space Center, workers are often exposed to cardiovascular and muscular stress in job-related activities which may require a high level of physical fitness in order to safely complete the work task. Similar tasks will be performed at other launch and landing facilities and in space for the Space Station. One such category includes workers who handle toxic propellants and must wear Self-Contained Atmospheric Protective Ensembles (SCAPE) that can weigh 56 lbs. with the air pack. These suits provide a significant physical challenge to many of the workers in terms of carrying this load while moving about and performing work. Furthermore, under some conditions, there is a significant thermal stress. The physical characteristics of these workers are, therefore, of consequence. The purpose of this study was to analyze the anthropometry, body composition, strength, power, endurance, flexibility, aerobic fitness, and blood variables of a representative sample of male KSC SCAPE workers and to compare them with characteristics of other male workers at KSC (total population N=110). Three separate comparisons were made.

Lasley, M. L.↗
NASA NTRSDate not supplied

Reconstructing Subgrid Cloud Variability Guided by CloudSat/CALIPSO Observations

Predicting realistic cloud subgrid variability remains a challenge for Global Climate Models (GCMs) even though it can play an important role for proper representation of processes pertaining to cloud microphysics, precipitation, and radiation, but also for comparisons with satellite observations which are of much higher resolution than model grids. The diagnostic approach of subgrid cloud variability is commonly handled by subcolumn cloud generators. For a specific GCM, one ideally wants the subgrid variability used for comparisons with satellite observations to be created by the same generator and with the same rules as the one used for model integration. With this in mind, we have embarked in an effort to test and improve cloud subcolumn generators appropriate for GCMs. For this purpose, we use cloud (hydrometeor) products from active observations by the CloudSat radar (CPR) and the CALIPSO lidar (CALIOP). Cloud products from active sensors while suffering signi􀂦cant sampling and coverage drawbacks have the advantage of resolving both horizontal and vertical variability. The main question is: given a profile of cloud condensate mean and variance, can we create a subgrid cloud field that is statistically similar to the observed subgrid cloud field? By “statistically”, we suggest that we do not aspire to reproduce “well” each individual subgrid cloud field of a GCM-scale region, but that our generator performs well for a large ensemble of cases. We simulate radar, passive imager, and radiation flux fields from the observed 2D cloud fields and create one-point statistics; we use the profiles of cloud condensate mean and variance as input to the generator to create subgrid cloud fields; we compare the statistics; we adjust the rules of the generator to create the best possible agreement between the radar, imager and radiation field statistics. In this process, the active observations have actually a dual role: they provide the actual subgrid cloud field which can be reduced to a profile of mean and variance used by the subcolumn generator, but they also provide the rules needed by the generator, such as measures (e.g, decorrelation length) of the vertical overlap of cloud fraction and of the condensate horizontal variability. This presentation will show our progress using Cloudsat and CALIPSO products in this dual fashion.

clouds↗
DOE OSTI2024

Latent Stochastic Differential Equations for Modeling Quasar Variability and Inferring Black Hole Properties

Quasars are bright and unobscured active galactic nuclei (AGN) thought to be powered by the accretion of matter around supermassive black holes at the centers of galaxies. The temporal variability of a quasar’s brightness contains valuable information about its physical properties. The UV/optical variability is thought to be a stochastic process, often represented as a damped random walk described by a stochastic differential equation (SDE). Upcoming wide-field telescopes such as the Rubin Observatory Legacy Survey of Space and Time (LSST) are expected to observe tens of millions of AGN in multiple filters over a ten year period, so there is a need for efficient and automated modeling techniques that can handle the large volume of data. Latent SDEs are machine learning models well suited for modeling quasar variability, as they can explicitly capture the underlying stochastic dynamics. In this work, we adapt latent SDEs to jointly reconstruct multivariate quasar light curves and infer their physical properties such as the black hole mass, inclination angle, and temperature slope. Our model is trained on realistic simulations of LSST ten year quasar light curves, and we demonstrate its ability to reconstruct quasar light curves even in the presence of long seasonal gaps and irregular sampling across different bands, outperforming a multioutput Gaussian process regression baseline. Our method has the potential to provide a deeper understanding of the physical properties of quasars and is applicable to a wide range of other multivariate time series with missing data and irregular sampling.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2006

Multichannel Phase and Power Detector

An electronic signal-processing system determines the phases of input signals arriving in multiple channels, relative to the phase of a reference signal with which the input signals are known to be coherent in both phase and frequency. The system also gives an estimate of the power levels of the input signals. A prototype of the system has four input channels that handle signals at a frequency of 9.5 MHz, but the basic principles of design and operation are extensible to other signal frequencies and greater numbers of channels. The prototype system consists mostly of three parts: An analog-to-digital-converter (ADC) board, which coherently digitizes the input signals in synchronism with the reference signal and performs some simple processing; A digital signal processor (DSP) in the form of a field-programmable gate array (FPGA) board, which performs most of the phase- and power-measurement computations on the digital samples generated by the ADC board; and A carrier board, which allows a personal computer to retrieve the phase and power data. The DSP contains four independent phase-only tracking loops, each of which tracks the phase of one of the preprocessed input signals relative to that of the reference signal (see figure). The phase values computed by these loops are averaged over intervals, the length of which is chosen to obtain output from the DSP at a desired rate. In addition, a simple sum of squares is computed for each channel as an estimate of the power of the signal in that channel. The relative phases and the power level estimates computed by the DSP could be used for diverse purposes in different settings. For example, if the input signals come from different elements of a phased-array antenna, the phases could be used as indications of the direction of arrival of a received signal and/or as feedback for electronic or mechanical beam steering. The power levels could be used as feedback for automatic gain control in preprocessing of incoming signals. For another example, the system could be used to measure the phases and power levels of outputs of multiple power amplifiers to enable adjustment of the amplifiers for optimal power combining.

Li, Samuel↗
NASA NTRS2007

Automated Hybridization of X-ray Absorber Elements-A Path to Large Format Microcalorimeter Arrays

In the design of microcalorimeters, it is often desirable to produce the X-ray absorber separately from the detector element. In this case, the attachment of the absorber to the detector element with the required thermal and mechanical characteristics is a major challenge. In such arrays, the attachment has been done by hand. This process is not easily extended to the large format arrays required for future X- ray astronomy missions such as the New x-ray Telescope or NeXT. In this paper we present an automated process for attaching absorber tiles to the surface of a large-scale X-ray detector array. The absorbers are attached with stycast epoxy to a thermally isolating polymer structure made of SU-8. SU-8 is a negative epoxy based photo resist produced by Microchem. We describe the fabrication of the X-ray absorbers and their suspension on a handle die in an adhesive matrix. We describe the production process for the polymer isolators on the detector elements. We have developed a new process for the alignment, and simultaneous bonding of the absorber tiles to an entire detector array. This process uses equipment and techniques used in the flip-chip bonding industry and approaches developed in the fabrication of the XRS-2 instrument. XRS-2 was an X-ray spectrometer that was launched on the Suzaku telescope in July 10, 2005. We describe the process and show examples of sample arrays produced by this process. Arrays with up to 300 elements have been bonded. The present tests have used dummy absorbers made of Si. In future work, we will demonstrate bonding of HgTe absorbers.

Moseley, S.↗
DOE OSTI2024

Efficient First-Order Algorithms for Large-Scale, Non-Smooth Maximum Entropy Models with Application to Wildfire Science

Maximum entropy (MaxEnt) models are a class of statistical models that use the maximum entropy principle to estimate probability distributions from data. Due to the size of modern data sets, MaxEnt models need efficient optimization algorithms to scale well for big data applications. State-of-the-art algorithms for MaxEnt models, however, were not originally designed to handle big data sets; these algorithms either rely on technical devices that may yield unreliable numerical results, scale poorly, or require smoothness assumptions that many practical MaxEnt models lack. In this paper, we present novel optimization algorithms that overcome the shortcomings of state-of-the-art algorithms for training large-scale, non-smooth MaxEnt models. Our proposed first-order algorithms leverage the Kullback–Leibler divergence to train large-scale and non-smooth MaxEnt models efficiently. For MaxEnt models with discrete probability distribution of n elements built from samples, each containing m features, the stepsize parameter estimation and iterations in our algorithms scale on the order of O(mn) operations and can be trivially parallelized. Moreover, the strong ℓ1 convexity of the Kullback–Leibler divergence allows for larger stepsize parameters, thereby speeding up the convergence rate of our algorithms. To illustrate the efficiency of our novel algorithms, we consider the problem of estimating probabilities of fire occurrences as a function of ecological features in the Western US MTBS-Interagency wildfire data set. Our numerical results show that our algorithms outperform the state of the art by one order of magnitude and yield results that agree with physical models of wildfire occurrence and previous statistical analyses of wildfire drivers.

Physics↗