Search NASA⌕ Search

SEARCH · Search NASA

Results for “deep generative models”

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 1,441 records · Page 80

Deep learning of dynamically responsive chemical Hamiltonians with semiempirical quantum mechanics

Conventional machine-learning (ML) models in computational chemistry learn to directly predict molecular properties using quantum chemistry only for reference data. While these heuristic ML methods show quantum-level accuracy with speeds several orders of magnitude faster than traditional quantum chemistry methods, they suffer from poor extensibility and transferability; i.e., their accuracy degrades on large or new chemical systems. Incorporating quantum chemistry frameworks into the ML models directly solves this problem. Here we take the structure of semiempirical quantum mechanics (SEQM) methods to construct dynamically responsive Hamiltonians. SEQM methods use empirical parameters fitted to experimental properties to construct reduced-order Hamiltonians, facilitating much faster calculations than ab initio methods but with compromised accuracy. By replacing these static parameters with machine-learned dynamic values inferred from the local environment, we greatly improve the accuracy of the SEQM methods. Trained on molecular energies and atomic forces, these dynamically generated Hamiltonian parameters show a strong correlation with atomic hybridization and bonding. Trained with only about 60,000 small organic molecular conformers, the resulting model retains interpretability, extensibility, and transferability when testing on much larger chemical systems and predicting various molecular properties. Overall, this work demonstrates the virtues of incorporating physics-based descriptions with ML to develop models that are simultaneously accurate, transferable, and interpretable.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Deep learning of experimental electrochemistry for battery cathodes across diverse compositions

Artificial intelligence (AI) has emerged as a tool for discovering and optimizing novel battery materials. However, the adoption of AI in battery cathode representation and discovery is still limited due to the complexity of optimizing multiple performance properties and the scarcity of high-fidelity data. Here, we present a machine learning model (DRXNet) for battery informatics and demonstrate the application in the discovery and optimization of disordered rocksalt (DRX) cathode materials. We have compiled the electrochemistry data of DRX cathodes over the past 5 years, resulting in a dataset of more than 19,000 discharge voltage profiles on diverse chemistries spanning 14 different metal species. Learning from this extensive dataset, our DRXNet model can capture critical features in the cycling curves of DRX cathodes under various conditions. Our approach offers a data-driven solution to facilitate the rapid identification of novel cathode materials, accelerating the development of next-generation batteries for carbon neutralization.

25 ENERGY STORAGE↗

Coupled Aerodynamic and Hydrodynamic Hybrid Simulation of Floating Offshore Wind Turbines

The development and innovation of floating offshore wind energy in the U.S. requires detailed high-fidelity observations and measurements of turbine and platform loading due to wind, waves, and currents. However, full-scale and quasi-full-scale experiments require significant financial and temporal investments for construction, experimental testing, and long-term field campaigns. To support the commercial advancement of the offshore wind energy industry, specialized wind tunnel and wave basin experimental facilities are critical to be able to test FOWT designs at small scale under controlled conditions prior to full-scale deployment. Oregon State University (OSU) is internationally known as a leader in water and energy research, development, and testing. The O.H. Hinsdale Wave Research Laboratory (HWRL) and the Wallace Energy Systems and Renewables Facility (WESRF) at OSU have extensive experience building, modeling, monitoring, controlling, and actuating scaled systems. Experiments on wave-structure interaction have been performed at the HWRL since its establishment in 1972. Studies have included the interaction of waves with coastal structures (breakwaters, seawalls, buildings, cylinders, bridges, fixed foundations of offshore wind turbines, etc.) and with floating structures (e.g., wave energy converters, maneuvering of vessels, etc.). Hinsdale is actively used by marine energy technology developers, both for private testing and OSU-collaborative research projects. However, despite the availability of several large-scale facilities for hydrodynamic testing (at OSU and elsewhere in the U.S.), existing experimental laboratories are generally limited in their ability to accurately generate combined wind and wave conditions. The simulation of both wind and waves in experimental testing is complicated due to a number of constraints, including: [i] incompatible similitude laws governing the wind and waves for scaled experiments, [ii] producing accurate wind over a large enough control volume via fans, and [iii] generating wind that reasonably represents the atmospheric boundary layer in existing wave basins/flumes. Hence, physical test data providing insight into the simultaneous wave- and wind-structure response of floating offshore wind components can be difficult to generate. Given the aforementioned challenges in classic hydrodynamic experiments, the motivation of this project is to establish a real-time hybrid simulation (RTHS) approach that can apply aero- and hydro-dynamic loading by augmenting wave-only experimental facilities with virtual aerodynamic forces through numerical models representing the remaining dynamic forces. RTHS is a physical-numerical approach that partitions a prototype system into physical and numerical sub-assemblies that interact with each other through actuators and sensors in real time. In coupling physical and numerical models, the hybrid simulation approach applied herein is ideal for problems with: (1) structures subjected to different scaling laws, such as floating offshore wind turbines subjected to combined aero/hydro-dynamic loading, (2) structures that are too large or complex to be tested entirely in a laboratory setting, such as deep-water mooring applications, and (3) component testing, where the behavior of a portion of the assembly is uncertain but still interacts with other portions of the structure, such as testing the fatigue life of turbine blades. Few U.S. experimental facilities are able to test simultaneous aero- and hydro-dynamic loading and none can accurately produce aero/hydro-dynamic response on scaled FOWT models due to conflicting similitude laws between the wind (commonly Reynolds) and the waves (commonly Froude). To aid in accelerating the development of the U.S. floating offshore industry, there is a significant need to develop a flexible, modular framework that can expand the capacities of existing wave-only laboratories. The project goal is to demonstrate a hydrodynamic real-time hybrid simulation (hydro-RTHS) framework that couples numerical wind and physical waves acting on a FOWT, thus representing simultaneous aero/hydro-dynamic loading. The FOWT is partitioned into a full-scale numerical sub-assembly associated with the aerodynamics and a model-scale physical sub-assembly associated with the hydrodynamics. The numerical-physical partition associated with hydro-RTHS mitigates scaling constraints by supplying different scaling laws to the physical and numerical sub-assemblies. Herein, length, force, and time are scaled and exchanged between the sub-assemblies using Froude scaling to represent the open-channel flow in the physical sub-assembly. Other similitude laws could also be utilized depending on the problem definition. It is envisioned that the ability to model FOWTs under waves and wind, with mitigation of similitude distortions, would result in reduced development costs (currently, FOWT concept development is performed with full-size pro- totypes at enormous expense and risk) and increase the reliability of the FOWT industry (since extreme wave and wind conditions and contingency events can be tested safely in a controlled environment).

16 TIDAL AND WAVE POWER↗

Downscaling atmospheric chemistry simulations with physically consistent deep learning

Abstract. Recent advances in deep convolutional neural network (CNN)-based super resolution can be used to downscale atmospheric chemistry simulations with substantially higher accuracy than conventional downscaling methods. This work both demonstrates the downscaling capabilities of modern CNN-based single image super resolution and video super-resolution schemes and develops modifications to these schemes to ensure they are appropriate for use with physical science data. The CNN-based video super-resolution schemes in particular incur only 39 % to 54 % of the grid-cell-level error of interpolation schemes and generate outputs with extremely realistic small-scale variability based on multiple perceptual quality metrics while performing a large (8×10) increase in resolution in the spatial dimensions. Methods are introduced to strictly enforce physical conservation laws within CNNs, perform large and asymmetric resolution changes between common model grid resolutions, account for non-uniform grid-cell areas, super-resolve lognormally distributed datasets, and leverage additional inputs such as high-resolution climatologies and model state variables. High-resolution chemistry simulations are critical for modeling regional air quality and for understanding future climate, and CNN-based downscaling has the potential to generate these high-resolution simulations and ensembles at a fraction of the computational cost.

54 ENVIRONMENTAL SCIENCES↗

Tracer evolution in winds generated by a global spectral mechanistic model

The lower boundary of a spectral mechanistic model is prescribed with 100 hPa geopotentials, and its performance during a November 1989 through March 1990 integration is compared with National Meteorological Center observations. Although the stratopause temperatures quickly become biased near the pole in both hemispheres, the model develops a residual mean circulation which shows significant descent over the winter pole and ascent in the tropics and over the summer pole at pressures less than 10 hPa. The daily correspondence of observed to modeled features in the upper stratosphere and mesosphere degrades after one month. However, the long-term variability qualitatively follows the observations. The results of off-line transport experiments are also described. A passive tracer is instantaneously injected into the flow over the poles and evolves in a manner which is consistent with the residual mean circulation. It demonstrates a significant cross-equatorial flux in the mesosphere near solstice, and air which originates in the southern hemisphere polar mesosphere can be found descending deep into the nothern polar stratosphere at the end of the integration. Nitrous oxide is also transported, and its ability to act as a dynamical tracer is evaluated by comparison to the evolution of the passive tracer.

Nielsen, J. E.↗

The Collection, Usage, and Preliminary Examination of the Apollo Sample Suite: Lessons for Artemis

Apollo Sample Collection and Usage: From 1969 to 1972 there were six Apollo missions to the surface of the Moon during which the astronauts collected 382 kg of rock and regolith (~2200 samples). The samples collected fall into these general categories: rocks (~66% by mass), rake samples (~4%), bulk regolith (~24%), and specialty regolith (deep drill cores, drive tubes, sealed bulk regolith) samples (~6%). In each category there are a variety of different subtypes available for study, e.g., among the bulk regolith samples there are also skim, trench, and (partially) shaded regolith samples each sampling unique types or depths of regolith. This variety of subsamples has enabled a multitude of different studies over the past 55 years (>3400 individual requests). We are still averaging ~50 unique requests and have allocated >500 individual Apollo samples annually for the past 10 years (2020 excepted). Looking at the 4,675 non-ANGSA (Apollo Next Generation Sample Analysis) samples allocated over the past 10 years, the proportions of allocated samples do not precisely align with the abundance (by mass) of those samples withing the collection: Rock (69.4 %); Rake (10.8 %); Bulk Regolith (15.5 %); Drive Tube (3.3 %); Core/Specialty (1.0 %). Apollo Preliminary Examination (PE): The PE process differs significantly for the various sample types enumerated above; we focus on regolith and rock samples here. During the Apollo mission era, the PE process evolved over the course of the missions; below is what was done for the Apollo 17 mission. For regolith samples, the PE process was: (1) documented bags containing regolith are opened, photographed, and described; (2) large rocks are removed and treated separately; (3) 25% to 33% of the bulk soil is scooped out, weighed, and stored in reserve; (4) the remaining sample is sieved to produce the size fractions <1, 1- 2, 2-4, and 4-10 mm, all of which are weighed. For rock samples, the process is: (1) removing rocks from the container(s) it came back from the Moon in; (2) rematching any materials that spalled off the rock to their original location; (3) numbering, weighing, and basic photographic documentation; (4) dusting with a gentle N2 gas jet; (5) Orthogonal photography; (6) detailed description of the textures and features of the rock; (7) rock modelling and measurement; (8) stereophotography; (9) determination of the orientation of the rock on the lunar surface. Drive tubes and deep drill cores were not characterized during PE beyond an initial weight and a sketch of the interior tube materials derived from 2D medical X-ray images. Catalogs: The ongoing utility of the Apollo samples is enabled by the robust cataloguing process for the samples [3-5], which allows the scientific community to accurately request samples uniquely suited to their proposed studies. A common misconception, however, is the amount of detail that goes into the initial catalog (e.g., [6]) for a collection from the preliminary examination (PE) period, versus what goes into the catalogs that come later in the life cycle of the samples from that mission (e.g., [7]). The only required data for a PE catalog is a weight, a basic photograph, and a description of the nature of the sample. Artemis PE: Over the past few years, the ongoing ANGSA project studied previously unopened Apollo 17 double drive tube samples 73001/2 [1], and a PE of the drive tubes was done. The PE took the existing core dissection process (developed during PE of Apollo cores in the 1970s, 1980s), and modernized it [7]. The main lesson from the ANGSA PE relative to future missions was that the physical work done during PE of lunar samples has not changed much over the past 5 decades. The use of “modern” technology during PE (e.g., XCT; multispectral analyses) resulted in an enhanced initial understanding of the 73001 and 73002 drive tubes, but greatly increased the time required. The lessons learned from recent astromaterial PEs (e.g., ANGSA and OREx) are important to consider when planning for Artemis, but the unique nature of the Artemis Campaign means many lessons learned from these mission will not be applicable. Given the time constraints (6 months) and likely number of samples that will be returned by Artemis (>200), the Artemis PE catalog will necessarily look much more like [4] than [6].

J Gross↗

Deep functional analysis of synII, a 770-kilobase synthetic yeast chromosome

INTRODUCTION Although much effort has been devoted to studying yeast in the past few decades, our understanding of this model organism is still limited. Rapidly developing DNA synthesis techniques have made a “build-to-understand” approach feasible to reengineer on the genome scale. Here, we report on the completion of a 770-kilobase synthetic yeast chromosome II (synII). SynII was characterized using extensive Trans-Omics tests. Despite considerable sequence alterations, synII is virtually indistinguishable from wild type. However, an up-regulation of translational machinery was observed and can be reversed by restoring the transfer RNA (tRNA) gene copy number. RATIONALE Following the “design-build-test-debug” working loop, synII was successfully designed and constructed in vivo. Extensive Trans-Omics tests were conducted, including phenomics, transcriptomics, proteomics, metabolomics, chromosome segregation, and replication analyses. By both complementation assays and SCRaMbLE (synthetic chromosome rearrangement and modification by loxP -mediated evolution), we targeted and debugged the origin of a growth defect at 37°C in glycerol medium. RESULTS To efficiently construct megabase-long chromosomes, we developed an I- Sce I–mediated strategy, which enables parallel integration of synthetic chromosome arms and reduced the overall integration time by 50% for synII. An I- Sce I site is introduced for generating a double-strand break to promote targeted homologous recombination during mitotic growth. Despite hundreds of modifications introduced, there are still regions sharing substantial sequence similarity that might lead to undesirable meiotic recombinations when intercrossing the two semisynthetic chromosome arm strains. Induction of the I- Sce I–mediated double-strand break is otherwise lethal and thus introduced a strong selective pressure for targeted homologous recombination. Since our strategy is designed to generate a markerless synII and leave the URA3 marker on the wild-type chromosome, we observed a tenfold increase in URA3 -deficient colonies upon I- Sce I induction, meaning that our strategy can greatly bias the crossover events toward the designated regions. By incorporating comprehensive phenotyping approaches at multiple levels, we demonstrated that synII was capable of powering the growth of yeast indistinguishably from wild-type cells (see the figure), showing highly consistent biological processes comparable to the native strain. Meanwhile, we also noticed modest but potentially significant up-regulation of the translational machinery. The main alteration underlying this change in expression is the deletion of 13 tRNA genes. A growth defect was observed in one very specific condition—high temperature (37°C) in medium with glycerol as a carbon source—where colony size was reduced significantly. We targeted and debugged this defect by two distinct approaches. The first approach involved phenotype screening of all intermediate strains followed by a complementation assay with wild-type sequences in the synthetic strain. By doing so, we identified a modification resulting from PCRTag recoding in TSC10 , which is involved in regulation of the yeast high-osmolarity glycerol (HOG) response pathway. After replacement with wild-type TSC10 , the defect was greatly mitigated. The other approach, debugging by SCRaMbLE, showed rearrangements in regions containing HOG regulation genes. Both approaches indicated that the defect is related to HOG response dysregulation. Thus, the phenotypic defect can be pinpointed and debugged through multiple alternative routes in the complex cellular interactome network. CONCLUSION We have demonstrated that synII segregates, replicates, and functions in a highly similar fashion compared with its wild-type counterpart. Furthermore, we believe that the iterative “design-build-test-debug” cycle methodology, established here, will facilitate progression of the Sc2.0 project in the face of the increasing synthetic genome complexity. SynII characterization. ( A ) Cell cycle comparison between synII and BY4741 revealed by the percentage of cells with separated CEN2-GFP dots, metaphase spindles, and anaphase spindles. ( B ) Replication profiling of synII (red) and BY4741 (black) expressed as relative copy number by deep sequencing. ( C ) RNA sequencing analysis revealed that the significant up-regulation of translational machinery in synII is induced by the deletion of tRNA genes in synII.

Shen, Yue↗

The Implications of 3D Thermal Structure on 1D Atmospheric Retrieval

Using the atmospheric structure from a 3D global radiation-hydrodynamic simulation of HD 189733b and the open-source Bayesian Atmospheric Radiative Transfer (BART) code, we investigate the difference between the secondary-eclipse temperature structure produced with a 3D simulation and the best-fit 1D retrieved model. Synthetic data are generated by integrating the 3D models over the Spitzer , the Hubble Space Telescope ( HST ), and the James Web Space Telescope ( JWST ) bandpasses, covering the wavelength range between 1 and 11 μ m where most spectroscopically active species have pronounced features. Using the data from different observing instruments, we present detailed comparisons between the temperature–pressure profiles recovered by BART and those from the 3D simulations. We calculate several averages of the 3D thermal structure and explore which particular thermal profile matches the retrieved temperature structure. We implement two temperature parameterizations that are commonly used in retrieval to investigate different thermal profile shapes. To assess which part of the thermal structure is best constrained by the data, we generate contribution functions for our theoretical model and each of our retrieved models. Our conclusions are strongly affected by the spectral resolution of the instruments included, their wavelength coverage, and the number of data points combined. We also see some limitations in each of the temperature parametrizations, as they are not able to fully match the complex curvatures that are usually produced in hydrodynamic simulations. The results show that our 1D retrieval is recovering a temperature and pressure profile that most closely matches the arithmetic average of the 3D thermal structure. When we use a higher resolution, more data points, and a parametrized temperature profile that allows more flexibility in the middle part of the atmosphere, we find a better match between the retrieved temperature and pressure profile and the arithmetic average. The Spitzer and HST simulated observations sample deep parts of the planetary atmosphere and provide fewer constraints on the temperature and pressure profile, while the JWST observations sample the middle part of the atmosphere, providing a good match with the middle and most complex part of the arithmetic average of the 3D temperature structure.

79 ASTRONOMY AND ASTROPHYSICS↗

Southern Ocean Response to NADW Changes

The possibility of North Atlantic Deep Water (NADW) changes in both past and future climates has raised the issue of how the Southern Ocean would respond. Recent experiments with the GISS coupled atmosphere-ocean model have shown that a "bipolar see-saw" between NADW production and Antarctic Bottom Water (AABW) production in the Weddell Sea can occur in conjunction with freshening of the North Atlantic. However, this effect operates not through a slow ocean response but via a rapid atmospheric mechanism. As NADW reduces, colder temperatures in the North Atlantic, and Northern Hemisphere in general, are associated with higher surface pressure (increased atmospheric mass). Reduced mass in the Southern Hemisphere occurs in response, with lower pressure over the South Pole (an EOF #1 effect, the "high phase" of the Antarctic Oscillation).The lower pressure is associated with stronger west winds that generate an intensified Antarctic Circumpolar Current (ACC), which leads to longitudinal heat divergence in the South Atlantic (and heat convergence in the Southern Indian Ocean). Colder temperatures in the Weddell Sea region lead to sea ice growth, increased salinity and surface water density, and greater Weddell Sea Bottom Water production. Increased poleward transport of heat occurs in the South Atlantic in conjunction with increased bottom water production, but its convergence at high latitudes is not sufficient to offset the longitudinal heat divergence due to the intensified ACC. The colder temperatures at high latitudes in the South Atlantic increase the latitudinal temperature gradient, baroclinic instability, eddy energy and eddy poleward transport of momentum, helping to maintain the lower pressure over the pole in an interactive manner. The heat flux convergence in the Indian Ocean provides a warming tendency in that region, and overall global production of AABW remains unchanged. These results have implications for the interpretation of the ice core records of the last deglaciation, but may also be relevant for changes during the Holocene and perhaps even in response to increased CO2 forcing,

Rind, David↗

NASA Tech Briefs, October 2008

Topics covered include: Control Architecture for Robotic Agent Command and Sensing; Algorithm for Wavefront Sensing Using an Extended Scene; CO2 Sensors Based on Nanocrystalline SnO2 Doped with CuO; Improved Airborne System for Sensing Wildfires; VHF Wide-Band, Dual-Polarization Microstrip-Patch Antenna; Onboard Data Processor for Change-Detection Radar Imaging; Using LDPC Code Constraints to Aid Recovery of Symbol Timing; System for Measuring Flexing of a Large Spaceborne Structure; Integrated Formation Optical Communication and Estimation System; Making Superconducting Welds between Superconducting Wires; Method for Thermal Spraying of Coatings Using Resonant-Pulsed Combustion; Coating Reduces Ice Adhesion; Hybrid Multifoil Aerogel Thermal Insulation; SHINE Virtual Machine Model for In-flight Updates of Critical Mission Software; Mars Image Collection Mosaic Builder; Providing Internet Access to High-Resolution Mars Images; Providing Internet Access to High-Resolution Lunar Images; Expressions Module for the Satellite Orbit Analysis Program Virtual Satellite; Small-Body Extensions for the Satellite Orbit Analysis Program (SOAP); Scripting Module for the Satellite Orbit Analysis Program (SOAP); XML-Based SHINE Knowledge Base Interchange Language; Core Technical Capability Laboratory Management System; MRO SOW Daily Script; Tool for Inspecting Alignment of Twinaxial Connectors; An ATP System for Deep-Space Optical Communication; Polar Traverse Rover Instrument; Expert System Control of Plant Growth in an Enclosed Space; Detecting Phycocyanin-Pigmented Microbes in Reflected Light; DMAC and NMP as Electrolyte Additives for Li-Ion Cells; Mass Spectrometer Containing Multiple Fixed Collectors; Waveguide Harmonic Generator for the SIM; Whispering Gallery Mode Resonator with Orthogonally Reconfigurable Filter Function; Stable Calibration of Raman Lidar Water-Vapor Measurements; Bimaterial Thermal Compensators for WGM Resonators; Root Source Analysis/ValuStream[Trade Mark] - A Methodology for Identifying and Managing Risks; Ensemble: an Architecture for Mission-Operations Software; Object Recognition Using Feature-and Color-Based Methods; On-Orbit Multi-Field Wavefront Control with a Kalman Filter; and The Interplanetary Overlay Networking Protocol Accelerator.

Source record↗

Baryon Number Violation Search

Understanding the fundamental forces and symmetries of nature has long been a central goal of particle physics. While the Standard Model (SM) provides a successful framework, it does not guarantee the conservation of baryon number B or lepton number L, thus motivating searches for their violation. Proton decay, a fundamental process violating B, has been at the forefront of experimental searches for decades.The discovery of the weak neutral current in 1973 unified the electromagnetic and weak forces and inspired the creation of Grand Unified Theories (GUTs) that also unify the strong force. In 1974, the first-ever GUT, proposed by Georgi and Glashow, naturally predicted proton decay. This prediction led to an experimental push to validate these theories, and a large underground detector boom was born. Initially designed for proton decay searches, these detectors later proved invaluable to neutrino physics.Although no evidence for proton decay has yet been observed, next-generation large detectors, such as the Deep Underground Neutrino Experiment (DUNE), offer the opportunity to improve on current experimental limits. Utilizing its Liquid Argon Time Projection Chamber (LArTPC) technology, DUNE is positioned to probe rare processes such as proton decay with increased sensitivity.This thesis presents a sensitivity study for the dominant proton decay mode predicted by Supersymmetric GUTs, p → K+ν, utilizing machine learning approaches. Two methods are explored in this thesis: a Boosted Decision Tree (BDT) analysis and a Graphical Neural Network (GNN) analysis with NuGraph. A lifetime limit of 5.36 ± 0.69 × 1033 years for 400 kt-yrs is found using the BDT, while the GNN achieves a lifetime limit of 6.19±1.26×1033 years for 400-kt-yrs. The NuGraph result offers better sensitivity compared to the current limit set by Super-Kamiokande of 5.90 × 1033 while the BDT result offers a slightly lower sensitivity.Additionally, this thesis discusses cross-section work, a first-ever foray into proton decay and atmospheric neutrinos in a vertical drift (VD) DUNE detector, and extensive hardware contributions to the DUNE Far Detector (FD) 1 Module-0, ProtoDUNE-2, which serves as a testbed for the final detector design and installation.

Stokes, Tyler D. [Louisiana State U.] (ORCID:00000↗

THE XMM CLUSTER SURVEY: THE BUILD-UP OF STELLAR MASS IN BRIGHTEST CLUSTER GALAXIES AT HIGH REDSHIFT

We present deep J- and K{sub s} -band photometry of 20 high redshift galaxy clusters between z = 0.8 and1.5, 19 of which are observed with the MOIRCS instrument on the Subaru telescope. By using near-infrared light as a proxy for stellar mass we find the surprising result that the average stellar mass of Brightest Cluster Galaxies (BCGs) has remained constant at {approx}9 x 10{sup 11} M {sub sun} since z {approx} 1.5. We investigate the effect on this result of differing star formation histories generated by three well-known and independent stellar population codes and find it to be robust for reasonable, physically motivated choices of age and metallicity. By performing Monte Carlo simulations we find that the result is unaffected by any correlation between BCG mass and cluster mass in either the observed or model clusters. The large stellar masses imply that the assemblage of these galaxies took place at the same time as the initial burst of star formation. This result leads us to conclude that dry merging has had little effect on the average stellar mass of BCGs over the last 9-10 Gyr in stark contrast to the predictions of semi-analytic models, based on the hierarchical merging of dark matter halos, which predict a more protracted mass build-up over a Hubble time. However, we discuss that there is potential for reconciliation between observation and theory if there is a significant growth of material in the intracluster light over the same period.

79 ASTRONOMY AND ASTROPHYSICS↗

A Quantitative Study of Deep Learning Training on Heterogeneous Supercomputers

Deep learning (DL) has become a key technique for solving complex problems in scientific research and discovery. DL training for science is substantially challenging because it has to deal with massive quantities of multi-dimensional data. High- performance computing (HPC) supercomputers are increasingly being employed for meeting the exponentially growing demand for DL. Multiple GPUs and high-speed interconnect network are needed for supporting DL on HPC systems. However, the excessive use of GPUs without considering effective benefits leads to inefficient resource utilization of these expensive setups. In this paper, we conduct a quantitative analysis to gauge the efficacy of DL workloads on the latest HPC system and identify viability of next-generation DL-optimized heterogeneous supercomputers for enabling researchers to develop more efficient resource management and distributed DL middleware. We evaluate well- known DL models with large-scale datasets using the popular TensorFlow framework, and provide a thorough evaluation in- cluding scalability, accuracy, variability, storage resource, GPU- GPU/GPU-CPU data transfer, and GPU utilization. Our analysis reveals that the latest heterogeneous supercomputing cluster shows varying performance trend as compared to the existing literature for single- and multi-node training. To the best of our knowledge, this is the first work to conduct such a quantitative and comprehensive study of DL training on a supercomputing system with multiple GPUs.

Han, Jingoo↗

Using machine learning to predict antimicrobial MICs and associated genomic features for nontyphoidal Salmonella

Nontyphoidal Salmonella species are the leading bacterial cause of food-borne disease in the United States. Whole genome sequences and paired antimicrobial susceptibility data are available for Salmonella strains because of surveillance efforts from public health agencies. In this study, a collection of 5,278 nontyphoidal Salmonella genomes, collected over 15 years in the United States, were used to generate XGBoost-based machine learning models for predicting minimum inhibitory concentrations (MICs) for 15 antibiotics. The MIC prediction models have an overall average accuracy of 95% within ± 1 two-fold dilution step (confidence interval of 95-95%), an average very major error rate of 2.7% (confidence interval of 2.4-3.0%) and an average major error rate of 0.1% (confidence interval of 0.1-0.2%). The model predicts MICs with no a priori information about the underlying gene content or resistance phenotypes of the strains. By selecting diverse genomes for training sets, we show that highly accurate MIC prediction models can be generated with fewer than 500 genomes. We also show that our approach for predicting MICs is stable over time despite annual fluctuations in antimicrobial resistance gene content in the sampled genomes. Finally, using feature selection, we explore the important genomic regions identified by the models for predicting MICs. To date, this is one of the largest MIC modeling studies to be published. Furthermore, our strategy for developing whole genome sequence-based models for surveillance and clinical diagnostics can be readily applied to other important human pathogens.

59 BASIC BIOLOGICAL SCIENCES↗

Transverse Mode Dynamics of VCSELs Undergoing Current Modulation

Transverse mode dynamics of a 20-micron-diameter vertical-cavity surface-emitting laser (VCSEL) undergoing gain switching by deep current modulation is studied numerically. The direct current (dc) level is set slightly below threshold and is modulated by a large alternating current (ac). The resulting optical pulse train and transverse-mode patterns are obtained numerically. The ac frequency is varied from 2.5 GHz to 10 GHz, and the ac amplitude is varied from one-half to four times that of the dc level. At high modulation frequencies, a regular pulse train is not generated unless the ac amplitude is large enough. At all modulation frequencies, the transverse spatial profile switches from single-mode to multiple-mode pattern as the ac pumping level is increased. Optical pulse widths vary in the range 5-30 ps. with the pulse width decreasing when either the frequency is increased or the ac amplitude is decreased. The numerical modeling uses an approximation form of the semiconductor Maxwell-Bloch equations. Temporal evolution of the spatial profiles of the laser (and of carrier density) is determined without any assumptions about the type or number of modes. Keywords: VCSELs, current modulation, gain switching, transverse mode dynamics, computational modeling

Goorjian, Peter M.↗

The Pulsed Fission-Fusion (PUFF) Concept for Deep Space Exploration and Terrestrial Power Generation

This team is exploring a modified Z-pinch geometry as a propulsion system, imploding a liner of liquid lithium onto a pellet containing both fission and fusion fuel. The plasma resulting from the fission and fusion burn expands against a magnetic nozzle, for propulsion, or a magnetic confinement system, for terrestrial power generation. There is considerable synergy in the concept; the lithium acts as a temporary virtual cathode, and adds reaction mass for propulsion. Further, the lithium acts as a radiation shield against generated neutrons and gamma rays. Finally, the density profile of the column can be tailored using the lithium sheath. Recent theoretical and experimental developments (e.g. tailored density profile in the fuel injection, shear stabilization, and magnetic shear stabilization) have had great success in mitigating instabilities that have plagued previous fusion efforts. This paper will review the work in evaluating the pellet sizes and z-pinch conditions for optimal PuFF propulsion. Trades of pellet size and composition with z-pinch power levels and conditions for the tamper and lithium implosion are evaluated. Current models, both theoretical and computational, show that a z-pinch can ignite a small (~1 cm radius) fission-fusion target with significant yield. Comparison is made between pure fission and boosted fission targets. Performance is shown for crewed spacecraft for high speed Mars round trip missions and near interstellar robotic missions. The PuFF concept also offers a solution for terrestrial power production. PuFF can, with recycling of the effluent, achieve near 100% burnup of fission fuel, providing a very attractive power source with minimal waste. The small size of PuFF relative to today's plants enables a more distributed power network and less exposure to natural or man-made disruptions.

Adams, Robert↗

Hydrogen Storage for Flexible Fossil Fuel Power Generation: Integration of Underground Hydrogen Storage with Gas Turbine (Final Report)

As the nation continues to encourage, through market structures and financial incentives, the proliferation of intermittent renewable electricity, how to optimize the ever-changing electric grid and identify means to retain and improve resilience, while ensuring continued reductions in GHG emissions, will be critical. According to Bloomberg, wind & solar generated 10.5% of US electricity in 2020 and that percentage continues to grow. In support of expanding renewable energy use, and to address its intermittent nature, this project will develop the Hydrogen Storage for Flexible Fossil Fuel Power Generation platform that is dispatchable, reliable, repeatable and have the ability to produce zero or negative carbon power while interfacing with geology capable of CO2 and hydrogen storage. GTI Energy (GTIE) and team members Illinois State Geological Survey (ISGS), Mitsubishi Heavy Industries America (MHIA), Ameren Illinois, Hexagon Purus, and the Low Carbon Resources Initiative (LCRI) completed a Phase I Conceptual Study under contract DE-FE0032012 for Hydrogen Storage for Flexible Fossil Fuel Power Generation: Integration of Underground Hydrogen Storage with Gas Turbine. The Hydrogen Storage for Flexible Fossil Fuel Power Generation platform addresses the intermittent nature of the expanding use of Variable Renewable Energy (VRE) generation. The low cost of the electricity (COE) generated results in greater dispatch and more operation at higher power levels (higher efficiency), fewer short intervals, and fewer start/stop cycles. The reliable, resilient system can produce zero carbon power and store hydrogen. It will demonstrate hydrogen storage in geologic formations like those used in natural gas underground storage thus enabling large scale storage of hydrogen in sedimentary strata across the United States rather than in geographically restricted salt caverns. The Phase I study confirmed the system is feasible and generates power at lower cost than other low carbon approaches. The demonstration defines the pathway for broad commercial application and will accelerate the development of larger systems suitable for centralized utility scale electricity production. The study advanced the maturity of the H 2 storage-based system with flexible power generation by completing a Pre-FEED study (Phase II). The Pre-FEED focused on the selected Energy Farm on the University of Illinois Urbana-Champaign (UIUC) site that includes above ground and underground hydrogen storage, low-carbon hydrogen production (GTI’s Compact Hydrogen Generator, CHG) with underground CO2 sequestration, and a 40-MW class gas turbine. The Pre-FEED addressed the entire system and its interconnection to the natural gas and electric grid and mitigation of key risks, such as storage behavior, load-following, and system operation. During Phase 1 of the project, the team completed key tasks, which moved the entire demonstration project, specific components and approaches closer to commercialization. These Phase I Accomplishments include: Completing System Requirements Review; Completing System Layout and Modeling - Heat & Mass Balance and Process Flow Diagram; Completing modelling of 9 turbine performance cases; Evaluating rock strata for underground storage of hydrogen and sequestration of carbon dioxide; Completing initial modelling of underground storage of hydrogen and withdrawal with evaluation of loss and water production; Identifying roadable storage for above ground hydrogen storage; Identifying existing electrical infrastructure for receiving/delivering electricity; Identifying existing gas supply infrastructure for receiving natural gas; Document concept design/development plans in required reports. Conclusions: The 12-month Feasibility study in Phase I study was completed and confirmed the system is feasible and generates power at lower cost than other low carbon approaches and even lower cost than the reference NGCC plant without carbon capture when taking advantage of 45Q carbon credits. The study enabled the fidelity of the concept to be improved and allowed identification of the requirements for the system. Defining the individual system and component requirements was performed via the system requirements review with the whole team. These requirements were then incorporated into and iterated with our Heat & Mass Balance process model and process flow diagrams were generated to reflect the overall system. This information was then used to complete the TEA and show economic feasibility. Large scale non-salt geologic storage of hydrogen is an enabling technology for a hydrogen-fired turbine that can be retrofitted into large-scale electric generating units (EGU). Our demonstration will include 428 MWh or ~4 hours full load of hydrogen storage (above and underground). Carbon capture inherent to the CHG process can capture 90% CO 2 (with upgrades to >98%). This system provides a COE of 23% savings relative to an NGCC with a post combustion amine system. Our proposed storage system decouples carbon capture and hydrogen production from power production; therefore, we expect our proposed system’s efficiency and variable COE to be superior resulting in overall higher dispatch and reduced deep cycling. Our demonstration will be full to multi-day hydrogen storage and has the potential for longer (seasonal) duration commercially. The demonstration defines the pathway for broad commercial application and will accelerate the development of larger systems suitable for centralized utility scale electricity production.

03 NATURAL GAS↗

Comprehensive Neural Posterior Estimation for Galaxy-Galaxy Strong Lensing

We present a deep learning model based on neural posterior estimation (NPE) for comprehensive extraction of astrophysical parameters from galaxy-scale strong gravitational lenses. The unprecedentedly large amount of galaxy-scale strong lenses expected in future cosmological surveys (${\cal O}(10^5)$) promises to enable valuable statistical constraints in various studies ranging from galaxy formation to the nature of dark matter, but it also poses a significant challenge for traditional modelling pipelines. To this end, our automated model includes several new, state-of-the-art features and approaches leveraging the framework of simulation-based inference (SBI). We infer a total of 20 parameters describing the mass and light profiles of both lens and source galaxies, using simulated raw multi-band data modelled under noise and observing conditions expected by the Legacy Survey of Space and Time (LSST), with its summary statistics generated by a residual network. We examine the efficacy of multi-band data in extracting nearly 20 model parameters simultaneous from strong lensing images including lens light. Finally, We perform a comprehensive set of diagnostics for SBI models, evaluating the model's prediction accuracy, stability, and uncertainty quantification.

Zhao, Roy J. [Chicago U., KICP]↗