Search NASA⌕ Search

SEARCH · Search NASA

Results for “Exploratory Modeling”

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 559 records · Page 31

Big Data Analytics for Scanning Transmission Electron Microscopy Ptychography

Electron microscopy is undergoing a transition; from the model of producing only a few micrographs, through the current state where many images and spectra can be digitally recorded, to a new mode where very large volumes of data (movies, ptychographic and multi-dimensional series) can be rapidly obtained. In this paper, we discuss the application of so-called “big-data” methods to high dimensional microscopy data, using unsupervised multivariate statistical techniques, in order to explore salient image features in a specific example of BiFeO 3 domains. Remarkably, k-means clustering reveals domain differentiation despite the fact that the algorithm is purely statistical in nature and does not require any prior information regarding the material, any coexisting phases, or any differentiating structures. While this is a somewhat trivial case, this example signifies the extraction of useful physical and structural information without any prior bias regarding the sample or the instrumental modality. Further interpretation of these types of results may still require human intervention. Finally, however, the open nature of this algorithm and its wide availability, enable broad collaborations and exploratory work necessary to enable efficient data analysis in electron microscopy.

36 MATERIALS SCIENCE↗

Connected Lighting Systems Efficiency Study: PoE Cable Energy Losses, Part 2

DOE published a report in November 2017 (herein referred to as “Part 1”) summarizing the results of an exploratory study investigating power losses in Ethernet cables used between PoE switches and luminaires in PoE connected lighting systems. Testing was conducted at the Pacific Northwest National Laboratory (PNNL) Connected Lighting Test Bed (CLTB) in September 2017. The results were analyzed to explore the impact of cable selection on PoE lighting system energy efficiency, as well as the effectiveness of guidelines published by the American National Standards Institute (ANSI) C137 Lighting Systems Committee in 2017. The guidance offered in ANSI C137.3-2017 was found to be effective in limiting cable energy losses to 5% in PoE lighting applications, provided that the average cable length on a project does not exceed 50 m. This Part 2 report summarizes the results of a continued investigation of the power losses in Ethernet cables used between PoE switches and luminaires in PoE connected lighting systems. Testing was conducted at the PNNL CLTB in July-August 2018. A test setup comprising a PoE switch, a set of luminaires, and a reference meter was again used to test multiple cable models of varying design. The results were analyzed to explore the impact of cable selection and installation practices (e.g., cable bending, cable bundling, and installation in conduit) on PoE lighting system energy efficiency, as well as the effectiveness of the ANSI C137.3 guidelines. Notably, three cables—two of which were shielded—were previously excluded from the Part 1 study due to compatibility issues. All three cables were included in this Part 2 study, along with a fourth model cable introduced in Part 2. In addition, whereas no patch cords were used in Part 1, patch cords were used in some Part 2 testing. The ANSI C137.3 guidance was again shown to effectively limit power losses to less than 5% in the cables tested (varying in AWG, Category, shielding, fire rating, and manufacturer), and AWG again proved to be the most important cable design variable for energy performance, as expected. However, these findings should not be construed as being representative of all cable models and installation practices. For example, energy losses would be greater when cables are connected to patch cords, bundled in conduit, and loaded with powered devices (PDs) approaching 100 W input power. The other key finding was that with 44 W luminaires as PDs and room ambient temperatures below 30°C, cable energy losses were not substantially increased by cable bending or bundling in uninsulated conduit. However, environments with higher ambient temperatures will have greater energy losses due to increased conductor DCR. In addition, product selection and installation practices will have increased importance as PDs approaching 100 W input power—conveyed by a single Ethernet cable—are introduced following publication of the forthcoming Institute of Electrical and Electronics Engineers (IEEE) Standard 802.3bt. The report also provides recommendations and next steps, which include plans for additional PoE cable energy losses testing with higher-power PDs to more fully explore the potential of PoE lighting systems.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

From microbial diversity to functional potential using dimensionality reduction

The high dimensionality of microbial diversity data from ‘omics observations can be reduced using Machine Learning, with many recent studies showcasing ML utility for exploratory ecological feature finding and process prediction. Here, we compare the Self Organizing Map (SOM) dimensionality reduction method to the well-documented sample-based Principal Coordinate Analysis (PCoA) and taxa-based Weighted Gene Correlation Network Analysis (WGCNA) using near daily 16S rRNA gene amplicon sequencing data from the 2019 to 2020 MOSAiC International Arctic Drift Expedition. We then map k-means clustering outputs from each method to available metagenomes, extracting functionally distinct seasonal microbial ecotypes in the surface Arctic Ocean. Our results indicate the SOM method better represented expected seasonal transitions and identified a greater number of metabolically distinct functional groups than the more traditional PCoA ordination. Ultimately, we identified four community ecotypes with distinct taxonomic and functional cut-offs driven by seasonality, water mass, and substrate turnover, highlighting the importance of succession in functional diversity for the central Arctic Ocean. These results reinforce ML dimensionality reduction as a meaningful translator in the mining of historical amplicon datasets to address modern mechanistic questions and potentially provide ’omics informed ecotype diversity to leverage in mechanistic biogeochemical models.

Arctic Ocean↗

Three-Flavor Neutrino Oscillations with MINOS+

This dissertation presents the analysis of the last two years of data from the MINOS+ long-baseline neutrino oscillation experiment. The analysis explores the data above the first oscillation maximum for neutrinos in the standard three-flavor neutrino oscillation model, adding significantly more neutrino events and constraining the model with increased precision. The analysis of the high-energy region of the MINOS+ data set, where previously there has been limited study, can help constrain alternative models or possibly show evidence of new phenomena. The predecessor to MINOS+, the MINOS experiment, measured $\nu_\mu$ disappearance and $\nu_{e}$ appearance using Fermilab's NuMI $\nu_\mu$ beam from 2005 to 2012. During this period the neutrino beam's energy spectrum was focused to peak near the first oscillation maximum. In addition to measuring accelerator beam neutrinos, the MINOS Far Detector collected a sample of atmospheric neutrinos from 2003 to 2011. With the se two samples, MINOS measured the atmospheric mass splitting, $\Delta{m}^2_{32}$, at the 5\% level and the value of $\sin^2\theta_{23}$ at the 15\% level. The data also constrained the CP violating parameter $\delta_{\rm CP}$. The MINOS+ experiment exposed the MINOS detectors to a neutrino beam peaked at energies above the oscillation maximum from 2013 to 2016. With this higher energy neutrino beam, MINOS+ measures the shape of the $\nu_\mu$ survival probability away from the oscillation maximum with unprecedented precision. Measuring the shape of the oscillation probability as a function of neutrino energy is an essential test of the three-flavor oscillation model. At these higher energies where the standard oscillation probability decreases, the neutrino energy spectrum is sensitive to potential perturbations from mixing with additional sterile neutrinos or non-standard neutrino interactions. This analysis of the complete data set from MINOS+ finds no significant deviatio ns from the three-flavor oscillation probability in the ener! gy region of 4 to 10\,GeV covered by the neutrino beam. This provides increased confidence in the three-flavor model in this region and provides new constraints of the atmospheric oscillation parameters, $\Delta{m}^2_{32}$ and $\sin^2\theta_{23}$, by combining the MINOS and MINOS+ data sets. The hierarchy of the neutrino masses is related to the sign of $\Delta{m}^2_{32}$. When $\Delta{m}^2_{32} > 0$, the mass hierarchy is classified as normal. When $\Delta{m}^2_{32} < 0$, the mass hierarchy is classified as inverted. This analysis cannot rule out either of the mass hierarchies. For the normal mass hierarchy, the combined analysis measures $\lvert{\Delta{m}^2_{32}}\rvert = 2.41\pm 0.09 \times 10^{-3}\,{\mathrm{eV}^2}$ at 68\% C.L. and $\sin^2\theta_{23} = 0.42_{-0.06}^{+0.23}$ at 90\% C.L. For the inverted mass hierarchy, this analysis measures $\lvert{\Delta{m}^2_{32}}\rvert = 2.47_{-0.10}^{+0.08} \times 10^{-3}\,{\mathrm{eV}^2}$ at 68\% C.L. and $\sin^2\theta_{23} = 0.42_{ -0.06}^{+0.23}$ at 90\% C.L. In addition to the analysis and measurement of the atmospheric oscillation parameters using MINOS+, this work updates the constraint on $\delta_{\rm CP}$ from MINOS and performs an exploratory combination using $\nu_\mu$ disappearance data from the NOvA experiment. The NOvA experiment uses the NuMI beam in an off-axis strategy that results in a narrow flux of muon neutrinos at the first oscillation maximum and allows precision measurements of the three-flavor parameters. Since NOvA and MINOS+ shared the same neutrino beam, their data can be combined taking advantage of common beam related uncertainties. Combined, the data from the MINOS, MINOS+, and NOvA experiments precisely map $\nu_\mu$ oscillation probabilities using the same neutrino beam in a way that could not be attempted before.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Compendium of NASA data base for the global tropospheric experiment's Pacific Exploratory Mission West-B (PEM West-B)

This compendium describes aircraft data that are available from NASA's Pacific Exploratory Mission West-B (PEM West-B). PEM West is a component of the International Global Atmospheric Chemistry's (IGAC) East Asia/North Pacific Regional Study (APARE) project. Objectives of PEM West are to investigate the atmospheric chemistry of ozone over the northwest Pacific -- natural budgets and the impact of anthropogenic/continental sources; and to investigate sulfur chemistry -- continental and marine sulfur sources. The PEM West program encompassed two expeditions. PEM West-A was conducted in September 1991 during which the predominance of tropospheric air was from mid-Pacific (marine) regions, but (at times) was modified by Asian outflow. PEM West-B was conducted during February 1994, a period characterized by maximum Asian outflow. Results from PEM West-A and B are public domain. PEM West-A data are summarized in NASA TM 109177 (published February 1995). Flight experiments were based at Guam, Hong Kong, and Japan. This document provides a representation of NASA DC-8 aircraft data that are available from NASA Langley's Distributed Active Archive Center (DAAC). The DAAC includes numerous other data such as meteorological and modeling products, results from surface studies, satellite observations, and sonde releases.

Gregory, Gerald L.↗

Compendium of NASA data base for the Global Tropospheric Experiment's Pacific Exploratory Mission West-A (PEM West-A)

This compendium describes aircraft data that are available from NASA's Pacific Exploratory Mission West-A (PEM West-A). PEM West is a component of the International Global Atmospheric Chemistry's (IGAC) East Asia/North Pacific Regional Study (APARE) project. The PEM- West program encompassed two expeditions to study contrasting meteorological regimes in the Pacific. Objectives of PEM West are to investigate the atmospheric chemistry of ozone over the northwest Pacific -- natural budgets and the impact of anthropogenic sources; and to investigate sulfur chemistry -- continental versus marine sulfur sources. PEM West-A was conducted in September 1991 during which the predominance of tropospheric air is from the mid-Pacific (marine) regions, but (at times) is modified/mixed with Asian continental outflow. PEM West-B was conducted during February 1994, a period characterized by maximum continental outflow. PEM-B data (not included) will become public domain during the Summer of 1995. PEM West-A flight experiments were based at Japan, Hong Kong, and Guam. This document provides a representation of NASA DC-8 aircraft data that are available from NASA Langley's Distributed Active Archive Center (DAAC), which include numerous data such as meteorological observations, modeling products, results from surface studies, satellite observations, and sonde releases.

Gregory, G. L.↗

Inverse reinforcement learning control for building energy management

Reinforcement learning (RL) based control is widely considered a promising approach in building automation and control as it has demonstrated the potential to deal with complex objectives in adjacent domains like robotics, autonomous vehicles, gaming applications, and advertisement recommendations. When applied to any environment, model-free RL learns to improve its control performance over time without requiring a control model, by receiving and then analyzing feedback from the building environment after each control action. Operational objectives are becoming increasingly complex through the simultaneous consideration of thermal comfort, carbon emissions, grid services, and indoor air quality. In this context, conventional rule-based control approaches are proving sub-optimal, mostly heuristic, and inadequate. The model-free and self learning nature of RL appears promising and attractive as it may address the scalability issues associated with advanced control approaches. However, it suffers from long training times and unstable control behavior during the early stages of its learning process, which makes it unsuitable to be applied directly to buildings. This paper addresses these issues using an inverse reinforcement learning approach (IRL), a technique utilized to learn the objective of a controller agent which is considered an expert in its respective domain. Here, we consider a rule-based control as the expert demonstrator. IRL is different from a direct imitation (i.e., direct mapping of states to actions) of control actions as it tries to find the underlying intent of an expert's policy, providing the controller with a better-generalized policy for unseen states or environments with slightly different dynamics. This approach propels the RL controller's policy to levels similar to or better than that of a rule-based policy before it starts learning by interacting with the building. This makes RL for building energy management applications more practical as it prevents the erratic and exploratory behavior in the initial training period, simultaneously speeding up the learning process when compared to applying an untrained RL agent directly to a building environment.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Impacts of air pollutants from fire and non-fire emissions on the regional air quality in Southeast Asia

Severe haze events in Southeast Asia caused by particulate pollution have become more intense and frequent in recent years, degrading air quality, threatening human health, and interrupting economic and societal activities. Widespread biomass burning activities are a major source of severe haze events in Southeast Asia. On the other hand, particulate pollutants from human activities other than biomass burning also play an important role in degrading air quality in Southeast Asia. In this study, numerical simulations have been conducted using the Weather Research and Forecasting (WRF) model coupled with a chemistry component (WRF-Chem) to quantitatively examine the contributions of aerosols emitted from fire (i.e., biomass burning) versus non-fire (including fossil fuel combustion, road and industrial dust, land use, and land change, etc.) sources to the degradation of air quality and visibility over Southeast Asia. These simulations cover a time period from 2002 to 2008 and were respectively driven by emissions from: (a) fossil fuel burning only, (b) biomass burning only, and (c) both fossil fuel and biomass burning. Across ASEAN 50 cities, these model results reveal that 39&thinsp;% of observed low visibility days can be explained by either fossil fuel burning or biomass burning emissions alone, a further 20&thinsp;% by fossil fuel burning alone, a further 8&thinsp;% by biomass burning alone, and a further 5&thinsp;% by a combination of fossil fuel burning and biomass burning. The remaining 28&thinsp;% of observed low visibility days remain unexplained, likely due to emissions sources that have not been accounted for. Further analysis of 24-hr PM 2.5 Air Quality Index (AQI) indicates that comparing to the simulated result of the case with stand-alone non-fire emissions, the case with coexisting fire and non-fire PM 2.5 can substantially increase the chance of AQI being in the moderate or unhealthy pollution level from 23&thinsp;% to 34&thinsp;%. The premature mortality among major Southeast Asian cities due to degradation of air quality by particulate pollutants is estimated to increase from ~&thinsp;4110 per year in 2002 to ~&thinsp;6540 per year in 2008. In addition, we demonstrate the importance of certain missing non-fire anthropogenic aerosol sources including anthropogenic fugitive and industrial dusts in causing urban air quality degradation. An exploratory experiment of using machine learning algorithms to forecasting the occurrence of haze events in Singapore is also demonstrated in this study. All these results suggest that besides minimizing biomass burning activities, an effective air pollution mitigation policy for Southeast Asia needs to consider controlling emissions from non-fire anthropogenic sources.

Lee, Hsiang-He (ORCID:0000000251407324)↗

Interface Consistency: Phase I Results & Phase II Status

Future exploration missions will rely on designing and developing vehicles and complex systems from within NASA and through multiple external commercial partners to meet mission goals. Despite existing consistency-related agency requirements, NASA’s approach to commercial spaceflight development encourages providers’ flexibility and innovation. This strategy is resulting in significant design diversity across Artemis vehicles. Design best practices and guidelines champion interface consistency to promote mental model development and knowledge transfer. However, research investigating the benefits of consistency is mixed, and little is known about its role in complex systems. Determining the level of risk that system diversity presents is difficult, as there is no established method for quantifying the degree of consistency within and across interfaces, nor is there information about the differential impacts of different types of inconsistency. Phase I of this project (Characterization and Measurement) served as a starting point to better understand the construct of consistency, its application, and the range of studies and methods for measuring it. The project team created a taxonomy of consistency to apply to interfaces as a framework to guide the development of tools to assess intersystem consistency. Checklist and cognitive walkthrough methods were developed for use by human factors (HF) and human-computer interaction (HCI) experts. The Intersystem Consistency Scale (ICS) was developed for interface evaluations with crew. A pilot study evaluated the methods’ ability to distinguish differences between Artemis-like prototype pairs exhibiting either high or low design consistency. In addition, click errors and time on task were collected within the ICS (crew-like) group. Results from our exploratory analysis and lessons learned from the pilot study will be discussed. The project team will also present the status of Phase II (Risk Assessment, Standards and Guidelines). This includes incorporating feedback to redesign the assessment tools, and inputs from displays and training Subject Matter Experts to update tasks and prototype designs. The team will present the risk assessment study design to identify the types and levels of inconsistency that pose the greatest risk to performance. Plans to apply these results toward agency standards and guideline recommendations will also be discussed.

Human-Computer Interaction↗

MASSES, RADII, AND CLOUD PROPERTIES OF THE HR 8799 PLANETS

The near-infrared colors of the planets directly imaged around the A star HR 8799 are much redder than most field brown dwarfs of the same effective temperature. Previous theoretical studies of these objects have concluded that the atmospheres of planets b, c, and d are unusually cloudy or have unusual cloud properties. Some studies have also found that the inferred radii of some or all of the planets disagree with expectations of standard giant planet evolution models. Here, we compare the available data to the predictions of our own set of atmospheric and evolution models that have been extensively tested against observations of field L and T dwarfs, including the reddest L dwarfs. Unlike some previous studies, we require mutually consistent choices for effective temperature, gravity, cloud properties, and planetary radius. This procedure thus yields plausible values for the masses, effective temperatures, and cloud properties of all three planets. We find that the cloud properties of the HR 8799 planets are not unusual but rather follow previously recognized trends, including a gravity dependence on the temperature of the L to T spectral transition-some reasons for which we discuss. We find that the inferred mass of planet b is highly sensitive to whether or not we include the H- and the K-band spectrum in our analysis. Solutions for planets c and d are consistent with the generally accepted constraints on the age of the primary star and orbital dynamics. We also confirm that, like in L and T dwarfs and solar system giant planets, non-equilibrium chemistry driven by atmospheric mixing is also important for these objects. Given the preponderance of data suggesting that the L to T spectral type transition is gravity dependent, we present an exploratory evolution calculation that accounts for this effect. Finally we recompute the bolometric luminosity of all three planets.

79 ASTRONOMY AND ASTROPHYSICS↗

Roadmap to an Engineering-Scale Nuclear Fuel Performance & Safety Code

Developing new fuels and qualifying them for large-scale deployment in power reactors is a lengthy and expensive process, typically spanning a period of two decades from concept to licensing. Nuclear fuel designers serve an indispensable role in the process, at the initial exploratory phase as well as in analysis of the testing results. In recent years fuel performance capabilities based on first principles have been playing more of a role in what has traditionally been an empirically dominated process. Nonetheless, nuclear fuel behavior is based on the interaction of multiple complex phenomena, and recent evolutionary approaches are being applied more on a phenomenon-by-phenomenon basis, targeting localized problems, as opposed to a systematic approach based on a fundamental understanding of all interacting parameters. Advanced nuclear fuels are generally more complex, and less understood, than the traditional fuels used in existing reactors (ceramic UO{sub 2} with burnable poisons and other minor additives). The added challenges are primarily caused by a less complete empirical database and, in the case of recycled fuel, the inherent variability in fuel compositions. It is clear that using the traditional approach to develop and qualify fuels over the entire range of variables pertinent to the U.S. Department of Energy (DOE) Office of Nuclear Energy on a timely basis with available funds would be very challenging, if not impossible. As a result the DOE Office of Nuclear Energy has launched the Nuclear Energy Advanced Modeling and Simulation (NEAMS) approach to revolutionize fuel development. This new approach is predicated upon transferring the recent advances in computational sciences and computer technologies into the fuel development program. The effort will couple computational science with recent advances in the fundamental understanding of physical phenomena through ab initio modeling and targeted phenomenological testing to leapfrog many fuel-development activities. Realizing the full benefits of this approach will likely take some time. However, it is important that the developmental activities for modeling and simulation be tightly coupled with the experimental activities to maximize feedback effects and accelerate both the experimental and analytical elements of the program toward a common objective. The close integration of modeling and simulation and experimental activities is key to developing a useful fuel performance simulation capability, providing a validated design and analysis tool, and understanding the uncertainties within the models and design process. The efforts of this project are integrally connected to the Transmutation Fuels Campaign (TFC), which maintains as a primary objective to formulate, fabricate, and qualify a transuranic-based fuel with added minor actinides for use in future fast reactors. Additional details of the TFC scope can be found in the Transmutation Fuels Campaign Execution Plan. This project is an integral component of the TFC modeling and simulation effort, and this multiyear plan borrowed liberally from the Transmutation Fuels Campaign Modeling and Simulation Roadmap. This document provides the multiyear staged development plan to develop a continuum-level Integrated Performance and Safety Code (IPSC) to predict the behavior of the fuel and cladding during normal reactor operations and anticipated transients up to the point of clad breach.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Directionally Sensitive Silicon Radiation Sensor (VCELL)

Sensors are a mission critical element in many NASA programs and require some very unique properties such as small size, low power, high reliability, low weight. Low cost sensors offer the possibility of technology transfer to the public domain for commercial applications. One sensor application that is important to many NASA programs is the ability to point at a radiation source, such as the sun. Such sensors may be an integral part of the guidance and control systems in space platforms and in remote exploratory vehicles. Sun/solar pointing is also important for ground-based systems such as solar arrays. These systems are not required to be small and lightweight. However, if a sensor with a sun pointing capability was developed that is very small, rugged, lightweight and at the same time low cost, it certainly could be used in existing and perhaps many new ground based applications. The objective of the VCELL (Directionally Sensitive Silicon Radiation Sensor) research is to develop a new and very unique silicon based directionally sensitive radiation sensor which can be fabricated using conventional monolithic IC technologies and which will meet the above requirements. The proposed sensor is a novel silicon chip that is directionally sensitive to incident radiation, providing azimuth and elevation information on the incident radiation. The resulting sensor chip will be appropriate for integration into a silicon IC or useful in a hybrid structure to be interfaced with a standard IEEE 1451 bus interface IC to create an Intelligent Sensor. It is presently estimated that it will require about three man-years of effort to complete the VCELL research and development. This includes the optical, electrical, mechanical and silicon fabrication and testing as well as computer simulations and theoretical analysis and modeling including testing in simulated space environments. This report summarizes the sensor research completed this summer as part of the Summer Faculty Fellowship Program. The primary effort was focused on activity necessary to fabricate prototype sensor.

Koy B. Cook↗

An Exploratory Exercise in Taguchi Analysis of Design Parameters: Application to a Shuttle-to-space Station Automated Approach Control System

The chief goals of the summer project have been twofold - first, for my host group and myself to learn as much of the working details of Taguchi analysis as possible in the time allotted, and, secondly, to apply the methodology to a design problem with the intention of establishing a preliminary set of near-optimal (in the sense of producing a desired response) design parameter values from among a large number of candidate factor combinations. The selected problem is concerned with determining design factor settings for an automated approach program which is to have the capability of guiding the Shuttle into the docking port of the Space Station under controlled conditions so as to meet and/or optimize certain target criteria. The candidate design parameters under study were glide path (i.e., approach) angle, path intercept and approach gains, and minimum impulse bit mode (a parameter which defines how Shuttle jets shall be fired). Several performance criteria were of concern: terminal relative velocity at the instant the two spacecraft are mated; docking offset; number of Shuttle jet firings in certain specified directions (of interest due to possible plume impingement on the Station's solar arrays), and total RCS (a measure of the energy expended in performing the approach/docking maneuver). In the material discussed here, we have focused on single performance criteria - total RCS. An analysis of the possibility of employing a multiobjective function composed of a weighted sum of the various individual criteria has been undertaken, but is, at this writing, incomplete. Results from the Taguchi statistical analysis indicate that only three of the original four posited factors are significant in affecting RCS response. A comparison of model simulation output (via Monte Carlo) with predictions based on estimated factor effects inferred through the Taguchi experiment array data suggested acceptable or close agreement between the two except at the predicted optimum point, where a difference outside a rule-of-thumb bound was observed. We have concluded that there is most likely an interaction effect not provided for in the original orthogonal array selected as the basis for our experimental design. However, we feel that the data indicates that this interaction is a mild one and that inclusion of its effect will not alter the location of the optimum.

Deal, Don E.↗

Single-Fuel Reactivity Controlled Compression Ignition Combustion Enabled by Onboard Fuel Reformation (Final Report)

Reactivity-Controlled Compression Ignition (RCCI) is a promising new advanced, low temperature combustion mode that offers efficiencies equal to or greater than conventional diesel combustion, but significantly lower soot and NOx emissions. Additionally, RCCI has enhanced controllability over the combustion process compared to some of the other advanced combustion modes. Therefore, RCCI combustion has the potential to be of great benefit to the public by increasing the efficiency and reducing the harmful pollutant emissions from transportation. However, RCCI combustion achieves this favorable level of performance through the use of two distinct fuels (e.g., diesel and gasoline, diesel and natural gas, or diesel and ethanol). In a vehicle application, RCCI therefore requires two different fuel tanks and fuel injection systems, which presents a significant drawback to the commercialization RCCI. This project proposed a new concept to enable RCCI combustion from a single parent fuel stored onboard the vehicle in a single fuel tank. Specifically, the parent fuel would be stored in the fuel tank and one branch of the fuel system would deliver the parent fuel, unaltered, to the cylinder, while a second branch of the fuel system would route the parent fuel through a fuel reformer to chemically react the fuel into a gaseous fuel mixture, called “reformate”, that has different properties from the original parent fuel. In this way, there can be two fuels at the engine that have different properties, but only one parent fuel needs to be stored and refilled in the vehicle. This initial, exploratory investigation into the feasibility of this promising new concept included reacting three potential parent fuels (diesel, gasoline, and natural gas) in a fuel reformer and determining the composition of the gaseous reformate fuel that exited the fuel reformer. There are different types of reformation processes, but since this project was a proof-of-concept investigation into the potential of Single-Fuel RCCI and the combustion process in the cylinder, a catalytic partial oxidation (CPOX) reformation process was selected, which is more straightforward to perform and control, and would potentially be easier to implement in a vehicle application compared to other reformation processes. Once the three parent fuels were reformed to varying degrees and their reformate compositions were determined, their autoignition tendencies were experimentally determined by comparing advanced combustion results of the reformate mixtures with known reference fuels. This showed that diesel fuel and its reformate fuel mixture exhibited the largest reactivity separation, which is necessary to enable RCCI combustion in the cylinder. Using diesel fuel as the parent fuel, experimental testing was conducted to provide a better understanding of the operational considerations, benefits, and drawbacks of Single-Fuel RCCI with onboard fuel reformation. The experimental testing was accompanied by computational fluid dynamics and system-level modeling to help complement the experiments and further expand the understanding. The main conclusions are that Single-Fuel RCCI using diesel fuel as the direct injected, high reactivity fuel and diesel’s gaseous reformate mixture as the premixed, low reactivity fuel is viable. The efficiencies were equal to, or sometimes greater than, RCCI with conventional fuel pairs, and the emissions were comparable as well. A key consideration is that gaseous reformate fuel contains a large amount of diluents in the fuel which displace air entering the cylinder and act as EGR inside the cylinder. Additionally, the reactivity separation between diesel and its reformate fuel mixture is larger than diesel-gasoline, but less than diesel-natural gas, and the combustion process therefore responded in a manner that was between the two conventional fuel pairs. Finally, although achievable in the cylinder, there are practical drawbacks or other disadvantages of Single-Fuel RCCI with onboard fuel reformation. First, this project used a CPOX reformation process which is exothermic, meaning that a portion of the fuel’s energy was released in the reformer where it could not be converted into useful work. This constituted an efficiency penalty of the system of on the order of 10% or 4 percentage points. Future research should therefore focus on alternate reforming processes including steam-methane reforming or autothermal reforming where there might not be any energy released in the reformer, or there could even be an energy gain in the reformer (through an endothermic process with waste heat recovery). Practical drawbacks that exist, but were beyond the scope of this preliminary proof-of-concept research project, are the added cost of the fuel reformer, which is a catalyst made of precious metals, transient response of the system, and the lifetime of the fuel reformer.

33 ADVANCED PROPULSION SYSTEMS↗

Data-centric framework for crystal structure identification in atomistic simulations using machine learning

Atomic-level modeling performed at large scales enables the investigation of mesoscale materials properties with atom-by-atom resolution. The spatial complexity of such cross-scale simulations renders them unsuitable for simple human visual inspection. Instead, specialized structure characterization techniques are required to aid interpretation. These have historically been challenging to construct, requiring significant intuition and effort. Here we propose an alternative framework for a fundamental structural characterization task: classifying atoms according to the crystal structure to which they belong. Our approach is data-centric and favors the employment of Machine Learning over heuristic rules of classification. A group of data-science tools and simple local descriptors of atomic structure are employed together with an efficient synthetic training set. We also introduce the first standard and publicly available benchmark data set for evaluation of algorithms for crystal-structure classification. Further, it is demonstrated that our data-centric framework outperforms all of the most popular heuristic methods—especially at high temperatures when lattices are the most distorted—while introducing a systematic route for generalization to new crystal structures. Moreover, through the use of outlier detection algorithms our approach is capable of discerning between amorphous atomic motifs (i.e., noncrystalline phases) and unknown crystal structures, making it uniquely suited for exploratory materials synthesis simulations.

36 MATERIALS SCIENCE↗

Complex Disorder in Type-I Clathrates: Synthesis and Structural Characterization of A8GaxSn46−x (A = Rb, Cs; 6.9 < x < 7.5)

Exploratory studies in the systems Rb–Ga–Sn and Cs–Ga–Sn yielded the cubic type-I clathrates with refined compositions Rb8GaxSn46−x and Cs8GaxSn46−x (6.9 < x < 7.5). Nearly single-phase materials with good crystallinity were obtained from stoichiometric reactions of the elements. The structures were characterized by means of single-crystal X-ray diffraction methods. Both Rb8GaxSn46−x and Cs8GaxSn46−x represents cases, where a Group 13 element randomly substitutes a Group 14 element in the structure. The extent of Ga/Sn mixing is apparently governed by the drive of the system to achieve an optimal valence electron count, and hence, Rb8GaxSn46−x and Cs8GaxSn46−x (x ≈ 8) can be regarded as Zintl phases. This notion is supported by structure refinements on a multitude of single-crystal X-ray diffraction data, which also confirm that both types of cages in the cubic type-I structure are fully occupied by Rb and Cs atoms. The open-framework, comprised of 46 nodes per formula unit, adapts to the incorporation of nearly eight Ga atoms within the matrix of Sn, whereby small, short-range distortions result. The exact nature of these effects is still unclear, as so far, the structural variations could only be modeled as both positional and occupational disorder at one of three framework sites. Since vacancies in the structures of the binary type-I clathrates A8Sn46−x☐x (A = Rb, Cs; ☐ = missing Sn atom) are also known to cause local distortions, the latter were also synthesized with the same protocols used for the synthesis of A8GaxSn46−x and structurally re-analyzed. The results from the latter studies confirm that homogeneity issues abound, and that the final structures/compositions are an intricate function of the experimental conditions.

36 MATERIALS SCIENCE↗

Exploring Applications of Machine Learning for Wildfire Monitoring and Detection using Unmanned Aerial Vehicles

Wildfires are increasing in frequency and severity around the world, including the United States. The losses caused by wildfires could be mitigated if high-risk areas, hotspots, and flare-ups could be monitored continuously, such as through the use of Unmanned Aerial Vehicles (UAVs). This paper documents exploratory efforts using machine learning to determine efficient flight paths for UAVs and to detect wildfires using image classification. On path planning, three machine learning techniques—Genetic Algorithm, Simulated Annealing, and Dynamic Programming—were explored. Genetic Algorithm was found to be an effective approach for path planning for wildfire monitoring and surveillance by UAVs. For a scenario of 25 locations in a circular arrangement, the algorithm was able to return the optimal path. The accuracy and execution time was found to be sensitive to the algorithm hyperparameters selected, which was especially evident in scenarios with hundreds or thousands of locations. Simulated Annealing was also found to be an effective approach for UAV path planning, with a major benefit of avoiding getting trapped in local minima and being straightforward to implement. Like Genetic Algorithm, the performance of Simulated Annealing was also found to be sensitive to the algorithm hyperparameters selected. By comparison, Dynamic Programming guarantees optimality for any number of locations, but it was found to be less practical in terms of execution time for scenarios with more than about a couple dozen locations. On wildfire detection, image classification using deep learning with a convolutional neural network was explored. Transfer learning was found to be a useful technique to efficiently train deep learning models. Also, it was determined that GPU processing can increase training speed by an order of magnitude, which enables significantly faster development. For a validation test set of 500 images, there were only two false negatives and zero false positives. These results demonstrate that detecting wildfires in static cameras using machine learning is feasible and establish a baseline for using images captured by UAVs in flight for wildfire detection.

Wildfire management↗

Algorithms and Libraries

This exploratory study initiated our inquiry into algorithms and applications that would benefit by latency tolerant approach to algorithm building, including the construction of new algorithms where appropriate. In a multithreaded execution, when a processor reaches a point where remote memory access is necessary, the request is sent out on the network and a context--switch occurs to a new thread of computation. This effectively masks a long and unpredictable latency due to remote loads, thereby providing tolerance to remote access latency. We began to develop standards to profile various algorithm and application parameters, such as the degree of parallelism, granularity, precision, instruction set mix, interprocessor communication, latency etc. These tools will continue to develop and evolve as the Information Power Grid environment matures. To provide a richer context for this research, the project also focused on issues of fault-tolerance and computation migration of numerical algorithms and software. During the initial phase we tried to increase our understanding of the bottlenecks in single processor performance. Our work began by developing an approach for the automatic generation and optimization of numerical software for processors with deep memory hierarchies and pipelined functional units. Based on the results we achieved in this study we are planning to study other architectures of interest, including development of cost models, and developing code generators appropriate to these architectures.

Dongarra, Jack↗