Search NASA⌕ Search

SEARCH · Search NASA

Results for “Tesla”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
NASA NTRS2008

Magnetohydrodynamic Augmented Propulsion Experiment

Over the past several years, efforts have been under way to design and develop an operationally flexible research facility for investigating the use of cross-field MHD accelerators as a potential thrust augmentation device for thermal propulsion systems. The baseline configuration for this high-power experimental facility utilizes a 1.5-MWe multi-gas arc-heater as a thermal driver for a 2-MWe MHD accelerator, which resides in a large-bore 2-tesla electromagnet. A preliminary design study using NaK seeded nitrogen as the working fluid led to an externally diagonalized segmented MHD channel configuration based on an expendable heat-sink design concept. The current status report includes a review of engineering/design work and performance optimization analyses and summarizes component hardware fabrication and development efforts, preliminary testing results, and recent progress toward full-up assembly and testing

Litchford, Ron J.↗
NASA NTRS2008

Electroconductive PET/SWNT Films by Solution Casting

The market for electrically conductive polymers is rapidly growing, and an emerging pathway for attaining these materials is via polymer-carbon nanotube (CNT) nanocomposites, because of the superior properties of CNTs. Due to their excellent electrical properties and anisotropic magnetic susceptibility, we expect CNTs could be easily aligned to maximize their effectiveness in imparting electrical conductivity to the polymer matrix. Single-walled carbon nanotubes (SWNT) were dispersed in a polyethylene terephthalate (PET) matrix by solution blending then cast onto a glass substrate to create thin, flexible films. Various SWNT loading concentrations were implemented (0.5, 1.0, and 3.0 wt.%) to study the effect of additive density. The processing method was repeated to produce films in the presence of magnetic fields (3 and 9.4 Tesla). The SWNTs showed a high susceptibility to the magnetic field and were effectively aligned in the PET matrix. The alignment was characterized with Raman spectroscopy. Impedance spectroscopy was utilized to study the electrical behavior of the films. Concentration and dispersion seemed to play very important roles in improving electrical conductivity, while alignment played a secondary and less significant role. The most interesting result proved to be the effect of a magnetic field during processing. It appears that a magnetic field may improve dispersion of unmodified SWNTs, which seems to be more important than alignment. It was concluded that SWNTs offer a good option as conductive, nucleating filler for electroconductive polymer applications, and the utilization of a magnetic field may prove to be a novel method for CNT dispersion that could lead to improved nanocomposite materials.

Steinert, Brian W.↗
NASA NTRS2004

Induced Anisotropy in FeCo-Based Nanocrystalline Ferromagnetic Alloys (HITPERM) by Very High Field Annealing

Very high magnetic field annealing is shown to affect the magnetic anisotropy in FeCo-base nanocrystalline soft ferromagnetic alloys. Alloys of composition Fe(44.5)Co(44.5)Zr(7)B(4) were prepared by melt spinning into amorphous ribbons, then wound to form toroidal bobbin cores. One set of cores was crystallized in a zero field at 600 deg. C for 1 h, then, field annealed at 17 tesla (T) at 480 deg. C for 1 h. Another set was crystallized in a 17-T field at 480 deg. C for 1 h. Field orientation was transverse to the magnetic path of the toroidal cores. An induced anisotropy is indicated by a sheared hysteresis loop. Sensitive torque magnetometry measurements with a Si cantilever sensor indicated a strong, uniaxial, longitudinal easy axis in the zero-field-crystallized sample. The source is most likely magnetoelastic anisotropy, caused by the residual stress from nanocrystallization and the nonzero magnetostriction coefficient for this material. The magnetostrictive coefficient lambda(5) is measured to be 36 ppm by a strain gage technique. Field annealing reduces the magnitude of the induced anisotropy. Core loss measurements were made in the zero-field-crystallized, zero-field-crystallized- than-field-annealed, and field-crystallized states. Core loss is reduced 30%-50% (depending on frequency) by field annealing. X-ray diffraction reveals no evidence of crystalline texture or orientation that would cause the induced anisotropy. Diffusional pair ordering is thought to be the cause of the induced anisotropy. However, reannealing the samples in the absence of a magnetic field at 480 deg. C does not completely remove the induced anisotropy.

FERROMAGNETIC ALLOYS↗
NASA NTRS2000

Helicon Plasma Injector and Ion Cyclotron Acceleration Development in the VASIMR Experiment

In the Variable Specific Impulse Magnetoplasma Rocket (VASIMR) radio frequency (rf) waves both produce the plasma and then accelerate the ions. The plasma production is done by action of helicon waves. These waves are circular polarized waves in the direction of the electron gyromotion. The ion acceleration is performed by ion cyclotron resonant frequency (ICRF) acceleration. The Advanced Space Propulsion Laboratory (ASPL) is actively developing efficient helicon plasma production and ICRF acceleration. The VASIMR experimental device at the ASPL is called VX-10. It is configured to demonstrate the plasma production and acceleration at the 10kW level to support a space flight demonstration design. The VX-10 consists of three electromagnets integrated into a vacuum chamber that produce magnetic fields up to 0.5 Tesla. Magnetic field shaping is achieved by independent magnet current control and placement of the magnets. We have generated both helium and hydrogen high density (>10(exp 18) cu m) discharges with the helicon source. ICRF experiments are underway. This paper describes the VX-10 device, presents recent results and discusses future plans.

Squire, Jared P.↗
NASA NTRS2010

The Effect of Magnetohydrodynamic (MHD) Energy Bypass on Specific Thrust for a Supersonic Turbojet Engine

This paper describes the preliminary results of a thermodynamic cycle analysis of a supersonic turbojet engine with a magnetohydrodynamic (MHD) energy bypass system that explores a wide range of MHD enthalpy extraction parameters. Through the analysis described here, it is shown that applying a magnetic field to a flow path in the Mach 2.0 to 3.5 range can increase the specific thrust of the turbojet engine up to as much as 420 N/(kg/s) provided that the magnitude of the magnetic field is in the range of 1 to 5 Tesla. The MHD energy bypass can also increase the operating Mach number range for a supersonic turbojet engine into the hypersonic flight regime. In this case, the Mach number range is shown to be extended to Mach 7.0.

Benyo, Theresa L.↗
NASA NTRS2009

A Simple Mechanism for Fractionating Oxygen Isotopes in the Solar Nebula

Lightning in the Solar Nebula is caused by the tribo-electric charging of dust grains carried by massive turbulent flows and driven by the accretion energy in the disk: it has long been one agent assumed responsible for the formation of chondrules. The degree to which charge separation can occur is dependent upon a number of factors, including the concentration of radioactive sources and the total level of ionization in the nebula, and these factors determine the maximum energy likely to be released by a single bolt. While chondrule formation requires a massive discharge, even a small lightning bolt can vaporize grains in the ionized discharge channel. Experimental studies have shown that silica, iron silicate and iron oxide grains formed from a high voltage discharge in hydrogen rich gas containing some oxygen produces solids that are enriched in O-17 and O-18 relative to the composition of the starting gas. Vaporization of silicates produces SiO, metal and free oxygen atoms in each discharge and these species will immediately begin to recondense from the hot plasma. Freshly condensed grains are incrementally enriched in heavy oxygen while the gas is enriched in O-16. Repeated evaporation and condensation of silicates in continuously occurring lightning discharges will monotonically increase the fractionation of oxygen isotopes between the O-17 and O-18 rich dust and the O-16 rich gas. The first mass independently fractionated refractory oxide particles were produced in the lab following the condensation of a flowing gas mixture containing variable amounts of hydrogen, silane, pentacarbonyl iron and oxygen that passed through a high voltage discharge powered by a Tesla coil. While the exact chemical pathway is still uncertain, the most probable reaction mechanisms involve oxidation of the growing refractory clusters by O3, OH or O atoms. This model has some interesting consequences for chemical processes in the early solar nebula. Chemical fractionation of recondensed dust evaporated via lightning discharges should be strongly time dependent. At earlier times, the accretion rate is maximal, thus driving strong turbulence, energetic grain-grain collisions, tribo-electric charging and charge separation, leading to frequent, powerful lightning discharges. As the accretion rate diminishes, turbulence decreases and lightning discharges will become both less powerful and less frequent, thus decreasing the rate of dust-gas fractionation. The most rapid increase in the formation of O-16 poor dust will occur early in nebular history. Generation of fractionated dust should be distributed throughout the inner disk. Once condensed, grain dispersal would average out any significant isotopic anomalies within the inner disk.

Nuth, Joseph A., III↗
NASA NTRS2011

High field superconducting magnets

A superconducting magnet includes an insulating layer disposed about the surface of a mandrel; a superconducting wire wound in adjacent turns about the mandrel to form the superconducting magnet, wherein the superconducting wire is in thermal communication with the mandrel, and the superconducting magnet has a field-to-current ratio equal to or greater than 1.1 Tesla per Ampere; a thermally conductive potting material configured to fill interstices between the adjacent turns, wherein the thermally conductive potting material and the superconducting wire provide a path for dissipation of heat; and a voltage limiting device disposed across each end of the superconducting wire, wherein the voltage limiting device is configured to prevent a voltage excursion across the superconducting wire during quench of the superconducting magnet.

Hait, Thomas P.↗
NASA NTRS2011

Flow Matching Results of an MHD Energy Bypass System on a Supersonic Turbojet Engine Using the Numerical Propulsion System Simulation (NPSS) Environment

Flow matching has been successfully achieved for an MHD energy bypass system on a supersonic turbojet engine. The Numerical Propulsion System Simulation (NPSS) environment helped perform a thermodynamic cycle analysis to properly match the flows from an inlet employing a MHD energy bypass system (consisting of an MHD generator and MHD accelerator) on a supersonic turbojet engine. Working with various operating conditions (such as the applied magnetic field, MHD generator length and flow conductivity), interfacing studies were conducted between the MHD generator, the turbojet engine, and the MHD accelerator. This paper briefly describes the NPSS environment used in this analysis. This paper further describes the analysis of a supersonic turbojet engine with an MHD generator/accelerator energy bypass system. Results from this study have shown that using MHD energy bypass in the flow path of a supersonic turbojet engine increases the useful Mach number operating range from 0 to 3.0 Mach (not using MHD) to a range of 0 to 7.0 Mach with specific net thrust range of 740 N-s/kg (at ambient Mach = 3.25) to 70 N-s/kg (at ambient Mach = 7). These results were achieved with an applied magnetic field of 2.5 Tesla and conductivity levels in a range from 2 mhos/m (ambient Mach = 7) to 5.5 mhos/m (ambient Mach = 3.5) for an MHD generator length of 3 m.

Benyo, Theresa L.↗
NASA NTRS2011

Annular MHD Physics for Turbojet Energy Bypass

The use of annular Hall type MHD generator/accelerator ducts for turbojet energy bypass is evaluated assuming weakly ionized flows obtained from pulsed nanosecond discharges. The equations for a 1-D, axisymmetric MHD generator/accelerator are derived and numerically integrated to determine the generator/accelerator performance characteristics. The concept offers a shockless means of interacting with high speed inlet flows and potentially offers variable inlet geometry performance without the complexity of moving parts simply by varying the generator loading parameter. The cycle analysis conducted iteratively with a spike inlet and turbojet flying at M = 7 at 30 km altitude is estimated to have a positive thrust per unit mass flow of 185 N-s/kg. The turbojet allowable combustor temperature is set at an aggressive 2200 deg K. The annular MHD Hall generator/accelerator is L = 3 m in length with a B(sub r) = 5 Tesla magnetic field and a conductivity of sigma = 5 mho/m for the generator and sigma= 1.0 mho/m for the accelerator. The calculated isentropic efficiency for the generator is eta(sub sg) = 84 percent at an enthalpy extraction ratio, eta(sub Ng) = 0.63. The calculated isentropic efficiency for the accelerator is eta(sub sa) = 81 percent at an enthalpy addition ratio, eta(sub Na) = 0.62. An assessment of the ionization fraction necessary to achieve a conductivity of sigma = 1.0 mho/m is n(sub e)/n = 1.90 X 10(exp -6), and for sigma = 5.0 mho/m is n(sub e)/n = 9.52 X 10(exp -6).

Schneider, Steven J.↗
NASA NTRS2012

NASA Tech Briefs, October 2012

Topics discussed include: Detection of Chemical Precursors of Explosives; Detecting Methane From Leaking Pipelines and as Greenhouse Gas in the Atmosphere; Onboard Sensor Data Qualification in Human-Rated Launch Vehicles; Rugged, Portable, Real-Time Optical Gaseous Analyzer for Hydrogen Fluoride; A Probabilistic Mass Estimation Algorithm for a Novel 7-Channel Capacitive Sample Verification Sensor; Low-Power Architecture for an Optical Life Gas Analyzer; Online Cable Tester and Rerouter; A Three-Frequency Feed for Millimeter-Wave Radiometry; Capacitance Probe Resonator for Multichannel Electrometer; Inverted Three-Junction Tandem Thermophotovoltaic Modules; Fabrication of Single Crystal MgO Capsules; Inflatable Hangar for Assembly of Large Structures in Space; Mars Aqueous Processing System; Hybrid Filter Membrane; Design for the Structure and the Mechanics of Moballs; Pressure Dome for High-Pressure Electrolyzer; Cascading Tesla Oscillating Flow Diode for Stirling Engine Gas Bearings; Compact, Low-Force, Low-Noise Linear Actuator; Ultra-Compact Motor Controller; Extreme Ionizing-Radiation-Resistant Bacterium; Wideband Single-Crystal Transducer for Bone Characterization; Fluorescence-Activated Cell Sorting of Live Versus Dead Bacterial Cells and Spores; Nonhazardous Urine Pretreatment Method; Laser-Ranging Transponders for Science Investigations of the Moon and Mars; Ka-Band Waveguide Three-Way Serial Combiner for MMIC Amplifiers; Structural Health Monitoring with Fiber Bragg Grating and Piezo Arrays; Low-Gain Circularly Polarized Antenna with Torus-Shaped Pattern; Stereo and IMU- Assisted Visual Odometry for Small Robots; Global Swath and Gridded Data Tiling; GOES-R: Satellite Insight; Aquarius iPhone Application; Monitoring of International Space Station Telemetry Using Shewhart Control Charts; Theory of a Traveling Wave Feed for a Planar Slot Array Antenna; Time Manager Software for a Flight Processor; Simulation of Oxygen Disintegration and Mixing With Hydrogen or Helium at Supercritical Pressure; A Superfluid Pulse Tube Refrigerator Without Moving Parts for Sub-Kelvin Cooling; Sapphire Viewports for a Venus Probe; The Mobile Chamber; Electric Propulsion Induced Secondary Mass Spectroscopy; and Radiation-Tolerant DC-DC Converters.

Source record↗
NASA NTRS2014

Magnet Architectures and Active Radiation Shielding Study - SR2S Workshop

Analyze new coil configurations with maturing superconductor technology -Develop vehicle-level concept solutions and identify engineering challenges and risks -Shielding performance analysis Recent advances in superconducting magnet technology and manufacturing have opened the door for re-evaluating active shielding solutions as an alternative to mass prohibitive passive shielding.Publications on static magnetic field environments and its bio-effects were reviewed. Short-term exposure information is available suggesting long term exposure may be okay. Further research likely needed. center dotMagnetic field safety requirements exist for controlled work environments. The following effects have been noted with little noted adverse effects -Magnetohydrodynamic (MHD) effects on ionized fluids (e.g. blood) creating an aortic voltage change -MHD interaction elevates blood pressure (BP) center dot5 Tesla equates to 5% BP elevation -Prosthetic devises and pacemakers are an issue (access limit of 5 gauss).

Westover, Shane↗
NASA NTRS2014

Visual Impairment/lntracranial Pressure Risk Clinical Care Data Tools

Prior to 2010, several ISS crewmembers returned from spaceflight with changes to their vision, ranging from a mild hyperopic shift to frank disc edema. As a result, NASA expanded clinical vision testing to include more comprehensive medical imaging, including Optical Coherence Tomography and 3 Tesla Brain and Orbit MRIs. The Space and Clinical Operations (SCO) Division developed a clinical practice guideline that classified individuals based on their symptoms and diagnoses to facilitate clinical care. For the purposes of clinical surveillance, this classification was applied retrospectively to all crewmembers who had sufficient testing for classification. This classification is also a tool that has been leveraged for researchers to identify potential risk factors. In March 2014, driven in part by a more comprehensive understanding of the imaging data and increased imaging capability on orbit, the SCO Division revised their clinical care guidance to outline in‐flight care and increase post‐flight follow up. The new clinical guidance does not include a classification scheme

Van Baalen, Mary↗
NASA NTRS2015

Evidence Based Medicine in Space Flight: Evaluation of Inflight Vision Data for Operational Decision-Making

Due to recently identified vision changes associated with space flight, JSC Space and Clinical Operations (SCO) implemented broad mission‐related vision testing starting in 2009. Optical Coherence Tomography (OCT), 3 Tesla Brain and Orbit MRIs, Optical Biometry were implemented terrestrially for clinical monitoring. While no inflight vision testing was in place, already available onorbit technology was leveraged to facilitate in‐flight clinical monitoring, including visual acuity, Amsler grid, tonometry, and ultrasonography. In 2013, on‐orbit testing capabilities were expanded to include contrast sensitivity testing and OCT. As these additional testing capabilities have been added, resource prioritization, particularly crew time, is under evaluation.

Van Baalen, Mary↗
NASA NTRS2016

Utilizing GPUs to Accelerate Turbomachinery CFD Codes

GPU computing has established itself as a way to accelerate parallel codes in the high performance computing world. This work focuses on speeding up APNASA, a legacy CFD code used at NASA Glenn Research Center, while also drawing conclusions about the nature of GPU computing and the requirements to make GPGPU worthwhile on legacy codes. Rewriting and restructuring of the source code was avoided to limit the introduction of new bugs. The code was profiled and investigated for parallelization potential, then OpenACC directives were used to indicate parallel parts of the code. The use of OpenACC directives was not able to reduce the runtime of APNASA on either the NVIDIA Tesla discrete graphics card, or the AMD accelerated processing unit. Additionally, it was found that in order to justify the use of GPGPU, the amount of parallel work being done within a kernel would have to greatly exceed the work being done by any one portion of the APNASA code. It was determined that in order for an application like APNASA to be accelerated on the GPU, it should not be modular in nature, and the parallel portions of the code must contain a large portion of the code's computation time.

computer programming↗
NASA NTRS2018

Integrated Computational-Experimental Development of Lithium-Air Batteries for Electric Aircraft

The primary obstacle to enable NASA's vision of Green Aviation is the extraordinary energy storage requirements for electric aircraft. Significant advances in high energy, rechargeable, safe batteries are required to enable electric aviation. Boeing's SUGAR and NASA studies have identified 400 Wh/kg as the threshold energy density for general aviation and 750 Wh/kg for commercial regional air service. State of the Art Lithium Ion Battery (LIB) technology currently has a density of 200 Wh/kg and is expected to plateau at 300 Wh/kg due to fundamental chemistry limitations making it unsuitable for future electric aircraft. Additional demanding requirements include high power, rechargeability, and high safety. Such battery technology does not currently exist. The recent considerable activity in battery research (DOE, Tesla Gigafactory, etc) overwhelmingly has been geared towards reducing cost and improving safety of LIB technology in order to promote the adoption of electric automobiles; and thus it is expected to have little impact on electric aviation development. New battery materials will be needed for the "Beyond Li Ion" (BLI) technologies required for high energy, safe electric aviation. Li-Air batteries have the highest known theoretical energy density (3400 Wh/kg) and therefore and if realized promises to transform the global transportation system. These high energy batteries have the potential to meet the energy storage challenges of current and future NASA aeronautics and space missions in addition to many terrestrial transportation applications as well. However, this technology requires significant components development and integration, as it is currently unable to achieve aircraft requirements. The objective of this project is to leverage modern computational materials methods combined with battery multiphysics tools to develop radically advanced compatible cathode and electrolyte materials, build several Li-Air cells, and flight-demonstrate the corresponding Li-Air battery packs. A significant problem for current Lithium-Air batteries is large scale decomposition of the battery electrolyte during operation leading to battery failure after a handful of charge/discharge cycles. Therefore, development of large scale, ultra-high energy, rechargeable, and safe Lithium-Air batteries require highly stable electrolytes that are resistant to decomposition under operating conditions. A NASA-based cross-organizational "dream team" of high-powered experts combined integrated supercomputer modeling, fundamental chemistry analysis, advanced material science, and battery cell development to tackle this very challenging, multidisciplinary problem. The ultimate goal for the team is to develop an integrated experimental/computational infrastructure to produce a reliable predictive capability for the selection of optimal components, their fabrication parameters, and "design rules" of novel cell components for advanced ultra-high energy batteries that can meet energy storage challenges of NASA missions and many terrestrial transportation applications.

Li-air battery↗
NASA NTRS2018

An Investigation of Parallel Programming Techniques Applied to Monte Carlo Simulations for Post-Flight Reconstruction of Spacecraft Trajectory

Parallelizing software to execute on multi-core central processing units (CPUs) and graphics processing units (GPUs) can be challenging. For some fields outside of Computer Science, this transition comes with new issues. For example, memory limitations can require modifications to code not initially developed to run on GPUs. This work applies the Open Multi-Processing (OpenMP) and Open Accelerators (OpenACC) directive-based parallelization strategies on a Monte Carlo simulation approach for trajectory reconstruction enabling it to run on multi-core CPUs and GPUs. Large matrix operations are the most common use of GPUs, which are not present in this algorithm; however, the natural parallelism of independent trajectories in Monte Carlo simulations is exploited. Benchmarking data are presented comparing execution times of the software for single-thread CPUs, multi-thread CPUs with OpenMP, and multi-thread GPUs using OpenACC. These data were collected using nodes with Intel® Xeon® E5-2670 (Sandy Bridge) CPUs enhanced with NVIDIA® Tesla® K40 GPUs on the Pleiades Supercomputer cluster at the National Aeronautics and Space Administration (NASA) Ames Research Center (ARC) and a local Intel® Xeon Phi™ node at NASA Langley Research Center (LaRC).

Williams, R. Anthony↗
NASA NTRS2019

Further Development Of Aperture: A Precise Extremely Large Reflective Telescope Using Re-Configurable Elements: NASA Innovative Advanced Concepts (NIAC) - Phase II

One of the pressing needs for space ultraviolet-visible astronomy is a design to allow larger mirrors than the James Webb Space Telescope primary. The diameter of the rocket fairing limits the mirror diameter such that all future missions calling for mirrors up to 16 meters in diameter or larger will require a mirror that is deployed post-launch. In response to the deployment requirement, we address the issues of this concept called "A Precise Extremely Large Reflective Telescope Using Reconfigurable Elements (APERTURE) with both hardware experiments and software simulations... We designed and built several fixtures with O-rings to hold a membrane. We established a coating process to make a membrane that was coated on one side with Cr and the other side with Cr-Terfenol-D-NiCo. The Terfenol-D (T-D hereafter) is the MSM (Magnetic Smart Memory) we use. We bought and established a procedure for measuring a deformation over time and purchased a Shack Hartmann system from Imagine Optic (https://www.imagine-optic.com). The first substrate we used was DuPont (TM) Kapton® polyimide film. Due to material creep, we found the stability over a 48-hour period with a Kapton substrate was not as good as desired (greater than 1 micron). We then switched to CP1 Polyimide. We found the CP1 much more stable to creep, being stable from about 3 hours to 48 hours to within a measurement error to below approximately 0.1 micron. We produced a 13 micron maximum deviation on a 50-millimeter-diameter piece of CP1 (25 microns thick). The T-D coating was about 2 microns, and the other layers, about 10 nanometers. The magnetic field at the base was about 0.1 teslas. We can make the T-D film at least 5 times thicker and the magnetic field at least 5 times stronger, and hence make deformations as much as 25 times larger. We have a formed a collaboration produced at the NIAC (NASA Innovative Advanced Concepts) mid-term review with Dr. Ron Shiri of Goddard Space Flight Center (GSFC) to explore making controlled deviations on lambda/14-lambda/20 scales which are required to bring a surface to the diffraction limit. We carried out only preliminary work on Si using a Coordinate Measuring Machine (CMM), which produced deviations on the 1 micron level. We are still working on a program to bring to GSFC a flat enough (radius of curvature greater than 10 microns) -coated a Si piece with Cr, T-D, NiCo. Then we plan to carry out tests with an interferometer. Further, we formed a new collaboration with Prof. Rajan Vaidyanathan of the University of Central Florida to replace the CP1 with a shape memory alloy (SMA). With his collaboration, we acquired new Federal funding outside of NASA to explore the use of SMAs (we use NiTi). Our preliminary results indicate that we can produce deformations greater than 1 micron on approximately 100 microns thick. Furthermore we have shown that the NiTi can deploy to better than 1 micron of its set original and then trained shape.

Re-Configurable Elements↗
NASA NTRS2019

Accelerated Simulation of Air Pollution Using NVIDIA RAPIDS

Atmospheric chemistry models are a central tool to study and forecast the impact of air pollution on the environment, vegetation, and human health. However, the numerical simulation of chemical kinetics is computationally expensive due to the stiffness of the system of ordinary differential equations that describes atmospheric chemistry. Here we present an alternative approach to the computation of atmospheric chemistry based on machine learning. Our training data set is produced using the NASA Goddard Earth Observing System (GEOS) model with GEOS-Chem chemistry, run on the NASA Center for Climate Simulation (NCCS) Discover supercomputing cluster on 384 Intel Xeon Haswell cores. This model spends more than 50% of total run time on solving atmospheric chemistry. The data set contains as input features the air pollution concentrations before solving the differential equations, together with some key physical parameters such as temperature and sun intensity. As target variables we define the air pollution concentrations after solving the differential equations. Using Dask-cuDF and Dask-XGBoost on the NVIDIA RAPIDS platform on 8 Tesla V100 GPUs, we generate from this training set gradient boosted decision tree models that can reproduce the simulation of chemical kinetics. We do this on the NCCS Advanced Data Analytics Platform (ADAPT) science cloud environment. Our application takes full advantage of recent advances in Dask-XGBoost, such as multi-node and multi-GPU scaling for distributed training with large data sets. The increase in training data size enabled by this is critical to capture the full range of chemical environments encountered across the globe and all annual seasons.The boosted tree models offer good predictability and show many of the features of the full chemistry reference simulation. Further improvements can be achieved through mass balance considerations and by accounting for error correlations. We incorporate the boosted tree models into the GEOS reference model using XGBoost's C API. This enables a seamless integration of the GPU trained models into GEOS-Chem, which is written in Fortran and optimized for use in a massively parallel CPU environment. We show the benefits of this approach and discuss the potential speedup of this machine learning accelerated atmospheric chemistry model.

Keller, Christoph A.↗