Search NASA⌕ Search

SEARCH · Search NASA

Results for “sampling size”

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 433 records · Page 24
All sourcesNASA NTRSDOE OSTI
NASA NTRS2004

Stardust Sample Collection at Wild 2 and Its Preliminary Examination

The primary objective of STARDUST is to collect coma samples from 81P/Wild 2. This was made on January 2, 2004. Before the encounter three significant model predictions existed for the number and size of samples to be captured. Three investigations during the Wild 2 encounter (Dust Flux Monitor, Comet and Interstellar Dust Analyzer and Dynamic Science) made in situ measurements of the dust. Spectacular images were captured of the Wild 2 nucleus and dust jets. This abstract compares the model predictions with the in situ measurements and Wild 2 images and assesses the likely samples to be returned for analysis on January 15, 2006. To give some lead time for sample analysts to prepare for the analyses of the returned samples, the organization of the Preliminary Examination is presented.

Tsou, P.↗
NASA NTRS2024

Plasma-Based Elemental Analyzer for Crewed Transit and Planetary Habitation

As NASA continues to perform research on the ISS and move towards the Moon and Mars, it is important for astronauts to have a technology capable of determining the composition of solid, liquid, and gas samples. Currently, samples are transported back to Earth for analysis, which may impact results due to the time delay of samples waiting to be returned to Earth as well as potential sample contamination from re-entry, change in environment, and various personnel handling. Therefore, researchers at the Applied Chemistry Lab at KSC have developed a proof-of-concept “Plasma-Based Elemental Analyzer,” which has the capability of testing solid, liquid, and gas samples for identification. Currently, the analysis is only qualitative, with quantitative identification left for future work. This technology is similar to ICP-OES, but differs in instrument size and sample preparation. Both technologies use plasma to ionize samples for light emission. The plasma elemental analyzer aims to be portable and lightweight, require only electrical power, and analyze solid, liquid, and gas samples. It does not use acid digestion or a nebulizer, requires less power than ICP-OES, and is able to operate with minimal to no usage of argon gas. Furthermore, the plasma elemental analyzer operates within low power constraints, which makes it feasible for usage on future missions with higher power constraints. The technology was tested at atmospheric pressure using mostly room air as the carrier gas for plasma production in the proof-of-concept study. The results showed the peak signals of the elements with an associated arbitrary signal intensity that allowed for qualitative identification. Solid samples that were tested included inedible biomass from the Golden Heirloom Cherry Tomato plant and Lunar Highlands Simulants (LHS-1). The liquid samples included elements of interest for the space crop production community. This work showed that a system smaller than the traditional ICP-OES equipment is able to be designed and implemented in future use cases, although it requires further development.

Kenneth Engeling↗
NASA NTRS1993

Extraction and isotopic analysis of medium molecular weight hydrocarbons from Murchison using supercritical carbon dioxide

The large variety of organic compounds present in carbonaceous chondrites poses particular problems in their analysis not the least of which is terrestrial contamination. Conventional analytical approaches employ simple chromatographic techniques to fractionate the extractable compounds into broad classes of similar chemical structure. However, the use of organic solvents and their subsequent removal by evaporation results in the depletion or loss of semi-volatile compounds as well as requiring considerable preparative work to assure solvent purity. Supercritical fluids have been shown to provide a powerful alternative to conventional liquid organic solvents used for analytical extractions. A sample of Murchison from the Field Museum was analyzed. Two interior fragments were used; the first (2.85 g) was crushed in an agate pestel and mortar to a grain size of ca. 50-100 micron, the second (1.80 g) was broken into chips 3-8 mm in size. Each sample was loaded into a stainless steel bomb and placed in the extraction chamber of an Isco supercritical fluid extractor maintained at 35 C. High purity (99.9995 percent) carbon dioxide was used and was pressurized using an Isco syringe pump. The samples were extracted dynamically by flowing CO2 under pressure through the bomb and venting via a 50 micron fused filica capillary into 5 mls of hexane used as a collection solvent. The hexane was maintained at a temperature of 0.5 C. A series of extractions were done on each sample using CO2 of increasing density. The principal components extracted in each fraction are summarized.

Gilmour, Iain↗
NASA NTRS2008

BRDF of Salt Pan Regolith Samples

Laboratory Bi-directional Reflectance Distribution Function (BRDF) measurements of salt pan regolith samples are presented in this study in an effort to understand the role of spatial and spectral variability of the natural biome. The samples were obtained from Etosha Pan, Namibia (19.20 deg S, 15.93 deg E, alt. 1100 m). It is shown how the BRDF depends on the measurement geometry - incident and scatter angles and on the sample particle sizes. As a demonstration of the application of the results, airborne BRDF measurements acquires with NASA's Cloud Absorption Radiometer (CAR) over the same general site where the regolith samples were collected are compared with the laboratory results. Good agreement between laboratory measured and field measured BRDF is reported.

Georgiev, Georgi T.↗
NASA NTRS2004

Transition from Forward Smoldering to Flaming in Small Polyurethane Foam Samples

Experimental observations are presented of the effect of the flow velocity and oxygen concentration, and of a thermal radiant flux, on the transition from smoldering to flaming in forward smoldering of small samples of polyurethane foam with a gas/solid interface. The experiments are part of a project studying the transition from smolder to flaming under conditions encountered in spacecraft facilities, i.e., microgravity, low velocity variable oxygen concentration flows. Because the microgravity experiments are planned for the International Space Station, the foam samples had to be limited in size for safety and launch mass reasons. The feasible sample size is too small for smolder to self propagate because of heat losses to the surrounding environment. Thus, the smolder propagation and the transition to flaming had to be assisted by reducing the heat losses to the surroundings and increasing the oxygen concentration. The experiments are conducted with small parallelepiped samples vertically placed in a wind tunnel. Three of the sample lateral-sides are maintained at elevated temperature and the fourth side is exposed to an upward flow and to a radiant flux. It is found that decreasing the flow velocity and increasing its oxygen concentration, and/or increasing the radiant flux enhances the transition to flaming, and reduces the delay time to transition. Limiting external ambient conditions for the transition to flaming are reported for the present experimental set-up. The results show that smolder propagation and the transition to flaming can occur in relatively small fuel samples if the external conditions are appropriate. The results also indicate that transition to flaming occurs in the char left behind by the smolder reaction, and it has the characteristics of a gas-phase ignition induced by the smolder reaction, which acts as the source of both gaseous fuel and heat.

Bar-Ilan, A.↗
NASA NTRS1974

Lunar Science Conference, 5th, Houston, Tex., March 18-22, 1974, Proceedings. Volume 1 - Mineralogy and petrology. Volume 2 Chemical and isotope analyses. Organic chemistry. Volume 3 - Physical properties

Numerous studies on the properties of the moon based on Apollo findings and samples are presented. Topics treated include ages of the lunar nearside light plains and maria, orange material in the Sulpicius Gallus formation at the southwestern edge of Mare Serenitatis, impact-induced fractionation in the lunar highlands, igneous rocks from Apollo 16 rake samples, experimental liquid line of descent and liquid immiscibility for basalt 70017, ion microprobe mass analysis of plagioclase from 'non-mare' lunar samples, grain size and the evolution of lunar soils, chemical composition of rocks and soils at Taurus-Littrow, the geochemical evolution of the moon, U-Th-Pb systematics of some Apollo 17 lunar samples and implications for a lunar basin excavation chronology, volatile-element systematics and green glass in Apollo 15 lunar soils, solar wind nitrogen and indigenous nitrogen in Apollo 17 lunar samples, lunar trapped xenon, solar flare and lunar surface process characterization at the Apollo 17 site, and the permanent and induced magnetic dipole moment of the moon. Individual items are announced in this issue.

Gose, W. A.↗
NASA NTRS1989

Problems in determining the surface density of the Galactic disk

A new method is presented for determining the local surface density of the Galactic disk from distance and velocity measurements of stars toward the Galactic poles. The procedure is fully three-dimensional, approximating the Galactic potential by a potential of Staeckel form and using the analytic third integral to treat the tilt and the change of shape of the velocity ellipsoid consistently. Applying the procedure to artificial data superficially resembling the K dwarf sample of Kuijken and Gilmore (1988, 1989), it is shown that the current best estimates of local disk surface density are uncertain by at least 30 percent. Of this, about 25 percent is due to the size of the velocity sample, about 15 percent comes from uncertainties in the rotation curve and the solar galactocentric distance, and about 10 percent from ignorance of the shape of the velocity distribution above z = 1 kpc, the errors adding in quadrature. Increasing the sample size by a factor of 3 will reduce the error to 20 percent. To achieve 10 percent accuracy, observations will be needed along other lines of sight to constrain the shape of the velocity ellipsoid.

Statler, Thomas S.↗
NASA NTRS2000

Fiber-Optic Imaging Probe Developed for Space Used to Detect Diabetes Through the Eye

Approximately 16 million Americans have diabetes mellitus, which can severely impair eyesight by causing cataracts, diabetic retinopathy, and glaucoma. Cataracts are 1.6 times more common in people with diabetes than in those without diabetes, and cataract extraction is the only surgical treatment. In many cases, diabetes-related ocular pathologies go undiagnosed until visual function is compromised. This ongoing pilot project seeks to study the progression of diabetes in a unique animal model by monitoring changes in the lens with a safe, sensitive, dynamic light-scattering probe. Dynamic light scattering (DLS), has the potential to diagnose cataracts at the molecular level. Recently, a new DLS fiber-optic probe was developed at the NASA Glenn Research Center at Lewis Field for noncontact, accurate, and extremely sensitive particle-sizing measurements in fluid dispersions and suspensions (ref. 1). This compact, portable, and rugged probe is free of optical alignment, offers point-and-shoot operation for various online field applications and challenging environments, and yet is extremely flexible in regards to sample container sizes, materials, and shapes. No external vibration isolation and no index matching are required. It can measure particles as small as 1 nm and as large as few micrometers in a wide concentration range from very dilute (waterlike) dispersions to very turbid (milklike) suspensions. It is safe and fast to use, since it only requires very low laser power (10 nW to 3 mW) with very short data acquisition times (2 to 10 sec).

Ansari, Rafat R.↗
NASA NTRS2012

Analysis of Direct Samples of Early Solar System Aqueous Fluids

Over the past three decades we have become increasingly aware of the fundamental importance of water, and aqueous alteration, on primitive solar-system bodies. Some carbonaceous and ordinary chondrites have been altered by interactions with liquid water within the first 10 million years after formation of their parent asteroids. Millimeter to centimeter-sized aggregates of purple halite containing aqueous fluid inclusions were found in the matrix of two freshly-fallen brecciated H chondrite falls, Monahans (1998, hereafter simply "Monahans") (H5) and Zag (H3-6) (Zolensky et al., 1999; Whitby et al., 2000; Bogard et al., 2001) In order to understand origin and evolution of the aqueous fluids inside these inclusions we much measure the actual fluid composition, and also learn the O and H isotopic composition of the water. It has taken a decade for laboratory analytical techniques to catch up to these particular nanomole-sized aqueous samples. We have recently been successful in (1) measuring the isotopic composition of H and O in the water in a few fluid inclusions from the Zag and Monahans halite, (2) mineralogical characterization of the solid mineral phases associated with the aqueous fluids within the halite, and (3) the first minor element analyses of the fluid itself. A Cameca ims-1270 equipped with a cryo-sample-stage of Hokkaido University was specially prepared for the O and H isotopic measurements. The cryo-sample-stage (Techno. I. S. Corp.) was cooled down to c.a. -190 C using liquid nitrogen at which the aqueous fluid in inclusions was frozen. We excavated the salt crystal surfaces to expose the frozen fluids using a 15 keV Cs+ beam and measured negative secondary ions. The secondary ions from deep craters of approximately 10 m in depth emitted stably but the intensities changed gradually during measurement cycles because of shifting states of charge compensation, resulting in rather poor reproducibility of multiple measurements of standard fluid inclusions of +/- 90 0/00(2 sigma) for delta D, and +/- 29 0/00 (2 sigma) for delta O-18. On the other hand, the reproducibility of Delta O-17 is plus or minus 8 /00 (2 sigma ) because the observed variations of isotope ratios follow a mass dependent fractionation law. Variations of delta D of the aqueous fluids range over sog,a 330(90; 2 sigma ) to +1200(90) 0/00 for Monahans and delta 300(96) 0/00 to +90(98)0/00 for Zag. Delta O-17 of aqueous fluids range over delta 16(22) 0/00 to +18(10) 0/00 for Monahans and +3(10) 0/00 to +27(11) 0/00 for Zag. These variations are larger than the reproducibility of standard analyses and suggest that isotope equilibria were under way in the fluids before trapping into halite. The mean values of delta D and Delta O-17 are +290 0/00 and +9 0/00, respectively. The mean values and the variations of these fluids are different from the representative values of ordinary chondrites, verifying our working hypothesis that the fluid inclusion-bearing halites were not indigenous to the H chondrite parent-asteroid but rather represent exogenous material delivered onto the asteroid from a separate cryovolcanically-active body. This initial isotopic work has demonstrated the feasibility of the measurements, but also revealed sample processing and analytical shortcomings that are now being addressed. Examination of solid mineral inclusions within Monahans and Zag halite grains by confocal Raman spectroscopy at the Carnegie Geophysical Laboratory has revealed them to be metal, magnetite, forsteritic olivine (Fo.98), macromolecular carbon (MMC), pyroxenes, feldspar with Raman spectral affinity to anorthoclase and, probably, fine-grained lepidocrocite (FeO(OH)). In addition, one inclusion features aliphatic material with Raman spectral features consistent with a mixture of short-chain aliphatic compounds. We have initiated analyses of the bulk composition of the fluids within the inclusions in Zag and Monahans halites at Virginia Tech by LA ICPMS using angilent 7500ce quadrupole ICPMS and a Lambda Physik GeoLas 193 nm excimer laser ablation system. Preliminary results reveal that the inclusion aqueous fluids contain highly charged cations of Ca, Mg and Fe. The minerals and compounds discovered thus far within Monahans/Zag halites are indicative of an originating body at least partly composed of unequilibrated anhydrous materials (high Fo olivine, pyroxenes, feldspars, possibly the metal) which were subjected to aqueous alteration (the halite parent brine) and containing a light organic component (the short-chain aliphatic compounds). This material was ejected from the originating body with little or no disruption, as evidenced with the presence of fluid inclusions. An actively geysering body similar to modern Enceladus (Postberg et al., 2011) may be a reasonable analogue in this respect. Also, the originating body should have been within close proximity to the H chondrite parent in order to generate the number of halite grains seen in Monahans and Zag. Other candidates for Monahans/Zag halite parent bodie(s) may include a young Ceres with its possible liquid ocean, or Main Belt comets.

Zolensky, Michael E.↗
NASA NTRS1974

Nondestructive activation analysis of sample of lunar surface material returned by Luna 16 automatic station

The composition of a sample of lunar surface material returned by the Luna 16 automatic station from the Sea of Fertility was studied, using nondestructive activation analysis. The structure of the returned surface material is inhomogeneous; the surficial material is thin, quite homogeneous, and the granularity increases with depth. Based on grain size, the sample was separated into five zones. The activation analysis was conducted on a sample taken from the friable surficial layer, zone A. The content of Al, Mn, Na, Cr, Co, Fe, and Sc was determined by nondestructive activation analysis of the sample. In determining Cr, Co, Fe, and Sc, the sample was irradiated for 24 hours and cooled for 10 days. Gamma spectra of the samples were recorded with a semiconductor Ge(Li)-detector and a multichannel analyzer, and measurement data were processed with an electronic computer.

Chayko, M.↗
NASA NTRS1986

Acoustic levitation methods for density measurements

The capability of ultrasonic levitators operating in air to perform density measurements has been demonstrated. The remote determination of the density of ordinary liquids as well as low density solid metals can be carried out using levitated samples with size on the order of a few millimeters and at a frequency of 20 kHz. Two basic methods may be used. The first one is derived from a previously known technique developed for acoustic levitation in liquid media, and is based on the static equilibrium position of levitated samples in the earth's gravitational field. The second approach relies on the dynamic interaction between a levitated sample and the acoustic field. The first technique appears more accurate (1 percent uncertainty), but the latter method is directly applicable to a near gravity-free environment such as that found in space.

Trinh, E. H.↗
NASA NTRS2013

A Mars Sample Return Sample Handling System

We present a sample handling system, a subsystem of the proposed Dragon landed Mars Sample Return (MSR) mission [1], that can return to Earth orbit a significant mass of frozen Mars samples potentially consisting of: rock cores, subsurface drilled rock and ice cuttings, pebble sized rocks, and soil scoops. The sample collection, storage, retrieval and packaging assumptions and concepts in this study are applicable for the NASA's MPPG MSR mission architecture options [2]. Our study assumes a predecessor rover mission collects samples for return to Earth to address questions on: past life, climate change, water history, age dating, understanding Mars interior evolution [3], and, human safety and in-situ resource utilization. Hence the rover will have "integrated priorities for rock sampling" [3] that cover collection of subaqueous or hydrothermal sediments, low-temperature fluidaltered rocks, unaltered igneous rocks, regolith and atmosphere samples. Samples could include: drilled rock cores, alluvial and fluvial deposits, subsurface ice and soils, clays, sulfates, salts including perchlorates, aeolian deposits, and concretions. Thus samples will have a broad range of bulk densities, and require for Earth based analysis where practical: in-situ characterization, management of degradation such as perchlorate deliquescence and volatile release, and contamination management. We propose to adopt a sample container with a set of cups each with a sample from a specific location. We considered two sample cups sizes: (1) a small cup sized for samples matching those submitted to in-situ characterization instruments, and, (2) a larger cup for 100 mm rock cores [4] and pebble sized rocks, thus providing diverse samples and optimizing the MSR sample mass payload fraction for a given payload volume. We minimize sample degradation by keeping them frozen in the MSR payload sample canister using Peltier chip cooling. The cups are sealed by interference fitted heat activated memory alloy caps [5] if the heating does not affect the sample, or by crimping caps similar to bottle capping. We prefer cap sealing surfaces be external to the cup rim to prevent sample dust inside the cups interfering with sealing, or, contamination of the sample by Teflon seal elements (if adopted). Finally the sample collection rover, or a Fetch rover, selects cups with best choice samples and loads them into a sample tray, before delivering it to the Earth Return Vehicle (ERV) in the MSR Dragon capsule as described in [1] (Fig 1). This ensures best use of the MSR payload mass allowance. A 3 meter long jointed robot arm is extended from the Dragon capsule's crew hatch, retrieves the sample tray and inserts it into the sample canister payload located on the ERV stage. The robot arm has capacity to obtain grab samples in the event of a rover failure. The sample canister has a robot arm capture casting to enable capture by crewed or robot spacecraft when it returns to Earth orbit

Wilson, David↗
NASA NTRS1990

HII regions in dwarf irregular galaxies of the local group

Deep, narrowband H alpha Charge Coupled Device (CCD) surveys of HII regions were carried out in several dwarf irregular galaxies in and near the local group. Data are now complete for these galaxies: NGC 6822, GR 8, IC 10, IC 1613, Sextans A, Sextans B, and Sag Irr. Observations are complete for DDO 47, 53, 167, 168 and 187. Details of some of the results for the surveys completed so far are discussed. For NGC 6822, CCD survey at H alpha resulted in the detection of 145 HII regions in the local group irregular galaxy NGC 6822. Most of them are newly detected, faint surface-brightness objects. Positions, maps and dimensions are being published elsewhere. For GR 8, a deep narrowband H alpha imaging of the nearby dwarf irregular galaxy GR 8 revealed a total of 32 HII regions. Positions, H alpha luminosities, and sizes of these objects were determined. The H alpha luminosity function has the same shape as that for more luminous galaxies, except for size of sample effects. Most HII regions detected are at the very low luminosity end of the general luminosity function. For IC 10, a deep CCD narrowband H alpha imaging of the local group dwarf irregular galaxy IC 10 revealed a total of 144 HII regions. Positions, H alpha luminosities, and sizes of these objects were determined. The H alpha luminosity function has the same shape as that for more luminous galaxies.

Hodge, Paul↗
NASA NTRS2024

Chemical and Physical Properties of Cometary Dust

Cometary dust particles are best preserved remnants of the matter present at the onset of the formation of the Solar System. In Space missions, telescopic observations and laboratory analyses advanced the knowledge on the properties of cometary dust. the only samples certified with a cometary origin were returned by the Stardust space mission from comet 81P/Wild2. The “chondritic porous” (here called “chondritic anhydrous”) interplanetary dust particles (CA-IDPs) and micrometeorites (CP-MMs), and the ultracarbonaceous Antarctic MMs (UCAMMs) also show strong evidence for a cometary origin. The elemental composition of cometary dust is generally consistent with the chondritic (CI) composition, with the notable exception of elevated contents in carbon and nitrogen compared with CI. The organic matter of cometary dust is mixed with minor amounts of crystalline (at least 25% of the minerals) and amorphous mineral phases. The most abundant crystalline minerals are ferromagnesian silicates (olivine and low-Ca pyroxenes), but High-Ca pyroxenes, refractory minerals and Low Ni Fe sulfides are also present. The crystalline olivine and low-Ca pyroxene compositions can vary from their Mg-rich end-member (forsterite and enstatite) to relatively Fe-rich compositions. Refractory minerals as well as secondary minerals likeLIME (low-iron, Mn-enriched) olivines, unusual Fe sulfides or mineral aggregates of specific compositions like Kosmochloric high-Ca pyroxene and FeO-rich olivine - KOOL grains) are also found. The presence of carbonates in cometary dust is still debated, but a phyllosilicate-like phase was observed in a UCAMM. The abundance of pyroxene to olivine (in numbers) is larger than in primitive meteorites (e.g. the Px/Ol ratio is usually larger than 1). GEMS phases (glass with embedded metals and sulfides) are abundant in cometary dust, although not systematically found. Some of the organic matter present in cometary dust particle resembles the insoluble organic matter (IOM) present in primitive meteorites, but amorphous carbon and exotic (e.g. N-rich) organic phases are also present. The hydrogen isotopic composition of cometary dust particles (in the organic matter) is usually rich in deuterium, tracing a formation at very low temperatures, either in the protosolar cloud or in the outer regions of the protoplanetary disk. The presolar dust concentration in cometary dust can reach about 1%, which is the most elevated value observed in extraterrestrial samples. The size distribution of cometary dust in comet trails is well represented by a power-law distribution (differential size distribution) with a mean power index N typically ranging from -3 to -4. Polarimetric and light scattering studies of cometary dust suggest mixtures of porous agglomerates of sub-micrometer minerals with organic matter, which is compatible with the in situ analyses of 67P/Churyumov-Gerasimenko by MIDAS (Rosetta) and with the studies of it Stardust samples, CA-IDPs, CP-MMs and UCAMMs. Cometary dust particles have low tensile strength, and low density.

C. Engrand↗
NASA NTRS1976

Inorganic analyses of Martian surface samples at the Viking landing sites

Elemental analyses of fines in the Martian regolith at two widely separated landing sites, Chryse Planitia and Utopia Planitia, produced remarkably similar results. At both sites, the uppermost regolith contains abundant Si and Fe, with significant concentrations of Mg, Al, S, Ca, and Ti. The S concentration is one to two orders of magnitude higher, and K (less than 0.25% by weight) is at least 5 times lower than the average for earth's crust. The trace elements Sr, Y, and possibly Zr have been detected at concentrations near or below 100 parts per million. Pebble-sized fragments sampled at Chryse contain more S than the bulk fines and are thought to be pieces of a sulfate-cemented duricrust.

Clark, B. C.↗
NASA NTRS1984

Structure of nonevaporating sprays - Measurements and predictions

Structure measurements were completed within the dilute portion of axisymmetric nonevaporating sprays (SMD of 30 and 87 microns) injected into a still air environment, including: mean and fluctuating gas velocities and Reynolds stress using laser-Doppler anemometry; mean liquid fluxes using isokinetic sampling; drop sizes using slide impaction; and drop sizes and velocities using multiflash photography. The new measurements were used to evaluate three representative models of sprays: (1) a locally homogeneous flow (LHF) model, where slip between the phases was neglected; (2) a deterministic separated flow (DSF) model, where slip was considered but effects of drop interaction with turbulent fluctuations were ignored; and (3) a stochastic separated flow (SSF) model, where effects of both interphase slip and turbulent fluctuations were considered using random sampling for turbulence properties in conjunction with random-walk computations for drop motion. The LHF and DSF models were unsatisfactory for present test conditions-both underestimating flow widths and the rate of spread of drops. In contrast, the SSF model provided reasonably accurate predictions, including effects of enhanced spreading rates of sprays due to drop dispersion by turbulence, with all empirical parameters fixed from earlier work.

Solomon, A. S. P.↗
NASA NTRSDate not supplied

Structured Covariance Gaussian Networks for Orion Crew Module Aerodynamic Uncertainty Quantification

In this paper we propose a new approach for nonlinear regression and uncertainty quantification. The method is based on a pair of neural networks which parameterize mean and dense covariance functions of a multivariate Gaussian process, trained together to maximize the log-likelihood of observing the given data. The covariance matrix is made positive definite at every input by construction. We also propose a sampling approach that produces viable surrogate function realizations from the Gaussian process. We call the proposed model a Structured Covariance Gaussian Network (SCGN). We illustrate the use of SCGNs for learning an aerodynamic response surface with built-in uncertainty for the Orion crew module. We find that SCGN provides an efficient and systematic way to learn nonlinear functional relationships and dense covariances. We compare results to a baseline Gaussian process regressor and observe that the SCGN provides comparable uncertainty descriptions with improved scalability to dataset size. The sample functions generated by SCGN are fast to evaluate online and are therefore convenient for use in trajectory simulations. These results suggest that SCGN may be a viable computational method for aerodynamic uncertainty quantification.

machine learning↗