Sublinear Algorithms for Massive Data Sets
This is the final SAND report for the Early-Career LDRD (# 158477) “Sublinear Algorithms for Massive Data Sets”. We provide a list of the various publications and achievements in the project. 3
SEARCH · Engineering Papers
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.
This is the final SAND report for the Early-Career LDRD (# 158477) “Sublinear Algorithms for Massive Data Sets”. We provide a list of the various publications and achievements in the project. 3
This SAND report summarizes the activities and outcomes of the Network and Ensemble Enabled Entity Extraction in Information Text (NEEEEIT) LDRD project, which addressed improving the accuracy of conditional random fields for named entity recognition through the use of ensemble methods.
A key challenge of nuclear forensics is reducing the time and manpower effort required to measure nuclear debris compositions. The overall motivation for this work is to explore development of a robust, automated system that can be used to concurrently analyze several elements/isotopes associated with the forensic signature of nuclear materials. The primary focus of this research has been to methodically investigate if rapid partial leaching of post-detonation debris can yield usable elemental and isotopic information for interpretation. The unique requirements of post-detonation nuclear forensics have not been fully adapted to or fully incorporated contemporary chemical separation techniques. Challenges include addressing the range of material matrices or mixed fission product and actinide compositions and concentrations that might be encountered. These include, but are not limited to, puddle melt glass, urban debris, seawater, air filters, iron-rich urban debris, asphalt, and silica sand. Separation of these elements and their subsequent measurement is a key element of related laboratory analysis activity. Existing practices at LLNL rely on proven but time-consuming and labor intensive processes. Significant time and labor savings are possible in chemical separations, however, if rapid processing methods can be adapted to post-detonation debris. Development of a simple and reliable leaching technique could shorten analytical times and would be useful as a field deployable method for the preliminary characterization of actinide isotopic ratios in soils. Measurement of isotopic ratios in the field using modern mass spectrometry capabilities such as Inductively Coupled Plasma Mass Spectrometry (ICP-MS) is desirable, taking advantage of the extended range of isotopic systems measureable using such instruments. Sample introduction to these types of mass spectrometry instruments requires partial leaching or full dissolution of a sample to remove isobaric (same mass) interferences, and, in some cases, to concentrate the elements(s) of interest. To develop a field-deployable mass spectrometry capability, therefore, automated and robust leaching of likely debris samples (ranging from silicates and oxides to metals and urban materials such as concrete and asphalt), followed by separation/purification through cation exchange column chemistry is necessary. In a post-detonation environment, analysis of melt glasses via rapid leaching and ICP-MS could be a viable route to the same goal. This report presents initial leaching experiments on ‘uncontaminated’ soils, as well as data from melt glass from a single nuclear weapons test. Samples were characterized by gamma spectrometry, then aliquoted for rapid leaching experiments. Experiments were conducted using two different rapid acid treatments to leach the soils. Following leaching, the leachate solutions were analyzed by ICP-MS to determine if U isotopic composition. We present these data to address the question as to whether or not rapid (~1 hr) leaching techniques have the potential to yield meaningful U isotopic compositions, without the need for a complete (time consuming) sample dissolution and separation.
This report documents the statistical analyses performed (by Pacific Northwest National Laboratory for Washington River Protection Solutions) on data from 26 tests conducted using two scaled tanks (43 and 120 inches) in the Small Scale Mixing Demonstration platform. The 26 tests varied several test parameters, including mixer-jet nozzle velocity, base simulant, supernatant viscosity, and capture velocity. For each test, samples were taken pre-transfer and during five batch transfers. The samples were analyzed for the concentrations (lbs/gal slurry) of four primary components in the base simulants (gibbsite, stainless steel, sand, and ZrO2). The statistical analyses including modeling the component concentrations as functions of test parameters using stepwise regression with two different model forms. The resulting models were used in an equivalent performance approach to calculate values of scaling exponents (for a simple geometric scaling relationship) as functions of the parameters in the component concentration models. The resulting models and scaling exponents are displayed in tables and graphically. The sensitivities of component concentrations and scaling exponents to the test parameters are presented graphically. These results will serve as inputs to subsequent work by other researchers to develop scaling relationships that are applicable to full-scale tanks.
Current vehicles use mostly ferrous components for structural applications. It is possible to reduce the weight of the vehicle by substituting these parts with those made from light metals such as aluminum and magnesium. Many alloys and manufacturing processes can be used to produce these light metal components and casting is known to be most economical. One of the high integrity casting processes is permanent mold casting which is the focus of this research report. Many aluminum alloy castings used in automotive applications are produced by the sand casting process. Also, aluminum-silicon (Al-Si) alloys are the most widely used alloy systems for automotive applications. It is possible that by using high strength aluminum alloys based on an aluminum-copper (Al-Cu) system and permanent mold casting, the performance of these components can be enhanced significantly. This will also help to further reduce the weight. However, many technological obstacles need to be overcome before using these alloys in automotive applications in an economical way. There is very limited information in the open literature on gravity and low-pressure permanent mold casting of high strength aluminum alloys. This report summarizes the results and issues encountered during the casting trials of high strength aluminum alloy 206.0 (Al-Cu alloy) and moderate strength alloy 535.0 (Al-Mg alloy). Five engineering components were cast by gravity tilt-pour or low pressure permanent mold casting processes at CanmetMATERIALS (CMAT) and two production foundries. The results of the casting trials show that high integrity engineering components can be produced successfully from both alloys if specific processing parameters are used. It was shown that a combination of melt processing and mold temperature is necessary for the elimination of hot tears in both alloys.
Inclusions in steel castings can cause rework, scrap, poor machining, and reduced casting performance, which can obviously result in excess energy consumption. Significant progress in understanding inclusion source, formation and control has been made. Inclusions can be defined as non-metallic materials such as refractory, sand, slag, or coatings, embedded in a metallic matrix. This research project has focused on the mold filling aspects to examine the effects of pouring methods and gating designs on the steel casting cleanliness through water modeling, computer modeling, and melting/casting experiments. Early in the research project, comprehensive studies of bottom-pouring water modeling and low-alloy steel casting experiments were completed. The extent of air entrainment in bottom-poured large castings was demonstrated by water modeling. Current gating systems are designed to prevent air aspiration. However, air entrainment is equally harmful and no prevention measures are in current practice. In this study, new basin designs included a basin dam, submerged nozzle, and nozzle extension. The entrained air and inclusions from the gating system were significantly reduced using the new basin method. Near the end of the project, there has been close collaboration with Wescast Industries Inc., a company manufacturing automotive exhaust components. Both computer modeling using Magma software and melting/casting experiments on thin wall turbo-housing stainless steel castings were completed in this short period of time. Six gating designs were created, including the current gating on the pattern, non-pressurized, partially pressurized, naturally pressurized, naturally pressurized without filter, and radial choke gating without filter, for Magma modeling. The melt filling velocity and temperature were determined from the modeling. Based on the simulation results, three gating designs were chosen for further melting and casting experiments on the same casting pattern using the lip pouring method. It was observed again that gating designs greatly influenced the melt filling velocity and the number of inclusion defects. The radial choked gating showed improvements in casting cleanliness and yield over the other gatings, even though no mold filters were used in the gating system.
The Zion Nuclear Power Station is being decommissioned. ZionSolutions proposes to leave much of the below grade structures in place and to fill them with a backfill to provide structural support. Backfills under consideration include “clean” concrete demolition debris from the above grade parts of the facility, a flowable grout, cinder block construction debris and sand. A previous study (Yim, 2012) examined the sorption behavior of five nuclides (Fe-55, Co-60, Ni-63, Sr-85, and Cs-137) on concrete and local soils. This study, commissioned by ZionSolutions and conducted by the Brookhaven National Laboratory (BNL) examines the sorption behavior on cinder block and grout materials. Specifically, this study measured the distribution coefficient for four radionuclides of concern using site-groundwater and cinder block from the Zion site and a flowable grout. The distributions coefficient is a measure of the amount of the radionuclide that will remain sorbed to the solid material that is present relative to the amount that will remain in solution. A high distribution coefficient indicates most of the radionuclide will remain on the solid material and will not be available for transport by the groundwater. The radionuclides examined in this set of tests were Co-60, Ni-63, Sr-85, and Cs-137. Tests were performed following ASTM C1733-10, Standard Test Methods for Distribution Coefficients of Inorganic Species by the Batch Method. Sr-85 was used in the testing as an analogue for Sr-90 because it behaves similarly with respect to sorption and has a gamma emission that is easier to detect than the beta emission from Sr-90.
During February and March 2014, Pacific Northwest National Laboratory conducted hydraulic (slug) tests at 200-DV-1 Operable Unit wells 299-E33-350 (C8914) and 299-E33-351 (C8915) as part of B-Complex Area Perched-Water characterization activities at the Hanford Site 200-East Area. During the construction/completion phase of each well, two overlapping depth intervals were tested within the unconfined perched-water aquifer contained in the silty-sand subunit of the Cold Creek Unit. The purpose of the slug-test characterization was to provide estimates of transmissivity and hydraulic conductivity for the perched-water aquifer at these selected well locations.
Monitoring infections in vectors such as mosquitoes,sand flies, tsetse flies, and ticks to identify human pathogens may serve as an early warning detection system to direct local government disease preventive measures. One major hurdle in detection is the ability to screen large numbers of vectors for human pathogens without the use of genotype-specific molecular techniques. Next generation sequencing (NGS) provides an unbiased platform capable of identifying known and unknown pathogens circulating within a vector population, but utilizing this technology is time-consuming and costly for vector-borne disease surveillance programs. To address this we developed cost-effective Ilumina® RNA-Seq library preparation methodologiesin conjunction with an automated computational analysis pipeline to characterize the microbial populations circulating in Culex mosquitoes (Culex quinquefasciatus, Culex quinquefasciatus/pipiens complex hybrids, and Culex tarsalis) throughout California. We assembled 20 novel and well-documented arboviruses representing members of Bunyaviridae, Flaviviridae, Ifaviridae, Mesoniviridae, Nidoviridae, Orthomyxoviridae, Parvoviridae, Reoviridae, Rhabdoviridae, Tymoviridae, as well as several unassigned viruses. In addition, we mapped mRNA species to divergent species of trypanosoma and plasmodium eukaryotic parasites and characterized the prokaryotic microbial composition to identify bacterial transcripts derived from wolbachia, clostridium, mycoplasma, fusobacterium and campylobacter bacterial species. We utilized these microbial transcriptomes present in geographically defined Culex populations to define spatial and mosquito species-specific barriers of infection. The virome and microbiome composition identified in each mosquito pool provided sufficient resolution to determine both the mosquito species and the geographic region in California where the mosquito pool originated. This data provides insight into the complexity of microbial species circulating in medically important Culex mosquitoes and their potential impact on the transmission of vector-borne human/veterinary pathogens in California.
The purpose of this study was to develop methodologies for establishing baseline datasets and statistical protocols for determining statistically significant changes between background concentrations and predicted concentrations that would be used to represent a contamination plume in the Gen II models being developed by NRAP’s Groundwater Protection team. The initial effort examined selected portions of two aquifer systems; the urban shallow-unconfined aquifer system of the Edwards-Trinity Aquifer System (being used to develop the ROM for carbon-rock aquifers, and the a portion of the High Plains Aquifer (an unconsolidated and semi-consolidated sand and gravel aquifer, being used to development the ROM for sandstone aquifers). Threshold values were determined for Cd, Pb, As, pH, and TDS that could be used to identify contamination due to predicted impacts from carbon sequestration storage reservoirs, based on recommendations found in the EPA’s ''Unified Guidance for Statistical Analysis of Groundwater Monitoring Data at RCRA Facilities'' (US Environmental Protection Agency 2009). Results from this effort can be used to inform a ''no change'' scenario with respect to groundwater impacts, rather than the use of an MCL that could be significantly higher than existing concentrations in the aquifer.
A portable instrument has been developed to assess the functionality of filter sand o-rings on nuclear material storage canisters, without requiring removal of the canister lid. Additionally, a set of fifteen filter standards were procured for verifying aerosol leakage and pressure drop measurements in the Los Alamos Filter Test System. The US Department of Energy uses several thousand canisters for storing nuclear material in different chemical and physical forms. Specialized filters are installed into canister lids to allow gases to escape, and to maintain an internal ambient pressure while containing radioactive contaminants. Diagnosing the condition of container filters and canister integrity is important to ensure worker and public safety and for determining the handling requirements of legacy apparatus. This report describes the In-Place-Filter-Tester, the Instrument Development Plan and the Instrument Operating Method that were developed at the Los Alamos National Laboratory to determine the “as found” condition of unopened storage canisters. The Instrument Operating Method provides instructions for future evaluations of as-found canisters packaged with nuclear material. Customized stainless steel canister interfaces were developed for pressure-port access and to apply a suction clamping force for the interface. These are compatible with selected Hagan-style and SAVY-4000 storage canisters that were purchased from NFT (Nuclear Filter Technology, Golden, CO). Two instruments were developed for this effort: an initial Los Alamos POC (Proof-of-Concept) unit and the final Los Alamos IPFT system. The Los Alamos POC was used to create the Instrument Development Plan: (1) to determine the air flow and pressure characteristics associated with canister filter clogging, and (2) to test simulated configurations that mimicked canister leakage paths. The canister leakage scenarios included quantifying: (A) air leakage due to foreign material (i.e. dust and hair) fouling of o-rings, (B) leakage through simulated cracks in o-rings, and (C) air leakage due to inadequately tightened canister lids. The Los Alamos POC instrument determined pertinent air flow and pressure quantities, and this knowledge was used to specify a customized Isaac® (Z axis, Salt Lake City, UT) leak test module. The final Los Alamos IPFT (incorporating the Isaac® leak test module) was used to repeat the tests in the Instrument Development Plan (with simulated filter clogging tests and canister leak pathway tests). The Los Alamos IPFT instrument is capable of determining filter clogging and leak rate conditions, without requiring removal of the container lid. The IPFT measures pressure decay rate from 1.7E-03 in WC/sec to 1.7E-01 in WC/sec. On the same unit scale, helium leak testing of canisters has a range from 5.7E-07 in WC/sec to 1.9E-03 in WC/sec. For a 5-quart storage canister, the IPFT measures equivalent leak flow rates from 0.03 to 3.0 cc/sec. The IPFT does not provide the same sensitivity as helium leak testing, but is able to gauge the assembled condition of as-found and in-situ canisters.
REALGASBRINE v1.0 is a numerical code that for the simulation of the behavior of gas-bearing porous and/fractured geologic media. It is an option of TOUGH+ v1.5 [Moridis, 2014], a successor to the TOUGH2 [Pruess et al., 1999] family of codes for multi-component, multiphase fluid and heat flow developed at the Lawrence Berkeley National Laboratory. REALGASBRINE v1.0 needs the TOUGH+ v1.5 core code in order to compile and execute. It is written in standard FORTRAN 95/2003, and can be run on any computational platform (workstation, PC, Macintosh) for which such compilers are available. REALGASBRINE v1.0 describes the non-isothermal two- (for pure water) or three-phase (for brine) flow of an aqueous phase and a real gas mixture in a gas-bearing medium, with a particular focus in ultra-tight (such as tight-sand and shale gas) systems. Up to 12 individual real gases can be tracked, and salt can precipitate as solid halite. The capabilities of the code include coupled flow and thermal effects, real gas behavior, Darcy and non-Darcy flow, several isotherm options of gas sorption onto the grains of the porous media, complex fracture descriptions, gas solubility into water, and geomechanical effects on flow properties. REALGASBRINE v1.0 allows the study of flow and transport of fluids and heat over a wide range of time frames and spatial scales not only in gas reservoirs, but also in any problem involving the flow of gases in geologic media, including the geologic storage of greenhouse gas mixtures, the behavior of geothermal reservoirs with multi-component condensable (H 2 O and CO 2 ) and non-condensable gas mixtures, the transport of water and released H 2 in nuclear waste storage applications, etc.
This technical report summarizes some recently developed approaches to studies of rock properties at a pore scale. Digital rock approach is complementary to laboratory and field studies. It can be especially helpful in situations where experimental data are uncertain, or are difficult or impossible to obtain. Digitized binary images of the pore geometries of natural rocks obtained by different imaging techniques are the input data. Computer-generated models of natural rocks can be used instead of images in a case where microtomography data are unavailable, or the resolution of the tools is insufficient to adequately characterize the features of interest. Simulations of creeping viscous flow in pores produce estimates of Darcy permeability. Maximal Inscribed Spheres calculations estimate two-phase fluid distribution in capillary equilibrium. A combination of both produce relative permeability curves. Computer-generated rock models were employed to study two-phase properties of fractured rocks, or tight sands with slit-like pores, too narrow to be characterized with micro-tomography. Various scenarios can simulate different fluid displacement mechanisms, from piston-like drainage to liquid dropout at the dew point. A finite differences discretization of Stokes equation is developed to simulate flow in the pore space of natural rocks. The numerical schemes are capable to handle both no-slip and slippage flows. An upscaling procedure estimates the permeability by subsampling a large data set. Capillary equilibrium and capillary pressure curves are efficiently estimated with the method of maximal inscribed spheres both an arbitrary contact angle. The algorithms can handle gigobytes of data on a desktop workstation. Customized QuickHull algorithms model natural rocks. Capillary pressure curves evaluated from computer-generated images mimic those obtained for microtomography data.
Sandia's Dakota software (available at http://dakota.sandia.gov) supports science and engineering transformation through advanced exploration of simulations. Specifically it manages and analyzes ensembles of simulations to provide broader and deeper perspective for analysts and decision makers. This enables them to enhance understanding of risk, improve products, and assess simulation credibility. This manual offers Consortium for Advanced Simulation of Light Water Reactors (LWRs) (CASL) partners a guide to conducting Dakota-based VUQ studies for CASL problems. It motivates various classes of Dakota methods and includes examples of their use on representative application problems. On reading, a CASL analyst should understand why and how to apply Dakota to a simulation problem. This SAND report constitutes the product of CASL milestone L3:VUQ.V&V.P8.01 and is also being released as a CASL unlimited release report with number CASL-U-2014-0038-000.
Geomechanics Technologies has completed a detailed characterization study of the Wilmington Graben offshore Southern California area for large-scale CO₂ storage. This effort has included: an evaluation of existing wells in both State and Federal waters, field acquisition of about 175 km (109 mi) of new seismic data, new well drilling, development of integrated 3D geologic, geomechanics, and fluid flow models for the area. The geologic analysis indicates that more than 796 MMt of storage capacity is available within the Pliocene and Miocene formations in the Graben for midrange geologic estimates (P50). Geomechanical analyses indicate that injection can be conducted without significant risk for surface deformation, induced stresses or fault activation. Numerical analysis of fluid migration indicates that injection into the Pliocene Formation at depths of 1525 m (5000 ft) would lead to undesirable vertical migration of the CO₂ plume. Recent well drilling however, indicates that deeper sand is present at depths exceeding 2135 m (7000 ft), which could be viable for large volume storage. For vertical containment, injection would need to be limited to about 250,000 metric tons per year per well, would need to be placed at depths greater than 7000ft, and would need to be placed in new wells located at least 1 mile from any existing offset wells. As a practical matter, this would likely limit storage operations in the Wilmington Graben to about 1 million tons per year or less. A quantitative risk analysis for the Wilmington Graben indicate that such large scale CO₂ storage in the area would represent higher risk than other similar size projects in the US and overseas.
This Report characterizes the defects in the defect reaction network in silicon-doped, n-type InAs predicted with first principles density functional theory. The reaction network is deduced by following exothermic defect reactions starting with the initially mobile interstitial defects reacting with common displacement damage defects in Si-doped InAs, until culminating in immobile reaction products. The defect reactions and reaction energies are tabulated, along with the properties of all the silicon-related defects in the reaction network. This Report serves to extend the results for the properties of intrinsic defects in bulk InAs as collated in SAND 2013 - 2477: ''Simple intrinsic defects in InAs : numerical predictions'' to include Si-containing simple defects likely to be present in a radiation-induced defect reaction sequence.
An important question for the Carbon Capture, Storage, and Utility program is “can we adequately predict the CO 2 plume migration?” For tracking CO 2 plume development, the Sleipner project in the Norwegian North Sea provides more time-lapse seismic monitoring data than any other sites, but significant uncertainties still exist for some of the reservoir parameters. In Part I, we assessed model uncertainties by applying two multi-phase compositional simulators to the Sleipner Benchmark model for the uppermost layer (Layer 9) of the Utsira Sand and calibrated our model against the time-lapsed seismic monitoring data for the site from 1999 to 2010. Approximate match with the observed plume was achieved by introducing lateral permeability anisotropy, adding CH 4 into the CO 2 stream, and adjusting the reservoir temperatures. Model-predicted gas saturation, CO 2 accumulation thickness, and CO 2 solubility in brine—none were used as calibration metrics—were all comparable with the interpretations of the seismic data in the literature.
This report provides supplemental information to the assessment of target markets provided in Appendix A of the 2012 Next Generation Nuclear Plant (NGNP) Industry Alliance (NIA) business plan [NIA 2012] for deployment of High Temperature Gas-Cooled Reactors (HTGRs) in the 2025 – 2050 time frame. This report largely reiterates the [NIA 2012] assessment for potential deployment of 400 to 800 HTGR modules (100 to 200 HTGR plants with 4 reactor modules) in the 600-MWt class in North America by 2050 for electricity generation, co-generation of steam and electricity, oil sands operations, hydrogen production, and synthetic fuels production (e.g., coal to liquids). As the result of increased natural gas supply from hydraulic fracturing, the current and historically low prices of natural gas remain a significant barrier to deployment of HTGRs and other nuclear reactor concepts in the U.S. However, based on U.S. Department of Energy (DOE) Energy Information Agency (EIA) data, U.S. natural gas prices are expected to increase by the 2030 – 2040 timeframe when a significant number of HTGR modules could be deployed. An evaluation of more recent EIA 2013 data confirms the assumptions in [NIA 2012] of future natural gas prices in the range of approximately $7/MMBtu to $10/MMBtu during the 2030 – 2040 timeframe. Natural gas prices in this range will make HTGR energy prices competitive with natural gas, even in the absence of carbon-emissions penalties. Exhibit ES-1 presents the North American projections in each market segment including a characterization of the market penetration logic. Adjustments made to the 2012 data (and reflected in Exhibit ES-1) include normalization to the slightly larger 625MWt reactor module, segregation between steam cycle and more advanced (higher outlet temperature) modules, and characterization of U.S. synthetic fuel process applications as a separate market segment.