Publications

Acoustic testing using commercial sound system components is becoming more popular as a cost effective way of generating the required environment both in and out of a reverberant chamber. This paper will present the development of such a sound system that uses a state-of-the-art random vibration controller to perform closed-loop control in the reverberant chamber at Sandia National Laboratories. Test data will be presented that demonstrates narrow-band controlability, performance and some limitations of commercial sound generation equipment in a reverberant chamber.

We present two ways in which dynamic self-assembly can be used to perform computation, via stochastic protein networks and self-assembling software. We describe our protein-emulating agent-based simulation infrastructure, which is used for both types of computations, and the few agent properties sufficient for dynamic self-assembly. Examples of protein-network-based computation and self-assembling software are presented. We describe some novel capabilities that are enabled by the inherently dynamic nature of the self-assembling executable code. © Springer-Verlag 2004.

Finite difference equations are derived for the simulation of dielectric waveguides using an Hz -Ez formulation defined on a nonuniform triangular grid. The resulting equations may be solved as a banded eigenproblem for waveguide structures of arbitrary shape composed of regions of piecewise constant isotropic dielectric, and all transverse fields then computed from the solutions. Benchmark comparisons are presented for problems with analytic solutions, as well as a sample calculation of the propagation loss of a hollow Bragg fiber.

The structure of laminar inverse diffusion flames (IDF) of methane and ethylene in air was studied using a cylindrical co-flowing burner. IDF were similar to normal diffusion flames, except that the relative positions of the fuel and oxidizer were reversed. Radiation from soot surrounding the IDF masked the reaction zone in visible images. As a result, flame heights determined from visible images were overestimated. The height of the reaction zone as indicated by OH LIF was a more relevant measure of height. The concentration and position of PAH and soot were observed using LIF and laser-induced incandescence (LII). PAH LIF and soot LII indicated that PAH and soot are present on the fuel side of the flame, and that soot is located closer to the reaction zone than PAH. Ethylene flames produced significantly higher PAH LIF and soot LII signals than methane flames, which was consistent with the sooting propensity of ethylene. The soot and PAH were present on the fuel side of the reaction zone, but the soot was closer to the reaction zone than the PAH. This is an abstract of a paper presented at the 30th International Symposium on combustion (Chicago, IL 7/25-30/2004).

This paper describes an array of in-plane piezoelectric actuator segments laminated onto a comer-supported substrate to create a thin bimorph for reflector applications. An electric field distribution over the actuator segments causes the segments to expand or contract, thereby effecting plate deflection. To achieve a desired bimorph shape, the shape is first expressed as a two-dimensional series expansion. Then, using coefficients from the series expansion, an inverse problem is solved that determines the electric field distribution realizing the desired plate shape. A static example is presented where the desired deflection shape is a paraboloid. Copyright © 2004 by ASME.

An important component of ubiquitous computing is the ability to quickly sense the dynamic environment to learn context awareness in real-time. To pervasively capture detailed information of movements, we present a decentralized algorithm for feature extraction within a wireless sensor network. By approaching this problem in a distributed manner, we are able to work within the real constraint of wireless battery power and its effects on processing and network communications. We describe a hardware platform developed for low-power ubiquitous wireless sensing and a distributed feature extraction methodology which is capable of providing more information to the user of events while reducing power consumption. We demonstrate how the collaboration between sensor nodes can provide a means of organizing large networks into information-based clusters. © Springer-Verlag 2004.

Algorithms for multivariate image analysis and other large-scale applications of multivariate curve resolution (MCR) typically employ constrained alternating least squares (ALS) procedures in their solution. The solution to a least squares problem under general linear equality and inequality constraints can be reduced to the solution of a non-negativity-constrained least squares (NNLS) problem. Thus the efficiency of the solution to any constrained least square problem rests heavily on the underlying NNLS algorithm. We present a new NNLS solution algorithm that is appropriate to large-scale MCR and other ALS applications. Our new algorithm rearranges the calculations in the standard active set NNLS method on the basis of combinatorial reasoning. This rearrangement serves to reduce substantially the computational burden required for NNLS problems having large numbers of observation vectors. Copyright © 2005 John Wiley & Sons, Ltd.

We report the cooling of nitric oxide molecules in a single collision between an argon atom and an NO molecule at collision energies of 5.65 ± 0.36 kJ/mol and 14.7 ± 0.9 kJ/mol in a crossed molecular beam apparatus. We have produced in significant numbers (∼108 molecules cm -3 per quantum state) translationally cold NO(2Π 1/2, v′ = 0, j′ = 7.5) molecules in a specific quantum state with an upper-limit laboratory-frame rms velocity of 14.8 ± 1.1 m/s, corresponding to a temperature of 406 ± 28 mK. The translational cooling results from the kinematic collapse of the velocity distribution of the NO molecules after collision. Increasing the collision energy by increasing the velocity of the argon atoms, as we do here, does shift the scattering angle at which the cold molecules appear, but does not result in an experimentally measurable change in the velocity spread of the cold NO. This is entirely consistent with our analysis of the kinematics of the scattering which predicts that the velocity spread will actually decrease with increasing argon atom velocity. © EDP Sciences, Società, Italiana di Fisica, Springer-Verlag 2004.

Abstract not provided.

It seems well understood that supercomputer simulation is an enabler for scientific discoveries, weapons, and other activities of value to society. It also seems widely believed that Moores Law will make progressively more powerful supercomputers over time and thus enable more of these contributions. This paper seeks to add detail to these arguments, revealing them to be generally correct but not a smooth and effortless progression. This paper will review some key problems that can be solved with supercomputer simulation, showing that more powerful supercomputers will be useful up to a very high yet finite limit of around 1021 FLOPS (1 Zettaflops. The review will also show the basic nature of these extreme problems. This paper will review work by others showing that the theoretical maximum supercomputer power is very high indeed, but will explain how a straightforward extrapolation of Moores Law will lead to technological maturity in a few decades. The power of a supercomputer at the maturity of Moores Law will be very high by todays standards at 1016-1019 FLOPS (100 Petaflops to 10 Exaflops, depending on architecture , but distinctly below the level required for the most ambitious applications. Having established that Moores Law will not be that last word in supercomputing, this paper will explore the nearer term issue of what a supercomputer will look like at maturity of Moores Law. Our approach will quantify the maximum performance as permitted by the laws of physics for extension of current technology and then find a design that approaches this limit closely. We study a "multi-architecture" for supercomputers that combines a microprocessor with other "advanced" concepts and find it can reach the limits as well. This approach should be quite viable in the future because the microprocessor would provide compatibility with existing codes and programming styles while the "advanced" features would provide a boost to the limits of performance.

Many practical combustion devices and uncontrolled fires involve high Reynolds number nonpremixed turbulent flames that feature non-equilibrium finite-rate chemistry effects, e.g., local flame extinction and reignition, where enhanced transport of mass and heat away from the flame due to rapid turbulent mixing exceeds the local burning rate. Probability density function methods have shown promise in predicting piloted nonpremixed CH4-air flames over a range of Reynolds numbers and varying degrees of flame extinction and reignition. A study was carried out to quantify and characterize the kinetics of localized extinction and reignition in the Sandia flames D, E, and F, for which detailed velocity and scalar data exists. PDF methods in large eddy simulation to predict the filtered mass density function (FMDF) was used. A simple idealized mixing simulation was performed of a nonpremixed turbulent fuel jet in an air co-flow. Mixing statistics from the Monte Carlo-based FMDF solution of the chemical species scalar were compared to those from a more traditional Eulerian mixing simulation using gradient transport-based subgrid closure models. The FMDF solution will be performed with the Euclidian minimum spanning tree mixing model that uses the phenomenological connection between physical space and state space for mixing events. This is an abstract of a paper presented at the 30th International Symposium on Combustion (Chicago, IL 7/25-30/2004).

The synthesis, characterization, and separations capability of defect-free, thin-film zeolite membranes were presented. The one-micron thick sodium-aluminosilicate films of Silicalite-1 and ZSM-5 were synthesized by hydrothermal methods on either disk- or tube-supports. Techniques for growing membranes on both Al2O3 substrates as well as oxide-coated stainless steel substrates were presented. The resulting defect-free zeolite films had high flux rates at room temperature (∼ 10-7 mole/Pa-sec-sq m) and showed selective separations (3-7) between pure gases of H2 and CH4, O2, N2, CO2, CO, SF6. Results from mixed gas studies showed similar flux rates as pure gases with enhanced selectivity (15-50) for H2. The selectivity through both Silicalite-1 and ZSM-5 membranes was compared and contrasted for several gas mixtures. Data comparisons for defect-free and "defect-filled" membranes were also discussed. Under operation, the flow through these membranes quickly reached its maximum value and was stable over long periods of time. Results from experiments at high temperatures, ≤ 300°C, were compared with the data obtained at room temperature. This is an abstract of a paper presented at the 228th ACS National Meeting (Philadelphia, PA, 8/22-26/2004).

A new constitutive model relating proton conductivity to water content in a polymer electrolyte or membrane is presented. Our constitutive model is based on Faraday's law and the Nernst-Einstein equation; and it depends on the molar volumes of dry membrane and water but otherwise requires no adjustable parameters. We derive our constitutive model in two different ways. Predictions of proton conductivity as a function of membrane water content computed from our constitutive model are compared with that from a representative correlation and other models as well as experimental data from the literature and those obtained in our laboratory using a 4-point probe. Copyright © 2004 by ASME.

A coupled-physics analysis code has been developed to simulate the electrical, thermal, and mechanical responses of surface micromachined (SMM) actuators. Our objective is to optimize the design and performance of these micro actuators. Since many new designs of these electro-thermal actuators have shuttles or platforms between beams, calculating the local Joule heating requires a multi-dimensional electrostatics analysis. Moreover, the electrical solution is strongly coupled to the temperature distribution since the electrical resistivity is temperature dependent. Thus, it is essential to perform a more comprehensive simulation that solves the coupled electrostatics, thermal, and mechanical equations. Results of the coupled-physics analyses will be presented. Copyright © 2004 by ASME.

The process of removing liquid water droplets in polymer electrolyte fuel cells (PEFC) is examined using a simple analytical model and two-dimensional simulations. Specifically, the stability of a droplet adhering to the wall of the cathode flow channel is examined as a function of the geometry of the flow channel, the applied pressure gradient, and the wetting properties. The result is a prediction of the critical droplet size as a function of the difference between the advancing and receding contact angles, or contact angle hysteresis. The analytical model is shown to qualitatively predict this stability limit when compared to two-dimensional simulation results. The simulations are performed using both Arbitrary Lagrangian Eulerian (ALE) methods and level set methods. The ALE and level set predictions are shown to be in good agreement. Copyright © 2004 by ASME.

Surface micromachined structures with high aspect ratios are often utilized as sensor platforms in microelectromechanical systems (MEMS) devices. These structures generally fail by suction or adhesion to the underlying substrate during operation, or related initial processing. Such failures represent a major disadvantage in mass production of MEMS devices with highly compliant structures. Fortunately, most suction failures can be prevented or repaired in a number of ways. Passive approaches implemented during fabrication or release include: (1) utilizing special low adhesion coatings and (2) processing with low surface energy rinse agents. These methods, however, increase both the processing time and cost and are not entirely effective. Active approaches, such as illuminating stiction-failed microstructures with pulsed laser irradiation, have proven to be very effective for stiction repair [1-5]. A more recent and promising method, introduced by Gupta et al. [6], utilized laser-induced stress waves to repair stiction-failed microstructures. This approach represents a logical extension of the laser spallation technique for debonding thin films from substrates [7-9]. The method transmits stress waves into MEMS structures by laser-irradiating the back side of the substrate opposite the stiction-failed structures. This paper presents an experimental study that compares the stress wave repair method with the thermomechanical repair method on identical arrays of stiction-failed cantilevers. Copyright © 2004 by ASME.

X-ray radiography has long been recognized as a valuable tool for detecting internal features and flaws. Recent developments in microfabrication and composite materials have extended inspection requirements to the resolution limits of conventional radiography. Our work has been directed toward pushing both detection and measurement capabilities to a smaller scale. Until recently, we have used conventional contact radiography, optimized to resolve small features. With the recent purchase of a nano-focus (sub-micron) x-ray source, we are now investigating projection radiography, phase contrast imaging and micro-computed tomography (μ-CT). Projection radiography produces a magnified image that is limited in spatial resolution mainly by the source size, not by film grain size or detector pixel size. Under certain conditions phase contrast can increase the ability to resolve small features such as cracks, especially in materials with low absorption contrast. Micro-computed tomography can provide three-dimensional measurements on a micron scale and has been shown to provide better sensitivity than simple radiographs. We have included applications of these techniques to small-scale measurements not easily made by mechanical or optical means. Examples include void detection in meso-scale nickel MEMS parts, measurement of edge profiles in thick gold lithography masks, and characterization of the distribution of phases in composite materials. Our work, so far, has been limited to film. Copyright © 2004 by ASME.

Three-dimensional seismic wave propagation within a heterogeneous isotropic poroelastic medium is simulated with an explicit, time-domain, finite-difference algorithm. A system of thirteen, coupled, first-order partial differential equations is solved for the velocity vector components, stress tensor components, and pressure associated with solid and fluid constituents of the composite medium. A massively parallel computational implementation, utilizing the spatial domain decomposition strategy, allows investigation of large-scale earth models and/or broadband wave propagation within reasonable execution times.

In Combinatorial Optimization, one is frequently faced with linear programming (LP) problems with exponentially many constraints, which can be solved either using separation or what we call compact optimization. The former technique relies on a separation algorithm, which, given a fractional solution, tries to produce a violated valid inequality. Compact optimization relies on describing the feasible region of the LP by a polynomial number of constraints, in a higher dimensional space. A commonly held belief is that compact optimization does not perform as well as separation in practice. In this paper,we report on an application in which compact optimization does in fact largely outperform separation. The problem arises in structural proteomics, and concerns the comparison of 3-dimensional protein folds. Our computational results show that compact optimization achieves an improvement of up to two orders of magnitude over separation. We discuss some reasons why compact optimization works in this case but not, e.g., for the LP relaxation of the TSP. © Springer-Verlag 2004.

This report summarizes a series of structural calculations that examine effects of raising the Waste Isolation Pilot Plant repository horizon from the original design level upward 2.43 meters. These calculations allow evaluation of various features incorporated in conceptual models used for performance assessment. Material presented in this report supports the regulatory compliance re-certification, and therefore begins by replicating the calculations used in the initial compliance certification application. Calculations are then repeated for grid changes appropriate for the new horizon raised to Clay Seam G. Results are presented in three main areas: 1. Disposal room porosity, 2. Disturbed rock zone characteristics, and 3. Anhydrite marker bed failure. No change to the porosity surface for the compliance re-certification application is necessary to account for raising the repository horizon, because the new porosity surface is essentially identical. The disturbed rock zone evolution and devolution are charted in terms of a stress invariant criterion over the regulatory period. This model shows that the damage zone does not extend upward to MB 138, but does reach MB 139 below the repository. Damaged salt would be expected to heal in nominally 100 years. The anhydrite marker beds sustain states of stress that promote failure and substantial marker bed deformation into the room assures fractured anhydrite will sustain in the proximity of the disposal rooms.

This report documents the results obtained during a one-year Laboratory Directed Research and Development (LDRD) initiative aimed at investigating coupled structural acoustic interactions by means of algorithm development and experiment. Finite element acoustic formulations have been developed based on fluid velocity potential and fluid displacement. Domain decomposition and diagonal scaling preconditioners were investigated for parallel implementation. A formulation that includes fluid viscosity and that can simulate both pressure and shear waves in fluid was developed. An acoustic wave tube was built, tested, and shown to be an effective means of testing acoustic loading on simple test structures. The tube is capable of creating a semi-infinite acoustic field due to nonreflecting acoustic termination at one end. In addition, a micro-torsional disk was created and tested for the purposes of investigating acoustic shear wave damping in microstructures, and the slip boundary conditions that occur along the wet interface when the Knudsen number becomes sufficiently large.

Motivated by observations about job runtimes on the CPlant system, we use a trace-driven microsimulator to begin characterizing the performance of different classes of allocation algorithms on jobs with different communication patterns in space-shared parallel systems with mesh topology. We show that relative performance varies considerably with communication pattern. The Paging strategy using the Hilbert space-filling curve and the Best Fit heuristic performed best across several communication patterns.

We have made progress in developing a new statistical mechanics approach to designing self organizing systems that is unique to SNL. The primary application target for this ongoing research has been the development of new kinds of nanoscale components and hardware systems. However, this research also enables an out of the box connection to the field of software development. With appropriate modification, the collective behavior physics ideas for enabling simple hardware components to self organize may also provide design methods for a new class of software modules. Our current physics simulations suggest that populations of these special software components would be able to self assemble into a variety of much larger and more complex software systems. If successful, this would provide a radical (disruptive technology) path to developing complex, high reliability software unlike any known today. This high risk, high payoff opportunity does not fit well into existing SNL funding categories, as it is well outside of the mainstreams of both conventional software development practices and the nanoscience research area that spawned it. This LDRD effort was aimed at developing and extending the capabilities of self organizing/assembling software systems, and to demonstrate the unique capabilities and advantages of this radical new approach for software development.

Biological systems create proteins that perform tasks more efficiently and precisely than conventional chemicals. For example, many plants and animals produce proteins to control the freezing of water. Biological antifreeze proteins (AFPs) inhibit the solidification process, even below the freezing point. These molecules bond to specific sites at the ice/water interface and are theorized to suppress solidification chemically or geometrically. In this project, we investigated the theoretical and experimental data on AFPs and performed analyses to understand the unique physics of AFPs. The experimental literature was analyzed to determine chemical mechanisms and effects of protein bonding at ice surfaces, specifically thermodynamic freezing point depression, suppression of ice nucleation, decrease in dendrite growth kinetics, solute drag on the moving solid/liquid interface, and stearic pinning of the ice interface. Stearic pinning was found to be the most likely candidate to explain experimental results, including freezing point depression, growth morphologies, and thermal hysteresis. A new stearic pinning model was developed and applied to AFPs, with excellent quantitative results. Understanding biological antifreeze mechanisms could enable important medical and engineering applications, but considerable future work will be necessary.

An estimate of the distribution of fatigue ranges or extreme loads for wind turbines may be obtained by separating the problem into two uncoupled parts, (1) a turbine specific portion, independent of the site and (2) a site-specific description of environmental variables. We consider contextually appropriate probability models to describe the turbine specific response for extreme loads or fatigue. The site-specific portion is described by a joint probability distribution of a vector of environmental variables, which characterize the wind process at the hub-height of the wind turbine. Several approaches are considered for combining the two portions to obtain an estimate of the extreme load, e.g., 50-year loads or fatigue damage. We assess the efficacy of these models to obtain accurate estimates, including various levels of epistemic uncertainty, of the turbine response.