Publications

This is meant as a place to put commentary on the whitepaper and is meant to be pretty much ad-hoc. Because the whitepaper describes a potential program in DOE ASCR and because it concerns many researchers in the field, these notes are meant to be extendable by anyone willing to put in the effort. Of course criticisms of the contents of the notes themselves are also welcome.

Abstract not provided.

Abstract not provided.

A recently established mathematical equivalence-between weakly perturbed Huygens fronts (e.g., flames in weak turbulence or geometrical-optics wave fronts in slightly nonuniform media) and the inviscid limit of white-noise-driven Burgers turbulence-motivates theoretical and numerical estimates of Burgers-turbulence properties for specific types of white-in-time forcing. Existing mathematical relations between Burgers turbulence and the statistical mechanics of directed polymers, allowing use of the replica method, are exploited to obtain systematic upper bounds on the Burgers energy density, corresponding to the ground-state binding energy of the directed polymer and the speedup of the Huygens front. The results are complementary to previous studies of both Burgers turbulence and directed polymers, which have focused on universal scaling properties instead of forcing-dependent parameters. The upper-bound formula can be heuristically understood in terms of renormalization of a different kind from that previously used in combustion models, and also shows that the burning velocity of an idealized turbulent flame does not diverge with increasing Reynolds number at fixed turbulence intensity, a conclusion that applies even to strong turbulence. Numerical simulations of the one-dimensional inviscid Burgers equation using a Lagrangian finite-element method confirm that the theoretical upper bounds are sharp within about 15% for various forcing spectra (corresponding to various two-dimensional random media). These computations provide a quantitative test of the replica method. The inferred nonuniversality (spectrum dependence) of the front speedup is of direct importance for combustion modeling. © 2008 The American Physical Society.

Filtered mass density functions (FMDFs) of mixture fraction and temperature are studied by analyzing experimental data obtained from one-dimensional Raman/Rayleigh/LIF measurements of nonpremixed CH4/H2/N2 turbulent jet flames at Reynolds numbers of 15,200 and 22,800 (DLR-A and -B). The experimentally determined FMDFs are conditioned on the Favré filtered values of the mixture fraction and its variance. Filter widths are selected as fixed multiples of the experimentally determined dissipation length scale at each measurement location. One-dimensional filtering using a top-hat filter is performed to obtain the filtered variables used for conditioning. The FMDFs are obtained by binning the mass and filter kernel weighted samples. Emphasis is placed on the shapes of the FMDFs in the fuel-rich, fuel-lean, and stoichiometric intervals for the Favré filtered mixture fraction, and low, medium, and high values for the Favré filtered mixture fraction variance. It is found that the FMDFs of mixture fraction are unimodal in samples with low mixture fraction variance and bimodal in samples with high variance. However, the FMDFs of mixture fraction at the smallest filter size studied are unimodal for all values of the variance. The FMDFs of temperature are unimodal in samples with low mixture fraction variance, and either unimodal or bimodal, depending on the mixture fraction mean, in samples with high variance. The influence of the filter size and the jet Reynolds number on the FMDFs is also considered. © 2008 The Combustion Institute.

The current trend in high performance computing is to aggregate ever larger numbers of processing and interconnection elements in order to achieve desired levels of computational power, This, however, also comes with a decrease in the Mean Time To Interrupt because the elements comprising these systems are not becoming significantly more robust. There is substantial evidence that the Mean Time To Interrupt vs. number of processor elements involved is quite similar over a large number of platforms. In this paper we present a system that uses hardware level monitoring coupled with statistical analysis and modeling to select processing system elements based on where they lie in the statistical distribution of similar elements. These characterizations can be used by the scheduler/resource manager to deliver a close to optimal set of processing elements given the available pool and the reliability requirements of the application. © 2008 IEEE.

The goal of this research was to examine foundational methods, both computational and theoretical, that can improve the veracity of entity-based complex system models and increase confidence in their predictions for emergent behavior. The strategy was to seek insight and guidance from simplified yet realistic models, such as cellular automata and Boolean networks, whose properties can be generalized to production entity-based simulations. We have explored the usefulness of renormalization-group methods for finding reduced models of such idealized complex systems. We have prototyped representative models that are both tractable and relevant to Sandia mission applications, and quantified the effect of computational renormalization on the predictive accuracy of these models, finding good predictivity from renormalized versions of cellular automata and Boolean networks. Furthermore, we have theoretically analyzed the robustness properties of certain Boolean networks, relevant for characterizing organic behavior, and obtained precise mathematical constraints on systems that are robust to failures. In combination, our results provide important guidance for more rigorous construction of entity-based models, which currently are often devised in an ad-hoc manner. Our results can also help in designing complex systems with the goal of predictable behavior, e.g., for cybersecurity.

Abstract not provided.

One of the authors previously conjectured that the wrinkling of propagating fronts by weak random advection increases the bulk propagation rate (turbulent burning velocity) in proportion to the 4/3 power of the advection strength. An exact derivation of this scaling is reported. The analysis shows that the coefficient of this scaling is equal to the energy density of a lower-dimensional Burgers fluid with a white-in-time forcing whose spatial structure is expressed in terms of the spatial autocorrelation of the flow that advects the front. The replica method of field theory has been used to derive an upper bound on the coefficient as a function of the spatial auto-correlation. High precision numerics show that the bound is usefully sharp. Implications for strongly advected fronts (e.g., turbulent flames) are noted.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.