Beyond 2D-grids: A dependence maximization view on image browsing

Ideally, one would like to perform image search using an intuitive and friendly approach. Many existing image search engines, however, present users with sets of images arranged in some default order on the screen, typically the relevance to a query, only. While this certainly has its advantages, arguably, a more flexible and intuitive way would be to sort images into arbitrary structures such as grids, hierarchies, or spheres so that images that are visually or semantically alike are placed together. This paper focuses on designing such a navigation system for image browsers. This is a challenging task because arbitrary layout structure makes it difficult - if not impossible - to compute cross-similarities between images and structure coordinates, the main ingredient of traditional layouting approaches. For this reason, we resort to a recently developed machine learning technique: kernelized sorting. It is a general technique for matching pairs of objects from different domains without requiring cross-domain similarity measures and hence elegantly allows sorting images into arbitrary structures. Moreover, we extend it so that some images can be preselected for instance forming the tip of the hierarchy allowing to subsequently navigate through the search results in the lower levels in an intuitive way. Copyright 2010 ACM.

The behaviour of laminar stratified methane/air flames in counterflow

Flames propagating through a mixture with a gradient of equivalence ratio have been previously demonstrated to travel faster or slower than their equivalent premixed flames. The present study aims to numerically investigate the response of strained laminar methane-air flames to such gradients. The flames are simulated in a counterflow configuration where a premixed reactant stream at equivalence ratio φR opposes a hot equilibrium stream at equivalence ratio φP. Premixed and stratified flames are compared with respect to the equivalence ratio φ* and the corresponding gradient ∇φ* at the point of peak heat release rate, for three strain rates, a=50, 300 and 500s-1 and a range of φ*. The effect of different stratification levels is also investigated by varying the ratio of φP to φR, Θ. Results indicate that, as long as flames stabilize within the diffusion layer and Θ>1, increased heat release rate Q is seen throughout the progress variable space in comparison to the premixed state. In contrast, an attenuation of heat release rate is seen for Θ<1. The enhancement (or attenuation) of heat release varies monotonically with Θ. The effect of stratification on flame behavior becomes more pronounced as the strain rate increases. The present study reveals the mechanisms for the propagation of quasi-steady stratified flames under lean and rich conditions: stratified flames are primarily dominated by the diffusion of heat under lean conditions, and diffusion of H2 under rich conditions. Thanks to species and thermal support, stratified flames continue to burn beyond the premixed lean and rich flammability limits. Further investigation on the unsteady response of flames to the fluctuating equivalence ratio implies that the steady results represent the unsteady response well, as long as φ* and ∇φ* are similar in both steady and unsteady cases. © 2013 The Combustion Institute.

Complex chemistry DNS of n-heptane spray autoignition at high pressure and intermediate temperature conditions

Direct Numerical Simulations (DNS) of turbulent n-heptane sprays autoigniting at high pressure (P=24bar) and intermediate air temperature (Tair=1000K) have been performed to investigate the physical mechanisms present under conditions where low-temperature chemistry is expected to be important. The initial turbulence in the carrier gas, the global equivalence ratio in the spray region, and the initial droplet size distribution of the spray were varied. Results show that spray ignition exhibits a spotty nature, with several kernels developing independently in those regions where the mixture fraction is close to its most reactive value ξMR (as determined from homogeneous reactor calculations) and the scalar dissipation rate is low. Turbulence reduces the ignition delay time as it promotes mixing between air and the fuel vapor, eventually resulting in lower values of scalar dissipation. High values of the global equivalence ratio are responsible for a larger number of ignition kernels, due to the higher probability of finding regions where ξ=ξMR. Spray polydispersity results in the occurrence of ignition over a wider range of mixture fraction values. This is a consequence of the inhomogeneities in the mixing field that characterize these sprays, where poorly mixed rich spots are seen to alternate with leaner ones which are well-mixed. The DNS simulations presented in this work have also been used to assess the applicability of the Conditional Moment Closure (CMC) method to the simulation of spray combustion. CMC is found to be a valid method for capturing spray autoignition, although care should be taken in the modelling of the unclosed terms appearing in the CMC equations. © 2013 The Combustion Institute.

Effect of operating conditions and fuel type on crevice hc emissions: Model results and comparison with experiments

A one-dimensional model for crevice HC post-flame oxidation is used to calculate and understand the effect of operating parameters and fuel type (propane and isooctane) on the extent of crevice hydrocarbon and the product distribution in the post flame environment. The calculations show that the main parameters controlling oxidation are: bulk burned gas temperatures, wall temperatures, turbulent diffusivity, and fuel oxidation rates. Calculated extents of oxidation agree well with experimental values, and the sensitivities to operating conditions (wall temperatures, equivalence ratio, fuel type) are reasonably well captured. Whereas the bulk gas temperatures largely determine the extent of oxidation, the hydrocarbon product distribution is not very much affected by the burned gas temperatures, but mostly by diffusion rates. Uncertainties in both turbulent diffusion rates as well as in mechanisms are an important factor limiting the predictive capabilities of the model. However, it seems well suited to sensitivity calculations about a baseline. Copyright © 1999 Society of Automotive Engineers, Inc.

Analysis of high pressure premixed flames using Equivalent Reactor Networks for predicting NOx emissions

This paper describes a computational study of lean premixed high pressure methane-air flames, using Computational Fluid Dynamics (CFD) together with a reactor network approach. A detailed chemical reaction mechanism is employed to predict pollutant concentrations, placing emphasis on nitrogen oxide emissions. The reacting flow field is divided into separate zones in which homogeneity of the physical and chemical conditions prevails. The defined zones are interconnected forming an Equivalent Reactor Network (ERN). Three flames are examined for which experimental data is available. Flame A is characterised by an equivalence ratio of 0.43 while Flames B and C are richer with equivalence ratios of 0.5 and 0.56 respectively. Computations are performed for a range of operating conditions, quantifying the effect in the emitted NOx levels. Model predictions are compared against the available experimental data. Sensitivity analysis is performed to investigate the effect of the network size, in order to define the optimum number of reactors for accurate predictions of the species mass fractions. © 2012 Elsevier Ltd. All rights reserved.

Low-rank optimization on the cone of positive semidefinite matrices

We propose an algorithm for solving optimization problems defined on a subset of the cone of symmetric positive semidefinite matrices. This algorithm relies on the factorization X = Y Y T , where the number of columns of Y fixes an upper bound on the rank of the positive semidefinite matrix X. It is thus very effective for solving problems that have a low-rank solution. The factorization X = Y Y T leads to a reformulation of the original problem as an optimization on a particular quotient manifold. The present paper discusses the geometry of that manifold and derives a second-order optimization method with guaranteed quadratic convergence. It furthermore provides some conditions on the rank of the factorization to ensure equivalence with the original problem. In contrast to existing methods, the proposed algorithm converges monotonically to the sought solution. Its numerical efficiency is evaluated on two applications: the maximal cut of a graph and the problem of sparse principal component analysis. © 2010 Society for Industrial and Applied Mathematics.

Low-order modelling of the response of ducted flames in annular geometries

The adoption of lean premixed prevaporised combustion systems can reduce NOx emissions from gas turbines, but unfortunately also increases their susceptibility to thermoacoustic instabilities. Initially, acoustic waves can produce heat release fluctuations by a variety of mechanisms, often by perturbing the equivalence ratio. If correctly phased, heat release fluctuations can subsequently generate more acoustic waves, which at high amplitude can result in significant structural damage to the combustor. The prediction of this phenomenon is of great industrial interest. In previous work, we have coupled a physics based, kinematic model of the flame with a network model to provide the planar acoustic response necessary to close the feedback loop and predict the onset and amplitude of thermoacoustic instabilities in a lab-scale, axisymmetric single burner combustor. The advantage of a time domain approach is that the modal interaction, the influence of harmonics, and flame saturation can be investigated. This paper extends this approach to more realistic, annular geometries, where both planar and circumferential modes must be considered. In lean premixed prevaporised combustors, fluctuations in equivalence ratio have been shown to be a dominant cause of unsteady combustion. These can occur, for example, due to velocity perturbations in the premix ducts, which can lead to equivalence ratio fluctuations at the fuel injectors, which are subsequently convected downstream to the flame surfaces. Here, they can perturb the heat release by locally altering the flame speed, enthalpy of combustion, and, indirectly, the flame surface area. In many gas turbine designs, particularly aeroengines, the geometries are composed of a ring of premix ducts linking a plenum and an annular combustor. The most unstable modes are often circumferential modes. The network model is used to characterise the flow response of the geometry to heat fluctuations at an appropriate location, such as the fuel injectors. The heat release at each flame holder is determined in the time domain using the kinematic flame model derived, as a function of the flow perturbations in the premix duct. This approach is demonstrated for an annular ring of burners on a in a simple geometry. The approach is then extended to an industrial type gas turbine combustor, and used to predict the limit cycle amplitudes. Copyright © 2012 by ASME.

Statistics of reaction progress variable and mixture fraction gradients from DNS of turbulent partially premixed flames

Statistically planar turbulent partially premixed flames for different initial intensities of decaying turbulence have been simulated for global equivalence ratios = 0.7 and 1.0 using three-dimensional, simplified chemistry-based direct numerical simulations (DNS). The simulation parameters are chosen such that the flames represent the thin reaction zones regime combustion. A random bimodal distribution of equivalence ratio is introduced in the unburned gas ahead of the flame to account for the mixture inhomogeneity. The results suggest that the probability density functions (PDFs) of the mixture fraction gradient magnitude |Δξ| (i.e., P(|Δξ|)) can be reasonably approximated using a log-normal distribution. However, this presumed PDF distribution captures only the qualitative nature of the PDF of the reaction progress variable gradient magnitude |Δc| (i.e., P(|Δc|)). It has been found that a bivariate log-normal distribution does not sufficiently capture the quantitative behavior of the joint PDF of |Δξ| and |Δc| (i.e., P(|Δξ|, |Δc|)), and the agreement with the DNS data has been found to be poor in certain regions of the flame brush, particularly toward the burned gas side of the flame brush. Moreover, the variables |Δξ| and |Δc| show appreciable correlation toward the burned gas side of the flame brush. These findings are corroborated further using a DNS data of a lifted jet flame to study the flame geometry dependence of these statistics. © 2013 Copyright Taylor and Francis Group, LLC.

Modal dynamics of self-excited azimuthal instabilities in an annular combustion chamber

In this paper we describe the time-varying amplitude and its relation to the global heat release rate of self-excited azimuthal instabilities in a simple annular combustor operating under atmospheric conditions. The combustor was modular in construction consisting of either 12, 15 or 18 equally spaced premixed bluff-body flames around a fixed circumference, enabling the effect of large-scale interactions between adjacent flames to be investigated. High-speed OH* chemiluminescence imaged from above the annulus and pressure measurements obtained at multiple locations around the annulus revealed that the limit cycles of the modes are degenerate in so much as they undergo continuous transitions between standing and spinning modes in both clockwise (CW) and anti-clockwise (ACW) directions but with the same resonant frequency. Similar behaviour has been observed in LES simulations which suggests that degenerate modes may be a characteristic feature of self-excited azimuthal instabilities in annular combustion chambers. By modelling the instabilities as two acoustic waves of time-varying amplitude travelling in opposite directions we demonstrate that there is a statistical prevalence for either standing m=1 or spinning m=±1 modes depending on flame spacing, equivalence ratio, and swirl configuration. Phase-averaged OH* chemiluminescence revealed a possible mechanism that drives the direction of the spinning modes under limit-cycle conditions for configurations with uniform swirl. By dividing the annulus into inner and outer annular regions it was found that the spin direction coincided with changes in the spatial distribution of the peak heat release rate relative to the direction of the bulk swirl induced along the annular walls. For standing wave modes it is shown that the globally integrated fluctuations in heat release rate vary in magnitude along the acoustic mode shape with negligible contributions at the pressure nodes and maximum contributions at the pressure anti-nodes. © 2013.

System combination and score normalization for spoken term detection

Spoken content in languages of emerging importance needs to be searchable to provide access to the underlying information. In this paper, we investigate the problem of extending data fusion methodologies from Information Retrieval for Spoken Term Detection on low-resource languages in the framework of the IARPA Babel program. We describe a number of alternative methods improving keyword search performance. We apply these methods to Cantonese, a language that presents some new issues in terms of reduced resources and shorter query lengths. First, we show score normalization methodology that improves in average by 20% keyword search performance. Second, we show that properly combining the outputs of diverse ASR systems performs 14% better than the best normalized ASR system. © 2013 IEEE.

Laminar propagation of lean premixed flames ignited in stratified mixture

Combustion in stratified mixtures is envisaged in practical energy systems such as direct-injection spark-ignited (DISI) car engines, gas turbines, for reducing CO2 and pollutant emissions while protecting their efficiency. The mixture gradients change the fundamental properties of the flame, especially by a difference in temperature and composition between the burnt gases and those of a flame consuming a homogeneous mixture. This paper presents an investigation of the properties of the flame propagating in a lean homogeneous mixture after ignition in a richer mixture according to the magnitude of the stratification. Three magnitudes of stratification are investigated. The local flame burning velocity is determined by an original PIV algorithm developed previously. The local equivalence ratio in the fresh gases is measured from anisole PLIF. From the simultaneous PIV-PLIF measurements, the flame burning velocities conditioned on the local stretch rate and equivalence ratio in fresh gases are measured. The flame propagating through the homogeneous lean mixture has properties depending on the ignition conditions in the stratified layer. The flame propagating in the lean mixture is back-supported longer for ignition under the richer condition. The change of stretch sensitivity and burning velocity of the flame in the lean mixture is measured over time for the three magnitudes of mixture stratification investigated. The ignition in richer mixtures compensates for the nonequidiffusion effect of lean propane flame and sustains its robustness to stretch. The flame propagation in the lean homogeneous mixture is enhanced by ignition in a richer stratified layer, as much by their robustness to stretch as by an increase in the flame speed or the burning velocity. The decay time of this influence of the stratification, called memory effect, is determined. © 2013 The Combustion Institute.

Low voltage ride-through of DFIG and brushless DFIG: Similarities and differences

The brushless doubly fed induction generator (BDFIG) has been proposed as a viable alternative in wind turbines to the commonly used doubly fed induction generator (DFIG). The BDFIG retains the benefits of the DFIG, i.e. variable speed operation with a partially rated converter, but without the use of brush gear and slip rings, thereby conferring enhanced reliability. As low voltage ride-through (LVRT) performance of the DFIG-based wind turbine is well understood, this paper aims to analyze LVRT behavior of the BDFIG-based wind turbine in a similar way. In order to achieve this goal, the equivalence between their two-axis model parameters is investigated. The variation of flux linkages, back-EMFs and currents of both types of generator are elaborated during three phase voltage dips. Moreover, the structural differences between the two generators, which lead to different equivalent parameters and hence different LVRT capabilities, are investigated. The analytical results are verified via time-domain simulations for medium size wind turbine generators as well as experimental results of a voltage dip on a prototype 250 kVA BDFIG. © 2014 Elsevier B.V.

Investigation of flame stretch in turbulent lifted jet flame

DNS data of a laboratory-scale turbulent lifted hydrogen jet flame has been analyzed to show that this flame has mixed mode combustion not only at the flame base but also in downstream locations. The mixed mode combustion is observed in instantaneous structures as in earlier studies and in averaged structure, in which the predominant mode is found to be premixed combustion with varying equivalence ratio. The non-premixed combustion in the averaged structure is observed only in a narrow region at the edge of the jet shear layer. The analyzes of flame stretch show large probability for negative flame stretch leading to negative surface averaged flame stretch. The displacement speed-curvature correlation is observed to be negative contributing to the negative flame stretch and partial premixing resulting from jet entrainment acts to reduce the negative correlation. The contribution of turbulent straining to the flame stretch is observed to be negative when the scalar gradient aligns with the most extensive principal strain rate. The physics behind the negative flame stretch resulting from turbulent straining is discussed and elucidated through a simple analysis of the flame surface density transport equation. © 2014 Copyright Taylor and Francis Group, LLC.

Femtosecond laser induced breakdown spectroscopy of air-methane mixtures

Femtosecond laser pulses are used in order to induce dielectric breakdown in gaseous mixtures, namely in some reactive air-methane mixtures. The light emitted from the laser induced plasma was analyzed while the main emission features are identified and assigned. From the analysis of the emission spectra, a linear relationship was found to hold between the intensity of some spectral features and methane content. Finally, the use of femtosecond laser induced breakdown as a tool for the in situ determination of the composition of gaseous mixtures (e.g., equivalence ratio) is also discussed. © 2013 Elsevier B.V. All rights reserved.

Femtosecond laser induced breakdown for combustion diagnostics

The focused beam of a 100 fs, 800 nm laser is used to induce a spark in some laminar premixed air-methane flames operating with variable fuel content (equivalence ratio). The analysis of the light escaping from the plasma revealed that the Balmer hydrogen lines, H α and H β, and some molecular origin emissions were the most prominent spectral features, while the CN (B 2Σ +-X 2Σ +) band intensity was found to depend linearly with methane content, suggesting that femtosecond laser induced breakdown spectroscopy can be a useful tool for the in-situ determination and local mapping of fuel content in hydrocarbon-air combustible mixtures. © 2012 American Institute of Physics.