Compression techniques applied to multiple speech recognition systems

Speech recognition systems typically contain many Gaussian distributions, and hence a large number of parameters. This makes them both slow to decode speech, and large to store. Techniques have been proposed to decrease the number of parameters. One approach is to share parameters between multiple Gaussians, thus reducing the total number of parameters and allowing for shared likelihood calculation. Gaussian tying and subspace clustering are two related techniques which take this approach to system compression. These techniques can decrease the number of parameters with no noticeable drop in performance for single systems. However, multiple acoustic models are often used in real speech recognition systems. This paper considers the application of Gaussian tying and subspace compression to multiple systems. Results show that two speech recognition systems can be modelled using the same number of Gaussians as just one system, with little effect on individual system performance. Copyright © 2009 ISCA.

Initial phase modelling in numerical explosion applied to process safety

The utilisation of computational fluid dynamics (CFD) in process safety has increased significantly in recent years. The modelling of accidental explosion via CFD has in many cases replaced the classical Multi Energy and Brake Strehlow methods. The benefits obtained with CFD modelling can be diminished if proper modelling of the initial phase of explosion is neglected. In the early stages of an explosion, the flame propagates in a quasi-laminar regime. Proper modelling of the initial laminar phase is a key aspect in order to predict the peak pressure and the time to peak pressure. The present work suggests a modelling approach for the initial laminar phase in explosion scenarios. Findings are compared with experimental data for two classical explosion test cases which resemble the common features in chemical process areas (confinement and congestion). A detailed analysis of the threshold for the transition from laminar to turbulent regime is also carried out. The modelling is implemented in a fully 3D Navier-Stokes compressible formulation. Combustion is treated using a laminar flamelet approach based on the Bray, Moss and Libby (BML) formulation. A novel modified porosity approach developed for the unstructured solver is also considered. Results agree satisfactorily with experiments and the modelling is found to be robust. © 2013 The Institution of Chemical Engineers.

Comparison among MCNP-based depletion codes applied to burnup calculations of pebble-bed HTR lattices

The double-heterogeneity characterising pebble-bed high temperature reactors (HTRs) makes Monte Carlo based calculation tools the most suitable for detailed core analyses. These codes can be successfully used to predict the isotopic evolution during irradiation of the fuel of this kind of cores. At the moment, there are many computational systems based on MCNP that are available for performing depletion calculation. All these systems use MCNP to supply problem dependent fluxes and/or microscopic cross sections to the depletion module. This latter then calculates the isotopic evolution of the fuel resolving Bateman's equations. In this paper, a comparative analysis of three different MCNP-based depletion codes is performed: Montburns2.0, MCNPX2.6.0 and BGCore. Monteburns code can be considered as the reference code for HTR calculations, since it has been already verified during HTR-N and HTR-N1 EU project. All calculations have been performed on a reference model representing an infinite lattice of thorium-plutonium fuelled pebbles. The evolution of k-inf as a function of burnup has been compared, as well as the inventory of the important actinides. The k-inf comparison among the codes shows a good agreement during the entire burnup history with the maximum difference lower than 1%. The actinide inventory prediction agrees well. However significant discrepancy in Am and Cm concentrations calculated by MCNPX as compared to those of Monteburns and BGCore has been observed. This is mainly due to different Am-241 (n,γ) branching ratio utilized by the codes. The important advantage of BGCore is its significantly lower execution time required to perform considered depletion calculations. While providing reasonably accurate results BGCore runs depletion problem about two times faster than Monteburns and two to five times faster than MCNPX. © 2009 Elsevier B.V. All rights reserved.

Acoustic analogy in swirling mean flow applied to predict rotor trailing-edge noise

The present study aims at accounting for swirling mean flow effects on rotor trailing-edge noise. Indeed, the mean flow in between the rotor and the stator of the fan or of a compressor stage is highly swirling. The extension of Ffowcs-Williams & Hawkings' acoustic analogy in a medium at rest with moving surfaces and of Goldstein's acoustic analogy in a circular duct with uniform mean flow to a swirling mean flow in an annular duct is introduced. It is first applied to tonal noise. In most cases, the swirl modifies the pressure distribution downstream of the fan. In several configurations, when the swirl is rather close to a solid body swirl, it is often sufficient to apply a simple Doppler effect correction when predicting the duct modes in uniform mean flow in order to predict accurately the noise radiated with swirl. However, in other realistic configurations, the swirling mean-flow effect cannot be addressed using this simple Doppler effect correction. Second, a rotor trailing-edge noise model accounting for both the effects of the annular duct and the swirling mean flow is developed and applied to a realistic fan rotor with different swirling and sheared mean flows (and as a result different associated blade stagger angles). The benchmark cases are built from the Boeing 18-inch Fan Rig Broadband Noise Test. In all cases the swirling mean flow has an effect. In some cases the a simple Doppler effect may address it, but, in other realistic configurations our acoustic analogy with swirl is needed. © 2012 by the authors. Published by the American Institute of Aeronautics and Astronautics, Inc.

A self-adaptive genetic algorithm applied to multi-objective optimization of an airfoil

Genetic algorithms (GAs) have been used to tackle non-linear multi-objective optimization (MOO) problems successfully, but their success is governed by key parameters which have been shown to be sensitive to the nature of the particular problem, incorporating concerns such as the numbers of objectives and variables, and the size and topology of the search space, making it hard to determine the best settings in advance. This work describes a real-encoded multi-objective optimizing GA (MOGA) that uses self-adaptive mutation and crossover, and which is applied to optimization of an airfoil, for minimization of drag and maximization of lift coefficients. The MOGA is integrated with a Free-Form Deformation tool to manage the section geometry, and XFoil which evaluates each airfoil in terms of its aerodynamic efficiency. The performance is compared with those of the heuristic MOO algorithms, the Multi-Objective Tabu Search (MOTS) and NSGA-II, showing that this GA achieves better convergence.