Automatic transcription of continuous speech into syllable-like units for Indian languages

The focus of this paper is to automatically segment and label continuous speech signal into syllable-like units for Indian languages. In this approach, the continuous speech signal is first automatically segmented into syllable-like units using group delay based algorithm. Similar syllable segments are then grouped together using an unsupervised and incremental training (UIT) technique. Isolated style HMM models are generated for each of the clusters during training. During testing, the speech signal is segmented into syllable-like units which are then tested against the HMMs obtained during training. This results in a syllable recognition performance of 42·6% and 39·94% for Tamil and Telugu. A new feature extraction technique that uses features extracted from multiple frame sizes and frame rates during both training and testing is explored for the syllable recognition task. This results in a recognition performance of 48·7% and 45·36%, for Tamil and Telugu respectively. The performance of segmentation followed by labelling is superior to that of a flat start syllable recogniser (27·8%and 28·8%for Tamil and Telugu respectively).     

Adsorption and penetration of hydrogen in W: A first principles study

In BCC crystals, such as Tungsten (W), slippage has been observed on the (1 1 0) and (1 1 2) planes. In this work, hydrogen diffusion paths from three different W surfaces ((1 0 0), (1 1 0) and (1 1 2)) have been calculated using first principles density functional theory. Equilibrium sites for H at the surfaces are identified. The energetics for H penetration from the surfaces to the solute site in the bulk is calculated. It is found that for our low surface coverage of H (3.4 × 10¹⁴ H/cm²), approximately 2 eV is required for an H atom to penetrate any of the W surfaces considered in this study.     

Identification and extraction of brain tumor from MRI using local statistics of Zernike moments

In this article, a novel method is proposed for the detection of brain tumor in magnetic resonance images (MRIs). The features of Zernike moments are used to analyze the MRIs. The image is divided into two parts from the center of the image based on the average value of the pixel located at the center boundary, and new image vectors are formed to extract the tumor. The local statistics values obtained from the low and high order Zernike moments are used to calculate the appropriate threshold value for efficient tumor extraction. The proposed method successfully analyzes the tumor part of the image on testing with different MRIs.     

An improved damage diagnostic technique based on Singular Spectrum Analysis and time series models

An approach to enhance the sensitivity of the damage index using the prediction errors of autoregressive – autoregressive exogenous models by augmenting Singular Spectrum Analysis (SSA) to detect and locate minor damage is presented. Numerical simulation studies are carried out by considering a simply supported beam with single and multiple cracks. An experimental study on a reinforced cement concrete beam has been carried out to validate the technique using different levels of damage. The studies emphasise that the SSA improves the sensitivity of the damage index for detection and also localisation, handling environmental/operational variability and measurement noise.     

A theoretical study of the performance of sub-micron MOSFET devices in the presence of edge potential

In the present communication we have presented a detailed theoretical analysis of the performance of the sub-micron device in the presence of the discontinuity at the Si–SiO₂ interface. It is assumed that due to interface discontinuity a potential develops at the edges (Source/Drain) in addition to the built-in-potential. This potential, called Edge Potential, measures directly the extent of the interface roughness. The effect of this potential is more critical in the case of short channel device where drain and source are in close proximity. Our analysis shows that the discontinuity is dominant at the edges but not in the channel. Drive current as well as saturation transconductance decreases in the presence of edge potential. These results suggest that the performance of the device degrades due to the interface roughness. Effect of interface roughness near the edges can be reduced at high gate voltage but it will result more interface roughness scattering.     

Diffusion of Impurity Excitations in Superfluid and Solid 3He–4He Mixtures

Starting from the system of linearized kinetic equations, the exact expressions for the spin and mass diffusion coefficients and thermal diffusion ratio in quantum mixtures of helium isotopes with any degree of degeneration is obtained. From the general expression the various limiting expressions have been obtained, which are determined by various relations between collision rates of quasiparticles. The results of the calculations are compared with available experimental data. It is shown that in the low-temperature region, where an impuriton gas is one-component, the relaxation of concentration of ³ He in superfluid ³ He– ⁴ He mixtures is determined by the acoustic and dissipative collective modes with an effective diffusion coefficient. From the general relation, the expression for a spin diffusion coefficient in solid ³ He– ⁴ He quantum mixtures is derived. The comparison of the results obtained with the experimental data makes it possible to find a numerical value for the width of the impuriton energy zone.     

Syntheses of allyl phenyl acrylates and their evaluation as reactive diluents in UV-curable coating compositions

para-Substituted 2-allyl phenyl acrylates, and diallyl, dipropenyl and dipropyz bisphenol-A diacrylates have been synthesised in high yield and purity and characterised by spectroscopic methods. UV-curable coating compositions have been formulated using these monomers as reactive diluents. Their film properties were studied.     

The use of aerosol formation, flammability, and explosion information for heat-transfer fluid selection

The devastating consequences of aerosol/mist explosions have been widely documented, and there are currently efforts to understand the mechanisms of formation and explosion of aerosols. Heat-transfer fluids (HTFs) are particularly susceptible to these hazards, because they are utilized under high pressures and below their flash points, making them more prone to leaking as aerosols. In fact, there is a critical need during design stages for a perception of explosion risks associated with the selection of HTFs. This paper discusses a novel scheme to integrate the knowledge of HTF aerosol formation from leaks in process equipment into the selection of HTFs during the design process. Hazards of aerosols formed from leaks are classified qualitatively using process pressure and droplet sizes.     

Fuzzy Reliability Assessment of Distribution System with Wind Farms and Plug-in Electric Vehicles

Abstract

The equipment failures are highly uncertain in nature and simple average failure rate will not reflect this uncertainty. The uncertainty level further increases in reliability evaluation due to the integration of wind farm (WF) because of the intermittent nature of wind speed and random charging patterns of plug-in electric vehicles (PEVs). In this work, the uncertain variables in the distribution system (failure rate, repair time, WF output, PEVs charging and system load factor) are represented as fuzzy numbers to handle the uncertainty. The available uncertain data are used to find the probability distribution function (PDF) of that parameter and is converted into fuzzy membership function using transformation techniques. Failure rate of equipment is converted into failure probability using Monte Carlo simulation (MCS) method. Sampling method is applied to create the PDF of a variable which has average value. Fuzzy severity index (FSI) is proposed to find the importance of an equipment on reliability and is evaluated by measuring the fuzzy distance between the fuzzy reliability indices. The proposed assessment method is validated on modified RBTS bus 2 by comparing with analytical and MCS methods. The proposed method has been tested with integration of WFs and PEVs.