Neural computing models for prediction of permeability coefficient of coarse-grained soils

Correlations are very significant from the earliest days; in some cases, it is essential as it is difficult to measure the amount directly, and in other cases it is desirable to ascertain the results with other tests through correlations. Soft computing techniques are now being used as alternate statistical tool, and new techniques such as artificial neural networks, fuzzy inference systems, genetic algorithms, and their hybrids were employed for developing the predictive models to estimate the needed parameters, in the recent years. Determination of permeability coefficient (k) of soils is very important for the definition of hydraulic conductivity and is difficult, expensive, time-consuming, and involves destructive tests. In this paper, use of some soft computing techniques such as ANNs (MLP, RBF, etc.) and ANFIS (adaptive neuro-fuzzy inference system) for prediction of permeability of coarse-grained soils was described and compared. As a result of this paper, it was obtained that the all constructed soft computing models exhibited high performance for predicting k. In order to predict the permeability coefficient, ANN models having three inputs, one output were applied successfully and exhibited reliable predictions. However, all four different algorithms of ANN have almost the same prediction capability, and accuracy of MLP was relatively higher than RBF models. The ANFIS model for prediction of permeability coefficient revealed the most reliable prediction when compared with the ANN models, and the use of soft computing techniques will provide new approaches and methodologies in prediction of some parameters in soil mechanics.     

Narrow-FSK optical packet labeling scheme for optical ethernet networks

In this work, an intensity modulation/frequency-shift keying (IM/FSK) orthogonal modulation scheme is achieved by direct modulation of a distributed feedback laser source using its adiabatic chirp characteristics. Optical frequencies for "1" and "0" bits are separated only 0.7 GHz, obtaining a narrow-FSK modulation and, accordingly, a low residual intensity modulation. Ethernet frames at 1.25 Gb/s (GbE) are transmitted with a label inserted using coded mark inversion codification at a 155-Mb/s rate. Error rates for the Ethernet payload and for the label have been measured for different payload extinction ratios, showing the viability of this scheme for distances up to 25 km.     

High extinction ratio and high-temperature 2.5-Gb/s floor-free 1.3-/spl mu/m transmission with a directly modulated quantum dot laser

For the first time with a directly modulated InAs-GaAs quantum-dot laser, high extinction ratio (up to 17 dB) and 25/spl deg/C-85/spl deg/C single-mode-fiber data floor-free transmissions are achieved at 2.5 Gb/s. Moreover, an interferometric technique showed a nearly constant Henry factor /spl sim/2 until a bias current six times the threshold current.     

Evolution of fretting fatigue damage and relaxation of residual stress in shot-peened Ti-6Al-4V

The evolution of fretting fatigue damage was investigated in shot-peened Ti-6Al-4V samples, by measuring the changes in the surface residual stress, using the X-ray diffraction technique. The surface residual stress was found to relax as the number of fretting fatigue cycles increased. The relaxation behavior of the residual stress with the increasing number of fretting fatigue cycles was observed to occur in three stages. In the first 20 pct of the fretting fatigue life, a drastic relaxation was observed. In the second part (between 20 and 70 pct), a gradually increasing behavior was observed. During the last 20 to 30 pct of the fretting fatigue life, a dramatic relaxation of the residual stress was found to occur. A complete relaxation of the residual stress occurred in the fracture region. A scanning electron microscope observation of the microstructure of the damaged region was used to examine the mechanisms leading to the relaxation of the residual stress. The development of delaminations at the early stages of the accumulation of the fretting fatigue damage was observed to be the main cause of the initial relaxation. The generation of microcracks from the voids left behind by the delaminations is responsible for the additional relaxation of the residual stress. The coalescence of the microcracks generated from different delaminated regions produced yet more relaxation of residual stress and, ultimately, the final fracture of the specimen.     

On solving large strain hyperelastic problems with the natural element method

In this paper, an extension of the natural element method (NEM) is presented to solve finite deformation problems. Since NEM is a meshless method, its implementation does not require an explicit connectivity definition. Consequently, it is quite adequate to simulate large strain problems with important mesh distortions, reducing the need for remeshing and projection of results (extremely important in three‐dimensional problems). NEM has important advantages over other meshless methods, such as the interpolant character of its shape functions and the ability of exactly reproducing essential boundary conditions along convex boundaries. The α‐NEM extension generalizes this behaviour to non‐convex boundaries. A total Lagrangian formulation has been employed to solve different problems with large strains, considering hyperelastic behaviour. Several examples are presented in two and three dimensions, comparing the results with the ones of the finite element method. NEM performs better showing its important capabilities in this kind of applications. Copyright © 2004 John Wiley & Sons, Ltd.     

Nonlinear blind source separation using kernels

We derive a new method for solving nonlinear blind source separation (BSS) problems by exploiting second-order statistics in a kernel induced feature space. This paper extends a new and efficient closed-form linear algorithm to the nonlinear domain using the kernel trick originally applied in support vector machines (SVMs). This technique could likewise be applied to other linear covariance-based source separation algorithms. Experiments on realistic nonlinear mixtures of speech signals, gas multisensor data, and visual disparity data illustrate the applicability of our approach.     

Innovative methodology to improve the quality of electronic engineering formation through teaching industrial computer engineering

An innovative educational methodology adapted to the requirements of a new era with new societal and industrial challenges for electronic engineers is proposed in this paper. This active methodology, known as the Educational Innovation Project (EIP), is being studied in the Electronic Engineering (EE) degree of the Higher Technical School of Design Engineering at the Polytechnic University of Valencia, Valencia, Spain. The main objective of the EIP methodology is to improve the process of teaching and learning in order to increase student success. To accomplish this objective, the EIP method addresses various issues. From an organizational viewpoint, different structural aspects of the EE degree have been adapted, such as balancing and integrating lectures and laboratory sessions, advancing into interdisciplinary studies coordinated among all the subjects of the course, and strengthening the work in teams to tackle real engineer problems. The industrial computer engineering (ICE) subject is taken as a reference to show how these aspects have been applied. Regarding the faculty, lecturers participate in an open and permanent process of further training; attitudes toward cooperation and exchanges of experience among them are promoted; and research and reflection on new methodologies is encouraged. One of the challenges of the implementation of the EIP project is the development of multidisciplinary projects by team workers. The knowledge acquired from all the subjects is put into practice through the development of a common project to undertake real engineering problems.     

Cu diffusion in α-Fe: determination of solute diffusivities using atomic-scale simulations

The nucleation and growth of radiation-induced Cu-rich clusters have proved an important topic in the study of microstructural and chemical evolution responsible for property changes in pressure vessel steels subjected to neutron irradiation. Small Cu clusters can act as precipitate precursors and contribute to the hardening of the material by impeding dislocation glide. Reliable Cu diffusion coefficients provide critical input to Monte Carlo and rate theory approaches to study the kinetics of diffusion and clustering that lead to the formation of such Cu-rich features. In this paper we report the results of a molecular dynamics study of Cu solute-atom diffusion in a Fe–0.9at%Cu alloy, using high-temperatures and high-vacancy concentrations to hasten the convergence of the calculation. We find a value of 0.61 eV for the migration energy of Cu in the b.c.c Fe matrix, in good agreement with available experimental results. The main diffusion mechanisms under the conditions simulated are identified and discussed, and the effect of vacancy and solute clustering on Cu migration is assessed. In addition, a rigorous computation of the solute enhancement factor, correlation factor and solute diffusivity is performed within the five-frequency theoretical framework and the harmonic approximation. The results obtained are compared with those yielded by the direct MD simulations and the observed differences discussed.     

Languages generated by P systems with active membranes

We propose an alternative approach to generate languages by means of P systems: building up an appropriate representation for a string by means of a corresponding membrane structure and then generating the string by visiting the membrane structure according to a well-specified strategy. To this aim, we consider P systems with active membranes, allowing membrane creation or division or duplication and dissolution, where the output of a computation may be obtained either by visiting the tree associated with the membrane structure, or by following the traces of a specific object, called traveller, or sending out the objects. For each of these approaches, we provide characterizations of recursively enumerable languages based on P systems that use different sets of operations for modifying the membrane structure. Francesco Bernardini: He started his Ph.D. at the University of Sheffield in December 2002 after having previously got a master degree in Computer Science from the University of Pisa in Italy. His research is dedicated to the study of theoretical aspects of membrane computing (P systems) and discrete models of biological systems based on P systems. Marian Gheorghe, Ph.D.: His main research interests are in computational models and their applications to software modelling and testing, formal specifications of agent based systems, software engineering. He was investigating computational power of various generative devices (regular, context-free, fully initial; grammar systems; L-systems and variants). He is currently interested in natural computing (membrane calculus) and biological modelling.