Mariangela Usai

A signal-processing tool for non-destructive testing of inaccessible pipes

The design of Non-Destructive-Testing systems for fault detection in long and not accessible pipelines is an actual task in the industrial and civil environment. At this purpose, the diagnosis based on the propagation of guided ultrasonic waves along the pipes offers an attractive solution for the fault identification and classification. The authors studied this problem by means of suitable Artificial Neural Network models. Numerical techniques have been used to simulate the guided wave propagation in the pipes. In particular, the finite element method has been used to model different kinds of pipes and faults, and to obtain several returning echoes containing the faults information. Torsional wave modes have been used as excitation waves. The obtained signals have been processed in order to reduce the data dimensionality, and to extract suitable features. The features selected from the signals can be further processed in order to limit the size of the Neural Network models without loss of information. At this purpose, the principal component analysis has been investigated. Finally, the selected features have been used as input for the neural network models. In this paper, traditional feed-forward, multi-layer perceptron networks have been used to obtain the information on size and location of localized notches.

Performance analysis of locally recurrent neural networks

Many practical applications of neural networks require the identification of non‐linear deterministic systems or chaotic systems. In these cases the use of a network architecture known as locally recurrent neural network (LRNN) is often preferable in place of standard feedforward multi‐layer perceptron (MLP) networks, or of globally recurrent neural network. In this paper locally recurrent networks are used to simulate the behaviour of the Chua’s circuit that can be considered a paradigm for studying chaos. It is shown that such networks are able to identify the underlying link among the state variables of the Chua’s circuit. Moreover, they are able to behave like an autonomous Chua’s double scroll, showing a chaotic behaviour of the state variables obtainable through a suitable circuit elements choice.

A greedy genetic algorithm for continuous variables electromagnetic optimization problems

A greedy genetic algorithm for continuous variables electromagnetic optimization problems is presented. The presented algorithm is characterized by the use of a nonlinear simplex method as a principal optimizer, and of a greedy genetic algorithm to explore the search space, realizing a balance between diversity and a bias toward fitter individuals. The resulting algorithm merges the efficiency typical of calculus-based search with the robustness typical of random methods. A detailed comparison of the performance obtained implementing several strategies is presented, using an electromagnetic design test problem.

Slab line impedances revisited

Accurate solutions for impedances and charge distributions in slab lines and rectangularly shielded lines are obtained by numerical inversion of the Schwarz-Christoffel conformal transformation. Circular inner conductors are considered, putting to the test the relatively simple numerical methods utilized, and results are successfully compared to the best data available from the literature. The method, besides supplying accurate global parameters, such as capacitances and impedances, is also shown to provide good evaluations for local charge densities. Equipotential and field lines can be easily derived, and accurate calculation of local field maps is shown to be possible, even from approximate geometries, when boundary conditions are not completely known. >

Artificial neural networks for non-destructive evaluation with ultrasonic waves in not accessible pipes

The design of non-destructive testing systems for fault detection in long and not accessible pipelines is an actual task in the industrial and civil environment. At this purpose the diagnosis based on the propagation of guided ultrasonic waves along the pipes offers an attractive solution for the fault identification and classification. The authors studied this problem by means of suitable artificial neural network models. Numerical techniques have been used to model different kinds of pipes and faults, and to obtain several returning echoes containing the fault information. These signals have been processed to filter the noise by using wavelets e blind separation methods and passed to a feature extractor system, whose purpose is to reduce the data dimensionality and to compute suitable features. The features selected from the signals have been further processed in order to limit the size of the neural network models without loss of information. At this purpose, the Garsons method and the principal component analysis have been investigated and compared. Finally, the extracted features have been used as input for the neural network models. In this paper, traditional feed-forward, multi layer perceptron networks have been used to classify position, width, and depth of the defects.

Voltage collapse prediction with locally recurrent neural networks

Voltage stability studies aim to evaluate the ability of a power system to keep acceptable value of voltages at all nodes either under normal or contingency conditions. Voltage instability involves generation, transmission, and distribution and includes a wide range of phenomena. When a power system is working close to its stability limit, perturbations can easily lead it to a voltage collapse. Among all the stability indicators available in literature, the one based on the minimum singular value of the Jacobian matrix is very common, but it requires a tedious and time consuming iterative solution of the dynamic load flow equations, especially in real size power systems, and therefore it cannot be used for on-line applications. In this paper a new methodology based on the use of artificial neural networks, which are characterized by fast computation and high ability to generalize, is proposed. The adoption of locally recurrent neural networks has permitted predicting the value of minimum singular value with high accuracy.

Tabu-search procedure for PAPR reduction in PLC channels

In this paper a Tabu Search based procedure for Peak to Average Power Ratio (PAPR) reduction in Power Line Communication (PLC) channels is presented. The Orthogonal Frequency Division Multiplexing has been assumed as communication technique. The Selected Mapping approach is applied in order to investigate the possibility to define with a limited computational cost a phase factors vector that represents a sub-optimal solution. The study has been developed referring to the real case of a PLC system implemented in a mega yacht.

Bit loading Optimization for naval PLC systems

In this work a procedure for the optimization of the bit loading in a PLC system implementing OFDM modulation is presented. The optimization strategy aims to find the best compromise between Signal Power, Bit Rate (BR), Bit Error Rate (BER) and Peak-to-Average Power Ratio (PAPR). The problem is approached from two points of view. First, the conflicting objectives of Signal Power, Bit Rate and Bit Error Rate are optimized off line by means of a Multi-Objective approach. A Tabu Search-based search engine has been adapted to the Multi Objective Optimization in order to find a set of Pareto solutions, among which the designer has to take the final choice on the basis of the custom requirements. Secondly, the Selected Mapping method is used on line to reduce the PAPR, finding the best compromise between maximum Bit Rate and minimum PAPR. A semi analytical procedure, described in the paper, is used to determine the optimal phase factor vector. The suitability of the approach has been evaluated referring to the real case of a PLC system implemented in a mega yacht.

Acoustic NDT on Building Materials Using Features Extraction Techniques

Ultrasonic materials analysis is based on the principle that the propagation of any wave is affected by the medium through which it travels. Thus, changes in measurable parameters associated with the passage of a wave through a material can be correlated with changes in physical properties of the material. In this paper, an ultrasonic technique has been experimented for non destructive testing of building structures. In particular, several parameters associated with acoustic waves propagating through a trachite masonry and a concrete pillar have been analyzed and discussed making reference to a Finite Element model of the structures.

Performance improvement of a hybrid optimization algorithm for electromagnetic devices design

In this paper, an investigation of the behavior of a recently defined hybrid algorithm for continuous variables electromagnetic optimization problems is presented. This algorithm makes use of the nonlinear simplex method as a principal optimizator, and a greedy genetic algorithm to explore the search space. The algorithm is greatly affected by the tuning of some parameters. Referring to a solenoids system optimal design problem, a detailed comparison of performances obtained implementing a metaheuristic tuning is presented.