International Journal of Applied Electromagnetics and Mechanics

Preface to the Selected papers of the 10th International Workshop on Optimization and Inverse Problems in Electromagnetism (OIPE 2008) special issue.

The paper deals with the design optimization of high temperature superconducting magnetic bearings (SMB). The described algorithm allows a∼multiobjective optimal design with respect to the geometries of the gap oriented excitation arrangement as well as the superconducting bulks.

Magnetic shielding and shunting in industrial devices is crucial for keeping stray losses and hot-spot appearances as low as possible. Nevertheless, the additional costs for magnetic shunts and conductive shields have to be taken into account. Therefore, an optimal trade-off between a remarkable reduction of the stray losses and the design and position of the shielding devices has to be found. A higher order evolution strategy is applied to perform this task. The relevant objectives are merged into a single number by applying a fuzzy membership function to each objective, respectively. These membership functions can be implemented as static ones or in a way, that they adjust themselves dynamically during the optimization process. In general multi-objective problems show a number of local solutions besides the global one and it is worth to investigate as many of the local solutions as possible as well. Therefore, a most desirable behaviour would be if the optimization strategy behaves globally and yields additional information about local minima detected on the way to the global solution. This goal can be achieved by clustering the population prior to the recombination process and by performing the recombination in a cluster sensitive way.

In this paper, the Integrated Optimal Design (IOD) approach for energetic system design is discussed. IOD aims at concurrently optimizing the architecture, the element sizing and the energy management in an energetic system. IOD leads to complex optimization problems (typically mixed variable problems with several constraints and multiple objectives) which can be solved with direct optimization methods. We illustrate the interest of this approach through the design of a hybrid environmentally friendly locomotive moved by four DC motors supplied by a diesel engine generator in association with electric storage elements (batteries and ultracapacitors).

In this paper, traction system design in carried out with a multilevel optimization technique called Target Cascading. This methodology is based on a hierarchical breaking down and allows solving a problem with multiple parallel optimizations. The large-scale system design problem is usually decomposed into several linked optimization subproblems according to project organization structure. Target Cascading gives a new approach to formulate and to solve the complex system design optimization problem in the electrical field. The problem formulation is more suitable to this kind of problem than conventional methods because it is closer to a work organisation of design engineers.

Scale reduction improves the amplitude of diamagnetism. Thus diamagnetic levitation is possible at the microscale with simple micromagnets. The present work aims at comparing the performances of this levitation at the microscale with other well established technologies. A device including a levitating magnet in a diamagnetic box is optimized in order to fulfil the requirements of three different markets in the acceleration sensing field. It is shown that this technology fulfils the requirements for medium range and short time response markets. However, the use of small proof mass limits the resolution of high precision sensors.

This paper deals with electromagnetic modeling dedicated to constrained optimization needs. A tool dedicated to magnetic devices with simple shapes is presented and applied to the design of magnetic MEMS. This tool allows semi-numerical modeling and provides symbolic gradient computation. A multi-level optimization strategy is used to ensure fast and global convergence of magnetic latch micro device.

We present the topology optimization of an assembly consisting of a piezoelectric layer attached to a plate with support. The optimization domain is the piezoelectric layer. Using the SIMP (Solid Isotropic Material with Penalization) method with forced vibrations by harmonic electrical excitation, we achieve a maximization of the dynamic displacement. We show that the considered objective function can be used under certain boundary conditions to optimize the sound radiation. The vibrational patterns resulting from the optimization are analysed in comparison with the modes from an eigenvalue analysis. Multiple-frequency optimization is achieved by adaptive weighted sums. As a second optimization criterion, a flat frequency response is integrated in the optimization process.

For high frequency switch-mode power converters and supplies, planar transformers allow good performances to be reached. They suit very well to embedded power electronics applications, where size and weight matter. This paper presents a modeling, simulation, and sizing tool based on a 1D analytical model. This tool aims at helping designers describe, simulate, and know frequency behavior of planar wound components. As soon as all measurable impedances given as the tool outputs, device behavior becomes predictable. Due to its analytical nature, this tool is very fast and it can be driven by optimization software to reduce transformer losses and, finally, reduce cooling system weight.

Neural networks algorithms are tested to replace an optimization module, for the energy management of a residential grid-connected PV system with storage. Multi-layer perceptrons are generic tools able to map non-linear relationships between inputs and outputs without knowing any information about the system model. This study shows multiple problems of the implementation of these techniques, despite their theoretical promising interests.

Magnetic induction tomography, highly promising imaging technique for medical purposes, applies a magnetic field from an exciter to induce eddy currents in the object of low conductivity, and the magnetic field from these is then detected by a sensor. The induced eddy currents are proportional to the conductivity distribution. This paper deals with the solution of the fundamental inverse problem existing in magnetic induction tomography, which consists in the reconstruction of the eddy currents distribution basing on the magnetic field measurements. The problem has been solved by the Tikhonov regularization method. The choice of the regularization parameter and the accuracy of the proposed method have also been discussed.

The goal of contactless inductive flow tomography (CIFT) is the velocity reconstruction in electrically conducting melts which are used in many metallurgical and crystal growth applications. In this paper, we discuss some recent methodological improvements of this method, in particular the automatic search for an optimum regularisation parameter and the amended treatment of the boundary integrals.

Regional hyperthermia is a cancer therapy aiming at heating tumors using phased array applicators. This article provides an overview over current mathematical challenges of delivering individually optimal treatments. The focus is on therapy planning and identification of technical as well as physiological quantities from magnetic resonance (MR) thermometry measurements.

Purpose: This paper proposes an approach for the coupled EEG-MEG data analysis that takes into account the anisotropy of brain tissues and the different sensitivities of EEG and MEG measurements.

Design/methodology/approach: The study is carried out by using a combination of analytical and numerical methods. In particular, where available, closed solutions are used to evaluate the numerical errors, while FEM models are used where anisotropy and geometrical characteristics make analytical solutions not applicable.

Findings: The impact of tissues anisotropy on coupled EEG-MEG analysis has been assessed, and the effectiveness of an approach based on smart exploitation of different EEG and MEG sensitivity to improve sources reconstruction has been demonstrated numerically.

Research limitations/implications: The analysis has been carried out using a spherical head model, and on simulated data.

Practical implications: The proposed approach allows to improve the localization of equivalent sources for the description of cellular brain activity.

Originality/value: The properties of objective functions used in brain sources reconstruction from EEG and MEG has been investigated taking into account tissues anisotropy in the spherical symmetric case. In addition a formulation of the joint source and tissues tensor conductivity estimation by exploiting simultaneously bioelectric and biomagnetic measurements based on the complementary sensitivity of those measurements has been presented.

Non-invasive localization of premature ventricular beat (PVB) foci is very important for medical treatment of numerous cardiac diseases. In this work a model-based method of reconstruction of ectopic center locations is investigated.

Within the scope of this method patient's multichannel ECG is used as a reference for optimization of an electrophysiological cardiac model. This model is based on the cellular automaton principle and utilizes anatomical data of the patient. Optimized are coordinates of the ectopic focus as well as excitation conduction velocity of ventricular myocardium. Initial values for these parameters are obtained by solving the linearized problem of electrocardiography in terms of activation times. Optimization is performed by minimization of discrepancy between the simulated and reference ECGs.

The aim of the current work is to estimate the quality of ectopic focus localization delivered by this method. Four sample ectopic beats have been simulated, with their foci located in different regions of the left ventricle. 1% Gaussian noise has been introduced into the resulting ECGs. In this way the "measured" ECG signals for this investigation have been obtained. Afterwards the origin of each ectopic beat has been reconstructed using the model-based approach. The method has demonstrated reliable localization of PVB foci, reconstruction errors have not exceeded 6.1 mm.

To meet the increasing fabrication quality standards and the high throughput requirements NDE techniques are reliant on efficient reconstruction tools and visualization tools. In this work we present an inverse algorithm for a modern electromagnetic non-destructive testing approach using a small GMR sensor array to inspect superconducting wires. Four sensitive GMR sensors are positioned around the wire. Small defects of 100 μm in size could be detected in a depth of 200 μm with a signal-to-noise ratio of better than 400. Surface defects could be detected with a SNR of up to 10,000. This remarkably SNR and the small extent of GMR sensors results in a spatial resolution which offers new visualisation techniques for defect localisation, defect characterization and future tomography-like mapping techniques. We developed several inverse algorithms based on either a Finite Element Method or an analytical approach leading to defect localization with an accuracy of a few 10 μm.

An adaptive sampling method based on simplex-mesh refinement is proposed for creating a database that facilitates the inverse problem solution in NDT. The resulting database is used as a generic interpolator of the forward problem. In order to obtain optimal sampling for the purpose of interpolation, physically based mesh generation technique relying on spring analogy is utilized. The method is tested with a four-parameter crack reconstruction problem. It is demonstrated that the optimized database provides high interpolation quality comparing to regular grid databases.

This paper presents the identification of the Jiles-Atherton and Preisach hysteresis models by means of a new heuristic: the Flock-of-Starlings Optimization (FSO). The FSO can be classified as an artificial life algorithm since it takes inspiration from the Particle Swarm Optimization (PSO) and from recent naturalistic observations on real flocks of common little European birds(starlings, Sturnus Vulgaris), performed by M. Ballerini et al. The one-to-one correspondence between the real flight of starlings searching food and the virtual flight of candidate solutions searching global optima is the core of the algorithm. Validations and comparisons with other heuristics have shown that the FSO gives good performances especially in those cases in which the solution space has a huge dimension. In fact, from the analysis of the obtained results by testing the FSO on hysteresis model identification, this heuristic has shown to be very attractive in comparison with other famous heuristics.