<p> This thesis contributes significantly to the advanced electromagnetic (EM) modeling of the interaction of the microwave field with human tissues. The proposed EM models achieve unprecedented computational efficiency, accuracy, and reliability. </p> <p> The EM modeling is crucial in (a) multiphysics and (b) EM analysis m support of optimization procedures with applications in design optimization and inverse problem solutions. The challenge in such applications stems from the fact that EM modeling requires extensive computational resources. Therefore, the reduction of these computational requirements is necessary in order to handle the complexity of multi physics modeling and microwave imaging. </p> <p> In this thesis, an efficient EM/thermal analysis of the interaction of the radiofield (RF) fields of mobile phones with the human eyes is presented. Another advanced application developed here is the solution of inverse problem in microwave imaging and detection by making use of response sensitivity analysis. </p> <p> Two methods are proposed for the evaluation of the maximum specific absorption rate (SAR) in the human eyes due to RF exposure from handheld devices. They account for the existence of resonance in the eye and are applied to the case of near-field exposure. The first method is semi-analytical. As an input. it requires the measured or simulated open-space near field of the device under test in the absence of the eye. As an output. depending on the mutual position and orientation of the eye and the device, it produces the maximum SAR value in the eye averaged over I and I 0 grams of tissue. The second method is experimental. It requires the fabrication of a simple eye phantom and relies on a measurement with an SAR robot. The proposed methods allow for the fast and reliable SAR evaluation of newly developed handheld devices in an industrial environment. Results concerning the temperature rise in the eyes are also presented. They are based on detailed simulation eye models. </p> <p> A conceptually new detection algorithm is proposed in this thesis for the localization of electrically small scatterers in a known background medium. The algorithm requires the knowledge of the electric field distribution inside the known background medium where no scatterers are present. It is based on a self-adjoint response sensitivity computation which can be performed in real time. Using the Efield distribution in the background medium. it provides three-dimensional maps of the Frechet derivative within the imaged volume. The peaks and dips in these maps identify the locations where the permittivity and conductivity of the measured medium differ from those in the background medium. The background medium can be heterogeneous. The performance of the detection algorithm is studied in terms of the number of transmission/reception points, the dielectric contrast of the scatterer compared to the background medium, and the size of the scatterer. Its resolution is also addressed. The proposed detection algorithm is successfully applied in breast cancer detection. </p> / Thesis / Doctor of Philosophy (PhD)

I detta arbete simulerades pressningsprocessen och värmeflödet hos en kabelsko genom att använda finite element analasis med hjälp av simuleringsprogrammet COMSOL Multiphysics. Syftet med värmeanalysen var för att undersöka om kabelskon får värmefläckar under drift som kan orsaka utmattning av metall. Först simulerades pressningsprocessen. Resultatet från pressnings-simuleringen användes sedan som underlag för att kunna genomföra värmeflödesanalysen av en pressad kabelsko i drift. Vid värmeanalysen hade kabeln en fast temperatur och kabelskons temperatur var beroende av tiden. Resultatet från simuleringen visade att det uppstår värmefläckar under drift men med väldigt låga temperaturskillnader. Däremot finns det osäkerheter i den valda metoden då många förenklingar gjordes i simuleringen. En simulering av värmeöverföring, där kabeln hade en fast temperatur, gjordes i ställetför att simulera en ström som går igenom kabeln och kabelskon. Val av detalj av meshen var enbart tillräckligt hög för att programmet skulle kunna genomföra beräkningen. En mer detaljerad mesh ger ett godare resultat. Modellen av kabelskon förenklades även till ett rör för att bespara beräkningstiden av programmet. För att möjligen öka noggrannheten på resultatet behövs en simulering med mer detaljerad mesh, med ström genom kabeln i stället för en fast temperatur och med en modell av den faktiska kabelskon i stället för en förenkling. En mer detaljerad mesh innebär mer detaljerad deformation i pressningsprocessen. Detta kan innebära fler och bättre kontaktytor mellan kabel och kabelsko och därmed en noggrannare resultat av värmeanalysen. Att simulera med ström och en icke-förenklad modell av en kabelsko i stället för fast temperatur och en förenklad modell innebär en mer verklighetsbaserad simulering där eventuella effekter av ojämn strömtäthet kan fångas. Vidare rekommenderas att en validering av simuleringen genomförs med hjälp av ett verklig test med exempelvis en värmekamera. / The crimping process and the heat flow on a cable lug were simulated in this project using finite element analysis and the simulation software COMSOL Multiphysics. The purpose of the heat analysis is to determine if there are any hot spots, that are prone to cause metal fatigue, on a cable lug while in use.Firstly, the crimping process were simulated. The result from the crimping process were then used to simulate the heat transfer while the cable lug is in use. The cable in the heat analysis had a fixed temperature and the temperature of the cable lug was time dependent. The simulation showed hot spots on the cable lug while in use although the temperature difference was modest. There were on the other hand uncertainties with the choice of method because simplifications were made in the simulation. A heat transfer where the cable hade a set value of temperature wassimulated instead of simulating a current that goes through the cable and the cable lug. The choice of mesh detail was just high enough for the software to be able to do calculations. A higher mesh quality would yield a better result. The geometry of the cable lug was simplified as a tube to save computation time. A new simulation with higher mesh quality, with current flowing through the cable and the cable lug instead of a set temperature and with the actual model of the cable lug is needed to get a more accurate result. A more detailed mesh means better approximations in crimping simulation which in turn gives better results in heat analysis due to better contact surface between the cable and the cable lug. Having current flow through the cable and the actual model of the cable lug instead of the simplifications that were made in the simulation is closer to reality and which can give a more accurate result. A validation of the simulation is recommended by using for example a thermal camera.

This study presents peridynamic field equations for mechanical deformation, thermal diffusion, moisture diffusion, electric potential distribution, porous flow and atomic diffusion in either an uncoupled or a coupled manner. It is a nonlocal theory with an internal length parameter. Therefore, it can capture physical phenomenon for the problems which include non-local effects and are not suitable for classical theories. Moreover, governing equations of peridynamics are based on integro-differential equations which permits the determination of the field variable in spite of discontinuities. Inherent with the nonlocal formulations, the imposition of the boundary conditions requires volume constraints. This study also describes the implementation of the essential and natural boundary conditions, and demonstrates the accuracy of their implementation. Solutions coupled field problems concerning plastic deformations, thermomechanics, hygrothermomechanics, hydraulic fracturing, thermal cracking of fuel pellet and electromigration are constructed. Their comparisons with the finite element predictions establish the validity of the PD field equations for coupled field analysis.

Corona discharge is one of the crucial problems related with high-voltage equipment. This paper focuses on the physical and numerical modelling of corona discharge in an Electrostatic precipitator (ESP). The model is based on Maxwells equations and the Finite element method (FEM) and is implemented with the COMSOL Multiphysics software.The simulation allows studying the electric charge distribution and the behaviour of the electric field inside the ESP. The work is focused primarily on time-dependent studies of the corona discharge.

Nos travaux portent sur l’étude du dimensionnement d’une chaîne d’actionnement au moyen d’une optimisation multi-physique. L’approche est de type « système », puisqu’il s’agit d’étudier l’association de filtres électroniques, d’un convertisseur de puissance DC-DC, d’un moteur à courant continu et d’une commande, en prenant en compte la transformation électromécanique de l’énergie ainsi que les comportements thermiques du système. L’intégration de la problématique de compatibilité électromagnétique (CEM) conduite est aussi abordée. Le travail consiste tout d’abord à modéliser chacune des physiques à prendre en compte pour chacun des sous-systèmes, puis de coupler les modèles obtenus afin de pouvoir mener des optimisations. L’algorithme d’optimisation utilisé est de type évolutionnaire. La démarche est appliquée sur le cas industriel d’une vanne EGR (Exhaust Gas Recirculation). Dans ce cadre, l’objectif est de minimiser l’encombrement du système tout en s’assurant que ses performances sont cohérentes avec le cahier des charges ainsi que les contraintes thermiques et CEM. Au-delà de l’application, la démarche permet de s’interroger sur les différents niveaux de modélisation et les moyens de couplages adéquats des modèles, dans le cadre d’une optimisation coûteuse en temps de calcul. / This thesis focuses on the design of a power conversion system using a multiphysic optimization. A system approach is applied, since the electronic filters, a DC-to-DC converter, a DC motor, and the system control are considered simultaneously. The electromechanical energy transformation is of prime importance, although the thermal behavior of the entire system is also significant. Conducted electromagnetic compatibility (EMC) issues are discussed and integrated in our approach. The first step of the methodology is the establishment of the various physical models for each sub-system. Then, these models must be coupled so that an optimization algorithm can be applied. In this work, an evolutionary algorithm is used. The approach is tested on an industrial study-case – an EGR (Exhaust Gas Recirculation) valve. In this framework, the objective is to minimize the geometrical size of the system, while ensuring that the performance imposed by the specifications is achieved and that we conform to the thermal and EMC constraints.Beyond this application, the underlying issues of the modeling level and the means of coupling the various models are addressed in the context of a time-consuming optimization.

Lightning damage resistance of unstitched carbon/epoxy laminates and a Pultruded Rod Stitched Efficient Unitized Structure (PRSEUS) panel were characterized by laboratory-scale lightning strike tests and multiphysics-based lightning strike finite element (FE) models. This dissertation combines three related research topics: (1) a three-dimensional (3D) heat transfer problem, (2) lightning damage resistance assessments of carbon/epoxy laminates, and (3) lightning damage resistance of PRSEUS panel. The first project deals with a 3D analytical heat transfer problem as a solid foundation for understanding the steady-state temperature distribution in an anisotropic composite heat spreader. The second project characterizes lightning damage to unprotected carbon/epoxy laminates and laminates with either copper mesh (CM) or pitch carbon fiber paper (PCFP) protection layers subjected to standard impulse current waveforms, consistent with actual lightning waveforms, with 50, 125, and 200 kA nominal peak currents. Multiphysics-based FE models were developed to predict matrix thermal decomposition (a primary form of lightning damage) in unprotected, CM-protected, and PCFP-protected carbon/epoxy laminates. The predicted matrix decomposition domains in the damaged laminates showed good agreement with experimental results available in the literature. Both the CM and the PCFP lightning protection layers successfully mitigated lightning damage development in the underlying composites. The third project includes lightning damage characterization of a PRSEUS panel. Laboratory-scale lightning strike tests with nominal 50, 125, and 200 kA peak currents were performed at the mid-bay, stringer, frame, and frame/stringer intersection locations of the PRSEUS panel. The elliptical regions of intense local damage were elongated along the outermost lamina’s carbon fiber direction, consistent with observations from the unstitched carbon/epoxy laminates. However, the damaged PRSEUS panel exhibited unique damage features due to use of warp-knitted fabrics and through-thickness VectranTM stitches. The polyester threads used to weave the warp-knitted laminates locally confined small-scale fiber damage. This resulted in somewhat periodic and scattered small tufts of carbon fibers near the lightning attachments. Through-thickness VectranTM stitches also confined intense local damage development at the stringer and frame locations. The polyester warp-knit fabric skins and through-thickness VectranTMstitches have a significant beneficial effect on lightning damage development on a PRSEUS panel.

We investigate (i) pull-in instabilities in a microelectromechanical (MEM) beam due to the Coulomb force and in MEM membranes due to the Coulomb and the Casimir forces, and (ii) thermomechanical instability in a heat-conducting thermoelastoviscoplastic solid due to thermal softening overcoming hardening caused by strain- and strain-rate effects. Each of these nonlinear multiphysics problems is analyzed by the meshless local Petrov-Galerkin (MLPG) method. The moving least squares (MLS) approximation is used to generate basis functions for the trial solution, and the basis for test functions is taken to be either the weight functions used in the MLS approximation, or the same as for the trial solution. In this case the method becomes Bubnov-Galerkin. Essential (displacement, temperature, electric potential) boundary conditions are enforced by the method of Lagrange multipliers. For the electromechanical problem, the pull-in voltage and the corresponding deflection are extracted by combining the MLPG method with either the displacement iteration pull-in extraction algorithm or the pseudoarclength continuation method. For the thermomechanical problem, the localization of deformation into narrow regions of intense plastic deformation is delineated. For every problem studied, computed results are found to compare well with those obtained either analytically or by the finite element (FE) method. For the same accuracy, the MLPG method generally requires fewer nodes but more CPU time than the FE method; thus additional computational cost is compensated somewhat by the increased efficiency of the MLPG method. / Ph. D.

Os biopolímeros comestíveis, tais como amido e alginato, podem ser utilizados na formulação de géis com elevada capacidade de retenção de água. A secagem convectiva pode ser utilizada para a remoção da umidade em géis. Os parâmetros da secagem (temperatura, velocidade do ar e umidade de equilíbrio) necessitam ser monitorados para a produção de um sólido seco e sem degradação de compostos ativos. A modelagem matemática da secagem pelo método numérico de elementos finitos em COMSOL Multiphysics tem sido utilizada para simular a secagem, usando um número reduzido de ensaios experimentais. Este trabalho teve como objetivo o desenvolvimento de modelos analíticos e numéricos para predizer a umidade e tamanho de fatias de géis de amido de milho e alginato de cálcio durante a secagem convectiva. O acoplamento entre a transferência de massa e encolhimento das fatias durante a secagem foi simulado e a difusividade mássica efetiva foi obtida pelo ajuste não linear aos dados experimentais. Três modelos foram utilizados como estudos de caso: A secagem de géis contendo 60% de água e 5.4% de amido gelatinizados (GC90), foram descritas pela solução analítica da segunda lei de Fick (R2 = 0.997-0.998); A secagem de géis contendo 60% de água e 5.4% de amido nativo (RC90), foram mais bem explicadas pelo modelo analítico com inclusão do termo de encolhimento (R2 = 0.992); O modelo numérico desenvolvido em COMSOL Multiphysics descreveu adequadamente a secagem de géis formulados com 86% de água e 34% amido gelatinizado e não-gelatinizado (GC50 e RC90), dando um R2 de 0.983-0.992. O encolhimento foi estimado a partir do fluxo molar de água, enquanto a deformação da geometria foi simulada pelo método arbitrário Lagrangian-Eulerian (ALE). A inclusão do termo de encolhimento modificou o perfil de taxa de secagem e o período de pseudo-taxa constante foi observado. O modelo desenvolvido neste trabalho pode ser aplicado em estudos de secagem de géis, alimentos e outros materiais que apresentam elevada razão de encolhimento. / Biopolymers, such as starch and alginate, can be used in the formulation of gels with high water retention. The convective drying can be applied to gel moisture removing. Drying parameters (temperature, air velocity and equilibrium moisture) should be monitored in view of producing a dry solid without degradation of active compounds. The mathematical modeling by the finite element method in COMSOL Multiphysics has been used to simulate drying profiles, with reduced experimental runs. This work aimed at developing of analytical and numerical models to predict the moisture and size of slices of gels containing cornstarch and calcium alginate, during convective drying. The coupling between mass transfer and shrinkage of slices during drying was simulated and the effective mass diffusivity was obtained by non-linear adjustment to the experimental data. Three models have been used as case studies obtained the effective mass diffusivity. Drying of gels containing 60% water and 5.4% gelatinized cornstarch (GC90 samples) as well fitted by the analytical solution of Fick\'s second law (R2 = 0.997-0.998). Drying of gels containing 60% water and 5.4% native starch (RC90 samples) as explained by Fick\'s analytical model while inclusion of the shrinkage term (R2 = 0.992). The numerical model developed in COMSOL Multiphysics adequately described the drying of gels formulated with 86% water and 34% of starch, gelatinized or non-gelatinized, (GC50 and RC90 samples), giving a R2 of 0.983-0.992. The shrinkage was estimated by the molar flux of water, while the geometry shrinkage was simulated by the Arbitrary Lagrangian-Eulerian (ALE) method. The inclusion of the shrinkage modified the drying rate profiles and a pseudo-constant rate period was observed. The model developed in this work can be applied to drying studies of gels, food and other materials that have a high shrinkage ratio.

Os biopolímeros comestíveis, tais como amido e alginato, podem ser utilizados na formulação de géis com elevada capacidade de retenção de água. A secagem convectiva pode ser utilizada para a remoção da umidade em géis. Os parâmetros da secagem (temperatura, velocidade do ar e umidade de equilíbrio) necessitam ser monitorados para a produção de um sólido seco e sem degradação de compostos ativos. A modelagem matemática da secagem pelo método numérico de elementos finitos em COMSOL Multiphysics tem sido utilizada para simular a secagem, usando um número reduzido de ensaios experimentais. Este trabalho teve como objetivo o desenvolvimento de modelos analíticos e numéricos para predizer a umidade e tamanho de fatias de géis de amido de milho e alginato de cálcio durante a secagem convectiva. O acoplamento entre a transferência de massa e encolhimento das fatias durante a secagem foi simulado e a difusividade mássica efetiva foi obtida pelo ajuste não linear aos dados experimentais. Três modelos foram utilizados como estudos de caso: A secagem de géis contendo 60% de água e 5.4% de amido gelatinizados (GC90), foram descritas pela solução analítica da segunda lei de Fick (R2 = 0.997-0.998); A secagem de géis contendo 60% de água e 5.4% de amido nativo (RC90), foram mais bem explicadas pelo modelo analítico com inclusão do termo de encolhimento (R2 = 0.992); O modelo numérico desenvolvido em COMSOL Multiphysics descreveu adequadamente a secagem de géis formulados com 86% de água e 34% amido gelatinizado e não-gelatinizado (GC50 e RC90), dando um R2 de 0.983-0.992. O encolhimento foi estimado a partir do fluxo molar de água, enquanto a deformação da geometria foi simulada pelo método arbitrário Lagrangian-Eulerian (ALE). A inclusão do termo de encolhimento modificou o perfil de taxa de secagem e o período de pseudo-taxa constante foi observado. O modelo desenvolvido neste trabalho pode ser aplicado em estudos de secagem de géis, alimentos e outros materiais que apresentam elevada razão de encolhimento. / Biopolymers, such as starch and alginate, can be used in the formulation of gels with high water retention. The convective drying can be applied to gel moisture removing. Drying parameters (temperature, air velocity and equilibrium moisture) should be monitored in view of producing a dry solid without degradation of active compounds. The mathematical modeling by the finite element method in COMSOL Multiphysics has been used to simulate drying profiles, with reduced experimental runs. This work aimed at developing of analytical and numerical models to predict the moisture and size of slices of gels containing cornstarch and calcium alginate, during convective drying. The coupling between mass transfer and shrinkage of slices during drying was simulated and the effective mass diffusivity was obtained by non-linear adjustment to the experimental data. Three models have been used as case studies obtained the effective mass diffusivity. Drying of gels containing 60% water and 5.4% gelatinized cornstarch (GC90 samples) as well fitted by the analytical solution of Fick\'s second law (R2 = 0.997-0.998). Drying of gels containing 60% water and 5.4% native starch (RC90 samples) as explained by Fick\'s analytical model while inclusion of the shrinkage term (R2 = 0.992). The numerical model developed in COMSOL Multiphysics adequately described the drying of gels formulated with 86% water and 34% of starch, gelatinized or non-gelatinized, (GC50 and RC90 samples), giving a R2 of 0.983-0.992. The shrinkage was estimated by the molar flux of water, while the geometry shrinkage was simulated by the Arbitrary Lagrangian-Eulerian (ALE) method. The inclusion of the shrinkage modified the drying rate profiles and a pseudo-constant rate period was observed. The model developed in this work can be applied to drying studies of gels, food and other materials that have a high shrinkage ratio.

This diploma thesis is concerned with a problem of the line heat simulation by different kinds of materials with computer - aided COMSOL Multiphysics. The thesis is composed of three thematic units. In the first part there is a line heat principle shortly described and his basic physical quantities which are necessary for correct defining of the simulation. The second part is concerned with the software, its function, auxiliary modulus and a description of single actions, which are essential for making of the model. In the last part there is the whole process of the multiphysical task creation that enables to simulate line heat in a specific material in reliance on time.