Author : M. Orhan Özyalçın
Publisher :
ISBN 13 :
Total Pages : pages
Book Rating : 4.:/5 (85 download)
Book Synopsis Modeling and Simulation of Electromagnetic Problems Via the Transmission Line Matrix Method by : M. Orhan Özyalçın
Download or read book Modeling and Simulation of Electromagnetic Problems Via the Transmission Line Matrix Method written by M. Orhan Özyalçın and published by . This book was released on 2002 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: Today's electromagnetic (EM) problems are very complex. Analytical solutions are available only for some canonical structures and this has lead to an increased interest in numerical electromagnetics. Today, parallel to the increase in computer's capacity and speed, numerical approaches have become rather popular . Improvements in computers have also made it possible to solve EM problems directly in time domain (TD), starting either or from field and network theories. That is why , the transmission line matrix (TLM) and finite difference time domain (FDTD) methods, have enjoyed widespread use in the last decade. TLM is a fast developing technique which was first introduced by P .B. Johns in 1971. At the beginning, three dimensional (3D) problems were simplified and reduced to a generalized 2D nodes (Expanded Node) in the TLM method. Towards the end of the 70's, 3D TLM began to be applied successfully to a wide variety of EM problems. This thesis can be considered as an attempt for increasing the realm of complex EM problems which can be satisfactorily addressed by the TLM method. We will consider two such problems concerning Electromagnetic Compatibility (EMC) and Specific Absorption Rate (SAR) calculations under realistic conditions for which hitherto it has not been possible to generate TLM solutions. Our numerical calculations will show clearly that TLM can satisfactorily be applied to these problem areas. We will also validate our solutions, albeit in a necessarily incomplete manner , by comparing our results with independently generated FDTD solutions of the same problems. It has also to be mentioned that both the TLM and the FDTD algorithms used for this purpose in our work were developed and coded by the author . TLM is based on network theory and involves TD lumped transmission line modeling of the Maxwell's equations in discretized spatial domain. This is an entirely different approach from that used in the FDTD, which relies on the direct discretization of the governing differential equations. TLM involves replacing a continuous system by a network or array of lumped elements. Interrelations and analogies between network equations and Maxwell ' s equations form the basis of this method, and as such it can be considered as being more physical than strictly mathematical discretization approach. Lumped parameters of the transmission line, such as, inductance and capacitance correspond to the electrical parameters relating to the permeability and permittivity distributions in the corresponding EM problem, respectively .Currents and voltages on the other hand, correspond to the magnetic and electrical field components in the system. There are many different TLM versions in the literature. Here, the most powerful of these approaches, namely the 3' symmetrical condensed node (SCN) TLM version is used. The main advantage of this node structure over the others and over the FDTD method is the symmetry it provides and the fact that the calculation of all the 6 field components is accomplished at the same time step. Each SCN- TLM node is represented by a scattering matrix, S, with which the reflected voltage pulses are related to the incident voltage pulses during the simulation time. On the other hand, there are two main drawbacks of the TLM method, which are the requirement of (İ) large computer memories, and (İİ) high simulation times. The organization of this work is as follows: .Chapter 2 is devoted to a fairly complete and detailed treatment of TLM method in 2D and 3D. In this chapter we critically investigate the various TLM algorithms based on different node structures and assess relative merits in relation to their respective computational requirements and the accuracies (numerical dispersion effects) they provide. .In Chapter 3, two canonical problems, (i) the Green's function representation in a PEC resonator and, (ii) radiation from an aperture, are considered for the purposes of validation (calibration) of our codes. The calibration is done via comparisons both in TD and frequency domain (FD). Analytical representations of these two canonical problems are derived in the FD, therefore comparisons in this domain are straightforward. The TLM results are transformed to the FD via discrete Fourier transformation (DFT). On the other hand, broad band TD comparisons are difficult and one needs to follow the steps given below: o The TLM results are obtained directly by using a broad band pulse as a source and the response is obtained directly in the TD. o FD results are calculated separately at chosen sampling frequencies of the broad band pulse used in TLM simulation via analytical representations. Frequency domain results are weighed with the source spectrum and inverse DFT is applied to obtain the TD analytical results. During this process, one must take the frequency resolution criteria into the consideration to get correct results. .In Chapter 4, after having successfully calibrated our TLM code ( and also FDTD which is used to obtain comparison solutions) we proceed to investigate the complex problems which constitute the main original contribution of this work. These are the Shielding Effectiveness (SE) and the Specific Absorption Rate (SAR) simulations for which, where no TLM solutions are available. SE is an effective parameter in EM compatibility (EMC) problems and is used as a criterion for assessing a structure's susceptibility to EM interference. As a realistic prototype of EMC problems in this thesis we have considered a resonator with an aperture for SE modeling. The second problem we investigated concerns SAR calculations. SAR is the only parameter in bio-EM where device-human tissue interaction is of interest. The determination of SAR is an extremely complex problem and can be addressed either via difficult to perform laboratory measurements or via numerical methods using simulated tissue prototypes. In this thesis we have considered the nowadays rather actual problem of calculating SAR distributions in human head models. Extensive calculations for different parameter regimes are done for both problems and the TLM results are compared against the FDTD results and, in the SE case, also with the results given in the literature as obtained via Method of Moments (MoM) and validated experimentally .In all cases our results were in rather good agreement with the comparison solutions used. .Finally , in Chapter 5 we present some concluding remarks together with suggestions for future work.
