Research On Electromagnetic Tomography Based On Finite Element Method And Improved Iterative Algorithm

Electromagnetic Tomography (EMT) is a new type of Process Tomography technology which is produced in recent years. It can attain the information of conductivity and permeability in the target space. EMT can avoid the effect which caused by the contact of electrodes and objects because of the non-contact and non-intervention characteristics, and it also has the other characteristics of simple structure, low-cost system and quick response. The application of EMT is concerned possible in the areas including impurity detection in medication, the measure of3-phase current of the oil field on the sea and food supervision.The research on the forward problem and the inverse problem is the key part of the research on EMT. This thesis based on the basic theories of electromagnetics related to EMT uses correlative software in analyzing and simulating the forward problem of EMT and contrastively analyze mesh methods in the forward problem. An improved algorithm based on Newton method is put forward, which intend to ameliorate the morbidity of the non-linear equation and to reduce the effect caused by the non-equality of the medium. Here is the work achieved:(1) The basic theory of EMT is analyzed and calculated through the basic theories of mathematics and electromagnetics. The concept of forward problem and inverse problem is summarized and Finite Element Method (FEM) is confirmed to be used in solving the forward problem.(2) Maxwell2D is used to solve forward problem based on FEM. The effect of magnetic induction caused by the medium of the target area, the properties of the objects, the number of the objects, the position of the objects and the magnitude and frequency of the current in the excitation coils are analyzed thoroughly through the simulation. The distribution of magnetic induction, the sensitivity in the middle of the field and the effect on nearby magnetic induction caused by the condition of multi-objects are studied from the research. Then, aiming at the self-adaptive mesh method which is used commonly in tackling forward problem, even mesh method is put forward to be used in solving forward problem with the intention to reduce the unconsistency between the mesh methods in forward problem and inverse problem.(3) The main algorithms including linear method, iterative method and optimal method are mentioned, especially LBP method and Newton method. Aiming at reducing the morbidity of the non-linear equation and effect caused by the non-equality of the medium, an improved Newton method based on Newton method is put forward. Though the analysis on the effect of image reconstruction caused by different mesh density, the method of24times mesh is used to simulate and contrast LBP, Newton method and improved Newton method. The improvement of orientation and precision produced by the improved Newton method is proved by the research on the objects with different positions and numbers.Based on the theory analyzing and simulation results, FEM and the improved Newton method are proved that they can be applied adequately in EMT.