Image Reconstruction And Upper Computer Software For TMR-EMT System

Gas-liquid-solid three-phase fluidized bed is an important reaction equipment in the fields of petrochemical,biological engineering,and so on.The accurate measurement of the distribution and flow characteristics of the internal phase media is of great significance to the characterization,design and optimization of gas-liquid-solid three-phase fluidized bed.As a new type of electrical tomography technology,Electromagnetic tomography technology（EMT）can reconstruct the distribution of media in the region of interest in real time based on the difference of the conductivity or permeability of each phase in the multiphase flow.In order to measure the distribution of magnetic catalyst in gas-liquid-solid three-phase fluidized bed,the upper computer software of tunneling magneto resistance（TMR）-EMT system is designed in this thesis.Meanwhile,aiming at the ill-conditioned problem ofimage reconstruction,the image reconstruction quality of several regularization methodsis compared,and a reconstruction strategy based on regularization theory and guided image filtering method is proposed.The main worksare summarized as follows:1.The upper computer software of TMR-EMT system is designed.Qt is used to design the interface of PC software,and C++language based on Visual Studio is used to achieve software development.The multi-thread design method,matrix calculation based on CMatrix class and Open GL image rendering technology are applied to upper computer software development,which simplifies the program design and improves the online imaging speed.2.Taking the correlation coefficient as an indicator,the effects of L₁regularization method,L₂regularization method and L_pregularization method with different p on improving the ill-conditioned problem of TMR-EMT inverse problem are studied and compared.It is found that,when p is 1.75,the image reconstruction can consider smoothness and sparsity better simultaneously.3.In order to take into account the advantages of L₁and L₂regularization methods,a new image reconstruction strategy based on guided image filtering is proposed.The images of L₁regularization method and L₂regularization method are used as the guided image and the input image of guided image filtering method,respectively.The parameters of the guided image filtering are determined by the simulation of different flow patterns.By the comparison of simulation and experiment,the anti-noise performance and the image quality of guided image filtering method are better than that of L₁and L₂regularization methods.4.The experiment based on gas-liquid-solid three-phase fluidized bed is carried out by using the developed upper computer software and the proposed reconstruction algorithm.The location and size of the agglomerated catalyst can be identified,and the change of solid holdup at different bed heights can be accurately reflected.