Optimization Of Detection System And Image Reconstruction Algorithm For Electrical Capacitance Tomography

Electrical Capacitance Tomography(ECT)is a promising process imaging technique.It mainly measures the capacitance data between a group of electrodes,and then visually reconstructs the permittivity distribution in the measurement area with the help of an appropriate image reconstruction algorithm.The ECT technology has the advantages of non-invasive measurement,low equipment cost,high imaging quality,fast imaging speed,radiation-free and simple structure,so it has gradually become an important technical means of process inspection in the industrial field.High-precision imaging quality is an important guarantee for ECT to obtain reliable measurement results,and provides strong scientific support for the study of the complex mechanism inside the imaging target.Therefore,this thesis mainly studies the detection system and image reconstruction algorithm to improve the imaging quality of ECT,and the main work is summarized as follows:(1)The mathematical mechanism of forward problem,sensitive field and inverse problem of ECT is analyzed in detail.The imaging principles,advantages and disadvantages of typical non-iterative and iterative imaging algorithms are introduced,and their numerical performances are quantitatively compared through numerical experiments.(2)A set of ECT system is designed and built to realize the online automatic imaging for the imaging targets.The hardware circuit integrating signal generator,electrode selection module,capacitance measurement circuit,programmable gain amplifier and data acquisition module is designed,and it effectively weakens the influence of noise in the environment on the imaging results.The signal noise ratio of each measurement channel is kept at a high level,indicating that the system has good anti-noise ability and stability.The multi-functional host computer interface of the ECT system is written in C++ language,which realizes the functions of online imaging and controlling the actions of the lower computer.The experimental results show that the ECT imaging system not only has high measurement sensitivity,signal-to-noise ratio and data acquisition accuracy,but also improves the data acquisition speed and on-line imaging ability.It can be used in the experimental research of multiphase flow imaging.(3)A new sensor structure with driving electrodes and a new excitation mode in ECT are proposed,which effectively solves the problem of poor ECT imaging quality caused by the small sensitive field value in the central area of the measured area and uneven distribution of the sensitivity map,and greatly improves the image reconstruction quality.The solution formula of sensitive field of driving electrode sensor structure is derived through electromagnetic field analysis.In addition,the change rule of the optimal driving voltage is studied when the excitation voltage of the excitation electrodes,the length of the driving electrodes and distance between adjacent measuring electrodes change.The numerical simulation and experimental results show that the sensor structure with driving electrodes can distinctively improve the distribution quality of the sensitivity map by selecting the optimal driving voltage,which brings significant improvement in the imaging quality and image spatial resolution.(4)An image reconstruction algorithm based on Split Bregman iterative is proposed,which improves the efficiency of solving the ECT inverse problem,enhances the robustness and stability of the algorithm,and reduces the gap between imaging results and imaging targets.Based on the Tikhonov Regularization method and variation domain sparsity theory,a new objective function is designed to simulate the image reconstruction problem of ECT,in which the L1 norm is deployed as data fidelity,the LP norm and the L1 norm of the variables are used as regularizer.Based on the fast iterative shrinkage thresholding technique and the soft thresholding method as subsolvers,the Split Bregman iterative method is designed as an effective solver for the proposed objective function.Numerical simulation and experimental results show that the proposed image reconstruction algorithm improves the image reconstruction quality.(5)An image reconstruction algorithm based on regularized extreme learning machine is proposed,which effectively alleviates the adverse effects caused by inaccurate capacitance data and low quality of sensitive field,enhances the imaging ability of ECT system when facing complex imaging targets,and improves the final imaging quality and stability.The implementation of the algorithm mainly includes two stages.At the first stage,i.e.,the learning stage,according to a large number of training samples,the regularized extreme learning machine model is solved by Split Bregman iterative algorithm to extract the mapping between the capacitance correlation coefficient and the imaging target.At the second stage,i.e.,the prediction stage,the correlation coefficient of the measured capacitance value of the ECT system is calculated,and it is used as the input of the regularized extreme learning machine model to predict the final imaging.The numerical simulation and experimental results show that the proposed image reconstruction algorithm enhances the stability of the numerical solution and improves the image reconstruction accuracy.