Image Reconstruction Algorithm Based On Fusion Of Particle Swarm Optimization And Fuzzy Inference For Electrical Capacitance Tomography

The Electrical Capacitance Tomography(ECT)technique is a process tomographic imaging technology used to visualize the processes of two-phase or multiphase flow in enclosed pipelines.This technique reconstructs the image of the measured medium distribution through the application of relevant algorithms,based on the capacitance data collected by sensors.ECT technology has several advantages,such as non-invasiveness,simplicity,low cost,and ease of installation,thus making it widely applicable in industrial production.As a non-destructive testing method,the performance of ECT system in practical applications is not only related to the detection means and hardware conditions,but also to the accuracy and speed of image reconstruction algorithms.This study focuses on the issues of gas-solid two-phase flow imaging using ECT technology,and its main works and achievements are as follows:Firstly,a detailed analysis was conducted on the basic principles and overall structure of the8-electrode ECT imaging system,along with the introduction of commonly used algorithms in ECT.Secondly,a mathematical model was established for ECT,and the direct and inverse problems of ECT were solved,with the process of ECT image reconstruction deduced.Finally,using the COMSOL finite element software,an experimental simulation environment was constructed to determine the ECT sensor structure and corresponding parameters suitable for gassolid two-phase flow in this study.Secondly,this thesis proposes an image reconstruction algorithm based on the fusion of fuzzy reasoning system and particle swarm optimization(PSO-Fuzzy)to improve the quality of ECT image reconstruction.The basic concept of fuzzy reasoning system is introduced,followed by a specific process for ECT image reconstruction based on PSO-fuzzy.This algorithm utilizes the nonlinear network structure of the fuzzy reasoning system to achieve a nonlinear relationship between capacitance value and dielectric constant.Furthermore,the parameters of PSO-Fuzzy are set,and the optimal parameters of fuzzy membership function are obtained by utilizing the basic characteristics of the PSO algorithm.Experimental results demonstrate that the images reconstructed by the PSO-Fuzzy algorithm are visually very close to the original images,with lower relative error,faster reconstruction speed,and fewer distortions and artifacts compared to traditional ECT image reconstruction algorithms.Thirdly,a novel image reconstruction algorithm(GSAPSO-Fuzzy)was proposed,integrating gravitational search algorithm(GSA)into the original particle swarm optimization(PSO)algorithm.GSA was introduced as a novel optimization algorithm,and incorporated with fuzzy inference system to improve the image reconstruction process.By replacing the single PSO algorithm with PSOGSA algorithm,the learning capability and global search ability were enhanced.The sensitivity updating process does not require solving the forward problem,leading to faster imaging speed.Experimental results indicate that the GSAPSO-Fuzzy algorithm yields reconstructed images with superior stability,effectively reducing reconstruction artifacts and improving reconstruction speed,making it an efficient method for image reconstruction.