| Electrical Capacitance Tomography Technology(ECT)is a multiphase flow detection technology developed gradually in the 1980 s.It calculates the dielectric constant of the material inside the pipe by measuring the capacitance between the electrodes on the periphery of the pipe.Electrical capacitance tomography is an emerging industrial technology with great development prospects.It has been widely used in many industrial fields such as petroleum and chemical industry due to its low cost,fast response,non-invasiveness,and simple structure.At the same time,it has been widely used in many industrial fields such as petroleum and chemical industry in recent years.It has become one of the hotspots of research by experts and scholars,but in practical applications,the nonlinear problem of ECT system is still not completely solved,which makes the image reconstruction effect of ECT system difficult to meet the needs of industry.In order to overcome the problems existing in ECT image reconstruction system,the ECT image reconstruction In-depth exploration of the algorithm is the key to solving the problem,and more experts and scholars are still required to study this technology in depth in the future tasks:1.Firstly,it analyzes the research background and significance of the topic selection of this article in detail,and introduces the basic principles and classification of process tomography technology,summarizes the development status of ECT system abroad and at home,and elaborates several traditional classic image reconstruction algorithms.The advantages and disadvantages of its algorithms are analyzed and compared.2.Then introduce the structural composition and basic principles of the ECT system in detail,and based on the mathematical model established by the Maxwell equations of electromagnetics,by using COMSOL finite element software,the ECT system is simulated and calculated,and then the capacitance data and sensitivity between the electrodes of the capacitance sensor are solved.Field,which provides prior information for subsequent ECT image reconstruction.3.Aiming at the nonlinear problem of ECT image reconstruction,this paper uses the adjacency matrix to truly reflect the mutual influence between the pixels of the ECT image,and proposes a method for constructing the adjacency matrix of the ECT image.Firstly,conduct theoretical research on the ECT system.Through the research,it is found that the ECT system’s object field distribution has a nonlinear relationship with the output of the capacitance sensor,and the correctness of the non-linear relationship of the ECT system is verified through simulation experiments.Then the problem of constructing ECT image adjacency matrix is studied.Since its non-linear relationship is mainly reflected in the similarity between pixel gray levels and the distance between pixels,the ECT image adjacency matrix is constructed accordingly for subsequent ECT image reconstruction based on graph neural network prepare.4.Aiming at the problem of low accuracy of reconstructed images using traditional ECT image reconstruction algorithms,a new ECT image reconstruction algorithm based on graph neural network is proposed.By extracting a large number of diverse flow pattern samples as the network training set,the network model is established and reconstructed by landweber algorithm.The image is used as the initial input of the network,and the GNN network model is used to reconstruct the ECT image of different types of flow patterns,and finally the reconstructed image result is output.5.Finally,through simulation experiments,two commonly used image evaluation indicators,image errors and correlation coefficients are introduced for image quality evaluation.Compared with the traditional algorithm,the image reconstruction result of the algorithm in this paper is smaller and the correlation coefficient is larger.It can be seen that the algorithm in this paper can better improve the quality of image reconstruction,and further proves the feasibility and effectiveness of the graph neural network in the ECT image reconstruction system. |
