Research On Iterative Correction Algorithm For Electrical Capacitance Tomography Based On Quasi-Newton

In the late 1980 s,Electrical Capacitance Tomography(ECT)technology gradually developed into a process tomography(PT)technology,ECT technology has a strong reliability,low cost,simple system structure and so on,so it is a promising application process tomography technology.However,there are typical pathological problems and "soft field" effect in ECT.In order to solve the above two problems,the quasi-Newton algorithm is used to reconstruct the ECT,and the convergence of the algorithm is analyzed by induction method.The feasibility of quasi-Newton algorithm applied to ECT is discussed.The quasi-Newton algorithm satisfies the convergence condition and the imaging precision is high.It can be seen from the experimental numerical results that the reconstructed image of the quasi-Newton algorithm is ideal and the quality of the image is better than Landweber,CG,LBP and SD algorithm,which provides an effective new method for ECT image reconstruction.At the same time,in order to make the reconstructed gray image,it can be close to the real image in the actual production environment,the edge of each medium of the gray image is reconstructed.The main contents of this paper are summarized as follows:(1)Detailing the composition of the electrical capacitance tomography system and the basic working principle of the capacitor layer,the inaccuracy of inverse problem of capacitance tomography is analyzed,for this problem,several commonly used algorithms are described,including LBP,PAA,Landweber and Conjugate Gradient algorithms.(2)In this paper,based on the analysis of quasi-Newton algorithm for ECT system,a DFP and a Broyden Correction image reconstruction algorithm are proposed for the typical pathological problem and "soft field" effect of ECT.The mathematic model of DFP correction and the mathematical model of Broyden family correction are deduced,and the convergence of the two algorithms is analyzed by induction method.On this basis,the feasibility of applying the two algorithms of ECT is discussed.Both the DFP algorithm and the Broyden algorithm satisfy the convergence condition and the error of reconstructing the image is small.Experimental data show that the use of DFP algorithm and Broyden to reconstruct the image,can get very good results,in terms of image quality,superior to Landweber,SD,CG and LBP algorithm;in each laminar flow,the error of the Broyden correction algorithm is slightly lower than the error of the DFP correction algorithm,that is,the image quality of the Broyden correction algorithm is slightly better than the DFP correction algorithm.This study provides a new idea for ECT image reconstruction.(3)The necessity of edge compensation for ECT is proved by solving the asymptotic optimal order according to the inverse problem of ECT.At the same time,the convex compensation method is used to clarify the edge compensation range,which proves the feasibility of ECT edge compensation technique.Taking into account the genetic ant colony algorithm is based on genetic and ant colony algorithm their own advantages and disadvantages of simply superimposing two algorithms,and no real integration.In order to improve the optimization and efficiency of the algorithm,the improved genetic ant colony algorithm(IG-AA)is proposed,and the algorithm is verified and analyzed experimentally.The experimental values show that the error in reconstructing the image is smaller after the edge compensation,and the result of the reconstruction is closer to the original flow pattern.