Analysis And Improvement Of Image Reconstruction Algorithm For Electrical Capacitance Tomography Used In Oil-gas Lubrication

Electrical capacitance tomography(ECT)is regarded as an effective method for parameters detection of oil-gas two-phase flow in oil-gas lubrication pipelines.Two and three dimensional image reconstruction algorithms for ECT used in oil-gas lubrication are studied in this thesis.The main research work and conclusions are as follows,(1)The algorithm expression of the positive problem for ECT is derived based on the basic principle of ECT,its parameter characteristics and ill-posed property are analyzed,the ill-posed degree of ECT image reconstruction is analyzed,employing the method of singular value decomposition,under the conditions of different diameters of oil-gas lubrication pipelines and different film thicknesses.The results show that the image reconstruction of ECT used in oil-gas lubrication is dispersed ill-posed problem;the ill-posed degree of the image reconstruction of ECT enhances with the decrease of film thickness when the diameter of pipeline is constant;the ill-posed degree of image reconstruction of ECT reduces with the decrease of the pipeline diameter when film thickness is constant.(2)The Tikhonov regularization parameters are determined about the ECT system used in oil-gas lubricating employing the L-curve method,the improved Tikhonov regularization algorithm improved by OTSU is proposed for image reconstruction about ECT used in oil-gas lubrication,distribution images of oil-gas two phase flow in the oil-gas lubricating pipeline,reconstructed by this algorithm and the Tikhonov regularization algorithm,compared and analyzed.The results show that the improved image reconstruction algorithm has better edge resolution ability and higher imaging accuracy.(3)An indirect 3D image reconstruction algorithm which is suitable for ECT used in oil and gas lubrication is proposed.First of all,two-dimensional image of the pipeline sections is reconstructed employing the improved Tikhonov regularization algorithm proposed in this thesis,then the pipeline axial film thickness distribution curve is acquired employing the three spline interpolation method,the three-dimensional image of oil and gas distribution in the pipeline is finally reconstructed after turning the plane,formed by the pipe axial film thickness distribution curve and pipe wall in effective pipe length interval,3600°.The simulation results show that the volume error of the oil and gas distribution image reconstructed by this 3D algorithm is little.