Research On Image Reconstruction Based On Planar ECT

Hypersonic vehicles are of great significance in military,national defense and civil applications.The intense friction between the high-speed flying body and the air will produce strong high temperature.If the bonding between the thermal insulation composite and the body surface is defective,there will be great potential safety hazards.Electrical capacitance tomography(ECT)technology can realize the visual detection of composite bonding defects by image reconstruction,but there are many artifacts in the reconstructed image and the defects are not clear enough,which is prone to misjudgment and omission.Aiming at this problem,this paper studies image reconstruction based on planar ECT.The main work of this paper is as follows:(1)The simulation model of planar ECT sensor is established by using COMSOL finite element simulation software.The structural composition of planar ECT sensor is analyzed,and the performance index of the sensor is determined.The effects of structural parameters such as electrode plate side length,electrode plate spacing,bottom shielding layer thickness,edge shielding layer to electrode plate distance and inter electrode shielding layer width on the performance of the sensor are studied by single factor rotation method.By analyzing the influence of sensor structure parameters on signal strength,measurement sensitivity,sensitivity uniformity and measurement dynamic range,the sensor structure parameters used in this paper are determined.(2)Image reconstruction based on sensitivity matrix is studied.Firstly,the selection of different layer sensitivity matrices required for two-dimensional reconstruction image is analyzed,and the best layer sensitivity matrix is determined;then,an improved method of sensitivity matrix is proposed.Through the threshold processing and absolute value processing of the negative sensitivity region,the ill conditioned and non-uniformity of the sensitivity matrix are improved,and the modeling error of ECT positive problem is reduced.The Tikhonov regularization algorithm and Landweber iterative algorithm are used to simulate and reconstruct a variety of defect distributions combined with the sensitivity matrix before and after the improvement.The results show that the proposed sensitivity matrix improvement method can significantly reduce the artifacts in the reconstructed image and improve the clarity of the reconstructed defects.(3)An Landweber iterative algorithm with model error correction is proposed.The model error predicted by limit learning machine is used to correct the capacitance data;polynomial acceleration strategy is used to improve the convergence of the iterative algorithm;using the automatic modeling method based on COMSOL WITH MATLAB,the sample data set required by the limit learning machine is obtained.The simulation results show that the proposed iterative reconstruction algorithm has better results than the traditional reconstruction algorithm.(4)The planar ECT sensor is made,and the host computer software is compiled.Combined with the capacitance data collector in the laboratory,a complete planar ECT experimental system is built.The system is used to verify the proposed method.