Research On Image Reconstruction Algorithm For Two Phase Flow Pattern Based On Compressed Sensing

Two-phase flow model widely exists in many basic physical processes in thermal power engineering,nuclear energy engineering,petrochemical industry,aerospace,refrigeration,food,metallurgy and other industrial fields.However,due to the complexity and changeability of two-phase flow model,parameter detection is difficult.Real-time and effective detection can reflect two-phase flow.The parameter information of the dynamic system is particularly important.Electrical capacitance tomography(ECT)technology provides an intuitive and effective detection method for flow pattern identification of two-phase flow.It has the characteristics of low cost,simple structure,good safety performance and quick response.It has become the development of flow tomography process parameter detection technology at home and abroadthe mainstream and research hotspot of the exhibition.Capacitance tomography image reconstruction technology is a process of reconstructing flow pattern and other two-phase parameters of pipeline cross-section from measured capacitance values.There are some problems in the reconstruction process,such as non-linearity and morbidity,which always limit the accuracy of image reconstruction.In 2006,Donoho and Candidates put forward the theory of Compressed Sensing(CS).The theory uses the sparse prior data of the original signal or image to reconstruct the signal accurately with a small number of sampling or observation values.This paper attempts to apply compressed sensing theory to image reconstruction technology of electrical capacitance tomography,and tries to design algorithms with high reconstruction accuracy and low computational complexity in order to accurately reconstruct image signals.The main work and innovations of this paper include:(1)In view of the defect that the atomic selection criterion of orthogonal matching algorithm can' t select the optimal atom in each iteration,the atomic selection criterion is optimized.The optimization criterion is theoretically analyzed,and the data simulation is carried out to verify the correctness of the improved idea.(2)If the reconstruction algorithm remains unchanged,the better the performance of the measurement matrix,the smaller the error value of the reconstruction signal.It is of great significance to optimize the performance of the existing measurement matrix.Optimizing the measurement matrix by minimizing the cross-correlation coefficient of the measurement matrix can improve the performance of the image reconstruction algorithm.(3)By studying many signal reconstruction algorithms based on compressed sensing theory,an improved SL0 algorithm is proposed,which is considered from two aspects:improving the continuous function and selecting the step factor.The results show that the improved algorithm can improve the quality of reconstructed image.