一种测量多相流流型的ECT图像重建算法  

An Image Reconstruction Algorithm to Testing Flow Patterns of Multiphase Flow for Electrical Capacitance Tomography

作　　者：刘靖[1] 徐文忠[1] 姜凡[2] 雷兢[3] 刘石[3] 

机构地区：[1]河南理工大学土木工程学院,河南焦作454000 [2]东华大学环境学院,上海201620 [3]华北电力大学能动学院,北京102206

出　　处：《工程热物理学报》2011年第4期602-604,共3页Journal of Engineering Thermophysics

基　　金：国家自然科学基金资助项目(No.50736002)

摘　　要：提出了一种新的电容层析成像图像重建算法-IMNSNOF算法。该算法在分析极小范数解的基础之上针对ECT逆问题的特点对其进行改进,并利用奇异值分解定理证明这种改进的数值稳定作用。在此基础上建立新的目标泛函,并在求解该泛函的过程中采用正则技巧对其进行改进确保数值解的稳定性。利用非闭合电极ECT传感器实测了多相流沿流向的流型,两种流型图像重建结果显示,IMNSNOF算法图像重建质量高于LBP算法、投影Landweber迭代法及标准的Tikhonov正则法。A novel image reconstruction algorithm for Electrical Capacitance Tomography（ECT） is presented.According to the ill posed characteristics of ECT inverse problem the minimum norm solution was improved on the basis of analyzing the algorithm,and the stabilization of the numerical solution was proved via singular value decomposition principle.In addition,a new objective functional was established;in the process of solving the functional regularization technique was vised to ensure the stabilization of numerical solution.The flow patterns of multiphase flow were tested using unclosed electrode ECT sensor.The image reconstruction results of two flow patterns indicated the quality of the reconstructed image using the improved algorithm is better than other image reconstruction algorithms such as LBP,projective Landweber iteration algorithm and standard Tikhonov regularization.

关 键 词：多相流流型 电容层析成像 传感器 图像重建算法 

分 类 号：TK123[动力工程及工程热物理—工程热物理]