摘要
在电阻层析成像的仿真计算中,当两相流的网格结构较简单时,重建图像与原始图像几乎相同,每个单元的电阻率仿真值几乎与设定值相同。当网格结构较复杂时,重建图像却不能达到上述效果,成像算法掉入局部最优点。提出一种局部自适应均值滤波算法,使原来的成像算法跳出局部最优点。并给出几个常见流型的图像重建结果。
In the simulation of electrical resistance tomography,when the partition mesh of a two-phase flow pipeline was simple,the reconstructed image was almost identical to the original image,and the obtained resistivity for each element was also almost identical to the set value.However,when the partitioning mesh structure became more complex,the reconstructed image and the resistivity value generally deviated from the above results owing to the iteration process trapped in a local optimum.The algorithm adopted an adaptively calculated regional mean resistivity filter performed on some regional elements.Such an adaptive filter intends to give a perturbation when the iteration process is idled in a local optimum and hopefully move to a better optimum.The image reconstruction results of several commonly adopted flow patterns were given which show the effectiveness of the algorithm.
出处
《化工自动化及仪表》
CAS
北大核心
2010年第9期57-60,共4页
Control and Instruments in Chemical Industry
基金
中国石油大学"211工程"重点建设项目(211SZ05)
关键词
电阻层析成像
MNR算法
自适应均值滤波
electrical resistance tomography
modified Newton Raphson algorithm
adaptive mean filte
