摘要
电容层析成像的成像数据采用不同的归一化模型,对重建图像的质量有重要影响.针对两相流的典型流型,研究了采用不同的归一化模型时,Tikhonov正则化算法性能与流型及相含率的相关性.仿真采用新的归一化模型与常规的模型处理电容测量值,重建原图像.结果表明,基于单一的归一化模型,不能在各流型分布下都有最好的成像效果.提出了基于流型识别、自动适应流型的归一化方法,新方法提高了Tikhonov正则化算法的性能.
The imaging data being used for electrical capacitance tomography （ECT） need be normalized and the reconstructed image quality greatly relies on the normalization method. In ECT, the relationship between the images reconstructed by Tikhonov regularization algorithm based on the different normalization models and the two-phrase flow patterns and phase concentrations, was investigated. The images reconstructed with the new normalization models were proposed and compared to those with conventional ones. A conclusion is drawn that one normalization model can＇t adapt to different flow patterns. So, a new normalization method adaptive to flow pattern recognition is presented to reconstruct images, thus improving the performance of the Tikohov regularization algorithm.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第7期932-935,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(60374056)
关键词
电容层析成像
归一化模型
自适应归一化
TIKHONOV正则化
流型识别
ECT （electrical capacitance tomography ）
normalization model
adaptivenormalization
Tikhonov regularization algorithm
flow pattern recognition
