基于部分边界阻抗频谱测量数据的双模态电学层析成像

This project studies the image reconstruction algorithms and hardware system in dual modality electrical tomography by using the impedance spectrum, which is measured from partial boundary. Typically, in electrical tomography, the measurement is implemented on the electrodes that are placed equal-distantly along the whole boundary of the sensing region. However, installation of the electrode array is usually permitted along a partial boundary of the sensing region, especially in industrial environment. Information capture ability of the electrode array on a partial boundary will be analyzed. Based on the soft field property of the electrical sensing region, the dual modality electrical tomography methods will be generalized to the case of partial boundary measurements, which means that the electrodes are placed in a part of the sensing region’s boundary. The parameter identification of the impedance network will be carried out with the help of the direct image reconstruction algorithms, which serve as constraints of the identification problem. The material distribution (e.g. permittivity and conductivity distributions) in the sensing region can be recovered in terms of the parameter estimations of the impedance network. Consequently, the software for the dual modality tomography will be developed. Different excitation frequencies will be applied to study the relationship between the number of independent measurements and frequency selection strategy. The hardware system will be developed to extract the global information from the electrodes installed on partial boundary. As a result, the dual modality parameters in the sensing region can be reconstructed by using the developed software and hardware systems. This project is expected to provide new insight to the global information extraction of the soft field, to give new ideas of solving the inverse problem of electrical tomography, and consequently to offer theoretical principles as well as design guidance of instruments for two phase flows.

本项目研究测量信息仅可在部分边界处获得时的双模态电学成像方法及硬件系统。工业应用中常要求电极阵列仅布置于被测区域的部分边界上，但电学传感器中的电极阵列通常需环绕被测区域一周。通过对部分边界上电极阵列的信息获取能力进行分析，结合电学敏感场的“软场”特性，将双模态电学层析成像方法推广至部分边界，即根据布置于部分边界的电极阵列上获得的频谱测量数据，融合图像重建的直接算法及阻抗网络的反演方法，重构被测区域内的物质分布（介电常数及电导率分布），并在此基础上开发双模态成像所需的软件系统。通过研究频率选择与独立测量数目的关系，构建一套双模态成像硬件系统，实现对部分边界的频谱测量数据的获取，并进行被测区域的双模态同步解耦重建，有望提供“软场”特性所包含的全局信息的获取方法，并为电学成像及相应的数理逆问题求解提供新的思路，进而为多相流的在线测量提供新手段。

在工业应用中，电学传感器电极阵列往往无法分布在被测区域完整边界上，但传统的电学成像系统通常要求电极阵列环绕被测区域一周，部分边界的限制对激励测量方式和图像重建方法带来了挑战。本项目针对部分边界限制所导致的测量与重建难题，提出了部分边界双模态模型，实现了部分电极信息获取能力的性能优化，提出了一系列图像重建算法，构建了高速、宽频谱双模态成像软硬件系统，并用于测量高压管道内CO2气液分布和本生灯火焰轴向复电导率的分布。.本项目执行过程中，项目组首先构建部分边界双模态模型，证明了电导率方程的共形变换不变性，完成部分电极信息获取能力的性能优化，为阻抗测量系统设计提供依据。然后，项目组提出基于闭环控制、捷径D-bar、 Radon变换三类图像重建算法，可以实现感兴趣区域的直接重建，有效提高了图像重建质量。针对宽动态范围的多路阻抗测量问题，项目组提出了差分式阻抗测量方法、高精度正弦信号发生方法、基于卡尔曼滤波和信息滤波的单频多频数字递推解调方法、基于压缩感知的宽频谱阻抗测量方法，提高了阻抗测量精度及系统响应速度。.项目组设计研制了多通道单频、多频激励高速数字阻抗测量系统和宽频谱阻抗测量系统；并将搭建的系统成功应用于高压管道内CO2气液运输过程状态监测，实现了CO2气液分布图像重建及含率数据的在线获取。利用所研究的双模态测量及重建方法，将成像系统应用于本生灯火焰双模态监测，分别获得了电导率和介电常数的分布图像和本生灯火焰轴向复电导率的分布规律。.受项目支持，项目组共发表学术论文14篇，其中SCI检索论文11篇。项目组已授权发明专利11项。本项目研究工作为电学成像方法及相应的数理逆问题求解提供了新思路，发明并研制的阻抗测量系统将进一步推进火焰燃烧机理研究和多相流在线测量技术的发展。