一种新型两相流检测ERT图像重建算法

Novel image reconstruction algorithm of electrical resistance tomography for two-phase flow

针对目前电阻层析成像(ERT)应用于两相流领域重建图像算法存在成像精度较低和速度较慢的问题,基于BFGS最优化方法,提出一种新型图像重建算法.为满足两相流在线参数测量的要求,对BFGS算法进行了改进.改进的BFGS算法(MBFGS)选取了不精确的迭代方向,并用单位矩阵近似代替目标函数的Hessian矩阵.数值实验证明,新算法较为稳定和快速,重建图像的质量优于目前常用的灵敏度系数法和线性反投影算法,算法以较小的速度代价重建较高质量的图像.

In order to resolve the image reconstruction problem of low reconstruction quality or reconstruction speed of electrical resistance tomography (ERT) for two-phase flow, a novel image reconstruction algorithm based on the BFGS (Broyden, Fletcher, Goldfarb, Shanno) optimization principle was proposed. BFGS algorithm was modified to enable the acquirement of on-line measurement of two-phase flow. In the modified BFGS (MBFGS) algorithm, the inexact iteration directions were selected and the Hessian matrix of the objection function was replaced by the identity matrix. Experimental results show that MBFGS algorithm is stable and fast. The quality of images reconstructed by MBFGS algorithm is higher than that of images obtained by the commonly used sensitivity coefficient method and the linear back projection method, while its reconstruction speed is only slightly slower.