Study Of Dynamic Mesh Image Reconstruction On A Stratified Two-Phase Flow Using Electrical Resistance Tomography

Electrical resistance tomography (ERT) is an advanced measurement technique and has a cutting edge for multiphase flow measurement with following advantages: dynamic flow visualization, non-intrusive measurement, capable of yielding a higher image rate, flexible measuing scheme, low cost, and operation safe etc. It is particulary attractive to applications in which the states of a liquid based multiphase flow dominates the very problems under question, such as those often encountered in many process industries.The work described in this paper,by adopting a parameterized adaptive meshing image reconstruction method to yield the interface between the two layered phases, aims to improve the existing static mesh based ERT measurement accuracy for a stratified two phase flow within a closed pipe, assuming that the phase interface is of a distinctive nature or its fuzziness in the boundary region omittable. Simulation shows that when the two-phase flow maintains a certain degree of electrical conductivity contrast, it is possible to identify the resistivity of each phase, as well as the position of interface through a mesh adaptive process. During the simulation, the noise influence on measurement data is taken into consideration and the iteration step for the mesh adjustment is designed to be variable to enhance the measurement accuracy but still with a reasonably fast speed. Also in the paper describes some preliminary results obtained from using neural network and genetic algorithms performed on a dynamic mesh model.