Electrical Resistance Tomographic Measurement Based On A Line Electrode Model And Adaptive Meshing Technology

Electrical Resistance Tomography (ERT) is one of Process Tomography (PT) techniques based on the principle of resistance measurement. Measurements are performed, under different current configuration, between electrodes mounted on an insulated circular pipe wall through which a moving multi-phase flow filled. Reconstruction of resistivity distribution of flow is then performed by means of some inversion algorithms on the data measured. ERT offers some excellent cutting edges in multiphase flow measurement such as capable of yielding process images, non-intrusive to the flow media, fast-responding, low-cost, non-radiation, flexible measurement options etc.Based on the equipotential electrode model developed, a study of comparison of impacts of different electrode models on ERT simulation package (both on forward solver and on image reconstruction) output was carried out in this research and is described in this thesis. In particular performances generated by a"point"electrode model and by a"line"electrode model are compared in following aspects: (1)the influence of electrode models on computational accuracy of an FEM based forward solver; (2)the effect of electrode models on image reconstructions and (3)the time incurred in computation with different electrode models. Simulation shows that the developed equipotential"line"electrode model has some unique advantages in high accuracy modeling of a physically sizes electrode thus is very useful to yield a finer mesh structure which is essential for better image reconstruction over the"point"electrode model.The second work described in this thesis is a work of dynamic meshing for ERT measurement. The dynamic mesh is adaptively obtained according to an analysis performed on the original reconstructed images, through which the inter-phase boundary is captured. Triangular elements of original mesh in the vicinity of captured inter-phased boundary region are divided into four smaller triangles. Hence it forms a new finite element mesh for both forward solver and image reconstruction. The primary objective of the work is to create a dynamic FEM model to allow a better image reconstruction with respect to a dynamic multiphase flow situation. Based on the new mesh, researches were carried out on three issues: (1) the impacts of new mesh on ERT measurement with both the"point"and the"line"electrode models; (2)the impacts of new mesh on image reconstruction; (3) the evaluation of performance of new mesh on the accuracy obtainable and the time required. The simulation results show that the accuracy obtained by the adaptive FEM model is better than that obtained by the non-adaptive FEM model. The time required for the adaptive FEM model for ERT measurement is shorter than that is required for an all space re-partitioned FEM model. Adaptive mesh FEM model improves the accuracy of ERT measurement at a cost of extra time in an acceptable range.