CURRENT DISTRIBUTION AND FINITE-ELEMENT MESH SELECTION FOR ELECTRICAL-IMPEDANCE TOMOGRAPHY

The method of Electrical Impedance Tomography was originally developed for clinical applications, and it also shows great potential for proc ess diagnostics. EIT enables the internal impedance distribution withi n a region to be reconstructed on the basis of electrical measurements taken on the region's periphery. It is limited to systems that consis t of a conducting fluid and less conducting solids, and is therefore w ell suited to most water based slurries that are encountered the miner al industries. Many existing systems rely on fast but often crude algo rithms. This study uses an iterative finite element approach, the effe ctiveness of which has been established in several studies, but, none have explored the combined choice of current projection and mesh geome try on the accuracy and speed of the reconstructions. A scheme is intr oduced in this paper to select the finite element mesh to improve the conditioning and the speed of the reconstruction. The effect of noise was also studied and the practice of using numerous active electrodes simultaneously was shown to have a superior noise handling capability over the use of two electrodes at a time, as used by other authors. Th e criteria for the selection of mesh/current projection also provide a n efficient framework for adaptive mesh generation.