Segmented excitation for electrical capacitance tomography

Abstract

Electrical Capacitance Tomography (ECT) system offers a non-invasive measurement for flow properties within pipelines. The resolution of the images can be improved by utilizing more electrodes within the sensing area which causes reduction in size, but this causes low resolution at the center of the reconstructed images due to the limitation of soft field sensors to distribute high electric field to the middle of the pipe. This project studies the performance of using segmented excitation of ECT system in order to overcome low resolution of image reconstruction. The segmented excitation method offers minimum modification of the hardware as it only require changes on the switching configuration. Three different ECT systems (8-electrode, 12- electrode and 16-electrode) are studied via simulation modeling using COMSOL Multiphysics to investigate the distribution of electrical potential and electrical field within the pipeline. Simulation shows 16-electrode ECT system yields the highest potential at the center of the pipe 1.74 ?10-9V, as compared to the 8-electrode and 12- electrode for maximum number of excitation, which is half of the total electrodes analysis. For various segmentation excitation with 4-electrode, the adjacent electrode configuration shows 16.6% improvement of electrical potential at the center of the pipe as compared to the opposite excited electrode configuration. The electrical field increases significantly for the segmented excitation as compared to single excitation. The segmented excitation method manages to overcome low resolution problem due to the increased number of excited electrodes. The thesis also describes the process of reconstructing the sensitivity map using COMSOL Multiphysics simulation data. The sensitivity map is reconstructed by extracting raw data and normalization process using standard deviation application. The reconstructed image from the simulation data matches the image reconstructed through the experimental data.