Font Size: a A A

Design Of A Matlab Based Simulation Platform For Electrical Resistance Tomography Measurement

Posted on:2008-03-09Degree:MasterType:Thesis
Country:ChinaCandidate:L ZhangFull Text:PDF
GTID:2178360218463556Subject:Detection Technology and Automation
Abstract/Summary:PDF Full Text Request
Electrical Resistance Tomography (ERT) is one of Process Tomography (PT) techniques based on the principle of impedance measurement, and has been developed very fast over the last decade. ERT is particularly suitable for the two-phase (or multi-phase) flow applications having a conductive continuous phase, and is able to provide 2D/3D visualization information of the parameters of a multi-phase mixture flowing inside a closed pipe or vessel. ERT offers some excellent cutting edges in measurement such as capable of yielding process images, non-intrusive to the flow media, fast-responding, low-cost, non-radiation, flexible measurement options etc. The work described in this paper addresses the research need of ERT initiated in China University of Petroleum, by designing a Matlab based simulation software package capable of performing some basic calculations such as ERT problems modeling, solving forward problems and providing some reconstruction algorithms. The package includes a forward solver developed by using the Finite Element Method (FEM), and several ERT inverse reconstruction modules such as the Linear Back Projection (LBP) method and the Modified-Newton-Raphson (MNR) method. Moreover. In this paper, the influences of the measuring electrode numbers, the magnitude of regularization factor, the initial values for iteration, and the noise impacts on the inverse problem are also discussed in details. Also in this paper, we proposed a method to use either a cross-correlation coefficient method or a variance method to quantitatively assess the inversed image quality on account of the likeness between the modeled image and the reconstructed image. This work allows us to be able to evaluate the improvements numerically in the modifications made to both forward and inverse modelings. In addition, the work also introduces two artificial intelligence optimization methods, genetic algorithm and particle swarm optimization algorithm, to the designed package to expand the image reconstruction power. It improves significantly the convergence capability and the reconstruction accuracy compared with results obtained from the algorithms without using artificial intelligence method.
Keywords/Search Tags:Electrical resistance tomography, Image reconstruction, Image quality evaluation, Artificial intelligence optimization method
PDF Full Text Request
Related items