Tomografia por resistência elétrica na formação de mapas de condutâncias

Abstract

Electrical Resistance Tomography is a technique aimed at estimating the inner conductivity distribution of a phantom from boundary measurements. In medical applications, the goal is to detect any medical condition. In industrial applications, the objective is to detect any anomalous condition. A notable application is the determination of gas temperature distribution. This work proposes an alternative approach to determine all conductances in a resistor network or grid where only boundary elements are accessible for measurements. Differently from most existing approaches, where the medium is considered continuum, this work proposes to discretize the medium in resistive elements. The basic principle is to inject a known electrical current in certain points and estimate all conductances in the network from only boundary voltage measurements. The proof-of-concept sensor is composed by inter-welded resistors. For modeling purposes, simulations were performed in Proteus. To validate the development, a current source and a LabView software were developed. The complete setup includes a reconstruction algorithm developed in Matlab. Since the relationship between conductance values and voltage measurements is nonlinear, the inverse problem requires an iterative linearization step, where a sensitivity matrix is calculated at each iteration. The main applications of the developed technique include any distributed sensor based on resistive transducers such as temperature, pressure, etc. The advantages of the proposed technique are the ability of design robust sensors as well as the simplicity of measurements.