Multiphase Flow Imaging by Capacitance Tomography Using Simulated Annealing Inversion

Electrical capacitance tomography (ECT) is a technique for obtaining crosssectional images of the electrical permittivity distribution inside an electrically non-conducting body. It can be used to map the composition of two-phase mixtures like gas-oil systems and provides a useful tool for multiphase flow visualization and measurement and potential applications in petroleum industry. The ECT sensor collects the measured capacitances which are then inverted with a suitable reconstruction algorithm in order to produce an image of the permittivity distribution. But the linear, iterative or regularizing inversion techniques commonly used introduce unwanted smoothing effects in the reconstructed images, may become unstable or may not converge towards the desired solution. In this work we apply the simulated annealing (SA) method to the reconstruction images from ECT measurements. The forward problem is calculated by using a finite volume space discretization mehod to avoid geometrical singularities (as occurs with a classical finite difference method), and to take advantage from its conservative formulation. We test the SA inversion method using static physical models and simulate the typical distribution patterns of two-component flows. This inversion technique has some advantages over approaches based on damping least-square inversions: they find good solutions starting with poor initial models, implement more easily complex a priori information, and do not introduce smoothing effects in the final permittivities.