Dynamic Phase Boundary Estimation in Two-phase Flows Based on Electrical Impedance Tomography

For the dynamic visualization of the phase boundary in two-phase flows, the electrical impedance tomography (EIT) technique is introduced. In EIT, a set of predetermined electrical currents is injected through the electrodes placed on the boundary of the flow passage and the induced electrical potentials are measured on the electrodes. With the relationship between the injected currents and the induced voltages, the electrical conductivity distribution across the flow domain is estimated through the image reconstruction algorithm where the conductivity distribution corresponds to the phase distribution. In the application of EIT to two-phase flows where there are only two conductivity values, the conductivity distribution estimation problem can be transformed into the boundary estimation problem. This paper considers phase boundary estimation with EIT in annular two-phase flows. As the image reconstruction algorithm, the unscented Kalman filter (UKF) is adopted since from the control theory it is reported that the UKF shows better performance than the extended Kalman filter (EKF) that has been commonly used. For the present problem, the formulation of UKF algorithm involved its incorporation in the adopted image reconstruction algorithm. Also, phantom experiments have been conducted to evaluate the improvement reported by UKF