Design and Performance Simulation of Magnetic Induction Tomography System for Image Reconstruction of Solid-liquid Two-phase Flow

Multiphase flow systems are important parts of industrial processes, including the food, pharmaceutical, oil and petrochemical industries and also agriculture. In order to monitor and control these systems, it is necessary to determine its characteristics such as mass flow of materials, mass concentration, volumetric concentration, density, velocity and homogeneity distribution of fluid. Magnetic induction tomography is one of the new methods of monitoring fluid in pipelines. Its main advantage is that MIT sensors have no contact with the fluid flow. In this research, design and performance simulation of applied current magnetic induction tomography (AC-MIT) system was studied. The system included two annular electrodes as a transmitter sensor and 196 coils as a receiver sensor. In order to performance simulation of the system, three positions (R=0, R=0.35, R=0.65) were considered for the target object and the image reconstructed using Landweber iterative algorithm with Tikhonov regularization method. To evaluate the quality of the reconstructed image, size error (SE) and root mean square error (RMSE) were calculated in different regularization coefficients. The results showed that RMSE increases with increasing regularization coefficient and decreases when the target object approaches the cross-sectional wall. SE parameter decreases as the target object approaches the cross-sectional wall. The minimum and maximum values of the calculated SE were related to R=0.65 and middle position, equal to 0.13 % and 5.72%, respectively.