Research On Application Of Electrical Capacitance Tomography In Testing Components And Dynamic Variety Of Frozen Soil

The safty of engineering on frozen soil region is influenced directly by the freezing-thawing processs of soil. In order to avoid the occurrence of catastrophe, a large number of studies on the basic theory and methods to prevent damages of frozen soil have been carried out by scholars worldwide. Up to now, the theories of soil's freezing-thawing process have not been thoroughly developed, and experiments are still the major means in this research. The measurement study using ECT on medium distribution and dynamic change of frozen soil will not only deepens the understanding, but also promotes the applications of ECT in porous and multi-phase medium measurement.A novel ECT sensor with unclosed electrodes was designed and made to satisfy the requirements of frozen soil testing and its characteristics were analyzed. A model for the sensor was established. The electric field was calculated using finite element method (FEM). The capacitance distribution characteristics of the ECT sensors with 10, 12 and 16 unclosed electrodes were simulated and analyzed and the optimal number of electrodes in the frozen soil's measurement is established. The sensitivity maps were computed, and the so-called "soft field" effect was discussed.A method of optimum design for the unclosed-electrode sensor was presented. Uniform design and non-linear partial least-square (NPLS) regression were employed to extract the relationship between the structure of the sensor and the optimization goals. And an objective functional was established. The optimal parameters of the sensor were obtained via solving the objective function using the partial swarm optimization algorithm (PSO). Setting an unclosed electrodes ECT sensor with 10 electrodes as the object for investigation, its geometric parameters were determined by optimized design. Then, an unclosed-electrode sensor was fabricated according to the optimized parameters. Simulations were conducted on image reconstruction and results indicated that the quality of the reconstructed images by the optimized sensor is better than that of the conventional sensors. Furthermore, the optimization method can shorten the time of sensor's design and improve the precision of the optimization solution.A novel image reconstruction algorithm for electrical capacitance tomography was developed. According to the nature of the inverse problems, the minimum norm solution was improved on the basis of algorithm analysis, and a new objective function was established. The regularization technique was employed to stabilize the numerical solution. The images using the unclosed electrodes sensors were reconstructed by simulation and experiment. The results of simulations and experiments indicate that the quality of the IMNSNOF images is better than other compared algorithms such as LBP algorithm, standard Tikhonov regularization and projective Landweber algorithm.An experiment system for frozen soil test was built. One dimension freezing of soil was realized. The permittivity distribution and variation of soil in the freezing process were tested using ECT, and the images were reconstructed displaying frozen region, unfrozen region and the boundary between them. The results obtained by ECT were compared with temperature distribution for validation.