Three-Dimensional Reconstruction of Porous Media Based on Simulated Annealing
Wang Meiling, Guo Ruoyu, Xiao Meifeng, Liu Tong*, and Fu Huijin
(School of Automation, Beijing Institute of Technology, Beijing 100081)
In order to study fluid flow in a porous transducer, a three-dimensional random reconstruction strategy of porous media is proposed based on simulated annealing. On the basis of dynamic temperature reduction factor, the simulated annealing was conducted using an improved status update method. Binary images of pore and solid phase, which were considered as reference model, were acquired by processing slice images collected by micro CT. The sampling, processing and reconstruction of the porous media were realized with the porosity of the porous media unchanged. Experimental results show that the proposed three-dimensional reconstruction strategy can increase the reconstruction speed and accuracy of permeability calculation, and the pore fraction dimension as well as two-point correlation function of the reconstructed model coincides with the reference model.