Rock Image Reconstruction And Electrical Properties Calculation Based On Improved MCMC Method

It is generally believed that the calculation accuracy of electrical parameters of rock samples depends on the modeling accuracy of pore structure.Therefore,digital core is committed to establishing a micro geometric model with high reducibility and good connectivity.Markov chain Monte Carlo(MCMC)method can use the state trend and sampling of pixels to construct the unknown part of the model.However,using the same method to reconstruct different kinds of core images will get unreasonable results,and the relevant optimization of MCMC algorithm is based on the assumption of sample isotropy,so it is not suitable for the reconstruction of heterogeneous anisotropic rock samples.In view of the diversity of reconstruction path and neighborhood selection,and the randomness and uncertainty of reconstruction results,this paper analyzes the influence of anisotropy on neighborhood and path selection,and the good effect of this influence on reconstruction results,Control variables are used to analyze the reconstruction effect of different reconstruction components and their applicability to different images.The main work is as follows:(1)In this paper,on the basis of previous studies,experiments are carried out on the isotropic assumption of the current MCMC algorithm in the reconstruction of anisotropic rock images.The results show that the transfer matrix of different scanning directions will show differences.It is proved that the improved MCMC algorithm based on the reconstruction path construction transfer matrix proposed in this paper has prominent application advantages in the rock image reconstruction with obvious anisotropic characteristics.(2)In order to explore the relationship between image properties and reconstruction components,the potential function is used to construct classical isotropic / anisotropic images,and the differences between the two kinds of images in pixel distribution and variation curve fitting are summarized.And then,the reconstruction effect of two kinds of images under different conditions is analyzed by controlling variables.In the reconstruction of core images with different characteristics,it is found that the characteristics of small pore images are basically consistent with isotropic images,which proves the isotropy of this kind of core images and its wide applicability in the selection of reconstruction components.(3)The known information is extracted from the original sequence imaging model as the initial condition of three-dimensional MCMC reconstruction,and a 3D(three-dimensional)rock sample model with good effect in porosity and variogram curve fitting is successfully constructed.The cubic mesh is divided and the 3D-FDM(Finite Difference Method)is used to calculate the equivalent electrical parameters.The results show that the equivalent permittivity of rock samples with different sizes obtained by sequential imaging and 3D MCMC reconstruction are consistent,which proves the feasibility and correctness of 3D MCMC.In conclusion,this paper proposes an improved MCMC modeling method,and successfully modeling the reservoir rock samples.It not only makes up the discomfort of the existing methods for anisotropic rock samples,improves the accuracy of core sample modeling and electrical parameter calculation,but also further enhances the accuracy of reservoir evaluation,which is of great significance for oil exploration.