Discrete fracture modeling method based on multi-point geostatistics

Abstract

The invention provides a multipoint geostatistics-based discrete fracture modeling method, which combines multipoint geostatistics modeling with a discrete fracture modeling method, and comprises thefollowing steps of: establishing a training image of multipoint geostatistics simulation consisting of discrete fracture pieces; representing the discrete fracture pieces by convex quadrilateral surface elements, representing the spatial configuration relationship of the fractures by data events formed by the fracture pieces, and calculating the similarity of the fracture piece data events in a manner of overall scaling of the fracture pieces and weighting of morphological parameters of the fracture pieces; and according to a sequential method, adopting a direct sampling mode to extract an appropriate crack mode from a training image to an actual model. According to the method, the structural property of crack space distribution can be guaranteed, the description of the crack scale is alsoreflected, and the configuration relationship of underground cracks of different scales in space can be objectively reproduced.

Description

technical field [0001] The invention relates to the field of oil and gas development, in particular to a discrete fracture modeling method based on multi-point geostatistics. Background technique [0002] Fractured reservoirs are widely distributed in the world and have important economic and strategic potentials. Fractured reservoirs in my country also play an important role in the development and production of oil and gas fields. Fracture geological modeling is a three-dimensional quantitative model that reflects the characterization parameters and spatial distribution of fractures in fractured reservoirs. It can not only reflect the distribution of fractures, but also meet the needs of reservoir engineering research. [0003] The discrete fracture network model displays and characterizes each fracture by fracture slices with different shapes, sizes, orientations, and inclinations in three-dimensional space. Multiple fracture slices with consistent characteristics form a f...

AI Technical Summary

Problems solved by technology

[0005] (1) The existing multi-point geostatistical modeling algorithms are all based on grid pixels, and the training images represent the spatial distribution pattern of geological bodies in the form of a three-dimensional grid model, which can simulate the space between multiple grids The positional relationship is very beneficial in characterizing the spatial distribution of complex sedimentary facies types. Generally, it is only used to characterize the matrix part of fractured reservoirs, and the traditional discrete fracture modeling method is still used for the fracture part, such as the literature "Fracture-cavity carbonate Modeling of rock reservoirs and research on remaining oil distribution" (Li Hongkai et al., China University of Geosciences (Beijing) 2012.6), the matrix model was established using multi-point geostatistics, and the fracture model was established in three directions using traditional discrete fracture modeling , the multi-point geostatistical method was not used in the fracture model; in the literature "Research on karst facies-controlled modeling methods in weathered karst areas of carbonate rocks" (Xinjiang Petroleum Geology, 2015.6), multi-point geostatistics was used to establish the karst facies model , using a deterministic method to build large-scale fractures, using random discrete fracture modeling to build a fracture model, and did not use the multi-point geostatistical method in the fracture model

Both methods do not take into account the complex spatial configuration of fractures

[0006] (2) In terms of the discrete fracture modeling algorithm, the predecessors only improved some links in the simulation framework. The simulation method first generates the fracture development value according to the fracture density field, and then generates the fracture occurrence according to the statistical fracture occurrence. Fractures are generated artificially given the fracture shape, such as in the literature "Stochastic Modeling Method for Discrete Fracture Networks" (Zheng Songqing et al., "Journal of Petroleum and Natural Gas", Vol. 31, No. 4, pp. 106-110, August 2009) , use the spatial variable probability method to determine the fracture location, use the circular statistical method to analyze the fracture occurrence, and use the Fisher distribution and fractal distribution to simulate the fracture occurrence and geometric parameters. It is necessary to give the location probability and fracture geometric parameters for different series of fractures. , the distribution of existing fractures cannot be considered in the simulation process, but this method cannot reflect the spatial configuration of discrete fractures; in the literature "Using seismic data to constrain the spatial distribution of small-scale discrete The density model is transformed into a probability model, and the probability model is used as the termination condition of the simulation to solve the matching problem between the simulated fracture density and the well point data, and this method cannot solve the problem of three-dimensional space configuration of multi-series fractures

[0007] (3) Scholars have tried to apply the idea of ​​multi-point geostatistics to fracture modeling, and the simulation results can reflect the complex spatial configuration of fractures. In Acta Sinica, 2017.6), multi-point geostatistics is used to establish a fracture model, which can consider the spatial configuration relationship of fractures, but the fractures are represented by a grid method, which cannot simulate fractures with a grid scale similar to or smaller than the grid scale. Unable to accurately characterize crack density

[0008] Therefore, the current technical status in this field is: multi-point geostatistical modeling is usually used to simulate 3D geological bodies with equal volumes of sediments in matrix properties; discrete fracture modeling methods only use parameters such as fracture density, occurrence, and shape The statistical law of each group of fractures is simulated separately, and the spatial configuration relationship of multi-scale fractures and multi-series fractures cannot be considered; in the prior art, the pixel grid is used to represent the spatial distribution of fractures, and the existing multi-point geostatistical method is used. Establishing a fracture model, using a pixel grid to represent discrete fractures limits the parameters such as the scale and opening of the fracture, which will result in the inability to simulate fractures smaller than the grid scale and the approximate size of the grid scale, and it is also impossible to simulate multi-scale fractures, which seriously affects the Accuracy of fracture simulation

Images