Study On Geological Modeling Techniques For Discrete Fracture Network

The key of effective development for the fractured reservoir is to describe the distribution of the fracture and build the fracture geological model. Now the fractured reservoir models are divided into equivalent continuous model and discrete model. But the former model simplifies the heterogeneity of the fracture and is uncertainty of the crossflow function,it is not able to satisfy the requirement for descriping the distribution of the fracture. In this paper, the theoretical foundation for neural network, fractal geometry, stochastic simulation, combinate the various information in geology, logging, well core, seism and field outcrop about the fracture and calcaulate the parameters of the fracture(fracture density, orientation , distribution and scale) with multi-discipline, then appropriate reservoir geological modeling nethod, this article combines deterministic modeling methods with stochastic modeling methods, finally, three-dimensional geologica model is established. Firstly, the statistics about the parameters of the fracture is obtained from core and FMI data. The part of values well-point in 2D space(fracture density and dimension) or inter-wells in 3D space(fracture density) is predicted with ANN. And then all vaules of the fractal dimension in the pay zone is calcaulated with sequential gaussian simulation. At last the position and orientation and dimensions of the fracture is calcaulated with the method of a stochastic simulaiton based on the density of the fracture, so the discrete fracture network is built. The reliability of the model is checked and analysed. The accomplishment of this paper will be helpful for the fractured reservoir development.