Fracture Network Model Of The Ordovician Carbonate Rock In The 6th Block,Tahe Oilfield

The petroleum reserves of Ordovician carbonate rock in Tahe Oilfield take up 90% of the total,and fractures are the main factors controlling the oil and gas accumulation and productivity,so it must be pointed out that it is very important to make clear the law of fracture development and realize the three-dimensional distribution of it.Based on the predecessors' research on carbonate karst fracture-cave in Tahe oilfield,the Ordovician in the 6th Block is selected as the study object to conduct the fracture network model research,which aims at establishing the fracture network model and characterizing the development and the distribution of fracture.Referencing the research results of predecessors and integrating the fracture information effectively from many sources,such as outcrop,core,logging and seismic data,identification and characterization of the 6th block of Tahe Oilfield Ordovician fracture is been required.The results of statistical analysis show that the fracture in the study area is dominated by the high angle fracture,and the fracture density is large,which mainly include NE-SW,NNW-SSE(near N-S direction)and NEE-SWW(near E-W direction)three groups.On the basis of effective identification of the fracture and statistical analysis,3D modeling of discrete fracture network is established through the method of combination of deterministic simulation and stochastic simulation.First of all,large-scale fracture has been identified and described by using seismic attribute.The large-scale fracture model in the study area is established mainly by the way of ant tracking technology and the distribution model of large scale fracture is built through the method of deterministic modeling.Then,the density of small-scale fracture around the wells which cannot be identified by the seismic data is predicted through artificial neural network(ANN)on the basis of interpretation of imaging well logging.Under ant body volume constraints,the density distribution model of small-scale fracture is set up by sequential Gaussian simulation.Finally,under the constraints of fracture occurrence statistical data and the fracture density distribution model,the location,the orientation and the size of fracture is simulated stochastically through marked point process simulation and simulated annealing algorithm.The discrete fracture network model has been established eventually,also has been checked and analyzed.The result illustrates that the characteristics and distribution of fracture is mainly affected by tectonic stress and tectonic position.The fracture developed very well,mostly in direction NE,is related to the regional NE compressional stress field for a long time,and on both sides of the fault and the local structural high position,high degree of fracture development.In this paper,many kinds of mathematical methods including neural network,ant tracking and statistical analysis are adopted,which have a good support and constraint to the geological information.At the same time,the modeling uncertainty has been effectively reduced and the problem of the fracture modeling is solved well,as a result of the method of through the ant tracking technology to establish a large-scale fracture model,using BP neural network and stochastic simulation to set up the small-scale fracture model.