法国教授 Noetinger学术报告

报告题目：Toward exact minimal discretization of flows in discrete fracture networks, application to modeling of shale gas reservoirs

报告人及简介：Benoit Noetinger，巴黎中央理工学院化石能源系主任、法国石油研究院“新能源”集团高级专家、法国巴黎第六大学教授，美国怀俄明州立大学特邀教授

时间：5月30日下午15：00-16：30

地点：力学一楼239

摘要： Modeling natural Discrete Fracture Networks (DFN) receives more and more attention in applied geosciences,from oil and gas industry,geothermal recovery, to CO2 sequestration issues and evidently shale oil or gas recovery. Accounting for the flow inside the fracture network, and accounting for the transfers between the matrix and the fractures,with the same level of accuracy is an important issue for calibrating the wells architecture and for setting up optimal resources recovery strategies. Recently, we proposed a method allowing to model transient pressure diffusion in the fracture network only. The matrix was assumed to be impervious.A systematic approximation scheme was built,allowing to model the initial DFN by a set of $N$ unknowns located at the intersection between fractures.The higher $N$, the higher the accuracy of the model.The lowest order approximation $N=1$ appears under the form of solving a transient problem in a resistor/capacitor network, a so-called pipe network. Its topology is the same as the network of geometrical intersections between fractures. Formal expressions were given for determining the capacity (or mass) matrix, and the conductivity matrix, that depends on both the local shape of the fracture and the geometry of the intersections.

Here, we generalize this approach in order to account for fluxes from matrix to fractures. We show that in the case of well separated time scales between matrix and fractures, the preceding model need only to be slightly modified in order to incorporate these fluxes.The additional knowledge of the so called matrix to fracture transfer function allows to modify the mass matrix that becomes a time convolution operator.This is reminiscent of existing space averaged transient dual porosity models.This helps us to develop of model of flow in DFN using a number of degrees of freedom close to the number of intersection between fractures,avoiding adding extra unknowns corresponding to a detailled meshing of the fractures.Application involving several millions of fractures can be done.

Finally, using some geomechanical coupling (relation fracture aperture stress tensor), we were able to study a real case of a gas shale reservoir naturally fractured hydraulically stimulated, and to get some predictive simulations for various stages of the hydraulic fracturing process.

Keywords: Flow in Fractured media, Discrete fracture network, Quasi steady state, Dual porosity, Transfer function, hydraulic fractures, geomechanical coupling, mico sismicity