CN-122021341-A - Fractured shale reservoir fluid injection fracture activation process simulation method

CN122021341ACN 122021341 ACN122021341 ACN 122021341ACN-122021341-A

Abstract

The invention discloses a simulation method of a fracture injection and fracture activation process of a fractured shale reservoir fluid, which belongs to the technical field of unconventional oil and gas exploitation, and comprises the steps of inputting a multi-scale fracture in-situ parameter characteristic data set into an inversion model, inverting multi-scale fracture in-situ parameter to obtain an inversion result, constructing a three-dimensional multi-scale fracture geometric model, fusing the inversion result into the geometric model, constructing a three-dimensional multi-scale discrete fracture network model, constructing a macroscopic heterogeneous shale reservoir matrix model based on a shale macroscopic rock mechanical parameter distribution function, embedding the three-dimensional multi-scale discrete fracture network model into the macroscopic heterogeneous shale reservoir matrix model to form a shale reservoir model, constructing the fracture injection and fracture activation model according to the shale reservoir model, and performing simulation characterization on a fracture activation sliding process. The fracture activation process numerical model constructed by the method can accurately describe the dynamic evolution process from the stable state to the unsteady activation of the crack.

Inventors

LIU ZHAOYI
LIANG PINGPING
HAN LINGLING
WANG XIANGYU
LI XIZHE
DUAN GUIFU
LI KUNLONG
Cui Hanzhuo
ZHAO WANCHUN
WU ZHENKAI

Assignees

东北石油大学

Dates

Publication Date: 20260512
Application Date: 20260403

Claims (9)

1. The method for simulating the injection fracture activation process of the fractured shale reservoir fluid is characterized by comprising the following steps of: acquiring logging imaging data, microseism data and seismic data of a shale reservoir in a research area, and constructing a multi-scale fracture in-situ parameter characteristic data set; Inputting a multi-scale fracture in-situ parameter characteristic data set into a pre-trained inversion model, inverting the multi-scale fracture in-situ parameter to obtain an inversion result, wherein the inversion model is based on a physical information neural network, and a residual loss item for coupling shale seepage law and mechanical strength constraint is added after an output layer; fusing the inversion result into the geometric model to construct a three-dimensional multi-scale discrete fracture network model; Embedding the three-dimensional multi-scale discrete fracture network model into the macroscopic heterogeneous shale reservoir matrix model to form a shale reservoir model reflecting the multi-scale discrete fracture network system and reservoir heterogeneity; And establishing a fracture activation judgment criterion by adopting a coulomb friction law according to a shale reservoir model, comprehensively considering thermal stress, thermal expansion caused by external fluid injection, an in-situ shale reservoir environment stress field and a thermal-fluid-solid-damage multi-physical field coupling effect, constructing a fracture activation model for injecting a fractured shale reservoir fluid into the fracture, and performing simulation characterization on a fracture activation sliding process.
2. The method for simulating the injection fracture activation process of the fractured shale reservoir fluid according to claim 1, wherein the multi-scale fracture in-situ parameter characteristic dataset is input into a pre-trained inversion model, inversion is carried out on the multi-scale fracture in-situ parameter to obtain an inversion result, and the inversion model is based on a physical information neural network, and a residual loss term for coupling shale seepage law and mechanical strength constraint is added after the output layer, and specifically comprises the following steps: determining the input parameters of the physical information neural network as The output parameters are obtained as follows: ; Wherein, the Representing the first physical information neural network The number of output items is chosen to be the number of output items, Representing the first physical information neural network I represents an index of in-situ characteristic parameters of the multi-scale fracture, Represent the first The weight coefficient corresponding to the in-situ characteristic parameters of the multi-scale cracks, N represents the dimension of the in-situ characteristic parameter dataset of the multi-scale fracture; the introduced residual loss term couples shale percolation law and mechanical strength constraints, where: the residual loss term of the percolation law constraint is: ; Wherein, the Indicating the permeability of the fracture and, Represent the first N represents the number of items and refers to the number of the whole group of data; the residual loss term of the mechanical strength constraint is: ; Wherein, the And Respectively represent the first in shale reservoirs Shear stress and positive stress corresponding to in-situ characteristic parameters of the multi-scale cracks, And Representing predicted cohesion and internal friction angle of the crack; according to the residual loss term constrained by the seepage law and the residual loss term constrained by the mechanical strength, defining a total loss function as follows: ; Wherein, the The super-parameter is represented by a parameter, Representing a predictive penalty of the bayesian deep learning network; optimizing output parameters of the physical information neural network through a total loss function to obtain inversion results, wherein the inversion results comprise mechanical parameters and seepage parameters under multi-scale crack differences; the mechanical parameters are shear stress and positive stress of shale, predicted cohesive force and internal friction angle of the crack, and the seepage parameters are permeability and pressure gradient of the crack.
3. The method for simulating the injection fracture activation process of the fractured shale reservoir fluid according to claim 1, wherein the method for constructing the three-dimensional multi-scale fracture geometric model comprises the steps of fusing inversion results into the geometric model to construct a three-dimensional multi-scale discrete fracture network model, and specifically comprises the following steps: based on an MATLAB platform, constructing a three-dimensional multi-scale crack geometric model by randomly generating a plurality of groups of disc-shaped cracks; Performing three-dimensional grid division on the constructed three-dimensional multi-scale fracture geometric model, distributing an inversion result to each divided grid node, and endowing corresponding rock mechanical parameters for each grid node so as to simulate the rock mechanical characteristics of the fracture and construct a three-dimensional multi-scale discrete fracture network model containing in-situ parameter characteristics.
4. The method for simulating the injection fracture activation process of a fractured shale reservoir fluid according to claim 1, wherein the constructing a macroscopic heterogeneous shale reservoir matrix model based on the shale macroscopic rock mechanical parameter distribution function specifically comprises: Based on Weibull distribution, the shale macroscopic rock mechanical parameter distribution function is constructed as follows: ; Wherein, the Respectively representing rock mechanical parameters, average values of the rock mechanical parameters and inhomogeneous coefficients; According to the shale macroscopic rock mechanical parameter distribution function, the spatial distribution of shale reservoir rock mechanical parameters is simulated by introducing a heterogeneous coefficient, and a heterogeneous shale reservoir geological model is generated, so that the macroscopic heterogeneous shale reservoir model is constructed.
5. The method for simulating a fracture activation process of fluid injection into a fractured shale reservoir according to claim 1, wherein the method is characterized by establishing a fracture activation criterion by adopting coulomb friction law according to a shale reservoir model, comprehensively considering thermal stress, thermal expansion, in-situ shale reservoir environmental stress field and thermal-fluid-solid-damage multi-physical field coupling effect caused by external fluid injection, constructing a fracture activation model of fluid injection into the fractured shale reservoir, and performing simulated characterization on a fracture activation sliding process, and specifically comprises the following steps: acquiring a plurality of groups of shale reservoir model parameters according to the macroscopic heterogeneous shale reservoir model; Taking a plurality of groups of shale reservoir model parameters as input parameters of a block discrete element development platform, constructing a mechanical and seepage parameter model considering different crack dimensions, and endowing corresponding mechanical properties and seepage properties for cracks of different dimensions; Describing fracture surface and reservoir rock deformation by adopting a discrete element method, discretizing a tetrahedral mesh of the reservoir, describing rock matrix deformation by adopting a linear elastic constitutive model, and constructing a stress field equation; Describing fluid flow in the crack based on a simplified Navier-Stokes equation, constructing a flow field equation, and establishing a relation between fluid pressure and crack volume change; Describing heat exchange between fluid and solid based on Fourier law and energy conservation equation, and constructing a temperature field equation; Based on a stress field equation, a flow field equation and a temperature field equation, introducing a thermal strain increment and pore pressure change caused by thermal expansion, constructing a thermal-fluid-solid coupling mechanism, coupling the thermal stress increment and the pore pressure increment into a mechanical constitutive equation, and constructing a thermal-fluid-solid-damage multi-physical field coupling control equation; Establishing a fracture activation judgment criterion by taking a coulomb friction law as a basic component model of fracture, and judging as tensile failure when normal stress exceeds tensile strength, and judging as shear failure when shear stress exceeds maximum allowable shear force based on cohesive force, an internal friction angle and normal stress; The normal and tangential stress increment caused by thermal stress is added into Coulomb friction law to correct critical conditions of crack activation; According to the mechanical properties and the seepage properties, a fracture activation model of the fractured shale reservoir fluid injection is constructed based on a thermal-flow-solid-damage multi-physical field coupling control equation, and a numerical simulation is carried out on a fracture activation sliding process to obtain a multi-scale fracture activation and expansion sliding diagram in the fracturing injection process.
6. The method of claim 5, wherein the thermal-fluid-solid-damage multi-physical field coupling control equation is: ; Wherein, the And Is the stress tensor component and the strain tensor component corresponding to the ith fracture to the jth fracture in the fluid injection process; is the kronecker operator symbol whose table below represents the ij-th component of the symmetric tensor matrix, where B is the Biot coefficient; Is the shear modulus of the rock; is the bulk modulus of the rock and, E is the elastic modulus of the rock, v is the Poisson's ratio of the rock; Is the linear thermal expansion coefficient of rock; is the sum of the strain tensors of the rock in different directions.
7. The method for simulating the fracture activation process of the fluid injection of the fractured shale reservoir according to claim 5, wherein the step of simulating and characterizing the fracture activation slip process further comprises the step of determining fracture activation distribution characteristics according to multi-scale fracture activation and expansion slip graphs obtained in the fracturing injection process in combination with microseism event evaluation, and specifically comprises the steps of: According to a clustering mechanism, when two sliding sections in the multi-scale crack activation and expansion sliding diagram are adjacent to each other and sliding duration time is overlapped, combining sliding nodes within a preset radius R range, and updating the microseism sliding sections; Determining the slip speed according to the updated microseism slip segment, and judging that the earthquake slip exists when the slip speed is larger than a speed threshold value, or judging that the earthquake slip does not exist; Aiming at the sliding section judged to have earthquake sliding, calculating earthquake moment and magnitude according to the sliding quantity, the sliding fracture area and the rock shear modulus of the crack, and generating a micro-seismic ball; And according to the space position of the microseism ball, establishing a one-to-one correspondence between fault sliding velocity and microseism events, and determining the space distribution characteristics of fracture activation.
8. The method for simulating the injection fracture activation process of a fractured shale reservoir according to claim 1, wherein the steps of obtaining logging imaging data, microseism data and seismic data of the shale reservoir in the research area and constructing a multi-scale fracture in-situ parameter characteristic data set comprise the following steps: collecting logging imaging data and microseism data of an oil and gas field construction site and earthquake data; carrying out gray level processing and binary processing on logging imaging data to obtain micron-centimeter-level in-situ characteristic parameters of the well periphery microcracks; inversion solution is carried out based on a seismic source mechanism according to microseism data, and in-situ characteristic parameters of the meter-scale cracks are obtained; processing the seismic data based on an ant body method to obtain in-situ characteristic parameters of hundred-meter-level fracture; And fusing in-situ characteristic parameters of the micro-cm-level well periphery micro-cracks, the meter-level cracks and the hundred-meter-level cracks to construct a multi-scale crack in-situ parameter characteristic data set.
9. A fractured shale reservoir fluid injection fracture activation process simulation system, comprising: The multi-scale data set construction module is used for acquiring logging imaging data, microseism data and seismic data of the shale reservoir in the research area and constructing a multi-scale fracture in-situ parameter characteristic data set; The parameter inversion module is used for inputting the multi-scale fracture in-situ parameter characteristic data set into a pre-trained inversion model, inverting the multi-scale fracture in-situ parameter to obtain an inversion result, wherein the inversion model is based on a physical information neural network, and a residual loss item for coupling shale seepage law and mechanical strength constraint is added after the inversion model is output; the shale reservoir model construction module is used for constructing a three-dimensional multi-scale fracture geometric model, fusing inversion results into the geometric model to construct a three-dimensional multi-scale discrete fracture network model, constructing a macroscopic heterogeneous shale reservoir matrix model based on a shale macroscopic rock mechanical parameter distribution function, embedding the three-dimensional multi-scale discrete fracture network model into the macroscopic heterogeneous shale reservoir matrix model to form a shale reservoir model reflecting the multi-scale discrete fracture network system and reservoir heterogeneity; And the crack activation slip process simulation module is used for establishing a crack activation judgment criterion by adopting a coulomb friction law according to a shale reservoir model, comprehensively considering thermal stress, thermal expansion, in-situ shale reservoir environment stress field and thermal-fluid-solid-damage multi-physical field coupling effect caused by external fluid injection, constructing a crack activation model for injecting a crack shale reservoir fluid, and carrying out simulation characterization on the crack activation slip process.

Description

Fractured shale reservoir fluid injection fracture activation process simulation method Technical Field The invention relates to the technical field of unconventional oil and gas exploitation, in particular to a method for simulating a fluid injection fracture activation process of a fractured shale reservoir. Background With the deep advancement of global energy structure transformation strategic targets, efficient and green development of unconventional oil and gas resources, gas storage and carbon storage technologies have become core issues in the fields of energy and environmental protection. Shale reservoirs are important unconventional oil and gas resources, and commercial development of the shale reservoirs is highly dependent on complex fracture networks formed by large-scale hydraulic fracturing technologies. The external fluid is injected into the fractured shale reservoir, so that the fault sliding is often activated and serious sleeve-changing risks are generated while the reservoir oil gas is effectively reformed and the diversion capacity is improved, and the single well yield is greatly influenced. However, currently, the activation characterization problem of fractures in the process of fracturing injection fluid is generally idealized and uniform under the condition of multi-scale fracture attribute parameter characterization, and the urgent requirements of the current technical attack and efficient development of complex deep unconventional energy are difficult to meet. Under the background, a multi-scale fracture stratum is constructed by deep learning, and the technology of the in-situ reservoir fracture activation slip process is characterized by a multi-object flow field method. Because the fracturing and seepage processes of external fluid injection into a fractured shale reservoir are complex nonlinear problems related to thermal-fluid-solid-damage multi-physical field intensity coupling, the traditional fractured shale reservoir fluid injection method mainly adopts a numerical simulation method, and most of fracture networks are simplified or idealized, the physical attribute difference of multistage fractures is not considered, a discrete fracture system under in-situ geological conditions is difficult to truly reflect, and large deviation exists in the simulation of fracture activation dynamics. Disclosure of Invention Aiming at the problems in the field, the invention provides a method and a system for simulating the injection fracture activation process of a fractured shale reservoir fluid, and the constructed fractured shale reservoir fluid injection fracture activation model has the advantages that the direct association relation between the stress change and the temperature and pore pressure change is clarified, and the dynamic evolution process from the stable state to the destabilization activation of the fracture can be accurately depicted. In order to solve the technical problems, the invention discloses a method for simulating the injection fracture activation process of a fractured shale reservoir fluid, which comprises the following steps: acquiring logging imaging data, microseism data and seismic data of a shale reservoir in a research area, and constructing a multi-scale fracture in-situ parameter characteristic data set; Inputting a multi-scale fracture in-situ parameter characteristic data set into a pre-trained inversion model, inverting the multi-scale fracture in-situ parameter to obtain an inversion result, wherein the inversion model is based on a physical information neural network, and a residual loss item for coupling shale seepage law and mechanical strength constraint is added after an output layer; fusing the inversion result into the geometric model to construct a three-dimensional multi-scale discrete fracture network model; Embedding the three-dimensional multi-scale discrete fracture network model into the macroscopic heterogeneous shale reservoir matrix model to form a shale reservoir model reflecting the multi-scale discrete fracture network system and reservoir heterogeneity; And establishing a fracture activation judgment criterion by adopting a coulomb friction law according to a shale reservoir model, comprehensively considering thermal stress, thermal expansion caused by external fluid injection, an in-situ shale reservoir environment stress field and a thermal-fluid-solid-damage multi-physical field coupling effect, constructing a fracture activation model for injecting a fractured shale reservoir fluid into the fracture, and performing simulation characterization on a fracture activation sliding process. Preferably, the multi-scale fracture in-situ parameter characteristic dataset is input into a pre-trained inversion model, inversion is carried out on the multi-scale fracture in-situ parameter characteristic dataset, and an inversion result is obtained, wherein the inversion model is based on a physical information neural network, and a r