Multi-objective optimization method and device for CO2 oil displacement storage of high-water-content fault block oil reservoir

Abstract

The invention discloses a high-water-cut fault block oil reservoir CO2 oil displacement storage multi-objective optimization method, which comprises the following steps of: establishing a high-water-cut fault block oil reservoir CO2 oil displacement storage fluid-solid coupling model, sampling based on a Latin hypercube sampling method to generate samples, and constructing an objective agent model by adopting a long-short-term memory neural network and the samples, and introducing a fast non-dominated sorting genetic multi-objective optimization algorithm, and optimizing the plurality of objective functions based on the constructed proxy model to obtain a Pareto solution set of the objective functions. Thus, based on a comprehensive framework of the three-dimensional oil reservoir seepage and geomechanics coupling model, the LSTM algorithm and the multi-objective optimization function NSGA-II, an injection-production well production system scheme can be obtained through optimization, the crude oil recovery rate and CO2 storage amount are maximized, and fault sliding displacement is minimized; and the CO2 oil displacement storage joint optimization efficiency and the result accuracy of the high-water-content fault block oil reservoir are improved.

Description

Technical Field

[0001] The present invention relates to the field of oil and gas field development, and in particular to a multi-objective optimization method and device for CO2 oil recovery and storage in a high-water-content fault block reservoir. Background Art

[0002] Old oil fields in eastern China (such as Daqing and Shengli oil fields) have entered the high water content development stage. Studies have shown that CO2 flooding in high water content reservoirs can generally extend the reservoir production life by 15 to 20 years and increase the recovery rate by 8% to 15%. In addition, during the process of CO2 flooding to enhance oil recovery (Carbon Dioxide Enhanced Oil Recovery, CO2-EOR), a portion of CO2 will be permanently stored underground in different forms. In order to obtain the maximum recovery rate and CO2 storage, the production strategy of injection and production wells must be optimized. In addition, since many oil fields usually have complex faults, in such reservoirs, it is also necessary to evaluate the geomechanical risks of fault activation, which transforms CO2 flooding and storage into a multi-objective optimization problem.

[0003] In the related technologies, the technical problems of CO2 oil recovery and sealing in high water-content fault block reservoirs are low CO2 oil recovery volume, small CO2 storage volume and high geological risk. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a multi-objective optimization method and device for CO2 oil recovery and storage in high-water-content fault block reservoirs, so as to solve the disadvantages of small CO2 oil recovery volume, low CO2 storage volume and high geological risk in the prior art of water-to-gas drive development of high-water-content oil reservoirs.

[0005] In order to achieve the above object, the technical solution of the embodiment of the present invention is:

[0006] A first aspect of the present invention provides a multi-objective method for CO2 flooding and oil storage in a high water-cut fault block reservoir, the method comprising:

[0007] Establish a fluid-solid coupling numerical simulation model for CO2 flooding and storage in high water-cut fault block reservoirs;

[0008] Based on the Latin hypercube sampling method, the injection well system is sampled as an input parameter, and the CO2 flooding and oil storage fluid-solid coupling numerical simulation model is used for calculation to obtain the output and fault slip displacement samples under different injection systems;

[0009] The constructed samples are input into the long short-term memory neural network and trained using the Adam optimization algorithm until the root mean square error meets the preset conditions to obtain the target proxy model;

[0010] Based on the target proxy model, a multi-objective optimization algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault slip displacement objective function to obtain the corresponding optimization result of each objective function.

[0011] A second aspect of the present invention provides a multi-objective optimization device for CO2 flooding and oil storage in a high water-cut fault block reservoir, the device comprising:

[0012] Establish a module for establishing a fluid-solid coupling numerical simulation model for CO2 flooding and storage in high water-cut fault block reservoirs;

[0013] A calculation module is used to sample the injection well system based on the Latin hypercube sampling method as an input parameter, and to calculate based on the CO2 flooding and oil storage fluid-solid coupling numerical simulation model to obtain output and fault slip displacement samples under different injection systems;

[0014] The training module is used to input the constructed samples into the long short-term memory neural network and use the Adam optimization algorithm for training until the root mean square error meets the preset conditions to obtain the target proxy model;

[0015] The optimization module is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault sliding displacement objective function based on the objective proxy model using a multi-objective optimization algorithm to obtain the corresponding optimization result of each objective function.

[0016] In some embodiments, the establishment module is further used to determine the linear relationship between the ultimate strength, shear stress and normal stress of each stress-bearing surface; the linear relationship is shown in the following formula;

[0017]

[0018] In the formula, τ n is the shear stress on the failure surface, σ n is the normal stress of the failure surface, C is the cohesion of the rock; is the internal friction angle of rock;

[0019] According to the shear stress σ of the fault plane t and σ s , determine the maximum shear stress τ max ; The maximum shear stress is shown in the following formula:

[0020]

[0021] According to the maximum shear stress, the sliding displacement of the fault plane is determined; the sliding displacement is shown in the following formula:

[0022]

[0023] In the formula, K t is the stiffness of the fault;

[0024] According to the value of the sliding displacement, the friction coefficient of the sliding friction is determined; the friction coefficient is shown in the following formula:

[0025]

[0026]

[0027]

[0028] In the formula, is the minimum friction angle, is the maximum value of fault friction, is the minimum value of fault friction;

[0029] The slip trend of the fault is determined according to the maximum shear stress and the friction coefficient; the slip trend is shown in the following formula:

[0030]

[0031] In the formula, cohes is the cohesion of the fault. When the pore pressure increases, the Mohr circle moves to the left.

[0032] In some embodiments, the calculation module is further used to assume a structure function function with K-dimensional basic variables, where N represents the selection of N groups x=(x1, x2, ..., x n ), the matrix P is an N*K matrix, the column elements of P are randomly generated by integers from 1 to N, and the value of each element is not equal, the matrix R is also an N*K matrix, the column elements of R are randomly generated by rational numbers from 0 to 1, and the value of each element is not equal;

[0033]

[0034]

[0035] In the formula, s ij is the i-th row and j-th column element of the matrix S, is the distribution function that the jth variable conforms to, is the jth variable of the i-th group of samples obtained by Latin hypercube sampling.

[0036] In some embodiments, the training module is further used to perform training using a long short-term memory neural network to define weight and bias parameters of the gate; the weight and bias are shown in the following formula:

[0037]

[0038] Among them, W represents weight, b represents bias, i represents input gate, f represents forget gate, and o represents output gate;

[0039] In a single hidden unit of the long short-term memory neural network, a Sigmoid activation function is used to determine a forget gate; the forget gate is shown in the following formula:

[0040] f t =σ(W f ·[h t-1 ,x t ]+b f ),

[0041] In the formula, f t is the forget gate, σ represents the activation function, and the expression is σ(x)=(1+e -x ) -1 , whose output range is from 0 to 1, x t is the input of the current time step, h t-1 is the input of the previous time step, W f With b f is the weight and bias of the forget gate;

[0042] The update of the input gate is determined according to the Sigmoid activation function, and at the current step length, whether the current input is written by the information of the memory cell is determined based on the input gate; the input gate is shown in the following formula:

[0043] i t =σ(W i ·[h t-1 ,x t ]+b i ),

[0044] In the formula, i t is the input gate, b i is the bias matrix of the input gate;

[0045]

[0046] In the formula, tanh represents a hyperbolic tangent function with an output range of -1 to 1;

[0047] According to the historical cell state, update the current cell state as shown in the following formula:

[0048]

[0049] In the formula, C t-1 is the historical cell state, C t is the current cell state, is the vector of new candidate values;

[0050] Output is made through the output gate, which controls whether the information stored in the memory cell affects the subsequent network layer;

[0051] o t =σ(W o [h t -1,x t ]+b o ),

[0052] In the formula, o t is the output gate, W o With b o are the weight and bias of the output gate, h t It is the output value of the current cell state and also the input value of the next cell state.

[0053] In some embodiments, the optimization module is further used to take the production system parameters of the injection and production wells as independent variables, and take the recovery rate, CO2 storage rate and fault sliding displacement as dependent variables, so as to maximize the recovery rate objective function and the CO2 storage rate objective function and minimize the fault sliding displacement objective function; define the following formula:

[0054] maxF(x)=[(f1(x),f2(x),-f3(x)) T ],

[0055]

[0056] Where OOIP is the initial recoverable reserves of the reservoir; is the oil production at time t; is the total amount of CO2 injected into the injection well at time t; is the total amount of CO2 produced by the production well at time t, T is the total time for reservoir development evaluation, and D s is the fault slip displacement;

[0057] The NSGA-Ⅱ algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault sliding displacement objective function to obtain the corresponding optimization results.

[0058] The third aspect of the present invention provides an electronic device, comprising: a memory for storing executable instructions; a processor for implementing the above-mentioned multi-objective optimization method for CO2 oil recovery and storage in high-water-content fault-block reservoirs when executing the executable instructions stored in the memory.

[0059] A fourth aspect of the present invention provides a computer-readable storage medium storing executable instructions for causing a processor to execute the executable instructions to implement the above-mentioned multi-objective optimization method for CO2 flooding and storage in high-water-content fault-block reservoirs.

[0060] The multi-objective optimization method for CO2 flooding and oil storage in high-water-content fault block reservoirs provided by the present invention first establishes a fluid-solid coupling numerical simulation model for CO2 flooding and oil storage in high-water-content fault block reservoirs, then performs sampling based on the Latin hypercube sampling method and generates samples by numerical simulator calculation, uses a long short-term memory neural network to build an agent model, introduces the NSGA-II multi-objective optimization algorithm to perform multi-objective optimization, and obtains the Pareto solution set of the objective function. In this way, on the one hand, the present invention, based on the comprehensive framework of the three-dimensional flow reservoir seepage and geomechanics coupling model, the artificial intelligence algorithm LSTM and the multi-objective optimization algorithm NSGA-II, can obtain the injection and production well production system plan through optimization, thereby achieving the maximization of crude oil recovery and CO2 storage and the minimization of fault sliding displacement. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a flow chart of a multi-objective optimization method for CO2 flooding and oil storage in a high water-cut fault block reservoir provided by an embodiment of the present invention;

[0062] Figure 2 is a schematic diagram of the Mohr-Coulomb yield criterion provided by an embodiment of the present invention;

[0063] Figure 3 is a schematic diagram of the structure of a long short-term memory neural network model provided by an embodiment of the present invention;

[0064] Figure 4 It is a flowchart of the NSGA-II algorithm provided by an embodiment of the present invention;

[0065] FIG5 is a diagram of a reservoir model attribute field model provided by an embodiment of the present invention, wherein (a) is a permeability field, and (b) is a porosity field;

[0066] Figure 6 is a schematic diagram of a network structure provided by an embodiment of the present invention;

[0067] Figure 7 1 is a diagram of prediction results of the long short-term memory neural network proxy model provided by an embodiment of the present invention, (a) is a diagram of recovery rate prediction results, (b) is a diagram of CO2 storage rate prediction results, and (c) is a diagram of fault slip displacement prediction results;

[0068] Figure 8 is a Pareto solution set after NSGA-II optimization provided by an embodiment of the present invention;

[0069] FIG9 is a diagram of NSGA-II optimization statistical results provided by an embodiment of the present invention, wherein (a) is a diagram of recovery factor statistical results, (b) is a diagram of sealing factor statistical results, and (c) is a diagram of fault sliding displacement statistical results;

[0070] Fig.10It is a schematic diagram of the composition structure of a multi-objective optimization device for CO2 oil recovery and storage in a high water-cut fault block reservoir provided by an embodiment of the present invention;

[0071] Fig.11 It is a schematic diagram of the composition structure of a multi-objective optimization device for CO2 oil recovery and storage in a high-water-content fault-block reservoir provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present invention. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present invention.

[0073] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of the present invention have the same meaning as those commonly understood by those skilled in the art to which the embodiments of the present invention pertain. The terms used in the embodiments of the present invention are only for the purpose of describing the embodiments of the present invention and are not intended to limit the present invention.

[0074] In high-water-content fault block reservoirs, due to long-term water injection development or other mining activities, the water content in the reservoir gradually increases, resulting in a decrease in crude oil recovery. In order to improve the recovery rate, a variety of methods have been adopted, among which CO2 oil recovery and storage technology is an effective means. This technology injects CO2 into the reservoir and uses the dissolution, expansion and viscosity reduction effects of CO2 to improve the fluidity of crude oil, thereby achieving effective crude oil extraction. At the same time, the injected CO2 can also be stored underground to achieve the purpose of reducing greenhouse gas emissions.

[0075] The following describes an exemplary application of a multi-objective optimization device for CO2 flooding and oil sealing in a high water-cut fault block reservoir according to an embodiment of the present invention. The multi-objective optimization device for CO2 flooding and oil sealing in a high water-cut fault block reservoir provided by an embodiment of the present invention can be implemented as a terminal or a server. In one implementation, the multi-objective optimization device for CO2 flooding and oil sealing in a high water-cut fault block reservoir provided by an embodiment of the present invention can be implemented as various types of terminals such as a laptop computer, a tablet computer, a desktop computer, and a mobile device; in another implementation, the multi-objective optimization device for CO2 flooding and oil sealing in a high water-cut fault block reservoir provided by an embodiment of the present invention can also be implemented as a server, wherein the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (CDN, Content Delivery Network), and big data and artificial intelligence platforms. The terminal and the server can be directly or indirectly connected by wired or wireless communication, which is not limited in the embodiment of the present invention. Next, an exemplary application will be described when the multi-objective optimization device for CO2 flooding and storage in a high water-cut fault block reservoir is a server.

[0076] The embodiment of the present invention provides a multi-objective optimization method for CO2 flooding and storage in a high water-cut fault block reservoir. Figure 1 , Figure 1 This is a flow chart of a multi-objective optimization method for CO2 flooding and oil storage in a high water-cut fault block reservoir provided by an embodiment of the present invention. Figure 1 The steps shown are explained.

[0077] Step S101, establishing a fluid-solid coupling numerical simulation model for CO2 flooding and oil storage in a high water-cut fault block reservoir.

[0078] In some embodiments, a high water content reservoir refers to a reservoir having a relatively high water content in the formation.

[0079] In some embodiments, a fault block refers to a rock block that is formed due to stress inside the earth's crust and undergoes significant displacement along a certain fracture surface. In oil and gas reservoir geology, a fault block reservoir refers to an oil and gas accumulation unit formed by the division and closure of faults.

[0080] In practical applications, when conducting fluid-solid coupling numerical simulations for CO2 flooding and oil storage in high-water-content fault-block reservoirs, it is necessary to fully consider the geological characteristics and fluid distribution characteristics of the fault-block reservoirs, as well as the interaction between fluid flow and reservoir deformation caused by CO2 injection. This interaction is called fluid-solid coupling.

[0081] In some embodiments, the fluid-solid coupling numerical simulation model of CO2 flooding and oil storage in high-water-content fault-block reservoirs aims to predict and evaluate the oil recovery rate, CO2 storage rate and the possibility and impact of fault activation by simulating the changes in fluid pressure in the reservoir, the mechanical response of the rock, and the geometric and physical properties of the fault during CO2 injection.

[0082] In some embodiments, the slip state of the fault can be evaluated based on the Mohr-Coulomb yield criterion on the basis of considering the reservoir seepage-stress coupling equation, and then the permeability at the fault can be inferred.

[0083] In the present invention, the CO2 flooding and oil storage fluid-solid coupling numerical simulation model can be a model that integrates fluid mechanics and geomechanics to simulate the fault sliding state. Of course, the CO2 flooding and oil storage fluid-solid coupling numerical simulation model can also be other types of models, and the present invention does not specifically limit this.

[0084] Step S102, based on the Latin hypercube sampling method, sampling the injection well system as an input parameter, and calculating based on the CO2 flooding and oil storage fluid-solid coupling numerical simulation model to obtain output and fault slip displacement samples under different injection systems.

[0085] In some embodiments, Latin Hypercube Sampling (LHS) is a method for approximate random sampling from a multivariate parameter distribution, which is a stratified sampling technique. By uniformly sampling in each dimension, the correlation between input variables is reduced, thereby improving the accuracy and efficiency of model prediction.

[0086] In the present invention, the injection well system is sampled based on Latin hypercube sampling as an input parameter, and the CO2 flooding and oil storage fluid-solid coupling numerical simulation model established in the above step S101 is calculated, and the output result, i.e., the sample, is read.

[0087] Step S103, input the constructed samples into the long short-term memory neural network, and use the Adam optimization algorithm to train until the root mean square error meets the preset conditions to obtain the target proxy model.

[0088] In some embodiments, the proxy model is constructed based on a long short-term memory neural network.

[0089] Step S104, based on the target proxy model, a multi-objective optimization algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault slip displacement objective function to obtain the corresponding optimization result of each objective function.

[0090] In some embodiments, the multi-objective optimization algorithm may be an NSGA-II algorithm or other multi-objective optimization algorithms, which are not specifically limited in this application.

[0091] The method provided by the present invention first establishes a numerical simulation model of fluid-solid coupling of CO2 flooding and oil storage in high-water-content fault block reservoirs, and then performs sampling based on the Latin hypercube sampling method and calculates and generates samples based on the fluid-solid coupling model of CO2 flooding and oil storage, uses a long short-term memory neural network to build an agent model, introduces the NSGA-II multi-objective optimization algorithm to perform multi-objective optimization, and obtains the Pareto solution set of the objective function. In this way, on the one hand, the present invention, based on the comprehensive framework of the three-dimensional reservoir seepage and geomechanics coupling model, the artificial intelligence algorithm LSTM and the multi-objective optimization algorithm NSGA-II, can obtain the production system scheme of the injection and production wells through optimization, thereby maximizing the crude oil recovery rate and CO2 storage capacity and minimizing the fault sliding displacement.

[0092] In some embodiments, the above step S101 may include the following contents:

[0093] For any load-bearing surface considered, the linear relationship between the ultimate strength, shear stress and normal stress is shown in formula (1);

[0094]

[0095] In the formula, τ n is the shear stress on the failure surface, σ n is the normal stress of the failure surface, C is the cohesion of the rock; is the internal friction angle of rock.

[0096] In some embodiments, when a point in a material slips (shears) on a plane, the shear stress τ acting on the plane is n In addition to overcoming the inherent cohesive force C of the material, it is also necessary to overcome the normal stress σ acting on the surface. n The friction created.

[0097] The stress component acting on the fault plane is the normal stress σ' on the fault surface n , according to the shear stress σ on the fault plane t and σ s , the maximum shear stress τ can be determined max for

[0098] According to the maximum shear stress, the sliding displacement of the fault plane is determined; the sliding displacement is shown in the following formula:

[0099]

[0100] In the formula, Kt is the stiffness of the fault. When the maximum shear stress reaches a certain level, the grids on both sides of the fault will be displaced. This movement will affect the sliding friction of the grid blocks.

[0101] According to the slip displacement D s The value of is used to determine the friction coefficient of sliding friction, as shown in equations (3) to (5):

[0102]

[0103] In the formula, is the minimum friction angle. When the fault surface starts to slip, the sliding friction becomes smaller. The fault displacement from the beginning to the sliding displacement is less than D smax period, is the maximum value of fault friction, is the minimum value of fault friction.

[0104] The slip trend of the fault is determined according to the maximum shear stress and the friction coefficient; the slip trend is shown in the following formula:

[0105]

[0106] Where cohes is the fault cohesion.

[0107] like Figure 2 As shown in the figure, when the pore pressure increases, the Mohr circle moves to the left. When the Mohr circle touches the failure envelope, the material fails. When the Mohr circle touches the failure envelope, the slip trend of the fault is 1. For engineering safety considerations, a safety factor F less than 1 is introduced in actual reservoirs. s , the slip trend of the fault is greater than F s ×ST, the fault is activated.

[0108] In some embodiments, the above step S102 may include the following contents:

[0109] Latin hypercube sampling is applied to the sampling of long short-term memory neural network training set. This sampling method can improve the reliability of numerical simulation and the representativeness of samples. Assume a structure function function with K-dimensional basic variables, N represents the selection of N groups x = (x1, x2, ..., x n ), the matrix P is an N*K matrix, the column elements of P are randomly generated by integers from 1 to N, and the value of each element is not equal, the matrix R is also an N*K matrix, the column elements of R are randomly generated by rational numbers from 0 to 1, and the value of each element is not equal;

[0110]

[0111]

[0112] In the formula, s ij is the i-th row and j-th column element of the matrix S, is the distribution function that the jth variable conforms to, is the jth variable of the i-th group of samples obtained by Latin hypercube sampling.

[0113] In some embodiments, the above step S103 may include the following contents:

[0114] The long short-term memory neural network is used for training to define the weight and bias parameters of the gate; the weight and bias are shown in the following formula:

[0115]

[0116] Among them, W represents weight, b represents bias, i represents input gate, f represents forget gate, and o represents output gate.

[0117] like Figure 3 As shown, in a single hidden unit, the first part is the forget gate, which will choose to forget the information of the previous time step, which will allow LSTM to retain more important information. In a single hidden unit of the long short-term memory neural network, the Sigmoid activation function is used to determine the forget gate; the forget gate is shown as follows:

[0118] f t =σ(W f ·[h t-1 ,x t ]+b f ) (10),

[0119] In the formula, f t is the forget gate, σ represents the activation function, and the expression is σ(x)=(1+e -x ) -1 , whose output range is from 0 to 1, x t is the input of the current time step, h t-1 is the input of the previous time step, W f With b f is the weight and bias of the forget gate;

[0120] The update of the input gate is determined according to the Sigmoid activation function, and at the current step length, whether the current input is written by the information of the memory cell is determined based on the input gate; the input gate is shown in the following formula:

[0121]

[0122] In the formula, i t is the input gate, b i is the bias matrix of the input gate;

[0123]

[0124] In the formula, tanh represents a hyperbolic tangent function with an output range of -1 to 1;

[0125] According to the historical cell state, update the current cell state as shown in the following formula:

[0126]

[0127] In the formula, C t-1 is the historical cell state, C t is the current cell state, is the vector of new candidate values;

[0128] Output is made through the output gate, which controls whether the information stored in the memory cell affects the subsequent network layer;

[0129] o t =σ(W o [h t -1,x t ]+b o ) (14),

[0130] In the formula, o t is the output gate, W o With b o are the weight and bias of the output gate, h t It is the output value of the current cell state and also the input value of the next cell state.

[0131] In some embodiments, the above step S104 may include the following contents:

[0132] In the CO2 injection development of high water-content fault block reservoirs, the well control parameters of the gas injection wells and production wells have a final impact on the final recovery rate, storage rate and fault activation geological risk of the reservoir. Therefore, the well control operation parameters of the injection and production wells are used as inputs, and the recovery rate f1(x), storage rate f2(x) and fault slip displacement f3(x) are used as three objective functions for output. The problem is transformed into the maximization of f1(x) and f2(x) and the minimization of f3(x), which are defined as follows:

[0133] maxF(x)=[(f1(x),f2(x),-f3(x)) T ] (15),

[0134]

[0135]

[0136] f3=Ds (18)

[0137] Where OOIP is the initial recoverable reserves of the reservoir; is the oil production at time t; is the total amount of CO2 injected into the injection well at any given moment; is the total amount of CO2 produced by the production well at time t, T is the total time for reservoir development evaluation, and D s is the fault slip displacement;

[0138] In solving multi-objective optimization problems, the NSGA-II algorithm is used, which is evolved from the genetic algorithm. The genetic algorithm generally includes three basic operations: selection, crossover, and mutation. NSGA-II introduces non-dominated sorting algorithm, crowding operator fitness strategy, elite strategy and other methods on the basis of genetic algorithm for improvement. The algorithm flow is as follows:

[0139] Set the crowding distance of each individual to 0, and set the two individuals d0 and d l The crowding degree is positive infinity, and all other individuals i = 1, 2, ..., l-1 are allocated as follows:

[0140]

[0141] In the formula, f j i+1 is the jth objective function value at point i+1, f j i-1 Represents the j-th objective function value at point i-1.

[0142] In some embodiments, the complexity of the NSGA-II algorithm depends on the sorting method. When all individuals are in a non-dominated set, the computational complexity is the highest, which is mNlgN.

[0143] In some embodiments, to ensure that the non-dominated solution set is more evenly distributed, the definition of the crowding operator is based on i d with i rank These two properties, for two individuals i and j, if i rank <j rank , then individual i is better. If i rank =j rank , it is better to choose individuals with smaller crowding. Individuals with smaller crowding mean that there are relatively fewer individuals around them, and they need to be retained to maintain diversity.

[0144] The process of NSGA-II algorithm is as follows Figure 4As shown in the figure: First, initialize the parent population P0 to generate a child population Q0. Secondly, merge all individuals in the population and sort them according to the non-dominated relationship. Calculate the crowding degree of individuals in each non-dominated layer. Then, select individuals with better non-dominated order and crowding degree according to the crowding degree comparison operator to form a new parent population P1. Finally, generate the next generation population Q1 by selecting a series of operations such as crossover and mutation, and merge P1 and Q1 into a new population. Repeat the above operations until the requirements set by the program are met.

[0145] The NSGA-II algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function, and the fault sliding displacement objective function to obtain the corresponding optimization results.

[0146] The following describes an exemplary application of an embodiment of the present invention in a practical application scenario.

[0147] The purpose of the present invention is to provide a joint optimization method for high water-cut reservoirs based on long short-term memory neural networks. Aiming at the deficiencies in the research on CO2 flooding and storage in high water-cut reservoirs, the present invention establishes a fluid-solid coupling numerical simulation model for CO2 flooding and storage in high water-cut fault block reservoirs through CMG, uses Latin hypercube sampling and simulator calculation to generate samples, uses long short-term memory neural networks to construct a proxy model, introduces the NSGA-II multi-objective optimization algorithm to perform multi-objective optimization on the basis of the proxy model, and obtains a Pareto solution set containing three objectives.

[0148] This embodiment involves a multi-objective optimization method for oil displacement and storage in a high water-cut fault block reservoir based on a long short-term memory neural network. The application of this method in an example reservoir is as follows:

[0149] Based on the method of the aforementioned embodiment, a numerical simulation model of flow and geomechanics coupling is established using the CMG-GEM module, in which the reservoir is set as a dual-porosity and dual-permeability model, as shown in Figure 5, and the grids for geomechanics and flow simulations are the same. In order to simulate the influence of overlying rocks, the initial conditions of overlying stress are given. The model has 52×49×1=2548 grids, including 4 injection wells and 9 production wells, and the well type is five-point. The porosity of the model is shown in Figure 5(a) and the permeability is distributed in Figure 5(b). The reservoir and fluid parameters are shown in Table 1. In the early water flooding process, in order to maintain an injection-production ratio of 1:1, the production rate of the production well located in the center of the model is 160m 3 / day, the production well in the center of the boundary is 80m 3 / day, 40m at the corner 3 / day, the injection-production ratio is 1:1. In order to simulate the development effect of gas drive at different water drive stages during the development process, the present invention uses the model after water drive development as the restart model and inputs it into the gas drive. Since the first contact miscible pressure is 25MPa and the multi-stage contact miscible pressure is 18MPa, the injection of CO2 in this reservoir belongs to CO2 miscible drive.

[0150] Table 1

[0151]

[0152] In the present invention, a production dynamic prediction model based on LSTM is established. First, the model framework needs to be designed. The optimal framework consists of four hidden layers, one input layer and one output layer, as shown in the figure. When inputting, the sequence input is selected, first entering the first LSTM layer, and after obtaining the output, it is input to ReLU to extract relevant features. After repeating the LSTM layer and ReLU layer, it is passed to the fully connected layer and finally output. The specific structure of the model is as follows Figure 6 shown.

[0153] Figure 7 The results show that the LSTM proxy model has a fitting coefficient R when the moisture content is 0.7. 2 The value of performs well in the training set, and the sealing rate is related to the fault slip displacement R 2 The value is high, while in the test set, R 2 They are all greater than 0.7, indicating that the true value has a good correlation with the predicted value.

[0154] Depend on Figure 8 As shown in the figure, when the population size is 100 and the number of iterations is 100, the optimization results of the NSGA-Ⅱ algorithm are shown in the figure. All solutions are non-dominated solutions, and the distribution form of the Pareto solution in three-dimensional space is approximately "L" shaped. The statistical results of the objective function are shown in Figure 9. It is not difficult to see from the figure that the distribution patterns of the Pareto optimal concentrated solutions under different water cuts are similar regardless of the recovery factor, sealing factor or fault sliding displacement.

[0155] Aiming at the deficiencies in the research on CO2 flooding and storage in high-water-content oil reservoirs, the present invention establishes a fluid-solid numerical simulation model for CO2 flooding and storage in high-water-content fault block oil reservoirs through CMG, generates samples through Latin hypercube sampling and simulator calculation, constructs a proxy model based on the samples using a long short-term memory neural network, introduces the NSGA-Ⅱ multi-objective optimization algorithm to perform three-objective optimization on the basis of the proxy model, and obtains a Pareto solution set containing the three objectives.

[0156] Fig.10 is a schematic diagram of the composition structure of a high water-cut oil reservoir joint optimization device based on a long short-term memory neural network provided by an embodiment of the present invention, such as Fig.10As shown, a multi-objective optimization device 1000 for CO2 flooding and oil storage in high water-cut fault block reservoirs based on long short-term memory neural network includes: an establishment module 1001, used to establish a fluid-solid coupling numerical simulation model for CO2 flooding and oil storage in high water-cut fault block reservoirs; a calculation module 1002, used to sample and input the injection well system of the CO2 flooding and oil storage model based on the Latin hypercube sampling method, and calculate based on the fluid-solid coupling numerical simulation of CO2 flooding and oil storage to obtain samples; a training module 1003, used to input the constructed samples into the long short-term memory neural network, and use the Adam optimization algorithm for training until the root mean square error meets the preset conditions to obtain a proxy model; an optimization module 1004, used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault slip displacement objective function based on the target proxy model using a multi-objective optimization algorithm to obtain corresponding optimization results.

[0157] In some embodiments, the establishment module 1001 is also used to determine the linear relationship between the ultimate strength, shear stress and normal stress of each stress-bearing surface; the linear relationship is shown in the following formula:

[0158]

[0159] In the formula, τ n is the shear stress on the failure surface, σ n is the normal stress of the failure surface, C is the cohesion of the rock; is the internal friction angle of rock;

[0160] According to the shear stress σ of the fault plane t and σ s , determine the maximum shear stress τ max ; The maximum shear stress is shown in the following formula:

[0161]

[0162] According to the maximum shear stress, the sliding displacement of the fault plane is determined; the sliding displacement is shown in the following formula:

[0163]

[0164] In the formula, K t is the stiffness of the fault;

[0165] According to the value of the sliding displacement, the friction coefficient of the sliding friction is determined; the friction coefficient is shown in the following formula:

[0166]

[0167]

[0168]

[0169] In the formula, is the minimum friction angle, is the maximum value of fault friction, is the minimum value of fault friction;

[0170] The slip trend of the fault is determined according to the maximum shear stress and the friction coefficient; the slip trend is shown in the following formula:

[0171]

[0172] In the formula, cohes is the cohesion of the fault. When the pore pressure increases, the Mohr circle moves to the left.

[0173] In some embodiments, the calculation module 1002 is further used to assume a structure function function with K-dimensional basic variables, where N represents the selection of N groups x=(x1, x2, ..., x n ), the matrix P is an N*K matrix, the column elements of P are randomly generated by integers from 1 to N, and the value of each element is not equal, the matrix R is also an N*K matrix, the column elements of R are randomly generated by rational numbers from 0 to 1, and the value of each element is not equal;

[0174]

[0175]

[0176] In the formula, s ij is the i-th row and j-th column element of the matrix S, is the distribution function that the jth variable conforms to, is the jth variable of the i-th group of samples obtained by Latin hypercube sampling.

[0177] In some embodiments, the training module 1003 is further used to perform training using a long short-term memory neural network to define weight and bias parameters of the gate; the weight and bias are shown in the following formula:

[0178]

[0179] Among them, W represents weight, b represents bias, i represents input gate, f represents forget gate, and o represents output gate;

[0180] In a single hidden unit of the long short-term memory neural network, a Sigmoid activation function is used to determine a forget gate; the forget gate is shown in the following formula:

[0181] f t =σ(W f ·[h t-1 ,x t]+b f ),

[0182] In the formula, f t is the forget gate, σ represents the activation function, and the expression is σ(x)=(1+e -x ) -1 , whose output range is from 0 to 1, x t is the input of the current time step, h t-1 is the input of the previous time step, W f With b f is the weight and bias of the forget gate;

[0183] The update of the input gate is determined according to the Sigmoid activation function, and at the current step length, whether the current input is written by the information of the memory cell is determined based on the input gate; the input gate is shown in the following formula:

[0184] i t =σ(W i ·[h t-1 ,x t ]+b i ),

[0185] In the formula, i t is the input gate, b i is the bias matrix of the input gate;

[0186]

[0187] In the formula, tanh represents a hyperbolic tangent function with an output range of -1 to 1;

[0188] According to the historical cell state, update the current cell state as shown in the following formula:

[0189]

[0190] In the formula, C t-1 is the historical cell state, C t is the current cell state, is the vector of new candidate values;

[0191] Output is made through the output gate, which controls whether the information stored in the memory cell affects the subsequent network layer;

[0192] o t =σ(W o [h t -1,x t ]+b o ),

[0193] In the formula, o t is the output gate, W o With bo are the weight and bias of the output gate, h t It is the output value of the current cell state and also the input value of the next cell state.

[0194] In some embodiments, the optimization module 1004 is further used to take the production system parameters of the injection-production well as independent variables, and take the recovery rate, CO2 storage rate and fault sliding displacement as dependent variables, to maximize the recovery rate objective function, the CO2 storage rate objective function and minimize the fault sliding displacement objective function; the definitions are as follows:

[0195] maxF(x)=[(f1(x),f2(x),-f3(x)) T ],

[0196]

[0197] Where OOIP is the initial recoverable reserves of the reservoir; is the oil production at time t; is the total amount of CO2 injected into the injection well at time t; is the total amount of CO2 produced by the production well at time t, T is the total time for reservoir development evaluation, and D s is the fault slip displacement;

[0198] The NSGA-Ⅱ algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault sliding displacement objective function to obtain the corresponding optimization results.

[0199] It should be noted that the description of the device of the embodiment of the present invention is similar to the description of the above method embodiment, and has similar beneficial effects as the same method embodiment, so it will not be repeated. For technical details not disclosed in the embodiment of the device, please refer to the description of the method embodiment of the present invention for understanding.

[0200] It should be noted that, in the embodiment of the present invention, if the above-mentioned multi-objective optimization method for CO2 flooding and oil storage in high-water-content fault block reservoirs is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a terminal to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM, Read Only Memory), a magnetic disk or an optical disk. In this way, the embodiment of the present invention is not limited to any specific combination of hardware and software.

[0201] Correspondingly, an embodiment of the present invention provides a multi-objective optimization device for CO2 flooding and storage in a high water-cut fault block reservoir. Fig.11 Schematic diagram of the structure of the high water content reservoir joint optimization device provided by the embodiment of the present invention. Fig.11 As shown, the high water-cut reservoir joint optimization device 1100 at least includes: a processor 1101 and a computer-readable storage medium 1102 configured to store executable instructions, wherein the processor 1101 generally controls the overall operation of the high water-cut reservoir joint optimization device. The computer-readable storage medium 1102 is configured to store instructions and applications executable by the processor 1101, and can also cache data to be processed or processed by each module in the processor 1101 and the high water-cut fault block reservoir CO2 flooding and storage multi-objective optimization device 1100, which can be implemented through a flash memory (FLASH) or a random access memory (RAM, Random Access Memory).

[0202] An embodiment of the present invention provides a storage medium storing executable instructions, wherein the executable instructions are stored. When the executable instructions are executed by a processor, the processor will be caused to execute the method provided by the embodiment of the present invention, for example, Figure 1 The method shown.

[0203] In some embodiments, the storage medium can be a computer-readable storage medium, for example, a ferroelectric random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disk, or a compact disk read-only memory (CD-ROM), etc.; it can also be various devices including one or any combination of the above memories.

[0204] In some embodiments, executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0205] As an example, executable instructions may, but need not necessarily, correspond to a file in a file system, may be stored as part of a file storing other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions). As an example, executable instructions may be deployed to be executed on one electronic device, or on multiple electronic devices located at one location, or on multiple electronic devices distributed at multiple locations and interconnected by a communication network.

[0206] The above description is only an embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement and improvement made within the spirit and scope of the present invention are included in the protection scope of the present invention.

[0207] It should be understood that "one embodiment" or "an embodiment" mentioned throughout the specification means that specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in various embodiments of the present invention, the size of the serial number of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention. The serial numbers of the above-mentioned embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments.

[0208] It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of further restrictions, an element defined by the statement "comprises one..." does not exclude the presence of other identical elements in the process, method, article or device including the element. In several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed.

[0209] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A multi-objective optimization method for CO2 flooding and storage in high water-cut fault-block reservoirs, characterized in that: The method comprises: Establish a fluid-solid coupling numerical simulation model for CO2 flooding and storage in high water-cut fault block reservoirs; Based on the Latin hypercube sampling method, the injection well system is sampled as an input parameter, and the CO2 flooding and oil storage fluid-solid coupling numerical simulation model is used for calculation to obtain the output and fault slip displacement samples under different injection systems; The constructed samples are input into the long short-term memory neural network and trained using the Adam optimization algorithm until the root mean square error meets the preset conditions to obtain the target proxy model; Based on the target proxy model, a multi-objective optimization algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault slip displacement objective function to obtain the corresponding optimization result of each objective function.

2. The method according to claim 1, characterized in that The establishment of a fluid-solid coupling numerical simulation model for CO2 flooding and oil storage in high water-cut fault block reservoirs includes: Determine the linear relationship between the ultimate strength, shear stress and normal stress of each load-bearing surface; the linear relationship is shown in the following formula; In the formula, τ n is the shear stress on the failure surface, σ n is the normal stress of the failure surface, C is the cohesion of the rock; is the internal friction angle of rock; According to the shear stress σ of the fault plane t and σ s , determine the maximum shear stress τ max ; The maximum shear stress is shown in the following formula: According to the maximum shear stress, the sliding displacement of the fault plane is determined; the sliding displacement is shown in the following formula: In the formula, K t is the stiffness of the fault; According to the value of the sliding displacement, the friction coefficient of the sliding friction is determined; the friction coefficient is shown in the following formula: In the formula, is the minimum friction angle, is the maximum value of fault friction, is the minimum value of fault friction; The slip trend of the fault is determined according to the maximum shear stress and the friction coefficient; the slip trend is shown in the following formula: In the formula, cohes is the cohesion of the fault. When the pore pressure increases, the Mohr circle moves to the left.

3. The method according to claim 1, characterized in that Based on the Latin hypercube sampling method, the injection well system is sampled and input as an input parameter, and the CO2 flooding and oil storage fluid-solid coupling numerical simulation model is used for calculation to obtain the output and fault slip displacement samples under different injection systems, including: Assume a structure function with K-dimensional basic variables, N represents the selection of N groups x = (x1, x2, ..., x n ), the matrix P is an N*K matrix, the column elements of P are randomly generated by integers from 1 to N, and the value of each element is not equal, the matrix R is also an N*K matrix, the column elements of R are randomly generated by rational numbers from 0 to 1, and the value of each element is not equal; In the formula, s ij is the i-th row and j-th column element of the matrix S, is the distribution function that the jth variable conforms to, is the jth variable of the i-th group of samples obtained by Latin hypercube sampling.

4. The method according to claim 1, characterized in that: The constructed samples are input into the long short-term memory neural network and trained using the Adam optimization algorithm until the root mean square error meets the preset conditions to obtain the target proxy model, including: The long short-term memory neural network is used for training to define the weight and bias parameters of the gate; the weight and bias are shown in the following formula: Among them, W represents weight, b represents bias, i represents input gate, f represents forget gate, and o represents output gate; In a single hidden unit of the long short-term memory neural network, a Sigmoid activation function is used to determine a forget gate; the forget gate is shown in the following formula: f t =σ(W f ·[h t-1 ,x t ]+b f ), In the formula, f t is the forget gate, σ represents the activation function, and the expression is σ(x)=(1+e -x ) -1 , whose output range is from 0 to 1, x t is the input of the current time step, h t-1 is the input of the previous time step, W f With b f is the weight and bias of the forget gate; The update of the input gate is determined according to the Sigmoid activation function, and at the current step length, whether the current input is written by the information of the memory cell is determined based on the input gate; the input gate is shown in the following formula: i t =σ(W i ·[h t-1 ,x t ]+b i ), In the formula, i t is the input gate, b i is the bias matrix of the input gate; In the formula, tanh represents a hyperbolic tangent function with an output range of -1 to 1; According to the historical cell state, update the current cell state as shown in the following formula: In the formula, C t-1 is the historical cell state, C t is the current cell state, is the vector of new candidate values; Output is made through the output gate, which controls whether the information stored in the memory cell affects the subsequent network layer; the t =σ(W o [h t -1.x t ]+b o ), In the formula, o t is the output gate, W o With b o are the weight and bias of the output gate, h t It is the output value of the current cell state and also the input value of the next cell state.

5. The method according to claim 1, characterized in that Based on the proxy model, a multi-objective optimization algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault slip displacement objective function, and the corresponding optimization results of each objective function are obtained, including: Taking the production system parameters of the injection-production well as independent variables, and the recovery rate, CO2 storage rate and fault sliding displacement as dependent variables, in order to maximize the recovery rate objective function and the CO2 storage rate objective function and minimize the fault sliding displacement objective function; define the following formula: max F(x)=[(f1(x),f2(x),-f3(x)) T ], <h2 style=";text-align:left;direction:ltr"> <h2 style=";text-align:left;direction:ltr"> f3=D<h2 style=";text-align:left;direction:ltr"> s <h2 style=";text-align:left;direction:ltr"> , Where OOIP is the initial recoverable reserves of the reservoir; is the oil production at time t; is the total amount of CO2 injected into the injection well at time t; is the total amount of CO2 produced by the production well at time t, T is the total time for reservoir development evaluation, and D s is the fault slip displacement; The NSGA-Ⅱ algorithm is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault sliding displacement objective function to obtain the corresponding optimization results.

6. A multi-objective optimization device for CO2 flooding and storage in high water-cut fault block reservoirs based on long short-term memory neural network, characterized in that: The device comprises: Establish a module for establishing a fluid-solid coupling numerical simulation model for CO2 flooding and storage in high water-cut fault block reservoirs; A calculation module is used to sample the injection well system based on the Latin hypercube sampling method as an input parameter, and to calculate based on the CO2 flooding and oil storage fluid-solid coupling numerical simulation model to obtain output and fault slip displacement samples under different injection systems; The training module is used to input the constructed samples into the long short-term memory neural network and use the Adam optimization algorithm for training until the root mean square error meets the preset conditions to obtain the target proxy model; The optimization module is used to optimize the recovery factor objective function, the CO2 storage rate objective function and the fault sliding displacement objective function based on the objective proxy model using a multi-objective optimization algorithm to obtain the corresponding optimization result of each objective function.

7. An electronic device, characterized in that: include: A memory for storing executable instructions; The processor is used to implement the multi-objective optimization method for CO2 flooding and storage in high water-cut fault block reservoirs as described in any one of claims 1 to 5 when executing the executable instructions stored in the memory.

8. A computer-readable storage medium storing executable instructions for causing a processor to execute the executable instructions to implement the multi-objective optimization method for CO2 flooding and storage in a high-water-cut fault-block reservoir as described in any one of claims 1 to 5.

Images