A method for natural gas hydrate reservoir layered balanced production regulation based on analog optimization

Abstract

The application discloses a natural gas hydrate reservoir layering balanced exploitation regulation and control method based on simulation optimization, S1: constructing a three-dimensional geological model according to well logging, seismic and core data of a target area; S2: establishing a heat-flow-force-chemical multi-field coupling numerical model and initializing; S3: defining a multi-dimensional engineering decision parameter space; S4: establishing a comprehensive objective function; S5: obtaining an optimal engineering decision parameter combination X_optimal through iterative search; S6: implementing drilling and well completion operations according to X_optimal, and dynamically verifying and feedback adjusting the model and exploitation system based on real-time monitoring data in the production process. The heat-flow-force-chemical multi-field coupling numerical model is established based on the three-dimensional geological model, a proxy model and multi-objective optimization are introduced, and dynamic verification and feedback adjustment are conducted in combination with monitoring data in the production period, so that interlayer contradictions are effectively inhibited, balanced utilization of reservoir energy is realized, and the overall recovery rate and development economy of the multi-layer hydrate reservoir are improved.

Description

Technical Field

[0001] This invention relates to the field of oil and gas field development technology, and in particular to a method for controlling the layered and balanced exploitation of natural gas hydrate reservoirs based on simulation optimization. Background Technology

[0002] Natural gas hydrates are a potential clean energy source found in marine or permafrost regions. Their depressurization extraction process involves complex physical processes such as hydrate decomposition and phase transition, multiphase flow, heat transfer, and sediment mechanical responses. In multi-layered natural gas hydrate reservoirs, due to the significant heterogeneity in porosity, permeability, net-to-gross ratio, and their vertical distribution, traditional single depressurization regimes or indiscriminate overall extraction methods often result in preferential depressurization and rapid gas production in high-permeability or well-connected sections, while low-permeability layers are difficult to effectively utilize. This leads to severe inter-layer interference, low overall recovery rate, high water production, and increased reservoir stability risks.

[0003] In existing technologies, most methods focus on numerical simulation analysis of single-layer hydrate reservoirs, or only perform one-time scheme optimization before mining. There is a lack of a systematic method that can combine real-time monitoring data during production to continuously revise the model of multi-layer hydrate reservoirs and implement stratified dynamic control. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a simulation-optimized method for the balanced exploitation and control of natural gas hydrate reservoirs.

[0005] The objective of this invention is achieved through the following technical solution: a method for regulating the layered and balanced exploitation of natural gas hydrate reservoirs based on simulation optimization, comprising the following steps:

[0006] S1: Construct a three-dimensional geological model based on well logging, seismic, and core data of the target area;

[0007] S2: Establish a multi-field coupled numerical model of thermo-fluid-mechanical-chemical processes based on a three-dimensional geological model, and initialize it;

[0008] S3: Define a multi-dimensional engineering decision parameter space;

[0009] S4: Establish the comprehensive objective function;

[0010] S5: Obtain the optimal combination of engineering decision parameters X_optimal through iterative search;

[0011] S6: Drilling and completion operations are carried out based on X_optimal, and the model and mining system are dynamically verified and adjusted based on real-time monitoring data during the production process.

[0012] Preferably, in step S4, a comprehensive objective function is established based on maximizing the total cumulative gas production, the inter-layer gas production balance, and economic benefit indicators.

[0013] ;

[0014] in, , and For weight parameters, For interlayer gas production equilibrium, The normalized total cumulative gas production The normalized net present value;

[0015] Normalized total cumulative gas production The calculation formula is:

[0016] ;

[0017] in, To simulate the actual total cumulative gas production, The theoretical recoverable quantity estimated based on geological reserves;

[0018] Interlayer gas production equilibrium The calculation formula is:

[0019] ;

[0020] in, Gas production at each level standard deviation Gas production at each level The arithmetic mean, No. The cumulative gas production of a hydrate reservoir over the entire extraction cycle = 1, 2, ..., N, where N is the total number of productive layers;

[0021] Normalized net present value The calculation formula is:

[0022] ;

[0023] in, For time, Cash inflow in year t Cash outflow in year t is the discount rate.

[0024] Preferably, step S5 further includes the following step:

[0025] S51: Generate an initial parameter sample set, call a multi-field coupled numerical model for each parameter sample to obtain the corresponding objective function value, and build a sample database;

[0026] S52: The parameter combination of each sample point is used to obtain the corresponding objective function value through a multi-field coupled numerical model, forming a sample database of parameter combination-simulation results;

[0027] S53: The simulation results of the sample set are used to train a proxy model through machine learning algorithms;

[0028] S54: Using the surrogate model as a fast evaluator, obtain the optimal combination of engineering decision parameters X_optimal.

[0029] Preferably, in step S51, an orthogonal experimental design method is used to extract a certain number of sample points in the engineering decision parameter space to generate an initial parameter sample set.

[0030] Preferably, in step S6, the real-time monitoring data is fed back to the multi-field coupled numerical model, and the uncertainty parameters in the model are inverted and corrected so that the model always maintains an accurate mapping of the underground situation.

[0031] The present invention has the following advantages: The present invention establishes a multi-field coupled numerical model of thermo-fluid-mechanical-chemical based on a three-dimensional geological model, and introduces a surrogate model and multi-objective optimization. Combined with monitoring data during production, dynamic verification and feedback adjustment are carried out, which effectively suppresses inter-layer contradictions, realizes the balanced utilization of reservoir energy, and improves the overall recovery rate and development economy of multi-layer hydrate reservoirs. Attached Figure Description

[0032] Figure 1 A schematic diagram of the process for controlling the layered and balanced exploitation of natural gas hydrate reservoirs;

[0033] Figure 2 This is a schematic diagram of the vertical heterogeneous distribution of net-to-gross ratio in a multi-layered hydrate reservoir geological model.

[0034] Figure 3 This is a schematic diagram of the vertical heterogeneous distribution of permeability in a geological model of a multi-layered hydrate reservoir.

[0035] Figure 4 This is a schematic diagram of the vertical heterogeneous distribution of porosity in a multi-layered hydrate reservoir geological model.

[0036] Figure 5 A schematic diagram showing the deployment of different horizontal well sections at vertical locations (bottom, middle, and top) in the reservoir;

[0037] Figure 6 A schematic diagram comparing the gas production of each layer before optimization;

[0038] Figure 7 A schematic diagram showing the comparison of gas production at each layer after optimization;

[0039] Figure 8 This is a schematic diagram of the process of a hybrid intelligent optimization algorithm based on an agent model. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0041] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0042] It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined with each other.

[0043] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0044] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use, or the orientation or positional relationship commonly understood by those skilled in the art. They are only used for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. In addition, the terms "first," "second," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0045] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0046] In this embodiment, as Figure 1 As shown, a simulation-optimized method for controlling the stratified and balanced exploitation of natural gas hydrate reservoirs includes the following steps:

[0047] S1: Construct a three-dimensional geological model based on well logging, seismic, and core data of the target area; specifically, step S1 is the foundation of this application, requiring the integration of high-resolution well logging curves, three-dimensional seismic inversion data, and core experimental data of the target block to construct a three-dimensional geological grid model that can realistically reflect the heterogeneity of the reservoir. For example... Figure 2 As shown, the model characterizes the vertical fluctuation of net gross to density (NTG) with depth in multi-layered hydrate reservoirs, with a mean value around 0.55; Figure 3 As shown, the permeability field exhibits significant heterogeneity, with an average permeability of 5 mD in the hydrate layer and differences of 1-2 orders of magnitude between layers. The coefficient of variation for permeability is set between 0.05 and 0.08. Figure 4 As shown, the porosity field exhibits a discontinuous distribution in the vertical direction, with a mean of approximately 0.55. These geological parameters together form the static basis for subsequent thermo-fluid-mechanical-chemical multi-field coupled simulations.

[0048] S2: A coupled numerical model of thermo-fluid-mechanical-chemical fields is established and initialized based on a three-dimensional geological model. Specifically, a grid system and non-uniform characterization are constructed first. The three-dimensional geological model from step S1 is discretized into a numerical simulation grid. To address the heterogeneity of hydrate reservoirs, Kriging interpolation is used to map static parameters such as net-to-gross ratio, porosity, permeability, and phase saturation to each grid cell. It is important to note that local grid refinement is performed in the peri-well region to accurately capture the high-pressure gradient and drastic changes in the thermal flow field at the hydrate decomposition front. Secondly, a composition and phase model is established. Specifically, five key components are defined in CMG-STARS: methane, water, hydrates, ice, and salinity. The Kim-Bishnoi model is introduced to describe the decomposition and synthesis of hydrates, setting decomposition kinetic constants and activation energies. An equilibrium pressure curve is established, and the phase transition rate of hydrates is calculated by judging the degree to which the pressure and temperature within the grid deviate from the equilibrium conditions. Next, thermal coupling is performed, which involves setting the thermophysical parameters of the formation, its caprock, and the underlying rocks, including heat conduction and convection: defining the specific heat capacity and thermal conductivity of the rock matrix and fluids; latent heat of phase change: specifying the exact values ​​of the heat absorbed by hydrate decomposition and the heat released by the water-ice phase transition. Then, mechanical coupling is performed, establishing the stress-strain relationship based on porous media mechanics theory, specifically parameter correlation: defining dynamic functions of porosity and permeability as a function of effective stress, such as the Palmer-Mansoori model; mechanical boundaries: setting Young's modulus, Poisson's ratio, and initial geopressure gradient (vertical and horizontal stress) to ensure the simulator can calculate in real time the compression of the reservoir framework and surface subsidence caused by hydrate decomposition. Before the simulation begins, a "balance calculation" is performed, which assigns values ​​based on the geothermal gradient and hydrostatic pressure gradient, and determines the initial spatial distribution of hydrates, gas, and water based on the gas-water interface and capillary pressure curves, ensuring the model is in a state of gravity-thermodynamic equilibrium when no disturbance is applied (no extraction). In this embodiment, the model used for the thermo-fluid-mechanical-chemical multi-field coupling numerical model is CMG STARS (version 2023.10).

[0049] S3: Define a multi-dimensional engineering decision parameter space; specifically, key engineering factors affecting mining performance are grouped into an engineering decision parameter set, which includes discrete and continuous variables, forming a high-dimensional optimization space. First, various engineering factors are transformed into mathematical expressions that can be processed by the optimization algorithm. Specifically: for stratified pressure drop gradients, pressure stabilization time, and throttle valve opening, their physically feasible upper and lower limits are set; for well type selection (vertical well = 0, horizontal well = 1, multi-branch well = 2) and the vertical position of the horizontal section (top = 0, middle = 1, bottom = 2), integer encoding is used to map them to discrete coordinate points in the high-dimensional space; for the selection of the vertical position of the horizontal well section, the specific decision logic loop is as follows: for reservoirs with positive rhythm (improving upwards), the candidate position is constrained to the upper region of the reservoir to utilize the high-permeability layer at the top for rapid pressure reduction and diffusion; for reservoirs with inverse rhythm (deteriorating upwards) or homogeneous permeability, combined with water avoidance and production stability requirements, the horizontal well section is optimized to be located in the middle or lower part of the reservoir, such as... Figure 5 As shown, the specific variables cover the vertical position of the horizontal well section within the target layer, and the specific decision logic is: for reservoirs where permeability exhibits a positive rhythm (such as... Figure 3 (As shown in the upper high-permeability zone), the horizontal well section is anchored in the top 1 / 4 of the reservoir to utilize gravity to promote gas phase confluence; for reservoirs with bottom water risk or homogeneous reservoirs, through... Figure 5 The simulation comparison shown indicates that horizontal wells are deployed in the middle or lower part of the reservoir (at least 10 meters away from the bottom water interface) to balance the oil drainage area and water avoidance requirements.

[0050] S4: Establish the comprehensive objective function and constraints; further, in step S4, a comprehensive objective function is established based on maximizing the total cumulative gas production, the inter-layer gas production balance, and economic benefit indicators.

[0051] ;

[0052] in, , and For weight parameters, For interlayer gas production equilibrium, The normalized total cumulative gas production The normalized net present value;

[0053] Normalized total cumulative gas production The calculation formula is:

[0054] ;

[0055] in, To simulate the actual total cumulative gas production, The theoretical recoverable reserves are estimated based on geological reserves, with the aim of making... The range of its value is usually within the interval [0, 1]; specifically, This refers to the total cumulative gas production over the entire production cycle, obtained by simulating a given combination of engineering decision parameters X (including bottom hole pressure, branch well length, perforation density, etc.) using an established thermo-fluid-mechanical-chemical multi-field coupled numerical model. The calculation method involves the simulator calculating the gas production of each grid cell in real time at each time step based on the mass conservation equation, energy conservation equation, hydrate decomposition kinetic equation, and multiphase flow equation, and then summing the results to output the final value. . The baseline values ​​are calculated based on the geological parameters of the target reservoir, and are derived from actual field logging data: resistivity logging, sonic logging, density logging, etc., used to calculate porosity and hydrate saturation; seismic data: 3D seismic inversion, used to delineate gas-bearing area and reservoir distribution; core experiments: directly measuring porosity, permeability and hydrate saturation, used to calibrate logging interpretation results, and the calculation formula is the volumetric method for calculating reserves recognized in the field of oil and gas reservoir engineering.

[0056] Interlayer gas production equilibrium The calculation formula is:

[0057] ;

[0058] in, Gas production at each level The standard deviation measures the magnitude of fluctuations in gas production across different layers. Gas production at each level The arithmetic mean, No. The cumulative gas production of a hydrate reservoir over the entire extraction cycle =1, 2, ..., N, where N is the total number of productive layers; specifically... and Based on the cumulative gas production of each layer The calculated statistics are used in the optimization design phase (step S5). Derived from the layered output results of numerical simulators such as CMG STARS, the simulator calculates the gas production of each grid cell based on multi-field coupling equations and accumulates them layer by layer, directly outputting the gas production of each layer. During the production implementation phase (step S6). Real-time downhole monitoring was used to obtain data, and production logging tools were used to measure the gas production contribution rate of each layer. Combined with the total gas production recorded by the wellhead flow meter ,calculate .

[0059] Normalized net present value The calculation formula is:

[0060] ;

[0061] in, For time, Cash inflow in year t (mainly natural gas sales revenue). Cash outflows in year t (including drilling investment, well completion costs, operation and maintenance fees, etc.) This is the discount rate (reflecting the cost of capital and risk). Specifically, The closer the value is to 1, the more balanced the gas production across layers, maximizing the gas production balance across hydrate zones. This balance is characterized by the ratio of the standard deviation to the mean of the cumulative gas production in each layer. Economic benefits are quantified using net present value or payback period. In this embodiment, constraints include minimizing cumulative water production or gas-water ratio constraints, reservoir subsidence or shear strain thresholds, pore pressure gradient constraints, and lower temperature limits or secondary hydrate regeneration risk indicators.

[0062] S5: Obtain the optimal combination of engineering decision parameters X_optimal through iterative search. Further, such as... Figure 8 As shown, step S5 also includes the following steps:

[0063] S51: Generate an initial parameter sample set, call a multi-field coupled numerical model for each parameter sample to obtain the corresponding objective function value, and construct a sample database; preferably, use an orthogonal experimental design method to extract a certain number of sample points in the engineering decision parameter space to generate an initial parameter sample set.

[0064] S52: The parameter combination of each sample point is used to obtain the corresponding objective function value through a multi-field coupled numerical model, forming a sample database of parameter combination-simulation results;

[0065] S53: The simulation results of the sample set are used to train an SVM surrogate model using a machine learning algorithm; specifically, this also includes the following steps:

[0066] S53.1: Using Latin hypercube sampling (LHS), generate N uniformly distributed candidate schemes within the parameter space defined in step S3, and call the multi-field coupling simulator in step S2 to perform batch calculations to obtain the gas production and equilibrium label corresponding to each scheme.

[0067] S53.2: Perform Min-Max mapping on the input features to normalize the parameter range to [0, 1], and use the Pearson correlation coefficient to remove redundant features;

[0068] S53.3: Support vector machine is used as the surrogate model architecture, mean squared error is used as the loss function, and the network weights are adjusted through backpropagation algorithm to fit the nonlinear mapping relationship between "engineering decision-making and mining effect".

[0069] S53.4: Introduce the test set to calculate the model's coefficient of determination R. 2 When R 2 When the error response surface is less than 0.9, the region with the large prediction deviation is located by using the error response surface, and local dense sampling is performed to supplement the sample until the model accuracy reaches the target.

[0070] S54: Using the surrogate model as a fast evaluator, obtain the optimal combination of engineering decision parameters X_optimal. Specifically, this also includes the following steps:

[0071] S54.1: Initialize the particle swarm P (of size N) of MOPSO p Each particle's position represents a set of production parameters. Velocity is randomly generated, and an empty SPEA2 external elite file A is initialized to store previously found non-dominated solutions. Specifically, production parameters include bottom hole flowing pressure / layer pressure drop regime parameters, well structure parameters, completion parameters, and production control parameters. The bottom hole flowing pressure / layer pressure drop regime parameters include the minimum bottom hole flowing pressure, pressure drop gradient, and stabilization time for each hydrate layer. If layered pressure drop is used, each layer can have its own independently set pressure drop curve. Well structure parameters include well type selection: discrete variables (vertical well, single-branch horizontal well, multi-branch well), horizontal section length, and the vertical position of the horizontal section within the target layer: discrete variables (top / middle / bottom). Completion parameters include the perforation start and end depths and perforation density for each hydrate layer. Production control parameters include the choke valve opening for each layer or the production allocation ratio for each layer.

[0072] S54.2: Perform iterative optimization; specifically including fitness evaluation, SPEA2 fitness allocation, updating the adaptation database, and MOPSO particle update. Fitness evaluation is performed by quickly assessing the current population P using a trained SVM surrogate model. t and external files A t All objective function values ​​for all individuals in the set P; t ∪A t Each individual in Calculate its strength value That is, the number of individuals it dominates. The initial fitness of each individual is calculated as R(i) = Σ_{j dominates i}S(j), to further distinguish individuals with the same... For each individual, a density estimate is introduced (usually using the reciprocal of the distance to the k-th nearest neighbor), and the final fitness is... , Smaller values ​​indicate better individual performance (non-dominant and dispersed). The current non-dominant individuals are copied to the next-generation adaptation database set A{t+1}. If the size of A{t+1} exceeds a preset value, then based on density information (i.e., ... To maintain the file size, individuals in crowded areas are truncated first; if the A{t+1} size is insufficient, individuals are truncated from the dominated individuals. Selectively supplement. For each particle... From its historical best position pBest i In the elite archive A{t+1}, the global leader gBest, selected through tournament or roulette, updates particle velocity and position:

[0073] ; ;

[0074] in, The first in the population One particle, For the current optimization iteration round, For particles In the The position vector of the substitute. For particles In the The velocity vector of the vector, where ω is the inertial weight. For individual learning factors, As a social learning factor, A random number within the range [0,1] For particles The optimal position, For external elite files The middle is a particle The globally optimal boot location is selected. For particles exist The new position of the generation. When the iteration reaches the maximum number of generations, or when the Pareto front of the outer archive A no longer improves significantly over several consecutive generations, the loop terminates and the final outer archive is output as the approximate Pareto optimal solution set.

[0075] S54.3: Construct a decision matrix, where rows represent schemes in the Pareto solution set and columns represent objective values. Normalize the objective values ​​and assign weights to each objective based on preferences (e.g., total gas production weight 0.5, equilibrium weight 0.3, economic weight 0.2). Calculate the Euclidean distance between each scheme and the positive ideal solution (composed of the optimal values ​​of each objective) and the negative ideal solution (composed of the worst values ​​of each objective). and ;

[0076] S54.4: Calculate the relative closeness of each option.

[0077] ;

[0078] choose The scheme with the largest value is taken as the best compromise mining scheme, that is, its content is the optimal combination of engineering decision parameters X_optimal.

[0079] S6: Drilling and completion operations are carried out based on X_optimal, and the model and production regime are dynamically verified and adjusted based on real-time monitoring data during the production process. The specific process of dynamic verification and feedback adjustment is as follows:

[0080] Initial stage: The X_optimal value obtained from step S5 is put into production, with an initial equilibrium degree E = 0.52 (e.g., ...). Figure 6 (as shown)

[0081] Day 30: Monitoring data shows that the gas production share of Layer 1 (high permeability layer) has increased to 70%, while the share of Layer 2 (low permeability layer) has decreased to 30%. The actual equilibrium E_actual=0.48, which is lower than expected.

[0082] Inversion correction: Inputting the monitoring data into the model for inversion revealed that the actual permeability of layer 1 was 20% higher than that of the geological model, while the permeability of layer 2 was 15% lower.

[0083] Rapid optimization: Based on the updated model, a new control scheme was obtained: the bottom flow pressure of layer 1 was increased from 4.5 MPa to 5.2 MPa (to limit gas production); the bottom flow pressure of layer 2 was decreased from 4.5 MPa to 3.8 MPa (to enhance pressure reduction); the total gas production allocation ratio was adjusted to 0.55:0.45.

[0084] Execution of control: The above adjustments are performed via the downhole throttle valve;

[0085] Day 60: Monitoring data shows that the gas production share of layer 1 decreased to 58%, while that of layer 2 increased to 42%, and the equilibrium index E improved to 0.91 (e.g., Figure 7As shown in the diagram, the desired effect was achieved. Specifically, through distributed fiber optic temperature / pressure measurement systems and production logging tools deployed downhole, the gas production profile and pressure recovery data of each layer were monitored in real time. Real-time monitoring data was fed back to a multi-field coupled numerical model, and the uncertain parameters in the model were inverted and corrected to ensure the model always maintained an accurate mapping of the subsurface conditions. Based on the updated model, rapid optimization cycles could be initiated periodically or in real-time to fine-tune the bottom hole flowing pressure or inter-layer production distribution, forming a closed loop of "monitoring-optimization-control" to cope with dynamic changes in the reservoir during production. This invention establishes a thermo-fluid-mechanical-chemical multi-field coupled numerical model based on a three-dimensional geological model, and introduces a surrogate model and multi-objective optimization. Combined with monitoring data during production, dynamic verification and feedback adjustments were performed, effectively suppressing inter-layer contradictions, achieving balanced utilization of reservoir energy, and improving the overall recovery rate and development economy of multi-layer hydrate reservoirs. Figure 6 As shown, in the traditional scheme before optimization, due to the difference in permeability of each layer (reference...), Figure 3 Data shows that high-permeability layers (such as the first layer) dominate gas production, resulting in an interlayer gas production equilibrium E of only about 0.45, and the utilization of low-permeability layers is extremely low; for example... Figure 7 As shown, by executing the optimal decision combination X_optimal searched in step S5 (such as increasing the pressure drop gradient for low-permeability layers and implementing throttling control for high-permeability layers), the gas production of each layer tends to be balanced after optimization, and the standard deviation is [missing information]. The pressure drop was significant, and the equilibrium E increased to over 0.85. This balanced utilization effectively avoided the risk of reservoir subsidence caused by excessive local pressure drop, and achieved an overall improvement in total gas production and recovery rate.

[0086] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for controlling the layered and balanced exploitation of natural gas hydrate reservoirs based on simulation optimization, characterized in that: Includes the following steps: S1: Construct a three-dimensional geological model based on well logging, seismic, and core data of the target area; S2: Based on the aforementioned three-dimensional geological model, establish a thermo-fluid-mechanical-chemical multi-field coupled numerical model and initialize it; S3: Define a multi-dimensional engineering decision parameter space; S4: Establish the comprehensive objective function; S5: Obtain the optimal combination of engineering decision parameters X_optimal through iterative search; S6: Drilling and completion operations are carried out based on X_optimal, and the model and mining system are dynamically verified and adjusted based on real-time monitoring data during the production process; In step S4, a comprehensive objective function is established based on maximizing the total cumulative gas production, the inter-layer gas production balance, and economic benefit indicators. ； in, , and For weight parameters, For interlayer gas production equilibrium, The normalized total cumulative gas production The normalized net present value; Normalized total cumulative gas production The calculation formula is: ； in, To simulate the actual total cumulative gas production, The theoretical recoverable quantity estimated based on geological reserves; Interlayer gas production equilibrium The calculation formula is: ； in, Gas production at each level standard deviation Gas production at each level The arithmetic mean, For the first The cumulative gas production of a hydrate reservoir over the entire extraction cycle = 1, 2, ..., N, where N is the total number of productive layers; Normalized net present value The calculation formula is: ； in, For time, Let t be the cash inflow in year t. Let t be the cash outflow in year t. is the discount rate.

2. The method for stratified and balanced exploitation and control of natural gas hydrate reservoirs based on simulation optimization according to claim 1, characterized in that: Step S5 also includes the following steps: S51: Generate an initial parameter sample set, call a multi-field coupled numerical model for each parameter sample to obtain the corresponding objective function value, and build a sample database; S52: The parameter combination of each sample point is used to obtain the corresponding objective function value through a multi-field coupled numerical model, forming a sample database of parameter combination-simulation results; S53: The simulation results of the sample set are used to train a proxy model through machine learning algorithms; S54: Using the surrogate model as a fast evaluator, obtain the optimal combination of engineering decision parameters X_optimal.

3. The method for stratified and balanced exploitation control of natural gas hydrate reservoirs based on simulation optimization according to claim 2, characterized in that: In step S51, an orthogonal experimental design method is used to extract a certain number of sample points in the engineering decision parameter space to generate an initial parameter sample set.

4. The method for stratified and balanced exploitation and control of natural gas hydrate reservoirs based on simulation optimization according to claim 3, characterized in that: In step S6, real-time monitoring data is fed back to the multi-field coupled numerical model, and the uncertainty parameters in the model are inverted and corrected so that the model always maintains an accurate mapping of the underground situation.

Images