A method, equipment and storage medium for intelligent control of water intrusion in gas reservoirs
By constructing a multi-objective dynamic data set of water intrusion and screening the main controlling factors using machine learning models, and combining multi-objective optimization algorithms, the optimal well pattern and well control parameters of the gas reservoir were determined. This solved the problem of poor drainage and gas production effect caused by the single basis of existing well pattern design, and realized stable production support for the gas reservoir.
Patent Information
- Application Number
- CN202411556306.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-04
- Publication Date
- 2025-11-14
- Estimated Expiration
- 2044-11-04
AI Technical Summary
The existing gas reservoir well network design is based on a single standard, resulting in lower-than-expected drainage and gas production effects, making it difficult to effectively mitigate water intrusion damage and affecting the stable production capacity and recovery rate of gas wells.
By constructing a multi-objective dynamic dataset of water intrusion, using machine learning models to screen the main controlling factors, establishing a comprehensive evaluation model for drainage wells, and solving the drainage and extraction regime through multi-objective optimization algorithms, the optimal well network and well control parameters are determined.
It enables intelligent control of water intrusion in gas reservoirs, improves the reliability of well network design, reduces computational costs, effectively mitigates water intrusion damage, and supports stable gas reservoir production.
Smart Images
Figure CN119494267B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of oil and gas development technology, specifically to a method, equipment, and storage medium for intelligent control of water intrusion in gas reservoirs. Background Technology
[0002] Natural gas is a crucial component of fossil fuels and is typically developed through depletion. However, natural gas reservoirs are often connected to groundwater bodies. As formation pressure gradually decreases during extraction, formation water can infiltrate the reservoir, leading to reduced gas-phase permeability and increased reservoir abandonment pressure. This severely impacts the stable production capacity and recovery rate of gas wells. Therefore, reservoir water management is of paramount importance for maintaining the efficient development of natural gas.
[0003] Drainage gas production is one of the effective methods for water management in gas reservoirs. The key to improving the effectiveness of drainage gas production lies in the design of the drainage well network and the optimization of well control parameters for existing production wells. Well network design mainly involves determining drainage wells, but existing design criteria only include one of the following: structural characteristics, water intrusion conditions, and economic benefits. Considering only a single factor can lead to the failure of crucial factors, resulting in a well network design effect that falls short of expectations. The determination of well control parameters for the drainage system can be viewed as a constrained nonlinear optimization problem. Current methods that artificially pursue a single objective by matching process parameters with the current development status are unlikely to yield optimal solutions, making it difficult to effectively mitigate water intrusion damage to gas reservoirs in field applications. Summary of the Invention
[0004] In view of this, the present invention proposes a method for intelligent control of water intrusion in gas reservoirs, which can achieve efficient optimization of well network construction and obtain accurate well control parameters.
[0005] To solve at least one of the above-mentioned technical problems, the present invention provides a method for intelligent control of water intrusion in gas reservoirs, comprising the following steps:
[0006] Step S1: Randomly generate multiple sets of well group production simulation models including different geological parameters, engineering parameters, development parameters and target parameters to construct a multi-objective water intrusion dynamic dataset, and normalize the dataset;
[0007] Step S2: Train a machine learning model using the normalized dataset, use the trained machine learning model to calculate the factors affecting water intrusion dynamics and rank their importance to obtain the main controlling factors affecting water intrusion dynamics and their corresponding weights.
[0008] Step S3: Establish a comprehensive evaluation model for drainage wells. Input the corresponding parameters of the production wells that produce water in the well network of the production area into the comprehensive evaluation model for drainage wells. Determine the production wells with an evaluation value greater than 0.5 calculated in the comprehensive evaluation model for drainage wells as drainage wells, and then determine the drainage well network composed of production wells and drainage wells.
[0009] The comprehensive evaluation model for drainage wells is shown in Equation (1):
[0010]
[0011] In equation (1), V is the comprehensive evaluation value of the drainage well, and t = 1, 2, ..., n;
[0012] Step S4: Based on the drainage well network, establish a multi-objective optimization mathematical model for the drainage system, wherein the objective function of the mathematical model is shown in equations (2)-(2.2):
[0013] minf(x), f(x) = [f EW -f EG (2)
[0014]
[0015] In equation (2), f EW For the cumulative water production of all gas wells, f EG G represents the cumulative gas production of all gas wells. i W represents the daily gas production of the i-th gas well. j This represents the daily water production of the j-th drainage well;
[0016] The constraints are the upper and lower limits of the daily gas production of the gas production well and the upper and lower limits of the daily water production of the drainage well, as shown in equation (3):
[0017]
[0018] In equation (3), i represents the i-th gas producing well, i = 1, 2, ..., k, G i G represents the daily gas production of the i-th gas well. imin G represents the minimum daily gas production of the i-th gas-producing well. imax W represents the maximum daily gas production of the i-th gas producing well, and j represents the j-th draining well, where j = 1, 2, ..., m. j W represents the daily water production of the j-th drainage well. jmin W represents the minimum daily water production of the j-th drainage well. jmax This represents the maximum daily water production of the j-th drainage well;
[0019] Step S5: Use a multi-objective optimization algorithm to solve the multi-objective optimization mathematical model of the drainage system, and obtain the Pareto solution set and well control parameters corresponding to different drainage system schemes.
[0020] Another object of the present invention is to provide a device for intelligent control of water intrusion in gas reservoirs, comprising,
[0021] processor;
[0022] A storage module stores a program for intelligent control of water intrusion in a gas reservoir that can run on the processor, wherein the program for intelligent control of water intrusion in a gas reservoir implements the steps described in the method when executed by the processor.
[0023] The output module is used to output the calculation results.
[0024] Furthermore, the present invention also provides a computer-readable storage medium storing processor-executable code, characterized in that the computer-readable storage medium includes multiple instructions configured to cause the processor to execute the aforementioned intelligent control method for water intrusion into gas reservoirs.
[0025] The technical effects achieved by this invention are:
[0026] This invention discloses a method, equipment, and storage medium for intelligent control of water intrusion in gas reservoirs. It can accurately obtain Pareto solution sets and corresponding well control parameters for different schemes through multi-objective optimization algorithms, effectively realizing intelligent control of water intrusion damage in gas reservoirs and providing support for stable gas reservoir production. Attached Figure Description
[0027] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.
[0028] Figure 1 This is a flowchart of the intelligent gas reservoir control method in this invention;
[0029] Figure 2 This is a schematic diagram of the conceptual numerical simulation model of the intelligent gas reservoir control method in this invention;
[0030] Figure 3 A graph showing the ranking of the importance of the influencing factors of the intelligent gas reservoir regulation method in this invention.
[0031] Figure 4 This is a schematic diagram of the optimal well pattern in the target area of the intelligent gas reservoir control method of this invention;
[0032] Figure 5 This is a schematic diagram of the Pareto front obtained by solving the intelligent gas reservoir control method in this invention;
[0033] Figure 6 These are the well control parameters for each individual well under the optimal drainage and production regime of the intelligent gas reservoir control method in this invention. Detailed Implementation
[0034] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings.
[0035] 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 a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present 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.
[0036] A method for intelligent control of water intrusion in gas reservoirs includes the following steps:
[0037] Step S1: Randomly generate multiple sets of well group production simulation models including different geological parameters, engineering parameters, development parameters and target parameters to construct a multi-objective water intrusion dynamic dataset, and normalize the dataset;
[0038] The geological parameters include porosity, permeability, and water saturation; the engineering parameters include the number of gas-producing wells and well spacing; the development parameters include formation pressure and water-to-gas ratio; and the target parameters include cumulative gas production and cumulative water production.
[0039] A large number of parameters, including different geological factors, engineering factors, and corresponding development status parameters, as well as multiple target values, are normalized to construct a dynamic water intrusion dataset containing multiple targets; the normalization formula is shown in Equation (5):
[0040]
[0041] In equation (5), S z The value is the normalized value, and S is the initial value. max S is the maximum value in the dataset. min It is the minimum value in the dataset.
[0042] Step S2: Train a machine learning model using the normalized dataset, use the trained machine learning model to calculate the factors affecting water intrusion dynamics and rank their importance to obtain the main controlling factors affecting water intrusion dynamics and their corresponding weights.
[0043] The machine learning model is one of Random Forest, GBDT, XGBoost, or LightGBM. Its training set, validation set, and test set are divided into the above-mentioned water intrusion dynamic dataset in an 8:1:1 ratio. The training set is used to train the basic machine learning model, the validation set is used to adjust the hyperparameters of the machine learning model, and the test set is used to test the performance of the machine learning model.
[0044] The method for ranking importance is the Pearson correlation coefficient method. The main control factors are the influencing factors with a correlation coefficient greater than 0.1. The method for calculating the weight of the main control factors is to sum the correlation coefficients of each main control factor, and then divide the correlation coefficient of each main control factor by the sum of the correlation coefficients of the main control factors to obtain the weight of each main control factor.
[0045] Step S3: Establish a comprehensive evaluation model for drainage wells. Input the corresponding parameters of the production wells that produce water in the well network of the production area into the comprehensive evaluation model for drainage wells. Determine the production wells with an evaluation value greater than 0.5 calculated in the comprehensive evaluation model for drainage wells as drainage wells, and then determine the drainage well network composed of production wells and drainage wells.
[0046] The comprehensive evaluation model for drainage wells is shown in Equation (1):
[0047]
[0048] In equation (1), V is the comprehensive evaluation value of the drainage well, and t = 1, 2, ..., n;
[0049] Step S4: Based on the drainage well network, establish a multi-objective optimization mathematical model for the drainage system, mainly including setting optimization variables, constraints, and optimization objectives; the optimization variables are the daily gas production of the gas production wells and the daily drainage volume of the drainage wells; the constraints are the upper and lower limits of the daily gas production of the gas production wells and the upper and lower limits of the daily water production of the drainage wells; the optimization objectives are to minimize the cumulative water production of all gas production wells and maximize the cumulative gas production of all gas production wells, wherein the objective function of the mathematical model is shown in equations (2)-(2.2):
[0050] minf(x), f(x) = [f EW -f EG (2)
[0051]
[0052] In equation (2), f EW For the cumulative water production of all gas wells, f EG G represents the cumulative gas production of all gas wells. i W represents the daily gas production of the i-th gas well. j This represents the daily water production of the j-th drainage well;
[0053] The constraints are the upper and lower limits of the daily gas production of the gas production well and the upper and lower limits of the daily water production of the drainage well, as shown in equation (3):
[0054]
[0055] In equation (3), i represents the i-th gas producing well, i = 1, 2, ..., k, G i G represents the daily gas production of the i-th gas well. imin G represents the minimum daily gas production of the i-th gas-producing well. imax W represents the maximum daily gas production of the i-th gas producing well, and j represents the j-th draining well, where j = 1, 2, ..., m. j W represents the daily water production of the j-th drainage well. jmin W represents the minimum daily water production of the j-th drainage well. jmax This represents the maximum daily water production of the j-th drainage well;
[0056] Step S5: Solve the multi-objective optimization mathematical model of the drainage system using a multi-objective optimization algorithm to obtain the Pareto solution set and well control parameters corresponding to different drainage system schemes. The multi-objective optimization algorithm is one of NSGA-II, MOPSO, or MOEA / D.
[0057] The performance of the multi-objective optimization algorithm and the performance of its solution set across multiple objectives are evaluated using the HV index, as shown in Equation (4):
[0058]
[0059] In equation (4), PS is the Pareto solution set, M is the reference point, and v is the hypercube.
[0060] This invention provides an intelligent control device for water intrusion in gas reservoirs, comprising,
[0061] processor;
[0062] A storage module stores a program for intelligent control of water intrusion in a gas reservoir that can run on the processor, wherein the program for intelligent control of water intrusion in a gas reservoir implements the steps described in the method when executed by the processor.
[0063] The output module is used to output the calculation results.
[0064] The processor includes:
[0065] The dataset construction module is used to construct a large amount of dynamic water intrusion data, which includes different geological factors, engineering factors, and corresponding development status parameter combinations and multiple target values.
[0066] The machine learning model training module is used to screen the main controlling factors affecting water intrusion dynamics and obtain their corresponding weights.
[0067] The well network determination module is used to construct a comprehensive evaluation model of drainage wells to determine the well network.
[0068] The module for constructing a multi-objective optimization mathematical model of the polling scheduling system is used to construct a multi-objective optimization mathematical model of the polling scheduling system that includes optimization variables, constraints, and optimization objectives.
[0069] The well control parameter optimization module is used to obtain Pareto solution sets and corresponding well control parameters for different schemes by solving a multi-objective optimization algorithm.
[0070] The present invention also provides a computer-readable storage medium storing processor-executable code, the computer-readable storage medium including a plurality of instructions configured to cause the processor to execute the above-described intelligent control method for water intrusion in gas reservoirs.
[0071] Example:
[0072] Taking a production block in a real gas field that has suffered water intrusion damage during depletion development as an example, 500 production simulation models were randomly generated, each with different porosity, permeability, water saturation, number of gas-producing wells, well spacing, formation pressure at the start of drainage and gas production, and water-gas ratio. A schematic diagram of the production simulation model is shown below. Figure 2 As shown, this mainly illustrates the relationship between formation depth and the model, and numerical simulations were run to obtain the cumulative gas production and cumulative water production under corresponding parameter combinations. After normalization, a dataset containing 500 different parameter combinations was obtained.
[0073] The XGBoost model was trained using a training set consisting of 400 sets of data, and the hyperparameters of the XGBoost model were tuned using a validation set consisting of 50 sets of data. The results of the hyperparameter tuning are shown in Table 1. The performance of the XGBoost model was tested using a test set consisting of 50 sets of data.
[0074] Table 1. Hyperparameters of the XGBoost model
[0075] Parameter name Value Parameter name Value Maximum tree depth 9 Number of independent trees that make up the set 2000 Learning rate 0.10 The minimum sample weights of child nodes 5
[0076] The ranking of the importance of the influencing factors is as follows: Figure 3 As shown, the water-to-gas ratio, permeability, well spacing, number of gas-producing wells, and porosity are the main controlling factors affecting gas and water production after water intrusion. The weights for the water-to-gas ratio, permeability, well spacing, number of gas-producing wells, and porosity are 0.20, 0.10, 0.20, 0.10, and 0.40, respectively.
[0077] The main control parameters of each well are input into the comprehensive evaluation model of drainage wells. The drainage wells deployed in the target area are TZ4, TZ5, TZ15, and the existing drainage well TZ18. The optimal well network diagram of the target area is shown below. Figure 4 As shown.
[0078] The mathematical model for optimizing the polling system was solved using NSGA-II, MOPSO, and MOEA / D. The resulting Pareto front diagram is shown below. Figure 5 As shown in Table 2, the HV index of different multi-objective optimization algorithms was calculated, and the results are shown in Table 2.
[0079] Table 2 HV index values for different algorithms
[0080] Optimization Algorithm NSGA-II MOEA / D MOPSO HV index value 0.273 0.142 0.202
[0081] NSGA-II has the highest HV index value, and the optimal scheme has a cumulative gas production of 700.93 × 10⁻⁶ over 10 years. 8 m 3 The total water production in the region was 11.61 × 10⁻⁶. 8 m 3 The corresponding optimized daily production rate for the gas reservoir is 1094 × 10⁻⁶. 4 m 3 / d, the daily drainage volume of the entire gas reservoir area is 1298m³. 3 / d, the well control parameters for each well under the optimal drainage regime are as follows: Figure 6 As shown.
[0082] As can be seen, this invention determines drainage wells by comprehensively considering various factors, increases the credibility of drainage well network design, and reduces the calculation cost of directly optimizing well locations; multi-objective intelligent optimization of well control parameters effectively avoids getting trapped in local optima, ensures the operability of optimization results, realizes intelligent control of water intrusion damage in gas reservoirs, and provides strong support for stable natural gas production.
[0083] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the embodiments of the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. A method for intelligent control of water intrusion in gas reservoirs, characterized in that, Includes the following steps: Step S1: Randomly generate multiple sets of well group production simulation models including different geological parameters, engineering parameters, development parameters and target parameters to construct a multi-objective water intrusion dynamic dataset, and normalize the dataset; Step S2: Train a machine learning model using the normalized dataset. Utilize the trained model to calculate the influencing factors affecting water intrusion dynamics and rank their importance to determine the dominant factors and their corresponding weights. The ranking method is the Pearson correlation coefficient method. Dominant factors are those with a correlation coefficient greater than 0.
1. The weights of dominant factors are calculated by summing the correlation coefficients of each dominant factor, then dividing each dominant factor's correlation coefficient by the sum of their correlation coefficients to obtain the weights of each dominant factor. Step S3: Establish a comprehensive evaluation model for drainage wells. Input the corresponding parameters of the production wells that produce water in the well network of the production area into the comprehensive evaluation model for drainage wells. Determine the production wells with an evaluation value greater than 0.5 calculated in the comprehensive evaluation model for drainage wells as drainage wells, and then determine the drainage well network composed of production wells and drainage wells. The comprehensive evaluation model for drainage wells is shown in Equation (1): In equation (1), V is the comprehensive evaluation value of the drainage well, and t = 1, 2, ..., n; Step S4: Based on the drainage well network, establish a multi-objective optimization mathematical model for the drainage system, wherein the objective function of the mathematical model is shown in equations (2)-(2.2): minf(x),f(x)=[f EW ,-f EG ](2) In equation (2), f EW For the cumulative water production of all gas wells, f EG G represents the cumulative gas production of all gas wells. i W represents the daily gas production of the i-th gas well. j This represents the daily water production of the j-th drainage well; The constraints are the upper and lower limits of the daily gas production of the gas production well and the upper and lower limits of the daily water production of the drainage well, as shown in equation (3): In equation (3), i represents the i-th gas producing well, i = 1, 2, ..., k, G i G represents the daily gas production of the i-th gas well. imin G represents the minimum daily gas production of the i-th gas-producing well. imax W represents the maximum daily gas production of the i-th gas producing well, and j represents the j-th draining well, where j = 1, 2, ..., m. j W represents the daily water production of the j-th drainage well. jmin W represents the minimum daily water production of the j-th drainage well. jmax This represents the maximum daily water production of the j-th drainage well; Step S5: Use a multi-objective optimization algorithm to solve the multi-objective optimization mathematical model of the drainage system, and obtain the Pareto solution set and well control parameters corresponding to different drainage system schemes.
2. The intelligent control method for water intrusion in a gas reservoir according to claim 1, characterized in that: The geological parameters include porosity, permeability, and water saturation; the engineering parameters include the number of gas-producing wells and well spacing; the development parameters include formation pressure and water-to-gas ratio; and the target parameters include cumulative gas production and cumulative water production.
3. The intelligent control method for water intrusion in a gas reservoir according to claim 1, characterized in that: The machine learning model mentioned in step S2 is one of Random Forest, GBDT, XGBoost, and LightGBM.
4. The intelligent control method for water intrusion in a gas reservoir according to claim 1, characterized in that: The multi-objective optimization algorithm mentioned in step S5 is one of NSGA-II, MOPSO, and MOEA / D.
5. A smart control device for water intrusion in a gas reservoir, characterized in that, include, processor; A storage module stores a program for intelligent control of water intrusion in a gas reservoir that can run on the processor, wherein the program for intelligent control of water intrusion in a gas reservoir, when executed by the processor, implements the steps described in any one of claims 1-4. The output module is used to output the calculation results.
6. A computer-readable storage medium storing a processor-executable code program, characterized in that, The computer-readable storage medium includes a plurality of instructions configured to cause a processor to execute the intelligent control method for water intrusion in a gas reservoir as described in any one of claims 1-4.