A multi-stage fractured horizontal well CO2 huff and puff injection-production optimization method

Abstract

This invention provides an optimization method for CO2 injection and production in multi-stage fractured horizontal wells. The optimization method includes: Step S1: Establishing a reservoir simulation calculation model required for CO2 injection and production optimization in multi-stage fractured horizontal wells; Step S2: Establishing the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fractured horizontal wells; Step S3: Establishing a framework for the optimization method of CO2 injection and production in multi-stage fractured horizontal wells based on PPO; Step S4: Applying the obtained optimization parameters for CO2 injection and production in multi-stage fractured horizontal wells to specific blocks and analyzing the reasons for the high total net present value of the injection and production optimization scheme. This method can simultaneously and dynamically optimize the injection and production rates and durations at different stages of injection and production, achieving variable injection and production rates and durations, which better meets the practical needs of injection and production in mining operations.

Description

Technical Field

[0001] This invention relates to the field of oil and gas field development, production and injection, and particularly to an optimized method for CO2 injection and production in multi-stage fractured horizontal wells. Background Technology

[0002] Tight oil reservoirs, due to their dense formation and extremely low permeability, exhibit excessively high flow resistance, making it difficult to achieve industrial production through conventional vertical well development. Therefore, multi-stage fracturing horizontal wells are necessary to stimulate the tight reservoir, with artificial fracture networks providing high-speed channels for crude oil seepage. While multi-stage fracturing horizontal wells offer high initial production, this decline is rapid. The near-wellbore fractured zone has limited drainage area, and the crude oil in the unstimulated areas further away cannot be utilized, leading to a sharp reduction in reservoir fluid supply capacity. CO2 huff and puff is an important alternative for the flexible development of multi-stage fracturing horizontal wells, and optimizing huff and puff injection-production is a crucial measure to improve its development effectiveness, offering advantages such as ease of construction, low cost, and significant oil production enhancement.

[0003] Currently, CO2 throughput injection-production optimization methods fall into two categories. One relies on the field experience of experts to determine the injection-production rate and production duration for each stage of throughput and production. This method is highly subjective, and its effectiveness is limited by the experience level of the field experts. The other method utilizes optimization theory to calculate the optimal injection-production parameters. Optimization-based methods can be categorized into two types based on whether gradient information is required: gradient-based methods, which have high computational efficiency but require the partial derivatives of the optimization target with respect to the injection-production optimization variables, making implementation difficult by rewriting reservoir simulator code and significantly limiting their application. Gradient-free methods overcome the shortcomings of gradient-based methods by not requiring partial derivative information, making them easier to implement and more widely applicable, but they have low computational efficiency. The Proximal Policy Optimization (PPO) method used in this patent is a gradient-free intelligent injection-production optimization method that does not require partial derivative information. Furthermore, by combining it with transfer learning, it has excellent generalization ability. When extended to other blocks for throughput injection-production optimization, only the neural network weights need to be fine-tuned, significantly saving the time required to train the neural network.

[0004] Current optimization methods for CO2 injection and production in multi-stage fractured horizontal wells have significant shortcomings. These methods fail to comprehensively consider key influencing factors such as fracture network morphology, stress-sensitive effects, and CO2 molecule diffusion, resulting in optimization outcomes that cannot effectively guide field practice. Existing CO2 injection and production optimization methods focus on optimizing the injection and production rates at different stages of the injection and production process, without optimizing the production duration at each stage. Summary of the Invention

[0005] In view of the above problems, the present invention is proposed to provide an optimized CO2 injection and production method for multi-stage fracturing horizontal wells that overcomes or at least partially solves the above problems.

[0006] According to one aspect of the present invention, an optimized method for CO2 huff-and-puff injection and production in multi-stage fractured horizontal wells is provided, the optimization method comprising:

[0007] Step S1: Establish a reservoir simulation calculation model for optimizing CO2 injection and production in multi-stage fractured horizontal wells;

[0008] Step S2: Establish the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fracturing horizontal wells;

[0009] Step S3: Establish a framework for optimizing CO2 injection and production in multi-stage fracturing horizontal wells based on PPO;

[0010] Step S4: Apply the obtained CO2 injection and production optimization parameters of the multi-stage fractured horizontal well to the specific block, and analyze the reasons for the high total net present value of the injection and production optimization scheme.

[0011] Optionally, step S1: establishing the reservoir simulation calculation model required for CO2 huff-and-puff injection-production optimization in multi-stage fracturing horizontal wells specifically includes:

[0012] Geological data of the study area were collected, and a geological model of the study area was established using stochastic modeling, taking into account the influence of the morphology of the pressure fracture network and the stress pressure sensitivity effect.

[0013] The fluid model adopts the CO2-driven component model, taking into account the effect of CO2 molecule diffusion;

[0014] Import the geological model and fluid model into the reservoir numerical simulation software.

[0015] Optionally, step S2: establishing the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fracturing horizontal wells specifically includes:

[0016] The optimization objective for CO2 injection and production optimization in multi-stage fractured horizontal wells is determined. Net present value is the optimization objective, and the cumulative reward of the agent is the total net present value of the entire injection and production optimization stage. The injection and production optimization aims to maximize the total net present value of the cumulative reward.

[0017] Determine the optimization variables for CO2 huff-and-puff injection and production optimization in multi-stage fractured horizontal wells;

[0018] Determine the constraints for optimizing CO2 injection and production in multi-stage fractured horizontal wells.

[0019] Optionally, the CO2 injection and production optimization variables for the multi-stage fracturing horizontal well include: injection rate, injection time, well shut-in time, backflow rate, and backflow time, reflecting the injection and production characteristics of the three stages of CO2 injection, shut-in, and backflow.

[0020] Optionally, the constraints include the bottom-hole flowing pressure of the injection well being lower than the rock fracturing pressure, and the bottom-hole flowing pressure of the production well being higher than the crude oil bubble point pressure, so that the injection-production optimization system meets the constraints of field production practice.

[0021] Optionally, the method framework specifically includes: an Actor action selection neural network and a Critic value evaluation neural network.

[0022] Optionally, step S3: establishing a framework for optimizing CO2 huff-and-puff injection and production in multi-stage fracturing horizontal wells based on PPO specifically includes:

[0023] Step S31: The Actor neural network selects the optimal injection and production parameters for CO2 huff and puff in the multi-stage fracturing horizontal well based on the injection and production environment status.

[0024] Step S32: The reservoir simulator E300 executes the optimal injection and production parameters for the current time step to obtain production data of CO2 injection and production wells in multi-stage fractured horizontal wells.

[0025] Step S33: Using the well production data obtained from simulation calculations, calculate the immediate reward for this injection-production adjustment time step, and proceed to the injection-production status of the next time step.

[0026] Step S34: Use the instantaneous reward, injection parameters, and injection state characterization parameters as samples (s) to train the neural network. t ,a t ,r t ,s t+1 Stored in the experience pool;

[0027] Step S35: The online neural network for Critic value evaluation assesses the cumulative reward corresponding to the injection and production parameters adopted in the current injection and production time step, and the target neural network for Critic value evaluation searches for the historical maximum cumulative reward in the experience pool for the current injection and production time step.

[0028] Step S36: Iterate repeatedly until the online cumulative reward at the current time step converges to the maximum historical cumulative reward. Each injection and extraction time step is adjusted to obtain the optimal throughput injection and extraction parameters, thus obtaining the optimal injection and extraction system for the entire injection and extraction optimization phase.

[0029] Optionally, step S4: applying the obtained CO2 huff-and-puff injection-production optimization parameters of the multi-stage fractured horizontal well to a specific block, and analyzing the reasons for the high total net present value of the injection-production optimization scheme, specifically includes:

[0030] Step S41: In the field injection and production practice, the optimal injection and production parameters for CO2 huff and puff in multi-stage fractured horizontal wells are adopted.

[0031] Step S42: Analyze the CO2 injection and production optimization scheme for multi-stage fractured horizontal wells, and the main mechanism that can improve the cumulative reward of the optimization target compared with the benchmark scheme.

[0032] Optionally, the injection and production status variables include: production time, daily gas injection volume, daily gas production volume, daily oil production volume, production gas-oil ratio, bottom hole flowing pressure, cumulative gas injection volume, cumulative gas production volume, and cumulative oil production volume.

[0033] Optionally, the optimization variables specifically include: the gas injection rate, gas injection time, well shut-in time, backflow rate, and backflow time of the well.

[0034] A CO2 huff-and-puff injection-and-production optimization system for multi-stage fractured horizontal wells, applying the aforementioned CO2 huff-and-puff injection-and-production optimization method for multi-stage fractured horizontal wells, characterized in that the optimization system specifically includes:

[0035] The reservoir simulation calculation model establishment module is used to establish the reservoir simulation calculation model required for CO2 huff and puff injection and production optimization in multi-stage fracturing horizontal wells;

[0036] The constraint establishment module is used to establish the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fracturing horizontal wells;

[0037] The injection-production optimization method framework establishment module is used to establish an injection-production optimization method framework for multi-stage fracturing horizontal wells based on PPO;

[0038] The block application module is used to apply the obtained CO2 injection and production optimization parameters of multi-stage fractured horizontal wells to specific blocks and analyze the reasons for the high total net present value of the injection and production optimization scheme.

[0039] A storage medium for optimizing CO2 injection and production in multi-stage fractured horizontal wells, storing an optimized CO2 injection and production system for multi-stage fractured horizontal wells.

[0040] This invention provides an optimization method for CO2 injection and production in multi-stage fractured horizontal wells. The optimization method includes: Step S1: Establishing a reservoir simulation calculation model required for CO2 injection and production optimization in multi-stage fractured horizontal wells; Step S2: Establishing the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fractured horizontal wells; Step S3: Establishing a framework for the optimization method of CO2 injection and production in multi-stage fractured horizontal wells based on PPO; Step S4: Applying the obtained optimization parameters for CO2 injection and production in multi-stage fractured horizontal wells to specific blocks and analyzing the reasons for the high total net present value of the injection and production optimization scheme. This method can simultaneously and dynamically optimize the injection and production rates and durations at different stages of injection and production, achieving variable injection and production rates and durations, which better meets the practical needs of injection and production in mining operations.

[0041] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 A flowchart illustrating an optimized CO2 huff-and-puff injection-production method for multi-stage fracturing horizontal wells, provided in an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of the framework for the PPO-based multi-stage fracturing horizontal well CO2 huff and puff injection optimization method provided in an embodiment of the present invention;

[0045] Figure 3 This is a convergence diagram of the target for CO2 huff-and-puff injection and production optimization in a multi-stage fracturing horizontal well based on PPO, provided in an embodiment of the present invention.

[0046] Figure 4 This is a comparison chart of cumulative production optimization for multi-stage fracturing horizontal wells based on PPO, provided in an embodiment of the present invention.

[0047] Figure 5 The graph shows the variation of optimized CO2 injection and production parameters for multi-stage fracturing horizontal wells based on PPO, as provided in an embodiment of the present invention. Detailed Implementation

[0048] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0049] The terms "comprising" and "having," and any variations thereof, in the specification, embodiments, claims, and drawings of this invention are intended to cover non-exclusive inclusion, such as including a series of steps or units.

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0051] Example 1

[0052] like Figure 1 and Figure 2 As shown, an optimized CO2 huff-and-puff injection-production method for multi-stage fracturing horizontal wells includes:

[0053] Step 1: Establish a reservoir simulation model for optimizing CO2 injection and production in multi-stage fractured horizontal wells. Collect geological data for the study block and establish a geological model of the study area using stochastic modeling, considering the influence of fracture network morphology and stress-sensitive effects. The fluid model adopts a CO2 flooding component model, considering the influence of CO2 molecule diffusion. Finally, import the geological model and fluid model into the reservoir numerical simulation software.

[0054] Step 2: Establish the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fractured horizontal wells.

[0055] Step 21: Determine the optimization objective for CO2 injection-production optimization in multi-stage fractured horizontal wells. Net present value (NPV) is the optimization objective, and the agent's cumulative reward is the total NPV of the entire optimization phase. The optimization aims to maximize the total NPV of the cumulative reward.

[0056] Step 22: Determine the optimization variables for CO2 injection and production optimization in multi-stage fractured horizontal wells. The optimization variables for CO2 injection and production optimization in multi-stage fractured horizontal wells include gas injection rate, gas injection time, well shut-in time, backflow rate, and backflow time, reflecting the injection and production characteristics of the three stages of CO2 injection and production: inhalation, shut-in, and outflow.

[0057] Step 23: Determine the constraints for optimizing CO2 injection and production in multi-stage fractured horizontal wells. These constraints include ensuring that the bottom-hole flowing pressure of the injection well is lower than the rock fracturing pressure, and the bottom-hole flowing pressure of the production well is higher than the crude oil bubble point pressure, so that the injection and production optimization regime meets the constraints of actual field production practices.

[0058] Step 3: Establish a framework for optimizing CO2 injection and production in multi-stage fracturing horizontal wells based on PPO. The framework mainly includes an Actor action selection neural network and a Critic value evaluation neural network.

[0059] Step 31: The Actor neural network selects the optimal injection and production parameters for CO2 influx and outflow in the multi-stage fracturing horizontal well based on the injection and production environment status.

[0060] Step 32: The reservoir simulator E300 executes the optimal injection and production parameters for the current time step to obtain CO2 injection and production data for multi-stage fractured horizontal wells.

[0061] Step 33: Using the well production data obtained from simulation calculations, calculate the immediate reward for this injection-production adjustment time step, and proceed to the injection-production status of the next time step.

[0062] Step 34: Use the representation parameters such as reward, injection parameters, and injection status as samples (s) to train the neural network. t ,a t ,r t ,s t+1 ) Stored in the experience pool.

[0063] Step 35: The Critic value evaluation online neural network assesses the cumulative reward corresponding to the injection and production parameters adopted in the current injection and production time step, and the Critic value evaluation target neural network searches for the historical maximum cumulative reward in the experience pool for the current injection and production time step.

[0064] Step 36: Iterate repeatedly until the online cumulative reward at the current time step converges to the maximum historical cumulative reward. Each injection and extraction time step is adjusted to obtain the optimal throughput injection and extraction parameters, thereby obtaining the optimal injection and extraction system for the entire injection and extraction optimization phase.

[0065] Step 4: Apply the obtained CO2 injection and production optimization parameters of the multi-stage fractured horizontal well to the specific block, and analyze the reasons for the high total net present value of the injection and production optimization scheme.

[0066] Step 41: In the field injection and production practice, the optimal injection and production parameters for CO2 huff and puff in multi-stage fractured horizontal wells are adopted.

[0067] Step 42: Analyze the CO2 injection and production optimization scheme for multi-stage fractured horizontal wells, and the main mechanism that can improve the cumulative reward of the optimization target compared with the benchmark scheme.

[0068] Step 1 establishes a reservoir simulation calculation model for CO2 injection-production optimization in multi-stage fractured horizontal wells. The key feature is that the numerical simulation model for CO2 injection-production optimization in multi-stage fractured horizontal wells comprehensively considers major influencing factors such as fracture network morphology, CO2 molecule diffusion, and stress pressure-sensitive effects. The numerical simulation model more closely reflects the actual conditions in the mining field, and the resulting injection-production optimization parameters can better guide field practice.

[0069] Step 2: Establish the optimization objective, optimization variables, and constraints for CO2 injection and production optimization in multi-stage fractured horizontal wells. The key feature is that the net present value (NPV) is used as the optimization objective, the cumulative reward is the total NPV of the entire optimization phase, and the immediate reward is the NPV corresponding to the current injection and production adjustment time step. The optimization parameters for CO2 injection and production in multi-stage fractured horizontal wells include injection rate, injection time, well shut-in time, backflow rate, and backflow time. Constraints include the injection well bottomhole flowing pressure being lower than the rock fracturing pressure and the production well bottomhole flowing pressure being higher than the crude oil bubble point pressure.

[0070] Step 3: Establish a framework for optimizing CO2 injection and production in multi-stage fracturing horizontal wells based on PPO. The key feature is that, in step 31, the Actor neural network selects the optimal injection and production parameters for the current injection and production adjustment time step. Specifically, the agent selects the injection and production parameters for the current injection and production adjustment time step that maximizes the cumulative reward based on the characteristics of the injection and production environment. These environmental characteristics are described by state variables, including production time, daily gas injection volume, daily gas production volume, daily oil production volume, production gas-oil ratio, bottomhole flowing pressure, cumulative gas injection volume, cumulative gas production volume, and cumulative oil production volume.

[0071] Step 32: The reservoir simulator E300 executes the optimal injection and production parameters for the current time step to obtain production data for the CO2 huff and puff well in the multi-stage fractured horizontal well. The key feature is that the multi-stage fractured horizontal well CO2 huff and puff reservoir numerical simulator E300 runs the optimal injection and production parameters for the current time step passed to the simulator, calculating the daily and cumulative production data for the huff and puff well at the current injection and production adjustment time step.

[0072] Step 33: Using the well production data obtained from simulation calculations, calculate the immediate reward for this injection-production adjustment time step, and proceed to the injection-production state of the next time step. The key feature is that calculating the net present value of the immediate reward stage requires considering the economic benefits brought by the cumulative oil production of the current injection-production time step and the cumulative gas injection costs of the stage, used to evaluate the stage economic benefits obtained from the injection-production parameters of this injection-production adjustment time step.

[0073] Step 34: Use the instant reward, throughput injection parameters, and injection state characterization parameters as samples (s) to train the neural network. t ,a t ,r t ,s t+1 The data is stored in the experience pool. Its characteristic is that the instantaneous reward, throughput injection parameters, injection status, and entry into the next injection state variable at the current injection adjustment time step are stored as a sample in the experience pool for training the Actor and Critic neural networks of the agent.

[0074] Step 35: The Critic value evaluation online network assesses the cumulative reward corresponding to the injection and production parameters adopted at the current injection and production time step. The Critic value evaluation target network searches for the historical maximum cumulative reward at the current injection and production time step in the experience pool. The key feature is that it compares the difference between the net present value of the online cumulative reward at the current injection and production adjustment time step and the net present value of the historical maximum cumulative reward. Subsequently, the Critic online value evaluation network continuously adjusts the weights to minimize the difference, and the corresponding online cumulative reward gradually approaches the historical maximum value.

[0075] Step 36: Iterate repeatedly until the online cumulative reward at the current time step gradually converges to the historical maximum cumulative reward. At this point, the optimal injection and extraction regime for the entire injection and extraction optimization phase is obtained. The key feature is that during the iterative learning process of the throughput well agent, the Critic value evaluation network adjusts the network weights to minimize the difference between the cumulative reward at the current injection and extraction adjustment time step and the historical maximum cumulative reward, causing the online cumulative reward to gradually approach the historical maximum. The Action Selection Actor network adjusts the neural network weights to maximize the cumulative reward of the injection and extraction parameters at the current time step. As the number of iterations increases, the optimal injection and extraction parameters for each injection and extraction adjustment time step are gradually found, and their combination constitutes the optimal injection and extraction regime for the entire injection and extraction optimization period.

[0076] Step 4 involves applying the obtained optimized CO2 injection and production parameters from multi-stage fractured horizontal wells to specific blocks. The key feature is that, in step 41, the optimal CO2 injection and production parameters from multi-stage fractured horizontal wells are used in the field injection and production practice. The injection and production rates and durations at different stages—intake, steaming, and release—change dynamically, better meeting the needs of actual field injection and production practices.

[0077] Step 42 analyzes the optimization scheme for CO2 injection and production in multi-stage fractured horizontal wells, identifying the main mechanisms by which the cumulative reward of the optimization target is improved compared to the baseline scheme. The key feature is that the injection and production parameters remain constant in the baseline scheme, while the parameters of the optimized scheme change dynamically. The total net present value of the optimization target in the optimized scheme is higher than that in the baseline scheme. By comparing the differences in cumulative oil production, cumulative gas injection volume, and cumulative gas production, the reasons for the higher total net present value of the optimal scheme are analyzed. The main mechanisms by which the injection and production scheme achieves higher economic benefits are identified, providing theoretical guidance for field injection and production practices.

[0078] Example 2

[0079] like Figure 1 As shown, the optimized CO2 huff-and-puff injection-production method for multi-stage fracturing horizontal wells based on PPO includes:

[0080] Step 1: Establish a reservoir simulation calculation model for optimizing CO2 injection and production in multi-stage fractured horizontal wells.

[0081] Specifically, the numerical simulation model comprehensively considers the main influencing factors such as the morphology of the hydraulic fracture network, CO2 molecule diffusion, and stress pressure sensitivity effect, and the resulting injection and production optimization parameters are closer to the actual field conditions.

[0082] Step 2: Establish the optimization objective, optimization variables, and constraints for CO2 injection and production optimization in multi-stage fractured horizontal wells.

[0083] Specifically, the optimization objective is determined by net present value, and the optimization variables include the gas injection rate, gas injection time, well shut-in time, flowback rate, and flowback time of the injection well. Constraints include that the bottomhole flowing pressure during the gas injection stage is lower than the rock fracturing pressure, and the bottomhole flowing pressure during the flowback stage is higher than the crude oil bubble point pressure.

[0084] Step 3: Establish a framework for optimizing CO2 injection and production in multi-stage fracturing horizontal wells based on PPO. The framework mainly includes an Actor action selection neural network and a Critic value evaluation neural network.

[0085] Specifically:

[0086] Step 31: The Actor selection network selects the optimal injection parameters for the current injection adjustment time step based on the characteristics of the injection environment at this injection adjustment time step.

[0087] Step 32: The reservoir component simulator E300 executes the optimal injection and production parameters for the current injection and production adjustment time step, and calculates and obtains the production data of the CO2 huff and puff well in the multi-stage fractured horizontal well.

[0088] Step 33: Using the production data of the multi-stage fractured horizontal well CO2 huff and puff well, calculate the instantaneous net present value of the injection-production adjustment time step, and proceed to the next injection-production adjustment time step.

[0089] Step 34: Use the instant reward, throughput injection parameters, and injection state characterization parameters as samples (s) to train the neural network. t ,a t ,r t ,s t+1 ) Stored in the experience pool.

[0090] Step 35: The Critic value evaluation online network assesses the cumulative reward corresponding to the injection and production parameters adopted at the current injection and production time step, and the Critic value evaluation target network searches for the historical maximum cumulative reward at the current injection and production time step in the experience pool.

[0091] Step 36, iterate repeatedly until the online accumulated reward at the current time step gradually converges to the maximum historical accumulated reward, such as... Figure 3 As shown, the optimal injection and production regime for the entire injection and production optimization phase is obtained at this point.

[0092] Step 4: Apply the obtained CO2 injection and production optimization parameters from the multi-stage fractured horizontal wells to specific blocks.

[0093] Specifically:

[0094] Step 41: In the field injection and production practice, the optimal injection and production parameters for CO2 huff and puff in multi-stage fractured horizontal wells are adopted. The injection and production rates and durations at different stages of huff and puff change dynamically, such as... Figure 5 As shown.

[0095] Step 42: Analyze the optimization scheme for CO2 injection and production in multi-stage fractured horizontal wells, and the main mechanism by which the cumulative net present value of the optimization target is improved compared with the benchmark scheme.

[0096] Specifically: Further analysis of the main reasons for the high net present value, comparing the differences in cumulative oil production, cumulative gas injection, and cumulative gas production, as shown in the appendix. Figure 4 As shown, the optimal injection-production scheme has higher cumulative oil production but lower cumulative gas injection and gas production compared to the baseline scheme. The injection-production scheme increases revenue by improving cumulative oil production and reduces costs by decreasing cumulative gas injection, thus achieving higher economic benefits. The conclusions drawn can provide theoretical guidance for injection-production practices in mining operations.

[0097] Beneficial effects: 1. The CO2 injection and production optimization method for multi-stage fracturing horizontal wells based on PPO established in this invention is a gradient-free intelligent optimization method with the advantages of easy implementation, wide applicability and strong generalization ability.

[0098] 2. The present invention is a CO2 huff and puff injection and production optimization method for multi-stage fracturing horizontal wells based on PPO. It can dynamically optimize the injection and production rate and injection and production duration at different stages of huff and puff, realize variable injection and production rate and injection and production duration, and better meet the practical needs of field huff and puff injection and production.

[0099] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An optimized method for CO2 huff-and-puff injection and production in multi-stage fractured horizontal wells, characterized in that, The optimization method includes: Step S1: Establish a reservoir simulation calculation model for optimizing CO2 injection and production in multi-stage fractured horizontal wells; Step S2: Establish the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fracturing horizontal wells; Step S3: Establish a framework for optimizing CO2 injection and production in multi-stage fracturing horizontal wells based on PPO; Step S4: Apply the obtained CO2 injection and production optimization parameters of the multi-stage fractured horizontal well to the specific block, and analyze the reasons for the high total net present value of the injection and production optimization scheme.

2. The method for optimizing CO2 huff-and-puff injection and production in a multi-stage fractured horizontal well according to claim 1, characterized in that, Step S1: Establishing the reservoir simulation calculation model required for CO2 huff-and-puff injection and production optimization in multi-stage fractured horizontal wells specifically includes: Geological data of the study area were collected, and a geological model of the study area was established using stochastic modeling, taking into account the influence of the morphology of the pressure fracture network and the stress pressure sensitivity effect. The fluid model adopts the CO2-driven component model, taking into account the effect of CO2 molecule diffusion; Import the geological model and fluid model into the reservoir numerical simulation software.

3. The optimized CO2 huff-and-puff injection method for multi-stage fracturing horizontal wells according to claim 1, characterized in that, Step S2: Establishing the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fracturing horizontal wells specifically includes: The optimization objective for CO2 injection and production optimization in multi-stage fractured horizontal wells is determined. Net present value is the optimization objective, and the cumulative reward of the agent is the total net present value of the entire injection and production optimization stage. The injection and production optimization aims to maximize the total net present value of the cumulative reward. Determine the optimization variables for CO2 huff-and-puff injection and production optimization in multi-stage fractured horizontal wells; Determine the constraints for optimizing CO2 injection and production in multi-stage fractured horizontal wells.

4. The optimized CO2 huff-and-puff injection method for multi-stage fracturing horizontal wells according to claim 3, characterized in that, The CO2 injection and production optimization variables for multi-stage fracturing horizontal wells include: injection rate, injection time, well shut-in time, backflow rate, and backflow time, reflecting the injection and production characteristics of the three stages of CO2 injection, shut-in, and backflow.

5. The optimized CO2 huff-and-puff injection method for multi-stage fracturing horizontal wells according to claim 3, characterized in that, The constraints include the bottom-hole flowing pressure of the gas injection well being lower than the rock fracture pressure, and the bottom-hole flowing pressure of the production well being higher than the crude oil bubble point pressure, so that the injection-production optimization system meets the constraints of field production practice.

6. The optimized CO2 huff-and-puff injection-production method for multi-stage fracturing horizontal wells according to claim 1, characterized in that, The method framework specifically includes: an Actor action selection neural network and a Critic value evaluation neural network.

7. The optimized CO2 huff-and-puff injection-production method for multi-stage fracturing horizontal wells according to claim 6, characterized in that, Step S3: Establishing a framework for optimizing CO2 huff-and-puff injection and production in multi-stage fracturing horizontal wells based on PPO specifically includes: Step S31: The Actor neural network selects the optimal injection and production parameters for CO2 huff and puff in the multi-stage fracturing horizontal well based on the injection and production environment status. Step S32: The reservoir simulator E300 executes the optimal injection and production parameters for the current time step to obtain production data of CO2 injection and production wells in multi-stage fractured horizontal wells. Step S33: Using the well production data obtained from simulation calculations, calculate the immediate reward for this injection-production adjustment time step, and proceed to the injection-production status of the next time step. Step S34: Use the instantaneous reward, injection parameters, and injection state characterization parameters as samples (s) to train the neural network. t ,a t ,r t ,s t+1 Stored in the experience pool; Step S35: The online neural network for Critic value evaluation assesses the cumulative reward corresponding to the injection and production parameters adopted in the current injection and production time step, and the target neural network for Critic value evaluation searches for the historical maximum cumulative reward in the experience pool for the current injection and production time step. Step S36: Iterate repeatedly until the online cumulative reward at the current time step converges to the maximum historical cumulative reward. Each injection and extraction time step is adjusted to obtain the optimal throughput injection and extraction parameters, thus obtaining the optimal injection and extraction system for the entire injection and extraction optimization phase.

8. The optimized CO2 huff-and-puff injection-production method for multi-stage fracturing horizontal wells according to claim 1, characterized in that, Step S4: Applying the obtained CO2 huff-and-puff injection-production optimization parameters of the multi-stage fractured horizontal well to a specific block, and analyzing the reasons for the high total net present value of the injection-production optimization scheme, specifically including: Step S41: In the field injection and production practice, the optimal injection and production parameters for CO2 huff and puff in multi-stage fractured horizontal wells are adopted. Step S42: Analyze the CO2 injection and production optimization scheme for multi-stage fractured horizontal wells, and the main mechanism that can improve the cumulative reward of the optimization target compared with the benchmark scheme.

9. The optimized CO2 huff-and-puff injection method for multi-stage fracturing horizontal wells according to claim 7, characterized in that, The injection and production status variables include: production time, daily gas injection volume, daily gas production volume, daily oil production volume, production gas-oil ratio, bottom hole flowing pressure, cumulative gas injection volume, cumulative gas production volume, and cumulative oil production volume.

10. The optimized CO2 huff-and-puff injection method for multi-stage fracturing horizontal wells according to claim 3, characterized in that, The optimization variables specifically include: gas injection rate, gas injection time, well shut-in time, backflow rate, and backflow time of the well.

11. A CO2 huff-and-puff injection-production optimization system for multi-stage fractured horizontal wells, employing the CO2 huff-and-puff injection-production optimization method for multi-stage fractured horizontal wells as described in any one of claims 1-10, characterized in that, The optimization system specifically includes: The reservoir simulation calculation model establishment module is used to establish the reservoir simulation calculation model required for CO2 huff and puff injection and production optimization in multi-stage fracturing horizontal wells; The constraint establishment module is used to establish the optimization objectives, variables, and constraints for CO2 injection and production in multi-stage fracturing horizontal wells; The injection-production optimization method framework establishment module is used to establish an injection-production optimization method framework for multi-stage fracturing horizontal wells based on PPO; The block application module is used to apply the obtained CO2 injection and production optimization parameters of multi-stage fractured horizontal wells to specific blocks and analyze the reasons for the high total net present value of the injection and production optimization scheme.

12. An optimized storage medium for CO2 injection and production in multi-stage fractured horizontal wells, characterized in that, The storage device contains a CO2 injection and production optimization system for a multi-stage fractured horizontal well as described in claim 11.

Images