A Machine Learning-Based Method and System for Evaluating Shale Oil Flowback Performance

Abstract

This invention discloses a machine learning-based method and system for evaluating the effectiveness of shale oil flowback, relating to the field of petroleum engineering technology. The method includes: determining the critical flow rate and critical pressure of the shale oil flowback regime to be evaluated based on well logging data from shale oil wells; dividing the target space into multiple category regions based on the critical flow rate and critical pressure; using a machine learning-based clustering method to divide multiple flowback data points of the shale oil flowback regime to be evaluated into multiple category regions; counting the number of flowback data points in each category region to obtain a quantity matrix; calculating a safe flowback evaluation index based on the quantity matrix and a preset weight matrix; and evaluating the flowback effectiveness of the shale oil flowback regime based on the safe flowback evaluation index. This invention alleviates the technical problem of existing technologies that cannot quantitatively and specifically guide the evaluation of shale oil flowback effectiveness.

Description

Technical Field

[0001] This invention relates to the field of petroleum engineering technology, specifically to a method and system for evaluating the flowback effect of shale oil based on machine learning. Background Technology

[0002] Currently, the focus of petroleum exploration and development research is shifting from conventional oil to unconventional oil. Shale oil is a key area of ​​unconventional oil and gas research and is gradually becoming a focus of actual field development in my country's oilfields. Shale oil development is highly challenging, as traditional technologies cannot achieve natural industrial oil flow. Key technologies such as horizontal well volumetric fracturing are required to enhance production efficiency. The hydraulic fracturing process requires the injection of large amounts of fracturing fluid and proppant to open and support hydraulic fractures. After the formation of the hydraulic fracture network or after well shut-in operations, fracturing fluid flowback is necessary to prevent formation damage. The flowback process directly affects the effectiveness of hydraulic fracturing, directly influencing the size of the effectively fractured fractures, and ultimately impacting production and economic benefits.

[0003] Optimizing the flowback regime design is a crucial part of shale oil production. Shale oil has low porosity and permeability, making fracturing stimulation technology challenging and costly; therefore, ensuring fracturing effectiveness is paramount. The flowback process is a complex coupling of formation, fractures, and wellbore, requiring comprehensive research from multiple perspectives to optimize the flowback regime. Therefore, a method is needed that comprehensively considers formation, fractures, and wellbore, providing a quantitative evaluation of flowback effectiveness to guide flowback operations and ensure fracturing results.

[0004] Some scholars have optimized flowback technology for tight sandstone gas reservoirs using experimental methods. Leveraging the colorless, odorless, and weakly soluble properties of elemental liquid nitrogen, liquid nitrogen fracturing experiments were conducted, and liquid nitrogen was also used to assist flowback. Water-lock damage was evaluated using core samples from various reservoirs in the Sichuan Basin, and gas permeability was measured. The results showed that higher pressure and longer pressurization time resulted in more severe water-lock damage during the flowback process; reducing the flowback pressure could reduce the degree of water-lock damage.

[0005] However, the above technical solutions are designed for tight sandstone gas reservoirs. Liquid nitrogen fracturing-assisted flowback methods have limited application in engineering practice, and the number of experiments is also small. No experiments have been conducted on shale oil wells. Only pressure parameters are considered, and various problems that may occur during the flowback process are not fully taken into account. The results are qualitative and cannot provide quantitative and targeted guidance for evaluating the flowback effect of shale oil. Summary of the Invention

[0006] The purpose of this invention is to provide a machine learning-based method and system for evaluating the shale oil flowback effect in order to solve at least one of the above-mentioned technical problems.

[0007] In a first aspect, embodiments of the present invention provide a machine learning-based method for evaluating the flowback effect of shale oil, comprising: determining the critical flow rate and critical pressure of the shale oil flowback regime to be evaluated based on logging data from the flowback wellbore of a shale oil well; the critical flow rate includes a maximum flow rate and a minimum flow rate, and the critical pressure includes a maximum bottom-hole flowing pressure and a minimum bottom-hole flowing pressure; dividing the target space into multiple category regions based on the critical flow rate and the critical pressure, using a machine learning-based clustering method; assigning multiple flowback data points of the shale oil flowback regime to be evaluated to the multiple category regions; counting the number of flowback data points in each category region to obtain a quantity matrix; calculating a safe flowback evaluation index based on the quantity matrix and a preset weight matrix; and evaluating the flowback effect of the shale oil flowback regime to be evaluated based on the safe flowback evaluation index.

[0008] Furthermore, the formula for calculating the minimum flow rate includes:

[0009]

[0010] in,

[0011] ρ f =ρ o f o +ρ w f w

[0012]

[0013] Q1 is the minimum flow rate, d s ρ is the diameter of the proppant particles. s ρ is the density of the proppant particles. f Let d be the fluid density in the wellbore. W C is the diameter of the return wellbore. D ρ is the drag coefficient. w ρ is the density of the water phase. o Let q be the density of the oil phase. w q is the aqueous phase flow rate. o This represents the oil phase flow rate.

[0014] Furthermore, the formula for calculating the maximum flow rate includes:

[0015]

[0016] Where Q2 is the maximum flow rate, n is the total number of wellbore holes, γ is the surface tension, μ is the fluid viscosity, and d s ρ is the diameter of the proppant particles. s ρ is the density of the proppant particles. f Let ρ be the fluid density in the wellbore, and ΔP be the effective stress on the proppant.

[0017] Further, determining the minimum bottom hole flowing pressure includes: calculating the flowback pressure of the flowback wellbore and the bubble point pressure of the shale oil based on the logging data; and determining the maximum value of the flowback pressure and the bubble point pressure of the shale oil as the minimum bottom hole flowing pressure.

[0018] Further, the flowback pressure is the sum of the wellhead pressure and the pressure drop of the flowback wellbore; wherein, the pressure drop includes:

[0019]

[0020] In the formula,

[0021]

[0022] p is the pressure inside the flowback wellbore, dp / dL is the pressure drop, L is the length of the flowback wellbore section, θ is the angle between the flowback wellbore and the horizontal direction, and R... e ρ is the Reynolds number. f Let d be the fluid density in the wellbore. W v is the diameter of the return wellbore. m q is the flow velocity of the mixture. w q is the aqueous phase flow rate. o This represents the oil phase flow rate.

[0023] Furthermore, the formula for calculating the bubble point pressure of shale oil includes:

[0024]

[0025] In the formula, R s For the dissolved gas-oil ratio, γ g γ0 is the relative density of dissolved gas, γ0 is the relative density of degassed crude oil, and p0 is the relative density of dissolved gas. exp The bubble point pressure of the shale oil is T, and the formation temperature is T.

[0026] Further, based on the quantity matrix and the preset weight matrix, the safety back-run evaluation index is calculated, including: calculating the Hadamard product of the quantity matrix and the preset weight matrix to obtain the product matrix; and calculating the safety back-run evaluation index based on the product matrix.

[0027] Secondly, embodiments of the present invention also provide a shale oil flowback effect evaluation system based on machine learning, comprising: a determination module, a partitioning module, a clustering module, a statistics module, a calculation module, and an evaluation module; wherein, the determination module is used to determine the critical flow rate and critical pressure of the shale oil flowback regime to be evaluated based on well logging data from the shale oil well flowback wellbore; the critical flow rate includes the maximum flow rate and the minimum flow rate, and the critical pressure includes the maximum bottom hole flowing pressure and the minimum bottom hole flowing pressure; the partitioning module is used to divide the target space into multiple category regions based on the critical flow rate and the critical pressure; the clustering module is used to partition multiple flowback data points of the shale oil flowback regime to be evaluated into the multiple category regions using a machine learning-based clustering method; the statistics module is used to count the number of flowback data points in each category region to obtain a quantity matrix; the calculation module is used to calculate a safe flowback evaluation index based on the quantity matrix and a preset weight matrix; and the evaluation module is used to evaluate the flowback effect of the shale oil flowback regime to be evaluated based on the safe flowback evaluation index.

[0028] Thirdly, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method provided in the embodiments of the present invention.

[0029] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the method provided in the embodiments of the present invention.

[0030] This invention provides a machine learning-based method and system for evaluating the shale oil flowback effect. It determines the acceptable flowback pressure and flow rate during the shale oil flowback process through well logging data, and combines the clustering principle of machine learning to form a flowback effect evaluation method, which alleviates the technical problem of existing technologies that cannot quantitatively and specifically guide the evaluation of shale oil flowback effect. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0032] Figure 1 A flowchart illustrating a machine learning-based method for evaluating shale oil flowback performance, provided as an embodiment of the present invention;

[0033] Figure 2 This is a schematic diagram of a wellbore proppant carrying method provided in an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of the stress on a crack proppant provided in an embodiment of the present invention;

[0035] Figure 4 A rectangular coordinate system diagram after partitioning is provided in an embodiment of the present invention;

[0036] Figure 5 Another rectangular coordinate system diagram after partitioning provided in an embodiment of the present invention;

[0037] Figure 6 This is a schematic diagram of a machine learning-based shale oil flowback effect evaluation system provided in an embodiment of the present invention. Detailed Implementation

[0038] 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. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0039] Example 1

[0040] Figure 1 This is a flowchart illustrating a machine learning-based method for evaluating the flowback effect of shale oil, provided by an embodiment of the present invention. Figure 1 As shown, the method specifically includes the following steps:

[0041] Step S102: Based on the logging data of the shale oil well flowback wellbore, determine the critical flow rate and critical pressure of the shale oil flowback regime to be evaluated; the critical flow rate includes the maximum flow rate and the minimum flow rate, and the critical pressure includes the maximum bottom hole flowing pressure and the minimum bottom hole flowing pressure.

[0042] Step S104: Based on critical flow rate and critical pressure, the target space is divided into multiple category regions with flow rate and pressure as dimensions.

[0043] Step S106: Using a clustering method based on machine learning, the multiple return data points of the shale oil return system to be evaluated are divided into multiple category regions.

[0044] Step S108: Count the number of returned data points in each category region to obtain the quantity matrix.

[0045] Step S110: Calculate the safety backflow evaluation index based on the quantity matrix and the preset weight matrix.

[0046] Step S112: Evaluate the flowback effect of the shale oil flowback system to be evaluated based on the safety flowback evaluation index.

[0047] Specifically, in step S102, the determination process for the maximum bottomhole flowing pressure is as follows: During hydraulic fracturing, high-pressure fluid is used to break the formation and generate hydraulic fractures, and proppant is carried by injected proppant fluid for hydraulic fracture support. During well shut-in, the pressure in the fracture continuously dissipates, and the proppant settles within the fracture. If the pressure dissipation in the fracture is insufficient during well opening, resulting in a fracture pressure higher than the fracture closure pressure, the fracture will be in an open state. Without the support of fracture stress, the proppant will be in a loose, settled state, leading to easy backflow of the proppant during well opening and flowback. Therefore, the bottomhole flowing pressure during well opening is used as one of the indicators for evaluating the flowback effect. The fracture fluid pressure during well opening should be lower than the fracture closure pressure; if it is higher, an unfavorable flowback regime is in place. For ease of application and practical considerations, the fracture fluid pressure can be considered equal to the bottomhole flowing pressure, and the fracture closure pressure can be considered equal to the minimum horizontal principal stress.

[0048] In actual flowback operations, the minimum horizontal principal stress can be directly obtained from well logging, while the bottomhole flowing pressure needs to be calculated from the wellhead pressure. The calculation formula is as follows:

[0049]

[0050] In the formula, p wfo The bottom hole flowing pressure is MPa; MD is the vertical depth of the well, in meters; p H The wellhead pressure is expressed in MPa; h w This represents the head loss due to friction along the friction path, expressed in meters (m).

[0051] Then, the maximum bottom hole pressure P2 allowed by the optimized design is determined based on the bottom hole flowing pressure.

[0052] Figure 2 This is a schematic diagram of proppant carrying in a wellbore according to an embodiment of the present invention. Figure 2 As shown, if residual proppant remains in the wellbore during shale oil fracturing or if sand production occurs during flowback, the residual proppant will adversely affect the shale oil flowback process. For example, proppant deposited at the bottom of the wellbore and at the build-up area will obstruct the flow channels, causing wellbore blockage or even halting production. Moving proppant will also erode the wellbore and downhole instruments, shortening their lifespan. Therefore, evaluating the flowback effect requires assessing whether the flowback rate exceeds the critical settling flow rate of the proppant in the wellbore, ensuring that the proppant is carried out of the wellhead in a timely manner. If the flow rate is less than this, the proppant cannot be carried out, which is considered an unfavorable flowback system.

[0053] Specifically, the formula for calculating the minimum flow rate includes:

[0054]

[0055] in,

[0056] ρ f =ρ o f o +ρ w f w

[0057]

[0058] Q1 is the minimum flow rate, d s ρ represents the diameter of the proppant particles, in meters (m). s This refers to the density of proppant particles, in kg / m³. 3 ;ρ f The density of the fluid in the wellbore is expressed in kg / m³. 3 ;d W The diameter of the return wellbore is in meters (m); C D ρ is the drag coefficient, dimensionless; w ρ is the density of the water phase. o Let q be the density of the oil phase. w q is the aqueous phase flow rate. o This represents the oil phase flow rate.

[0059] Among them, C D The possible values ​​are shown in Table 1:

[0060] Table 1. Traction Coefficient Values

[0061]

[0062] in,

[0063]

[0064] In the formula, u f The fluid velocity in the wellbore is expressed in m / s; μ f The viscosity of the fluid in the wellbore is expressed in Pa·s.

[0065] Figure 3 This is a schematic diagram of the stress on a crack proppant according to an embodiment of the present invention. Figure 3 As shown, it is necessary to determine whether the current backflow regime can prevent the proppant in the crack from being flushed out of the crack. If the current backflow regime is greater than the flow rate, the proppant in the crack will flow back, which is an unfavorable phenomenon.

[0066] Specifically, the formula for calculating the maximum flow rate includes:

[0067]

[0068] Where Q2 is the maximum flow rate, n is the total number of wellbore holes, γ is the surface tension (0.03 N / m), μ is the fluid viscosity (Pa·s), and d s ρ is the diameter of the proppant particles. s ρ is the density of the proppant particles. f Let ρ be the fluid density in the wellbore, and ΔP be the effective stress on the proppant.

[0069] Specifically, in step S102, determining the minimum bottom hole flowing pressure includes:

[0070] Based on well logging data, the flowback pressure in the flowback wellbore and the bubble point pressure of shale oil were calculated respectively.

[0071] The maximum value between the flowback pressure and the shale oil bubble point pressure is determined as the minimum bottom hole flowing pressure.

[0072] In this embodiment of the invention, a certain backflow pressure is required to ensure smooth backflow. This backflow pressure is the minimum bottom hole flowing pressure. Therefore, this index is established to determine whether the pressure during the backflow process is greater than this backflow pressure. If it is less than this backflow pressure, the backflow fluid cannot return to the wellhead, and the backflow cannot proceed smoothly. In the case of multiphase and homogeneous oil-water flow, the backflow pressure is the sum of the wellhead pressure and pressure drop in the backflow wellbore; wherein, the pressure drop includes:

[0073]

[0074] In the formula,

[0075]

[0076] p is the pressure inside the flowback wellbore, dp / dL is the pressure drop, L is the length of the flowback wellbore section, θ is the angle between the flowback wellbore and the horizontal direction, and R... e ρ is the Reynolds number. f Let d be the fluid density in the wellbore. W v is the diameter of the return wellbore. m q is the flow velocity of the mixture. w q is the aqueous phase flow rate. o This represents the oil phase flow rate.

[0077] Meanwhile, to ensure that the crude oil in the flowback wellbore does not undergo phase change during the flowback process, the bottomhole flowing pressure needs to be greater than the shale oil bubble point pressure. The formula for calculating the shale oil bubble point pressure includes:

[0078]

[0079] In the formula, R s For the dissolved gas-oil ratio, γ gγ0 is the relative density of dissolved gas (air = 1), γ0 is the relative density of degassed crude oil (water = 1), p exp ρ represents the bubble point pressure of shale oil, in MPa; T represents the formation temperature. All the above parameters were obtained from field fluid PVT physical property parameter experiments.

[0080] In an optional embodiment provided by this invention, step S104 includes:

[0081] First, a rectangular coordinate system is plotted with flow rate as the horizontal axis and pressure as the vertical axis, resulting in a target space with flow rate and pressure as dimensions. Then, the rectangular coordinate system is divided into 9 categories, with critical flow rates Q1 and Q2 and critical pressures P1 and P2 as boundaries. The coordinates of the center points of each category are as follows: Figure 4 As shown.

[0082] Then, in the Cartesian coordinate system diagram, calculate the Euclidean distance between each back-row data point and each center point, and classify each back-row data point into the category region of the nearest (i.e., the one with the smallest Euclidean distance) center point.

[0083] Then, the number of returned data points within each category region is counted to obtain a count matrix. Optionally, the count matrix M is a 3×3 matrix.

[0084] In an optional embodiment provided by this invention, the preset weight matrix is ​​as follows:

[0085]

[0086] Specifically, step S110 also includes the following steps:

[0087] Step S1101: Calculate the Hadamard product of the scalar matrix and the preset weight matrix to obtain the product matrix; specifically, the product matrix...

[0088] Step S1102: Calculate the safety backtracking evaluation index based on the product matrix. Specifically, the calculation formula for the safety backtracking evaluation index F is as follows:

[0089]

[0090] In the formula, h ij These are elements of the product matrix H.

[0091] In this embodiment of the invention, the closer the safety flowback evaluation index F value is to 9, the better the flowback system. However, due to practical engineering limitations and the need to shorten the flowback operation period as much as possible, the F value of some shale oil wells cannot be equal to 9. Based on actual flowback operation experience, an F value greater than 2 is within the acceptable range for practical engineering.

[0092] As described above, the embodiments of the present invention provide a machine learning-based method for evaluating the shale oil flowback effect. By using well logging data to determine the acceptable flowback pressure and flow rate during the shale oil flowback process, and combining the machine learning clustering principle, a flowback effect evaluation method is formed, which alleviates the technical problem of existing technologies that cannot quantitatively and specifically guide the evaluation of shale oil flowback effect.

[0093] Example 2

[0094] Taking a shale oil well as an example, this well underwent fracturing testing and flowback operations in 2023. Well logging data shows that the average minimum horizontal principal stress of the well was 73.5 MPa, meaning the maximum bottomhole flowing pressure P2 = 73.5 MPa; the boundary flow rate of the proppant deposition in the wellbore, i.e., the minimum flowback flow rate Q1 = 68 m³ / s. 3 / d; Critical backflow rate of crack proppant, i.e., maximum backflow rate Q2 = 104 m³ / d. 3 / d, plot the actual return flow rate and pressure points on a rectangular coordinate system with flow rate as the horizontal axis and pressure as the vertical axis, and divide the data into the nine sections mentioned above, such as... Figure 5 As shown.

[0095] According to the method provided in the embodiments of the present invention, the Euclidean distance between each point in the coordinate graph and each center point is calculated, and each back-row data point is assigned to the category of the nearest (smallest Euclidean distance) center point.

[0096] Perform cluster analysis and record the number of returned data points in each region as a 3×3 matrix M.

[0097]

[0098] F>2 is within the safe flowback range that is actually acceptable in engineering practice, the flowback system is reasonable, and the well did not experience any adverse production phenomena such as sand production, production stoppage, or sudden reduction in output during the flowback process.

[0099] Example 3

[0100] Figure 6 This is a schematic diagram of a machine learning-based shale oil flowback performance evaluation system provided by an embodiment of the present invention. Figure 6 As shown, the system includes: a determination module 10, a partitioning module 20, a clustering module 30, a statistics module 40, a calculation module 50, and an evaluation module 60.

[0101] Specifically, module 10 is used to determine the critical flow rate and critical pressure of the shale oil flowback regime to be evaluated based on the logging data of the shale oil well flowback wellbore. The critical flow rate includes the maximum flow rate and the minimum flow rate, and the critical pressure includes the maximum bottom hole flowing pressure and the minimum bottom hole flowing pressure.

[0102] The partitioning module 20 is used to divide the target space into multiple category regions based on critical flow and critical pressure.

[0103] Clustering module 30 is used to divide multiple return data points of the shale oil return system to be evaluated into multiple category regions using machine learning-based clustering methods.

[0104] The statistics module 40 is used to count the number of returned data points in each category area to obtain a quantity matrix.

[0105] The calculation module 50 is used to calculate the safety back-out evaluation index based on the quantity matrix and the preset weight matrix.

[0106] Evaluation module 60 is used to evaluate the effectiveness of the shale oil return system based on the safety return evaluation index.

[0107] Specifically, the calculation module 50 is also used to: calculate the Hadamard product of the quantity matrix and the preset weight matrix to obtain the product matrix; and calculate the safety back-out evaluation index based on the product matrix.

[0108] The present invention also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method provided in the embodiments of the present invention.

[0109] The present invention also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the method provided in the embodiments of the present invention.

[0110] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0111] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A machine learning-based method for evaluating the flowback effect of shale oil, characterized in that, include: Based on well logging data from shale oil well flowback wellbore, the critical flow rate and critical pressure of the shale oil flowback regime to be evaluated are determined; the critical flow rate includes the maximum flow rate and the minimum flow rate, and the critical pressure includes the maximum bottom hole flowing pressure and the minimum bottom hole flowing pressure. Based on the critical flow rate and the critical pressure, the target space is divided into multiple category regions with flow rate and pressure as the dimensions. Using a machine learning-based clustering method, multiple return data points of the shale oil return regime to be evaluated are divided into multiple category regions; The number of returned data points within each category region is counted to obtain a quantity matrix; Based on the quantity matrix and the preset weight matrix, calculate the safety backflow evaluation index; Based on the aforementioned safety return evaluation indicators, the return effect of the shale oil return system to be evaluated is assessed. The formula for calculating the minimum flow rate includes: in, Q1 is the minimum flow rate, d s The diameter of the proppant particles, For proppant particle density, Let d be the fluid density in the wellbore. W The diameter of the return wellbore. This is the drag coefficient. The density of the aqueous phase, The density of the oil phase is... For water phase flow rate, ρ is the oil phase flow rate; g is the gravitational acceleration; The formula for calculating the maximum flow rate includes: Where Q2 is the maximum flow rate, n is the total number of wellbore holes, γ is the surface tension, μ is the fluid viscosity, and d s The diameter of the proppant particles, For proppant particle density, Let ρ be the fluid density in the wellbore, and ΔP be the effective stress on the proppant. Determining the minimum bottom hole flowing pressure includes: Based on the well logging data, the flowback pressure and shale oil bubble point pressure of the flowback wellbore are calculated respectively. The maximum value of the backflow pressure and the shale oil bubble point pressure is determined as the minimum bottom hole flowing pressure.

2. The method according to claim 1, characterized in that: The return pressure is the sum of the wellhead pressure and the pressure drop of the return wellbore; wherein, the pressure drop includes: In the formula, p is the pressure inside the flowback wellbore, dp / dL is the pressure drop, L is the length of the flowback wellbore section, θ is the angle between the flowback wellbore and the horizontal direction, and R... e The Reynolds number is... Let d be the fluid density in the wellbore. W v is the diameter of the return wellbore. m The flow rate of the mixture, For water phase flow rate, This represents the oil phase flow rate.

3. The method according to claim 1, characterized in that: The formula for calculating the bubble point pressure of shale oil includes: ； In the formula, For dissolved gas-oil ratio, The relative density of the dissolved gas. The relative density of degassed crude oil. The bubble point pressure of the shale oil is T, and the formation temperature is T.

4. The method according to claim 1, characterized in that: Based on the quantity matrix and the preset weight matrix, the safety backflow evaluation index is calculated, including: Calculate the Hadamard product of the scalar matrix and the preset weight matrix to obtain the product matrix; The safety backflow evaluation index is calculated based on the product matrix.

5. A machine learning-based shale oil flowback performance evaluation system, characterized in that, include: The module includes a determination module, a partitioning module, a clustering module, a statistics module, a calculation module, and an evaluation module; among which, The determining module is used to determine the critical flow rate and critical pressure of the shale oil flowback regime to be evaluated based on the logging data of the shale oil well flowback wellbore; the critical flow rate includes the maximum flow rate and the minimum flow rate, and the critical pressure includes the maximum bottom hole flowing pressure and the minimum bottom hole flowing pressure; The partitioning module is used to divide the target space into multiple category regions based on the critical flow rate and the critical pressure. The clustering module is used to divide multiple return data points of the shale oil return system to be evaluated into multiple category regions using a machine learning-based clustering method. The statistics module is used to count the number of returned data points in each category area to obtain a quantity matrix; The calculation module is used to calculate the safety back-out evaluation index based on the quantity matrix and the preset weight matrix; The evaluation module is used to evaluate the return effect of the shale oil return system to be evaluated based on the safety return evaluation index. The formula for calculating the minimum flow rate includes: in, Q1 is the minimum flow rate, d s The diameter of the proppant particles, For proppant particle density, Let d be the fluid density in the wellbore. W The diameter of the return wellbore. This is the drag coefficient. The density of the aqueous phase, The density of the oil phase is... For water phase flow rate, ρ is the oil phase flow rate; g is the gravitational acceleration; The formula for calculating the maximum flow rate includes: Where Q2 is the maximum flow rate, n is the total number of wellbore holes, γ is the surface tension, μ is the fluid viscosity, and d s The diameter of the proppant particles, For proppant particle density, Let ρ be the fluid density in the wellbore, and ΔP be the effective stress on the proppant. Determining the minimum bottom hole flowing pressure includes: Based on the well logging data, the flowback pressure and shale oil bubble point pressure of the flowback wellbore are calculated respectively. The maximum value of the backflow pressure and the shale oil bubble point pressure is determined as the minimum bottom hole flowing pressure.

6. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method as described in any one of claims 1-4.

Images