Method, device and equipment for determining energy supplementing liquid amount before horizontal well refracturing

Abstract

The invention provides a method, a device and equipment for determining the amount of energy supplementing liquid before re-fracturing of a horizontal well. Comprising the following steps: quickly building a numerical simulation model, massively building a data sample set, screening key characteristic parameters which influence a pressure recovery level, training a machine learning model to obtain a pressure level prediction model before and after energy supplementation of a horizontal well, building an agent model for optimizing the re-fracturing energy supplementation parameters of the horizontal well, and optimizing the re-fracturing energy supplementation parameters of the horizontal well by taking maximization of the pressure recovery level as a target. And an optimization mathematical model is established, an optimization algorithm is called, and the corresponding relation between the pre-pressure energy supplementing liquid amount and the pressure recovery level under different geological, engineering and production conditions is obtained. According to the method, the optimal energy supplementing liquid amount parameter before the horizontal well refracturing is rapidly and accurately determined, the problems that the productivity numerical simulation means is single, key characteristic parameters cannot be clarified and the like are solved, on one hand, a large number of expensive simulation tests can be reduced, and economic benefits are remarkably improved; and on the other hand, the calculation efficiency is extremely high, and the test time cost can be greatly reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas field development, and in particular to a method, device, equipment and storage medium for determining the amount of replenishing fluid before repeated fracturing of a horizontal well. Background Art

[0002] At present, unconventional oil reservoirs have become an important strategic inherited energy source. Due to their physical properties, they generally have no natural production capacity, and fracturing transformation must be implemented to obtain industrial oil flow. "Horizontal well + volume fracturing" is an effective measure to achieve commercial exploitation. However, due to geological, engineering and production factors, the development process has led to serious declines in single well production and reservoir energy exhaustion. Repeated fracturing of horizontal wells is an effective measure to increase single well production and extend the stable production cycle. Studies have shown that recharging the reservoir before repeated fracturing of horizontal wells can effectively raise the formation energy and thus restore the production of single wells. Therefore, it is very important to optimize and determine the amount of recharging fluid before repeated fracturing of horizontal wells.

[0003] At present, the determination and optimization of pre-re-fracturing energy replenishment parameters for horizontal wells are mainly based on the principle of material balance. The energy replenishment fluid volume is designed with reference to the produced fluid volume after the initial fracturing. The production experience of field staff is mainly used. The optimization process does not consider the influence of different reservoir geological conditions and fracturing construction parameters. The optimization results do not directly reflect the recovery of reservoir pressure, and there is a problem of not being able to obtain the optimal parameters.

[0004] In the long run, the optimization of the amount of replenishing fluid before repeated fracturing basically relies on the material balance equation, which requires a large amount of simulation tests and is costly. The optimization process is extremely cumbersome, complex, time-consuming and labor-intensive. Summary of the invention

[0005] In order to overcome the defects of the existing technology, a method, device and equipment for determining the amount of energy replenishment fluid before repeated fracturing of horizontal wells are proposed. By integrating the application of "reservoir numerical simulation method + machine learning theory", a proxy model between geological, initial fracturing and energy replenishment parameters and formation pressure maintenance level is constructed, which solves the problem of optimizing the amount of energy replenishment fluid before repeated fracturing and overcomes the technical difficulty that the existing methods cannot accurately predict the formation pressure maintenance level. It is of great significance to guide the design of energy replenishment parameters before repeated fracturing.

[0006] To achieve the above objectives, some embodiments of the present application provide the following aspects:

[0007] In a first aspect, some embodiments of the present application further provide a method for determining the amount of replenishing fluid before refracturing of a horizontal well, comprising:

[0008] Step a: Collect the geological, engineering, and production dynamic data of the target area. Based on the collected data, use geological modeling algorithms to establish a geological model. Import the geological model, rock, and fluid physical property data into a reservoir numerical simulator for simulation. Taking a typical well in the target area as an example, set the horizontal well, the primary hydraulic fracture, and the corresponding operating regime to form a reservoir numerical simulation model coupled with the hydraulic fracture. On this basis, by calling the sequential Gaussian data generator, generate n sets of random combinations of geological and primary fracturing parameter schemes, and then construct a large number of numerical simulation models of fractured horizontal wells;

[0009] Step b: Based on each numerical simulation model established in Step a, call the simulator to complete the simulation calculation. First, calculate the production dynamics for a period of time after the primary fracturing. On this basis, set different energy supplement fluid volume parameters and calculate the reservoir production dynamics to complete the pre-fracturing energy supplement simulation. After the calculation is completed, count the basic parameters and calculation results of the model, collect the reservoir geological data of each model to establish a geological characteristic parameter set, collect the primary fracturing engineering parameter data of each model to establish an engineering characteristic parameter set, and collect the dynamic production parameters of each model to establish a production characteristic parameter set. Finally, set all the geological, engineering, and production characteristic parameters and the energy supplement fluid volume parameter as input variables, and the average reservoir pressure after the energy supplement ends as the output variable to complete the construction of the learning sample set;

[0010] Step c: Screen out the key characteristic parameters that affect the pressure recovery level for the learning sample set constructed in Step b;

[0011] Step d: Use the screened key characteristic parameters and the energy supplement fluid volume parameter as input variables, and the pressure level after the pre-fracturing energy supplement as the output variable to construct a data set. Use the data set to train the initial machine learning model to obtain a prediction model for the pressure level after the pre-fracturing energy supplement of the horizontal well;

[0012] Step e: Use the prediction model for the pressure level after the pre-fracturing energy supplement of the horizontal well obtained in Step d to obtain the pressure recovery level corresponding to the energy supplement fluid volume before repeated fracturing of the horizontal well under different geological, engineering, and production conditions. With the maximization of the pressure recovery level as the goal, determine the energy supplement fluid volume before repeated fracturing of the horizontal well.

[0013] In a second aspect, some embodiments of the present application also provide a device for determining the energy supplement fluid volume before repeated fracturing of a horizontal well. The device includes:

[0014] A numerical simulation model building module, which is used to collect data. According to the collected data, a geological model is established by using a geological modeling algorithm. The geological model, rock and fluid physical property data are imported into a reservoir numerical simulator for simulation. A horizontal well, a primary hydraulic fracture and corresponding working systems are set to form a reservoir numerical simulation model coupled with a hydraulic fracture. By calling a sequential Gaussian data generator, n groups of random geological and primary fracturing parameter combination schemes are generated and input into the reservoir numerical simulation model coupled with the hydraulic fracture, thereby constructing several numerical simulation models of fractured horizontal wells; the data includes: geological data, engineering parameter data and production dynamic data of the target area;

[0015] A data sample set building module, which is used to call a simulator to complete simulation calculations based on each of the constructed numerical simulation models of fractured horizontal wells. First, the production dynamics in a preset time period after primary fracturing are calculated. On this basis, different energy supplement fluid volume parameters are set to calculate the reservoir production dynamics and complete the pre-fracture energy supplement simulation. After the calculation is completed, the basic parameters and calculation results of the numerical simulation models of fractured horizontal wells are statistically analyzed. The geological feature parameter set is established by collecting the reservoir geological data of each model, the engineering feature parameter set is established by collecting the primary fracturing engineering parameter data of each model, and the production feature parameter set is established by collecting the dynamic production parameters of each model. Finally, all geological feature parameters, engineering feature parameters, production feature parameters and energy supplement fluid volume parameters are set as input variables, and the average reservoir pressure after energy supplement is set as the output variable to complete the construction of the learning sample set;

[0016] The data sample set building module is also used to screen out the key feature parameters that affect the pressure recovery level for the constructed learning sample set;

[0017] The data sample set building module is also used to use the screened key feature parameters and the energy supplement fluid volume parameters as input variables, and the pressure level after pre-fracture energy supplement as the output variable to construct a data set, and use the data set to train an initial machine learning model to obtain a prediction model for the pressure level after pre-fracture energy supplement of a horizontal well;

[0018] A pre-fracture energy supplement fluid volume determination module, which is used to use the prediction model for the pressure level after pre-fracture energy supplement of a horizontal well to obtain the pressure recovery level corresponding to the pre-fracture energy supplement fluid volume under different geological, engineering and production conditions, and determine the pre-fracture energy supplement fluid volume for repeated fracturing of a horizontal well with the goal of maximizing the pressure recovery level.

[0019] In a third aspect, some embodiments of the present application further provide a computer device, which includes: one or more processors; and a memory storing computer program instructions, and when the computer program instructions are executed, the processors execute the above method.

[0020] Fourthly, some embodiments of the present application further provide a computer-readable medium, on which computer program instructions are stored, and the computer program instructions can be executed by a processor to implement the method as described above.

[0021] Compared with the prior art, the method for optimizing and determining the energy supplement liquid volume before refracturing of horizontal wells provided by the embodiments of the present application constructs a rich sample set through production capacity numerical simulation, and quickly and accurately determines the optimal energy supplement liquid volume parameters before refracturing of horizontal wells by combining key feature screening and machine learning algorithms, solving the problems of single means of production capacity numerical simulation and inability to clarify key feature parameters, and accurately and efficiently obtaining the optimal energy supplement liquid volume parameters before fracturing. The method for determining the energy supplement liquid volume parameters before refracturing provided by the present invention can, on the one hand, reduce a large number of high-cost simulation tests and significantly improve economic benefits; on the other hand, the traditional optimization process is extremely cumbersome and complex, and time-consuming and laborious, while the calculation efficiency of the present invention is extremely high, which can greatly reduce the time cost of testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings forming a part of this invention are used to provide a further understanding of the invention. The schematic embodiments and descriptions thereof of the invention are used to explain the invention and do not constitute an improper limitation of the invention. In the drawings:

[0023] Figure 1 It is a schematic diagram for generating a basic proxy model provided by an embodiment of the present application;

[0024] Figure 2 It is the value distribution of typical characteristic parameters provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. It should be noted that, without conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0026] The following detailed descriptions are all exemplary descriptions, aiming to provide further detailed descriptions of the present invention. Unless otherwise specified, all technical terms adopted by the present invention have the same meaning as commonly understood by those of ordinary skill in the art to which the present invention belongs. The terms used in the present invention are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention.

[0027] A method for determining the energy supplement liquid volume before refracturing of horizontal wells includes:

[0028] Step a: Collect geological, engineering, and production dynamic data of the target area. Based on the collected data, use geological modeling algorithms to establish a geological model. Import the geological model, rock, and fluid property data into a reservoir numerical simulator for simulation. Taking a typical well in the target area as an example, set the horizontal well, primary hydraulic fracture, and corresponding operating regime to form a reservoir numerical simulation model coupled with hydraulic fractures. On this basis, by calling a sequential Gaussian data generator, generate n sets of random combinations of geological and primary fracturing parameter schemes, and then construct a large number of numerical simulation models for fractured horizontal wells;

[0029] Step b: Based on each numerical simulation model established in Step a, call the simulator to complete the simulation calculation. First, calculate the production dynamics for a period of time after the primary fracturing. On this basis, set different energy supplement fluid volume parameters and calculate the reservoir production dynamics to complete the pre-fracturing energy supplement simulation. After the calculation is completed, statistically analyze the basic parameters and calculation results of the model. Collect the reservoir geological data of each model to establish a geological characteristic parameter set, collect the primary fracturing engineering parameter data of each model to establish an engineering characteristic parameter set, and collect the dynamic production parameters of each model to establish a production characteristic parameter set. Finally, set all geological, engineering, production characteristic parameters, and energy supplement fluid volume parameters as input variables, and set the average reservoir pressure after the energy supplement ends as the output variable to complete the construction of the learning sample set;

[0030] Step c: Screen out the key characteristic parameters that affect the pressure recovery level for the learning sample set constructed in Step b;

[0031] Step d: Use the screened key characteristic parameters and energy supplement fluid volume parameters as input variables, and the pressure level after pre-fracturing energy supplement as the output variable to construct a data set. Use the data set to train an initial machine learning model to obtain a prediction model for the pressure level after pre-fracturing energy supplement of horizontal wells;

[0032] Step e: Use the prediction model for the pressure level after pre-fracturing energy supplement of horizontal wells obtained in Step d to obtain the pressure recovery levels corresponding to different geological, engineering, and production conditions for the pre-fracturing energy supplement fluid volume. With the goal of maximizing the pressure recovery level, determine the pre-fracturing energy supplement fluid volume for horizontal well refracturing.

[0033] In some embodiments of the present application, the geological data in Step a includes but is not limited to: isopach maps of sand body thickness distribution, effective thickness distribution, porosity distribution, permeability distribution, original formation pressure, temperature, pressure coefficient data, original oil / gas / water distribution, original oil-water interface and oil-gas interface, geological reserve reports, etc. for the target horizon in the study block; rock compressibility, fluid and rock laboratory analysis reports; engineering parameter data includes but is not limited to horizontal well footage, fracture half-length, fracture section spacing, fracturing fluid volume injected into the formation, sand addition volume, fracture monitoring and interpretation data, etc.; production dynamic data includes but is not limited to: oil production, liquid production, bottom-hole flowing pressure, etc.

[0034] In some embodiments of the present application, step c specifically includes: for the learning sample library constructed in step b, training is performed using models such as random forest and Pearson correlation analysis; taking random permutation importance as the evaluation index, and screening out the key characteristic parameters affecting the pressure recovery level by comprehensively evaluating the results of models such as random forest and Pearson correlation analysis.

[0035] The characteristic parameters affecting the pressure recovery level after energy supplementation before well testing in horizontal wells include the average reservoir pressure value, the average reservoir thickness value, the average permeability value, the average porosity value, the average oil saturation value, the horizontal well footage value, the average fracture spacing, the fracture length value, the bottom hole flowing pressure value, and the energy supplementation fluid volume value;

[0036] For different types of oil reservoirs and different horizontal wells, the key characteristic parameters affecting the pressure recovery level are different. In order to make the prediction model more accurate, it is necessary to screen out the key characteristic parameters.

[0037] In some embodiments of the present application, preferably, the machine learning models in step d include BP neural network, support vector machine, and gradient boosting decision tree. Other conventional machine learning models can be added or used for simple replacement, and the steps are similar, so they will not be elaborated here.

[0038] (1) Using the screened key characteristic parameters and energy supplementation fluid volume parameters as input variables, and the pressure level after energy supplementation before well testing as the output variable, a data set is constructed

[0039] (2) Randomly divide the entire data set into a training set and a prediction set;

[0040] (3) Training different machine learning algorithm models including BP neural network, support vector machine, gradient boosting decision tree, etc. with the divided training set;

[0041] (4) Calculating the coefficient of determination between the actual pressure level of the training set and the pressure level calculated by the model;

[0042] (5) Selecting the machine learning algorithm model with the largest coefficient of determination to construct a prediction model for the pressure level after energy supplementation before well testing in horizontal wells.

[0043] In some embodiments of the present application, preferably, after step d(5), it further includes: calculating the error of the prediction model to complete the construction of the prediction model for the pressure level after energy supplementation before well testing in horizontal wells. The specific steps are as follows:

[0044] Input the characteristic parameters of the test set and the energy supplement liquid volume parameters into the pressure level prediction model, calculate the model output (i.e., the pressure recovery level), and calculate the errors between the digital simulation oil production increment and the oil production increment calculated by the model of the test set, including the determination coefficient, relative error, and absolute error; if the model error is small, use the selected key characteristic parameters as input variables to complete the pressure level prediction model after energy supplement before horizontal well fracturing; if the model error is large, check the accuracy of the learning database. If the data is accurate, adjust the threshold of the key parameter model to re-screen the key parameters, and at the same time re-establish the pressure recovery level prediction model to calculate the model error again.

[0045] In some embodiments of the present application, step e constructs a surrogate model for optimizing the energy supplement parameters of horizontal well refracturing. With the maximization of the pressure recovery level as the goal, an optimization mathematical model is established, and an optimization algorithm is called to obtain the corresponding relationship between the energy supplement liquid volume before fracturing and the pressure recovery level under different geological, engineering, and production conditions.

[0046] An apparatus for determining the energy supplement liquid volume before horizontal well refracturing, the apparatus includes:

[0047] A numerical simulation model construction module, configured to collect data, establish a geological model using a geological modeling algorithm according to the collected data, import the geological model, rock, and fluid physical property data into a reservoir numerical simulator for simulation, set the horizontal well, the primary fracturing hydraulic fracture, and the corresponding working system to form a reservoir numerical simulation model coupled with the hydraulic fracture, and generate n groups of random geological and primary fracturing parameter combination schemes by calling a sequential Gaussian data generator and input them into the reservoir numerical simulation model coupled with the hydraulic fracture, thereby constructing several fracturing horizontal well numerical simulation models; the data includes: geological data, engineering parameter data, and production dynamic data of the target area;

[0048] A data sample set construction module, configured to, based on each of the constructed fracturing horizontal well numerical simulation models, call the simulator to complete the simulation calculation. First, calculate the production dynamics in a preset time period after the primary fracturing. On this basis, set different energy supplement liquid volume parameters, calculate the reservoir production dynamics, and complete the energy supplement simulation before fracturing. After the calculation is completed, count the basic parameters and calculation results of the fracturing horizontal well numerical simulation models, collect the geological feature parameter sets by collecting the reservoir geological data of each model, collect the engineering feature parameter sets by collecting the primary fracturing engineering parameter data of each model, collect the dynamic production parameters of each model to establish a production feature parameter set. Finally, set all the geological feature parameters, engineering feature parameters, production feature parameters, and energy supplement liquid volume parameters as input quantities, and set the average reservoir pressure after the energy supplement is completed as the output quantity to complete the construction of the learning sample set;

[0049] The data sample set construction module is further configured to screen out the key characteristic parameters that affect the pressure recovery level for the constructed learning sample set;

[0050] The data sample set construction module is further configured to use the selected key feature parameters and the energy supplement liquid volume parameters as input variables, and the pressure level after energy supplement before fracturing as the output variable to construct a data set, and use the data set to train an initial machine learning model to obtain a prediction model for the pressure level after energy supplement before fracturing of horizontal wells;

[0051] The energy supplement liquid volume determination module before fracturing is used to use the prediction model for the pressure level after energy supplement before fracturing of horizontal wells to obtain the pressure recovery level corresponding to the energy supplement liquid volume before fracturing under different geological, engineering and production conditions, and determine the energy supplement liquid volume before repeated fracturing of horizontal wells with the maximization of the pressure recovery level as the goal.

[0052] The beneficial effects of the technical solution provided by the implementation of the present invention are as follows:

[0053] A method for optimizing and determining the energy supplement liquid volume before repeated fracturing of horizontal wells is provided. This method combines production capacity numerical simulation, machine learning and intelligent optimization algorithms to quickly and accurately determine the optimal energy supplement liquid volume parameters before repeated fracturing of horizontal wells. Combining production capacity numerical simulation with optimization algorithms solves the problems of single production capacity numerical simulation means and inability to clarify key feature parameters, and accurately and efficiently obtains the optimal energy supplement liquid volume parameters before fracturing. The method for determining the energy supplement liquid volume parameters before repeated fracturing provided by the present invention can, on the one hand, reduce a large number of high-cost simulation tests and significantly improve economic benefits; on the other hand, the traditional optimization process is extremely cumbersome and time-consuming, while the calculation efficiency of the present invention is extremely high, which can greatly reduce the time cost of testing.

[0054] As is known by technical common sense, the present invention can be implemented through other embodiments that do not deviate from its spiritual essence or necessary features. Therefore, the above-disclosed embodiments are illustrative in all aspects and are not the only ones. All changes within the scope of the present invention or within the scope equivalent to the present invention are encompassed by the present invention.

[0055] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0056] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0057] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0058] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific embodiments of the present invention, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. A method for determining the energy supplement liquid volume before refracturing in horizontal wells, characterized in that Including: Step a: Collect data, including geological data, engineering parameter data, and production performance data of the target area. According to the collected data, establish a geological model using a geological modeling algorithm, import the geological model, rock, and fluid physical property data into a reservoir numerical simulator for simulation, set horizontal wells, primary hydraulic fracture, and corresponding operating systems to form a reservoir numerical simulation model coupled with hydraulic fractures. By calling a sequential Gaussian data generator, generate n sets of random geological and primary fracture parameter combination schemes and input them into the reservoir numerical simulation model coupled with hydraulic fractures, and then construct several numerical simulation models of fractured horizontal wells; Step b: Based on each numerical simulation model of fractured horizontal wells established in Step a, call the simulator to complete the simulation calculation. First, calculate the production performance during a preset time period after the primary fracture. On this basis, set different energy supplement fluid volume parameters and calculate the reservoir production performance to complete the pre-fracture energy supplement simulation. After the calculation is completed, count the basic parameters and calculation results of the numerical simulation models of fractured horizontal wells, collect the reservoir geological data of each model to establish a geological characteristic parameter set, collect the primary fracture engineering parameter data of each model to establish an engineering characteristic parameter set, collect the dynamic production parameters of each model to establish a production characteristic parameter set. Finally, set all geological characteristic parameters, engineering characteristic parameters, production characteristic parameters, and energy supplement fluid volume parameters as input variables, and set the average reservoir pressure after the energy supplement is completed as the output variable to complete the construction of the learning sample set; Step c: Screen out the key characteristic parameters that affect the pressure recovery level for the learning sample set constructed in Step b; Step d: Use the screened key characteristic parameters and the energy supplement fluid volume parameters as input variables, and use the pressure level after pre-fracture energy supplement as the output variable to construct a data set. Use the data set to train an initial machine learning model to obtain a prediction model for the pressure level after pre-fracture energy supplement of horizontal wells; Step e: Use the prediction model for the pressure level after pre-fracture energy supplement of horizontal wells obtained in Step d to obtain the pressure recovery level corresponding to the energy supplement fluid volume before re-fracturing under different geological, engineering, and production conditions. With the goal of maximizing the pressure recovery level, determine the energy supplement fluid volume before re-fracturing of horizontal wells.

2. The method according to claim 1, characterized in that, The specific process of using the data set to train the initial machine learning model to obtain the prediction model for the pressure level after pre-fracture energy supplement of horizontal wells is as follows: Randomly divide the data set into a training set and a prediction set; use the divided training set to train three machine learning models: BP neural network, support vector machine, and gradient boosting decision tree, calculate the determination coefficient between the actual pressure level of the training set and the pressure level calculated by the model, and select the machine learning model with the largest determination coefficient to construct the prediction model for the pressure level after pre-fracture energy supplement of horizontal wells.

3. The method according to claim 2, characterized in that After constructing the prediction model for the pressure level after pre-fracture energy supplement of horizontal wells and before obtaining the prediction model for the pressure level after pre-fracture energy supplement of horizontal wells, it also includes: Calculate the error of the pressure level prediction model after energy supplementation before fracturing of the horizontal well. If the error of the pressure level model after energy supplementation before fracturing of the horizontal well is large, check the accuracy of the learning sample set; if it is accurate, re-screen the key characteristic parameters affecting the pressure recovery level, and at the same time re-establish the pressure level prediction model after energy supplementation before fracturing of the horizontal well and then calculate the model error.

4. The method according to claim 1, wherein The key characteristic parameters include the average reservoir pressure value, the average reservoir thickness value, the average permeability value, the average porosity value, the average oil saturation value, the horizontal well footage value, the average fracture spacing, the fracture length value, the bottom-hole flowing pressure value, and the energy supplementation fluid volume value.

5. The method according to claim 1, characterized in that, Specifically, step c is as follows: For the learning sample set constructed in step b, use the random forest and Pearson correlation analysis models for training; use the random permutation importance as the evaluation index, and comprehensively screen the key characteristic parameters affecting the pressure recovery level based on the evaluation results of the random forest and Pearson correlation analysis models.

6. The method according to claim 1, wherein The geological data includes at least one of the isopach map of the sand body thickness distribution of the target horizon in the study block, the isopach map of the effective thickness distribution, the porosity distribution isopach map, the permeability distribution isopach map, the original formation pressure, temperature, pressure coefficient data, the original oil / gas / water distribution, the original oil-water interface and oil-gas interface, the geological reserve report, the rock compressibility coefficient, and the fluid and rock laboratory analysis report; the engineering parameter data includes at least one of the horizontal well footage, the fracture half-length, the fracture section spacing, the fracturing fluid volume injected into the ground, the sand addition volume, and the fracture monitoring and interpretation data; the production performance data includes at least one of the oil production rate, the liquid production rate, and the bottom-hole flowing pressure.

7. A device for determining the energy supplement fluid volume before refracturing in horizontal wells, characterized in that, The device includes: A numerical simulation model building module, which is used to collect data, and based on the collected data, use a geological modeling algorithm to establish a geological model, import the geological model, rock and fluid physical property data into an oil reservoir numerical simulator for simulation, set the horizontal well, the primary fracturing hydraulic fracture and the corresponding operating regime, form an oil reservoir numerical simulation model coupled with the hydraulic fracture, and generate n groups of random geological and primary fracturing parameter combination schemes by calling a sequential Gaussian data generator and input them into the oil reservoir numerical simulation model coupled with the hydraulic fracture, thereby constructing a number of fractured horizontal well numerical simulation models; the data includes: geological data, engineering parameter data and production performance data of the target area; A data sample set construction module, which is used to call the simulator to complete the simulation calculation based on each group of the constructed fractured horizontal well numerical simulation models. First, calculate the production performance during a preset time period after the primary fracturing. On this basis, set different energy supplementation fluid volume parameters, calculate the oil reservoir production performance, complete the simulation of energy supplementation before fracturing. After the calculation, count the basic parameters and calculation results of the fractured horizontal well numerical simulation models, collect the geological data of each model's oil reservoir to establish a geological characteristic parameter set, collect the primary fracturing engineering parameter data of each model to establish an engineering characteristic parameter set, collect the dynamic production parameters of each model to establish a production characteristic parameter set. Finally, set all the geological characteristic parameters, engineering characteristic parameters, production characteristic parameters and energy supplementation fluid volume parameters as input quantities, and set the average reservoir pressure after the energy supplementation ends as the output quantity to complete the construction of the learning sample set; The data sample set construction module is further configured to screen out the key feature parameters affecting the pressure recovery level for the constructed learning sample set; The data sample set construction module is further configured to use the screened key feature parameters and the energy replenishment liquid volume parameter as input variables, and the pressure level after energy replenishment before fracturing as the output variable to construct a data set, and use the data set to train an initial machine learning model to obtain a prediction model for the pressure level after energy replenishment before fracturing in horizontal wells; The energy replenishment liquid volume determination module before fracturing is configured to use the prediction model for the pressure level after energy replenishment before fracturing in horizontal wells to obtain the pressure recovery level corresponding to the energy replenishment liquid volume before fracturing under different geological, engineering, and production conditions, and determine the energy replenishment liquid volume before repeated fracturing in horizontal wells with the goal of maximizing the pressure recovery level.

8. The device according to claim 7, characterized in that, The step of using the data set to train an initial machine learning model to obtain a prediction model for the pressure level after energy replenishment before fracturing in horizontal wells is specifically as follows: Randomly divide the data set into a training set and a prediction set; use the divided training set to train three machine learning models, namely BP neural network, support vector machine, and gradient boosting decision tree, calculate the determination coefficient between the actual pressure level of the training set and the pressure level calculated by the model, and select the machine learning model with the largest determination coefficient to construct a prediction model for the pressure level after energy replenishment before fracturing in horizontal wells.

9. A computer device, characterized in that, The device includes: One or more processors; and a memory storing computer program instructions, which when executed cause the processor to execute the method according to any one of claims 1-6.

10. A computer-readable medium having computer program instructions stored thereon, which can be executed by a processor to implement the method according to any one of claims 1-6.

Images