A shale oil carbon dioxide fracturing and huff and puff production dynamic intelligent prediction method

Abstract

The present application relates to a kind of shale oil carbon dioxide fracturing production dynamic intelligent prediction method, the method is applied to the improved AHP-GRA-EWM method quantitatively calculate the influence weight of different categories factors to different stage carbon dioxide fracturing-throughput shale oil well production dynamic parameter, obtain the main control influence factor of different production stage production dynamic parameter;The main control influence factor of different production parameters of different production stages is respectively introduced into different artificial neural network system, nonlinear polynomial regression model system and different percolation mathematical model, establishes prediction model, selects the model with highest prediction accuracy to correct improvement, combination generates the production dynamic intelligent prediction model of the entire production cycle of target shale reservoir block carbon dioxide fracturing-throughput shale oil well;Based on the intelligent prediction model, the different production parameter values of different production stages of new well are calculated, and the production dynamic curve is drawn.The method designed by the present application can realize the fine and efficient prediction of carbon dioxide fracturing-throughput shale oil well production dynamics, provide basis for the development of development plan and optimization, improve the single-well production and recovery ratio of target shale reservoir block.

Description

Technical Field

[0001] This invention relates to the field of oil and gas field development technology, and in particular to a method for dynamic intelligent prediction of shale oil carbon dioxide fracturing huff and puff production. Background Technology

[0002] Shale oil reservoirs have low permeability. The main approach is to use hydraulic fracturing technology to create high-conductivity channels between the wells and the reservoir, enabling the wells to achieve production and maintain stable production. In some blocks, due to severe damage from hydraulic fracturing, low formation energy, high crude oil viscosity, and high seepage resistance, carbon dioxide fracturing and huff-and-puff development technologies are used to improve single-well production and reservoir utilization. Existing methods for characterizing the seepage and production dynamics of shale oil wells mostly draw on conventional reservoir seepage models, or improve conventional reservoir seepage models based on some characteristics of shale oil reservoirs, or establish production prediction models for a certain period of time based on machine learning methods. However, the geological and engineering characteristics of shale oil reservoirs differ significantly from those of conventional reservoirs. The seepage characteristics and production dynamics of shale oil wells under CO2 fracturing-huffling at different stages are complex and vary. Existing seepage models are insufficient to predict the production dynamics of shale oil wells under CO2 fracturing-huffling, and a method for establishing a refined prediction model of CO2 fracturing-huffling production dynamics based on machine learning has not yet been developed. Therefore, it is necessary to invent an intelligent characterization method for shale oil CO2 fracturing-huffling production dynamics, and to establish a refined and efficient prediction model of the production dynamics of shale oil wells under CO2 fracturing-huffling at different stages and time points based on field data. This would enable intelligent, accurate, and rapid prediction of the production dynamics of target CO2 fracturing-huffling shale oil wells, providing a basis for the formulation and optimization of development plans.

[0003] Searching the Derwent database, the State Intellectual Property Office patent database, and the Tsinghua Tongfang CNKI journal database, some studies have been conducted on the seepage characteristics, production capacity prediction, and production dynamics of shale oil reservoirs. However, there are certain limitations, mainly in the following aspects: (1) Based on the geological and engineering characteristics of the reservoir, improved seepage models and numerical simulation methods are used to predict the production capacity and production dynamics of oil wells. However, based on a thorough study of the influencing factors of carbon dioxide fracturing-huffing-puffing production dynamics of shale oil wells, machine learning methods are not used to establish prediction models for carbon dioxide fracturing-huffing-puffing production dynamics at different stages. For example, the patent "Method and Device for Predicting Condensate Oil Production in Shale Condensate Gas Reservoirs Based on Production Dynamics Data (CN202211362018.7)" establishes an unsteady seepage mathematical model, linearizes and normalizes the production dynamics data, and uses the linearized seepage mathematical model, the flow rate solution under constant production pressure, and the... Normalized production dynamic data is used to predict shale oil well production. The patent "Semi-steady-state flow stage production capacity calculation method for shale oil and gas fracturing (CN202111565157.5)" applies the concepts of optimal fracture conductivity and wellbore equivalent radius to derive a semi-steady-state seepage model and predict fracturing well production capacity. The literature "Simulation of complex fracture network in tight reservoir of Block S7 and optimization of carbon dioxide huff and puff parameters" establishes a dual-pore, dual-permeability component model based on production dynamic fitting, simulates multiphase flow of reservoir fluid after fracturing, fits post-fracturing production capacity, and studies the influence of different production parameters on production capacity based on numerical simulation methods. The retrieved patents and literature mainly use mathematical model analysis and numerical simulation methods to analyze the influence of different factors on shale oil well production capacity. However, compared to machine learning methods, the factors that can be considered and studied are limited, making them unsuitable for predicting the production dynamics of carbon dioxide fracturing-huff and puff in shale oil wells with complex seepage characteristics. (2) Machine learning algorithms are used to identify and analyze the influencing factors of bottom hole pressure and production, and a prediction model is established. However, based on a comprehensive analysis of geological properties, fluids, rock mechanics, fracturing construction, and production factors, a refined and efficient model and method for accurately and quickly predicting the production capacity of oil wells at different time periods and points in time has not been formed. For example, the literature "Parallel Calculation and Parameter Identification of Formation Pressure in Multi-stage Fracturing Horizontal Wells" establishes the flow equation based on the GPU parallel computing method, obtains the bottom hole pressure data, imports it into the neural network model, and uses the PSO algorithm to optimize the neural network parameters and SVM parameters to obtain the calculation model. The literature "Production Capacity Prediction of Tight Reservoirs Based on Machine Learning" uses random forest and XGBoost algorithms to screen the main influencing factors of production capacity in the three stages of initial high production, rapid decline, and slow decline, and uses BP neural network and support vector machine to establish a production capacity prediction model.While the above literature applies machine learning algorithms to oil well production prediction, it has not yet developed an efficient and accurate method for predicting the dynamic production of shale oil wells using carbon dioxide fracturing and huff-and-puff. Further research is needed to establish refined prediction models for the dynamic production of shale oil wells at different production stages and time points using different machine learning methods. These models should be validated and screened to predict the dynamic production of shale oil wells using carbon dioxide fracturing and huff-and-puff, enabling rapid and accurate plotting of production dynamic curves. Summary of the Invention

[0004] To address the aforementioned technical problems, at least one embodiment of the present invention provides an intelligent prediction method for the dynamic production of shale oil wells using carbon dioxide fracturing and huff-and-puff. This method analyzes the influence weights of different categories of factors on the production capacity of shale oil wells at different production stages, identifies the main controlling influencing factors, and then applies different machine learning intelligent systems to establish intelligent prediction models for the dynamic production of shale oil wells at different production stages, thereby accurately and efficiently predicting the production dynamics of carbon dioxide fracturing-huff-and-puff shale oil wells.

[0005] In some optional embodiments, the method mainly includes the following steps:

[0006] Collect geological, engineering, and production data of the target wells and establish a database analysis system;

[0007] Using the database analysis system, the impact of different factors on the production dynamics of the target well at different production stages is analyzed, and the main controlling influencing factors of different production parameters at different production stages are determined based on the improved AHP-GRA-EWM method.

[0008] The main influencing factors of different production parameters at different production stages are imported into different mathematical models to establish prediction models for different production parameters of the target well at different production stages. The prediction models of different production parameters at different production stages are then combined into a production dynamic intelligent prediction model for the entire production cycle of the target well.

[0009] The parameter values ​​of the main influencing factors at different production stages of the new well are imported into the intelligent prediction model of the production dynamics of the target well throughout its entire production cycle. The values ​​of different production parameters at different production stages of the new well are calculated to complete the production dynamic prediction of the new well.

[0010] In some optional embodiments, the step of using the database analysis system to analyze the impact of different factors on the production dynamics of the target well at different production stages includes:

[0011] Using the database analysis system, a preliminary analysis of the changing patterns of the target well's production dynamic parameters over time is conducted, and the entire production cycle is divided into different production stages.

[0012] Regression analysis was used to study the influence of different factors on the dynamic parameters of target well production at different production stages, and factors that meet the preset requirements were selected as candidate factors.

[0013] The improved AHP-GRA-EWM method was applied to calculate the influence weights of different categories of alternative factors on the production capacity of the target well at different production stages. Based on the weights, the main controlling influencing factors of different production dynamic parameters at different production stages were determined.

[0014] In some optional embodiments, the regression analysis method includes one of linear regression, exponential regression, and multinomial regression.

[0015] In some optional embodiments, the regression analysis method is used to study the influence of different factors on the dynamic parameters of target well production at different production stages, and factors that meet preset requirements are selected as candidate factors, including:

[0016] During the analysis, the correlation coefficient R between each factor and the production parameters of each production stage was taken. 2 The largest curve, mining the correlation coefficient R 2 Factors exceeding a preset threshold are considered as alternative factors.

[0017] In some alternative embodiments, the mathematical model includes one of an artificial neural network model, a nonlinear polynomial regression model, and a seepage mathematical model.

[0018] In some optional embodiments, the step of importing the main controlling influencing factors of different production parameters at different production stages into different mathematical models to establish a prediction model for the target well at different production stages and with different production parameters includes:

[0019] The production dynamics data for different production stages predicted using the aforementioned prediction model are fitted with the actual on-site data.

[0020] The prediction model is corrected based on the fitting results until the accuracy of the prediction model meets the preset accuracy conditions.

[0021] In some optional embodiments, the method further includes:

[0022] Production dynamic prediction curves were plotted based on the values ​​of different production parameters at different production stages of the new well.

[0023] At least one embodiment of the present invention also provides a dynamic intelligent prediction device for shale oil carbon dioxide fracturing huff and puff production, characterized in that it includes:

[0024] The analysis system establishment module is used to collect geological, engineering, and production data of the target well and establish a database analysis system;

[0025] The main control factor analysis module is used to analyze the impact of different factors on the production dynamics of the target well at different production stages using the database analysis system, and to determine the main control factors of different production parameters at different production stages based on the improved AHP-GRA-EWM method.

[0026] The prediction model building module is used to import the main influencing factors of different production parameters at different production stages into different mathematical models, establish prediction models of the target well at different production stages and for different production parameters, and combine the prediction models of different production stages and for different production parameters into a production dynamic intelligent prediction model for the entire production cycle of the target well.

[0027] The production dynamic prediction module is used to import the parameter values ​​of the main influencing factors of different production stages of the new well into the production dynamic intelligent prediction model of the entire production cycle of the target well, calculate the values ​​of different production parameters of the new well at different production stages, and complete the production dynamic prediction of the new well.

[0028] At least one embodiment of the present invention also provides an electronic device, characterized in that it comprises:

[0029] At least one processor; and,

[0030] A memory communicatively connected to the at least one processor; wherein,

[0031] The memory stores instructions that can be executed by the at least one processor, which enables the at least one processor to perform the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method as described above.

[0032] At least one embodiment of the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for dynamic intelligent prediction of shale oil carbon dioxide fracturing huff and puff production.

[0033] At least one embodiment of the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method as described above.

[0034] This invention provides an intelligent prediction method for the dynamic production of shale oil CO2 fracturing and huff-and-puff wells. This method aims to accurately and efficiently predict the production dynamics of shale oil wells undergoing CO2 fracturing and huff-and-puffing. It comprehensively applies an improved hierarchical analysis-grey relational analysis-entropy method (AHP-GRA-EWM), different types of artificial neural network systems, nonlinear polynomial regression, different seepage mathematical models, and a type-two fuzzy logic system to establish an intelligent prediction model for the dynamic production of shale oil CO2 fracturing and huff-and-puffing. This invention solves the technical problems in existing technologies, such as the numerous categories of influencing factors, complex relationships, differences in the main controlling influencing factors at different stages, and the difficulty in quickly and accurately predicting production at different time periods. This method quantitatively calculates the influence weights of different categories of factors on shale oil well productivity at different production stages based on different machine learning algorithm systems, establishing a refined, efficient, and intelligent prediction model for the dynamic production of CO2 fracturing and huff-and-puffing shale oil wells. This model accurately and quickly predicts the production dynamics of new wells at different time periods, providing a basis for the formulation and optimization of development plans. Attached Figure Description

[0035] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.

[0036] Figure 1 This is a flowchart of the steps of the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method according to Embodiment 1 of the present invention;

[0037] Figure 2 This is a flowchart of the database system establishment and production dynamic influencing factor analysis in Embodiment 1 of the present invention;

[0038] Figure 3 This is a flowchart of the analysis of influencing factors of production dynamic master control based on the improved AHP-GRA-EWM in Embodiment 1 of the present invention;

[0039] Figure 4 This is a schematic diagram of a multi-factor hierarchical structure model of different production stages in Embodiment 1 of the present invention;

[0040] Figure 5 This is a flowchart of the establishment of the intelligent prediction model for CO2 fracturing-swallowing shale oil well production dynamics in Embodiment 1 of the present invention;

[0041] Figure 6 This is a schematic diagram of the regression analysis of porosity and daily oil production in Embodiment 2 of the present invention;

[0042] Figure 7 This is a schematic diagram of the regression analysis of oil saturation and daily oil production in Embodiment 2 of the present invention;

[0043] Figure 8This is a schematic diagram of the regression analysis of crude oil viscosity and daily oil production in Embodiment 2 of the present invention;

[0044] Figure 9 This is a schematic diagram of the regression analysis of the horizontal principal stress difference and daily oil production in Embodiment 2 of the present invention;

[0045] Figure 10 This is a schematic diagram of the regression analysis of fracturing fluid volume and daily oil production in Embodiment 2 of the present invention;

[0046] Figure 11 This is a schematic diagram of the regression analysis of the total number of fracturing clusters and daily oil production in Embodiment 2 of the present invention;

[0047] Figure 12 This is a schematic diagram of the GRA-EWM weight values ​​in the depletion-driven stable production stage of Embodiment 2 of the present invention;

[0048] Figure 13 This is a schematic diagram of the GRA-EWM weight values ​​in the rapid decline phase of depletion-type development according to Embodiment 2 of the present invention;

[0049] Figure 14 This is a schematic diagram of the GRA-EWM weight values ​​in the slow decline stage of depletion development according to Embodiment 2 of the present invention;

[0050] Figure 15 This is a schematic diagram of the GRA-EWM weight values ​​during the stable production stage of CO2 throughput in Embodiment 2 of the present invention;

[0051] Figure 16 This is a schematic diagram of the GRA-EWM weight values ​​during the rapid decrease phase of CO2 throughput in Embodiment 2 of the present invention;

[0052] Figure 17 This is a schematic diagram of the GRA-EWM weight values ​​during the slow decrease phase of CO2 throughput in Embodiment 2 of the present invention;

[0053] Figure 18 This is a list of the top 80% of factors in terms of GRA-EWM weight values ​​for different production stages in Embodiment 2 of the present invention;

[0054] Figure 19 This is a list of the hierarchical influence weights of various factors on output at different production stages during the depletion development period according to Embodiment 2 of the present invention.

[0055] Figure 20 This is a list of the hierarchical influence weights of various factors on output during different production stages in Embodiment 2 of the present invention.

[0056] Figure 21 This is a list of different model prediction accuracies at each stage of Embodiment 2 of the present invention;

[0057] Figure 22 This is a schematic diagram of the production dynamic curve of well JH25 in Embodiment 2 of the present invention;

[0058] Figure 23 This is a schematic diagram of the production dynamic curve of the JH25 well optimization scheme in Embodiment 2 of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the embodiments of the present invention to facilitate a better understanding of the invention. However, the technical solutions claimed in the present invention can be implemented even without these technical details and various variations and modifications based on the following embodiments. The division of the following embodiments is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.

[0060] As mentioned earlier, some methods have been developed for predicting the dynamic production of oil wells. However, in the entire CO2 fracturing-huff and puff production cycle of shale oil wells, the factors affecting the production capacity of oil wells at different stages and time points are numerous, complex, and differ. It is necessary to comprehensively apply different machine learning algorithms to analyze the main controlling factors affecting the production capacity at different stages, establish a dynamic production prediction model for CO2 fracturing-huff and puff shale oil wells using multiple intelligent systems, and select efficient and accurate models to quickly and accurately predict the production dynamics of new wells at different stages. This will provide a basis for the formulation, optimization, and adjustment of fracturing operations and development plans, effectively improving oil well production and the recovery rate of target shale oil reservoirs.

[0061] Example 1

[0062] This invention proposes a dynamic intelligent prediction method for shale oil carbon dioxide fracturing huff and puff production. For example... Figure 1 As shown, the implementation process of this method mainly consists of the following four stages:

[0063] Phase 1: Establishment of a database analysis system for carbon dioxide fracturing-swirling shale oil wells;

[0064] Phase Two: Analysis of Influencing Factors of Production Dynamics Control Based on Improved AHP-GRA-EWM;

[0065] Phase 3: Establishment of a dynamic intelligent prediction model for carbon dioxide fracturing-swallowing shale oil well production;

[0066] Phase 4: Prediction and curve plotting of production dynamics at different stages of new wells.

[0067] The following details the operational process of each of the four stages mentioned above.

[0068] Phase 1: Establishment of a database analysis system for carbon dioxide fracturing-swallowing shale oil wells.

[0069] Geological, engineering, and production data of CO2 fracturing-huff-puff wells in the target shale oil reservoir were collected, filtered, integrated, and analyzed to establish a database analysis system. A preliminary analysis of the seepage characteristics and production variation over time in CO2 fracturing-huff-puff wells in the target shale oil reservoir was conducted. Production was divided into n stages, and regression analysis methods such as linear regression, exponential regression, and multinomial regression were comprehensively applied to study the influence of different factors on production parameters such as production rate and bottom hole flowing pressure of the fracturing wells in the target shale oil reservoir at different stages. During the analysis, the correlation coefficients (R²) between each factor and the production parameters of each stage were calculated. 2 The highest curve, dig out R 2 The factor α will be used for the next step of quantitative analysis; α can be taken as 0.001 to 0.05. The database system establishment and production dynamics influencing factor analysis process are detailed below. Figure 2 .

[0070] Phase 2: Analysis of influencing factors of production dynamic control based on improved AHP-GRA-EWM.

[0071] An improved AHP-GRA-EWM method was used to quantitatively calculate the influence weights of different categories of factors on the production dynamic parameters of shale oil wells at different stages of carbon dioxide fracturing-huff-puff. The main controlling influencing factors of production dynamic parameters at different production stages were obtained. The algorithm's operation flow is described in [link to algorithm]. Figure 3 First, the dynamic production parameters of carbon dioxide fracturing-huff and puff shale oil wells were normalized:

[0072]

[0073] In the formula, P ij Let be the value of the i-th production dynamic parameter in the j-th sample; Let m be the normalized value of the i-th production dynamic parameter in the j-th sample, and m be the number of sample wells.

[0074] Production dynamic parameters typically include average daily oil production at different production stages, daily oil production at different times, cumulative oil production at different times, average bottom hole pressure at different production stages, and bottom hole pressure at different times.

[0075] The correlation coefficient R between the production dynamic parameters extracted by regression analysis method 2 After normalizing all factors of α, the normalized factors that are positively correlated with the corresponding production dynamic parameters are characterized as follows:

[0076]

[0077] In the formula, x is the normalized value of the k-th influencing factor of the i-th production dynamic parameter in the j-th sample; kij Let be the value of the k-th influencing factor of the i-th production dynamic parameter in the j-th sample.

[0078] Normalized factors that are negatively correlated with the corresponding production dynamic parameters are characterized as follows:

[0079]

[0080] The improved grey relational-entropy method (GRA-EWM) is used to quantitatively calculate the influence weights of different factors on various production dynamic parameters of shale oil wells. The grey relational weight of the k-th influencing factor on the i-th production dynamic parameter is represented as follows:

[0081]

[0082] The entropy weight of the k-th influencing factor on the i-th production dynamic parameter is represented as:

[0083]

[0084] In the formula, nki represents the number of factors influencing the i-th production dynamic parameter.

[0085] The GRA-EWM weighting of the improvement of the k-th influencing factor on the i-th production dynamic parameter is represented as follows:

[0086]

[0087] Based on the weight calculation results of the improved GRA-EWM method, the top 80% of factors with the greatest influence on different production parameters at different production stages were selected, and the influencing factors were classified into geological physical property factors, fluid factors, rock mechanics factors, fracturing construction factors, fracturing fracture parameters, carbon dioxide injection parameters, and production factors. Geological physical properties include porosity, permeability, oil saturation, effective reservoir thickness, natural fracture development, clay content, and formation pressure; fluid factors include crude oil density, crude oil viscosity, formation water salinity, and crude oil volume compressibility; rock mechanics factors include horizontal principal stress difference, brittleness index, minimum horizontal principal stress, reservoir-interstitial stress difference, Young's modulus, Poisson's ratio, and fracture toughness; fracturing operation factors include fracturing fluid volume, carbon dioxide ratio, operation displacement, and sand ratio; fracturing fracture parameters include fracturing volume, fracture complexity, main fracture half-length, fracture conductivity, fracture height, number of fracturing stages, and total number of fracturing clusters; carbon dioxide injection parameters include carbon dioxide injection rate, injection cycle, well shut-in time, timing of carbon dioxide injection, and injection rate; production factors include production time, bottom hole pressure, daily oil production, and cumulative oil production. Based on the factor classification, a multi-factor hierarchical structure model for different production stages is established. The model structure is shown in [reference needed]. Figure 4 In the diagram, n p n represents the number of dynamic production parameters. M,s,i n represents the number of geological and physical property influencing factors for the i-th production dynamic parameter in the s-th production stage; F,s,i n represents the number of fluid influencing factors for the i-th production dynamic parameter in the s-th production stage; R,s,i n represents the number of rock mechanics influencing factors for the i-th production dynamic parameter in the s-th production stage; T,s,i The number of factors influencing fracturing construction for the i-th dynamic parameter of the s-th production stage; n A,s,i The number of factors influencing the fracturing fracture parameters in the i-th production dynamic parameter of the s-th production stage; n X,s,i n represents the number of factors influencing the carbon dioxide throughput of the i-th production dynamic parameter in the s-th production stage; D,s,i The number of production influencing factors for the i-th production dynamic parameter in the s-th production stage.

[0088] Using the improved GRA-EWM weights of each factor at different production stages as quantitative scaling values, pairwise comparison matrices are established. The pairwise comparison matrix of the i-th production dynamic parameter at the s-th production stage is represented as follows:

[0089]

[0090] Randomly construct null comparison matrices The consistency index of the u-th comparison matrix of the i-th production dynamic parameter in the s-th production stage, obtained by solving for the largest eigenvalue, is represented as:

[0091]

[0092] In the formula, n ki,s λ represents the number of factors influencing the i-th production dynamic parameter in the s-th production stage. iu,s,max Let be the largest eigenvalue of the u-th comparison matrix for the i-th production dynamic parameter in the s-th production stage. The average consistency index of the i-th production dynamic parameter in the s-th production stage is characterized as:

[0093]

[0094] The consistency ratio is characterized as:

[0095]

[0096] Consistency ratio O iu,s When γ < 0.005, the weights of each factor on different production parameters at different production stages of carbon dioxide fracturing-huff and puff shale oil wells are obtained through consistency tests. The top 70% of factors are selected as the main controlling factors, and γ can be taken as 0.001-0.05.

[0097] Phase 3: Establishment of a dynamic intelligent prediction model for carbon dioxide fracturing-swallowing shale oil well production.

[0098] The main influencing factors of different production parameters at different production stages are imported into different artificial neural network systems, nonlinear multinomial regression model systems, and different seepage mathematical models to establish prediction models for different production parameters at different production stages of CO2 fracturing-huff-puff shale oil wells. The production dynamic parameters predicted by each model at different stages are fitted with actual field data. The model with the highest prediction accuracy is selected for each stage for correction and improvement until the average relative error of the prediction is <β, generating a prediction model for the corresponding production dynamic parameters at the corresponding stage. β can be taken as 0.01 to 0.05. During the model improvement process, the production stages can be re-divided, and the influence weight of each factor on the production dynamic parameters in the newly divided production stages can be calculated to update the main influencing factors. Different models can be selected to represent different production parameters at different stages, and combined to generate a precise, efficient, and intelligent prediction model for the production dynamics of CO2 fracturing-huff-puff shale oil wells. The model establishment process is described in [link to documentation]. Figure 5 During model building, incorporating time into the production dynamic parameter prediction model at different stages allows for the acquisition of production dynamic parameter values ​​at different time points, enabling the plotting of production dynamic curves. However, data collection often encounters issues such as data uncertainty and incompleteness. Combining various model generation systems with a type-two fuzzy logic system addresses the uncertainty of input parameters.

[0099] The artificial neural network system consists of interconnected nodes that compute activation functions. The initial input parameters are the values ​​of the main influencing factors of various dynamic production parameters at different production stages, and the final output parameters are the values ​​of the dynamic production parameters at different production stages. The output value of each node's activation function is represented as follows:

[0100]

[0101] In the formula, This represents the output value of the i-th production dynamic parameter at the s-th production stage in computation node J. The weights assigned to node J in the neural network system. f(x) J The input parameter value is the function value with the input parameter value as the independent variable. It can be the transpose of the input vector or the function value calculated in a type II fuzzy logic system based on the range of the input parameters. This serves as a correction constant. Commonly used artificial neural network systems include backpropagation (BP) neural networks, recurrent neural networks, convolutional neural networks, and adaptive linear neural networks.

[0102] The nonlinear polynomial regression model of the i-th production dynamic parameter in the s-th production stage is characterized as follows:

[0103]

[0104] Take the partial derivative of both sides of the equation with respect to the coefficient set ak:

[0105]

[0106] Pick The following system of equations was obtained:

[0107]

[0108] The following matrix is ​​then obtained:

[0109]

[0110] By solving the matrix, the nonlinear fitting equations for each production dynamic parameter at different production stages can be obtained.

[0111] The mathematical models of seepage include linear flow models, multi-zone composite models, and discrete crack models. Combined with Laplace transform and Stehfest numerical inversion methods, the dynamic parameters of each production stage can be calculated and obtained.

[0112] Prediction models for different production dynamic parameters at different production stages are established by applying artificial neural networks, nonlinear polynomial regression, and seepage mathematical models respectively. These models are then fitted with actual field data. The model with the highest prediction accuracy is selected for correction and improvement until the average relative error of the prediction is <β. Prediction models for the corresponding production dynamic parameters at each stage are obtained. Finally, these models are combined to generate full-cycle prediction models for each production dynamic parameter.

[0113] Phase 4: Prediction and curve plotting of production dynamics at different stages of new wells.

[0114] The parameters of the main influencing factors at different production stages of the new well are imported into the intelligent prediction model of the production dynamics of the carbon dioxide fracturing-huff and puff well in the target shale oil block. The model calculates and predicts the different production parameter values ​​at different production stages of the new well, plots the production dynamic curve, and optimizes and adjusts the fracturing construction and development plan based on the predicted production dynamics of the new well.

[0115] In practical applications, the implementation steps of the above method are as follows:

[0116] Collect geological, engineering, and production data of CO2 fracturing-in-out wells in the target shale oil reservoir, filter and integrate all data, and establish a database analysis system.

[0117] Based on the database analysis system established in step (1), a preliminary analysis was conducted on the variation of the dynamic parameters of the carbon dioxide fracturing-huff and puff wells in the target shale oil reservoir over time, and the production was divided into n stages.

[0118] Regression analysis was used to study the influence of different factors on the production dynamic parameters of fractured wells in target shale oil reservoirs at different stages, and to explore the correlation coefficient R. 2 The factor α will be used as a factor for further quantitative analysis.

[0119] The improved AHP-GRA-EWM method was applied to quantitatively calculate the influence weights of different categories of factors on the production capacity of carbon dioxide fracturing-huff and puff shale oil wells at different stages, and to obtain the main controlling influencing factors of different production dynamic parameters at different production stages.

[0120] The main influencing factors of different production parameters at different production stages obtained in step (4) are imported into different artificial neural network systems, nonlinear polynomial regression model systems and different seepage mathematical models to establish prediction models for different production parameters at different production stages of carbon dioxide fracturing-spit shale oil wells.

[0121] The production dynamic parameters of different stages calculated by the production parameter prediction model established in step (5) are fitted with the actual field data. The model with the highest prediction accuracy is selected for each stage for correction and improvement. When the average relative error of prediction is >β, steps (2) to (5) are repeated until the average prediction error of all production parameters in all production stages is <β. The selected prediction models of all production stages are combined to generate a production dynamic intelligent prediction model for the entire production cycle of carbon dioxide fracturing-huff and puff shale oil wells in the target shale oil reservoir.

[0122] The main influencing factor parameter values ​​of different production dynamic parameters at different production stages of the new well are imported into the intelligent prediction model of the production dynamics of the target shale oil reservoir block carbon dioxide fracturing-huff and puff shale oil well throughout the entire production cycle established in step (6), to predict the different production parameter values ​​at different production stages of the new well and draw the production dynamic curve.

[0123] Based on the predicted production dynamics of the new well in step (7), optimize and adjust the fracturing construction and development plan, guide on-site construction and production, and comprehensively record and integrate the geological properties, fluids, rock mechanics, fracturing construction, fracturing fractures, carbon dioxide injection and production data of the new well, and update the database analysis system established in step (1).

[0124] Repeat steps (1) to (8) to gradually complete the production dynamic prediction of all carbon dioxide fracturing-huff and puff shale oil wells in the target shale oil reservoir block.

[0125] Example 2

[0126] The technical details and effects of the above-described method of the present invention are explained in detail below with reference to examples.

[0127] Taking Block J as an example, a database analysis system for carbon dioxide fracturing-huff and puff shale oil wells was established. Based on the improved AHP-GRA-EWM method, the main influencing factors of production at different production stages were predicted. A dynamic intelligent prediction model for the production of carbon dioxide fracturing-huff and puff shale oil wells in Block J was established by comprehensively applying multiple machine learning intelligent systems, as detailed below:

[0128] Block J is a shale oil reservoir with well-developed natural fractures. Formation pressure ranges from 25.7 to 31.9 MPa, with a pressure coefficient of 0.7 to 0.9, classifying it as an abnormally low-pressure reservoir. Carbon dioxide fracturing-huff-puff technology can effectively connect natural fractures, increase the complexity of the fracture network, and replenish formation energy, thereby enhancing well production. Geological, engineering, and production data from 32 wells in Block J were collected, and a database analysis system was established. By analyzing the seepage characteristics and production changes over time in the sample wells, the production stages were divided: during the depletion development period after fracturing, and within each carbon dioxide huff-puff cycle, the wells generally experience a stable production stage, a rapid decline stage, and a slow decline stage. Assuming there are m... tIf the carbon dioxide throughput is calculated, then the production can be divided into n = 3 + 3m. t In this study, linear regression, exponential regression, and multinomial regression methods were comprehensively applied to investigate the influence of different factors on the production of CO2 fracturing-huff-puff wells in the target shale oil reservoir at different stages. The regression analysis results of porosity, oil saturation, crude oil viscosity, horizontal principal stress difference, fracturing fluid volume, total number of fracturing clusters, and production are shown in [the table below]. Figures 6-11 The daily oil production figure is the average daily oil production of the shale oil well in the 30 days prior to production after fracturing. Regression analysis was performed on the factors affecting oil production at each stage of depletion development and CO2 huff and puff development, with correlation coefficients R0. 2 Factors influencing a value >0.01 include porosity, permeability, oil saturation, effective reservoir thickness, natural fracture density, clay content, formation pressure, crude oil density, crude oil viscosity, crude oil volume compressibility, horizontal principal stress difference, brittleness index, minimum horizontal principal stress, reservoir-interstitial stress difference, Young's modulus, Poisson's ratio, fracturing fluid volume, carbon dioxide ratio, operational displacement, sand ratio, fracturing stimulation volume, fracture complexity, main fracture half-length, fracture conductivity, fracture height, number of fracturing stages, total number of fracturing clusters, carbon dioxide injection rate, huff and puff cycles, well shut-in time, carbon dioxide injection timing, carbon dioxide injection rate, production time, bottomhole pressure, and cumulative oil production. The next step will be to quantitatively calculate the correlation coefficient R. 2 We used factors with an impact weight of >0.01 on output at different production stages to screen the main controlling factors at different production stages.

[0129] Based on the improved AHP-GRA-EWM method, the influence weights of various factors on the production of carbon dioxide fracturing-huff-puff shale oil wells at different production stages were quantitatively calculated. First, the improved GRA-EWM method was applied to quantitatively calculate the correlation coefficient R. 2 The weights of factors with an impact greater than 0.01 on output at different production stages, such as Figures 12-17 As shown. The top 80% of factors at different production stages were selected, and the results are as follows. Figure 18 As shown.

[0130] Factors at different production stages, selected based on GRA-EWM weight values, are categorized into geological physical properties, fluid factors, rock mechanics factors, fracturing construction factors, fracturing fracture parameters, carbon dioxide injection parameters, and production factors. A multi-factor hierarchical structure model for each production stage is established. The GRA-EWM weight values ​​of each factor at different production stages are used as quantitative scaling values. A pairwise comparison matrix is ​​established, and the maximum eigenvalue is calculated. After consistency testing, the hierarchical influence weight values ​​of each factor on output at different production stages are obtained. (See...) Figure 19 , Figure 20The top 70% of factors at each stage were selected as the main controlling factors. During the stable production stage of depletion-type development, the main controlling geological properties were porosity, permeability, oil saturation, effective reservoir thickness, and natural fracture density; the main controlling fluid factor was crude oil viscosity; the main controlling rock mechanics factor was brittleness index; the main controlling fracturing operation factors were fracturing fluid volume and carbon dioxide ratio; the main controlling fracturing fracture parameters were fracturing stimulation volume, fracture complexity, main fracture half-length, fracture conductivity, number of fracturing stages, and total number of fracturing clusters; and the main controlling production factor was bottomhole pressure. During the rapid decline stage of depletion-type development, the main controlling geological properties were porosity, permeability, oil saturation, effective reservoir thickness, and natural fracture density; the main controlling fluid factor was crude oil viscosity; and the main controlling fracturing operation factors were... The main controlling factors are fracturing fluid volume and carbon dioxide ratio. The main controlling fracturing fracture parameters are fracturing stimulation volume, fracture complexity, main fracture half-length, fracture conductivity, fracture height, number of fracturing stages, and total number of fracturing clusters. The main controlling production factor is production time. In the depletion-type development slow decline stage, the main controlling geological and physical property factors are porosity, permeability, oil saturation, effective reservoir thickness, natural fracture density, and formation pressure. The main controlling fluid factor is crude oil viscosity. The main controlling fracturing construction factors are fracturing fluid volume and carbon dioxide ratio. The main controlling fracturing fracture parameters are fracturing stimulation volume, main fracture half-length, fracture conductivity, fracture height, number of fracturing stages, and total number of fracturing clusters. The main controlling production factor is bottom hole pressure.

[0131] During the stable production phase of the carbon dioxide huff and puff period, the main controlling geological properties are porosity, permeability, oil saturation, effective reservoir thickness, and natural fracture density. The main controlling fluid factor is crude oil viscosity, the main controlling rock mechanics factor is brittleness index, and the main controlling fracturing operation factors are fracturing fluid volume and carbon dioxide ratio. The main controlling fracturing fracture parameters are fracturing stimulation volume, fracture complexity, fracture conductivity, number of fracturing stages, and total number of fracturing clusters. The main controlling carbon dioxide huff and puff parameters are carbon dioxide injection rate, huff and puff cycles, and well shut-in time. The main controlling production factors are production time and bottom hole pressure. During the rapid decline phase of the carbon dioxide huff and puff period, the main controlling geological properties are porosity, permeability, oil saturation, effective reservoir thickness, and natural fracture density. The main controlling fluid factor is crude oil viscosity, and the main controlling fracturing operation factors are fracturing fluid volume and carbon dioxide ratio. The main controlling fracturing parameters are fracturing volume, fracture complexity, main fracture half-length, fracture conductivity, number of fracturing stages, and total number of fracturing clusters. The main controlling carbon dioxide injection parameters are carbon dioxide injection rate, injection cycle, and well shut-in time. The main controlling production factor is cumulative oil production. During the slow decline phase of carbon dioxide injection, the main controlling geological properties are porosity, permeability, oil saturation, effective reservoir thickness, natural fracture density, and formation pressure. The main controlling fluid factor is crude oil viscosity. The main controlling fracturing operation factors are fracturing fluid volume and carbon dioxide ratio. The main controlling fracturing parameters are fracturing volume, main fracture half-length, fracture conductivity, fracture height, number of fracturing stages, and total number of fracturing clusters. The main controlling carbon dioxide injection parameters are carbon dioxide injection rate, injection cycle, and well shut-in time. The main controlling production factor is bottomhole pressure.

[0132] The main influencing factors of production output at each stage were imported into BP neural network, recurrent neural network, convolutional neural network, adaptive linear neural network, nonlinear multinomial regression model system, linear flow model, and multi-zone composite seepage model, respectively, to establish production prediction models for different production stages of carbon dioxide fracturing-huff-puff shale oil wells. The production output predicted by each model at different stages was fitted with actual field data. The prediction accuracy of different models at each production stage is as follows: Figure 21 As shown.

[0133] The model with the highest prediction accuracy for each production stage is selected for correction and improvement until the average relative error of the prediction is <0.05, at which point the corresponding stage production prediction model is generated. The production prediction models of each stage are combined to generate the production dynamic prediction model for the entire cycle of the carbon dioxide fracturing-through-spit shale oil wells in Block J.

[0134] Well JH25 is a horizontal well in Block J. The reservoir porosity is 11.9%-16.5%, permeability is 0.009-0.08 mD, oil saturation is 52.16%-60.72%, and crude oil viscosity is 5 mPa·s. Carbon dioxide dry fracturing was used, with 30 fractures created, each segment containing 2-4 fracture clusters, totaling 95 clusters. Following fracturing, one round of carbon dioxide huff and puff was conducted, with an injection volume of 400 × 10⁴ m³. The main influencing factors of well JH25's production at each stage were imported into a production dynamic prediction model for the entire cycle of carbon dioxide fracturing-huff and puff shale oil wells in Block J. The production values ​​at different stages of well JH25 were calculated and predicted, and production dynamic curves were plotted, as shown below. Figure 22 As shown, the daily oil production is the average daily oil production over 5 days.

[0135] Well JH25 has been producing for 180 days after starting carbon dioxide huff and puff, with a daily oil production of less than 3 tons. Further carbon dioxide huff and puff operations could be considered to increase well production. The design proposes adding two huff and puff cycles, with a 15-day shut-in period per cycle. The total carbon dioxide injection volume for the three huff and puff cycles would be 1000 × 10⁴ m³. Figure 23 To optimize the production dynamic curve.

[0136] The data above shows that the optimized scheme increased the cumulative oil production by 57.1% over 1080 days compared to the original scheme. Therefore, it is recommended to add two more CO2 huff and puff cycles, referring to the optimized scheme.

[0137] The above embodiments illustrate that, in response to the problems of numerous influencing factors, complex relationships, and differences in the main controlling factors at different stages in the production dynamics of carbon dioxide fracturing-huffling shale oil wells, this invention provides an intelligent prediction method for the production dynamics of carbon dioxide fracturing-huffling shale oil wells. This method includes establishing a database analysis system for carbon dioxide fracturing-huffling shale oil wells, analyzing the main controlling influencing factors of production dynamics based on an improved AHP-GRA-EWM model, establishing an intelligent prediction model for the production dynamics of carbon dioxide fracturing-huffling shale oil wells, and predicting and plotting the production dynamics of new wells at different stages. Based on the improved AHP-GRA-EWM method, the influence weights of different categories of factors on the production dynamic parameters of CO2 fracturing-huff-puff shale oil wells at different stages are quantitatively calculated to obtain the main controlling influencing factors of production dynamic parameters at different production stages. These main controlling influencing factors of different production parameters at different production stages are then imported into different artificial neural network systems, nonlinear polynomial regression model systems, and different seepage mathematical models to establish, optimize, and correct prediction models. These models are then combined to generate an intelligent prediction model of production dynamics for the entire production cycle of CO2 fracturing-huff-puff shale oil wells in the target shale oil reservoir. Based on this intelligent prediction model, the values ​​of different production parameters at different production stages of new wells are calculated, and production dynamic curves are plotted. The method designed in this invention can achieve precise and efficient prediction of the production dynamics of CO2 fracturing-huff-puff shale oil wells, providing a basis for the formulation and optimization of development plans and improving the single-well production and recovery rate of the target shale oil reservoir.

[0138] Example 3

[0139] Another embodiment of the present invention relates to a dynamic intelligent prediction device for shale oil carbon dioxide fracturing huff and puff production, comprising:

[0140] The analysis system establishment module is used to collect geological, engineering, and production data of the target well and establish a database analysis system;

[0141] The main control factor analysis module is used to analyze the impact of different factors on the production dynamics of the target well at different production stages using the database analysis system, and to determine the main control factors of different production parameters at different production stages based on the improved AHP-GRA-EWM method.

[0142] The prediction model building module is used to import the main influencing factors of different production parameters at different production stages into different mathematical models, establish prediction models of the target well at different production stages and for different production parameters, and combine the prediction models of different production stages and for different production parameters into a production dynamic intelligent prediction model for the entire production cycle of the target well.

[0143] The production dynamic prediction module is used to import the parameter values ​​of the main influencing factors of different production stages of the new well into the production dynamic intelligent prediction model of the entire production cycle of the target well, calculate the values ​​of different production parameters of the new well at different production stages, and complete the production dynamic prediction of the new well.

[0144] Example 4

[0145] Another embodiment of the present invention relates to an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method of the above embodiments.

[0146] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0147] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0148] Example 5

[0149] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method of the above embodiments.

[0150] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0151] Example 6

[0152] Another embodiment of the present invention relates to a computer program product, including a computer program that, when executed by a processor, implements the steps of the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method of the above embodiments.

[0153] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing the present invention, and in practical applications, various changes in form and detail may be made without departing from the spirit and scope of the present invention.

Claims

1. A method for dynamically and intelligently predicting shale oil carbon dioxide fracturing huff and puff production, characterized in that, Includes the following steps: Collect geological, engineering, and production data of the target wells and establish a database analysis system; Using the database analysis system, the impact of different factors on the production dynamics of the target well at different production stages is analyzed, and the main controlling influencing factors of different production parameters at different production stages are determined based on the improved AHP-GRA-EWM method. The main influencing factors of different production parameters at different production stages are imported into different mathematical models to establish prediction models for different production parameters of the target well at different production stages. The prediction models of different production parameters at different production stages are then combined into a production dynamic intelligent prediction model for the entire production cycle of the target well. The parameter values ​​of the main influencing factors at different production stages of the new well are imported into the intelligent prediction model of the production dynamics of the target well throughout its entire production cycle. The values ​​of different production parameters at different production stages of the new well are calculated to complete the production dynamic prediction of the new well.

2. The shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method according to claim 1, characterized in that, The analysis of the impact of different factors on the production dynamics of the target well at different production stages using the database analysis system includes: Using the database analysis system, a preliminary analysis of the changing patterns of the target well's production dynamic parameters over time is conducted, and the entire production cycle is divided into different production stages. Regression analysis was used to study the influence of different factors on the dynamic parameters of target well production at different production stages, and factors that meet the preset requirements were selected as candidate factors. The improved AHP-GRA-EWM method was applied to calculate the influence weights of different categories of alternative factors on the production capacity of the target well at different production stages. Based on the weights, the main controlling influencing factors of different production dynamic parameters at different production stages were determined.

3. The shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method according to claim 2, characterized in that, The regression analysis method includes one of linear regression, exponential regression, and multinomial regression.

4. The intelligent prediction method for shale oil carbon dioxide fracturing huff and puff production dynamics according to claim 1, characterized in that, The regression analysis method is used to study the influence of different factors on the dynamic parameters of target well production at different production stages, and factors that meet the preset requirements are selected as candidate factors, including: During the analysis, the correlation coefficient R between each factor and the production parameters of each production stage was taken. 2 The largest curve, mining the correlation coefficient R 2 Factors exceeding a preset threshold are considered as alternative factors.

5. The shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method according to claim 4, characterized in that, The mathematical model includes one of the following: artificial neural network model, nonlinear polynomial regression model, and seepage mathematical model.

6. The shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method according to claim 5, characterized in that, The process involves importing the main controlling influencing factors of different production parameters at different production stages into different mathematical models to establish a prediction model for the target well at different production stages and with different production parameters, including: The production dynamics data for different production stages predicted using the aforementioned prediction model are fitted with the actual on-site data. The prediction model is corrected based on the fitting results until the accuracy of the prediction model meets the preset accuracy conditions.

7. The intelligent prediction method for shale oil carbon dioxide fracturing huff and puff production dynamics according to claim 1, characterized in that, The method further includes: Production dynamic prediction curves were plotted based on the values ​​of different production parameters at different production stages of the new well.

8. A dynamic intelligent prediction device for shale oil carbon dioxide fracturing huff and puff production, characterized in that, include: The analysis system establishment module is used to collect geological, engineering, and production data of the target well and establish a database analysis system; The main control factor analysis module is used to analyze the impact of different factors on the production dynamics of the target well at different production stages using the database analysis system, and to determine the main control factors of different production parameters at different production stages based on the improved AHP-GRA-EWM method. The prediction model building module is used to import the main influencing factors of different production parameters at different production stages into different mathematical models, establish prediction models of the target well at different production stages and for different production parameters, and combine the prediction models of different production stages and for different production parameters into a production dynamic intelligent prediction model for the entire production cycle of the target well. The production dynamic prediction module is used to import the parameter values ​​of the main influencing factors of different production stages of the new well into the production dynamic intelligent prediction model of the entire production cycle of the target well, calculate the values ​​of different production parameters of the new well at different production stages, and complete the production dynamic prediction of the new well.

9. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the shale oil carbon dioxide fracturing huff and puff production dynamic intelligent prediction method as described in any one of claims 1 to 7.

Images