Shale gas fracturing construction parameter combination optimization method, device, equipment and medium

Abstract

The invention discloses a shale gas fracturing construction parameter combination optimization method and device, equipment and a medium. The method comprises the steps that a tree-shaped fractal fracture network mathematical model of single-phase flow and two-phase flow of a shale gas fracture network fracturing horizontal well is established; on the basis of the tree-shaped fractal fracture network mathematical model, the effective fracture network volume of shale gas is calculated; based on the effective fracture network volume, establishing a unit effective fracture network volume artificial neural network prediction model; and performing fracturing construction parameter optimization on an input data set by applying a genetic algorithm and combining with the effective fracture network volume artificial neural network prediction model to obtain an optimal combination solution of fracturing construction parameters. The method is used for optimizing shale gas well fracturing construction parameters, improving the fracturing transformation effect, improving the single-well yield of the shale gas well and promoting efficient development of shale gas resources.

Description

Technical Field

[0001] The present invention belongs to the technical field of unconventional oil and gas stimulation and transformation, and particularly relates to a method, device, equipment and medium for optimizing the combination of shale gas fracturing construction parameters. Background Art

[0002] Fracture network fracturing is the key technology for current shale reservoir exploitation. The effective fracture network of shale gas is the main seepage channel for fracturing fluid flowback and shale gas production, which determines the level of fracturing fluid flowback rate, the gas production capacity of shale gas wells, and the size of technically recoverable reserves. As the flow channel for fracturing fluid and shale gas, it means that the flowback production data of shale gas wells must carry the characteristic information of the effective fracture network of shale gas. By not only focusing on geological engineering parameters but also accurately utilizing the flowback production data, important reservoir characteristic parameters such as the effective fracture network volume can be obtained, and further optimize the fracturing construction parameters.

[0003] Currently, a large number of scholars' research mainly focuses on the conventional parameters affecting fracturing construction. Lei Meng (2023) et al. combined the static physical properties of the reservoir, analyzed the correlation between various factors and cumulative production, and used the random forest algorithm to establish a production capacity prediction model. Then, taking the average value of block geological parameters as the benchmark index, through the particle swarm algorithm, the fracturing construction parameter combination with the optimal production of the model was found. Guo Dali (2022) et al. first initially selected the parameters affecting the hydraulic fracturing effect, sorted the parameters using the grey correlation method, constructed an optimization model for daily gas production using the particle swarm algorithm, and then inversely deduced the optimal fracturing construction parameters. Cheng Zhenghua (2022) reasonably simplified the fracturing cracks under the fracturing construction method and arranged them into the actual geological model to predict the production capacity after horizontal well segmented multi-cluster fracturing.

[0004] In recent years, the scale of shale gas development has expanded rapidly. However, most of the existing fracturing construction parameter optimization methods applied in the field use relatively single geological and engineering parameters for optimization, ignoring the characteristic information of the effective fracture network of shale gas carried in the flowback production data of shale gas wells, failing to optimize the parameters with the effective fracture network volume as the goal, and unable to obtain good fracturing effects and control economic costs. Summary of the Invention

[0005] The purpose of the present invention is to overcome the defects of the existing technology and provide a method, device, equipment and medium for optimizing the combination of shale gas fracturing construction parameters. By comprehensively considering the dynamic and static characteristics of shale reservoir geology, engineering and flowback, establishing a tree-shaped fractal fracture network model for horizontal wells in shale gas fracture network fracturing, an inversion method for the effective fracture network volume of shale gas, and an artificial neural network prediction model for the unit effective fracture network volume, optimize the combination of fracturing construction parameters, which has guiding significance for guiding the fracture network fracturing of shale gas wells and plays an important role in improving the post-fracturing effect of shale gas wells.

[0006] The object of the present invention is achieved by the following technical solutions:

[0007] A method for optimizing shale gas fracturing construction parameters, the method comprising:

[0008] Establishing a tree-shaped fractal fracture network mathematical model for single-phase flow and two-phase flow of horizontal wells in shale gas seam network fracturing;

[0009] Calculating the effective fracture network volume of shale gas based on the tree-shaped fractal fracture network mathematical model;

[0010] Based on the effective fracture network volume, establishing an artificial neural network prediction model for the unit effective fracture network volume;

[0011] Applying a genetic algorithm, and combining the artificial neural network prediction model of the effective fracture network volume to optimize the fracturing construction parameters of the input data set, and obtaining the optimal combined solution of the fracturing construction parameters.

[0012] Further, the establishment of the tree-shaped fractal fracture network mathematical model for single-phase flow and two-phase flow of horizontal wells in shale gas seam network fracturing specifically includes:

[0013] The mathematical model of the single-phase flow tree-shaped fractal fracture network is:

[0014] W f0c =(1 - C f ΔP f )W f0 ;

[0015] Among them, β represents the correction coefficient, μ represents the fluid viscosity, l 0 , W f0 , h f0 represent the initial length, width and height of the tree-shaped fractal fracture, R L , R W , R h represent the fracture length, width and height ratio, n represents the number of branches of the fractal fracture, m represents the fracture order, ΔP represents the total pressure difference of the tree-shaped fractal fracture network, R represents the total flow resistance of the tree-shaped fractal fracture network, C f represents the fracture compressibility, ΔP f represents the pressure drop of the fracture system, W f0c represents the width of the tree-shaped fractal single fracture under the current fracture pressure;

[0016] The mathematical model of the two-phase flow tree-shaped fractal fracture network is:

[0017]

[0018] Among them, P fis the average pressure of the fracture system; P wf is the bottom-hole flowing pressure of the horizontal wellbore; i is the fluid type, and i takes the value of water or gas; B i is the volume coefficient; K ri (S w ) is the gas / water relative permeability in the tree-shaped fracture network.

[0019] Further, calculating the effective fracture network volume of shale gas based on the above-mentioned mathematical model of the tree-shaped fractal fracture network specifically includes:

[0020] Input the basic parameters of the shale gas well and calculate the bottom-hole flowing pressure;

[0021] Form the initial population index of the fitting parameters and calculate the water production Q w and the gas production Q g ;

[0022] Construct the fitness function and determine whether the preset termination condition is reached. If so, store the fitting parameters, calculate the effective fracture network volume of shale gas and end the calculation. Otherwise, generate a new population index according to gene crossover and mutation and continue to calculate Q w and Q g until the effective fracture network volume of shale gas is obtained.

[0023] Further, establishing an artificial neural network prediction model for the unit effective fracture network volume based on the above-mentioned effective fracture network volume specifically includes:

[0024] Normalize the variables of the effective fracture network volume, and the normalized variables follow a normal distribution;

[0025] Use the coefficient of variation to judge the degree of dispersion of the normalized variables, and use the Pearson correlation coefficient to reflect the linear correlation degree between the variable eigenvalue X and the label value Y;

[0026] The variable comprehensive score calculation model includes:

[0027] Score = w 1 ·S PCC + w 2 ·S MIC + w 3 ·S RF + w 4 ·S CatBoost + w 5 ·S CV ;

[0028] where, w 1 ~w 5 represent weights, w 5 = 0.01, w 1 ~w4 The sum is 0.99, S PCC is the Pearson correlation coefficient between feature X and label value Y, S MIC is the maximum information coefficient, S RF is the importance result calculated by the random forest, S catBoost is the importance result calculated by the CatBoost model, S CV is the divergence calculation result;

[0029] Select variables with a comprehensive score ranking in the top X% as the basic variables for predicting the effective fracture network volume per unit, where X% is a preset threshold;

[0030] Perform principal component analysis on the constructed input variables, and build an ANN algorithm program for the effective fracture network volume prediction model. After dividing the collected feature dataset and label set into training set and test set, apply the built algorithm program to train and test the model.

[0031] Further, the optimization of the fracturing construction parameters for the input dataset by combining the artificial neural network prediction model of the effective fracture network volume to obtain the optimal combination solution of the fracturing construction parameters specifically includes:

[0032] Keep other factors affecting the unit ESRV except the fracturing construction parameters unchanged. The unit ESRV prediction ANN model is expressed as:

[0033]

[0034] Express all the features for calculating the unit ESRV as:

[0035]

[0036] where x i1 ~x i7 represent geological factors, flowback factors and shut-in pressure, and x i8 ~x i11 represent construction displacement, number of clusters per stage, cluster spacing and fluid consumption intensity;

[0037] By adjusting the 4 construction parameters of construction displacement, number of clusters per stage, cluster spacing and fluid consumption intensity, make the unit ESRV after fracturing reach the maximum. The optimization objective is:

[0038]

[0039]

[0040] where q1, q2, q3 and q4 respectively represent the optimization target values of construction displacement, number of clusters per stage, cluster spacing and fluid consumption intensity, and obtain the optimal solution q opt =(q1 opt ,q 2 opt ,q 3 opt ,q 4 opt ) is the optimal combination solution of the fracturing construction parameters.

[0041] On the other hand, the present invention also provides a shale gas fracturing construction parameter optimization device, and the device includes:

[0042] A tree-shaped fractal fracture network establishment module, which establishes a mathematical model of the tree-shaped fractal fracture network for single-phase flow and two-phase flow of horizontal wells with shale gas seam network fracturing;

[0043] A shale gas effective seam network volume inversion module, which calculates the shale gas effective seam network volume based on the mathematical model of the tree-shaped fractal fracture network;

[0044] An effective fracture network prediction module, which establishes an artificial neural network prediction model of the unit effective fracture network volume based on the effective fracture network volume;

[0045] A parameter optimization module, which applies a genetic algorithm and combines the artificial neural network prediction model of the effective fracture network volume to optimize the fracturing construction parameters of the input data set, and obtains the optimal combination solution of the fracturing construction parameters.

[0046] On the other hand, the present invention also provides a computer device, which includes a processor and a memory. A computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement any one of the above shale gas fracturing construction parameter combination optimization methods.

[0047] On the other hand, the present invention also provides a computer-readable storage medium, in which a computer program is stored, and the computer program is loaded and executed by a processor to implement any one of the above shale gas fracturing construction parameter combination optimization methods.

[0048] The beneficial effects of the present invention are as follows:

[0049] The present invention fully considers the average geological and flowback parameters of the reservoir in the shale gas block and optimizes the combination of fracturing construction parameters. First, a mathematical model of a tree-shaped fractal fracture network for single-phase flow and two-phase flow in a horizontal well with shale gas fracture network fracturing is established. Based on the tree-shaped fractal fracture network and combined with the genetic algorithm, a method for inverting the volume of the effective fracture network of shale gas is established. Based on the calculated volume of the effective fracture network, an artificial neural network prediction model for the volume of the unit effective fracture network is established. Applying the genetic algorithm and combining the obtained ANN prediction model for the effective fracture network volume of shale gas, based on the average geological and flowback parameters of the reservoir in a certain shale gas block in the research dataset, the combination of fracturing construction parameters is optimized, which can achieve the purpose of obtaining the optimal effective fracture network volume of shale gas fracturing, improve the construction efficiency, and improve the stimulation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flowchart of the method for optimizing the combination of shale gas fracturing construction parameters provided by an embodiment of the present invention;

[0051] Figure 2 is a histogram of the effective fracture network volume distribution of a certain shale gas well in an embodiment of the present invention;

[0052] Figure 3 is a sorting diagram of variable comprehensive indexes in an embodiment of the present invention;

[0053] Figure 4 is a schematic diagram of the optimal unit ESRV at different displacement rates under a fixed section length and fluid consumption intensity in an embodiment of the present invention;

[0054] Figure 5 is a relationship diagram between the unit ESRV and the fluid consumption intensity under the optimized displacement rate, number of clusters per single section, and cluster spacing conditions in an embodiment of the present invention;

[0055] Figure 6 is a structural block diagram of the device for optimizing the combination of shale gas fracturing construction parameters provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] The following specifically illustrates the embodiments of the present invention through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0057] All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0058] Most of the existing methods for optimizing fracturing construction parameters in mines use relatively single geological and engineering parameters for optimization, ignoring the fact that the backflow production data of shale gas wells carry information on the characteristics of the effective fracture network of shale gas, failing to optimize parameters with the volume of the effective fracture network as the goal, and unable to obtain good fracturing effects and control economic costs.

[0059] To solve the above technical problems, the following various embodiments of the method, device, equipment and medium for optimizing the combination of shale gas fracturing construction parameters of the present invention are proposed.

[0060] Embodiment 1

[0061] Referring to Figure 1 , as Figure 1 shown is the flow block diagram of the method for optimizing the combination of shale gas fracturing construction parameters provided in this embodiment. The method specifically includes the following steps:

[0062] Step 1, establish a mathematical model of a tree-shaped fractal fracture network for single-phase flow and two-phase flow of horizontal wells in shale gas seam network fracturing.

[0063] Considering the fracture closure effect, the flow rate of the 1 / 2 single-cluster seam network single-phase flow tree-shaped fractal fracture network is:

[0064]

[0065] W f0c =(1 - C f ΔP f )W f0

[0066] where β represents the correction coefficient, μ represents the fluid viscosity, l 0 , W f0 , h f0 represent the initial length, width and height of the tree-shaped fractal fracture, R L , R W , R h represent the fracture length, width and height ratio, n represents the number of branches of the fractal fracture, m represents the fracture order, ΔP represents the total pressure difference of the tree-shaped fractal fracture network, R represents the total flow resistance of the tree-shaped fractal fracture network, C f represents the fracture compressibility, ΔP f represents the pressure drop of the fracture system, W f0c represents the width of the tree-shaped fractal single fracture at the current fracture pressure.

[0067] Considering the relative permeability in the single-phase flow tree-shaped fracture network flow model, the gas / water two-phase flow rate calculation model for the 1 / 2 single-cluster tree-shaped fracture network is:

[0068]

[0069] In the formula: P f is the average pressure of the fracture system (varying with fluid production), MPa; P wf is the bottom-hole flowing pressure of the horizontal wellbore, MPa; i is w or g (water or gas). When i is w, Q w represents the water production. When i is g, Q g represents the gas production; B i is the volume factor, m 3 / m 3 ; K ri (S w ) is the relative permeability of gas / water in the tree-shaped fracture network, dimensionless.

[0070] The material balance equation in the fracture system of the shale reservoir and the material balance equation in the matrix system of the shale reservoir are established and solved.

[0071] The material balance equation in the fracture system of the shale reservoir is:

[0072]

[0073] The material balance equation in the matrix system of the shale reservoir is:

[0074]

[0075] Step 2: Based on the tree-shaped fractal fracture network, combined with the genetic algorithm, an inversion method for the effective fracture network volume of shale gas is established. Taking a shale gas fracturing well as an example, the inversion of the effective fracture network volume of 162 post-fracture flowback wells is carried out to obtain a basic data set suitable for predicting the effective fracture network volume of shale gas. Referring to Figure 2 , as Figure 2 shown is the histogram of the effective fracture network volume distribution of a shale gas well in this embodiment.

[0076] ① Input the basic parameters of the shale gas well; ② Calculate the bottom-hole flowing pressure using the Beggs-Brill model; ③ Form the initial population index of the fitting parameters; ④ Calculate Qw and Qg; ⑤ Calculate the fitness function; ⑥ Determine whether the termination condition is reached. If so, store the fitting parameters, calculate the effective fracture network volume of shale gas, and end. If not, select genes, perform gene crossover and mutation to generate a new population, and repeat step ④.

[0077] The fitness function constructed in this embodiment is:

[0078]

[0079] Where:

[0080]

[0081]

[0082] The decision variables are:

[0083]

[0084] where:

[0085] x 1 = l 0

[0086] x 2 = W f0

[0087] x 3 = m

[0088] x 4 = R W

[0089] x 5 = R h

[0090] x 6 = R L

[0091] x 7 = θ

[0092] x 8 = p BT

[0093] x 9 = p fi

[0094] x 10 = α mf

[0095] x 11 = C f

[0096] x 12 = I imb

[0097] The objective function is:

[0098]

[0099] The range of x is:

[0100] LB i ≤ x i ≤ UB i (i = 1, 2, …, 12)

[0101] The constraint conditions are:

[0102] Vfi ≤TIV

[0103] Wherein:

[0104]

[0105] Step 3: Based on the calculated effective fracture network volume, establish an artificial neural network prediction model for the unit effective fracture network volume.

[0106] ① Data normalization

[0107] The normalized data follows a normal distribution:

[0108]

[0109] In the formula: min(x) — the minimum value of data x; max(x) — the maximum value of data x; x * — the value of data x after normalization.

[0110] ② Data divergence judgment

[0111] Use the coefficient of variation to judge the degree of dispersion:

[0112]

[0113] In the formula: σ x is the standard deviation of the feature; μ x is the average value of the feature.

[0114] ③ Data correlation judgment

[0115] Use the Pearson correlation coefficient to reflect the degree of linear correlation between two variables. The correlation coefficient is represented by ρ X,Y and is the covariance cov( X,Y ) of two variables X and Y divided by their standard deviations σ X and σ Y . The result is between -1 and 1. The larger the absolute value of ρ, the stronger the correlation.

[0116]

[0117] In the formula: X is the value of the feature; Y is the label value (predicted variable); σ X is the standard deviation of X; σ Y is the standard deviation of Y; cov(X,Y) is the covariance of two variables X and Y.

[0118] ④ Data importance judgment

[0119] Use two embedded feature evaluation methods: random forest and CatBoost ensemble tree model to calculate the variable importance.

[0120] ⑤ The variable comprehensive score calculation model is as follows:

[0121] Score = w 1 ·S PCC + w 2 ·S MIC + w 3 ·S RF + w 4 ·S CatBoost + w 5 ·S CV

[0122] Among them: the weight w 5 = 0.01, and the remaining weights are evenly distributed to w 1 ~w 4 . Among them, S PCC is the Pearson correlation coefficient between the feature X and the label value Y, S MIC is the maximum information coefficient MIC, S RF is the calculation result of the random forest, S catBoost is the calculation result of CatBoost, and S CV is the calculation result of the divergence degree.

[0123] ⑥ Prediction model feature construction

[0124] Select variables with a cumulative comprehensive index greater than 85%. It can be considered that these features contain 85% of the total features of the feature design and meet the basic requirements for optimizing the fracturing construction parameters. Therefore, select variables with a cumulative comprehensive index greater than 85% as the basic variables for predicting the unit effective fracture network volume (unit ESRV).

[0125] ⑦ Effective fracture network prediction

[0126] Perform principal component analysis on the constructed 11 input variables, and then based on the basic principles and processes of the neural network algorithm, build an ANN algorithm program for predicting the effective fracture network volume. For the collected strong feature dataset and label set, divide the training set and the test set into proportions of 70% and 30% respectively, and apply the built program to train and test the model.

[0127] Regarding the understanding of shale gas fracture network fracturing, 20 features including engineering factors, geological factors, and flowback factors that may affect the size of the effective fracture network volume formed by shale fracturing were collected and designed, as shown in Table 1.

[0128] Table 1 Feature codes and meanings

[0129]

[0130]

[0131] The samples are 162 on-site fracturing wells. The data is processed through the following steps: ① data normalization; ② data divergence judgment; ③ data correlation judgment; ④ data importance judgment; ⑤ calculation of the comprehensive score of variables. Refer to Figure 3 , as Figure 3 shown in the comprehensive index ranking diagram of variables in this embodiment, which shows the calculation results of the comprehensive scores of variables.

[0132] Based on the comprehensive scores, the input variables of the prediction model are constructed: the cumulative comprehensive index of the top 11 features in the comprehensive index of features is 85.2 (greater than 85). It can be considered that these 11 features contain 85% of the information of the 20 features in the feature design. Considering that these 11 features include 5 engineering features (displacement, number of clusters per stage, cluster spacing, fluid consumption intensity, and pump shut-off pressure), 4 geological features (porosity, Young's modulus, Poisson's ratio, and brittleness index), and 2 flowback features (shut-in time and gas breakthrough flowback rate), the features include engineering, geological, and flowback factors, and are mainly engineering factors, meeting the basic requirements for optimizing fracturing construction parameters later. Therefore, these 11 features are selected as the basic features for predicting the unit effective fracture network volume (unit ESRV).

[0133] After a large number of data training and test experiments, an ideal ANN model for predicting the effective fracture network volume is obtained. The correlation coefficient R of the training set of the model is 0.946, the determination coefficient R 2 = 0.895, and the root mean square error RMSE = 0.117×10 4 m 3 / m. The correlation coefficient R of the test set is 0.930, the determination coefficient R 2 = 0.864, and the root mean square error RMSE = 0.129×10 4 m 3 / m.

[0134] Step 4: Apply the genetic algorithm, combined with the ANN prediction model of the shale gas effective fracture network volume obtained in Step 3, to optimize the combination of fracturing construction parameters for the average geological and flowback parameters of the reservoir in the aforementioned shale gas block based on the research dataset.

[0135] Based on the given geological and flowback characteristic conditions, by optimizing the combination of 4 construction parameters, namely construction displacement, number of clusters per stage, cluster spacing, and fluid consumption intensity, the post-fracture unit ESRV is maximized. The fracturing construction parameters corresponding to the predicted maximum unit ESRV are the optimal construction parameter combination. Keeping other factors affecting the unit ESRV unchanged except for the fracturing construction parameters, the ANN prediction model of the unit ESRV can be expressed as:

[0136]

[0137] All the characteristics of the calculation unit ESRV are expressed as:

[0138]

[0139] where x i1 ~x i7 represent geological factors (porosity, Young's modulus, Poisson's ratio, and brittleness index), flowback factors (shut-in time and gas breakthrough flowback rate), and shut-in pressure, and x i8 ~x i11 represent construction displacement, number of clusters per single stage, cluster spacing, and fluid consumption intensity, which are represented by q 1 、q 2 、q 3 and q 4 respectively.

[0140] The optimization objective is:

[0141]

[0142]

[0143] The optimal solution q opt =(q 1 opt ,q 2 opt ,q 3 opt ,q 4 opt ) is the optimal combination solution of the fracturing construction parameters. In this embodiment, the initial population parameters q i of the genetic algorithm include fracturing construction displacement, number of clusters per single stage, cluster spacing, and fluid consumption intensity.

[0144] The average engineering, geological, and flowback conditions of 162 wells in the studied shale gas block are shown in Table 2.

[0145] Table 2 Average Conditions of a Certain Shale Gas

[0146]

[0147]

[0148] There are 4 construction parameters to be optimized, namely displacement, number of clusters per single stage, cluster spacing, and fluid consumption intensity. First, in the current average single-stage length of 67.3 m and fluid consumption intensity of 30.7 m 3 / m, under different displacements (14 m 3 / min~20 m 3 / min), the number of clusters per single stage and cluster spacing are optimized to obtain the optimal unit ESRV under different displacements.

[0149] Apply the genetic algorithm-based construction parameter optimization workflow established above to optimize the number of clusters per single stage and the cluster spacing, and obtain the maximum unit ESRV at different displacement rates. Refer to Figure 4 , as Figure 4 shown is the schematic diagram of the optimal unit ESRV at different displacement rates under the fixed stage length and fluid consumption intensity in this embodiment. It can be seen from this figure that the optimal construction displacement rate is 18 m 3 / min. Therefore, select a displacement rate of 18 m 3 / min, the corresponding optimal number of clusters per single stage of 5 clusters, and a cluster spacing of 13.47 m. Gradually increase the fluid consumption intensity, and apply the established unit ESRV prediction model (ANN prediction model) to calculate the change in unit ESRV at different fluid consumption intensities. Refer to Figure 5 , as Figure 5 shown is the relationship diagram between the unit ESRV and the fluid consumption intensity under the optimized displacement rate, number of clusters per single stage, and cluster spacing conditions in this embodiment. It can be seen from this figure that as the fluid consumption intensity increases, the unit ESRV gradually increases. When the fluid consumption intensity reaches 44 m 3 / m, the curve trend becomes flat, and the maximum unit ESRV can reach 3.534×10 4 m 3 / m, which is 264.7% higher than the unit ESRV of 0.969×

[0150] 10 4 m 3 / m under the current average conditions. Under the optimized displacement rate of 18 m 3 / min, the number of clusters per single stage of 5 clusters, the cluster spacing of 13.47 m, and the corresponding average geological and flowback conditions of the dataset, the optimal fluid consumption intensity is 44 m 3 / m.

[0151] Apply the construction parameter combination genetic algorithm workflow established above to optimize the displacement rate, the number of clusters per single stage, the cluster spacing, and the fluid consumption intensity simultaneously. The value ranges of these 4 construction parameters are shown in Table 3.

[0152] Table 3 Value ranges of construction parameters

[0153]

[0154] The displacement rate is extended from the upper limit of the dataset of 16.25 m 3 / min to 20 m 3 / min, the number of clusters per single stage is extended from the upper limit of the dataset of 8 clusters to 10 clusters, and the cluster spacing and fluid consumption intensity are not extended.

[0155] Compare and study a certain average construction parameter level of the dataset, as shown in Table 4 specifically.

[0156] Table 4 Comparison of results of different optimization schemes

[0157]

[0158]

[0159] The optimization plan mainly adopts the strategies of increasing the displacement, increasing the number of clusters, reducing the cluster spacing and increasing the fluid injection intensity to optimize the maximum unit ESRV. Compared with the average construction parameter level of the data set, the unit ESRV can be increased by about 2.843×10 4 m 3 / m, and the percentage increase is about 293.4%. This will increase the cost of sectioning and fluids. In this embodiment, only from the engineering perspective, the construction parameters are optimized to achieve the maximum unit ESRV, without considering the construction cost.

[0160] This embodiment fully considers the average geological and flowback parameters of the reservoir in the shale gas block of the layer, and optimizes the combination of fracturing construction parameters. First, a mathematical model of the tree-shaped fractal fracture network for single-phase flow and two-phase flow of horizontal wells in shale gas fracture network fracturing is established. Based on the tree-shaped fractal fracture network, combined with the genetic algorithm, a method for inverting the effective fracture network volume of shale gas is established. Based on the calculated effective fracture network volume, an artificial neural network prediction model for the unit effective fracture network volume is established. Applying the genetic algorithm, combined with the obtained ANN prediction model of the effective fracture network volume of shale gas, based on the average geological and flowback parameters of the reservoir in a certain shale gas block of the studied data set, the combination of fracturing construction parameters is optimized, which can achieve the purpose of obtaining the optimal effective fracture network volume of shale gas fracture network fracturing, improve the construction efficiency, and improve the stimulation effect.

[0161] Embodiment 2

[0162] Refer to Figure 6 as Figure 6 shown in the structural block diagram of the device for optimizing the combination of shale gas fracturing construction parameters provided by this embodiment. The device specifically includes the following structures:

[0163] A tree-shaped fractal fracture network establishment module, which establishes a mathematical model of the tree-shaped fractal fracture network for single-phase flow and two-phase flow of horizontal wells in shale gas fracture network fracturing;

[0164] A shale gas effective fracture network volume inversion module, which calculates the effective fracture network volume of shale gas based on the mathematical model of the tree-shaped fractal fracture network;

[0165] An effective fracture network prediction module, which establishes an artificial neural network prediction model for the unit effective fracture network volume based on the effective fracture network volume;

[0166] Parameter optimization module. The parameter optimization module applies the genetic algorithm and combines the artificial neural network prediction model of the effective fracture network volume to optimize the fracturing construction parameters of the input data set, and obtains the optimal combination solution of the fracturing construction parameters.

[0167] Embodiment 3

[0168] This preferred embodiment provides a computer device, which can implement the steps in any embodiment of the shale gas fracturing construction parameter combination optimization method provided by the embodiments of the present application. Therefore, the beneficial effects of the shale gas fracturing construction parameter combination optimization method provided by the embodiments of the present application can be achieved. For details, see the previous embodiments and will not be repeated here.

[0169] Embodiment 4

[0170] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by controlling related hardware through instructions. The instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. For this reason, an embodiment of the present invention provides a storage medium, in which multiple instructions are stored, and the instructions can be loaded by a processor to execute the steps in any embodiment of the shale gas fracturing construction parameter combination optimization method provided by the embodiments of the present invention.

[0171] Among them, the storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk, etc.

[0172] Since the instructions stored in the storage medium can execute the steps in any embodiment of the shale gas fracturing construction parameter combination optimization method provided by the embodiments of the present invention, the beneficial effects that can be achieved by any shale gas fracturing construction parameter combination optimization method provided by the embodiments of the present invention can be achieved. For details, see the previous embodiments and will not be repeated here.

[0173] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for optimizing shale gas fracturing construction parameters, characterized in that, the method includes: establishing a tree-shaped fractal fracture network mathematical model for single-phase flow and two-phase flow of shale gas fractured horizontal wells with a network of fractures; calculating the effective fracture network volume of shale gas based on the tree-shaped fractal fracture network mathematical model; establishing an artificial neural network prediction model for the unit effective fracture network volume based on the effective fracture network volume; applying a genetic algorithm, and optimizing the fracturing construction parameters for the input data set in combination with the artificial neural network prediction model for the effective fracture network volume to obtain the optimal combination solution of the fracturing construction parameters.

2. The method for optimizing shale gas fracturing construction parameters according to claim 1, characterized in that, the establishment of the tree-shaped fractal fracture network mathematical model for single-phase flow and two-phase flow of shale gas fractured horizontal wells specifically includes: The mathematical model of the tree-shaped fractal fracture network for single-phase flow is: W f0c = (1 - C f ΔP f )W f0 ; Among them, β represents the correction coefficient, μ represents the fluid viscosity, l 0 , W f0 , h f0 represent the initial length, width and height of the tree-shaped fractal crack, R L , R W , R h represent the crack length, width and height ratio, n represents the number of branches of the fractal crack, m represents the crack order, ΔP represents the total pressure difference of the tree-shaped fractal crack network, R represents the total flow resistance of the tree-shaped fractal crack network, C f represents the crack compressibility, ΔP f represents the pressure drop of the crack system, W f0c represents the width of a single tree-shaped fractal crack under the current crack pressure. The mathematical model of the tree-shaped fractal fracture network for two-phase flow is: Among them, P f is the average pressure of the fracture system; P wf is the bottom-hole flowing pressure of the horizontal wellbore; i is w or g. When i is w, Q w represents the water production. When i is g, Q g represents the gas production; B i is the volume factor; K ri (S w ) is the gas / water relative permeability in the tree-shaped fracture network.

3. The method for optimizing shale gas fracturing construction parameters according to claim 2, characterized in that, the calculation of the effective fracture network volume of shale gas based on the tree-shaped fractal fracture network mathematical model specifically includes: inputting the basic parameters of the shale gas well and calculating the bottom-hole flowing pressure; Form an initial population index of fitting parameters and calculate the water production Q w and the gas production Q g ; Construct an adaptation function and determine whether the preset termination condition is met. If so, store the fitting parameters, calculate the effective fracture network volume of shale gas, and end the calculation. Otherwise, generate new population indices based on gene crossover and mutation and continue to calculate Q w and Q g until the effective fracture network volume of shale gas is obtained.

4. The method for optimizing shale gas fracturing construction parameters according to claim 1, characterized in that, the establishment of the artificial neural network prediction model for the unit effective fracture network volume based on the effective fracture network volume specifically includes: normalizing the variables of the effective fracture network volume, and the variables after normalization follow a normal distribution; using the coefficient of variation to judge the degree of dispersion of the variables after normalization, and using the Pearson correlation coefficient to reflect the linear correlation degree between the variable eigenvalue X and the label value Y; The variable comprehensive score calculation model includes: Score=w 1 ·S PCC +w 2 ·S MIC +w 3 ·S RF +w 4 ·S CatBoost +w 5 ·S CV ; Among them, w 1 ~w 5 represent weights, w 5 = 0.01, and the sum of w 1 ~w 4 is 0.

99. S PCC is the Pearson correlation coefficient between feature X and label value Y, S MIC is the maximum information coefficient, S RF is the importance result obtained by random forest calculation, S catBoost is the importance result obtained by CatBoost model calculation, S CV is the divergence calculation result; selecting the variables with the top X% comprehensive score as the basic variables for predicting the unit effective fracture network volume, where X% is a preset threshold; performing principal component analysis on the constructed input variables, building an ANN algorithm program for the effective fracture network volume prediction model, dividing the collected feature data set and label set into a training set and a test set, and applying the built algorithm program to train and test the model.

5. The method for optimizing shale gas fracturing construction parameters according to claim 1, characterized in that, the optimization of the fracturing construction parameters for the input data set in combination with the artificial neural network prediction model for the effective fracture network volume to obtain the optimal combination solution of the fracturing construction parameters specifically includes: controlling the factors other than the fracturing construction parameters that affect the unit ESRV to be unchanged, and the unit ESRV prediction ANN model is expressed as: representing all the characteristics for calculating the unit ESRV as: where x i1 ~x i7 represent geological factors, flowback factors and shut-in pressure, and x i8 ~x i11 represent construction displacement, number of clusters per stage, cluster spacing and fluid consumption intensity; by adjusting the four construction parameters of construction displacement, number of clusters per stage, cluster spacing, and fluid consumption intensity, making the post-fracture unit ESRV reach the maximum, and the optimization objective is: Among them, q1, q2, q3, and q4 respectively represent the optimization target values of the construction displacement, the number of clusters per single stage, the cluster spacing, and the fluid consumption intensity, and the optimal solution q opt =(q 1 opt , q 2 opt , q 3 opt , q 4 opt ) is the optimal combination solution of the fracturing construction parameters.

6. A device for optimizing shale gas fracturing construction parameters, characterized in that, the device includes: a tree-shaped fractal fracture network establishment module, which establishes a tree-shaped fractal fracture network mathematical model for single-phase flow and two-phase flow of shale gas fractured horizontal wells with a network of fractures; Shale gas effective fracture network volume inversion module, the shale gas effective fracture network volume inversion module calculates the shale gas effective fracture network volume based on the tree-shaped fractal fracture network mathematical model; Effective fracture network prediction module, the effective fracture network prediction module establishes an artificial neural network prediction model for the unit effective fracture network volume based on the effective fracture network volume; Parameter optimization module, the parameter optimization module applies the genetic algorithm, combines the artificial neural network prediction model of the effective fracture network volume to optimize the fracturing construction parameters of the input data set, and obtains the optimal combination solution of the fracturing construction parameters.

7. A computer device, characterized in that, the computer device includes a processor and a memory, a computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement the shale fracturing construction parameter combination optimization method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, a computer program is stored in the storage medium, and the computer program is loaded and executed by a processor to implement the shale fracturing construction parameter combination optimization method according to any one of claims 1-5.

Images