A real-time three-dimensional fitting system and method for downhole fractures in shale gas fracturing

Abstract

The present invention relates to the technical field of shale gas development, and discloses a real-time three-dimensional fitting system and method for downhole fractures in shale gas fracturing, including a model construction module for constructing a numerical simulation model by a simplified three-dimensional displacement discontinuity method; a model training module for using historical geological parameters as the input of the numerical simulation model, outputting corresponding predicted fracturing curves and predicted fracture morphologies, constructing historical fracture morphologies through historical microseismic data, and iteratively training the numerical simulation model based on the predicted fracturing curves and predicted fracture morphologies, as well as historical fracturing curves and historical fracture morphologies, through a preset model training strategy until the maximum number of iterations is satisfied; a downhole real-time acquisition module for real-time obtaining current geological parameters corresponding to the current real formation; and a downhole real-time fitting module for using the current geological parameters as the input of the trained numerical simulation model and outputting corresponding current fracturing curves and current fracture morphologies.

Description

Technical Field

[0001] The present invention relates to the technical field of shale gas development, and particularly relates to a real-time three-dimensional fitting system and method for downhole fractures in shale gas fracturing. Background Art

[0002] In the forefront field of shale gas development, hydraulic fracturing technology is the key to unlocking natural gas resources in deep shale. However, in the face of the complexity of underground geological structures and the uncertainty of natural fracture networks, the current real-time fitting technology for the fracturing process has shown significant limitations, which not only affect the exploitation efficiency but also limit the accurate assessment and effective development of underground resources.

[0003] Specifically, the problems existing in the prior art mainly include: Disconnection between model simplification and reality: Traditional fracturing simulation technologies often ignore the complexity of underground geological structures, especially the distribution, orientation of natural fractures and their direct impact on the fracturing effect. This simplified treatment leads to a large gap between the model and the actual formation characteristics, and the simulation results are difficult to accurately reflect the real fracturing process, reducing the pertinence and effectiveness of exploitation strategies.

[0004] When considering geological complexity, the model calculation amount increases sharply, and the existing calculation methods and hardware facilities are often difficult to achieve efficient data processing and analysis while ensuring accuracy. This makes it difficult to control and optimize the fracturing process with high real-time requirements, affecting the timing of on-site decision-making.

[0005] Based on this, there is an urgent need for a real-time three-dimensional fitting system and method for downhole fractures in shale gas fracturing, which can ensure high efficiency while achieving real-time and accurate three-dimensional fitting of downhole fractures. Summary of the Invention

[0006] One of the purposes of the present invention is to provide a real-time three-dimensional fitting system and method for downhole fractures in shale gas fracturing, which can ensure high efficiency while achieving real-time and accurate three-dimensional fitting of downhole fractures.

[0007] To achieve the above purpose, a real-time three-dimensional fitting system for downhole fractures in shale gas fracturing is provided, including:

[0008] A model construction module, configured to construct a numerical simulation model by a simplified three-dimensional displacement discontinuity method; only one unit is discretized in the fracture height direction in the simplified three-dimensional displacement discontinuity method;

[0009] A model training module, which is used to obtain historical geological parameters, historical fracturing curves, and historical microseismic data corresponding to historical real formations from a database, use the historical geological parameters as inputs to a numerical simulation model, output corresponding predicted fracturing curves and predicted fracture morphologies, construct historical fracture morphologies through the historical microseismic data, and based on the predicted fracturing curves and predicted fracture morphologies, as well as the historical fracturing curves and historical fracture morphologies, perform iterative training on the numerical simulation model through a preset model training strategy until the maximum number of iterations is satisfied;

[0010] A downhole real-time acquisition module, which is used to acquire current geological parameters corresponding to the current real formation in real time;

[0011] A downhole real-time fitting module, which is used to use the current geological parameters as inputs to the numerically simulated model after training, and output corresponding current fracturing curves and current fracture morphologies.

[0012] The technical principle and effect of this solution: In this solution, the construction of the corresponding numerical simulation model is first realized through the simplified three-dimensional displacement discontinuity method, and only one unit is discretized in the fracture height direction in this simplified three-dimensional displacement discontinuity method. Compared with the three-dimensional displacement discontinuity method, through a slight simplification of excluding non-vertical fractures and the vertical component of shear stress and eliminating the discretization in the vertical height direction, while ensuring the accuracy of fracture morphology simulation, the computational efficiency of fracture simulation is also improved.

[0013] After that, through the historical geological parameters, historical fracturing curves, and historical microseismic data of the historical real formation in the database. Among them, the historical geological parameters are input into the numerical simulation model to output predicted fracturing curves and predicted fracture morphologies. Historical fracture morphologies are constructed through the historical geological data. Then, through the predicted fracturing curves and predicted fracture morphologies, as well as the historical fracturing curves and historical fracture morphologies, a preset model training strategy is used to perform iterative training on the numerical simulation model until the maximum number of iterations is satisfied, thereby realizing the training and optimization of the numerical simulation model, and greatly improving the reliability and accuracy of intelligent fitting.

[0014] After that, the current geological parameters of the current real formation are collected in real time and input into the numerically simulated model after training, so as to realize the real-time three-dimensional fitting of the current real formation and output corresponding current fracturing curves and current fracture morphologies.

[0015] 1. The simplified three-dimensional displacement discontinuity method is adopted to construct the numerical simulation model. This simplified three-dimensional displacement discontinuity method can not only calculate the displacement discontinuity and induced stress of a single crack, but also calculate the displacement discontinuity and induced stress of multiple three-dimensional cracks. Compared with the existing three-dimensional displacement discontinuity method, through a slight simplification of excluding non-vertical cracks and the vertical component of shear stress (along the x2 direction), and eliminating the discretization in the vertical (height) direction, the calculation efficiency is increased by more than a thousand times, greatly improving the simulation accuracy and reliability of simulating complex fracture networks such as natural fractures.

[0016] 2. By predicting the fracturing curve and fracture morphology, as well as the historical fracturing curve and historical fracture morphology, iterative training of the constructed numerical simulation model is realized, thereby improving the reliability of the numerical simulation model in the subsequent real-time three-dimensional fitting process, greatly reducing the manual parameter adjustment time, and improving the fitting efficiency of real-time three-dimensional fitting.

[0017] Furthermore, the model training module includes:

[0018] The historical data acquisition input module is used to obtain the historical geological parameters, historical fracturing curve, and historical microseismic data corresponding to the historical real formation from the database, and use the historical geological parameters as the input of the numerical simulation model;

[0019] The initial model parameter module is used to initialize the parameters of the JAYA algorithm and the numerical simulation model. The JAYA algorithm is designed with a population size of 20, the algorithm termination condition is to reach the maximum number of iterations of 100, and the algorithm design variable is 2; for the numerical simulation model, the regularization parameter μ and the kernel function parameter σ are optimized through multiple experiments, and the value ranges are [0.1, 100] and [0.1, 500] respectively;

[0020] The objective function definition module is used to define the calculation formula of the objective function corresponding to the numerical simulation model;

[0021] The calculation formula of the objective function is:

[0022]

[0023] In the formula, MSE is the mean square error between the predicted fracturing curve and the historical fracturing curve of the numerical simulation model, is the predicted value corresponding to the i-th point in the predicted fracturing curve, yi is the true value corresponding to the i-th point in the historical fracturing curve; n is the total number of points corresponding to the fracturing curve;

[0024] The random assignment module is used to randomly generate a set of vectors [γ, δ] within the value ranges of the regularization parameter μ and the kernel function parameter σ, and assign a set of (γ, δ) to each individual in the population of the JAYA algorithm.

[0025] An objective function calculation module, configured to calculate the optimal solution X of the current population based on the training result of the numerical simulation model and the objective function calculation formula j，best,i and the worst solution X j，worst,i , where the optimal solution X j，best,i is the solution with the largest MSE, and the worst solution X j，worst,i is the solution with the smallest MSE;

[0026] A correction module, configured to correct the current solution of the objective function in the iterative process according to the calculated optimal solution X j，best,i and the worst solution X j，worst,i of the current population, and form a corrected solution;

[0027] The calculation logic of the corrected solution is as follows:

[0028] X′ j，k,i = X j，k,i + r 1j，i (X j，best,i - |X j，k,i |) - r 2j，i (X j，worst,i - |X j，k,i |)

[0029] In the formula, X′ j，k，i is the corrected solution, and X j，k，i is the current solution;

[0030] A judgment module, configured to judge whether the corrected solution X′ j,k,i is better than the current solution X j，k，i . If so, the corrected solution is retained; otherwise, the current solution is retained. All the saved solutions are used as the input for the next iteration;

[0031] A loop output module, configured to judge whether the current iteration number reaches the maximum iteration. If not, continue the iteration, continue to calculate the optimal solution and the worst solution of the population. Otherwise, output the optimal solution X j,best,i , and output the corresponding optimal parameter combination (γ, δ).

[0032] Beneficial effects: In this solution, by using the objective function value as the actual response and combining JAYA to optimize the training of algorithm parameters, it quickly converges to the optimal solution of the model parameters, greatly reducing the calculation time of the algorithm and inversion research, improving the inversion accuracy. At the same time, through the output of the optimal parameter combination, the optimization of the numerical simulation model is realized, providing a more reliable model for subsequent fracture morphology prediction and fracturing curve prediction.

[0033] Further, the construction of the historical fracture morphology from the historical microseismic data includes the following steps:

[0034] Perform data preprocessing on the collected historical microseismic data, where the data preprocessing includes noise filtering operations, false alarm elimination operations, and time correction;

[0035] Determine the three-dimensional spatial coordinates corresponding to each historical microseismic data according to the arrival time difference corresponding to each historical microseismic data;

[0036] Perform spatial clustering analysis on the microseismic data with determined three-dimensional spatial coordinates, and group the microseismic data based on the spatial distribution density between the microseismic data. Each group is a corresponding potential fracture segment;

[0037] Perform fitting on the microseismic data within each clustering group, fit out a plane or curved surface model representing the fracture surface, and calculate the occurrence parameters of the fracture. The occurrence parameters include fracture strike, fracture dip, and fracture dip angle;

[0038] Construct the three-dimensional geometric shape of a single fracture according to the fitted plane or curved surface model and the calculated occurrence parameters of the fracture, and integrate to form a complete discrete fracture network model, and use this discrete fracture network model as the corresponding historical fracture morphology.

[0039] Beneficial effects: In this solution, by accurately identifying and analyzing the fracture network distribution, extension direction, and connectivity, the results of the fracturing operation can be evaluated more accurately, and the accuracy of the fracturing effect evaluation can be improved.

[0040] Furthermore, the geological parameters include natural fracture dip angle, natural fracture length, and natural fracture number, as well as other geological engineering parameters, where the natural fracture dip angle, natural fracture length, and natural fracture number are variables, and other geological engineering parameters are constants.

[0041] Beneficial effects: In this solution, the natural fracture dip angle, natural fracture length, and natural fracture number are variables, which can accurately simulate the complex results and distribution patterns of fracture networks under different geological conditions, help to understand the spatial distribution characteristics of the fracture system, and accurately simulating the characteristics of natural fractures is particularly crucial for formulating effective fracturing strategies, greatly improving the reliability of subsequent fracturing strategies.

[0042] The present invention also provides a method for real-time three-dimensional fitting of downhole fractures based on shale gas fracturing, using the above-mentioned real-time three-dimensional fitting system for downhole fractures based on shale gas, including the following steps:

[0043] S1. Construct a numerical simulation model through a simplified three-dimensional displacement discontinuity method; the simplified three-dimensional displacement discontinuity method discretizes only one unit in the fracture height direction;

[0044] S2. Obtain the historical geological parameters, historical fracturing curves, and historical microseismic data corresponding to the historical true formation from the database, use the historical geological parameters as the input of the numerical simulation model, output the corresponding predicted fracturing curves and predicted fracture morphologies, construct the historical fracture morphology through the historical microseismic data, and based on the predicted fracturing curves and predicted fracture morphologies, as well as the historical fracturing curves and historical fracture morphologies, perform iterative training on the numerical simulation model through a preset model training strategy until the maximum number of iterations is satisfied;

[0045] S3. Obtain the current geological parameters corresponding to the current true formation in real time;

[0046] S4. Use the current geological parameters as the input of the trained numerical simulation model, and output the corresponding current fracturing curves and current fracture morphologies. Description of the Drawings

[0047] Figure 1 It is a logic block diagram of the downhole fracture real-time three-dimensional fitting system based on shale gas in the first embodiment of the present invention;

[0048] Figure 2 It is a flow chart of the downhole fracture real-time three-dimensional fitting method based on shale gas in the first embodiment of the present invention. Detailed Embodiments

[0049] The following is further detailed through specific embodiments:

[0050] Embodiment 1

[0051] A downhole fracture real-time three-dimensional fitting system based on shale gas fracturing is basically as Figure 1 shown, including:

[0052] A model construction module, configured to construct a numerical simulation model by a simplified three-dimensional displacement discontinuity method; only one unit is discretized in the fracture height direction in the simplified three-dimensional displacement discontinuity method;

[0053] A model training module, configured to obtain the historical geological parameters, historical fracturing curves, and historical microseismic data corresponding to the historical true formation from the database, use the historical geological parameters as the input of the numerical simulation model, output the corresponding predicted fracturing curves and predicted fracture morphologies, construct the historical fracture morphology through the historical microseismic data, and based on the predicted fracturing curves and predicted fracture morphologies, as well as the historical fracturing curves and historical fracture morphologies, perform iterative training on the numerical simulation model through a preset model training strategy until the maximum number of iterations is satisfied; the geological parameters include the natural fracture dip angle, natural fracture length, and natural fracture number, as well as other geological engineering parameters, where the natural fracture dip angle, natural fracture length, and natural fracture number are variables, and other geological engineering parameters are constants.

[0054] The model training module includes:

[0055] A historical data acquisition input module, configured to acquire historical geological parameters, historical fracturing curves, and historical microseismic data corresponding to historical real formations from a database, and use the historical geological parameters as inputs to a numerical simulation model;

[0056] An initialization model parameter module, configured to initialize parameters for the JAYA algorithm and the numerical simulation model. The JAYA algorithm is designed with a population size of 20, an algorithm termination condition of reaching a maximum number of iterations of 100, and 2 design variables; for the numerical simulation model, the regularization parameter μ and the kernel function parameter σ are optimized through multiple experiments, and the value ranges are [0.1, 100] and [0.1, 500] respectively;

[0057] An objective function definition module, configured to define the calculation formula for the objective function corresponding to the numerical simulation model;

[0058] The calculation formula for the objective function is:

[0059]

[0060] In the formula, MSE is the mean square error between the predicted fracturing curve and the historical fracturing curve of the numerical simulation model, is the predicted value corresponding to the i-th point in the predicted fracturing curve, yi is the true value corresponding to the i-th point in the historical fracturing curve; n is the total number of points corresponding to the fracturing curve;

[0061] A random assignment module, configured to randomly generate a set of vectors [γ, δ] within the value ranges of the regularization parameter μ and the kernel function parameter σ, and assign a set of (γ, δ) to each individual in the population of the JAYA algorithm;

[0062] An objective function calculation module, configured to calculate the optimal solution X of the current population based on the calculation formula of the objective function according to the training results of the numerical simulation model j，best,i and the worst solution X j，worst,i , where the optimal solution X j，best,i is the solution with the largest MSE, and the worst solution X j，worst,i is the solution with the smallest MSE;

[0063] A correction module, configured to correct the current solution of the objective function during the iteration process according to the calculated optimal solution X j，best,i and the worst solution X j，worst,i of the current population, and form a corrected solution;

[0064] The calculation logic for the corrected solution is:

[0065] X′ j,k,i = Xj，k,i +r 1j，i (X j，best,i -|X j，k,i |)-r 2j，i (X j，worst,i -|X j，k,i |)

[0066] Wherein, X' j，k，i is the corrected solution, and X j，k，i is the current solution;

[0067] A judgment module, which is used to judge whether the corrected solution X' j,k,i is better than the current solution X j，k，i . If so, the corrected solution is retained; otherwise, the current solution is retained, and all the saved solutions are used as the input for the next iteration;

[0068] A loop output module, which is used to judge whether the current iteration number reaches the maximum iteration. If not, continue the iteration, continue to calculate the optimal solution and the worst solution of the population. Otherwise, output the optimal solution X j,best,i , and output the corresponding optimal parameter combination (γ, δ).

[0069] An underground real-time acquisition module, which is used to acquire the current geological parameters corresponding to the current real formation in real time;

[0070] An underground real-time fitting module, which is used to take the current geological parameters as the input of the trained numerical simulation model and output the corresponding current fracturing curve and current fracture morphology.

[0071] Constructing the historical fracture morphology through the historical microseismic data includes the following steps:

[0072] Perform data preprocessing on the collected historical microseismic data. The data preprocessing includes noise filtering operation, false alarm elimination operation, and time correction;

[0073] According to the arrival time differences corresponding to each historical microseismic data, determine the three-dimensional space coordinates corresponding to each microseismic data; in this embodiment, the double-difference positioning technology is used to determine the three-dimensional space coordinates corresponding to each microseismic data based on the arrival time differences of the historical microseismic data recorded by the sensors.

[0074] Perform spatial clustering analysis on the microseismic data with determined three-dimensional space coordinates. Based on the spatial distribution density among the microseismic data, group the microseismic data, and each group is a corresponding potential fracture segment; in this embodiment, the DBSCAN algorithm is used for spatial clustering, that is, according to the spatial distribution density of the microseismic data, cluster the densely distributed data points, identify the possible fracture segments, and at the same time eliminate the isolated noise points.

[0075] Fit the microseismic data within each clustering group to obtain a plane or curved surface model representing the fracture surface, and calculate the attitude parameters of the fractures. The attitude parameters include fracture strike, fracture dip direction, and fracture dip angle. In this embodiment, the RANSAC algorithm is used to fit the plane or curved surface model of the fracture surface, and this process can effectively exclude the interference of outliers.

[0076] Based on the fitted plane or curved surface model and the calculated attitude parameters of the fractures, construct the three-dimensional geometric shape of a single fracture, and integrate it to form a complete discrete fracture network model. Take this discrete fracture network model as the corresponding historical fracture morphology. In this embodiment, the alpha-shape algorithm is applied. By adjusting the shape parameter alpha, generate the three-dimensional geometric shape representing a single fracture. This method can effectively distinguish closely adjacent fracture segments and form a more accurate discrete fracture network. Dynamically adjust the alpha value to adapt to the complexity of the fracture network under different geological conditions, ensuring the flexibility and accuracy of fracture boundary recognition.

[0077] In this embodiment, in the simplified three-dimensional displacement discontinuity method, according to the normal displacement discontinuity D1 and the shear displacement discontinuity D3, based on the preset stress component calculation formula, calculate the corresponding stress component data and the corresponding spatial displacement;

[0078] The stress component calculation formula is:

[0079]

[0080] where, σ mn is the corresponding stress component, μ is the shear modulus, v is the Poisson's ratio, I ,ijk (i, j, k = 1, 2, 3) is the derivative of the kernel function.

[0081] As Figure 2 shown, this embodiment also discloses a real-time three-dimensional fitting method for downhole fractures in shale gas fracturing. Using the above real-time three-dimensional fitting system for downhole fractures in shale gas, it includes the following steps:

[0082] S1. Construct a numerical simulation model by the simplified three-dimensional displacement discontinuity method; the simplified three-dimensional displacement discontinuity method discretizes only one unit in the fracture height direction;

[0083] S2. Obtain the historical geological parameters, historical fracturing curves, and historical micro-seismic data corresponding to the historical real formation from the database, use the historical geological parameters as the input of the numerical simulation model, output the corresponding predicted fracturing curves and predicted fracture morphologies, construct the historical fracture morphology through the historical micro-seismic data, and based on the predicted fracturing curves and predicted fracture morphologies, as well as the historical fracturing curves and historical fracture morphologies, perform iterative training on the numerical simulation model through a preset model training strategy until the maximum number of iterations is satisfied;

[0084] S3. Obtain the current geological parameters corresponding to the current real formation in real time;

[0085] S4. Use the current geological parameters as the input of the trained numerical simulation model and output the corresponding current fracturing curves and current fracture morphologies.

[0086] The above are only embodiments of the present invention. Specific structures and common knowledge such as characteristics in the solution are described in too much detail here. Those of ordinary skill in the art know all the common technical knowledge in the technical field to which the invention belongs before the application date or priority date, can know all the existing technologies in this field, and have the ability to apply the conventional experimental means before this date. Those of ordinary skill in the art can, under the inspiration given in this application, combine their own abilities to improve and implement this solution. Some typical well-known structures or well-known methods should not become obstacles for those of ordinary skill in the art to implement this application. It should be noted that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can still be made, and these should also be regarded as the protection scope of the present invention, and these will not affect the implementation effect of the present invention and the practicality of the patent. The protection scope required by this application should be based on the content of its claims, and the specific implementation manners and the like recorded in the specification can be used to explain the content of the claims.

Claims

1. A real-time three-dimensional fitting system for downhole fractures in shale gas fracturing, characterized in that: Including: A model construction module for constructing a numerical simulation model by means of a simplified three-dimensional displacement discontinuity method; In the simplified three-dimensional displacement discontinuity method, only one unit is discretized in the fracture height direction; A model training module for obtaining historical geological parameters, historical fracturing curves, and historical microseismic data corresponding to historical real formations from a database, using the historical geological parameters as inputs to the numerical simulation model, outputting corresponding predicted fracturing curves and predicted fracture morphologies, constructing historical fracture morphologies through the historical microseismic data, and iteratively training the numerical simulation model based on the predicted fracturing curves and predicted fracture morphologies, as well as the historical fracturing curves and historical fracture morphologies, by means of a preset model training strategy until the maximum number of iterations is satisfied; A downhole real-time acquisition module for real-time obtaining current geological parameters corresponding to the current real formation; A downhole real-time fitting module for using the current geological parameters as inputs to the numerically simulated model after training, and outputting corresponding current fracturing curves and current fracture morphologies; The model training module includes: A historical data acquisition and input module for obtaining historical geological parameters, historical fracturing curves, and historical microseismic data corresponding to historical real formations from a database, and using the historical geological parameters as inputs to the numerical simulation model; Initialization model parameter module, which is used to initialize the parameters of the JAYA algorithm and the numerical simulation model. The designed population size of the JAYA algorithm is 20, the algorithm termination condition is to reach the maximum number of iterations of 100, and the designed variables of the algorithm are 2. The regularization parameter and the kernel function parameter After multiple experiments, the optimized value ranges are respectively and ; An objective function definition module for defining the calculation formula of the objective function corresponding to the numerical simulation model; The calculation formula of the objective function is: Wherein, is the mean square error of the predicted fracturing curve and the historical fracturing curve by the numerical simulation model, is the predicted value corresponding to the -th point in the predicted fracturing curve, is the true value corresponding to the -th point in the historical fracturing curve; is the total number of points corresponding to the fracturing curve; A random allocation module for randomly generating a set of vectors within the value ranges of the regularization parameter and the kernel function parameter and allocating a set of to each individual in the population of the JAYA algorithm ; An objective function calculation module, configured to calculate the optimal solution and the worst solution of the current population based on the calculation formula of the objective function according to the training result of the numerical simulation model and the worst solution , the optimal solution is the solution with the minimum value, and the worst solution is the solution with the maximum value; A correction module for correcting the current solution of the objective function during the iteration process according to the calculated optimal solution and the worst solution of the current population to form a corrected solution; and the worst solution ; The calculation logic of the corrected solution is: In the formula, is the corrected solution, is the current solution; A judgment module for judging the corrected solution to determine whether it is better than the current solution ; if so, retain the corrected solution, otherwise retain the current solution, and use all the saved solutions as the input for the next iteration The loop output module is used to determine whether the current iteration count has reached the maximum iteration. If not, continue the iteration, continue to calculate the optimal solution and the worst solution of the population. Otherwise, output the optimal solution and output the corresponding optimal parameter combination .

2. The real-time three-dimensional fitting system for downhole fractures in shale gas fracturing according to claim 1, wherein: Constructing the historical fracture morphology through the historical microseismic data includes the following steps: Performing data preprocessing on the collected historical microseismic data, where the data preprocessing includes noise filtering operations, false alarm elimination operations, and time correction; Determining the three-dimensional spatial coordinates corresponding to each historical microseismic data according to the arrival time differences corresponding to the respective historical microseismic data; Performing spatial clustering analysis on the microseismic data with determined three-dimensional spatial coordinates, grouping the microseismic data based on the spatial distribution density between the microseismic data, and each group being a corresponding potential fracture segment; Fitting the microseismic data within each clustering group to fit a plane or curved surface model representing the fracture surface, and calculating the occurrence parameters of the fracture, where the occurrence parameters include fracture strike, fracture dip, and fracture dip angle; Constructing the three-dimensional geometric morphology of a single fracture according to the fitted plane or curved surface model and the calculated occurrence parameters of the fracture, and integrating to form a complete discrete fracture network model, and using this discrete fracture network model as the corresponding historical fracture morphology.

3. The real-time three-dimensional fitting system for downhole fractures in shale gas fracturing according to claim 2, wherein: The geological parameters include natural fracture dip angle, natural fracture length, and natural fracture number, as well as other geological engineering parameters, where the natural fracture dip angle, natural fracture length, and natural fracture number are variables, and the other geological engineering parameters are constants.

4. A real-time three-dimensional fitting method for downhole fractures in shale gas fracturing, using a real-time three-dimensional fitting system for downhole fractures in shale gas fracturing according to any one of claims 1 to 3 above, characterized in that: Including the following steps: S1. Constructing a numerical simulation model by means of a simplified three-dimensional displacement discontinuity method; in the simplified three-dimensional displacement discontinuity method, only one unit is discretized in the fracture height direction; S2. Obtain the historical geological parameters, historical fracturing curves, and historical microseismic data corresponding to the historical real formation from the database, use the historical geological parameters as the input of the numerical simulation model, output the corresponding predicted fracturing curves and predicted fracture morphologies, construct the historical fracture morphology through the historical microseismic data, and based on the predicted fracturing curves and predicted fracture morphologies, as well as the historical fracturing curves and historical fracture morphologies, perform iterative training on the numerical simulation model through a preset model training strategy until the maximum number of iterations is satisfied; S3. Obtain the current geological parameters corresponding to the current real formation in real time; S4. Use the current geological parameters as the input of the trained numerical simulation model and output the corresponding current fracturing curves and current fracture morphologies.

Images