Method, device and equipment for obtaining artificial crack parameters

Abstract

This application discloses a method, device, and apparatus for acquiring artificial fracture parameters. The method comprises obtaining a shale gas well model; obtaining an initial probability distribution of artificial fracture parameters in the shale gas well model; obtaining multiple first sample parameters based on the initial probability distribution; obtaining multiple second sample parameters based on the shale gas well model, the multiple first sample parameters, and a reference fitting error; and obtaining target artificial fracture parameters based on the multiple second sample parameters and the shale gas well model. This method can reduce the manual workload during the artificial fracture parameter acquisition process and improve the accuracy of the artificial fracture parameter acquisition results.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of shale gas well mining, and in particular to a method, device, and equipment for obtaining artificial fracture parameters. Background Art

[0002] Shale gas is a man-made reservoir, and extraction requires the creation of artificial fractures through large-scale hydraulic fracturing. The productivity of shale gas wells is closely related to these artificial fractures, and determining the parameters of these fractures is a key step in assessing shale gas well productivity.

[0003] In related technologies, a shale gas well model is established, and then the artificial fracture parameters are continuously modified manually. The calculation results of the model are used to fit the actual production indicators of the gas well, such as gas production and liquid production. If the calculation results are very close to the actual observation results, it is considered that the artificial fracture parameters in the model are consistent with the actual situation of the shale gas well, and the artificial fracture parameters are obtained in this way.

[0004] However, due to the plethora of factors that affect the production capacity of shale gas wells, the related technology relies heavily on manual parameter adjustment, which will lead to great uncertainty in the fitting results and the low accuracy of the obtained artificial fracture parameters. In addition, due to the large number of samples and the huge workload of manual parameter adjustment, the related technology is not suitable for actual situations and is inefficient. Summary of the Invention

[0005] The embodiments of the present application provide a method, apparatus, and device for obtaining artificial crack parameters, which are used to reduce the manual workload in the process of obtaining artificial crack parameters and improve the accuracy of the results of obtaining artificial crack parameters.

[0006] In a first aspect, an embodiment of the present application provides a method for obtaining artificial fracture parameters, the method comprising: obtaining a shale gas well model; obtaining an initial probability distribution of artificial fracture parameters in the shale gas well model; obtaining multiple first sample parameters based on the initial probability distribution; obtaining multiple second sample parameters based on the shale gas well model, the multiple first sample parameters, and a reference fitting error; and obtaining target artificial fracture parameters based on the multiple second sample parameters and the shale gas well model.

[0007] In a possible implementation, obtaining the first sample parameters based on the initial probability distribution includes: performing random sampling in the initial probability distribution using a Markov chain Monte Carlo method to obtain a plurality of first sample parameters.

[0008] In one possible implementation, multiple second sample parameters are obtained based on a shale gas well model, multiple first sample parameters, and a reference fitting error, including: obtaining multiple first sample models based on the multiple first sample parameters and the shale gas well model; obtaining multiple first fitting errors based on the multiple first sample models through a historical fitting error function; establishing a first proxy model based on the multiple first sample parameters and the multiple first fitting errors; and obtaining multiple second sample parameters based on the first proxy model and the reference fitting error.

[0009] In one possible implementation, a first proxy model is established based on multiple first sample models and multiple first fitting errors, including: establishing a second proxy model based on multiple first sample parameters and multiple first fitting errors through a K-nearest neighbor classification (KNN) algorithm; obtaining multiple third sample parameters based on the second proxy model through a Markov chain Monte Carlo method; and iterating the second proxy model based on the multiple third sample parameters and a historical fitting error function to obtain a first proxy model.

[0010] In one possible implementation, the second proxy model is iterated based on multiple third sample parameters and a historical fitting error function to obtain a first proxy model, including: iterating the second proxy model based on multiple third sample parameters and a historical fitting error function, and using the iterated second proxy model that meets a first condition as the first proxy model, wherein the first condition is that the difference between the fitting error obtained by the historical fitting error function for a reference number of third sample parameters and the fitting error obtained by the iterated second proxy model is less than a reference threshold.

[0011] In one possible implementation, obtaining multiple second sample parameters based on the first proxy model and the reference fitting error includes: obtaining multiple sample parameters that meet the reference fitting error based on the first proxy model, and using the multiple sample parameters that meet the reference fitting error as the multiple second sample parameters.

[0012] In one possible implementation, target artificial fracture parameters are obtained based on multiple second sample parameters and a shale gas well model, including: obtaining a target probability distribution of the artificial fracture parameters based on the multiple second sample parameters; and inverting the target artificial fracture parameters based on the target probability distribution and the shale gas well model to obtain the target artificial fracture parameters.

[0013] In a possible implementation, the artificial fracture parameters include at least one of the length of the artificial fracture, the height of the artificial fracture, the water saturation of the artificial fracture, the width of the artificial fracture, and the conductivity coefficient of the artificial fracture.

[0014] The method for obtaining artificial fracture parameters provided in the embodiment of the present application obtains an initial probability distribution of the artificial fracture parameters, obtains multiple first sample parameters based on the probability distribution, further iterates and selects the first sample parameters to obtain more representative second sample parameters, and then obtains the artificial fracture parameters based on the second sample parameters. This makes the final artificial fracture parameters more accurate, effectively solving the current problems of shale gas well artificial fracture parameters that cannot be accurately obtained and the large amount of manual work.

[0015] In a second aspect, an embodiment of the present application provides a device for acquiring artificial fracture parameters, which includes: a first acquisition module for acquiring a shale gas well model; a second acquisition module for acquiring an initial probability distribution of artificial fracture parameters in the shale gas well model; a third acquisition module for acquiring multiple first sample parameters based on the initial probability distribution; a fourth acquisition module for acquiring multiple second sample parameters based on the shale gas well model, multiple first sample parameters and a reference fitting error; and a fifth acquisition module for acquiring target artificial fracture parameters based on multiple second sample parameters and the shale gas well model.

[0016] In a possible implementation, the third acquisition module is configured to perform random sampling in the initial probability distribution using a Markov Chain Monte Carlo method to obtain a plurality of first sample parameters.

[0017] In one possible implementation, the fourth acquisition module is used to obtain multiple first sample models based on multiple first sample parameters and a shale gas well model; obtain multiple first fitting errors based on the multiple first sample models through a historical fitting error function; establish a first proxy model based on the multiple first sample parameters and the multiple first fitting errors; and obtain multiple second sample parameters based on the first proxy model and the reference fitting error.

[0018] In one possible implementation, the fourth acquisition module is used to establish a second proxy model based on multiple first sample parameters and multiple first fitting errors through the K-nearest neighbor classification KNN algorithm; obtain multiple third sample parameters based on the second proxy model through the Markov chain Monte Carlo method; iterate the second proxy model based on the multiple third sample parameters and the historical fitting error function to obtain the first proxy model.

[0019] In one possible implementation, the fourth acquisition module is used to iterate the second proxy model based on multiple third sample parameters and a historical fitting error function, and use the iterated second proxy model that meets the first condition as the first proxy model. The first condition is that the difference between the fitting error obtained by the historical fitting error function for a reference number of third sample parameters and the fitting error obtained by the iterated second proxy model is less than a reference threshold.

[0020] In a possible implementation, the fourth acquisition module is configured to acquire a plurality of sample parameters that satisfy a reference fitting error based on the first proxy model, and use the plurality of sample parameters that satisfy the reference fitting error as the plurality of second sample parameters.

[0021] In one possible implementation, the fifth acquisition module is configured to obtain a target probability distribution of artificial fracture parameters based on multiple second sample parameters; and invert the target artificial fracture parameters based on the target probability distribution and a shale gas well model to obtain the target artificial fracture parameters.

[0022] In a possible implementation, the target artificial fracture parameter includes at least one of the length of the artificial fracture, the height of the artificial fracture, the water saturation of the artificial fracture, the width of the artificial fracture, and the conductivity coefficient of the artificial fracture.

[0023] In a third aspect, an embodiment of the present application provides a computer device, comprising a processor and a memory, wherein the memory stores at least one instruction, and when the at least one instruction is executed by the processor, the method for obtaining artificial crack parameters as described in any one of the first aspects above is implemented.

[0024] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which at least one instruction is stored. When the at least one instruction is executed, the method for obtaining artificial crack parameters as described in any one of the first aspects above is implemented.

[0025] In a fifth aspect, an embodiment of the present application provides a computer program (product), which includes: computer program code, which, when executed by a computer, enables the computer to execute the method for obtaining artificial crack parameters in the above aspects. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0027] Figure 1 This is a flow chart of a method for obtaining artificial crack parameters provided in an embodiment of the present application;

[0028] Figure 2 This is a shale gas well model provided in an embodiment of the present application;

[0029] Figure 3 This is a shale gas well model provided in an embodiment of the present application;

[0030] Figure 4 is a screening result of a second sample model provided in an embodiment of the present application;

[0031] Figure 5 is a screening result of a second sample model provided in an embodiment of the present application;

[0032] Figure 6 is a screening result of a second sample model provided in an embodiment of the present application;

[0033] Figure 7 is a screening result of a second sample model provided in an embodiment of the present application;

[0034] Figure 8 is a screening result of a second sample model provided in an embodiment of the present application;

[0035] Figure 9 is a screening result of a second sample model provided in an embodiment of the present application;

[0036] Figure 10 This is a visual inversion result provided by an embodiment of the present application;

[0037] Figure 11 This is a visual inversion result provided by an embodiment of the present application;

[0038] Figure 12 This is a visual inversion result provided by an embodiment of the present application;

[0039] Figure 13 This is a visual inversion result provided by an embodiment of the present application;

[0040] Figure 14 This is a visual inversion result provided by an embodiment of the present application;

[0041] Figure 15 This is a visual inversion result provided by an embodiment of the present application;

[0042] Figure 16 This is a visual inversion result provided by an embodiment of the present application;

[0043] Figure 17 This is a visual inversion result provided by an embodiment of the present application;

[0044] Figure 18 This is a visual inversion result provided by an embodiment of the present application;

[0045] Figure 19 This is a visual inversion result provided by an embodiment of the present application;

[0046] Figure 20This is a visual inversion result provided by an embodiment of the present application;

[0047] Figure 21 This is a visual inversion result provided by an embodiment of the present application;

[0048] Figure 22 This is a visual inversion result provided by an embodiment of the present application;

[0049] Figure 23 This is a visual inversion result provided by an embodiment of the present application;

[0050] Figure 24 This is a visual inversion result provided by an embodiment of the present application;

[0051] Figure 25 This is a visual inversion result provided by an embodiment of the present application;

[0052] Figure 26 This is a visual inversion result provided by an embodiment of the present application;

[0053] Figure 27 This is a visual inversion result provided by an embodiment of the present application;

[0054] Figure 28 This is a visual inversion result provided by an embodiment of the present application;

[0055] Figure 29 This is a visual inversion result provided by an embodiment of the present application;

[0056] Figure 30 This is a schematic diagram of a device for acquiring artificial crack parameters provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.

[0058] Please refer to Figure 1 , which shows a flow chart of a method for obtaining artificial crack parameters provided by an embodiment of the present application. The method provided by an embodiment of the present application may include the following steps:

[0059] Step 101: Obtain a shale gas well model.

[0060] The shale gas well model refers to a gas-liquid two-phase model of a shale gas multi-stage fractured horizontal well. This model includes two-phase fluid flow. Optionally, the two-phase fluids can be natural gas and formation water.

[0061] In one possible implementation, a shale gas well model is established based on existing information about the shale gas well to obtain the shale gas well model. The existing information about the shale gas well includes information about existing geology, gas reservoirs, fluids, fractures, and wells.

[0062] Illustratively, establishing a shale gas well model based on existing information of the shale gas well includes but is not limited to the following sub-steps 1011-1012:

[0063] 1011. Create artificial fractures in shale gas wells.

[0064] In one possible implementation, embedded discrete fracture module (EDFM) technology is used to create artificial fractures in shale gas wells.

[0065] 1012. Based on the artificial fractures and the existing information of the shale gas well, a shale gas well model is established to obtain a shale gas well model.

[0066] In one possible implementation, based on artificial fractures and according to existing information of shale gas wells, a gas-liquid two-phase numerical model of a shale gas well is established, and the gas-liquid two-phase numerical model of a shale gas well is used as a shale gas well model.

[0067] In another possible implementation, based on artificial fractures and according to existing information of shale gas wells, a gas-liquid two-phase simulation model of a shale gas well is established, and the gas-liquid two-phase simulation model of a shale gas well is used as a shale gas well model.

[0068] In order to make the shale gas well model closer to the actual situation of the shale gas well, in any of the above implementation methods, natural fractures can also be added to the shale gas well model. The natural fractures can be established through embedded discrete fracture technology.

[0069] Please refer to Figure 2 , Figure 2 This is a shale gas well model provided in the embodiment of this application. Figure 2 As shown, in this embodiment, a shale gas well gas-liquid two-phase simulation model is established based on artificial fractures and existing shale gas well information as a shale gas well model. This simulation model includes information such as wellbore length, perforation locations, and basic fracture morphology, but does not include information on the distribution of natural fractures.

[0070] Optionally, the shale gas well model can also include natural fractures, see Figure 3 , Figure 3 This is a shale gas well model provided in the embodiment of this application. Figure 3As shown, in this embodiment, a shale gas well gas-liquid two-phase simulation model is established based on artificial fractures and existing information about the shale gas well as the shale gas well model. The shale gas well gas-liquid two-phase simulation model also includes distribution information of natural fractures. Alternatively, the distribution information of natural fractures is obtained through stochastic simulation.

[0071] Step 102: Obtain the initial probability distribution of artificial fracture parameters in the shale gas well model.

[0072] The shale gas well model includes deterministic parameters and uncertain parameters. The deterministic parameters include the length of the horizontal well, the number of perforation clusters, the length of the reservoir, the width of the reservoir, the thickness of the reservoir, and the matrix permeability; the uncertain parameters include the length of the artificial fracture, the height of the artificial fracture, the water saturation of the artificial fracture, the width of the artificial fracture, and the conductivity coefficient of the artificial fracture. The uncertain parameters are the artificial fracture parameters of the embodiment of the present application.

[0073] In one possible implementation, obtaining the initial probability distribution of artificial fracture parameters in the shale gas well model includes: initially setting the deterministic parameters and artificial fracture parameters of the shale gas well model, and obtaining the values ​​of the deterministic parameters and the initial probability distribution of the artificial fracture parameters.

[0074] The deterministic parameter values ​​are fixed model parameter values ​​and can be set as empirical values. For the initial setting of the artificial fracture parameters, a prior distribution can be assigned to the artificial fracture parameters to obtain the initial probability distribution of the artificial fracture parameters.

[0075] Prior distribution refers to a probability distribution estimated for the target sample based on other relevant parameters before testing or sampling. Optionally, in this embodiment, a random, uniform estimated probability distribution is assigned to the artificial crack parameters as the distribution for initial sampling.

[0076] Optionally, if the shale gas well model also includes natural fractures, the natural fracture parameters are deterministic parameters in the shale gas well model. The embodiment of the present application may also use a set of fixed parameter combinations to initially set the natural fracture parameters.

[0077] Step 103: Acquire a plurality of first sample parameters based on the initial probability distribution.

[0078] In a possible implementation, the initial probability distribution of the artificial crack is sampled, and a plurality of artificial crack parameters obtained by sampling are used as first sample parameters.

[0079] Alternatively, a Markov Chain Monte Carlo method can be used to randomly sample from the initial probability distribution to obtain multiple first sample parameters. The Markov Chain Monte Carlo method is a method that uses computer simulation and sampling within the framework of Bayesian theory.

[0080] Alternatively, a plurality of first sample parameters can be obtained by sampling from the initial probability distribution using the Latin hypercube sampling method in an orthogonal experiment. The Latin hypercube sampling method can evenly distribute the first sample parameters in a high-dimensional space, thereby ensuring unbiased sampling. The number of first sample parameters can be adjusted appropriately based on the number of artificial crack parameters, which is not limited in this embodiment of the present invention.

[0081] Step 104 : Acquire multiple second sample parameters based on the shale gas well model, the multiple first sample parameters, and the reference fitting error.

[0082] Due to the inaccuracy of the first sample parameters, directly using the first sample parameters to obtain artificial fracture parameters will result in large errors in the results. Therefore, the embodiment of the present application combines the shale gas well model and the reference fitting error on the basis of the obtained first sample parameters to obtain more representative and accurate second sample parameters.

[0083] In a possible implementation, obtaining a plurality of second sample parameters based on a shale gas well model, a plurality of first sample parameters, and a reference fitting error includes the following steps.

[0084] 1041. Acquire multiple first sample models based on multiple first sample parameters and a shale gas well model.

[0085] In one possible implementation, obtaining multiple first sample models based on multiple first sample parameters and a shale gas well model includes: inputting the multiple first sample parameters into the shale gas well model to obtain multiple shale gas well models corresponding to the multiple first sample parameters, and using the multiple shale gas well models corresponding to the multiple first sample parameters as the first sample models. The first sample parameters correspond one-to-one to the first sample models.

[0086] For example, the first sample parameters are used as input parameters of the reservoir simulator, and the first sample model is obtained based on the shale gas well model and combined with the embedded discrete fracture preprocessor.

[0087] 1042 : Obtain a plurality of first fitting errors using a history fitting error function based on the plurality of first sample models.

[0088] The historical fitting error function is used to evaluate the accuracy of the sample model. Optionally, the historical fitting error function is defined as follows:

[0089]

[0090] Where i is the sequence of actual data points, j is the sequence of history fitting objective functions, n is the number of actual points, m is the number of history fitting objective functions, and x ij,modelRepresents the simulation results of the model, that is, the first sample model, x ij,history is the actual production data, NF j is the normalized value, defined as the maximum difference between the simulation result and the actual production data, w ij Represents the weight of the history matching data.

[0091] Based on the multiple first sample models, the first fitting error corresponding to each first sample model is calculated by using the historical fitting error function. As mentioned above, the first sample parameters correspond to the first sample models one-to-one, so the first sample parameters also correspond to the first fitting error one-to-one.

[0092] 1043 : Establish a first proxy model based on the plurality of first sample parameters and the plurality of first fitting errors.

[0093] In a possible implementation, establishing a first proxy model based on a plurality of first sample parameters and the plurality of first fitting errors includes the following steps.

[0094] 10431. Establish a second proxy model by using a K-nearest neighbor (KNN) algorithm based on the multiple first sample parameters and the multiple first fitting errors.

[0095] Optionally, based on the plurality of first sample parameters and the plurality of first fitting errors, a K-Nearest Neighbor (KNN) classification algorithm is used to train data to establish a second proxy model representing the first fitting errors and the first sample parameters. The first sample parameters are used as independent variables input to the second proxy model, and the first fitting errors are used as dependent variables output by the second proxy model. The second proxy model is generated using this training data.

[0096] Alternatively, the basic concept of the K-nearest neighbor classification algorithm is as follows:

[0097]

[0098] θ i It is the uncertainty parameter combination of the k nearest observation points, that is, k first sample parameters, i represents each first sample parameter, y(θ i ) is the target variable value of the k nearest observation points, that is, the first fitting error corresponding to the first sample parameter of the k nearest observation points, θ0 is the uncertainty parameter combination, that is, the sample parameter as the independent variable, y(θ0) is the target variable value to be predicted, that is, the fitting error as the dependent variable.

[0099] The KNN algorithm effectively ensures that the target variable's value is representative of the distribution in a high-dimensional space. Unlike polynomial algorithms, the KNN algorithm does not suffer from overfitting. Furthermore, the KNN algorithm ensures efficient computation, increasing computational time by 5-20 times compared to other algorithms.

[0100] 10432, obtaining multiple third sample parameters based on the second agent model through the Markov chain Monte Carlo method.

[0101] Based on the second agent model, a Markov chain Monte Carlo method can be used to regenerate a large number of sample parameters, and the sample parameters are used as the third sample parameters.

[0102] 10433. Iterate the second proxy model based on the multiple third sample parameters and the history fitting error function to obtain a first proxy model.

[0103] In one possible implementation, iterating the second proxy model based on the plurality of third sample parameters and the historical fitting error function to obtain the first proxy model includes iterating the second proxy model based on the plurality of third sample parameters and the historical fitting error function, and using the iterated second proxy model that meets a first condition as the first proxy model. The first condition is that a difference between a fitting error obtained by using the historical fitting error function for a reference number of third sample parameters and a fitting error obtained by using the iterated second proxy model is less than a reference threshold.

[0104] The second proxy model is iterated based on multiple third sample parameters and a historical fitting error function, including: inputting the multiple third sample parameters into the second proxy model to obtain fitting errors of the third sample parameters obtained by the second proxy model, and using the fitting errors of the third sample parameters obtained by the second proxy model as the second fitting errors; selecting the third sample parameter with the smallest second fitting error, and obtaining a third sample model based on the shale gas well model; and repeating steps 1042 and 10431 based on the third sample model to obtain an iterated second proxy model. If the iterated second model does not meet the first condition, then continuing to obtain new third sample parameters based on the iterated second model according to step 10432 until the iterated second proxy model meets the first condition, no new third sample parameters are obtained, and the iterated second proxy model at this time is used as the first proxy model. The first proxy model thus obtained can sufficiently accurately describe the relationship between the fitting errors and the sample parameters.

[0105] It should be noted that the fitting error of the third sample parameter in the first condition obtained by the historical fitting error function and the fitting error of the third sample parameter obtained by the iterative second proxy model are obtained according to the following method.

[0106] Input multiple third sample parameters into the shale gas well model to obtain multiple shale gas well models corresponding to the multiple third sample parameters; input the multiple shale gas well models corresponding to the multiple third sample parameters into the historical fitting error function to obtain the fitting error of the third sample parameters obtained by the historical fitting error function; input the multiple third sample parameters into the second proxy model to obtain the fitting error of the third sample parameters obtained by the second proxy model.

[0107] 1044 , obtain a plurality of second sample parameters based on the first proxy model and the reference fitting error.

[0108] In one possible implementation, obtaining multiple second sample parameters based on the first proxy model and the reference fitting error includes: obtaining multiple sample parameters that meet the reference fitting error based on the first proxy model, and using the multiple sample parameters that meet the reference fitting error as the multiple second sample parameters.

[0109] Optionally, a plurality of second sample models satisfying a reference fitting error are acquired based on the first proxy model, and sample parameters corresponding to the plurality of second sample models satisfying the reference fitting error are used as the plurality of second sample parameters.

[0110] The second sample model is selected from the shale gas well model corresponding to the third sample parameter last generated in step 1043. The second fitting error is the fitting error calculated using the second proxy model using the third sample parameter last generated in step 1043. The reference fitting error is a specific fitting error threshold. Optionally, the reference fitting error can be set based on engineering experience. For example, the reference fitting error can be obtained based on the fitting results of representative simulation curves (bottomhole flowing pressure curve, water production curve).

[0111] After the second sample model is screened out, the artificial crack parameters corresponding to the second sample model are counted, and the artificial crack parameters corresponding to the second sample model are used as second sample parameters.

[0112] Optionally, the second sample model obtained by screening may or may not include natural fracture distribution.

[0113] Please refer to Figure 4-9 , Figure 4-9 The second sample model screening result provided by the embodiment of the present application is shown. Figure 4 、 Figure 6 、 Figure 8 To ignore the results of natural cracks, Figure 5 、 Figure 7 、 Figure 9 Results considering a fixed set of natural fracture parameters.

[0114] Step 105 : Obtain target artificial fracture parameters based on the multiple second sample parameters and the shale gas well model.

[0115] In a possible implementation, obtaining target artificial fracture parameters based on multiple second sample parameters and a shale gas well model includes the following steps.

[0116] 1051. Obtain a target probability distribution of artificial crack parameters based on multiple second sample parameters.

[0117] Statistics are performed on the plurality of second sample parameters to obtain a target probability distribution of the artificial crack parameters.

[0118] 1052. Invert the target artificial fracture parameters based on the target probability distribution and the shale gas well model to obtain the target artificial fracture parameters.

[0119] The target artificial fracture parameters include at least one of the length of the artificial fracture, the height of the artificial fracture, the water saturation of the artificial fracture, the conductivity of the artificial fracture, and the width of the artificial fracture.

[0120] The inversion of artificial fracture parameters refers to the use of an established shale gas well model, using computer programs to obtain the distribution law of parameters such as artificial fracture geometry and conductivity through numerical simulation calculations and production data history fitting.

[0121] Production history matching involves using recorded shale gas reservoir static parameters to calculate key dynamic indicators during shale gas development. The calculated results are then compared with the observed key dynamic indicators of the gas wells, such as wellhead pressure and production. If there are discrepancies between the two, the reservoir static parameters are modified, and the modified static parameters are used for recalculation and comparison until the calculated results match the measured dynamic parameters. In the embodiments of this application, the artificial fracture parameters of the shale gas wells to be inverted are equivalent to the static parameters to be modified in the above method.

[0122] Optionally, based on the target probability distribution and the shale gas well model, a statistical method is used to gradually invert the target artificial fracture parameters, which may include any one or more of the following methods, which are not limited in the embodiments of the present application.

[0123] Method 1: Based on the target probability distribution and shale gas well model, the length of the artificial fracture is inverted. This parameter depends on geological factors and construction results, and the empirical value of the artificial fracture length is generally in the range of 80-120 meters.

[0124] Method 2: Based on the target probability distribution and shale gas well model, the fracture height of the artificial fracture is inverted. This parameter depends on the thickness of the specific shale gas reservoir. The fracture height of the artificial fracture fluctuates widely, with the lowest value being less than 10 meters and the highest reaching around the height of the gas reservoir.

[0125] Method 3: Based on the target probability distribution and the shale gas well model, the water saturation of artificial fractures in shale gas wells is inverted. Due to the frequent occurrence of fracturing water backflow in the early stages of shale gas development, the water saturation of artificial fractures generally fluctuates between 0.6% and 0.7%. However, different reservoir conditions can also result in different water saturations in artificial fractures (sometimes dropping below 0.5%).

[0126] Method 4: Based on the target probability distribution and the shale gas well model, the width of the artificial fracture in the shale gas well is inverted. The equivalent value of the width of the artificial fracture is generally 0.2-0.8 meters, and this parameter is closely related to water production.

[0127] Method 5: Based on the target probability distribution and the shale gas well model, the conductivity coefficient of the artificial fracture in the shale gas well is inverted. The value of this parameter depends on the effectiveness of the fracturing operation and has a large fluctuation range (10-8 millidarsim).

[0128] Optionally, after inverting the target artificial fracture parameters based on the target probability distribution and the shale gas well model, the method further includes: generating a visual inversion result based on the inverted artificial fracture parameters. Optionally, the inverted artificial fracture parameters may include one or more of the inverted artificial fracture length, the inverted fracture height of the shale gas well artificial fracture, the inverted water saturation of the shale gas well artificial fracture, the inverted artificial fracture width, and the inverted conductivity of the shale gas well artificial fracture.

[0129] In one possible implementation, a combined histogram and probability distribution plot are plotted based on the inverted artificial fracture length, inverted artificial fracture height, inverted artificial fracture water saturation, inverted artificial fracture width, and inverted artificial fracture conductivity. Optionally, the visualization results may or may not include natural fracture parameters.

[0130] Please refer to Figure 10-19 , which shows a visual inversion result provided by an embodiment of the present application. Figure 10-19 In the figure, a combined bar chart and probability distribution chart are used to represent the inversion results, where the bar chart represents the frequency of occurrence of artificial fracture parameters corresponding to this range, and the curve represents the probability distribution curve fitted according to the frequency bar chart. Figure 10 、 Figure 12 、 Figure 14 、 Figure 16 、 Figure 18 represents the inversion result without considering natural fractures, Figure 11 、 Figure 13 、 Figure 15 、 Figure 17 、 Figure 19 Represents the inversion results considering a fixed set of natural fracture parameters.

[0131] In another possible implementation, a cumulative probability distribution graph is generated based on the inverted length of the artificial fracture in the shale gas well, the fracture height value of the artificial fracture in the shale gas well, the water saturation of the artificial fracture in the shale gas well, the width of the artificial fracture in the shale gas well, and the conductivity coefficient of the artificial fracture in the shale gas well. Optionally, the visualization result may include or exclude natural fracture parameters.

[0132] The cumulative probability distribution plot is a visualization of the P10, P50, and P90 probabilities. The P10 and P90 values ​​help understand the representative range of a particular artificial fracture parameter, while the P50 value determines the average representative value of a particular artificial fracture parameter. The cumulative probability distribution plot helps better define the range of inversion results.

[0133] Please refer to Figure 20-29 It shows a visual inversion result provided by an embodiment of the present application. Figure 20-29 In the figure, the cumulative probability distribution diagram is used to represent the inversion results, where Figure 20 、 Figure 22 、 Figure 24 、 Figure 26 、 Figure 28 represents the cumulative probability distribution without considering the natural fracture parameters, Figure 21 、 Figure 23 、 Figure 25 、 Figure 27 、 Figure 29 represents the cumulative probability distribution considering a fixed set of natural fracture parameters.

[0134] The method for obtaining artificial fracture parameters provided by the above-mentioned embodiments effectively solves the current problems of inaccurately obtaining artificial fracture parameters in shale gas wells and the high manual workload. By establishing a shale gas well model, setting a prior probability distribution range for the artificial fracture parameters, and iteratively optimizing the artificial fracture parameters of shale gas horizontal wells using a Markov Chain-Monte Carlo algorithm, the most representative artificial fracture values ​​and a numerical model that is as close to the actual situation as possible are obtained. This makes the production performance prediction of shale gas horizontal wells closer to actual production, reducing the uncertainty and manual workload caused by a large number of unrepresentative samples. By statistically analyzing all optimized artificial fracture parameters, the posterior probability distribution of the parameters, i.e., the target probability distribution, is obtained. The distribution trend of the artificial fracture geometry (length, width, fracture height, etc.) and conductivity obtained by this determination method is an infinite approximation of the actual underground situation, enabling researchers and decision makers to gain a new quantitative understanding of the underground fracture network. This is of great significance for evaluating the fracturing effect of deep shale gas wells, predicting the ultimate recoverable reserves, and optimizing reasonable well spacing.

[0135] Based on the same technical concept, please refer to Figure 30 , which shows a schematic diagram of a device for obtaining artificial crack parameters provided by an embodiment of the present application, the device includes but is not limited to the following modules 701-705:

[0136] The first acquisition module 701 is used to acquire a shale gas well model.

[0137] The second acquisition module 702 is used to obtain the initial probability distribution of artificial fracture parameters in the shale gas well model.

[0138] The third acquisition module 703 is configured to acquire a plurality of first sample parameters based on the initial probability distribution.

[0139] In a possible implementation, the third acquisition module 703 is configured to perform random sampling in the initial probability distribution by using a Markov chain Monte Carlo method to obtain a plurality of first sample parameters.

[0140] The fourth acquisition module 704 is configured to acquire a plurality of second sample parameters based on the shale gas well model, the plurality of first sample parameters, and the reference fitting error.

[0141] In one possible implementation, the fourth acquisition module 704 is used to obtain multiple first sample models based on multiple first sample parameters and the shale gas well model; obtain multiple first fitting errors based on the multiple first sample models through a historical fitting error function; establish a first proxy model based on the multiple first sample parameters and the multiple first fitting errors; and obtain multiple second sample parameters based on the first proxy model and the reference fitting error.

[0142] Optionally, the fourth acquisition module 704 establishes a second proxy model based on multiple first sample parameters and multiple first fitting errors through the K-nearest neighbor classification KNN algorithm; obtains multiple third sample parameters based on the second proxy model through the Markov chain Monte Carlo method; iterates the second proxy model based on the multiple third sample parameters and the historical fitting error function to obtain the first proxy model.

[0143] Optionally, the fourth acquisition module 704 is used to iterate the second proxy model based on multiple third sample parameters and the historical fitting error function, and use the iterated second proxy model that meets the first condition as the first proxy model, wherein the first condition is that the difference between the fitting error obtained by the historical fitting error function of the reference number of third sample parameters and the fitting error obtained by the iterated second proxy model is less than a reference threshold.

[0144] Optionally, the fourth acquisition module 704 is configured to acquire a plurality of sample parameters that satisfy a reference fitting error based on the first proxy model, and use the plurality of sample parameters that satisfy the reference fitting error as a plurality of second sample parameters.

[0145] The fifth acquisition module 705 is configured to acquire target artificial fracture parameters based on the plurality of second sample parameters and the shale gas well model.

[0146] In one possible implementation, the fifth acquisition module 705 is configured to obtain a target probability distribution of artificial fracture parameters based on multiple second sample parameters; and invert the target artificial fracture parameters based on the target probability distribution and the shale gas well model to obtain the target artificial fracture parameters.

[0147] The target artificial fracture parameters include at least one of the length of the artificial fracture, the height of the artificial fracture, the water saturation of the artificial fracture, the width of the artificial fracture, and the conductivity coefficient of the artificial fracture.

[0148] It should be noted that the apparatus provided in the above embodiments is merely illustrated by the division of the above functional modules when implementing its functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0149] In an exemplary embodiment, a computer device is further provided, comprising a processor and a memory, wherein the memory stores at least one instruction, wherein the at least one instruction is configured to be executed by one or more processors to implement any of the above-mentioned methods for obtaining artificial fracture parameters.

[0150] In an exemplary embodiment, a storage medium is further provided, in which at least one program code is stored. The at least one program code is loaded and executed by a processor to enable a computer to implement any of the above-mentioned methods for obtaining artificial fracture parameters.

[0151] Optionally, the storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, or the like.

[0152] In an exemplary embodiment, a computer program or computer program product is also provided, wherein the computer program or computer program product stores at least one computer instruction, which is loaded and executed by a processor to cause a computer to implement any of the above-mentioned methods for obtaining artificial fracture parameters. The serial numbers of the above-mentioned embodiments of the present application are for descriptive purposes only and do not represent the merits or demerits of the embodiments.

[0153] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of the modules is merely a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, or can be an electrical, mechanical or other form of connection.

[0154] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0155] In addition, the functional modules in the various embodiments of the present application may be integrated into a processing module, or each module may exist physically separately, or two or more modules may be integrated into a single module. The above-mentioned integrated modules may be implemented in the form of hardware or software functional modules.

[0156] It should also be understood that in the various embodiments of the present application, the size of the serial number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0157] In this application, the term "at least one" means one or more, and the term "plurality" means two or more. For example, a plurality of data refers to two or more data.

[0158] It should be understood that the terminology used in the description of the various examples herein is for the purpose of describing particular examples only and is not intended to be limiting. As used in the description of the various examples and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0159] The above description is merely an exemplary embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for obtaining parameters of artificial cracks, characterized in that: The method comprises: Obtain shale gas well models; Obtaining an initial probability distribution of artificial fracture parameters in the shale gas well model; Based on the initial probability distribution, obtaining a plurality of first sample parameters; Acquire a plurality of first sample models based on the plurality of first sample parameters and the shale gas well model; Acquire a plurality of first fitting errors based on the plurality of first sample models by using a history fitting error function; establishing a first proxy model based on the plurality of first sample parameters and the plurality of first fitting errors; Acquire a plurality of second sample parameters based on the first proxy model and a reference fitting error; Acquire target artificial fracture parameters based on the multiple second sample parameters and the shale gas well model; The step of obtaining the initial probability distribution of the artificial fracture parameters in the shale gas well model comprises: initially setting the deterministic parameters and the artificial fracture parameters of the shale gas well model, obtaining the values ​​of the deterministic parameters and the initial probability distribution of the artificial fracture parameters, wherein the values ​​of the deterministic parameters are fixed model parameter values ​​and are empirical values, and the initial probability distribution of the artificial fracture parameters is to assign a random and uniform estimated probability distribution to the artificial fracture parameters; The establishing of the first proxy model based on the plurality of first sample parameters and the plurality of first fitting errors includes: establishing a second proxy model based on the plurality of first sample parameters and the plurality of first fitting errors by using a K-nearest neighbor classification (KNN) algorithm; obtaining a plurality of third sample parameters based on the second proxy model by using a Markov chain Monte Carlo method; and iterating the second proxy model based on the plurality of third sample parameters and the historical fitting error function to obtain the first proxy model. The acquiring the plurality of second sample parameters based on the first proxy model and the reference fitting error includes: acquiring a plurality of sample parameters that satisfy the reference fitting error based on the first proxy model, and using the plurality of sample parameters that satisfy the reference fitting error as the plurality of second sample parameters; The obtaining of target artificial fracture parameters based on the multiple second sample parameters and the shale gas well model includes: obtaining a target probability distribution of the artificial fracture parameters based on the multiple second sample parameters; and inverting the target artificial fracture parameters based on the target probability distribution and the shale gas well model to obtain the target artificial fracture parameters.

2. The method according to claim 1, characterized in that The obtaining of a plurality of first sample parameters based on the initial probability distribution includes: Random sampling is performed in the initial probability distribution using a Markov chain Monte Carlo method to obtain the multiple first sample parameters.

3. The method according to claim 1, characterized in that The iterating the second proxy model based on the plurality of third sample parameters and the history fitting error function to obtain the first proxy model includes: The second proxy model is iterated based on the multiple third sample parameters and the historical fitting error function, and the iterated second proxy model that meets a first condition is used as the first proxy model, wherein the first condition is that the difference between the fitting error obtained by the historical fitting error function for a reference number of third sample parameters and the fitting error obtained by the iterated second proxy model is less than a reference threshold.

4. The method according to any one of claims 1 to 3, characterized in that: The target artificial fracture parameters include at least one of the length of the artificial fracture, the fracture height of the artificial fracture, the water saturation of the artificial fracture, the width of the artificial fracture, and the conductivity coefficient of the artificial fracture.

5. A device for obtaining parameters of artificial cracks, characterized in that: The device comprises: A first acquisition module is used to acquire a shale gas well model; A second acquisition module is used to obtain the initial probability distribution of artificial fracture parameters in the shale gas well model; A third acquisition module, configured to acquire a plurality of first sample parameters based on the initial probability distribution; a fourth acquisition module, configured to acquire a plurality of first sample models based on the plurality of first sample parameters and the shale gas well model; acquire a plurality of first fitting errors based on the plurality of first sample models using a historical fitting error function; establish a first proxy model based on the plurality of first sample parameters and the plurality of first fitting errors; and acquire a plurality of second sample parameters based on the first proxy model and a reference fitting error; a fifth acquisition module, configured to acquire target artificial fracture parameters based on the plurality of second sample parameters and the shale gas well model; The second acquisition module is configured to initially set the deterministic parameters and artificial fracture parameters of the shale gas well model, and obtain the values ​​of the deterministic parameters and the initial probability distribution of the artificial fracture parameters, wherein the values ​​of the deterministic parameters are fixed model parameter values ​​and are empirical values, and the initial probability distribution of the artificial fracture parameters is to assign a random and uniform estimated probability distribution to the artificial fracture parameters; The fourth acquisition module is configured to: establish a second proxy model using a K-nearest neighbor (KNN) algorithm based on the multiple first sample parameters and the multiple first fitting errors; obtain multiple third sample parameters using a Markov chain Monte Carlo method based on the second proxy model; and iterate the second proxy model based on the multiple third sample parameters and the historical fitting error function to obtain the first proxy model; The fourth acquisition module is configured to: acquire, based on the first proxy model, a plurality of sample parameters that satisfy the reference fitting error, and use the plurality of sample parameters that satisfy the reference fitting error as the plurality of second sample parameters; The fifth acquisition module is used to: obtain a target probability distribution of artificial fracture parameters based on the multiple second sample parameters; and invert the target artificial fracture parameters based on the target probability distribution and the shale gas well model to obtain the target artificial fracture parameters.

6. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one instruction, and when the at least one instruction is executed by the processor, the method for obtaining artificial crack parameters according to any one of claims 1 to 4 is implemented.

Images