Anisotropy parameter determination method, initial model training method and device

Abstract

The invention discloses an anisotropic parameter determination method and an initial model training method and device. The initial model training method comprises the following steps: acquiring longitudinal wave first arrival, transverse wave first arrival and anisotropic parameter data of each orthogonal elastic symmetry plane of grid sample points, wherein transverse waves comprise horizontal polarization transverse waves and vertical polarization transverse waves; fitting the longitudinal wave first arrival data and the transverse wave first arrival data to obtain fitting data of each axis required by a corresponding fitting ellipse; constructing a sample data set based on the required fitting data of each axis and the corresponding anisotropic parameter data; and based on the sample data set, training the preset anisotropic initial model corresponding to each orthogonal elastic symmetric plane to obtain the trained anisotropic initial model of each orthogonal elastic symmetric plane. And the solving efficiency of the anisotropic parameters can be improved.

Description

Technical Field

[0001] This invention relates to the field of geophysical exploration technology, and in particular to a method for determining anisotropic parameters, an initial model training method, and an apparatus. Background Technology

[0002] Existing geophysical exploration methods rely on isotropic models, which assume that subsurface rocks have similar physical properties in all directions. However, in the complex conditions of the actual Earth's surface, subsurface geological structures are typically anisotropic, meaning their properties vary significantly in different directions. Anisotropy has a significant impact on the propagation of subsurface sound waves, electromagnetic waves, and other electromagnetic waves. Initial modeling of anisotropic parameters provides an important model foundation for subsequent fine-grained velocity modeling; therefore, constructing an accurate anisotropic geological model is crucial for subsequent migration imaging.

[0003] Existing anisotropy parameter values ​​are mainly obtained through seismic wave travel time or amplitude inversion iteration, which has certain limitations. For example, existing techniques such as the non-hyperbolic time-distance curve equation method based on the accurate anisotropic medium equation and the phase slowness fitting estimation method have high computational complexity, resulting in high computational resource and time requirements when solving for anisotropy parameters. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide an anisotropic parameter determination method, initial model training method and apparatus to overcome or at least partially solve the above problems.

[0005] This invention provides an anisotropic initial model training method, comprising:

[0006] Acquire the longitudinal wave first arrival, transverse wave first arrival, and anisotropic parameter data of each orthogonal elastic symmetry plane of the grid sample points. The transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves.

[0007] Fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse;

[0008] A sample dataset is constructed based on the fitting data for each required axis and the corresponding anisotropic parameter data;

[0009] Based on the sample dataset, the preset anisotropic initial model corresponding to each of the orthogonal elastic symmetry surfaces is trained to obtain the trained anisotropic initial model for each of the orthogonal elastic symmetry surfaces.

[0010] In an optional embodiment, the P-wave first arrival and S-wave first arrival data are fitted to obtain the required fitting data for each axis of the corresponding fitted ellipse, including:

[0011] The longitudinal wave initial arrival is fitted to obtain the longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data of the longitudinal wave fitting ellipse; the horizontally polarized transverse wave initial arrival is fitted to obtain the first transverse wave fitting major axis data and transverse wave fitting minor axis data of the first transverse wave fitting ellipse; the vertically polarized transverse wave initial arrival is fitted to obtain the second transverse wave fitting major axis data and transverse wave fitting c-axis data of the second transverse wave fitting ellipse.

[0012] Accordingly, a sample dataset is constructed based on the fitting data for each required axis and the corresponding anisotropic parameter data; including:

[0013] A sample dataset is constructed based on the longitudinal wave fitting major axis data, longitudinal wave fitting minor axis data, first transverse wave fitting major axis data, transverse wave fitting minor axis data, second transverse wave fitting major axis data, transverse wave fitting c-axis data, and the corresponding anisotropic parameter data.

[0014] In an optional embodiment, the anisotropy parameters are Thomsen parameters, including: a first anisotropy parameter ε, a second anisotropy parameter γ, and a third anisotropy parameter δ;

[0015] A sample dataset is constructed based on the longitudinal wave fitted major axis data, longitudinal wave fitted minor axis data, first transverse wave fitted major axis data, transverse wave fitted minor axis data, second transverse wave fitted major axis data, and transverse wave fitted c-axis data, along with the corresponding anisotropic parameter data; including:

[0016] Based on the longitudinal wave fitting major axis data and the longitudinal wave fitting minor axis data, as well as the first anisotropy parameter ε data, a first sample dataset is constructed.

[0017] A second sample dataset is constructed based on the first shear wave fitted major axis data, the shear wave fitted minor axis data, and the second anisotropic parameter γ data.

[0018] A third sample dataset is constructed based on the first shear wave fitting long axis data, the shear wave fitting c-axis data, and the third anisotropic parameter δ data.

[0019] In an optional embodiment, the preset anisotropic initial model includes a first parameter solution model, a second parameter solution model, and a third parameter solution model;

[0020] Based on the sample dataset, a preset anisotropic initial model corresponding to each of the orthogonal elastic symmetry surfaces is trained to obtain the trained anisotropic initial model for each of the orthogonal elastic symmetry surfaces, including:

[0021] Based on the first sample dataset, the first parameter solution model is trained and the first anisotropic parameter data is output; based on the second sample dataset, the second parameter solution model is trained; based on the third sample dataset and the first anisotropic parameter data output by the first parameter solution model, the third parameter solution model is trained to obtain the trained anisotropic initial model.

[0022] In an optional embodiment, the anisotropic initial model training method provided by this invention further includes:

[0023] A pre-defined anisotropic initial model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis, first transverse wave fitting major axis, transverse wave fitting minor axis, second transverse wave fitting major axis, and transverse wave fitting c-axis, as well as the corresponding anisotropic parameters.

[0024] In an optional embodiment, the anisotropy parameters include: a first anisotropy parameter ε, a second anisotropy parameter γ, and a third anisotropy parameter δ. Correspondingly, the preset anisotropy initial model includes a first parameter solution model, a second parameter solution model, and a third parameter solution model.

[0025] A pre-defined anisotropic initial model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis, first transverse wave fitting major axis, transverse wave fitting minor axis, second transverse wave fitting major axis, and transverse wave fitting c-axis, along with the corresponding anisotropic parameters. This model includes:

[0026] An expression for the first parameter solution model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis and the first anisotropy parameter ε;

[0027] An expression for the second parameter solution model is constructed based on the first fitted major axis of the shear wave, the fitted minor axis of the shear wave, and the second anisotropic parameter γ.

[0028] The expression for the third parameter solution model is constructed based on the second shear wave fitting major axis, the shear wave fitting c-axis, and the third anisotropic parameter δ.

[0029] In an optional embodiment, the expression for the first parameter solution model is:

[0030]

[0031] In the formula, a1 is the major axis of the longitudinal wave fitting, b1 is the minor axis of the longitudinal wave fitting, ε is the first anisotropy parameter, and D... * (θ) is a related complex function;

[0032] The expression for solving the model using the second parameter is:

[0033]

[0034] In the formula, a2 is the first shear wave fitting major axis, b2 is the shear wave fitting minor axis, and γ is the second anisotropy parameter;

[0035] The expression for the third parameter solution model is:

[0036] The expression for the third anisotropic parameter solution model is as follows:

[0037]

[0038] In the formula, a3 is the major axis of the second shear wave fitting, c is the minor axis of the shear wave fitting, β0 is the longitudinal wave velocity, α0 is the shear wave velocity, ε is the first anisotropy parameter, and δ is the third anisotropy parameter.

[0039] In an optional embodiment, the first arrival of the longitudinal wave and the first arrival of the transverse wave are fitted using the least squares method;

[0040] Training the pre-defined anisotropic initial model is achieved using a convolutional neural network method.

[0041] Based on the same inventive concept, embodiments of the present invention also provide a method for determining anisotropy parameters, including:

[0042] Based on the orthogonal anisotropy property, the target grid points are divided into three orthogonal elastic symmetry planes;

[0043] The first arrivals of longitudinal and transverse waves are picked up from each orthogonal elastic symmetry plane; the transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves.

[0044] Fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse;

[0045] The fitting data of each required axis is input into the anisotropic initial model, and the anisotropic parameter data of each orthogonal elastic symmetry plane is obtained through the anisotropic initial model.

[0046] Based on the anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point, the anisotropic parameter data of the target grid point is obtained;

[0047] The anisotropic initial model corresponding to each orthogonal elastic symmetry plane is obtained using the training method of the anisotropic initial model described in any one of claims 1-8.

[0048] In an optional embodiment, the P-wave first arrival and S-wave first arrival data are fitted to obtain the required fitting data for each axis of the corresponding fitted ellipse; including:

[0049] The longitudinal wave initial arrival is fitted to obtain the longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data of the longitudinal wave fitting ellipse; the horizontally polarized transverse wave initial arrival is fitted to obtain the first transverse wave fitting major axis data and transverse wave fitting minor axis data of the first transverse wave fitting ellipse; the vertically polarized transverse wave initial arrival is fitted to obtain the second transverse wave fitting major axis data and transverse wave fitting c-axis data of the second transverse wave fitting ellipse.

[0050] Accordingly, the required fitting data for each axis is input into the anisotropic initial model, including:

[0051] The longitudinal wave fitting major axis data, longitudinal wave fitting minor axis data, first transverse wave fitting major axis data, transverse wave fitting minor axis data, second transverse wave fitting major axis data, and transverse wave fitting c-axis data are input into the anisotropic initial model.

[0052] In an optional embodiment, the anisotropic parameter data of each orthogonal elastic symmetry surface is output through the anisotropic initial model; including:

[0053] The first anisotropy parameter data is obtained by using the first parameter solution model based on the input longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data;

[0054] The second anisotropy parameter data is obtained by using the second parameter solution model based on the input first shear wave fitted major axis data and shear wave fitted minor axis data.

[0055] The third anisotropic parameter data is obtained by using the third parameter solution model based on the input second shear wave fitting major axis data, shear wave fitting c-axis data and the first anisotropic parameter data output by the first parameter solution model.

[0056] Based on the same inventive concept, embodiments of the present invention also provide a training apparatus for an anisotropic initial model, comprising:

[0057] The first acquisition module is used to acquire the longitudinal wave first arrival, transverse wave first arrival, and anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point. The transverse wave includes: horizontally polarized transverse wave and vertically polarized transverse wave.

[0058] The fitting module is used to fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse.

[0059] The construction module is used to construct a sample dataset based on the fitting data of each required axis and the corresponding anisotropic parameter data;

[0060] The training module is used to train the preset anisotropic initial model corresponding to each of the orthogonal elastic symmetry surfaces based on the sample dataset, so as to obtain the trained anisotropic initial model of each of the orthogonal elastic symmetry surfaces.

[0061] Based on the same inventive concept, embodiments of the present invention also provide an apparatus for determining anisotropy parameters, comprising:

[0062] The partitioning module is used to divide the target mesh points into three orthogonal elastic symmetry planes according to their orthogonal anisotropy properties.

[0063] The second acquisition module is used to pick up the first arrival of longitudinal waves and the first arrival of transverse waves on each orthogonal elastic symmetry plane of the target grid point; the transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves.

[0064] Fitting module: Fits the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse;

[0065] Calculation module: Inputs the fitting data of each required axis into the anisotropic initial model, and obtains the anisotropic parameter data of each orthogonal elastic symmetry plane through the anisotropic initial model;

[0066] Determination module: used to determine the anisotropic parameter data of the target grid point based on the anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point;

[0067] The anisotropic initial model corresponding to each orthogonal elastic symmetry plane is obtained using the training method of the anisotropic initial model described above.

[0068] Based on the same inventive concept, this embodiment of the invention also provides a computer storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described anisotropic initial model training method and the above-described anisotropic parameter determination method.

[0069] Based on the same inventive concept, this invention also provides a computer device, characterized in that it includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described anisotropic initial model training method and the above-described anisotropic parameter determination method.

[0070] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:

[0071] The anisotropic initial model training method provided in this embodiment of the invention, after obtaining the P-wave and S-wave first arrivals of each orthogonal elastic symmetry plane of the target grid point, obtains corresponding fitted ellipses by fitting the P-wave and S-wave first arrivals. Based on the data of each axis of the fitted ellipses as feature data and the corresponding anisotropic parameter data as label data, a sample dataset is constructed. This dataset is then used to train a preset anisotropic initial model, resulting in a trained anisotropic initial model. In this embodiment, the sample dataset can be directly and conveniently obtained from the fitted ellipses of the P-wave and S-wave first arrivals, with low processing difficulty and complexity. Furthermore, during the solution process using the initial model, the input data for calculating the anisotropic initial parameters can be obtained by fitting the P-wave and S-wave first arrivals. After inputting the input data into the anisotropic initial model, the anisotropic parameter data can be obtained using the anisotropic initial model. Compared with existing methods that use seismic wave travel time or amplitude inversion iteration, this method improves the efficiency of anisotropic parameter calculation.

[0072] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0073] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0074] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0075] Figure 1 This is a schematic diagram of the orthogonal anisotropic elastic symmetry plane of the grid points in an embodiment of the present invention;

[0076] Figure 2 This is a schematic diagram of the anisotropic initial model training method in an embodiment of the present invention;

[0077] Figure 3 This is a schematic diagram of the fitted ellipse obtained by fitting the first solstice in an embodiment of the present invention;

[0078] Figure 4 This is a schematic diagram of the neural network structure in an embodiment of the present invention;

[0079] Figure 5 This is a schematic diagram illustrating the definition of the phase angle in an embodiment of the present invention;

[0080] Figure 6This is a flowchart illustrating the method for determining anisotropy parameters in an embodiment of the present invention.

[0081] Figure 7 This is a schematic diagram of the anisotropic initial model training device in an embodiment of the present invention;

[0082] Figure 8 This is a schematic diagram of the anisotropy parameter determination device in an embodiment of the present invention. Detailed Implementation

[0083] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0084] It should be noted that the grid points (including grid sample points and target grid points) in this embodiment of the invention are obtained by constructing a three-dimensional model based on the real ground surface and then dividing the model into grids. Each grid point has three orthogonal elastic symmetry planes that are mutually orthogonal, for example... Figure 1 The [x1,x2], [x1,x3], and [x2,x3] symmetry planes are symmetric about x3, x2, and x1 axes, respectively, and are used to represent the VTI properties of the formation and the HTI properties of vertical fractures. The anisotropy parameters are solved in the same way for the three orthogonal elastic surfaces. Therefore, the training method for the anisotropic initial model in the following embodiments is applicable to each orthogonal elastic symmetry plane. The difference lies in that the training data for the anisotropic initial model used for different orthogonal elastic symmetry planes are the P-wave first arrival data and S-wave first arrival data of the corresponding symmetry plane. Furthermore, since the accuracy of the S-wave and P-wave first arrival measurements will have some error as the surface depth increases, the anisotropic initial model training method provided in this embodiment is particularly suitable for training near-surface anisotropic initial models.

[0085] To address the problems existing in the prior art, embodiments of the present invention provide an anisotropic initial model training method, referring to... Figure 2 As shown, it includes:

[0086] Step S101: Obtain the longitudinal wave first arrival, transverse wave first arrival, and anisotropic parameter data of each orthogonal elastic symmetry plane of the grid sample point, wherein the transverse wave includes: horizontally polarized transverse wave and vertically polarized transverse wave;

[0087] Step S102: Fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse;

[0088] Step S103: Construct a sample dataset based on the fitting data of each required axis and the corresponding anisotropic parameter data;

[0089] Step S104: Based on the sample dataset, train the preset anisotropic initial model corresponding to each orthogonal elastic symmetry surface to obtain the trained anisotropic initial model for each orthogonal elastic symmetry surface.

[0090] The anisotropic initial model training method provided in this embodiment of the invention, after obtaining the P-wave and S-wave first arrivals of each orthogonal elastic symmetry plane of the target grid point, obtains corresponding fitted ellipses by fitting the P-wave and S-wave first arrivals. Based on the data of each axis of the fitted ellipses as feature data and the corresponding anisotropic parameter data as label data, a sample dataset is constructed. This dataset is then used to train a preset anisotropic initial model, resulting in a trained anisotropic initial model. In this embodiment, the sample dataset can be directly and conveniently obtained from the fitted ellipses of the P-wave and S-wave first arrivals, with low processing difficulty and complexity. Furthermore, during the solution process using the initial model, the input data for calculating the anisotropic initial parameters can be obtained by fitting the P-wave and S-wave first arrivals. After inputting the input data into the anisotropic initial model, the anisotropic parameter data can be obtained using the anisotropic initial model. Compared with existing methods that use seismic wave travel time or amplitude inversion iteration, this method improves the efficiency of anisotropic parameter calculation.

[0091] Specifically, the first arrivals of the P-wave and S-wave are picked up using software such as CGG (Compagnie Généralede Géophysique) and iPreSeis.

[0092] Furthermore, in step S102, the first arrival data of the longitudinal wave and the first arrival data of the transverse wave are fitted to obtain the fitting data for each axis of the corresponding fitted ellipse, including:

[0093] The longitudinal wave first arrival is fitted to obtain the longitudinal wave fitted major axis data and longitudinal wave fitted minor axis data of the longitudinal wave fitted ellipse; the horizontally polarized transverse wave first arrival is fitted to obtain the first transverse wave fitted major axis data and transverse wave fitted minor axis data of the first transverse wave fitted ellipse; the vertically polarized transverse wave first arrival is fitted to obtain the second transverse wave fitted major axis data and transverse wave fitted c-axis data of the second transverse wave fitted ellipse.

[0094] Accordingly, in step S103, a sample dataset is constructed based on the required fitting data for each axis and the corresponding anisotropic parameter data; including:

[0095] A sample dataset is constructed based on the longitudinal wave fitted major axis data, longitudinal wave fitted minor axis data, first transverse wave fitted major axis data, transverse wave fitted minor axis data, second transverse wave fitted major axis data, transverse wave fitted c-axis data, and the corresponding anisotropic parameter data.

[0096] In one embodiment, the fitting of the P-wave first arrival and the S-wave first arrival is performed using the least squares method, and the expression for the fitted ellipse can be found by referring to... Figure 3 As shown in the figure, the horizontal and vertical axes represent the lengths of the cross section recorded by a single shot in two mutually perpendicular directions. The blue dots represent a certain number of sample points randomly selected from the initial to final files, and the red dots represent the shape of the equation fitted based on the sample points, i.e., the fitted ellipse.

[0097] Among them, the anisotropy parameters are Thomsen parameters, including: the first anisotropy parameter ε, the second anisotropy parameter δ, and the third anisotropy parameter γ;

[0098] A sample dataset is constructed based on the longitudinal wave fitted major axis data, longitudinal wave fitted minor axis data, first transverse wave fitted major axis data, transverse wave fitted minor axis data, second transverse wave fitted major axis data, and transverse wave fitted c-axis data, along with the corresponding anisotropy parameter data; including:

[0099] A first sample dataset is constructed based on the longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data, as well as the first anisotropy parameter ε data;

[0100] A second sample dataset is constructed based on the first shear wave fitting major axis data and the shear wave fitting minor axis data, as well as the second anisotropic parameter δ data.

[0101] A third sample dataset is constructed based on the first shear wave fitting long axis data and the shear wave fitting c axis data, as well as the third anisotropic parameter γ data.

[0102] Furthermore, the preset anisotropic initial model provided in this embodiment of the invention includes a first parameter solution model, a second parameter solution model, and a third parameter solution model;

[0103] The next step, based on the sample dataset, trains a pre-defined anisotropic initial model corresponding to each orthogonal elastic symmetry surface to obtain a trained anisotropic initial model for each orthogonal elastic symmetry surface, including:

[0104] Based on the first sample dataset, the first parameter solution model is trained and the first anisotropic parameter data is output. Based on the second sample dataset, the second parameter solution model is trained. Based on the third sample dataset and the first anisotropic parameter data output by the first parameter solution model, the third parameter solution model is trained to obtain the trained anisotropic initial model.

[0105] The training of the pre-defined anisotropic initial model is achieved using a convolutional neural network (CNN). A CNN is a deep learning model, or a multilayer perceptron similar to an artificial neural network, that can effectively predict parameters by fitting equations through appropriate architecture design and loss function settings.

[0106] The training process of the model is described in detail below, taking the training of the model for solving the first parameter as an example:

[0107] Seismic data was extracted and a sample dataset was constructed using 3D forward modeling. The sample dataset includes: the first anisotropy parameter ε data, as well as the P-wave fitted major axis data and P-wave fitted minor axis data obtained by fitting the first arrival of the P-wave using the least squares method. The sample dataset also includes: a training set and a test set.

[0108] The model for solving the first parameter is trained using the training set to train the relationship between the parameters and obtain the initial parameter model.

[0109] The model is validated using test set data. Through continuous iteration, the loss function is minimized to obtain a stable parameter model.

[0110] The convolutional neural network structure of each parameter solving model in the embodiments of the present invention refers to... Figure 4 As shown, it includes: an input layer, an output layer, four activation layers (ReLU layers in the diagram), and two fully connected layers.

[0111] In one embodiment, the anisotropic initial model training method provided by this invention may further include, before step S104:

[0112] A pre-defined anisotropic initial model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis, first transverse wave fitting major axis, transverse wave fitting minor axis, second transverse wave fitting major axis, and transverse wave fitting c-axis, as well as the corresponding anisotropic parameters.

[0113] Furthermore, since the anisotropy parameters include: a first anisotropy parameter ε, a second anisotropy parameter δ, and a third anisotropy parameter γ, the preset anisotropy initial model accordingly includes a first parameter solution model, a second parameter solution model, and a third parameter solution model;

[0114] The pre-defined anisotropic initial model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis, first transverse wave fitting major axis, transverse wave fitting minor axis, second transverse wave fitting major axis, and transverse wave fitting c-axis, along with the corresponding anisotropic parameters. This model includes:

[0115] An expression for the first parameter solution model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis and the first anisotropy parameter ε;

[0116] An expression for the second parameter solution model is constructed based on the first shear wave fitted major axis, the shear wave fitted minor axis, and the second anisotropic parameter;

[0117] The expression for the third parameter solution model is constructed based on the second shear wave fitting major axis, the shear wave fitting c-axis, and the third anisotropic parameter.

[0118] Specifically, the expression for the first parameter solution model is as follows:

[0119]

[0120] In equation (1), a1 is the major axis of the longitudinal wave fitting, b1 is the minor axis of the longitudinal wave fitting, ε is the first anisotropy parameter, and D * (θ) is a related complex function.

[0121] The expression for the second anisotropic parameter solution model is as follows:

[0122]

[0123] In the formula, a2 is the first major axis of the shear wave fitting, b2 is the minor axis of the shear wave fitting, and γ is the second anisotropy parameter.

[0124] The expression for the third anisotropic parameter solution model is as follows:

[0125]

[0126] In the formula, a3 is the major axis of the second shear wave fitting, c is the minor axis of the shear wave fitting, β0 is the longitudinal wave velocity, α0 is the shear wave velocity, ε is the first anisotropy parameter, and δ is the third anisotropy parameter.

[0127] It should be noted that, when constructing the model, the major and minor axes of the longitudinal wave fitting are parameters related to the longitudinal wave first arrival fitted ellipse; the major and minor axes of the first transverse wave fitting are parameters related to the horizontally polarized transverse wave fitted ellipse; and the minor and minor axes of the transverse wave fitting are parameters related to the vertically polarized transverse wave fitted ellipse. Furthermore, the expressions for each parameter solving module in this embodiment are merely illustrative and do not constitute a limitation on specific expressions.

[0128] The process of constructing the expression for the first parameter solution model is as follows: According to the Thomsen parameter expression, the first anisotropy parameter ε and the P-wave phase velocity v p The relationship between (θ) and the longitudinal wave velocity α0 is as follows:

[0129]

[0130] In the formula, D * (θ) is a complex function that is related to δ, ε, α0, and β0.

[0131] D * The formula for (θ) is as follows:

[0132]

[0133] in,

[0134] In this embodiment, θ is a predefined phase angle, as referenced. Figure 5 As shown, the specific definition of the phase angle is the angle between the line perpendicular to the wavefront and the vertical direction. For a detailed definition, please refer to Thomsen's paper Weak Elastic Anistropy. This embodiment of the invention will not elaborate on it.

[0135] In formula (4), when θ = 0, both sides of the formula are multiplied by the square of time t. 2 In this case, the following formula is obtained:

[0136]

[0137] Similarly, in formula (4), When the formula is multiplied by the square of time t, then... 2 In this case, the following formula is obtained:

[0138]

[0139] based on and The ratio of the first parameter to the first parameter solution model can be obtained by finding the expression (1).

[0140] Similarly, the expression for the second parameter solution model is constructed as follows: According to the Thomsen parameter expression, the second anisotropy parameter γ is related to the phase velocity v of the horizontally polarized transverse wave. sh The relationship between (θ) and the transverse wave velocity β0 is as follows:

[0141]

[0142] Similar to the expression for constructing the first parameter solution model, based on the fitting of an ellipse to a horizontally polarized transverse wave, let...

[0143]

[0144] Let θ = 0,

[0145] based on and The ratio of the two parameters can be used to obtain the expression (2) for the second parameter solution model.

[0146] The process of constructing the expression for the third parameter solution model is as follows: According to the Thomsen parameter expression, the third anisotropy parameter δ is related to the vertically polarized transverse wave phase velocity v. sv The relationship between (θ) and the transverse wave velocity β0 is as follows:

[0147]

[0148] Similar to the expression for constructing the first parameter solution model, based on the fitting of an ellipse with vertically polarized transverse waves, let...

[0149]

[0150]

[0151] Based on the ratio of c and a, the expression (3) for solving the second parameter model can be obtained.

[0152] When training the anisotropic initial model for each orthogonal elastic symmetry surface based on the sample dataset, the specific training involves the correspondence between the longitudinal wave fitting major axis, the longitudinal wave fitting minor axis and the first anisotropic parameter; the correspondence between the first transverse wave fitting major axis, the transverse wave fitting minor axis and the second anisotropic parameter; and the correspondence between the second transverse wave fitting major axis, the transverse wave fitting c-axis and the third anisotropic parameter. This ensures that after the model is trained, the required axis data of the relevant fitting ellipse of each orthogonal elastic symmetry surface can be used as input, and the anisotropic parameter data of the corresponding orthogonal elastic symmetry surface can be output through the anisotropic parameter.

[0153] When solving for anisotropy parameters, especially when calculating the first anisotropy parameter ε, regarding D * (θ), under weak anisotropy conditions, this term can be omitted; however, in complex practical cases, this term cannot be ignored in determining the anisotropy parameter, since D * The expression for (θ) is complex and involves many parameters, making it difficult to solve directly from a mathematical perspective. Therefore, this embodiment constructs an anisotropic initial model and trains the model using sample data. The training method establishes the relationship between the anisotropic parameters and the longitudinal wave fitting major axis, longitudinal wave fitting minor axis, and the first transverse wave fitting major axis, transverse wave fitting minor axis, second transverse wave fitting major axis, and transverse wave fitting c-axis of the fitted ellipse. This results in a well-trained anisotropic initial model. By using the well-trained anisotropic initial model to calculate the anisotropic parameters, the efficiency of solving the first anisotropic parameter ε can be improved, thereby improving the overall efficiency of the anisotropic parameters.

[0154] Based on the same inventive concept, this invention also provides a method for determining anisotropy parameters, the flowchart of which is shown below. Figure 6 As shown, it includes the following steps:

[0155] Step S201: Divide the target mesh points into three orthogonal elastic symmetry planes according to the orthogonal anisotropy property;

[0156] Step S202: Pick up the first arrival of longitudinal waves and transverse waves of each orthogonal elastic symmetry plane; transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves;

[0157] Step S203: Fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse;

[0158] Step S204: Input the fitting data of each required axis into the anisotropic initial model, and obtain the anisotropic parameter data of each orthogonal elastic symmetry plane through the anisotropic initial model;

[0159] Step S205: Based on the anisotropic parameters of each orthogonal elastic symmetry plane of the target grid point, obtain the anisotropic parameter data of the target grid point;

[0160] The anisotropic initial model corresponding to each orthogonal elastic symmetry plane is obtained using the training method of the anisotropic initial model described above.

[0161] In step S203, the first arrival data of the P-wave and the first arrival data of the S-wave are fitted to obtain the fitting data for each axis of the corresponding fitted ellipse; including:

[0162] The longitudinal wave first arrival is fitted to obtain the longitudinal wave fitted major axis data and longitudinal wave fitted minor axis data of the longitudinal wave fitted ellipse; the horizontally polarized transverse wave first arrival is fitted to obtain the first transverse wave fitted major axis data and transverse wave fitted minor axis data of the first transverse wave fitted ellipse; the vertically polarized transverse wave first arrival is fitted to obtain the second transverse wave fitted major axis data and transverse wave fitted c-axis data of the second transverse wave fitted ellipse.

[0163] Accordingly, in step S204, the required fitting data for each axis is input into the anisotropic initial model, including:

[0164] Input the longitudinal wave fitting major axis data, longitudinal wave fitting minor axis data, first transverse wave fitting major axis data, transverse wave fitting minor axis data, second transverse wave fitting major axis data, and transverse wave fitting c-axis data into the anisotropic initial model.

[0165] Specifically, the anisotropic parameter data of each orthogonal elastic symmetry surface is output through the anisotropic initial model; including:

[0166] The first anisotropy parameter data is obtained by using the first parameter solution model based on the input longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data;

[0167] The second anisotropy parameter data is obtained by using the second parameter solution model based on the input first shear wave fitted major axis data and shear wave fitted minor axis data.

[0168] The third anisotropic parameter data is obtained by using the third parameter solution model based on the input second shear wave fitting major axis data, shear wave fitting c-axis data, and the first anisotropic parameter data output by the first parameter solution model.

[0169] For step S205, for example, the anisotropic parameters of seismic data with [x1, x3] as the orthogonal elastic symmetry plane for a certain target grid point are denoted as ε1, δ1, and γ1. Similarly, the anisotropic parameters of seismic data with [x2, x3] as the orthogonal elastic symmetry plane are ε2, δ2, and γ2, and the anisotropic parameters of seismic data with [x1, x2] as the orthogonal elastic symmetry plane are ε3, δ3, and γ3. Then, the relevant data of each orthogonal elastic symmetry plane are summarized to obtain the anisotropic parameter data of the target grid point.

[0170] Based on the same inventive concept, referring to Figure 7 As shown, this embodiment of the invention also provides a training apparatus for an anisotropic initial model, comprising:

[0171] The first acquisition module 11 is used to acquire the longitudinal wave first arrival, transverse wave first arrival and anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point. The transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves.

[0172] Fitting module 12 is used to fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse.

[0173] Module 13 is used to construct a sample dataset based on the fitting data for each required axis and the corresponding anisotropic parameter data.

[0174] Training module 14 is used to train the preset anisotropic initial model corresponding to each orthogonal elastic symmetry surface based on the sample dataset, so as to obtain the trained anisotropic initial model of each orthogonal elastic symmetry surface.

[0175] Based on the same inventive concept, embodiments of the present invention also provide a device for determining anisotropy parameters, referring to... Figure 8 As shown, it includes:

[0176] Module 21: Used to divide the target grid points into three orthogonal elastic symmetry planes according to the orthogonal anisotropy property;

[0177] Second acquisition module 22: used to pick up the first arrival of longitudinal waves and the first arrival of transverse waves on each orthogonal elastic symmetry plane of the target grid point; the transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves;

[0178] Fitting module 23: Fits the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse;

[0179] Calculation module 24: Inputs the fitting data of each required axis into the anisotropic initial model, and obtains the anisotropic parameter data of each orthogonal elastic symmetry plane through the anisotropic initial model;

[0180] Determining module 25: used to determine the anisotropic parameter data of the target grid point based on the anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point;

[0181] The anisotropic initial model corresponding to each orthogonal elastic symmetry plane is obtained using the training method of the anisotropic initial model described above.

[0182] Based on the same inventive concept, embodiments of the present invention also provide a computer storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described anisotropic initial model training method and the above-described anisotropic parameter determination method.

[0183] Based on the same inventive concept, this embodiment of the invention also provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-described anisotropic initial model training method and the above-described anisotropic parameter determination method.

[0184] Regarding the anisotropic initial model training device, anisotropic parameter determination device, computer storage medium, and computer equipment also provided in the embodiments of the present invention, since the principles by which these devices and clients solve the problems are similar to the aforementioned anisotropic initial model training and anisotropic parameter determination methods, the implementation of these devices and clients can refer to the implementation of the aforementioned methods, and repeated details will not be repeated.

[0185] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0186] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0187] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0188] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0189] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for training an anisotropic initial model, characterized in that, include: Acquire the longitudinal wave first arrival, transverse wave first arrival, and anisotropic parameter data of each orthogonal elastic symmetry plane of the grid sample points. The transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves. Fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse; A sample dataset is constructed based on the fitting data for each required axis and the corresponding anisotropic parameter data; Based on the sample dataset, the preset anisotropic initial model corresponding to each of the orthogonal elastic symmetry surfaces is trained to obtain the trained anisotropic initial model for each of the orthogonal elastic symmetry surfaces.

2. The anisotropic initial model training method as described in claim 1, characterized in that, Fitting the first arrival data of the P-wave and S-wave yields the fitting data for each axis of the corresponding fitted ellipse, including: The longitudinal wave initial arrival is fitted to obtain the longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data of the longitudinal wave fitting ellipse; the horizontally polarized transverse wave initial arrival is fitted to obtain the first transverse wave fitting major axis data and transverse wave fitting minor axis data of the first transverse wave fitting ellipse; the vertically polarized transverse wave initial arrival is fitted to obtain the second transverse wave fitting major axis data and transverse wave fitting c-axis data of the second transverse wave fitting ellipse. Accordingly, a sample dataset is constructed based on the fitting data for each required axis and the corresponding anisotropic parameter data; including: A sample dataset is constructed based on the longitudinal wave fitting major axis data, longitudinal wave fitting minor axis data, first transverse wave fitting major axis data, transverse wave fitting minor axis data, second transverse wave fitting major axis data, transverse wave fitting c-axis data, and the corresponding anisotropic parameter data.

3. The anisotropic initial model training method as described in claim 2, characterized in that, The anisotropy parameters are Thomsen parameters, including: a first anisotropy parameter ε, a second anisotropy parameter γ, and a third anisotropy parameter δ; A sample dataset is constructed based on the longitudinal wave fitted major axis data, longitudinal wave fitted minor axis data, first transverse wave fitted major axis data, transverse wave fitted minor axis data, second transverse wave fitted major axis data, and transverse wave fitted c-axis data, along with the corresponding anisotropic parameter data; including: Based on the longitudinal wave fitting major axis data and the longitudinal wave fitting minor axis data, as well as the first anisotropy parameter ε data, a first sample dataset is constructed. A second sample dataset is constructed based on the first shear wave fitted major axis data, the shear wave fitted minor axis data, and the second anisotropic parameter γ data. A third sample dataset is constructed based on the first shear wave fitting long axis data, the shear wave fitting c-axis data, and the third anisotropic parameter δ data.

4. The anisotropic initial model training method as described in claim 3, characterized in that, The preset anisotropic initial model includes a first parameter solution model, a second parameter solution model, and a third parameter solution model; Based on the sample dataset, a preset anisotropic initial model corresponding to each of the orthogonal elastic symmetry surfaces is trained to obtain the trained anisotropic initial model for each of the orthogonal elastic symmetry surfaces, including: Based on the first sample dataset, the first parameter solution model is trained and the first anisotropic parameter data is output; based on the second sample dataset, the second parameter solution model is trained; based on the third sample dataset and the first anisotropic parameter data output by the first parameter solution model, the third parameter solution model is trained to obtain the trained anisotropic initial model.

5. The anisotropic initial model training method as described in claim 2, characterized in that, Also includes: A pre-defined anisotropic initial model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis, first transverse wave fitting major axis, transverse wave fitting minor axis, second transverse wave fitting major axis, and transverse wave fitting c-axis, as well as the corresponding anisotropic parameters.

6. The anisotropic initial model training method as described in claim 5, characterized in that, The anisotropy parameters include: a first anisotropy parameter ε, a second anisotropy parameter γ, and a third anisotropy parameter δ. Correspondingly, the preset anisotropy initial model includes a first parameter solution model, a second parameter solution model, and a third parameter solution model. A pre-defined anisotropic initial model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis, first transverse wave fitting major axis, transverse wave fitting minor axis, second transverse wave fitting major axis, and transverse wave fitting c-axis, along with the corresponding anisotropic parameters. This model includes: An expression for the first parameter solution model is constructed based on the longitudinal wave fitting major axis, longitudinal wave fitting minor axis and the first anisotropy parameter ε; An expression for the second parameter solution model is constructed based on the first fitted major axis of the shear wave, the fitted minor axis of the shear wave, and the second anisotropic parameter γ. The expression for the third parameter solution model is constructed based on the second shear wave fitting major axis, the shear wave fitting c-axis, and the third anisotropic parameter δ.

7. The anisotropic initial model training method as described in claim 6, characterized in that, The expression for the first parameter solution model is: In the formula, a1 is the major axis of the longitudinal wave fitting, b1 is the minor axis of the longitudinal wave fitting, ε is the first anisotropy parameter, and D... * (θ) is a related complex function; The expression for solving the model using the second parameter is: In the formula, a2 is the first shear wave fitting major axis, b2 is the shear wave fitting minor axis, and γ is the second anisotropy parameter; The expression for the third parameter solution model is: The expression for the third anisotropic parameter solution model is as follows: In the formula, a3 is the major axis of the second shear wave fitting, c is the minor axis of the shear wave fitting, β0 is the longitudinal wave velocity, α0 is the shear wave velocity, ε is the first anisotropy parameter, and δ is the third anisotropy parameter.

8. The anisotropic initial model training method as described in any one of claims 1-7, characterized in that, The fitting of the first arrival of the P-wave and the first arrival of the S-wave is performed using the least squares method; Training the pre-defined anisotropic initial model is achieved using a convolutional neural network method.

9. A method for determining anisotropic parameters, characterized in that, include: Based on the orthogonal anisotropy property, the target grid points are divided into three orthogonal elastic symmetry planes; The first arrivals of longitudinal and transverse waves are picked up from each orthogonal elastic symmetry plane; the transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves. Fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse; The fitting data of each required axis is input into the anisotropic initial model, and the anisotropic parameter data of each orthogonal elastic symmetry plane is obtained through the anisotropic initial model. Based on the anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point, the anisotropic parameter data of the target grid point is obtained; The anisotropic initial model corresponding to each orthogonal elastic symmetry plane is obtained using the training method of the anisotropic initial model described in any one of claims 1-8.

10. The method for determining anisotropy parameters as described in claim 9, characterized in that, Fitting the first arrival data of the P-wave and S-wave yields the fitting data for each axis of the corresponding fitted ellipse; including: The longitudinal wave initial arrival is fitted to obtain the longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data of the longitudinal wave fitting ellipse; the horizontally polarized transverse wave initial arrival is fitted to obtain the first transverse wave fitting major axis data and transverse wave fitting minor axis data of the first transverse wave fitting ellipse; the vertically polarized transverse wave initial arrival is fitted to obtain the second transverse wave fitting major axis data and transverse wave fitting c-axis data of the second transverse wave fitting ellipse. Accordingly, the required fitting data for each axis is input into the anisotropic initial model, including: The longitudinal wave fitting major axis data, longitudinal wave fitting minor axis data, first transverse wave fitting major axis data, transverse wave fitting minor axis data, second transverse wave fitting major axis data, and transverse wave fitting c-axis data are input into the anisotropic initial model.

11. The method for determining anisotropy parameters as described in claim 10, characterized in that, The anisotropic parameter data of each orthogonal elastic symmetry surface is output through the anisotropic initial model, including: The first anisotropy parameter data is obtained by using the first parameter solution model based on the input longitudinal wave fitting major axis data and longitudinal wave fitting minor axis data; The second anisotropy parameter data is obtained by using the second parameter solution model based on the input first shear wave fitted major axis data and shear wave fitted minor axis data. The third anisotropic parameter data is obtained by using the third parameter solution model based on the input second shear wave fitting major axis data, shear wave fitting c-axis data and the first anisotropic parameter data output by the first parameter solution model.

12. A training device for an anisotropic initial model, characterized in that, include: The first acquisition module is used to acquire the longitudinal wave first arrival, transverse wave first arrival, and anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point. The transverse wave includes: horizontally polarized transverse wave and vertically polarized transverse wave. The fitting module is used to fit the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse. The construction module is used to construct a sample dataset based on the fitting data of each required axis and the corresponding anisotropic parameter data; The training module is used to train the preset anisotropic initial model corresponding to each of the orthogonal elastic symmetry surfaces based on the sample dataset, so as to obtain the trained anisotropic initial model of each of the orthogonal elastic symmetry surfaces.

13. A device for determining anisotropy parameters, characterized in that, include: The partitioning module is used to divide the target mesh points into three orthogonal elastic symmetry planes according to their orthogonal anisotropy properties. The second acquisition module is used to pick up the first arrival of longitudinal waves and the first arrival of transverse waves on each orthogonal elastic symmetry plane of the target grid point; the transverse waves include: horizontally polarized transverse waves and vertically polarized transverse waves. Fitting module: Fits the first arrival data of the longitudinal wave and the first arrival data of the transverse wave to obtain the fitting data of each axis of the corresponding fitting ellipse; Calculation module: Inputs the fitting data of each required axis into the anisotropic initial model, and obtains the anisotropic parameter data of each orthogonal elastic symmetry plane through the anisotropic initial model; Determination module: used to determine the anisotropic parameter data of the target grid point based on the anisotropic parameter data of each orthogonal elastic symmetry plane of the target grid point; The anisotropic initial model corresponding to each orthogonal elastic symmetry plane is obtained using the training method of the anisotropic initial model described in any one of claims 1-8.

14. A computer storage medium, characterized in that, The computer storage medium stores computer-executable instructions, which, when executed by a processor, implement the anisotropic initial model training method according to any one of claims 1-8 and the anisotropic parameter determination method according to claims 9-11.

15. A computer device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the anisotropic initial model training method according to any one of claims 1-8 and the anisotropic parameter determination method according to any one of claims 9-11.

Images