Multi-scale fracture earthquake comprehensive prediction method and device, electronic equipment and medium

Abstract

The invention discloses a multi-scale fracture earthquake comprehensive prediction method and device, electronic equipment and a medium. The method comprises the following steps: predicting a large-scale fault associated fracture according to seismic data; predicting a mesoscale fold longitudinal seam according to the seismic dip angle and azimuth angle information; fourier series expansion is carried out on the Ruger equation, and small-scale interlayer micro-cracks are predicted through a second-order Fourier series coefficient under the credibility constraint; and performing spatial fusion on the multi-scale fractures according to a principal component analysis technology to realize comprehensive prediction of the multi-scale fractures. According to the method, the tight sandstone fracture is divided into three scales, prediction is carried out by utilizing the deconvolution generalized S transform-based frequency division ant body, the tectonic deformation strength attribute and the Fourier series coefficient based on the credibility constraint, and finally fusion is carried out by utilizing the principal component analysis method, so that multi-scale fracture comprehensive prediction is realized, the fracture prediction precision is improved, and the fracture prediction efficiency is improved. And a technical support is provided for exploration of the compact sandstone reservoir.

Description

Technical Field

[0001] The present invention relates to the field of petroleum exploration technology, and more specifically, to a multi-scale fracture earthquake comprehensive prediction method, device, electronic equipment and medium. Background Art

[0002] Fractures are an important factor in achieving high production in tight sandstone gas reservoirs. The development degree and scale of fractures provide guarantee for reservoir activation. Due to the superposition of multi-stage tectonic stress, fractures show multi-stage, multi-type and multi-scale characteristics. Studies have shown that tight sandstone fractures can be divided into large-scale fault-associated fractures controlled by faults, medium-scale fold longitudinal fractures controlled by folding, and small-scale interlayer micro-fractures controlled by lithology. They are usually characterized by the superposition of fractures of different scales in the same area. Due to the small scale and multiple types of fractures, a certain fracture prediction technology can only have good prediction accuracy for fractures of one scale. The commonly used attribute analysis technologies such as coherence, curvature, ant body, and maximum likelihood have certain effects on the prediction of large-scale fault-associated fractures, but it is difficult to accurately describe small and medium-scale fractures far away from faults, and the prediction accuracy of fold longitudinal fractures generated by folds is low; the pre-stack fracture prediction technology based on seismic anisotropy characteristics has a good effect on the prediction of small-scale interlayer microfractures, but it is affected by the quality of seismic gathers and the azimuthal anisotropy fitting method, and the prediction has strong multi-solutions; the use of multi-scale fracture prediction technologies such as RGB fusion only enhances the description effect of fractures of a certain scale, and it is difficult to accurately describe the spatial distribution characteristics of fractures of multiple scales. In summary, there is currently a lack of a seismic comprehensive prediction technology that can finely describe large-scale fault-associated fractures, medium-scale fold longitudinal fractures, and small-scale interlayer microfractures, and effectively fuse the three scale fractures. Therefore, it is impossible to accurately describe the development scale and spatial distribution of tight sandstone fractures, which restricts the exploration and evaluation in this field.

[0003] At present, a comprehensive prediction method for multi-scale fracture earthquakes needs to be developed.

[0004] The information disclosed in the background technology section of the present invention is only intended to deepen the understanding of the general background technology of the present invention, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0005] The present invention proposes a multi-scale fracture seismic comprehensive prediction method, device, electronic equipment and medium, which divides tight sandstone fractures into three scales: large-scale fault associated fractures, medium-scale fold longitudinal tension fractures and small-scale interlayer microcracks, and predicts them using frequency-divided ant bodies based on deconvolution generalized S transform, structural deformation intensity attributes and Fourier series coefficients based on credibility constraints. Finally, the principal component analysis method is used for fusion to achieve multi-scale fracture comprehensive prediction, improve the fracture prediction accuracy, provide technical support for the exploration of tight sandstone reservoirs, and have high industrial practical value and promotion and application prospects.

[0006] In a first aspect, the present disclosure provides a multi-scale fracture earthquake comprehensive prediction method, comprising:

[0007] Based on seismic data, predict large-scale fault-associated fractures;

[0008] Based on the earthquake dip and azimuth information, the longitudinal joints of mesoscale folds are predicted;

[0009] Fourier series expansion is performed on the Ruger equation, and small-scale interlayer microcracks are predicted by the second-order Fourier series coefficients under credibility constraints;

[0010] The above multi-scale cracks are spatially fused according to the principal component analysis technology to achieve comprehensive prediction of multi-scale cracks.

[0011] Preferably, predicting the associated fractures of a large-scale fault includes:

[0012] The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain multiple single-frequency data volumes.

[0013] Determine the single-frequency data volume that responds reasonably to the faults or cracks in the target layer, carry out ant tracking, and obtain multiple frequency-divided ant volumes;

[0014] The wavelet transform is used to perform frequency fusion on the multiple frequency-divided ant bodies to obtain a high-precision ant tracking result and predict the associated fractures of the large-scale fault.

[0015] Preferably, predicting the longitudinal fractures of mesoscale folds according to the seismic dip and azimuth information includes:

[0016] The seismic dip and azimuth information are used to construct structural deformation intensity attributes, calculate the degree of local structural deformation of the stratum, and predict the longitudinal tension fractures of the mesoscale folds.

[0017] Preferably, constructing the structural deformation strength attribute includes:

[0018] The dip and azimuth of a point P0 on a stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0).i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), P0 and P i The vector norm between i :

[0019]

[0020] With P0 as the center and R as the radius, calculate the average norm between P0 and other points on the same level.

[0021]

[0022] Then P0 to P i Deform i for:

[0023]

[0024] Preferably, the Fourier series expansion for the Ruger equation is:

[0025]

[0026] in, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic intensity, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic intensity.

[0027] Preferably, the confidence level is used to constrain the second-order Fourier series coefficients:

[0028] Assume that the signal-to-noise ratio of a sampling point of 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0;

[0029] The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i, plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for:

[0030] b 2i '=b 2i *Credibility i .

[0031] Preferably, spatially merging the multi-scale fractures according to the principal component analysis technique to achieve comprehensive prediction of multi-scale fractures includes:

[0032] Establish a sample matrix in three-dimensional space based on multi-scale fractures;

[0033] The sample matrix is ​​standardized and the χ ik Transformation;

[0034] Calculate the correlation coefficient matrix of the sample matrix;

[0035] The correlation coefficient matrix is ​​solved to determine the principal components of the multi-scale cracks, so as to fuse the components with the largest contribution rate to the multi-scale cracks and realize the comprehensive prediction of the multi-scale cracks.

[0036] In a second aspect, the embodiment of the present disclosure further provides a multi-scale fracture earthquake comprehensive prediction device, comprising:

[0037] Large-scale fracture prediction module, which predicts large-scale fault-associated fractures based on seismic data;

[0038] The mesoscale fracture prediction module predicts the longitudinal fractures of mesoscale folds based on the earthquake dip and azimuth information;

[0039] Small-scale crack prediction module, which performs Fourier series expansion on the Ruger equation and predicts small-scale interlayer microcracks through the second-order Fourier series coefficients under credibility constraints;

[0040] The spatial fusion module spatially fuses the above multi-scale cracks based on the principal component analysis technology to achieve comprehensive prediction of multi-scale cracks.

[0041] Preferably, predicting the associated fractures of a large-scale fault includes:

[0042] The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain multiple single-frequency data volumes.

[0043] Determine the single-frequency data volume that responds reasonably to the faults or cracks in the target layer, carry out ant tracking, and obtain multiple frequency-divided ant volumes;

[0044] The wavelet transform is used to perform frequency fusion on the multiple frequency-divided ant bodies to obtain a high-precision ant tracking result and predict the associated fractures of the large-scale fault.

[0045] Preferably, predicting the longitudinal fractures of mesoscale folds according to the seismic dip and azimuth information includes:

[0046] The seismic dip and azimuth information are used to construct structural deformation intensity attributes, calculate the degree of local structural deformation of the stratum, and predict the longitudinal tension fractures of the mesoscale folds.

[0047] Preferably, constructing the structural deformation strength attribute includes:

[0048] The dip and azimuth of a point P0 on a stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0). i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), P0 and P i The vector norm between i :

[0049]

[0050] With P0 as the center and R as the radius, calculate the average norm between P0 and other points on the same level.

[0051]

[0052] Then P0 to P i Deform i for:

[0053]

[0054] Preferably, the Fourier series expansion for the Ruger equation is:

[0055]

[0056] in, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic intensity, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic intensity.

[0057] Preferably, the confidence level is used to constrain the second-order Fourier series coefficients:

[0058] Assume that the signal-to-noise ratio of a sampling point of 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0;

[0059] The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i , plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for:

[0060] b 2i '=b 2i *Credibility i .

[0061] Preferably, spatially merging the multi-scale fractures according to the principal component analysis technique to achieve comprehensive prediction of multi-scale fractures includes:

[0062] Establish a sample matrix in three-dimensional space based on multi-scale fractures;

[0063] The sample matrix is ​​standardized and the χ ik Transformation;

[0064] Calculate the correlation coefficient matrix of the sample matrix;

[0065] The correlation coefficient matrix is ​​solved to determine the principal components of the multi-scale cracks, so as to fuse the components with the largest contribution rate to the multi-scale cracks and realize the comprehensive prediction of the multi-scale cracks.

[0066] In a third aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:

[0067] A memory storing executable instructions;

[0068] A processor runs the executable instructions in the memory to implement the multi-scale fracture earthquake comprehensive prediction method.

[0069] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium storing a computer program, which implements the multi-scale fracture earthquake comprehensive prediction method when executed by a processor.

[0070] Its beneficial effects are:

[0071] (1) The prediction of large-scale fault-associated fractures is carried out using the frequency-divided ant body based on the deconvolution generalized S transform. This method has stronger noise resistance than the conventional S transform and comprehensively utilizes the advantages of ant bodies with different frequencies. It is more accurate than the conventional ant body and improves the prediction accuracy of large-scale fault-associated fractures.

[0072] (2) The structural deformation intensity attribute was constructed to predict the longitudinal fractures of mesoscale folds. This method uses seismic dip and azimuth data to calculate the local fold deformation degree of the stratum, and then predict the development degree of the longitudinal fractures of the folds. Compared with conventional curvature attributes, the prediction accuracy is greatly improved.

[0073] (3) The prediction of small-scale interlayer microcracks is carried out using Fourier series coefficients based on credibility constraints. Compared with the traditional Ruger azimuthal anisotropic ellipse fitting, the Fourier series coefficient method is more accurate and adds credibility constraints, which further improves the prediction accuracy and reliability of small-scale interlayer microcracks.

[0074] (4) The principal component analysis technique is used to effectively fuse the cracks of three scales. Compared with the conventional RGB fusion technique, this technique can better reflect the contribution of cracks of different scales to a certain structural position in space.

[0075] The methods and apparatus of the present invention have other features and advantages that will be apparent from, or will be described in detail in, the accompanying drawings and subsequent detailed descriptions incorporated herein, which together serve to explain the specific principles of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0077] Figure 1 A flow chart showing the steps of the multi-scale fracture earthquake comprehensive prediction method according to the present invention.

[0078] Figure 2 A flowchart showing the steps of a multi-scale fracture earthquake comprehensive prediction method according to an embodiment of the present invention.

[0079] Figure 3a , Figure 3b Schematic diagrams of high-precision ant tracking attribute slicing and conventional coherent attribute slicing according to an embodiment of the present invention are respectively shown.

[0080] Figure 4a , Figure 4b Schematic diagrams of a wrinkle deformation strength attribute profile and a curvature attribute profile according to an embodiment of the present invention are respectively shown.

[0081] Figure 5a , Figure 5b Schematic diagrams of a prediction plane of a C-well region using a Fourier series coefficient method and a prediction plane of a C-well region using a Ruger ellipse fitting method are respectively shown according to an embodiment of the present invention.

[0082] Figure 6a , Figure 6b Schematic diagrams of a multi-scale fracture fusion profile through Well B and imaging logging data of Well B according to an embodiment of the present invention are respectively shown.

[0083] Figure 7 A block diagram of a multi-scale fracture earthquake comprehensive prediction device according to an embodiment of the present invention is shown.

[0084] Description of reference numerals:

[0085] 201. Large-scale crack prediction module; 202. Medium-scale crack prediction module; 203. Small-scale crack prediction module; 204. Spatial fusion module. DETAILED DESCRIPTION

[0086] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0087] Figure 1 A flow chart showing the steps of the multi-scale fracture earthquake comprehensive prediction method according to the present invention.

[0088] like Figure 1 As shown, the present invention provides a multi-scale fracture earthquake comprehensive prediction method, comprising:

[0089] Step 101, predicting large-scale fault associated fractures based on seismic data;

[0090] Step 102, predicting the longitudinal joints of mesoscale folds based on earthquake dip and azimuth information;

[0091] Step 103, performing Fourier series expansion on the Ruger equation, and predicting small-scale interlayer microcracks through second-order Fourier series coefficients under credibility constraints;

[0092] Step 104, spatially fuse the multi-scale fractures according to the principal component analysis technology to achieve comprehensive prediction of multi-scale fractures.

[0093] In one example, predicting fractures associated with large-scale faults includes:

[0094] The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain multiple single-frequency data volumes.

[0095] Determine the single-frequency data volume that responds reasonably to the faults or cracks in the target layer, carry out ant tracking, and obtain multiple frequency-divided ant volumes;

[0096] Wavelet transform is used to perform frequency fusion on multiple frequency-divided ant bodies to obtain high-precision ant tracking results and predict large-scale fault-associated fractures.

[0097] In one example, based on earthquake dip and azimuth information, predicting the longitudinal fractures of mesoscale folds includes:

[0098] The seismic dip and azimuth information is used to construct structural deformation intensity attributes, calculate the degree of local structural deformation of the strata, and predict the longitudinal tension fractures of mesoscale folds.

[0099] In one example, constructing a construction deformation strength attribute includes:

[0100] The dip and azimuth of a point P0 on a stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0). i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), P0 and P i The vector norm between i :

[0101]

[0102] With P0 as the center and R as the radius, calculate the average norm between P0 and other points on the same level.

[0103]

[0104] Then P0 to P i Deformi for:

[0105]

[0106] In one example, the Fourier series expansion for the Ruger equation is:

[0107]

[0108] in, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic intensity, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic intensity.

[0109] In one example, the confidence level is used to constrain the second-order Fourier series coefficients:

[0110] Assume that the signal-to-noise ratio of a sampling point of 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0;

[0111] The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i , plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for:

[0112] b 2i '=b 2i *Credibility i .

[0113] In one example, the multi-scale fractures are spatially fused according to the principal component analysis technology to achieve comprehensive prediction of multi-scale fractures, including:

[0114] Establish a sample matrix in three-dimensional space based on multi-scale fractures;

[0115] The sample matrix was standardized and the χ ik Transformation;

[0116] Calculate the correlation coefficient matrix of the sample matrix;

[0117] The correlation coefficient matrix is ​​solved to determine the principal components of multi-scale fractures, and the components with the largest contribution rate to multi-scale fractures are integrated to achieve comprehensive prediction of multi-scale fractures.

[0118] Specifically, the seismic data are processed by frequency division based on the deconvolution generalized S transform, and then the frequency division body is optimized for ant tracking. Finally, the fusion is completed by combining with wavelet transform to carry out large-scale fault associated fracture prediction.

[0119] The fractures associated with large-scale faults are related to the distribution of fractures in the study area. Ant tracking technology is a commonly used method for characterizing faults and fractures. The ant tracking algorithm has the advantages of fast speed and less human intervention, but its noise resistance is poor and its resolution is relatively low. Therefore, a high-precision frequency-division ant technology based on deconvolution generalized S transform was studied to improve the prediction accuracy of fractures associated with large-scale faults.

[0120] First, the deconvolution generalized S transform technology is used to perform frequency division processing on seismic data. Its basic expression is:

[0121]

[0122] Among them, u(t) is the seismic signal in the time domain, h(t-τ,f) is the improved S-transform Gaussian window function, which is defined as:

[0123]

[0124] In the above formula, λ and p (λ, p>0) are the time-frequency adjustment parameters of the function, and τ is the time position of the wavelet function corresponding to the frequency f. When λ=1 and p=1, the above formula is the S-transform wavelet function, and when p=0, the formula is the short-time Fourier transform.

[0125] As a nonlinear time-frequency analysis method, Wigner distribution has high resolution and time-frequency focusing. The Wigner distribution of the generalized S-transform spectrum of the seismic signal u(t) is:

[0126]

[0127] On this basis, according to the Lucy-Richardson deconvolution algorithm, the optimal time-frequency distribution of the original signal is obtained through iterative calculation, while suppressing the cross-term interference caused by the Wigner distribution. The expression of the algorithm is:

[0128]

[0129] Where W x (k) is the kth iteration result of the Wigner distribution of the seismic signal x(t), W his the Wigner distribution of its window function h(t,f), the superscript “-” indicates the result of the last iteration, k is the number of iterations, and * is the correlation operator. is a two-dimensional convolution operator. The initial boundary condition of this formula is: W x - (0) = GST x .

[0130] The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain single-frequency data volumes.

[0131] Secondly, the single-frequency data volume with reasonable response to the development of faults or cracks in the target layer is selected for ant tracking to obtain the divided-frequency ant tracking results.

[0132] Finally, wavelet transform is used to perform frequency fusion on multiple frequency-divided ant bodies to obtain high-precision ant tracking results and predict large-scale fault-associated fractures.

[0133] The structural deformation intensity attributes were constructed using seismic dip and azimuth information, the degree of local structural deformation of the strata was calculated, and the prediction of mesoscale fold longitudinal fractures was carried out.

[0134] Mesoscale fold longitudinal fractures are generally distributed in the anticline axis, and their scale is smaller than that of fault-associated fractures. In particular, the continuity of seismic phase axes at the top of relatively wide and gentle anticlines is usually good, and it is difficult to identify them using conventional continuity-based post-stack attributes. Therefore, in order to better predict mesoscale fold longitudinal fractures, a structural deformation intensity attribute is constructed based on seismic dip and azimuth information. The attribute is defined as follows:

[0135] Assume that the dip and azimuth of a point P0 on a certain stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0), and the adjacent point P i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), define P0 and P i The vector norm between i :

[0136]

[0137] With P0 as the center and R as the radius, calculate the norm between P0 and other points on the same level, and calculate the average norm.

[0138]

[0139] Then P0 to P i Deform i It can be defined as:

[0140]

[0141] The structural deformation intensity attribute Deform is used to characterize the degree of local structural deformation. It can be used to detect the amplitude of fold deformation in anticline and syncline structures. The amplitude of fold deformation is proportional to the degree of development of longitudinal fold joints. Therefore, the structural deformation intensity attribute can be used to characterize the degree of development of longitudinal fold joints.

[0142] The Ruger equation is expanded using Fourier series, and Fourier series coefficient fitting is carried out. Under the constraint of credibility, the second-order Fourier series coefficients are used to characterize the anisotropic strength and predict small-scale interlayer microcracks.

[0143] At present, the prediction of fractures using seismic azimuthal anisotropy is mainly based on the Ruger equation:

[0144]

[0145] Where: θ is the incident angle, φ is the azimuth angle, iso is the orientation of the crack, R(θ,φ) is the longitudinal wave reflection coefficient related to the incident angle θ and the azimuth angle φ, and A iso , B iso , C iso is the isotropic term, B ani is the anisotropic term, δ (V) , ε (V) is the anisotropy coefficient of the medium, and Δ[·] represents the difference between the physical quantities of the upper and lower interfaces.

[0146] When the incident angle is small, the last term in the above formula can be ignored, and the above formula can be approximated as:

[0147] R(θ,φ)≈A iso +(B iso +B ani cos 2 (φ-φ iso ))sin 2 θ

[0148] At present, the above formula is usually used to fit the seismic azimuthal anisotropy ellipse to carry out small-scale crack prediction, but the ellipse fitting method often reduces the crack prediction accuracy due to the presence of singular points in the fitting points. In order to further improve the crack prediction accuracy, the Ruger equation is first expanded by Fourier series:

[0149]

[0150] In the formula, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic strength, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic strength. Since b4 is easily affected by seismic noise and the result is unstable, b2 is generally used to predict the direction and density of fracture development.

[0151] Usually, the second-order Fourier coefficient b2 fitting is directly carried out using the azimuth seismic data. Since the signal-to-noise ratio of the seismic data is greatly reduced near large fault zones and in the boundary areas of three-dimensional seismic data, there are abnormal values ​​in the b2 fitting, which affects the accuracy of fracture prediction. Therefore, the present invention adds a credibility constraint.

[0152] Assume that the signal-to-noise ratio of a sampling point in 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0.

[0153] The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i , plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for:

[0154] b 2i '=b 2i *Credibility i

[0155] Using the credibility constraint b 2i 'To carry out crack prediction and improve the stability and reliability of interlayer microcrack prediction.

[0156] The principal component analysis technique is used to spatially integrate the predicted fault-associated fractures, fold longitudinal fractures and interlayer microcracks to achieve comprehensive prediction of multi-scale fractures.

[0157] Assuming that the high-precision frequency division ant body is Y1, the wrinkle strength body is Y2, the second-order Fourier series coefficient body is Y3, and the total number of sampling points in three-dimensional space is N, then the N micro-sample matrix Y:

[0158]

[0159] The sample matrix is ​​first standardized, and the χ ik Transformation:

[0160]

[0161]

[0162]

[0163] Compute the correlation matrix for a data sample:

[0164]

[0165] Solving the correlation coefficient matrix characteristic equation yields three non-negative eigenvalues, and the corresponding eigenvectors corresponding to the eigenvalues ​​are {g 13 ,g 23 ,…g m3};satisfy:

[0166]

[0167] And there exists:

[0168]

[0169] Then the new variables x1, x2, …x m are called the first, second, ..., mth principal components of the original attributes. m is the eigenvector {g 13 ,g 23 ,…g m3}Corresponding eigenvalue. Ratio Reflects the component x i Contribution rate to the overall information: the larger the ratio, the more original information data is retained.

[0170] Select the component x that contributes most to Y1, Y2, and Y3 respectively i ,x j ,x k Fusion is performed to carry out comprehensive prediction of multi-scale cracks.

[0171] The present invention also provides a multi-scale fracture earthquake comprehensive prediction device, comprising:

[0172] Large-scale fracture prediction module, which predicts large-scale fault-associated fractures based on seismic data;

[0173] The mesoscale fracture prediction module predicts the longitudinal fractures of mesoscale folds based on the earthquake dip and azimuth information;

[0174] Small-scale crack prediction module, which performs Fourier series expansion on the Ruger equation and predicts small-scale interlayer microcracks through the second-order Fourier series coefficients under credibility constraints;

[0175] The spatial fusion module spatially fuses the above multi-scale cracks based on the principal component analysis technology to achieve comprehensive prediction of multi-scale cracks.

[0176] In one example, predicting fractures associated with large-scale faults includes:

[0177] The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain multiple single-frequency data volumes.

[0178] Determine the single-frequency data volume that responds reasonably to the faults or cracks in the target layer, carry out ant tracking, and obtain multiple frequency-divided ant volumes;

[0179] Wavelet transform is used to perform frequency fusion on multiple frequency-divided ant bodies to obtain high-precision ant tracking results and predict large-scale fault-associated fractures.

[0180] In one example, based on earthquake dip and azimuth information, predicting the longitudinal fractures of mesoscale folds includes:

[0181] The seismic dip and azimuth information is used to construct structural deformation intensity attributes, calculate the degree of local structural deformation of the strata, and predict the longitudinal tension fractures of mesoscale folds.

[0182] In one example, constructing a construction deformation strength attribute includes:

[0183] The dip and azimuth of a point P0 on a stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0). i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), P0 and P i The vector norm between i :

[0184]

[0185] With P0 as the center and R as the radius, calculate the average norm between P0 and other points on the same level.

[0186]

[0187] Then P0 to P iDeform i for:

[0188]

[0189] In one example, the Fourier series expansion for the Ruger equation is:

[0190]

[0191] in, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic intensity, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic intensity.

[0192] In one example, the confidence level is used to constrain the second-order Fourier series coefficients:

[0193] Assume that the signal-to-noise ratio of a sampling point of 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0;

[0194] The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i , plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for:

[0195] b 2i '=b 2i *Credibility i .

[0196] In one example, the multi-scale fractures are spatially fused according to the principal component analysis technology to achieve comprehensive prediction of multi-scale fractures, including:

[0197] Establish a sample matrix in three-dimensional space based on multi-scale fractures;

[0198] The sample matrix was standardized and the χ ik Transformation;

[0199] Calculate the correlation coefficient matrix of the sample matrix;

[0200] The correlation coefficient matrix is ​​solved to determine the principal components of multi-scale fractures, and the components with the largest contribution rate to multi-scale fractures are integrated to achieve comprehensive prediction of multi-scale fractures.

[0201] Specifically, the seismic data are processed by frequency division based on the deconvolution generalized S transform, and then the frequency division body is optimized for ant tracking. Finally, the fusion is completed by combining with wavelet transform to carry out large-scale fault associated fracture prediction.

[0202] The fractures associated with large-scale faults are related to the distribution of fractures in the study area. Ant tracking technology is a commonly used method for characterizing faults and fractures. The ant tracking algorithm has the advantages of fast speed and less human intervention, but its noise resistance is poor and its resolution is relatively low. Therefore, a high-precision frequency-division ant technology based on deconvolution generalized S transform was studied to improve the prediction accuracy of fractures associated with large-scale faults.

[0203] First, the deconvolution generalized S transform technology is used to perform frequency division processing on seismic data. Its basic expression is:

[0204]

[0205] Among them, u(t) is the seismic signal in the time domain, h(t-τ,f) is the improved S-transform Gaussian window function, which is defined as:

[0206]

[0207] In the above formula, λ and p (λ, p>0) are the time-frequency adjustment parameters of the function, and τ is the time position of the wavelet function corresponding to the frequency f. When λ=1 and p=1, the above formula is the S-transform wavelet function, and when p=0, the formula is the short-time Fourier transform.

[0208] As a nonlinear time-frequency analysis method, Wigner distribution has high resolution and time-frequency focusing. The Wigner distribution of the generalized S-transform spectrum of the seismic signal u(t) is:

[0209]

[0210] On this basis, according to the Lucy-Richardson deconvolution algorithm, the optimal time-frequency distribution of the original signal is obtained through iterative calculation, while suppressing the cross-term interference caused by the Wigner distribution. The expression of the algorithm is:

[0211]

[0212] Where W x (k) is the kth iteration result of the Wigner distribution of the seismic signal x(t), W his the Wigner distribution of its window function h(t,f), the superscript “-” indicates the result of the last iteration, k is the number of iterations, and * is the correlation operator. is a two-dimensional convolution operator. The initial boundary condition of this formula is: W x - (0) = GST x .

[0213] The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain single-frequency data volumes.

[0214] Secondly, the single-frequency data volume with reasonable response to the development of faults or cracks in the target layer is selected for ant tracking to obtain the divided-frequency ant tracking results.

[0215] Finally, wavelet transform is used to perform frequency fusion on multiple frequency-divided ant bodies to obtain high-precision ant tracking results and predict large-scale fault-associated fractures.

[0216] The structural deformation intensity attributes were constructed using seismic dip and azimuth information, the degree of local structural deformation of the strata was calculated, and the prediction of mesoscale fold longitudinal fractures was carried out.

[0217] Mesoscale fold longitudinal fractures are generally distributed in the anticline axis, and their scale is smaller than that of fault-associated fractures. In particular, the continuity of seismic phase axes at the top of relatively wide and gentle anticlines is usually good, and it is difficult to identify them using conventional continuity-based post-stack attributes. Therefore, in order to better predict mesoscale fold longitudinal fractures, a structural deformation intensity attribute is constructed based on seismic dip and azimuth information. The attribute is defined as follows:

[0218] Assume that the dip and azimuth of a point P0 on a certain stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0), and the adjacent point P i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), define P0 and P i The vector norm between i :

[0219]

[0220] With P0 as the center and R as the radius, calculate the norm between P0 and other points on the same level, and calculate the average norm.

[0221]

[0222] Then P0 to P i Deform i It can be defined as:

[0223]

[0224] The structural deformation intensity attribute Deform is used to characterize the degree of local structural deformation. It can be used to detect the amplitude of fold deformation in anticline and syncline structures. The amplitude of fold deformation is proportional to the degree of development of longitudinal fold joints. Therefore, the structural deformation intensity attribute can be used to characterize the degree of development of longitudinal fold joints.

[0225] The Ruger equation is expanded using Fourier series, and Fourier series coefficient fitting is carried out. Under the constraint of credibility, the second-order Fourier series coefficients are used to characterize the anisotropic strength and predict small-scale interlayer microcracks.

[0226] At present, the prediction of fractures using seismic azimuthal anisotropy is mainly based on the Ruger equation:

[0227]

[0228] Where: θ is the incident angle, φ is the azimuth angle, iso is the orientation of the crack, R(θ,φ) is the longitudinal wave reflection coefficient related to the incident angle θ and the azimuth angle φ, and A iso , B iso , C iso is the isotropic term, B ani is the anisotropic term, δ (V) , ε (V) is the anisotropy coefficient of the medium, and Δ[·] represents the difference between the physical quantities of the upper and lower interfaces.

[0229] When the incident angle is small, the last term in the above formula can be ignored, and the above formula can be approximated as:

[0230] R(θ,φ)≈A iso +(B iso +B ani cos 2 (φ-φ iso ))sin 2 θ

[0231] At present, the above formula is usually used to fit the seismic azimuthal anisotropy ellipse to carry out small-scale crack prediction, but the ellipse fitting method often reduces the crack prediction accuracy due to the presence of singular points in the fitting points. In order to further improve the crack prediction accuracy, the Ruger equation is first expanded by Fourier series:

[0232]

[0233] In the formula, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic strength, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic strength. Since b4 is easily affected by seismic noise and the result is unstable, b2 is generally used to predict the direction and density of fracture development.

[0234] Usually, the second-order Fourier coefficient b2 fitting is directly carried out using the azimuth seismic data. Since the signal-to-noise ratio of the seismic data is greatly reduced near large fault zones and in the boundary areas of three-dimensional seismic data, there are abnormal values ​​in the b2 fitting, which affects the accuracy of fracture prediction. Therefore, the present invention adds a credibility constraint.

[0235] Assume that the signal-to-noise ratio of a sampling point in 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0.

[0236] The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i , plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for:

[0237] b 2i '=b 2i *Credibility i

[0238] Using the credibility constraint b 2i 'To carry out crack prediction and improve the stability and reliability of interlayer microcrack prediction.

[0239] The principal component analysis technique is used to spatially integrate the predicted fault-associated fractures, fold longitudinal fractures and interlayer microcracks to achieve comprehensive prediction of multi-scale fractures.

[0240] Assuming that the high-precision frequency division ant body is Y1, the wrinkle strength body is Y2, the second-order Fourier series coefficient body is Y3, and the total number of sampling points in three-dimensional space is N, then the N micro-sample matrix Y:

[0241]

[0242] The sample matrix is ​​first standardized, and the χ ik Transformation:

[0243]

[0244]

[0245]

[0246] Compute the correlation matrix for a data sample:

[0247]

[0248] Solving the correlation coefficient matrix characteristic equation yields three non-negative eigenvalues, and the corresponding eigenvectors corresponding to the eigenvalues ​​are {g 13 ,g 23 ,…g m3};satisfy:

[0249]

[0250] And there exists:

[0251]

[0252] Then the new variables x1, x2, …x m are called the first, second, ..., mth principal components of the original attributes. m is the eigenvector {g 13 ,g 23 ,…g m3}Corresponding eigenvalue. Ratio Reflects the component x i Contribution rate to the overall information: the larger the ratio, the more original information data is retained.

[0253] Select the component x that contributes most to Y1, Y2, and Y3 respectively i ,x j ,x k Fusion is performed to carry out comprehensive prediction of multi-scale cracks.

[0254] The present invention also provides an electronic device, which includes: a memory storing executable instructions; and a processor running the executable instructions in the memory to implement the above-mentioned multi-scale fracture earthquake comprehensive prediction method.

[0255] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the multi-scale fracture earthquake comprehensive prediction method is implemented.

[0256] To facilitate understanding of the solutions and effects of the embodiments of the present invention, four specific application examples are given below. Those skilled in the art should understand that the examples are only for facilitating understanding of the present invention, and any specific details thereof are not intended to limit the present invention in any way.

[0257] Example 1

[0258] Figure 2 A flowchart showing the steps of a multi-scale fracture earthquake comprehensive prediction method according to an embodiment of the present invention.

[0259] like Figure 2 As shown in the figure, the tight sandstone fractures are divided into three scales: large-scale fault-associated fractures, medium-scale fold longitudinal fractures and small-scale interlayer microcracks. They are predicted using frequency-divided ant bodies based on deconvolution generalized S transform, structural deformation intensity attributes and Fourier series coefficients based on credibility constraints. Finally, they are fused using principal component analysis to achieve comprehensive prediction of multi-scale fractures.

[0260] Figure 3a , Figure 3b Schematic diagrams of high-precision ant tracking attribute slices and conventional coherent attribute slices according to an embodiment of the present invention are respectively shown. In this embodiment, compared with conventional coherent attribute slices, high-precision frequency-divided ant tracking attributes can accurately depict the distribution range of faults and their associated fractures.

[0261] Figure 4a , Figure 4b The schematic diagrams of the fold deformation intensity attribute section and the curvature attribute section according to an embodiment of the present invention are respectively shown. In this example, compared with the curvature attribute, the structural deformation intensity attribute can reflect the degree of fold deformation in the anticline axis, and further predict the development intensity of the fold longitudinal fracture. The yellow box in Well A is the fracture development layer section of the actual drilling, and the prediction result is consistent with the actual drilling.

[0262] Figure 5a , Figure 5b Schematic diagrams of the prediction plane of the Fourier series coefficient method and the prediction plane of the Ruger ellipse fitting method in the C well area according to an embodiment of the present invention are respectively shown. In this example, the Fourier series coefficient method with credibility constraint predicts that the fractures in a certain layer section of the C well area are relatively developed, while the Ruger ellipse fitting method predicts that the fracture development degree of this section is poor, but the measured imaging logging data of this layer section of the C well shows that the fractures in this section are developed, and the test obtains a medium-yield industrial gas flow, which is consistent with the prediction result of the Fourier series coefficient method with credibility constraint.

[0263] Figure 6a , Figure 6bThe schematic diagrams of the multi-scale fracture fusion section of Well B and the imaging logging data of Well B according to an embodiment of the present invention are respectively shown. In this example, the three-scale fractures are fused by principal component analysis, which can depict the spatial distribution of fractures inside the dense sandstone. The first component of the principal component analysis is mainly red on both sides of the fault zone, reflecting the distribution of large-scale fault associated fractures, the second component is mainly yellow in the fold axis, reflecting the distribution of medium-scale fold longitudinal fractures, and the third component is mainly blue between layers, reflecting the distribution of small-scale interlayer microfractures. Figure 4 (b) is the imaging logging data of the red frame section of Well B. The first track is the FMI static image, the second track is the measured depth, the third track is the FMI dynamic image and various geological interfaces picked up on the FMI dynamic image, and the fourth track is the true dipole tadpole map picked up on the FMI image. It can be clearly seen that the development degree of fractures of different scales in this section is high, which is consistent with the comprehensive prediction result of multi-scale fractures.

[0264] In summary, the present invention is based on the multi-scale fracture prediction theory, and uses the frequency division ant based on the deconvolution generalized S transform, the structural deformation intensity and the Fourier series coefficient based on the credibility constraint to respectively predict the fault associated fractures, the fold longitudinal fractures and the interlayer micro-cracks, and then uses the principal component analysis technology to fuse the three scale fractures, and comprehensively predict the spatial distribution of multi-scale fractures in the tight sandstone. It has high industrial practical value and promotion and application prospects.

[0265] Example 2

[0266] Figure 7 A block diagram of a multi-scale fracture earthquake comprehensive prediction device according to an embodiment of the present invention is shown.

[0267] like Figure 7 As shown, the multi-scale fracture earthquake comprehensive prediction device comprises:

[0268] A large-scale fracture prediction module 201 predicts large-scale fault associated fractures based on seismic data;

[0269] A mesoscale fracture prediction module 202 predicts mesoscale fold longitudinal fractures based on earthquake dip and azimuth information;

[0270] A small-scale crack prediction module 203 performs Fourier series expansion on the Ruger equation and predicts small-scale interlayer microcracks through second-order Fourier series coefficients under credibility constraints;

[0271] The spatial fusion module 204 performs spatial fusion on the multi-scale fractures according to the principal component analysis technology to achieve comprehensive prediction of multi-scale fractures.

[0272] As an alternative, the prediction of large-scale fault associated fractures includes:

[0273] The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain multiple single-frequency data volumes.

[0274] Determine the single-frequency data volume that responds reasonably to the faults or cracks in the target layer, carry out ant tracking, and obtain multiple frequency-divided ant volumes;

[0275] Wavelet transform is used to perform frequency fusion on multiple frequency-divided ant bodies to obtain high-precision ant tracking results and predict large-scale fault-associated fractures.

[0276] As an optional solution, based on the seismic dip and azimuth information, the prediction of mesoscale fold longitudinal fractures includes:

[0277] The seismic dip and azimuth information is used to construct structural deformation intensity attributes, calculate the degree of local structural deformation of the strata, and predict the longitudinal tension fractures of mesoscale folds.

[0278] Optionally, construct the tectonic deformation strength properties including:

[0279] The dip and azimuth of a point P0 on a stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0). i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), P0 and P i The vector norm between i :

[0280]

[0281] With P0 as the center and R as the radius, calculate the average norm between P0 and other points on the same level.

[0282]

[0283] Then P0 to P i Deform i for:

[0284]

[0285] As an alternative, the Fourier series expansion of the Ruger equation is:

[0286]

[0287] in, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic intensity, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic intensity.

[0288] As an alternative, use the confidence constraint on the second-order Fourier series coefficients:

[0289] Assume that the signal-to-noise ratio of a sampling point of 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0;

[0290] The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i , plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for:

[0291] b 2i '=b 2i *Credibility i .

[0292] As an optional solution, the above multi-scale fractures are spatially fused according to the principal component analysis technology to achieve comprehensive prediction of multi-scale fractures, including:

[0293] Establish a sample matrix in three-dimensional space based on multi-scale fractures;

[0294] The sample matrix was standardized and the χ ik Transformation;

[0295] Calculate the correlation coefficient matrix of the sample matrix;

[0296] The correlation coefficient matrix is ​​solved to determine the principal components of multi-scale fractures, and the components with the largest contribution rate to multi-scale fractures are integrated to achieve comprehensive prediction of multi-scale fractures.

[0297] Example 3

[0298] The present disclosure provides an electronic device, which includes: a memory storing executable instructions; a processor running the executable instructions in the memory to implement the above-mentioned multi-scale fracture earthquake comprehensive prediction method.

[0299] An electronic device according to an embodiment of the present disclosure includes a memory and a processor.

[0300] The memory is used to store non-temporary computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0301] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In one embodiment of the present disclosure, the processor is used to run the computer-readable instructions stored in the memory.

[0302] Those skilled in the art should be able to understand that in order to solve the technical problem of how to obtain a good user experience, the present embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the protection scope of the present disclosure.

[0303] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0304] Example 4

[0305] An embodiment of the present disclosure provides a computer-readable storage medium storing a computer program, which implements the multi-scale fracture earthquake comprehensive prediction method when executed by a processor.

[0306] According to the computer-readable storage medium of the embodiment of the present disclosure, non-transitory computer-readable instructions are stored thereon. When the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the method of each embodiment of the present disclosure are executed.

[0307] The above-mentioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or mobile hard disk), media with built-in rewritable non-volatile memory (e.g., memory card) and media with built-in ROM (e.g., ROM box).

[0308] Those skilled in the art should understand that the purpose of the above description of the embodiments of the present invention is only to exemplarily illustrate the beneficial effects of the embodiments of the present invention, and is not intended to limit the embodiments of the present invention to any given examples.

[0309] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A multi-scale fracture earthquake comprehensive prediction method, characterized in that: include: Based on seismic data, predict large-scale fault-associated fractures; Based on the earthquake dip and azimuth information, the longitudinal joints of mesoscale folds are predicted; Fourier series expansion is performed on the Ruger equation, and small-scale interlayer microcracks are predicted by the second-order Fourier series coefficients under credibility constraints. The above multi-scale cracks are spatially fused according to the principal component analysis technology to achieve comprehensive prediction of multi-scale cracks.

2. The multi-scale fracture earthquake comprehensive prediction method according to claim 1, wherein: Prediction of large-scale fault associated fractures includes: The seismic data is processed by frequency division using the deconvolution generalized S transform algorithm to obtain multiple single-frequency data volumes. Determine the single-frequency data volume that responds reasonably to the faults or cracks in the target layer, carry out ant tracking, and obtain multiple frequency-divided ant volumes; The wavelet transform is used to perform frequency fusion on the multiple frequency-divided ant bodies to obtain a high-precision ant tracking result and predict the associated fractures of the large-scale fault.

3. The multi-scale fracture earthquake comprehensive prediction method according to claim 1, wherein: According to the information of earthquake dip and azimuth, the prediction of mesoscale fold longitudinal fractures includes: The seismic dip and azimuth information are used to construct structural deformation intensity attributes, calculate the degree of local structural deformation of the stratum, and predict the longitudinal tension fractures of the mesoscale folds.

4. The multi-scale fracture earthquake comprehensive prediction method according to claim 3, wherein: Build tectonic deformation strength properties include: The dip and azimuth of a point P0 on a stratum in three-dimensional space are Dip0 and Azim0 respectively, denoted as P0(Dip0, Azim0). i The formation dip and azimuth are Dip i , Azim i , denoted as P i (Dip i , Azim i ), P0 and P i The vector norm between i : With P0 as the center and R as the radius, calculate the average norm between P0 and other points on the same level. Then P0 to P i Deform i for:

5. The multi-scale fracture earthquake comprehensive prediction method according to claim 1, wherein: The Fourier series expansion for the Ruger equation is: in, is the equivalent elastic parameter at each azimuth, is the earthquake azimuth, is the anisotropic symmetry axis, b0 is the zero-order Fourier series coefficient, representing the isotropic term, b2 is the second-order Fourier series coefficient, representing the anisotropic intensity, and b4 is the fourth-order Fourier series coefficient, representing the high-order term of the anisotropic intensity.

6. The multi-scale fracture earthquake comprehensive prediction method according to claim 5, wherein: Use credibility constraints on the second-order Fourier series coefficients: Assume that the signal-to-noise ratio of a sampling point of 3D seismic is SNR i , according to the seismic data of the target layer, different signal-to-noise ratio thresholds SNR0 are set, and the credibility of the point is i Defined as 0 or 1, if the SNR i ≥SNR0, then Credibility i =1, otherwise Credibility i =0; The second-order Fourier series coefficient b obtained by fitting any sampling point of 3D seismic 2i , plus Credibility i Constraint, then the constrained second-order Fourier series coefficient b 2i 'for: b 2i '=b 2i *Credibility i 。 7. The multi-scale fracture earthquake comprehensive prediction method according to claim 1, wherein: The above multi-scale fractures are spatially fused according to the principal component analysis technology to achieve comprehensive prediction of multi-scale fractures, including: Establish a sample matrix in three-dimensional space based on multi-scale fractures; The sample matrix is ​​standardized and the χ ik Transformation; Calculate the correlation coefficient matrix of the sample matrix; The correlation coefficient matrix is ​​solved to determine the principal components of the multi-scale cracks, so as to fuse the components with the largest contribution rate to the multi-scale cracks and realize the comprehensive prediction of the multi-scale cracks.

8. A multi-scale fracture earthquake comprehensive prediction device, characterized in that: include: Large-scale fracture prediction module, which predicts large-scale fault-associated fractures based on seismic data; The mesoscale fracture prediction module predicts the longitudinal fractures of mesoscale folds based on the earthquake dip and azimuth information; Small-scale crack prediction module, which performs Fourier series expansion on the Ruger equation and predicts small-scale interlayer microcracks through the second-order Fourier series coefficients under credibility constraints; The spatial fusion module spatially fuses the above multi-scale cracks based on the principal component analysis technology to achieve comprehensive prediction of multi-scale cracks.

9. An electronic device, characterized in that: The electronic device comprises: A memory storing executable instructions; A processor, wherein the processor runs the executable instructions in the memory to implement the multi-scale fracture earthquake comprehensive prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the multi-scale fracture earthquake comprehensive prediction method according to any one of claims 1 to 7 is implemented.

Images