Azimuthal seismic inversion method for orthotropic media based on combined dimensionality reduction strategy
Through the azimuth seismic inversion method of orthogonal anisotropic media based on a combined dimensionality reduction strategy, the problems of highly pathological and uncertainty in the inversion process in the prior art are solved, and the stability and accuracy of orthogonal anisotropic media are achieved. Parameter prediction, suitable for media where vertical fractures develop in horizontal transverse isotropic backgrounds.
Patent Information
- Application Number
- CN202411178089.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-27
AI Technical Summary
When using conventional seismic multi-parameter simultaneous inversion method, the reflection coefficient equation for orthogonal anisotropic media contains 7 unknown parameters, resulting in the inversion process being highly morbid and uncertain, and the combined parameters lack clarity. Physical significance is difficult to meet actual production needs.
The orthogonal anisotropic medium azimuth seismic inversion method based on the combined dimensionality reduction strategy was used to construct the corrected PP wave reflection coefficient equations directly expressed by attributes A, attribute B, attribute C, normal weakness and tangential weakness, and Using iterative reweighted least squares method and Bayesian theoretical framework, combined with low-frequency regularization constraints, reduce model parameters and stably predict the inversion result.
The parameter determinism improvement for seismic inversion of orthogonal anisotropic media is achieved, reducing uncertainty, and each parameter has strict physical significance, suitable for orthogonal isotropic development of vertical fractures in horizontal transverse isotropic background The opposite-sex medium improves the accuracy and stability of the reflection coefficient equation.
Smart Images

Figure CN118707604B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of oil and gas exploration and development, and in particular to an orthogonal anisotropic medium azimuthal seismic inversion method based on a combined dimension reduction strategy. Background Art
[0002] Naturally developed fractures are important storage spaces and seepage channels for oil and gas, and are important factors in controlling oil and gas distribution and determining oil and gas recovery rates. Effective prediction of underground fractures plays a vital role in reservoir characterization, reservoir management, well location optimization and other fields. When seismic waves propagate in fractured rocks, their kinematic and dynamic parameters usually show azimuthal anisotropy differences. Wide-azimuth seismic data containing rich azimuthal information provide data support for fracture detection, and using the amplitude information of azimuthal seismic data to conduct anisotropy analysis has become an important means of fracture detection.
[0003] Affected by sedimentation and tectonic stress fields, the actual distribution of underground fractures is relatively complex. Field outcrop and imaging logging data confirm that rocks with vertical or near-vertical fractures developed in a vertical transverseisotropic (VTI) background (for example, fine stratification, horizontal fractures, vertically symmetrical shales, etc.) are common underground. Under the seismic long-wavelength approximation, such rocks can usually be equivalent to orthorombic anisotropy (ORT) media.
[0004] Under the equivalent medium theory and weak anisotropy assumption, for ORT media composed of vertical or near-vertical fractures in VTI background, its equivalent elastic properties can be divided into the sum of VTI background and fracture induction parameters. Existing studies usually use anisotropy parameters and fracture weakness parameters to characterize VTI characteristics and fracture-induced anisotropy respectively, plus three elastic parameters describing the isotropic background, which leads to the reflection coefficient equation describing ORT media usually containing 7 unknown parameters, and the anisotropy parameters and fracture weakness parameters have a small numerical contribution to the reflection coefficient. These problems make the inverse problem solution of these equations using conventional seismic multi-parameter simultaneous inversion methods highly ill-conditioned and have high uncertainty. The existing method of combining parameters of the reflection coefficient equation can reduce the number of parameters to be inverted and alleviate the ill-conditioned nature of multi-parameter inversion, but the large-angle information items are usually discarded in the combination operation process, and the combined parameters have no clear physical meaning, which is difficult to meet the actual production needs of orthogonal fracture reservoirs. Summary of the invention
[0005] In order to solve the above technical problems, the present invention proposes a new azimuthal seismic inversion method for orthotropic media. The method is based on a parameter combination dimensionality reduction strategy, uses fewer model parameters, and considers low-frequency regularization constraints. It can weaken the uncertainty of seismic inversion in orthotropic media and stably and reasonably predict model parameters.
[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: an orthogonal anisotropic medium azimuthal seismic inversion method based on a combined dimensionality reduction strategy, the method comprising the following steps:
[0007] Step 1, constructing the orthotropic medium modified PP wave reflection coefficient equation directly expressed by attribute A, attribute B, attribute C, normal weakness and tangential weakness;
[0008] Step 2, constructing the target functional that the parameters of the model to be inverted obey;
[0009] Step 3, select iterative reweighted least squares method to solve the target functional and obtain the inversion results of attribute A, attribute B, attribute C, normal weakness and tangential weakness;
[0010] Step 4: Use the logging data to conduct intersection analysis of attribute A, attribute C, reservoir gas content and reservoir type, and produce an interpretation version;
[0011] Step 5, project the inversion results of the target reservoir attributes A and C into the interpretation volume to predict the gas content and reservoir type of the reservoir; use the normal weakness inversion results to indicate the degree of fracture development.
[0012] Furthermore, in step 1, when the vertical transverse isotropic background contains vertical or nearly vertical cracks, it is equivalent to an orthotropic medium, and the azimuthal seismic reflection coefficient equation describing the medium is expressed as:
[0013] (1),
[0014] In equation (1), It represents the azimuth seismic reflection coefficient; represents the angle of incidence, Indicates azimuth; and represents the longitudinal and transverse moduli of the isotropic background, Indicates density; and represents the anisotropy parameter perpendicular to the transversely isotropic background; and Indicates the normal weakness parameter and tangential weakness parameter of the crack; the symbol “—” at the top of the model parameter indicates the average value of the model parameters of the upper and lower media; the symbol It means to find the difference between the model parameters of the upper and lower media; ;
[0015] Equation (1) is expressed as:
[0016] (2),
[0017] In equation (2), and denote the longitudinal wave velocity and the shear wave velocity, respectively. , equation (2) is expressed as:
[0018] (3),
[0019] When it is assumed that the difference in the parameters of the upper and lower medium models is small, , the model parameters of equation (3) are combined and simplified to obtain:
[0020] (4),
[0021] In equation (4), the subscripts p and p+1 represent the pth layer and the p+1th layer respectively, and:
[0022] (5),
[0023] Equation (4) is the modified PP wave reflection coefficient equation for orthotropic media, in which: attribute A represents the longitudinal wave impedance; attribute B represents the comprehensive response of the vertical shear modulus and the background elliptical anisotropy. When the anisotropy parameter is 0, attribute B represents the isotropic background shear modulus; attribute C represents the horizontal longitudinal wave phase velocity; and It indicates that the normal weakness and tangential weakness of the crack are positively correlated with the degree of crack development.
[0024] Furthermore, in step 2,
[0025] Using the convolution model, equation (4) is convolved with the azimuth seismic wavelet to generate the corresponding seismic record. For a wide azimuth seismic data volume containing q incident angles, n azimuths and i sampling points, it is expressed in matrix form as follows:
[0026] (6),
[0027] In equation (6),
[0028] (7),
[0029] In equation (7), represents the wide-azimuth seismic data vector, The incident angle is and the azimuth is Single-channel seismic records, represents the seismic record of the i-th sampling point, represents the model parameter vector to be inverted, represents the forward operator, represents the wavelet matrix, represents the difference operation matrix, represents the coefficient matrix of the t-th model parameter, Denotes a diagonal matrix, with the superscript Represents the transpose of a matrix;
[0030] Equation (6) can be expressed as follows after decorrelation:
[0031] (8),
[0032] In equation (8), and denote the forward operator and model parameter vector after decorrelation, respectively; represents the covariance matrix, which is obtained from the well curves obtained from rock physics modeling;
[0033] In the framework of Bayesian theory, the Cauchy prior distribution is used as the prior probability density distribution function. Assuming that the seismic noise obeys Gaussian distribution, the posterior probability density distribution function of the model parameters is expressed as:
[0034] (9),
[0035] In equation (9), and denote the variance of model parameters and seismic noise respectively, represents the kth model parameter;
[0036] Introducing the low-frequency model regularization constraint into equation (9), the target functional is obtained as:
[0037] (10),
[0038] In equation (10), represents the low-frequency model of the kth model parameter; represents the regularization coefficient of the kth low-frequency model.
[0039] Compared with the prior art, the present invention has the following beneficial effects:
[0040] The modified PP wave reflection coefficient equation proposed in the present invention is more suitable for orthotropic media with vertical cracks developed in a horizontal transverse isotropic background. The equation has almost the same accuracy as the existing equation. The new equation contains fewer model parameters, and each parameter in the equation has a strict physical meaning.
[0041] The azimuthal seismic inversion method based on the modified reflection coefficient equation proposed in the present invention takes into account the low-frequency regularization constraint, can weaken the uncertainty of seismic inversion in orthogonal anisotropic media, and stably and reasonably predict the model parameters in the modified reflection coefficient equation. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0043] Figure 1 The present invention is a flow chart of the azimuthal seismic inversion method for orthotropic media based on the combined dimensionality reduction strategy.
[0044] Figure 2 Schematic diagram of the orthotropic medium model.
[0045] Figure 3 It is a schematic diagram for comparing the orthotropic medium modified PP wave reflection coefficient equation of the present invention with the equation of the prior art in the azimuth domain.
[0046] Figure 4 It is a schematic diagram for comparing the orthotropic medium modified PP wave reflection coefficient equation of the present invention with the equation of the prior art in the incident angle domain.
[0047] Figure 5 The schematic cross-sectional diagrams of the inversion results of various parameters for testing the method of the present invention, wherein (a) is a schematic cross-sectional diagram of the inversion result of attribute A, (b) is a schematic cross-sectional diagram of the inversion result of attribute B, (c) is a schematic cross-sectional diagram of the inversion result of attribute C, (d) is a schematic cross-sectional diagram of the inversion result of normal weakness parameter, and (e) is a schematic cross-sectional diagram of the inversion result of tangential weakness parameter.
[0048] Figure 6 Schematic diagram of logging intersection analysis of attribute A and attribute C. DETAILED DESCRIPTION
[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0050] Figure 1 Flow chart of the method of the present invention. Figure 1 As shown, the orthotropic medium azimuthal seismic inversion method based on the combined dimension reduction strategy of the present invention comprises the following steps:
[0051] Step 1, constructing the orthotropic medium modified PP wave reflection coefficient equation directly expressed by attribute A, attribute B, attribute C, normal weakness and tangential weakness;
[0052] Step 2, constructing the target functional that the parameters of the model to be inverted obey;
[0053] Step 3, select iterative reweighted least squares method to solve the target functional and obtain the inversion results of attribute A, attribute B, attribute C, normal weakness and tangential weakness;
[0054] Step 4: Use the logging data to conduct intersection analysis of attribute A, attribute C, reservoir gas content and reservoir type, and produce an interpretation version;
[0055] Step 5, project the inversion results of the target reservoir attributes A and C into the interpretation volume plate to predict the gas content and reservoir type of the reservoir; use the normal weakness inversion results to indicate the degree of fracture development.
[0056] In step 1, the PP wave reflection coefficient equation of the ORT medium is reconstructed mainly through the idea of parameter combination and dimensionality reduction.
[0057] like Figure 2 As shown in Figure 2, when the VTI background contains vertical or nearly vertical fractures, it can be equivalent to ORT medium. The azimuthal seismic reflection coefficient equation describing the medium can be expressed as:
[0058] (1)
[0059] in, It represents the azimuth seismic reflection coefficient; represents the angle of incidence, Indicates azimuth; and represents the longitudinal and transverse moduli of the isotropic background, Indicates density; and represents the anisotropy parameter of the VTI background; and Indicates the normal weakness parameter and tangential weakness parameter of the crack; the symbol “—” at the top of the model parameter indicates the average value of the model parameters of the upper and lower media; the symbol It means to find the difference between the model parameters of the upper and lower media; .
[0060] Equation (1) can be reformulated as:
[0061] (2)
[0062] in, and denote the longitudinal wave velocity and the transverse wave velocity, respectively. Note that, , so equation (2) can be re-expressed as:
[0063] (3)
[0064] Furthermore, when it is assumed that the difference in the parameters of the upper and lower medium models is small, there is a mathematical (Basic operation of the industry). Introducing this idea into equation (3), combining and simplifying the model parameters, we can get:
[0065] (4)
[0066] Among them, the subscripts p and p+1 represent the pth layer and the p+1th layer respectively, and:
[0067] (5)
[0068] The parameters in equation (4) contain only five unknown parameters and are called the ORT medium modified PP wave reflection coefficient equation. Each attribute in the equation has a clear physical meaning: attribute A represents the longitudinal wave impedance; attribute B represents the comprehensive response of the vertical shear modulus and the background elliptical anisotropy. When the anisotropy parameter is 0, attribute B can be expressed as the isotropic background shear modulus; attribute C represents the horizontal longitudinal wave phase velocity; and The normal weakness parameter and tangential weakness parameter representing the crack are both positively correlated with the degree of crack development and can be used to estimate the crack density.
[0069] Equation (4), equation (1) and The accuracy of the equation is verified by comparing it with the reflection coefficient equation of arbitrary weak anisotropic media proposed by Martins, such as Figure 3 and Figure 4 As shown in the figure, the solid line, dashed line and dotted line represent equation (1), equation (4), From the reflection coefficient curve calculated by the Martins equation, it can be seen that the ORT medium modified PP wave reflection coefficient equation (4) proposed in the present invention has a high degree of consistency with equation (1).
[0070] In step 2, based on a Bayesian framework, an azimuthal seismic inversion method is developed to estimate the five model parameters in equation (4) by considering the Cauchy prior information and the low-frequency regularization constraint.
[0071] Using the convolution model, equation (4) is convolved with the azimuth seismic wavelet to generate the corresponding seismic record. For a wide azimuth seismic data volume containing q incident angles, n azimuths, and i sampling points, the positive problem can be expressed in matrix form as follows:
[0072] (6)
[0073] in,
[0074] (7)
[0075] In the formula, represents the wide-azimuth seismic data vector, The incident angle is and the azimuth is Single-channel seismic records, represents the seismic record of the i-th sampling point, represents the model parameter vector to be inverted, represents the forward operator, represents the wavelet matrix, represents the difference operation matrix, represents the coefficient matrix of the t-th model parameter, Denotes a diagonal matrix, with the superscript Represents the transpose of a matrix.
[0076] Generally speaking, there is a strong correlation between model parameters, which may affect the stability of the model parameter inversion results. Therefore, it is necessary to decorrelate the model parameters before inversion. After decorrelation, equation (6) can be expressed as:
[0077] (8)
[0078] in, and represent the forward operator and model parameter vector after decorrelation, respectively. represents the covariance matrix, which can be obtained from the well curves obtained from rock physics modeling.
[0079] In the framework of Bayesian theory, the Cauchy prior distribution is used as the prior probability density distribution function. Assuming that the seismic noise obeys Gaussian distribution, the posterior probability density distribution function of the model parameters can be expressed as:
[0080] (9)
[0081] In the formula, and denote the variance of model parameters and seismic noise respectively, represents the kth model parameter. Then the low-frequency model regularization constraint is introduced into equation (9) to improve the stability of the inversion result. Finally, the target functional can be expressed as:
[0082] (10)
[0083] in, represents the low-frequency model of the kth model parameter; represents the regularization coefficient of the kth low-frequency model. This parameter usually controls the similarity between the inversion result and the low-frequency model.
[0084] In step 3, the iterative reweighted least squares algorithm is selected to solve the target functional (10), so that all parameters to be inverted can be reasonably obtained, and the inversion results of attribute A, attribute B, attribute C, normal weakness and tangential weakness can be obtained. The iterative reweighted least squares algorithm is an existing algorithm, and the specific steps of using the algorithm to solve the target functional (10) are not repeated.
[0085] On this basis, the fracture normal weakness parameter can be used to predict the degree of fracture development, that is, the larger the fracture normal weakness parameter and the tangential weakness parameter, the more developed the fracture; the intersection analysis of attribute A and attribute C can be used to identify the gas content and reservoir type of the reservoir. Specifically, firstly, the intersection analysis of attribute A, attribute C, reservoir gas content and reservoir type is performed using logging data, and an interpretation plate is prepared. Then, the inversion results of the target reservoir attribute A and attribute C are projected into the interpretation plate, and the prediction of reservoir gas content and reservoir type can be realized. Performing intersection analysis and preparing an interpretation plate is an existing technology, and the specific steps of using this technology to predict reservoir gas content and reservoir type will not be repeated.
[0086] Verify the rationality of the inversion method of the present invention:
[0087] A gas-bearing shale reservoir in western my country was selected to test the azimuth seismic inversion method of the present invention. The geological and logging interpretation data of the study area show that the target layer has a large number of horizontal bedding, high gas abundance, low permeability, low porosity, and a large number of vertical and near-vertical fractures. Therefore, the gas-bearing shale target reservoir can be regarded as an ORT medium with vertical fractures developed under the background of VTI. The azimuths of the seismic data are 15°, 45°, 75°, 105°, 135° and 165°, respectively. The three superimposed average incident angles of near, medium and far corresponding to each azimuth are 8°, 16° and 24°, respectively. The azimuth seismic inversion method proposed in the present invention is used to estimate the model parameters.
[0088] The inversion results are as follows Figure 5 As shown in the figure, the curve is a high-cut filter display of the true value of the logging curve at the well location. Figure 5 It can be seen that the inversion results of the target layer have good lateral continuity. The values of attribute A, attribute B and attribute C of the target layer are relatively low; at the same time, the values of fracture normal weakness and fracture tangential weakness are relatively high, indicating that the fractures are relatively developed, which is consistent with the geological prior information of the work area; in addition, the inversion results of the model parameters at the well location are basically consistent with the well curve, which further confirms the rationality and effectiveness of the inversion results. In addition, considering factors such as brittleness, organic matter content, and gas abundance, according to the logging data, the target layer is divided into Class I reservoirs (gas-bearing reservoirs with the highest brittleness, organic matter content, and gas abundance), Class II reservoirs (gas-bearing reservoirs with the second highest brittleness, organic matter content, and gas abundance), Class III reservoirs (gas-bearing reservoirs with the lowest brittleness, organic matter content, and gas abundance) and dry layers. The logging data is used to conduct an intersection analysis of attribute A, attribute C, reservoir gas content, and reservoir type, and an interpretation version is produced, such as Figure 6 As shown in the figure, according to the interpretation volume, the gas content and reservoir type of the reservoir can be effectively identified by using attributes A and C, that is, the inversion results of the target reservoir attributes A and C are projected onto the interpretation volume ( Figure 6 ), the inversion results distributed in area ① (the area filled with gray), area ② (the area filled with grid lines), area ③ (the area filled with black dots), and area ④ (the area filled with horizontal lines) represent that the target reservoirs are type I gas-bearing reservoirs, type II gas-bearing reservoirs, type III gas-bearing reservoirs, and dry layers, respectively.
Claims
1. A method for azimuthal seismic inversion in orthotropic media based on a combined dimensionality reduction strategy, characterized in that: The following steps are involved: Step 1: Construct a modified PP wave reflection coefficient equation for orthotropic media directly expressed by attribute A, attribute B, attribute C, normal weakness, and tangential weakness; Step 2: construct the target functional that the parameters of the model to be inverted obey; Step 3: Select the iterative reweighted least squares method to solve the target functional and obtain the inversion results of attribute A, attribute B, attribute C, normal weakness and tangential weakness; Step 4: Use the logging data to conduct an intersection analysis of attribute A, attribute C, reservoir gas content, and reservoir type, and produce an interpretation version; Step 5: Project the inversion results of the target reservoir attributes A and C into the interpretation volume to predict the gas content and reservoir type of the reservoir; use the normal weakness inversion results to indicate the degree of fracture development; In step 1, when the vertical transversely isotropic background contains vertical or nearly vertical cracks, it is equivalent to an orthotropic medium. The azimuthal seismic reflection coefficient equation describing the medium is expressed as: (1), In equation (1), represents the azimuthal seismic reflection coefficient; represents the angle of incidence, Indicates azimuth; and represents the longitudinal and transverse wave moduli of the isotropic background, Indicates density; and represents the anisotropy parameter perpendicular to the transversely isotropic background; and Indicates the normal weakness parameter and tangential weakness parameter of the crack; the symbol "—" on the top of the model parameter indicates the average value of the model parameters of the upper and lower media; the symbol Indicates the difference between the model parameters of the upper and lower media; ; Equation (1) is expressed as: (2), In equation (2), and denote the longitudinal wave velocity and the shear wave velocity, respectively. , equation (2) is expressed as: (3), When it is assumed that the difference in the upper and lower layer medium model parameters is small, , the model parameters of equation (3) are combined and simplified to obtain: (4), In equation (4), the subscripts p and p+1 represent the p-th layer and the p+1-th layer, respectively, and: (5), Equation (4) is the modified PP wave reflection coefficient equation for orthotropic media. In this equation, attribute A represents the longitudinal wave impedance; attribute B represents the comprehensive response of the vertical shear modulus and the background elliptical anisotropy. When the anisotropy parameter is 0, attribute B represents the isotropic background shear modulus. Attribute C represents the horizontal longitudinal wave phase velocity; and Indicates the normal weakness and tangential weakness of the crack; In step 2, the convolution model is used to convolve Equation (4) with the azimuth seismic wavelet to generate the corresponding seismic record. For a wide azimuth seismic data volume containing q incident angles, n azimuth angles, and i sampling points, it is expressed in matrix form as follows: (6), In equation (6), (7), In equation (7), represents the wide-azimuth seismic data vector, The incident angle is and the azimuth is Single-channel seismic records, represents the earthquake record of the i-th sampling point, represents the parameter vector of the model to be inverted, represents the forward operator, represents the wavelet matrix, represents the difference operation matrix, represents the coefficient matrix of the t-th model parameter, Indicates taking a diagonal matrix, the superscript Represents the transpose of a matrix; Equation (6) can be expressed as follows after decorrelation operation: (8), In equation (8), and denote the forward operator and model parameter vector after decorrelation, respectively; represents the covariance matrix, which is obtained from the well curves obtained from rock physics modeling; In the framework of Bayesian theory, the Cauchy prior distribution is used as the prior probability density distribution function. Assuming that the seismic noise obeys Gaussian distribution, the posterior probability density distribution function of the model parameters is expressed as: (9), In equation (9), and denote the variance of model parameters and seismic noise respectively, represents the kth model parameter; Introducing the low-frequency model regularization constraint into Equation (9), the target functional is obtained as: (10), In equation (10), represents the low-frequency model of the kth model parameter; represents the regularization coefficient of the kth low-frequency model.
