A rational impedance kernel-driven multi-parameter pre-stack inversion method and apparatus for TTI dielectrics

By constructing a direction-correlated qP phase velocity background in the pre-stack inversion of PP waves in TTI medium and combining iterative inversion, the sensitivity overlap and parameter crosstalk problems between elastic parameters and anisotropic parameters are solved, thus improving the stability and accuracy of the inversion.

CN123113000APending Publication Date: 2026-11-17CHENGDU UNIVERSITY OF TECHNOLOGY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611350881.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-02
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the existing TTI medium PP wave pre-stack inversion, the direction-correlated qP phase velocity background is only used as an additional perturbation term, resulting in severe sensitivity overlap and parameter crosstalk between elastic parameters and anisotropic parameters, and insufficient stability of multi-parameter inversion.

Method used

By establishing the geometric relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium, a direction-dependent qP phase velocity background is constructed and introduced into the denominator normalization structure of the direction-equivalent qP impedance. Combined with multi-angle and multi-directional pre-stack data, iterative inversion is performed to form a closed-loop update of the reflection kernel and the model.

Benefits of technology

It enhances the ability of the reflection coefficient to characterize the orientation-related features of TTI media, reduces the mutual influence between elastic parameters and anisotropic parameters, and improves the stability and accuracy of multi-parameter inversion.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113000A_ABST
    Figure CN123113000A_ABST
Patent Text Reader

Abstract

This invention relates to the field of seismic exploration and pre-stack seismic inversion technology, and discloses a rational impedance kernel-driven multi-parameter pre-stack inversion method and apparatus for TTI media. Addressing the problem in existing TTI media PP wave pre-stack inversions where the direction-correlated qP phase velocity background is not included in the impedance denominator normalization structure, leading to severe crosstalk between elastic parameters and anisotropic parameters, this method includes: acquiring multi-angle and multi-azimuth pre-stack PP wave seismic data and prior information about the TTI medium; establishing the geometric coupling relationship between the qP wave propagation direction and the tilted symmetry axis of the TTI medium to obtain the direction-correlated qP phase velocity background factor; introducing this background factor into the directional equivalent qP impedance while retaining the denominator normalization structure, performing first-order linearization on the elastic and anisotropic parameters, and constructing a phase velocity-constrained rational PP wave reflection coefficient; constructing a multi-angle and multi-azimuth joint forward modeling relationship based on this reflection coefficient and performing iterative inversion to output the inversion results.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the fields of seismic exploration, reservoir prediction, pre-stack seismic inversion, and anisotropic medium parameter inversion, and provides a rational impedance kernel-driven multi-parameter pre-stack inversion method and apparatus for TTI media. Background Technology

[0002] In complex structural zones, fractured reservoirs, areas with well-developed inclined bedding, shale reservoirs, and deep unconventional reservoirs, seismic wave propagation velocity and reflection amplitude typically exhibit significant direction dependence. When bedding, fracture groups, or stress-induced structures have a dominant direction, and this direction is inclined relative to the observation coordinate system, the subsurface medium can be approximated as a tilted transverse isotropic (TTI) medium. Compared to isotropic or vertically transversely isotropic (VTI) media, the reflection amplitude of P-waves in TTI media is influenced not only by P-wave velocity, S-wave velocity, and density interface contrast, but also by the Thomsen anisotropy parameter, the qP-wave phase propagation direction, the observation azimuth angle, and the geometric relationship of the tilted symmetry axis.

[0003] Existing pre-stack inversion methods for PP waves are typically based on approximations of isotropic or weakly anisotropic properties that can be neglected. In these methods, the reflection amplitude is often expressed as a linear combination of P-wave impedance, S-wave impedance, and density reflectivity. For conventional AVO analysis and elastic parameter inversion, this expression has advantages such as simplicity, high computational efficiency, and ease of constructing the Jacobian matrix. However, in TTI media, the qP wave phase velocity itself is a direction-dependent quantity, the magnitude of which depends on the angle between the qP wave phase propagation direction and the tilted axis of symmetry. This angle is determined by the incident angle, observation azimuth, axis of symmetry tilt, and axis of symmetry orientation. Therefore, if isotropic background velocity or conventional P-wave impedance is still used as the background term for the reflection coefficient, it is difficult to fully express the coupling relationship between angle, azimuth, and axis of symmetry tilt in TTI media.

[0004] Existing approximations of weakly anisotropic reflection coefficients typically start from a reference background medium, performing a first-order expansion of weak interface contrast and anisotropic perturbations. This approach facilitates the formation of linear inversion equations, but it also results in the direction-dependent qP background only entering the reflection coefficient as an additional perturbation term, failing to incorporate it into the normalized structure of background impedance and reflection kernel function. For TTI medium inversions that simultaneously estimate multiple parameters such as P-wave velocity, S-wave velocity, density, and Thomsen parameters ε, δ, and γ, insufficient expression of the direction dependence of the qP background phase velocity can enhance sensitivity overlap and parameter crosstalk between elastic and anisotropic parameters, reducing the inversion stability of anisotropic parameters.

[0005] Therefore, there is an urgent need for a new method for constructing PP wave reflection coefficients, which can, while maintaining the linearity of weak interface contrast, analytical differentiability, and computational operability, introduce the TTI medium direction-correlated qP phase velocity background into the denominator normalization structure of the reflection coefficients, and further combine multi-angle and multi-azimuth pre-stack data to establish a stable multi-parameter inversion framework. Summary of the Invention

[0006] The purpose of this invention is to solve the technical problems caused by the directional correlation qP phase velocity background in the pre-stack inversion of PP waves in existing TTI media, which only enters the reflection coefficient as an additional perturbation term and does not participate in the normalization structure of the background impedance denominator, resulting in sensitivity overlap between elastic parameters and anisotropic parameters, severe parameter crosstalk, and insufficient stability of multi-parameter inversion.

[0007] This invention further recognizes that simply introducing a direction-dependent qP phase velocity correction into the reflection coefficient cannot fully solve the above problems. On the one hand, if the direction-dependent qP phase velocity only participates in the reflection coefficient as an additional perturbation term, its description of the propagation direction of the TTI medium cannot directly change the background normalization scale of the reflection amplitude. On the other hand, even if the background impedance is constructed using the direction-dependent qP phase velocity, if the direction-dependent background and its corresponding reflection kernel remain unchanged throughout the inversion process, the directional background corresponding to the fixed reflection kernel will gradually deviate from the current model state as the P-wave velocity, S-wave velocity, density, and anisotropy parameters are continuously updated, causing inconsistencies between the reflection kernel function and the current TTI medium model.

[0008] Based on the above considerations, this invention does not treat the calculation of direction-dependent qP phase velocity, the construction of rational reflection coefficients, and multi-parameter iterative inversion as independent processes, but rather couples them together: First, a direction-dependent qP phase velocity background is constructed using the geometric relationship between the qP wave phase propagation direction and the TTI tilted symmetry axis; then, this direction-dependent background is introduced into the directional equivalent qP impedance, and its denominator normalization structure is retained to construct a phase velocity-constrained rational PP wave reflection kernel; further, during the iterative inversion process, the direction-dependent qP phase velocity background and the rational impedance kernel are recalculated based on the current model. Through this combination, the construction of the direction-dependent background, the normalization of the reflection amplitude background, and the iterative update of the model parameters form a mutually feedback closed-loop process.

[0009] To achieve the above objectives, the present invention employs the following technical means:

[0010] The invention provides a rational impedance kernel-driven multi-parameter pre-stack inversion method for TTI dielectrics, comprising the following steps:

[0011] Step 1: Obtain multi-angle, multi-azimuth pre-stack PP wave seismic data and TTI medium prior information for the target work area;

[0012] Step 2: Establish the geometric coupling relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium. Based on the rotational weak anisotropy qP phase velocity approximation, obtain the direction-dependent qP phase velocity background factor determined by the incident angle, observation azimuth angle, and symmetry axis tilt angle.

[0013] Step 3: Introduce the direction-dependent qP phase velocity background factor into the direction-equivalent qP impedance. While retaining the direction-equivalent qP impedance as the denominator normalization structure, perform first-order linearization on the interface comparison of elastic parameters and anisotropic parameters to construct the phase velocity-constrained rational PP wave reflection coefficient.

[0014] Step 4: Based on the phase velocity constrained rational PP wave reflection coefficient, construct a multi-angle, multi-azimuth joint forward modeling relationship, and perform iterative inversion including elastic parameters and anisotropic parameters; during the iteration process, update the direction-related qP phase velocity background factor according to the current model, and reconstruct the directional equivalent qP impedance, azimuth projection factor, and phase velocity constrained rational PP wave reflection kernel based on the updated direction-related qP phase velocity background factor, so that the direction-related background and reflection kernel function maintain a dynamic correspondence with the current inversion model;

[0015] Step 5: Output the inversion results of the elasticity parameters and anisotropy parameters.

[0016] In the above scheme, step 1 includes:

[0017] Step 1.1: Acquire multi-angle, multi-azimuth pre-stack PP wave seismic data, seismic wavelets, well logging curves, stratigraphic interpretation results, low-frequency initial model, and prior knowledge of the dip and azimuth axis of the TTI medium for the target work area; wherein, the pre-stack PP wave seismic data includes angle gathers, azimuth angle gathers, or pre-stack gathers with both angle and azimuth indices; the well logging curves include P-wave velocity, S-wave velocity, density, and Thomsen anisotropy parameters. , and curve;

[0018] Step 1.2: The pre-stack PP wave seismic data is processed by angle domain or azimuth domain gathers to form multidimensional observation data indexed by incident angle and observation azimuth angle;

[0019] Step 1.3: Perform time-depth conversion, resampling, outlier removal, and smoothing on the well logging curves to make them consistent with the sampling interval of the pre-stack PP wave seismic data, the layer interpretation results, and the inversion grid;

[0020] Step 1.4: Construct the low-frequency initial model based on the well logging curves, geological strata, and existing velocity or impedance interpretation results;

[0021] Step 1.5: Perform time-domain or depth-domain registration, scale unification, and noise suppression on the pre-stack PP wave seismic data, well logging curves, and low-frequency initial model; normalize parameters of different dimensions to ensure that P-wave velocity, S-wave velocity, density, and Thomsen anisotropy parameters are uniform. , and They have similar update scales during the inversion process;

[0022] Step 1.6: Based on the structural interpretation results of the work area, bedding dip angle, offset imaging results, or external geological priors, establish the priors for the dip angle and azimuth orientation of the TTI medium.

[0023] In the above scheme, step 2 includes:

[0024] Step 2.1: Establish a local TTI reflection geometry coordinate system for the target reflection interface, where the local coordinate axes are... , The normal direction of the reflection interface. Located within the reflecting plane, Orthogonal to the reflecting plane;

[0025] Step 2.2: Define the unit vector of the tilt symmetry axis of the TTI medium. Unit vector of seismic survey line direction and the unit vector of the incident phase direction of the qP wave Calculate the angle between the phase propagation direction of the qP wave and the tilted axis of symmetry. ;

[0026] Step 2.3: Based on the rotational weak anisotropy qP phase velocity approximation, according to the included angle Constructing the directionally correlated qP phase velocity background factor .

[0027] In the above scheme, in step 2.2, the included angle satisfy:

[0028]

[0029]

[0030]

[0031]

[0032] in, The incident angle of the PP wave. To observe the azimuth angle, The angle of inclination of the TTI axis of symmetry relative to the interface normal direction;

[0033] In step 2.3, the direction-related qP phase velocity background factor Represented as:

[0034]

[0035]

[0036] Or, under the weak anisotropy first-order approximation, it can be expressed as:

[0037]

[0038] in, Let qP be the phase velocity of the wave. The vertical longitudinal wave velocity, and For Thomsen anisotropy parameters.

[0039] In the above scheme, step 3 includes:

[0040] Step 3.1: Based on the direction-related qP phase velocity background factor Constructing the directional equivalent qP impedance :

[0041]

[0042] in, For density, The vertical longitudinal wave velocity, The background factor for the direction-dependent qP phase velocity;

[0043] Step 3.2: Construct the azimuth projection factor of shear-related anisotropy parameters :

[0044]

[0045] in, The unit vector for the direction of the seismic survey line. The unit vector of the tilt symmetry axis of the TTI medium. The tilt angle is the axis of symmetry. To observe the azimuth angle; using the azimuth projection factor For Thomsen parameters The relevant items are weighted by directional visibility;

[0046] Step 3.3: Under the assumptions of weak anisotropy and weak interface contrast, retain the denominator normalization structure of the directional equivalent qP impedance, and only normalize the P-wave velocity, S-wave velocity, density, and Thomsen parameter. , and The interface comparison is linearized to the first order to construct the phase velocity constrained rational type PP wave reflection coefficient.

[0047] In the above scheme, in step 3.3, the reflection coefficient of the phase velocity-constrained rational PP wave is expressed as:

[0048]

[0049] in, , , Centralized background quantity for the interface's upper and lower medium parameters; , , , , , Differences in interface parameters; , , , , , The analytical sensitivity kernel function is determined by the incident angle, observation azimuth angle, TTI symmetry axis tilt angle, direction-dependent qP phase velocity background factor, direction-equivalent qP impedance, and azimuth projection factor.

[0050] For the The six-parameter interface comparison vector for the reflective interface is:

[0051]

[0052] For the For each incident angle-observation azimuth angle sampling point, the reflection kernel function vector is:

[0053]

[0054] The corresponding reflection coefficient is .

[0055] In the above scheme, step 4 includes:

[0056] Step 4.1: Convert the layer parameter model using the difference operator Converted into an interface comparison model, the reflection coefficient sequence is processed by the seismic wavelet convolution operator. Converted to synthetic seismic records; for the first The forward modeling relationship for each incident angle-observation azimuth sampling point is expressed as:

[0057]

[0058] in, To synthesize pre-stack records, The reflection kernel matrix is ​​composed of the reflection coefficients of the phase velocity-constrained rational type PP wave. The model for the layer parameters to be inverted is D, and the layer parameter difference operator is D.

[0059] By combining the forward modeling relationships of all incident angles and observation azimuth angles, we obtain:

[0060]

[0061] in This is a pre-stack record of the combined synthesis of all angle-azimuth sampling points. It is a joint forward modeling operator that includes a reflection kernel function, a layer parameter difference operator, and a wavelet convolution operator;

[0062] Step 4.2: Construct a multi-parameter inversion objective function that includes data fitting terms, model smoothing constraint terms, and prior model constraint terms:

[0063]

[0064] in, To observe pre-stack PP wave seismic data, For the current iteration model or the initial model, The prior model is constructed from well logging data, geological stratigraphy, and low-frequency trends. It is a first-order or second-order difference smoothing operator. As a smoothing constraint factor, These are prior constraint factors;

[0065] Step 4.3: Solve the model increment using the damped Gauss-Newton iterative method. For each outer iteration, recalculate the geometric coupling relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium based on the current inversion model, update the direction-dependent qP phase velocity background factor, and reconstruct the directional equivalent qP impedance, azimuth projection factor, and phase velocity-constrained rational PP wave reflection kernel function based on the updated direction-dependent qP phase velocity background factor. Under the condition that the direction-dependent background and reflection kernel function corresponding to the current outer iteration remain unchanged, solve the six-parameter interface comparison or layer parameter increment through inner iteration. After completing the model update, update the direction-dependent background and reflection kernel function again until the preset iteration termination condition is met.

[0066] In the above scheme, in step 4.3, at the first In the next iteration, for the current model The forward response at a given point is linearized to the first order to construct the Jacobian matrix. The model increment is solved using the following formula:

[0067]

[0068] in, The damping factor, The identity matrix is ​​used; the model is updated as follows:

[0069]

[0070] in, For the first The model of the next iteration For the first The model of the next iteration This is the step size factor, determined based on the descent of the objective function, changes in data residuals, or the line search strategy.

[0071] In the above scheme, step 5 includes:

[0072] Step 5.1: Determine whether a preset iteration termination condition is met, wherein the termination condition includes at least one of the following:

[0073] (1) The data residual between the observed data and the synthetic data is less than the preset threshold;

[0074] (2) The decrease in the objective function between two consecutive iterations is less than a preset threshold;

[0075] (3) The model update amount in two consecutive iterations is less than the preset threshold;

[0076] (4) The damping factor or step size factor reaches the preset range;

[0077] (5) The number of iterations reaches the preset maximum number;

[0078] Step 5.2: When the termination condition is met, output the TTI medium multi-parameter inversion results for the target work area. The inversion results include P-wave velocity volume, S-wave velocity volume, density volume, and Thomsen anisotropy parameters. Volume, Thomsen anisotropy parameters Volume and Thomsen anisotropy parameters body;

[0079] Step 5.3: Output at least one quality control result from the following: synthetic record, residual profile, objective function convergence curve, parameter update amount, well-seismic comparison curve, layer constraint comparison result, and angle-azimuth domain amplitude fitting result, to evaluate the reliability of the inversion results.

[0080] This invention also provides a rational impedance kernel-driven TTI dielectric multi-parameter pre-stack inversion device, comprising:

[0081] The data acquisition module is used to acquire pre-stack PP wave seismic data from multiple angles and orientations, as well as TTI medium prior information.

[0082] The phase propagation direction-tilted symmetry axis coupled phase velocity construction module is used to establish the geometric coupling relationship between the phase propagation direction of the qP wave and the tilted symmetry axis of the TTI medium, and to obtain the direction-related qP phase velocity background factor based on the rotational weak anisotropy qP phase velocity approximation.

[0083] The rational impedance kernel construction module is used to introduce the direction-dependent qP phase velocity background factor into the directional equivalent qP impedance. Under the premise of retaining the directional equivalent qP impedance as the denominator normalization structure, the interface comparison of elastic parameters and anisotropic parameters is linearized in the first order to construct the phase velocity-constrained rational PP wave reflection coefficient.

[0084] The TTI medium multi-parameter pre-stack inversion module is used to construct a multi-angle, multi-azimuth joint forward modeling relationship based on the phase velocity-constrained rational PP wave reflection coefficient, and to perform iterative inversion including elastic parameters and anisotropic parameters. During the iteration process, the direction-related qP phase velocity background factor is updated according to the current inversion model, and the directional equivalent qP impedance, azimuth projection factor, and phase velocity-constrained rational PP wave reflection kernel function are reconstructed based on the updated directional-related qP phase velocity background factor, so that the direction-related background, directional equivalent qP impedance, and reflection kernel function are updated synchronously with the inversion model.

[0085] The result output module is used to output the inversion results of the elasticity parameters and anisotropy parameters.

[0086] Because the present invention employs the above-mentioned technical means, it has the following beneficial effects:

[0087] (1) This invention determines the direction-dependent qP phase velocity background based on the geometric relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium, and introduces the direction-dependent qP phase velocity background into the normalization term of the denominator of the directional equivalent qP impedance. By adopting the above processing method, the influence of the incident angle, observation azimuth angle, and symmetry axis tilt angle on the qP wave propagation characteristics can be directly involved in the calculation of the PP wave reflection coefficient background, and further reflected in the reflection kernel function corresponding to each parameter. This enhances the characterization ability of the reflection coefficient on the direction-dependent characteristics of the TTI medium, and helps to reduce the mutual influence between elastic parameters and anisotropic parameters.

[0088] (2) This invention combines the construction of the direction-dependent qP phase velocity background and the rational impedance kernel with multi-parameter iterative inversion. During the inversion process, the direction-dependent qP phase velocity background is updated according to the current model, and the directional equivalent qP impedance and the phase velocity-constrained rational PP wave reflection kernel are recalculated using the updated direction-dependent qP phase velocity background. This allows the reflection kernel function to be adjusted accordingly with the update of the inversion model, reducing the inconsistency between the fixed background and the fixed reflection kernel and the current model, and improving the stability of the multi-parameter iterative inversion process.

[0089] (3) This invention adopts an iterative approach of updating the directionally relevant background and reflection kernel in the outer layer and solving the parameter increments in the inner layer. In the outer layer iteration, the directionally relevant qP phase velocity background, directional equivalent qP impedance, and reflection kernel function are updated according to the current model; in the inner layer iteration, the six-parameter interface comparison or layer parameter increments are solved. By adopting the above approach, while preserving the analytical differentiability of the reflection kernel function and facilitating the construction of the Jacobian matrix, the reflection kernel can be adapted to the current TTI medium background, while avoiding the large computational load caused by directly solving the complete TTI elastic boundary conditions.

[0090] (4) This invention combines multi-angle, multi-azimuth pre-stack PP wave data with a direction-correlated rational impedance kernel for six-parameter joint inversion. For different incident angles and observation azimuths, based on the relative relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium, the direction-correlated qP phase velocity background, the directional equivalent qP impedance, and the corresponding reflection kernel function are calculated, and the azimuth projection factor is used to characterize the response differences of different parameters under different observation azimuths. By adopting the above method, the directional information in multi-angle, multi-azimuth data can be fully utilized, the constraint ability of different parameters can be enhanced, and the parameter crosstalk in multi-parameter inversion can be reduced, thereby improving the stability of the inversion results. Attached Figure Description

[0091] Figure 1 This is a flowchart of the overall process flow for rational impedance kernel-driven TTI dielectric multi-parameter pre-stack inversion according to an embodiment of the present invention.

[0092] Figure 2 This is a schematic diagram of the coupling relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium, and the construction of the direction-dependent phase velocity background, according to an embodiment of the present invention.

[0093] Figure 3 This is a comparison diagram of the amplitude response of the accurate reflection coefficient, the traditional approximation, and the phase velocity-constrained rational approximation reflection coefficient according to an embodiment of the present invention.

[0094] Figure 4 This is a comparison diagram of the low-frequency model input to the one-dimensional synthetic seismic record inversion of a layered model according to an embodiment of the present invention and the real model.

[0095] Figure 5 This is a comparison chart of the inversion results of elastic parameters and anisotropic parameters and an error comparison chart between the method of the present invention and the conventional method according to an embodiment of the present invention.

[0096] Figure 6 This is an example of the actual pre-stack angle gather input (a) and the six-parameter inversion result profile (b) of an embodiment of the present invention. Detailed Implementation

[0097] The embodiments of the present invention will be described in detail below. Although the present invention will be described and illustrated in conjunction with some specific embodiments, it should be noted that the present invention is not limited to these embodiments. On the contrary, any modifications or equivalent substitutions made to the present invention should be covered within the scope of the claims of the present invention.

[0098] Furthermore, to better illustrate the present invention, numerous specific details are set forth in the following detailed embodiments. Those skilled in the art will understand that the present invention can be practiced without these specific details.

[0099] To address the problems in existing technologies, such as the tendency to weaken the direction-dependent qP phase velocity background in the approximation of the reflection coefficient of PP waves in TTI media, the difficulty in explicitly expressing the coupling relationship between the incident angle, observation azimuth angle, and symmetry axis tilt angle using conventional isotropic or VTI reflection kernels, and the significant crosstalk between elastic and anisotropic parameters in multi-parameter pre-stack inversion, this invention provides a rational impedance kernel-driven multi-parameter pre-stack inversion method and apparatus for TTI media. This technical solution comprises three interconnected core steps: construction of the phase propagation direction-tilted symmetry axis coupled phase velocity, construction of the rational impedance kernel, and multi-parameter pre-stack inversion of TTI media. The direction-dependent qP phase velocity background constructed in the previous step serves as the background input for the directional equivalent qP impedance and reflection kernel function in the subsequent step. The updated model obtained in the inversion step is then used to recalculate the direction-dependent qP phase velocity background and rational impedance kernel, thus forming a closed-loop relationship between the construction of the directional background, the construction of the reflection kernel, and the model update.

[0100] like Figure 1 The diagram shown is an overall process flow chart of pre-stack inversion of TTI dielectric multi-parameter driven by rational impedance kernel according to an embodiment of the present invention, which includes the following five steps:

[0101] S1: Data Acquisition and Data Preprocessing

[0102] Acquire multi-angle, multi-azimuth pre-stack PP wave seismic data for the target work area. The pre-stack PP wave seismic data may include angle gathers, azimuth angle gathers, or pre-stack gathers with both angle and azimuth indices. Further acquire seismic wavelets, well logging curves, stratigraphic interpretation results, geological structural interpretation results, low-frequency initial models, and prior knowledge of the dip and azimuth axis of the TTI medium.

[0103] The logging curves include P-wave velocity, S-wave velocity, density, and parameters that can be constructed from rock physics relationships, prior interpretations, or inversion results. , and The low-frequency initial model can be established by combining well logging data, stratigraphic interpretation results, geological trend constraints, and existing inversion results.

[0104] During preprocessing, data from different sources are unified to the same time or depth domain, and sampling interval matching, outlier removal, smoothing, angle and azimuth range filtering, wavelet estimation, and parameter normalization are performed. For multi-parameter inversion, since the dimensions and numerical ranges of P-wave velocity, S-wave velocity, density, and Thomsen parameters are different, it is preferable to perform scale normalization on each parameter to reduce the impact of parameter scale differences on the damped Gauss-Newton solution.

[0105] S2: Construct a phase propagation direction-tilted symmetry axis coupled phase velocity background

[0106] like Figure 2 The diagram shown illustrates the coupling relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium, as well as the construction of the direction-dependent phase velocity background, according to an embodiment of the present invention. Local coordinate axes are established for each target reflecting interface. ,in The normal direction of the reflection interface. Located within the reflecting plane, Orthogonal to the reflection plane. Define the unit vector of the TTI medium axis of symmetry, the unit vector of the seismic survey line direction, and the unit vector of the qP wave incident phase direction as follows:

[0107]

[0108]

[0109]

[0110] in, The tilt angle is the axis of symmetry. To observe the azimuth angle, The angle of incidence. The angle between the phase propagation direction of the qP wave and the tilted axis of symmetry. Calculated by the following formula:

[0111]

[0112] The direction-dependent qP phase velocity background factor mentioned in step S2 is obtained by approximating the rotationally weakly anisotropic qP phase velocity, which is expressed as:

[0113]

[0114] in, For vertical or background P-wave velocity, and Here are the Thomsen anisotropy parameters; the direction-dependent qP background factor is expressed as:

[0115]

[0116] Or, under the weak anisotropy first-order approximation, it can be expressed as:

[0117]

[0118] Through the above steps, the qP phase velocity background is no longer a single vertical velocity or isotropic velocity, but a direction-dependent background quantity determined by the incident angle, observation azimuth angle, symmetry axis tilt angle, and anisotropic parameters.

[0119] The direction-dependent qP phase velocity background obtained in step S2 is not used as an independent velocity calculation result. If only the direction-dependent qP phase velocity is calculated and the PP wave reflection coefficient is still normalized using a conventional isotropic background impedance, the above propagation direction information will still mainly be an additional correction and cannot directly change the background normalization scale of the reflection amplitude. Therefore, this invention further uses the direction-dependent qP phase velocity background obtained in step S2 as the input for constructing the directional equivalent qP impedance in step S3, so that the incident angle, observation azimuth angle, and TTI symmetry axis geometric relationship are further transferred from the phase velocity background to the denominator normalization structure and parameter sensitivity kernel of the PP wave reflection coefficient.

[0120] S3: Construct a rational impedance kernel

[0121] like Figure 3 As shown, this invention demonstrates, through a comparison of the amplitude responses of the reflection coefficients of the accurate reflection coefficient, the traditional linear approximation, and the phase velocity-constrained rational approximation, that the rational impedance kernel approximation has good amplitude fitting capability within the commonly used incident angle range. In the specific construction process, the directional equivalent qP impedance is constructed based on the direction-dependent qP background factor:

[0122]

[0123] in, For density, The vertical longitudinal wave velocity, This is the direction-dependent qP background factor. This directional equivalent qP impedance describes the directional background impedance perceived by the PP wave at the current incident angle and observation azimuth. It is not a strictly accurate impedance under the complete TTI elastic boundary conditions, but rather an equivalent impedance used to normalize the reflection coefficient background under the assumptions of weak anisotropy and weak interface contrast.

[0124] Furthermore, to describe the shear-related anisotropy parameters Visibility under different observation azimuths is used to construct the azimuth projection factor:

[0125]

[0126] in, The unit vector for the direction of the seismic survey line. The unit vector of the tilt symmetry axis of the TTI medium. The tilt angle is the axis of symmetry. To observe the azimuth angle, the Thomsen parameters are adjusted using the azimuth projection factor. The relevant items are weighted by azimuth visibility to make The contribution of parameters to the reflection amplitude of PP waves varies with the observation azimuth and the direction of the TTI symmetry axis.

[0127] Under weak interface contrast conditions, the centralized average value of the upper and lower medium parameters of the interface is used as the background quantity, and the difference between the upper and lower medium parameters is used as the interface contrast quantity. For the first... The interface comparison vector for the six-parameter interface is as follows:

[0128]

[0129] When constructing the reflection coefficient, the denominator normalization structure of the directional equivalent qP impedance is retained, and the background impedance is not fully expanded into a linear polynomial; only the interface parameter comparison is linearized to the first order. This yields the phase velocity-constrained rational PP wave reflection coefficient:

[0130]

[0131] in, The reflection coefficient of a phase velocity-constrained rational PP wave. , , This is a centralized background quantity for the interface's upper and lower medium parameters. , , , , , Due to differences in interface parameters, , , , , , The analytical sensitivity kernel function is determined by the incident angle, observation azimuth angle, TTI symmetry axis tilt angle, direction-dependent qP background factor, direction-equivalent qP impedance, and shear projection factor.

[0132] in, For the first The first interface, the first The analytical reflection kernel function vector corresponding to each angle-azimuth sampling point is represented as:

[0133]

[0134] The corresponding reflection coefficient is written as:

[0135]

[0136] Each kernel function is determined by the incident angle, observation azimuth angle, angle between the qP phase propagation direction and the tilted symmetry axis, direction-dependent qP background factor, direction-equivalent qP impedance, shear projection factor, and current background parameters. During the construction of the reflection coefficient kernel function, the direction-equivalent qP impedance or its normalized form is retained in the denominator, giving the reflection coefficient a rational background normalization structure, rather than being fully expanded into a linear polynomial with perturbations of each parameter.

[0137] Furthermore, since the directional equivalent qP impedance and the phase velocity-constrained rational PP wave reflection kernel in step S3 are determined by the current P-wave velocity, S-wave velocity, density, anisotropy parameters, and propagation direction geometry, the original directionally correlated background and the medium state corresponding to the reflection kernel change as the inversion model is updated. If the initial directional background and reflection kernel remain unchanged throughout the entire inversion process, inconsistencies will gradually arise between the reflection kernel and the current inversion model. Therefore, this invention further embeds steps S2 and S3 into the iterative process of step S4, recalculating the directionally correlated qP phase velocity background and reconstructing the rational impedance kernel after the model is updated.

[0138] S4: Joint forward and inverse modeling of multiple parameters in TTI media based on rational impedance kernel

[0139] like Figure 4 and Figure 5 As shown, the inversion input conditions and technical effectiveness of the method can be verified by comparing the real model of the layered model with the low-frequency initial model and by comparing the inversion results of different methods.

[0140] In the inversion implementation, the layer parameter model Through the difference operator The model is converted to an interface contrast model, and the reflection coefficient sequence is convolved with the seismic wavelet. For the th The composite seismic record is obtained from 1 angle-azimuth sampling point:

[0141]

[0142] in, For layer parameter difference operators, For seismic wavelet convolution operators, The reflection kernel matrix is ​​composed of phase velocity-constrained rational impedance kernel functions. By jointly arranging all angle-azimuth sampling points, the overall forward modeling relation is obtained:

[0143]

[0144] in This is a pre-stack record of the combined synthesis of all angle-azimuth sampling points. It is a joint forward modeling operator that includes layer parameter difference, rational impedance kernel function and wavelet convolution.

[0145] Construct the inversion objective function:

[0146]

[0147] in, To observe pre-stack PP wave seismic data, For the current iteration model or the initial model, The prior model is constructed from well logging data, geological stratigraphy, and low-frequency trends. It is a first-order or second-order difference smoothing operator. and These are the smoothing constraint weights and the prior constraint weights, respectively.

[0148] In the In the next iteration, the forward response at the current model is linearized to the first order to construct the Jacobian matrix. And solve for the model increment:

[0149]

[0150] The model is updated as follows:

[0151]

[0152] in, The damping factor, Step size factor Let be the identity matrix. If the objective function decreases significantly, decrease the damping factor to improve convergence efficiency; if the objective function does not decrease significantly or the model update is unstable, increase the damping factor to suppress divergence.

[0153] In each outer iteration, the direction-dependent qP background factor, the direction-equivalent qP impedance, and the shear projection factor are updated, and in each inner iteration, the six-parameter interface comparison or layer parameter increment is solved.

[0154] S5: Output TTI medium multi-parameter inversion results and quality control

[0155] like Figure 6 As shown, this invention can also be applied to pre-stack angle gather data in actual survey areas, and outputs a six-parameter inversion profile of the actual data. When a preset termination condition is met, the six-parameter inversion results are output, including P-wave velocity. transverse wave velocity ,density Thomsen parameters , and The termination condition includes at least one of the following:

[0156] (1) The data residual between the observed data and the synthetic data is less than the preset threshold;

[0157] (2) The decrease in the objective function between two consecutive iterations is less than a preset threshold;

[0158] (3) The model update amount in two consecutive iterations is less than the preset threshold;

[0159] (4) The damping factor or step size factor reaches the preset range;

[0160] (5) The number of iterations reaches the preset maximum number.

[0161] Further outputs include synthetic records, residual profiles, objective function curves, well-seismic comparison results, angle-azimuth domain amplitude fitting results, and parameter update values, which can be used to evaluate the reliability of the inversion results.

[0162] To illustrate the feasibility and technical effects of the method of the present invention, the following can be used: Figure 3 The comparison of single-interface reflection coefficient amplitude response is shown. Figure 4 The input conditions of the one-dimensional layered discontinuous model shown are as follows: Figure 5 The six-parameter inversion results shown are compared and verified. Among them, Figure 3 Used to characterize the fitting relationship between the rational impedance kernel approximation and the accurate reflection coefficient of the present invention; Figure 4 Used to characterize the difference between the real model required for inversion and the low-frequency initial model; Figure 5 Used to characterize the traditional linear method and the method of this invention in terms of P-wave velocity, S-wave velocity, density, and Thomsen parameter. , , The difference in inversion effect.

[0163] Furthermore, it is possible to adopt Figure 6The actual pre-stack angle gather and six-parameter inversion results profile shown illustrate the application effect of the method of this invention in actual industrial data. The application of actual data may include pre-stack angle gather input, prior constraints of structural dip angle or TTI symmetry axis, six-parameter joint inversion output, and well-seismic consistency or residual fitting quality control results.

[0164] This embodiment provides a TTI medium PP wave multi-parameter pre-stack inversion device, including:

[0165] The data acquisition and preprocessing module is used to acquire multi-angle, multi-azimuth pre-stack PP wave seismic data, well logging curves, seismic wavelets, stratigraphic interpretation results, low-frequency initial models, and TTI medium symmetry axis dip angle and azimuth priors, and to complete data registration, resampling, normalization and quality control.

[0166] The phase propagation direction-tilted symmetry axis coupled phase velocity construction module is used to establish a local TTI reflection geometric coordinate system, calculate the angle between the qP phase propagation direction and the tilted symmetry axis, and construct the direction-related qP background factor based on the rotational weak anisotropy qP phase velocity approximation.

[0167] The rational impedance kernel construction module is used to construct the directional equivalent qP impedance, the azimuth projection factor of the shear-related anisotropic parameter, and the phase velocity-constrained rational PP wave reflection coefficient and analytical sensitivity kernel function while retaining the denominator normalization structure of the directional equivalent qP impedance.

[0168] The TTI medium multi-parameter pre-stack inversion module is used to combine layer parameter difference operators, rational impedance kernel function matrices, and seismic wavelet convolution operators to form multi-angle and multi-azimuth joint forward modeling relationships. It constructs the objective function, solves the model increment, adjusts the damping factor and step size factor, and updates the inversion model. After the model is updated, the updated model is fed back to the phase propagation direction-tilted symmetry axis coupled phase velocity construction module and the rational impedance kernel construction module to recalculate the direction-related qP phase velocity background, direction-equivalent qP impedance, azimuth projection factor, and phase velocity-constrained rational PP wave reflection kernel. The reconstructed reflection kernel is then used for the next model iteration.

[0169] The results output and quality control module is used to output P-wave velocity, S-wave velocity, density, and other parameters. , and The inversion results of six parameters are output, along with at least one quality control result from the following: synthetic record, residual profile, objective function curve, well-seismic comparison results, angle-azimuth domain amplitude fitting results, and parameter update amount.

[0170] This embodiment provides a computer-readable storage medium storing at least one instruction adapted to be loaded by a processor and to execute the steps of the method described in Embodiment 1.

[0171] This embodiment also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program adapted to be loaded by the processor and executed according to the method steps. The electronic device may be a workstation, server, cloud computing node, seismic data processing and interpretation platform, or other device with computing capabilities.

[0172] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that, without departing from the core concept of the present invention, equivalent substitutions or modifications can be made to the parameter forms, angle definition methods, regularization term forms, damping adjustment strategies, wavelet estimation methods, low-frequency model construction methods, and output quality control methods in the above technical solutions. All equivalent changes made based on the phase propagation direction-tilted symmetry axis coupled phase velocity background construction, directional equivalent qP impedance rational normalization, six-parameter interface comparison linearization, and damped Gaussian-Newton multi-angle multi-directional PP wave inversion framework of the present invention should fall within the scope of protection of the present invention.

Claims

1. A rational impedance kernel-driven pre-stack inversion method for multi-parameter TTI dielectrics, characterized in that, Includes the following steps: Step 1: Obtain multi-angle, multi-azimuth pre-stack PP wave seismic data and TTI medium prior information for the target work area; Step 2: Establish the geometric coupling relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium. Based on the rotational weak anisotropy qP phase velocity approximation, obtain the direction-dependent qP phase velocity background factor determined by the incident angle, observation azimuth angle, and symmetry axis tilt angle. Step 3: Introduce the direction-dependent qP phase velocity background factor into the direction-equivalent qP impedance. While retaining the direction-equivalent qP impedance as the denominator normalization structure, perform first-order linearization on the interface comparison of elastic parameters and anisotropic parameters to construct the phase velocity-constrained rational PP wave reflection coefficient. Step 4: Based on the phase velocity-constrained rational PP wave reflection coefficient, construct a multi-angle, multi-directional joint forward modeling relationship, and perform iterative inversion including elastic parameters and anisotropic parameters; During the iteration process, the direction-related qP phase velocity background factor is updated according to the current model, and the direction-equivalent qP impedance and the phase velocity-constrained rational PP wave reflection kernel are reconstructed based on the updated direction-related qP phase velocity background factor, so that the direction-related qP phase velocity background, the direction-equivalent qP impedance and the reflection kernel function are updated synchronously with the inversion model. Step 5: Output the inversion results of the elasticity parameters and anisotropy parameters.

2. The method according to claim 1, characterized in that, Step 1 includes: Step 1.1: Acquire multi-angle, multi-azimuth pre-stack PP wave seismic data, seismic wavelets, well logging curves, stratigraphic interpretation results, low-frequency initial model, and prior knowledge of the dip and azimuth axis of the TTI medium for the target work area; wherein, the pre-stack PP wave seismic data includes angle gathers, azimuth angle gathers, or pre-stack gathers with both angle and azimuth indices; the well logging curves include P-wave velocity, S-wave velocity, density, and Thomsen anisotropy parameters. , and curve; Step 1.2: The pre-stack PP wave seismic data is processed by angle domain or azimuth domain gathers to form multidimensional observation data indexed by incident angle and observation azimuth angle; Step 1.3: Perform time-depth conversion, resampling, outlier removal, and smoothing on the well logging curves to make them consistent with the sampling interval of the pre-stack PP wave seismic data, the layer interpretation results, and the inversion grid; Step 1.4: Construct the low-frequency initial model based on the well logging curves, geological strata, and existing velocity or impedance interpretation results; Step 1.5: Perform time-domain or depth-domain registration, scale unification, and noise suppression on the pre-stack PP wave seismic data, well logging curves, and low-frequency initial model; normalize parameters of different dimensions to ensure that P-wave velocity, S-wave velocity, density, and Thomsen anisotropy parameters are uniform. , and They have similar update scales during the inversion process; Step 1.6: Based on the structural interpretation results of the work area, bedding dip angle, offset imaging results, or external geological priors, establish the priors for the dip angle and azimuth of the TTI medium's symmetry axis.

3. The method according to claim 1, characterized in that, Step 2 includes: Step 2.1: Establish a local TTI reflection geometry coordinate system for the target reflection interface, where the local coordinate axes are... , The normal direction of the reflection interface. Located within the reflecting plane, Orthogonal to the reflecting plane; Step 2.2: Define the unit vector of the tilt symmetry axis of the TTI medium. Unit vector of seismic survey line direction and the unit vector of the incident phase direction of the qP wave Calculate the angle between the phase propagation direction of the qP wave and the tilted axis of symmetry. ; Step 2.3: Based on the rotationally weakly anisotropic qP phase velocity approximation, according to the included angle... Constructing the directionally correlated qP phase velocity background factor .

4. The method according to claim 3, characterized in that, In step 2.2, the included angle satisfy: in, The incident angle of the PP wave. To observe the azimuth angle, The angle of inclination of the TTI axis of symmetry relative to the interface normal direction; In step 2.3, the direction-related qP phase velocity background factor Represented as: Or, under the weak anisotropy first-order approximation, it can be expressed as: in, Let qP be the phase velocity of the wave. The vertical longitudinal wave velocity, and For Thomsen anisotropy parameters.

5. The method according to claim 1, characterized in that, Step 3 includes: Step 3.1: Based on the direction-related qP phase velocity background factor Constructing the directional equivalent qP impedance : in, For density, The vertical longitudinal wave velocity, The background factor for the direction-dependent qP phase velocity; Step 3.2: Construct the azimuth projection factor of shear-related anisotropy parameters : in, The unit vector for the direction of the seismic survey line. The unit vector of the tilt symmetry axis of the TTI medium. The tilt angle is the axis of symmetry. To observe the azimuth angle; using the azimuth projection factor For Thomsen parameters The relevant items are weighted by directional visibility; Step 3.3: Under the assumptions of weak anisotropy and weak interface contrast, retain the denominator normalization structure of the directional equivalent qP impedance, and only normalize the P-wave velocity, S-wave velocity, density, and Thomsen parameter. , and The interface comparison is linearized to the first order to construct the phase velocity constrained rational type PP wave reflection coefficient.

6. The method according to claim 5, characterized in that, In step 3.3, the phase velocity-constrained rational type PP wave reflection coefficient is expressed as: in, , , Centralized background quantity for the interface's upper and lower medium parameters; , , , , , Differences in interface parameters; , , , , , The analytical sensitivity kernel function is determined by the incident angle, observation azimuth angle, TTI symmetry axis tilt angle, direction-dependent qP phase velocity background factor, direction-equivalent qP impedance, and azimuth projection factor. For the The six-parameter interface comparison vector for the reflective interface is: For the For each incident angle-observation azimuth angle sampling point, the reflection kernel function vector is: The corresponding reflection coefficient is .

7. The method according to claim 1, characterized in that, Step 4 includes: Step 4.1: Convert the layer parameter model using the difference operator Converted into an interface comparison model, the reflection coefficient sequence is processed by the seismic wavelet convolution operator. Converted to synthetic seismic records; for the first The forward modeling relationship for each incident angle-observation azimuth sampling point is expressed as: in, To synthesize pre-stack records, The reflection kernel matrix is ​​composed of the reflection coefficients of the phase velocity-constrained rational type PP wave. The model for the layer parameters to be inverted is D, and the layer parameter difference operator is D. By combining the forward modeling relationships of all incident angles and observation azimuth angles, we obtain: in This is a pre-stack record of the combined synthesis of all angle-azimuth sampling points. It is a joint forward modeling operator that includes a reflection kernel function, a layer parameter difference operator, and a wavelet convolution operator; Step 4.2: Construct a multi-parameter inversion objective function that includes data fitting terms, model smoothing constraint terms, and prior model constraint terms: in, To observe pre-stack PP wave seismic data, For the current iteration model or the initial model, The prior model is constructed from well logging data, geological stratigraphy, and low-frequency trends. It is a first-order or second-order difference smoothing operator. As a smoothing constraint factor, These are prior constraint factors; Step 4.3: Solve the model increment using the damped Gauss-Newton iterative method. For each outer iteration, recalculate the geometric coupling relationship between the qP wave phase propagation direction and the tilted symmetry axis of the TTI medium based on the current inversion model, update the direction-dependent qP phase velocity background factor, and reconstruct the directional equivalent qP impedance, azimuth projection factor, and phase velocity-constrained rational PP wave reflection kernel function based on the updated direction-dependent qP phase velocity background factor. Under the condition that the direction-dependent background and reflection kernel function corresponding to the current outer iteration remain unchanged, solve the six-parameter interface comparison or layer parameter increment through inner iteration. After completing the model update, update the direction-dependent background and reflection kernel function again until the preset iteration termination condition is met.

8. The method according to claim 7, characterized in that, In step 4.3, in the first In the next iteration, for the current model The forward response at a given point is linearized to the first order to construct the Jacobian matrix. The model increment is solved using the following formula: in, The damping factor, The identity matrix is ​​used; the model is updated as follows: in, For the first The model of the next iteration For the first The model of the next iteration This is the step size factor, determined based on the descent of the objective function, changes in data residuals, or the line search strategy.

9. The method according to claim 1, characterized in that, Step 5 includes: Step 5.1: Determine whether a preset iteration termination condition is met, wherein the termination condition includes at least one of the following: (1) The data residual between the observed data and the synthetic data is less than the preset threshold; (2) The decrease in the objective function between two consecutive iterations is less than a preset threshold; (3) The model update amount in two consecutive iterations is less than the preset threshold; (4) The damping factor or step size factor reaches the preset range; (5) The number of iterations reaches the preset maximum number; Step 5.2: When the termination condition is met, output the TTI medium multi-parameter inversion results for the target work area. The inversion results include P-wave velocity volume, S-wave velocity volume, density volume, and Thomsen anisotropy parameters. Volume, Thomsen anisotropy parameters Volume and Thomsen anisotropy parameters body; Step 5.3: Output at least one quality control result from the following: synthetic record, residual profile, objective function convergence curve, parameter update amount, well-seismic comparison curve, layer constraint comparison result, and angle-azimuth domain amplitude fitting result, to evaluate the reliability of the inversion results.

10. A rational impedance kernel-driven TTI dielectric multi-parameter pre-stack inversion device, implementing the method as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire pre-stack PP wave seismic data from multiple angles and orientations, as well as TTI medium prior information. The phase propagation direction-tilted symmetry axis coupled phase velocity construction module is used to establish the geometric coupling relationship between the phase propagation direction of the qP wave and the tilted symmetry axis of the TTI medium, and to obtain the direction-related qP phase velocity background factor based on the rotational weak anisotropy qP phase velocity approximation. The rational impedance kernel construction module is used to introduce the direction-dependent qP phase velocity background factor into the directional equivalent qP impedance. Under the premise of retaining the directional equivalent qP impedance as the denominator normalization structure, the interface comparison of elastic parameters and anisotropic parameters is linearized in the first order to construct the phase velocity-constrained rational PP wave reflection coefficient. The TTI medium multi-parameter pre-stack inversion module is used to construct a multi-angle, multi-azimuth joint forward modeling relationship based on the phase velocity-constrained rational PP wave reflection coefficient, and to perform iterative inversion including elastic parameters and anisotropic parameters. During the iteration process, the direction-related qP phase velocity background factor is updated according to the current inversion model, and the directional equivalent qP impedance, azimuth projection factor, and phase velocity-constrained rational PP wave reflection kernel function are reconstructed based on the updated directional-related qP phase velocity background factor, so that the direction-related background, directional equivalent qP impedance, and reflection kernel function are updated synchronously with the inversion model. The result output module is used to output the inversion results of the elasticity parameters and anisotropy parameters.