A method for extracting a deep domain azimuthal seismic wavelet considering phase

Abstract

The application discloses a kind of phase's deep domain azimuth seismic wavelet extraction method, comprising the following steps: S1, the characteristics of depth domain non-stationary seismic signal is extracted instantaneous phase and instantaneous wave number information by Hilbert transform;S2, on the basis of traditional S transform, introduce nonlinear wave number weight and depth-dependent instantaneous wave number term, construct the optimization generalized S transform considering phase effect, realize depth-wave number domain high-resolution spectral decomposition;S3, restore depth domain seismic record by inverse transform, and extract seismic wavelet varying with depth, incidence angle and azimuth in combination with initial reflection coefficient model.According to the application, the variation law of depth domain seismic wavelet with depth, incidence angle and azimuth is provided, and reliable geophysical basis is provided for depth domain seismic inversion and anisotropic reservoir prediction.

Description

Technical Field

[0001] This invention relates to the technical field of seismic exploration and geophysical inversion, and particularly to a method for extracting depth-domain azimuth seismic wavelets that considers phase. Background Technology

[0002] With the development of pre-stack depth migration technology, seismic data interpretation is gradually shifting from the time domain to the depth domain. Compared to time-domain seismic data, depth-domain seismic data can more realistically reflect the characteristics of subsurface structures and directly correspond with well logging data in depth space, thereby improving the accuracy of reservoir description and parameter inversion. However, due to the propagation effects of seismic waves in complex media, depth-domain seismic data usually exhibits significant non-stationary characteristics, with seismic wavelets changing with depth, making traditional methods that assume wavelet invariance difficult to apply.

[0003] Seismic wavelet extraction is a crucial step in seismic inversion. Various depth-domain wavelet extraction methods have been proposed in previous studies. For example, Zhang et al. (2017) estimated depth-domain wavelets based on compressed sensing, and Cai et al. (2024) used spectral decomposition to construct a wavelet dictionary, improving the accuracy of wavelet extraction. These methods have improved the depth-domain wavelet extraction effect to some extent, but most focus on amplitude information and do not adequately consider phase information, which can easily lead to wavelet estimation errors under complex geological conditions.

[0004] In signal analysis methods, the S-transform is widely used in the analysis of non-stationary seismic signals due to its excellent time-frequency localization ability and ability to preserve absolute phase information (Stockwell et al., 1996). Zhang and Deng (2018, 2023) introduced the S-transform into depth-domain wavelet extraction, improving the ability to characterize wavelet variations with depth. However, the traditional S-transform has certain limitations in wavenumber resolution and depth resolution, and is prone to spectral aliasing under complex reservoir conditions.

[0005] Furthermore, the phase information of seismic signals has a significant impact on the morphology of seismic wavelets. The instantaneous attribute analysis method proposed by Taner et al. (1979) shows that phase information plays an important role in describing the characteristics of seismic signals. However, in existing depth-domain wavelet extraction methods, phase information is usually not fully utilized, resulting in poor stability of wavelet extraction results when phase distortion or non-stationary propagation effects are present.

[0006] In summary, existing methods for depth-domain seismic wavelet extraction still have shortcomings: firstly, insufficient utilization of wavelet phase information; secondly, difficulty in simultaneously achieving depth resolution and wavenumber resolution; and thirdly, limited stability under low signal-to-noise ratio conditions. Therefore, it is necessary to propose a new depth-domain seismic wavelet extraction method that can improve resolution while fully utilizing phase information, thereby enhancing the accuracy and stability of wavelet extraction. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the present invention aims to provide a phase-considered depth-domain azimuth seismic wavelet extraction method, providing a foundation for characterizing the variation of depth-domain seismic wavelets with depth, incident angle, and azimuth angle, and simultaneously providing reliable geophysical evidence for depth-domain seismic inversion and anisotropic reservoir prediction. To achieve the above-mentioned objectives and other advantages of the present invention, a phase-considered depth-domain azimuth seismic wavelet extraction method is provided, comprising: Starting from the characteristics of non-stationary seismic signals in the depth domain, we first construct a complex representation of the seismic signal using the Hilbert transform to extract instantaneous phase and instantaneous wavenumber information. Second, based on the traditional S-transform, we introduce nonlinear wavenumber weights and depth-dependent instantaneous wavenumber terms to construct an optimized generalized S-transform that considers phase effects, achieving high-resolution spectral decomposition in the depth-wavenumber domain. Finally, we recover the depth domain seismic record through inverse transform and, combined with the initial reflection coefficient model, extract seismic wavelets that vary with depth, incident angle, and azimuth angle.

[0008] Preferably, it includes the following steps: 1) Input pre-stack seismic gathers with different incident angles and azimuth angles, construct the complex representation of the seismic signal based on the Hilbert transform, extract the instantaneous phase and instantaneous wavenumber information, and introduce them into the S-transform window function to establish an optimized generalized S-transform expression containing depth-related terms. 2) Input pre-stack seismic gathers with different incident angles and azimuth angles, perform depth-wavenumber domain spectral decomposition using optimized generalized S-transform, and introduce phase information to obtain a complex spectral representation containing amplitude and phase. 3) The depth-domain seismic record is recovered by inverse Fourier transform, and combined with the initial reflection coefficient model, a depth-varying wavelet matrix is ​​constructed to extract the seismic wavelet that varies with depth, incident angle, and azimuth angle. Compared with the prior art, the advantages and positive effects of the present invention are: By incorporating instantaneous wavenumber information to improve the S-transform window function, the depth resolution is enhanced, effectively reducing spectral aliasing. Secondly, considering both amplitude and phase information during wavelet extraction allows for more accurate recovery of seismic wavelet morphology compared to traditional methods that only utilize amplitude spectra. Thirdly, this method can extract seismic wavelets that vary with azimuth angle, preserving anisotropic response characteristics and providing support for azimuth seismic analysis. Finally, the established method exhibits good stability and anti-interference capabilities under noisy conditions, making it effective for depth-domain seismic data processing and interpretation under complex geological conditions. Attached Figure Description

[0009] Figure 1A flowchart of the depth-domain azimuth seismic wavelet extraction method considering phase according to the present invention; Figure 2 Figure a) shows the application of the depth-domain azimuth seismic wavelet extraction method considering the phase of the present invention to a simple model depth-domain synthetic seismic data; Figure b) shows the depth-domain seismic record; Figure c) shows the depth wavenumber spectrum extracted by S-transform.

[0010] Figure 3 Figure 1 shows a comparison between the seismic wavelet extracted from a simple model and the actual wavelet, based on the phase-considered depth-domain azimuth seismic wavelet extraction method of the present invention; Figure 2 shows the Ricker wavelet with k=20 / km; Figure 3 shows the Ricker wavelet with k=15 / km; Figure 4 shows the Ricker wavelet with k=10 / km.

[0011] Figure 4 To illustrate the application of the phase-considered depth-domain azimuth seismic wavelet extraction method of the present invention to actual depth-domain seismic data, wavelet extraction is performed, and a comparison map of seismic records and actual seismic records is synthesized. Detailed implementation method. The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] Reference Figure 1 A depth-domain azimuth seismic wavelet extraction method considering phase, comprising: 1) Instantaneous phase and instantaneous wavenumber extraction is performed using depth-domain seismic data as input. The specific calculation is as follows: First, the input pre-stack azimuth seismic data is subjected to a Hilbert transform, converting the real-valued seismic signal into a complex signal form. Its expression is as follows: (1); in, Represents the depth axis. The original seismic signal, For the Hilbert transform operator, This is a re-earthquake signal.

[0013] After obtaining the complex seismic signal, the instantaneous phase and instantaneous wavenumber are further calculated: (2); (3); in, For instantaneous phase, The instantaneous wavenumber is used to characterize the local frequency characteristics of seismic signals as they vary with depth.

[0014] 2) Considering the phase effect, the optimized generalized S-transform is constructed by introducing nonlinear wavenumber weights and instantaneous wavenumber terms on the basis of the traditional S-transform, and improving the window function.

[0015] The traditional depth-domain S-transform expression is: (4); The Gaussian window function is: (5); in, For depth delay axis, For wave number, Representing the incident angle, the traditional S-transform has low depth and wavenumber resolution, making it unable to accurately describe the changes in depth-domain seismic data.

[0016] To improve depth resolution, the window width of the traditional S-transform is... Optimization, introducing improved window width: (6); in, As a scale factor, Wavenumber weighted index As a stable term, Here are the instantaneous wavenumber weighting coefficients. The improved window function is: (7); Meanwhile, based on the optimized generalized S-transform, we consider using pre-stack seismic data in the depth domain for wavelet extraction to construct an optimized generalized S-transform: (8); in, The spectrum obtained by optimizing the generalized S-transform is represented. Representing the azimuth angle, and considering the phase introduced by the S-transform, the depth wavenumber spectrum of the S-transform is expressed in complex form: (9); in, The representative is the amplitude spectrum. This represents taking the phase spectrum.

[0017] 3) Depth-wavenumber domain spectral reconstruction Integrating both sides of the result of optimizing the generalized S-transform in formula (8) with respect to the depth delay axis, since the integral of the Gaussian window function is 1, we obtain the wavenumber spectrum in the depth domain: (10); The depth-domain seismic signal is recovered by inverse Fourier transform of both sides of equation (10): (11).

[0018] 4) Depth-domain seismic wavelet extraction In seismic interpretation, convolutional models are used to synthesize seismic records. To address the non-stationarity of depth-domain seismic data, Margarave (1998) proposed a non-stationary convolutional method to represent depth-domain seismic data.

[0019] (12); Combining formulas (8) and (12), the expression for the depth-domain azimuth seismic wavelet can be obtained: (13); in The approximate value of the reflection coefficient varies with the incident angle and azimuth angle. However, in general, we cannot obtain accurate reflection coefficient values ​​from seismic data without well logging information. Therefore, we can use an initial model of the reflection coefficient. Using a more precise reflection coefficient instead, good wavelet extraction results can still be obtained.

[0020] The experimental simulation results are as follows: To verify the effectiveness and stability of this invention in depth domain seismic wavelet extraction and reconstruction, the invention was applied to synthetic seismic data and actual seismic data for analysis. Figure 2 The following table shows the results of processing depth-domain synthetic seismic data using a simple model, where (a) is the depth-domain seismic record and (b) is the depth wavenumber spectrum extracted by S-transform. It can be seen that this invention can effectively characterize the wavenumber features of seismic signals in the depth domain, with a clear spectral energy distribution and prominent dominant frequency information, providing a reliable basis for subsequent wavelet extraction.

[0021] Figure 3 To illustrate the application of this invention to wavelet extraction from model data and compare it with actual wavelets, (a), (b), and (c) show Ricker wavelets with wavenumbers of k=20 / km, k=15 / km, and k=10 / km, respectively. It can be seen that the wavelets extracted by this invention exhibit high consistency with the theoretical wavelets under different wavenumber conditions, showing good matching in both waveform morphology and dominant frequency characteristics. This demonstrates that the method possesses high accuracy and good adaptability in depth domain wavelet recovery.

[0022] Figure 4To apply this invention to wavelet extraction from actual depth-domain seismic data, a comparison was made between the synthesized seismic record obtained from the extracted wavelet and the actual seismic record. It can be seen that the wavelet obtained using this invention can reconstruct the original seismic record well. The synthesized record and the actual record show good consistency in amplitude characteristics and phase axis continuity, indicating that this method also has good stability and practicality in actual data processing.

[0023] The number of devices and processing scale described herein are for simplification purposes. Applications, modifications, and variations of this invention will be readily apparent to those skilled in the art. Although embodiments of the invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. It can be applied to various fields suitable for this invention, and further modifications can be readily implemented by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, this invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A depth-domain azimuth seismic wavelet extraction method considering phase, characterized in that, Includes the following steps: S1. Extract instantaneous phase and instantaneous wavenumber information from the characteristics of non-stationary seismic signals in the depth domain using Hilbert transform; S2. Based on the traditional S-transform, nonlinear wavenumber weights and depth-dependent instantaneous wavenumber terms are introduced to construct an optimized generalized S-transform that considers phase effects, thereby achieving high-resolution spectral decomposition in the depth-wavenumber domain. S3. The depth domain seismic record is recovered by inverse transformation, and the seismic wavelet that varies with depth, incident angle and azimuth angle is extracted by combining the initial reflection coefficient model.

2. The depth-domain azimuth seismic wavelet extraction method considering phase as described in claim 1, characterized in that, Step S1 specifically includes: Input pre-stack seismic gathers with different incident angles and azimuth angles; Based on the Hilbert transform, a complex representation of the seismic signal is constructed, and instantaneous phase and instantaneous wavenumber information are extracted; A depth-dependent S-transform window function is constructed using instantaneous phase and instantaneous wavenumber information, and an optimized generalized S-transform expression containing depth-dependent terms is established.

3. The optimized generalized domain azimuth seismic wavelet extraction method considering phase effects as described in claim 2, characterized in that, Step S2 specifically includes performing depth-wavenumber domain spectral decomposition through optimized generalized S-transform, while introducing phase information to obtain a complex spectrum representation containing amplitude and phase.

4. The depth-domain azimuth seismic wavelet extraction method considering phase as described in claim 3, characterized in that, When performing depth-wavenumber domain spectral decomposition, both amplitude and phase information are preserved to construct a complex spectral representation.

5. The method for extracting depth-domain azimuth seismic wavelet considering phase as described in claim 1, characterized in that, Input pre-stack seismic gather data with different incident and azimuth angles, perform optimized generalized S-transform processing respectively, and obtain depth-wavenumber spectra that vary with azimuth and incident angle.

6. The method for extracting depth-domain azimuth seismic wavelet considering phase as described in claim 1, characterized in that, The depth-wavenumber spectrum is first integrated along the depth delay axis, then inverse Fourier transform is performed to recover the depth domain seismic record, and the wavelet matrix is ​​constructed by combining the initial reflection coefficient model.

Images