Fracture identification method based on spectrum variance coherence

Abstract

The invention belongs to the field of seismic data processing, and particularly relates to a fracture identification method based on spectrum variance coherence. The fracture identification method comprises the following steps: (1) selecting a data point from a three-dimensional seismic data volume, and constructing a seismic data sub-volume by taking the data point as a center; (2) calculating an amplitude spectrum of the seismic data sub-body through a time-frequency analysis method, and solving a variance attribute to obtain a spectrum variance attribute body; (3) solving an amplitude spectrum of the spectral variance attribute body; and (4) constructing a covariance matrix by using the amplitude spectrum obtained in the step (3), solving a characteristic value of the covariance matrix, calculating a coherence class attribute, solving a coherence class attribute body of the whole data body, and then carrying out fracture identification. The fracture identification method based on spectrum variance coherence provided by the invention can effectively suppress interference caused by inclined formation factors, improves the identification precision of a fracture system, and plays an important guiding role in exploration and development of fault block fractured oil and gas reservoirs.

Description

Technical Field

[0001] The invention belongs to the field of seismic data processing, and in particular relates to a fault identification method based on frequency spectrum variance coherence. Background Art

[0002] The precise prediction and identification of faults and fractures plays an important role in sedimentary phase research, analysis of oil and gas drainage process and reservoir prediction of fault-type reservoirs, and is very important for oil field well deployment and improving recovery. For fault and fracture identification, a variety of technologies have been formed in seismic methods, such as coherence, curvature, edge detection, etc. Many commercial software also include fault and fracture identification modules.

[0003] The Chinese invention patent application with publication number CN104181587A proposes a method for obtaining the coherence value of the amplitude spectrum of seismic data. The patent constructs a covariance matrix by using the amplitude spectrum of the seismic effective frequency band data volume calculated by FFT, and then calculates the characteristic roots of the J-dimensional covariance matrix, where the ratio of the maximum characteristic root to the sum of all characteristic roots is the coherence value of a certain sampling point.

[0004] The Chinese invention patent with announcement number CN108957530B proposes an automatic crack detection method based on seismic coherence volume slices. This patent can directly use seismic coherence volume slice data to extract crack information and realize the azimuth calculation of cracks, greatly improving the efficiency of seismic slice crack interpretation and saving a lot of manpower and material costs.

[0005] The Chinese invention patent with the announcement number CN111239817B proposes a method for improving the accuracy of fault likelihood attribute analysis, which is conducive to highlighting the imaging of fault anomalies, obtaining more accurate underground fault structure information, improving the ability to identify fault structures, and enhancing the seismic imaging effect of faults.

[0006] The Chinese invention patent application with publication number CN113325470A proposes an adaptive multi-directional coherence fusion method, which integrates the advantages of different coherence bodies in describing discontinuous geological features in different directions, solves the shortcoming that traditional coherence cannot fully extract effective information, improves the description ability and application efficiency of the coherence body, and provides a reliable means for understanding underground structures and sedimentary characteristics.

[0007] The Chinese invention patent with the announcement number CN109490952B discloses a seismic coherence volume analysis method and system, which improves the accuracy of seismic coherence volume analysis in revealing small faults and micro-faults by integrating the respective advantages of coherent data volumes with different trace arrangements.

[0008] At present, the main coherence calculation methods are to calculate the cross-correlation between traces, multi-channel similarity, and covariance matrix eigenvalues. The input is the original seismic data, which contains phase information. The phase information is very sensitive to the structural dip. Therefore, the interference caused by the inclined stratum factor can often be seen in the recognition effect diagram of the coherence attribute, which affects the recognition effect of faults and cracks. Summary of the invention

[0009] The purpose of the present invention is to provide a fracture identification method based on spectral variance coherence to solve the problem that the accuracy and precision of fault and crack identification in existing methods need to be improved.

[0010] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0011] A fracture identification method based on spectral variance coherence comprises the following steps:

[0012] (1) Select a data point from the 3D seismic data volume and construct a seismic data subvolume with the data point as the center;

[0013] (2) calculating the amplitude spectrum of the seismic data sub-volume by a time-frequency analysis method, obtaining the variance attribute, and obtaining a spectrum variance attribute volume;

[0014] (3) Obtaining the amplitude spectrum of the spectral variance attribute body;

[0015] (4) Use the amplitude spectrum obtained in step (3) to construct a covariance matrix, obtain the eigenvalues ​​of the covariance matrix, calculate the coherence attributes, and then obtain the coherence attribute body of the entire data body, and then perform fracture identification.

[0016] The fault identification method based on spectrum variance coherence provided by the present invention can effectively suppress the interference caused by the tilted formation factor, improve the fault system identification accuracy, and play an important guiding role in the exploration and development of fault-block fracture type oil and gas reservoirs.

[0017] Preferably, in step (4), different coherent characterizations are used to calculate coherent attributes, and a preferred coherent attribute body is obtained according to the identification result, and fracture identification is performed according to the preferred coherent attribute body. The advantages and disadvantages of several coherent characterization methods can be utilized, and the optimal selection can be made according to the actual data to calculate the coherent attributes, so as to improve the accuracy of fault and fracture identification description.

[0018] In order to better reflect the adaptability of different coherent attributes to actual data, preferably, the different coherent representations include: Among them, λ j represents the jth eigenvalue of the covariance matrix, and λ1, λ2, and λ3 are the first three eigenvalues ​​of the covariance matrix, respectively.

[0019] Preferably, in step (2), the time-frequency analysis method is a fast Fourier transform method. In step (3), the amplitude spectrum of the spectral variance attribute body is obtained using a fast Fourier transform method. The amplitude spectrum of the data is calculated using the fast Fourier transform method (FFT), which ensures the accuracy of the amplitude spectrum while improving its calculation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flow chart of the fracture identification method based on spectrum variance coherence in the present invention;

[0021] Figure 2 The recognition results of different methods of the Marmousi2 model are shown; among them, 2a is the original velocity model diagram, 2b is the recognition result using conventional coherence, 2c is the coherence attribute recognition result based on spectral variance using the C3 coherence expression of the present invention; 2d is the coherence attribute recognition result based on spectral variance using the donias coherence expression of the present invention;

[0022] Figure 3 The identification results of different methods for actual data of a certain work area are shown; among them, 3a is a certain original seismic profile, 3b is the conventional coherence identification result, and 3c is the coherence technology identification result based on spectral variance of the present invention. DETAILED DESCRIPTION

[0023] The fault identification and detection strategy of the present invention is as follows:

[0024] The FFT spectrum analysis method is used to transform seismic data into the time-frequency domain to obtain a time-frequency spectrum, which can separate the influence of phase information and will not be affected by the time delay information corresponding to the inclined structure of the adjacent seismic trace; the variance attribute is sensitive to slight changes in the data, and the difference in amplitude spectrum is expressed, which is more sensitive to fault identification; different coherence representations are used to calculate the coherence-like attributes based on the spectral variance data, and the best identification results are selected to better identify faults.

[0025] The fracture identification method based on spectrum variance coherence of the present invention has a flow chart as shown in FIG. Figure 1 As shown, follow the steps below to implement it:

[0026] Step 1: First, select a data point from the 3D seismic data volume and construct a data sub-volume with the point as the center;

[0027] Step 2: Secondly, the amplitude spectrum of the seismic data subvolume is calculated using the fast Fourier transform (FFT) method;

[0028] Step 3: Utilize the characteristic that the variance attribute is relatively sensitive to the earthquake dislocation response to obtain the variance information of the data sub-body amplitude spectrum;

[0029] Step 4: Perform fast Fourier transform (FFT) again on the calculated amplitude spectrum variance to calculate the amplitude spectrum of the amplitude spectrum variance attribute;

[0030] Step 5: Use the amplitude spectrum of the spectral variance to construct the covariance matrix and calculate the eigenvalues ​​of the matrix;

[0031] Step 6: Calculate the coherence attributes using different coherence representations such as the third-generation eigenvalue coherence and donias coherence;

[0032] Step 7: Finally, iterate to obtain the eigenvalue-related attribute body of the entire data body.

[0033] Through experiments and actual data application research, the present invention can improve the accuracy of fault identification and detection, so it has very important guiding significance for the exploration and development of fault-block oil and gas reservoirs.

[0034] The implementation process of the present invention is described in detail below in conjunction with specific embodiments.

[0035] Example 1

[0036] The fracture identification method based on spectral variance coherence of this embodiment includes the following steps:

[0037] Step 1: Select a seismic data subvolume with a certain number of channels (generally 5-11, the larger the value, the better the accuracy, but the more calculation) centered at the analysis point from the 3D seismic data volume. Take a subvolume containing a data in J (here J=5) channels, marked as j, as an example. The inline direction of the j-th channel data is numbered as x, the crossline direction is numbered as y, the sampling time is t, and the sampling interval is dt. Therefore, each sampling point of the j-th channel data can be defined as s j (x,y,t).

[0038] s j (x,y,t)=(s 1j ,…s nj ,…s Nj ) T ,n=1,2,…N (1)

[0039] Where n represents the sampling point coordinates in the time direction; N represents the total number of sampling points in each channel; and the superscript T represents the transpose of the matrix.

[0040] Step 2: Define Z calculated by time-frequency analysis method (FFT method) j (x,y,t,ω) as s j Amplitude spectrum of (x,y,t).

[0041] Z j (x,y,t,ω)=(z1,ω ,…z n,ω ,…z N,ω ) T ,n=1,2,…N (2)

[0042]

[0043] Where NFT is the number of frequency points.

[0044] Step 3: Calculate Z j The variance of (x,y,t,ω) is defined as v j (x,y,t)

[0045] v j (x,y,t)=(v 1,j ,…v n,j ,…v N,j ) T (4)

[0046]

[0047] In the formula, Represents the average amplitude spectrum of the nth analysis point.

[0048] Step 4: Calculate v from equation (5) j (x, y, t) is transformed by FFT again to calculate its amplitude spectrum and obtain the amplitude spectrum a j (x,y,t,ω) is:

[0049] a j (x,y,t,ω)=(a 1,ω ,…a n,ω ,…a N,ω ) T ,n=1,2,…N (6)

[0050] Step 5: Use the amplitude spectrum of the spectral variance to define the matrix A as follows:

[0051] A(x,y,t)=[a1,…a j ,…a J ] (7)

[0052] The covariance matrix C(x,y,t) is:

[0053] C(x,y,t)=A T A (8)

[0054] Step 6: Obtain the first three eigenvalues ​​of the covariance matrix C(x,y,t), denoted as λ1, λ2, and λ3.

[0055] Step 7: Calculate the coherent class attributes using the coherent class attribute expressions shown in Table 1.

[0056] Table 1 Different expressions of eigenvalue coherence

[0057]

[0058]

[0059] Step 8: According to the degree of suppression of the inclined stratum interference in the recognition results and whether the recognized faults are clear, continuous and reliable, the recognition effects of several coherent attributes are judged, and the attribute with the best effect is selected as the final output result.

[0060] The classic Marmousi2 model with a combination of steeply inclined strata and complex faults in the field of seismic processing and interpretation was selected to represent the data of an actual work area with inclined strata and complex faults to test the recognition effect of the method. Figure 2 The identification results of different methods of the Marmousi2 model are shown; among them, 2a is the original geological model map, 2b is the identification result using conventional C3 coherence (the conventional C3 coherence method is mainly based on amplitude information for calculation. The amplitude value within a certain size window of seismic data is used to construct a covariance matrix, and the maximum eigenvalue information is obtained to represent the similarity of the local area of ​​the data to identify faults. The main steps are: 1) First, select a data point from the three-dimensional seismic data volume, and construct a data sub-volume with this point as the center; 2) Write the data sub-volume in matrix form, and then multiply the matrix with its transposed matrix to generate a covariance matrix; 3) Obtain the eigenvalues ​​of the covariance matrix, and then use the C3 coherence expression to calculate the coherence attributes to obtain the C3 coherence volume. ), 2c is the coherence attribute recognition result based on spectral variance using the C3 coherence expression; 2d is the coherence attribute recognition result based on spectral variance using the donias coherence expression.

[0061] from Figure 2 It can be seen that for the Marmousi2 model, different expressions have different recognition effects and applicability: Figure 2 b The fault layers are all identified, but Figure 2 b The inclined strata in the area pointed by the red arrow on the right are not suppressed; the spectral variance coherence attribute based on the C3 coherence expression Figure 2 In c, the faults are also identified, the inclined strata in the area indicated by the red arrow on the right are suppressed, and the boundaries of the orange-yellow special geological bodies in the geological model are more clearly depicted; the coherence-based attributes based on the spectral variance using the Donias expression Figure 2In d, the faults are also identified, the inclined strata in the area indicated by the red arrow on the right are suppressed, and the boundaries of the orange-yellow special geological bodies in the geological model are relatively clear. In general, all three methods can identify faults, but the present invention has a better suppression effect on relatively gently inclined strata. For special geological bodies, the method used in the present invention has a better effect, among which the better recognition result is the one based on the spectral variance coherence attribute using the C3 coherence expression. Figure 2 c.

[0062] Furthermore, the method of Example 1 is used to identify the fractures in a certain work area. Figure 3 The recognition results of different methods are shown, among which, Figure 3 a is the original seismic section, Figure 3 b is the attribute graph obtained by calculating the C3 coherent attributes, Figure 3 c is the fault identification result of the present invention (preferably the Bakker coherence attribute with the best identification effect is used for calculation).

[0063] Depend on Figure 3 It can be seen that compared with conventional coherence technology, the coherence technology based on spectral variance suppresses some background interference and the influence of inclined strata, and can identify faults and breakpoints that are relatively vague on conventional coherence profiles (the positions indicated by arrows in the figure), which can effectively suppress interference factors and improve the accuracy of fault system identification.

Claims

1. A fracture identification method based on spectral variance coherence, characterized in that: The following steps are involved: (1) Select a data point from the 3D seismic data volume and construct a seismic data subvolume with the data point as the center; (2) calculating the amplitude spectrum of the seismic data sub-volume by a time-frequency analysis method, obtaining the variance attribute, and obtaining a spectrum variance attribute volume; (3) Obtaining the amplitude spectrum of the spectral variance attribute body; (4) Use the amplitude spectrum obtained in step (3) to construct a covariance matrix, obtain the eigenvalues ​​of the covariance matrix, calculate the coherence attributes, and then obtain the coherence attribute body of the entire data body, and then perform fracture identification.

2. The fracture identification method based on spectral variance coherence according to claim 1, characterized in that: In step (4), different coherent representations are used to calculate coherent attributes, and a preferred coherent attribute body is obtained according to the identification result, and fracture identification is performed according to the preferred coherent attribute body.

3. The fracture identification method based on spectral variance coherence according to claim 2, characterized in that: The different coherent representations include: Among them, λ j represents the jth eigenvalue of the covariance matrix, and λ1, λ2, and λ3 are the first three eigenvalues ​​of the covariance matrix, respectively.

4. The fracture identification method based on spectral variance coherence according to any one of claims 1 to 3, characterized in that: In step (2), the time-frequency analysis method is a fast Fourier transform method.

5. The fracture identification method based on spectral variance coherence according to claim 4 is characterized in that: In step (3), the amplitude spectrum of the spectral variance attribute body is obtained using a fast Fourier transform method.

Images