Method for identifying hidden strike-slip fracture in different directions

Abstract

The invention discloses a method for identifying hidden strike-slip fracture in different directions. The method comprises the following steps: acquiring a sub-azimuth data volume of a vertical fracture direction in a research area; based on the sub-azimuth data body, obtaining an effective frequency band range and a main frequency of seismic data of a target stratum through spectral analysis; combining the effective frequency band range and the main frequency, and adopting frequency division filtering to obtain a narrow frequency band seismic data volume with the main frequency as the center; carrying out attribute analysis on the narrow-band seismic data volume to identify strike-slip fracture; identifying the hidden strike-slip fracture of the target horizon by using a deep learning fracture description technology; the problem that the hidden strike-slip fracture is difficult to effectively recognize in the prior art is effectively solved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of oil exploration and development, and particularly to a method for identifying concealed strike-slip faults by direction based on high-quality seismic data. Background Art

[0002] In practical applications, conventional seismic data is generally used to depict and identify strike-slip faults. The commonly used strike-slip fault analysis methods include the following: seismic profile fault structure interpretation, coherent attribute analysis technology, curvature attribute analysis technology, ant body tracking technology, etc. The basic ideas of these methods are similar, and they all analyze the geometric characteristics of strike-slip faults from perspectives such as profiles or planes, including evaluating the continuity of seismic events, the correlation between traces, and the curvature of horizons, etc., to clearly depict the geometric shape of strike-slip faults on seismic profiles.

[0003] The above classical methods have high requirements for the quality of seismic data. If the fault scale is small and its seismic response characteristics are weak, it is difficult to identify strike-slip faults. In practical applications, the strike-slip fault plane is vertical and the fault throw is small, there is no obvious fault throw on the seismic profile, the strike-slip fault characteristics of conventional seismic data are not clear, and the seismic events can only show characteristics such as distortion and flexure. At the same time, the strike-slip fault is buried deep and has strong concealment, the specific position characteristics of the fault plane are not clear, and the fault identification has strong multi-solution. It is difficult to identify concealed strike-slip faults by conventional methods, and there are no obvious characteristics in conventional attributes such as coherence curvature. Therefore, it is difficult to accurately depict and identify concealed strike-slip faults based on conventional methods. Summary of the Invention

[0004] In view of this, the present disclosure provides a method for identifying concealed strike-slip faults by direction, which solves the problem that it is difficult to effectively identify concealed strike-slip faults in the prior art.

[0005] To achieve the above invention purpose, the method for identifying concealed strike-slip faults by direction includes: Obtain a split-azimuth data volume perpendicular to the fault strike in the study area; Based on the split-azimuth data volume, obtain the effective frequency band range and dominant frequency of the seismic data of the target layer through spectral analysis; combine the effective frequency band range and the dominant frequency, and use frequency division filtering to obtain a narrow-band seismic data volume centered on the dominant frequency; Perform attribute analysis on the narrow-band seismic data volume to identify strike-slip faults; Use deep learning fault characterization technology to identify the concealed strike-slip faults of the target horizon.

[0006] In the present disclosure and possible embodiments, the method for obtaining the split-azimuth data volume perpendicular to the fault strike in the study area includes: Obtain the interpreted strike of the backbone faults in the study area, determine the azimuth division scheme of the pre-stack gathers, and obtain the azimuthal data volume perpendicular to the fault strike.

[0007] In the present disclosure and possible embodiments, the method for obtaining the strike of the backbone faults in the study area includes: Obtain the interpreted strike of the backbone faults in the study area from the seismic data, fault data, and fault plane distribution data of the study area.

[0008] In the present disclosure and possible embodiments, the method for determining the azimuth division scheme of the pre-stack gathers includes: Obtain the pre-stack time migration gathers of the target layer in the OVT domain, and in combination with the interpreted strike of the backbone faults in the study area, based on the basic understanding that the azimuthal stacking seismic data perpendicular to the fault strike has the strongest fault identification ability, while considering the signal-to-noise ratio, divide the overall OVT domain data into four azimuth segments of 0° - 45°, 45° - 90°, 90° - 135°, and 135° - 180° to obtain the azimuthal data volume perpendicular to the fault strike in the study area.

[0009] In the present disclosure and possible embodiments, the method for determining the effective frequency band range and main frequency of the seismic data of the target layer includes: Based on the azimuthal data volume, in combination with the principle in signal analysis theory that most of the energy of a signal is concentrated in a relatively narrow frequency band, determine the effective frequency band range and main frequency of the seismic data of the target layer according to the part with the highest frequency of occurrence in the signal.

[0010] In the present disclosure and possible embodiments, the method for performing attribute analysis on the narrow-band seismic data volume to identify strike-slip faults includes: Extract seismic attributes based on the narrow-band seismic data volume, and identify the strike-slip faults through multi-attribute fusion interpretation.

[0011] In the present disclosure and possible embodiments, the seismic attributes include curvature, coherence, and dip.

[0012] In the present disclosure and possible embodiments, based on the narrow-band seismic data volume, use the deep learning fault characterization technology of the CNN network image segmentation algorithm based on Unet to identify the hidden strike-slip faults.

[0013] The method for identifying hidden strike-slip faults in different directions in the present disclosure adopts the method of azimuthal stacking seismic data, highlighting the characteristics of hidden strike-slip faults in the seismic section, and successfully solving the key problem that the strike-slip faults are not clearly characterized in conventional seismic data; at the same time, using the deep learning fault characterization technology to finely identify hidden strike-slip faults; compared with the conventional method, this method has higher characterization ability and identification accuracy, providing reliable support for oil and gas resource evaluation and exploration deployment. Description of the Drawings

[0014] Through the description of the embodiments of the present disclosure with reference to the following drawings, the above and other objects, features, and advantages of the present disclosure will become clearer. In the drawings: Figure 1 is a flowchart of a method for identifying hidden strike-slip faults in different directions according to an embodiment of the present disclosure; Figure 2-1 is a cross-sectional view of the processing results of conventional seismic data according to an embodiment of the present disclosure; Figure 2-2 is a cross-sectional view of the processing results of seismic data based on the vertical fracture azimuth according to an embodiment of the present disclosure; Figure 3-1 is a curvature attribute map of the target layer of conventional seismic data according to an embodiment of the present disclosure; Figure 3-2 is a curvature attribute map of the target layer in the dominant frequency band based on azimuthal stacking in the vertical fracture direction according to an embodiment of the present disclosure; Figure 4-1 is a seismic plan view of the target layer of the deep learning fusion data volume according to an embodiment of the present disclosure; Figure 4-2 is a seismic cross-sectional view of the target layer of the deep learning fusion data volume according to an embodiment of the present disclosure; Figure 5 is a fracture outline map of the target layer according to an embodiment of the present disclosure. Embodiments

[0015] The following describes the present disclosure based on embodiments. However, it should be noted that the present disclosure is not limited to these embodiments. In the following detailed description of the present disclosure, some specific details are described in detail. However, for the parts that are not described in detail, those skilled in the art can also fully understand the present disclosure.

[0016] In addition, those of ordinary skill in the art should understand that the provided drawings are only for illustrating the objects, features, and advantages of the present disclosure, and the drawings are not actually drawn to scale. At the same time, unless the context clearly requires otherwise, the words "including", "comprising", and similar words throughout the specification and claims should be interpreted as having an inclusive meaning rather than an exclusive or exhaustive meaning; that is, it is the meaning of "including but not limited to".

[0017] To make the objects, technical solutions, and advantages of the present invention clearer and more understandable, the following further describes the present invention in detail with reference to the drawings and by way of examples.

[0018] Figure 1 is a flowchart of a method for identifying hidden strike-slip faults in different directions according to an embodiment of the present disclosure; shown by Figure 1 as follows: The method for identifying hidden strike-slip faults in different directions specifically includes the following steps: S1. Collect basic data of the study area, including high-quality seismic data, seismic interpretation horizons, fault data, fracture plane distribution data, and geological overview of the relevant area of the target interval.

[0019] S2. Combine the interpreted strike of the backbone faults in the study area to determine the azimuth division scheme of the pre-stack gathers, and obtain the split-azimuth stacking data volume perpendicular to the fault strike; specifically as follows: 1. Conduct preprocessing work such as denoising and correction on the collected seismic data, and perform time migration processing on the original pre-stack seismic data using the pre-stack migration algorithm to extract the pre-stack time migration gathers of the target interval based on the OVT domain. 2. Divide the faults in the study area into two categories, major faults and minor faults, according to the fault data and fracture plane distribution data collected based on S1, and further screen out the backbone faults from the major faults to obtain the interpreted strike of the backbone faults in the study area. 3. Obtain the pre-stack time migration gathers of the target interval in the study area based on the OVT (Offset Vector Tile) domain. Then, combine the interpreted strike of the backbone faults in the study area. Finally, based on the basic understanding that the azimuth stacking seismic data perpendicular to the fault strike has the strongest fault recognition ability, while considering the signal-to-noise ratio, divide the overall OVT domain data into four azimuth segments of 0° - 45°, 45° - 90°, 90° - 135°, and 135° - 180°. Compare the ability of the data volumes in different azimuth segments to depict strike-slip faults, and finally select the split-azimuth data volume perpendicular to the fault direction with the strongest fault depiction ability.

[0020] The azimuth of the backbone faults is 90° - 135°, so the azimuth perpendicular to the backbone faults is selected as 0° - 45°.

[0021] In this embodiment, split-azimuth stacking in the direction perpendicular to the faults is carried out for the seismic data gathers, and the split-azimuth stacking data is compared with the conventional full-azimuth stacking data, as Figure 2-1 and Figure 2-2 shown.

[0022] S3. Based on the split-azimuth data volume, determine the main frequency of the seismic data of the target interval in combination with the spectrum analysis results, and use frequency division filtering to obtain narrow-band seismic data centered on the main frequency to further enhance the structural change of the in-phase axis; specifically as follows: Perform spectral analysis on the azimuthal data volume in the vertical fracture direction extracted in step S2. Combining with the theory that most of the energy of the signal in signal analysis is concentrated in a relatively narrow frequency band, find the part with the highest amplitude occurrence frequency in the azimuthal data volume to determine that the main frequency of the seismic data of the target layer is 40 Hz, and the effective frequency band range is 4 - 75 Hz. Secondly, to highlight the amplitude attributes within the effective frequency band range, determine the cut-off frequency of 30 - 50 Hz, and perform frequency division filtering on the seismic data to extract the narrow-band high-amplitude seismic data centered on the main frequency, highlighting the seismic identification characteristics of the concealed strike-slip fault.

[0023] S4. Based on the extracted narrow-band seismic data volume, conduct conventional attribute analysis to identify large-scale strike-slip faults; specifically as follows: Based on the extracted narrow-band seismic data volume, perform smoothing processing to reduce the influence of noise. Then, extract various conventional seismic attributes such as curvature, coherence, and dip angle respectively: Calculate the first and second gradients of the data volume through the finite difference method to obtain the curvature attribute of the seismic data volume; Calculate the cross-correlation function of the data volume, and use the cross-correlation function to calculate the coherence spectrum. Then, perform threshold processing on the coherence spectrum to highlight the frequency components with stronger coherence to obtain the coherence attribute of the seismic data volume; Use the reflected waves in the seismic data and detect and extract the formation reflection interface through methods such as threshold and phase. After determining the formation position in combination with the formation interpretation data, calculate the slope between adjacent reflection interfaces to obtain the bottom dip angle attribute of the seismic data volume. Finally, conduct multi-attribute fusion interpretation to more accurately identify large-scale strike-slip faults.

[0024] In this embodiment, aiming at the curvature attribute of the target layer of the extracted azimuthal stacked narrow-band data and comparing it with the curvature attribute of the target layer extracted from the conventional seismic data, further verify the recognition accuracy of small-offset strike-slip faults, as Figure 3-1 and Figure 3-2 shown.

[0025] S5. Use the deep learning fracture characterization technology to accurately identify the concealed strike-slip faults and obtain the fracture distribution of the target horizon; specifically as follows: Construct a Unet model, select the cross-entropy loss function and the stochastic gradient descent optimizer, use the narrow-band seismic data volume extracted in S3 to train the Unet model for fracture characterization and recognition, and verify and adjust the results of the trained model. Repeatedly measure the matching degree between the generated results of the model and the actual labels to complete the work of accurately identifying the concealed strike-slip faults using the deep learning fracture characterization technology based on the CNN network of the Unet model and obtain the fracture distribution of the target horizon.

[0026] In this embodiment, based on the deep learning fracture characterization technology, give Figure 4-1 and Figure 4-2The seismic plan view and cross-section view of the target layer of the deep learning fusion data volume shown can accurately identify the hidden strike-slip faults in the target layer. And draw Figure 5 The fault outline map of the target layer shown to complete the comprehensive identification of hidden strike-slip faults.

[0027] The above embodiments are only for expressing the implementation modes of the present disclosure, and the description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the present disclosure. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present disclosure, several deformations, equivalent replacements, improvements, etc. can be made, and these all belong to the protection scope of the present disclosure. Therefore, the protection scope of the patent of the present disclosure should be subject to the appended claims.

Claims

1. A method for identifying concealed strike-slip faults in different directions, characterized in that, Including: Obtaining a directional data volume of the vertical fracture strike in the study area; Based on the directional data volume, obtaining the effective frequency band range and dominant frequency of the seismic data of the target layer through spectral analysis; combining the effective frequency band range and the dominant frequency, and using frequency division filtering to obtain a narrow frequency band seismic data volume centered on the dominant frequency; Performing attribute analysis on the narrow frequency band seismic data volume to identify strike-slip faults; Using deep learning fracture characterization technology to identify the hidden strike-slip faults at the target horizon.

2. The method for identifying concealed strike-slip faults in different directions according to claim 1, characterized in that, The method for obtaining the directional data volume of the vertical fracture strike in the study area includes: Obtaining the interpreted strike of the backbone faults in the study area, determining the azimuth angle division scheme of the prestack gather, and obtaining the directional data volume of the vertical fracture strike.

3. The method for identifying concealed strike-slip faults in different directions according to claim 2, characterized in that, The method for obtaining the strike of the backbone faults in the study area includes: Obtaining the interpreted strike of the backbone faults in the study area from the seismic data, fault data, and fracture plane distribution data in the study area.

4. The method for identifying concealed strike-slip faults in different directions according to claim 3, characterized in that, The method for determining the azimuth angle division scheme of the prestack gather includes: Obtaining the prestack time migration gather of the target layer based on the OVT domain, combining the interpreted strike of the backbone faults in the study area, based on the basic understanding that the azimuthal stacked seismic data perpendicular to the fracture strike has the strongest fracture identification ability, and while considering the signal-to-noise ratio, dividing the overall OVT domain data into four azimuth segments of 0° - 45°, 45° - 90°, 90° - 135°, and 135° - 180° to obtain the directional data volume of the vertical fracture strike in the study area.

5. The method for identifying concealed strike-slip faults in different directions according to any one of claims 1-4, characterized in that, The method for determining the effective frequency band range and dominant frequency of the seismic data of the target layer includes: Based on the directional data volume, combining the principle in signal analysis theory that most of the energy of a signal is concentrated in a relatively narrow frequency band, and determining the effective frequency band range and dominant frequency of the seismic data of the target layer according to the part with the highest frequency appearance in the signal.

6. The method for identifying concealed strike-slip faults in different directions according to claim 5, characterized in that, The method for performing attribute analysis on the narrow frequency band seismic data volume to identify strike-slip faults includes: Extracting seismic attributes based on the narrow frequency band seismic data volume, and identifying the strike-slip faults through multi-attribute fusion interpretation.

7. The method for identifying concealed strike-slip faults in different directions according to claim 6, characterized in that: The seismic attributes include curvature, coherence, and dip angle.

8. The method for identifying concealed strike-slip faults in different directions according to any one of claims 1-4, 6 or 7, characterized in that: Based on the narrow frequency band seismic data volume, using deep learning fracture characterization technology based on the Unet-based CNN network image segmentation algorithm to identify the hidden strike-slip faults.

Images