Fracture identification description method and device based on frequency and maximum likelihood attribute, and computer program product
By using frequency-division phase analysis and maximum likelihood volume calculation, fractures of different scales are identified, overcoming the shortcomings of coherence volume technology in identifying micro-faults and achieving higher-precision fracture characterization to meet the needs of oil and gas reservoir development.
Patent Information
- Application Number
- CN202410707293.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-03
- Publication Date
- 2025-12-05
AI Technical Summary
In existing technologies, coherence volume technology has difficulty effectively identifying micro-faults below the resolution of seismic data and is sensitive to noise, leading to false fault identification and affecting the accuracy of oil and gas reservoir development.
Seismic data is processed using frequency-division phase analysis technology to generate multiple frequency-division data volumes. Through filtering and maximum likelihood calculation, combined with the fault plane and dip direction scales, faults of different scales are identified.
It improves the clarity of fault planes and the resolution of fracture identification, meets the accuracy requirements of hydrocarbon accumulation demonstration, and improves work efficiency.
Smart Images
Figure CN121069479A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of oil reservoir development, in particular to a fracture identification and characterization method based on frequency and maximum likelihood attribute, a device and a computer program product. BACKGROUND
[0002] Fracture is a geological structure that has important influence on the generation, migration, accumulation and preservation of oil and gas. For example, fracture can serve as a channel for oil and gas to migrate from source rock to reservoir. Driven by high pressure difference underground, oil and gas migrates upward through the fractures and pores of the fracture and finally accumulates in the reservoir to form an oil and gas reservoir. Fracture activity can cause damage to the oil and gas reservoir, such as displacement of the fault which can cut off the oil and gas reservoir, resulting in oil and gas leakage.
[0003] The characteristics of the fracture system directly affect the enrichment and distribution of oil and gas, so in the development process of oil reservoir resources, it is often necessary to finely characterize the fracture. In the prior art, the fracture is usually identified by using the coherence volume technology. The coherence volume technology determines the fault in the stratum by analyzing the similarity of adjacent seismic trace signals in the three-dimensional seismic data volume. However, the coherence volume technology has poor identification effect for small-scale faults, especially for micro-faults below the resolution of seismic data. Moreover, the coherence volume technology is sensitive to noise, and random noise in seismic data can affect the calculation result of the coherence volume, resulting in the identification of false faults. SUMMARY
[0004] The purpose of the embodiments of the present application is to provide a fracture identification and characterization method based on frequency and maximum likelihood attribute, a device and a computer program product, which can solve or at least partially solve the above-mentioned defects of the prior art.
[0005] In order to achieve the above-mentioned purpose, the first aspect of the embodiments of the present application provides a fracture identification and characterization method based on frequency and maximum likelihood attribute, which comprises:
[0006] processing the seismic data corresponding to the target area by using the frequency division phase analysis technology to generate a plurality of frequency division data volumes;
[0007] filtering the plurality of frequency division data volumes to generate a corresponding filter volume;
[0008] calculating a maximum likelihood volume according to the filter volume, and dividing the maximum likelihood volume based on the fault plane direction scale and the fault dip direction scale to generate a plurality of fracture data volumes; and
[0009] determining the target fracture area based on the fracture data volume.
[0010] Optionally, the plurality of frequency division data volumes comprise seismic data volumes corresponding to different frequencies.
[0011] Optionally, the seismic data bodies corresponding to different frequencies are used to identify different sizes of fault surfaces, wherein the seismic data body corresponding to a low frequency is used to identify large-scale fault surfaces;
[0012] the seismic data body corresponding to a medium frequency is used to identify large-scale fault surfaces with a large fault throw;
[0013] the seismic data body corresponding to a high frequency is used to identify small-scale fault surfaces.
[0014] Optionally, the processing of the seismic data corresponding to the target region by using the frequency division phase analysis technology to generate a plurality of frequency division data bodies comprises:
[0015] The seismic data in a time window is converted into a frequency domain by discrete Fourier transform or maximum entropy spectrum estimation, and after the influence of a wavelet is removed, a frequency-phase spectrum is used to identify the boundary of a discontinuous geological body.
[0016] Optionally, the filtering processing of the plurality of frequency division data bodies to generate corresponding filter bodies comprises:
[0017] The frequency division data bodies are subjected to structure-oriented filtering processing to generate corresponding filter bodies.
[0018] Optionally, the filtering processing of the plurality of frequency division data bodies to generate corresponding filter bodies comprises:
[0019] The frequency division data bodies are subjected to average processing by using a nonlinear filtering algorithm.
[0020] Optionally, the calculation of the maximum likelihood body according to the filter bodies comprises:
[0021] first iteration operation is performed on the filter bodies to generate a first intermediate result; and
[0022] second iteration operation is performed based on the first intermediate result to generate a maximum likelihood body.
[0023] The second aspect of the embodiment of the present application provides a device for identifying a fault, the device comprising a processor, a memory, and a program stored on the memory and executable on the processor, and the processor executes the program to implement the fault identification and characterization method based on frequency and maximum likelihood attribute.
[0024] The third aspect of the embodiment of the present application provides a computer program product comprising a computer program, and the computer program implements the fault identification and characterization method based on frequency and maximum likelihood attribute when executed by a processor.
[0025] By the technical scheme, the fracture identification and delineation method based on frequency and maximum likelihood attribute provided by the embodiment of the application firstly processes seismic data corresponding to a target region by using frequency division phase analysis technology, then further smoothes the data by filtering processing, then calculates a maximum likelihood volume, divides the maximum likelihood volume based on a fault plane direction scale and a fault dip direction scale, and finally determines a target fracture region according to the division result, which on the one hand improves the clarity of the identified fault plane and the resolution of the fracture and other discontinuities, and on the other hand meets the accuracy requirement of fracture delineation in oil and gas accumulation demonstration, and the method technical process is easy to implement in the industry, and the work efficiency can be improved.
[0026] Other features and advantages of the present application will be further illustrated in the following detailed description of embodiments. BRIEF DESCRIPTION OF DRAWINGS
[0027] The accompanying drawings are included to provide a further understanding of the embodiments of the application, and constitute a part of the specification, and are used together with the following detailed description of the embodiments to explain the embodiments of the application, but do not constitute a limitation of the embodiments of the application. In the drawings:
[0028] Figure 1 is a flowchart of the fracture identification and delineation method based on frequency and maximum likelihood attribute provided by the embodiment of the application;
[0029] Figure 2 is a profile graph of a low-frequency data volume;
[0030] Figure 3 is a profile graph of a medium-frequency data volume;
[0031] Figure 4 is a profile graph of a high-frequency data volume;
[0032] Figure 5a 、 Figure 5b 、 Figure 5c is a calculation process diagram of a fault likelihood provided by the embodiment of the application;
[0033] Figure 6a 、 Figure 6b is a fracture identification effect diagram in a fracture data volume generated after different fault plane direction scale parameters and different fault dip direction scale parameters divide the maximum likelihood volume;
[0034] Figure 7 is a 20Hz frequency division volume calculating a 3200ms maximum likelihood volume slice;
[0035] Figure 8 is a 40Hz frequency division volume calculating a 3200ms maximum likelihood volume slice;
[0036] Figure 9It is a 60Hz frequency divider that calculates the maximum likelihood slice over 3200ms. Detailed Implementation
[0037] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.
[0038] Figure 1 This is a flowchart illustrating the fracture identification and characterization method based on frequency and maximum likelihood attributes provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:
[0039] Step S101: Process the seismic data corresponding to the target area using frequency division phase analysis technology to generate multiple frequency division data volumes.
[0040] Specifically, seismic data typically refers to the large amount of seismic waveform data collected during seismic exploration. After processing and interpretation, this data can be used to infer the underground geological structure and reservoir characteristics. In seismic exploration, seismic sources and receivers are deployed on the surface or underwater to record the waveforms of seismic waves propagating underground. This waveform data is converted into a digital format, forming a three-dimensional data volume containing rich information about the underground structure.
[0041] Furthermore, frequency-division phase analysis technology analyzes faults based on the frequency domain, enabling detailed imaging of discontinuous seismic bodies. On one hand, frequency-division phase analysis utilizes the most advantageous frequency band of fault response to highlight different fault orders, i.e., searching for the optimal frequency band that reveals faults of varying order. On the other hand, it leverages the sensitivity of phase technology to changes in stratigraphic attitude, overcoming the limitations of traditional coherent analysis in identifying faults with small vertical displacements.
[0042] In some embodiments, the seismic data is post-stack data. Post-stack data refers to seismic data obtained after stacking processing in the seismic data processing flow. In seismic exploration, in order to improve the signal-to-noise ratio and resolution, seismic records from multiple seismic traces at common shot points are stacked to form a post-stack data volume.
[0043] In some embodiments, the plurality of frequency-divided data volumes include seismic data volumes corresponding to different frequencies. Specifically, after frequency division, the seismic data corresponding to the target area will generate seismic data subsets of different frequencies, which are generally referred to as frequency-divided data or frequency band data. Through frequency division processing, different frequency components in the seismic data can be separated, and each frequency component can then be finely processed and analyzed, thereby improving the resolution of seismic data and enhancing the ability to identify small-scale geological bodies.
[0044] Furthermore,Figure 2 This is a cross-sectional view of the low-frequency data volume, see reference. Figure 2 Low-frequency seismic data volumes are used to identify large-scale fault sections. Figure 3 This is a cross-sectional view of the intermediate frequency data volume, see reference. Figure 3 The seismic data volume corresponding to the intermediate frequency (intermediate frequency data volume) is used to identify large-scale fault sections with large fault displacement. Figure 4 This is a cross-sectional view of the high-frequency data volume, see reference. Figure 4 High-frequency seismic data volumes (high-frequency data volumes) are used to identify small-scale fault sections. Specifically, the imaging characteristics of the same fault differ on high, medium, and low-frequency profiles, and also differ from full-frequency data. Low-frequency data volumes ( Figure 2 The overall energy is relatively strong, the seismic axis is relatively continuous, and large fault sections are relatively easy to identify. The mid-frequency data volume ( Figure 3 The energy difference between deep and shallow fractures is significant, and large-scale fracture surfaces with large displacements are relatively clear and easy to identify. (High-frequency data volume) Figure 4 The energy difference between the deep and shallow parts is also quite large. The amplitude of the deep part is significantly weaker, but the fracture surface of the shallow part is clear, and small fractures are relatively easy to identify.
[0045] Furthermore, phase analysis can be performed on subsets of seismic data at different frequencies to calculate the corresponding phase angle, phase difference, and other parameters. Based on the calculated phase angle and phase difference parameters, the phase change of seismic waves propagating in the underground medium can be determined, and the structural characteristics of the target area can be interpreted.
[0046] In some embodiments, processing seismic data corresponding to the target area using frequency-division phase analysis technology to generate multiple frequency-division data volumes includes: converting seismic data within a short time window into the frequency domain through discrete Fourier transform or maximum entropy spectrum estimation; removing the influence of wavelet; and using the frequency-phase spectrum to identify the boundaries of discontinuous geological bodies. Its advantage lies in using seismic data and Fourier transform to perform fine imaging and mapping of discontinuous geological bodies, solving problems that cannot be solved in the time domain. This technique utilizes short-time-window discrete Fourier transform to convert seismic data from the time domain to the frequency domain, calculates phase slices of the target layer, and uses these phase slices to observe the longitudinal and transverse variation characteristics of faults.
[0047] A time window refers to a specific interval selected in the time dimension for analyzing a specific portion of a seismic signal. Time windows are widely used in local analysis, signal enhancement, amplitude variation analysis, velocity analysis, imaging, data processing, and geological interpretation of seismic data.
[0048] In some embodiments, coherence analysis (frequency-division coherence) can also be performed on the basis of the frequency-division phase volume, and coherence slices can be used to identify fault and fracture combination features.
[0049] Step S102: Filter the multiple frequency division data volumes to generate corresponding filter volumes.
[0050] In some embodiments, the step of filtering the plurality of frequency-division data volumes to generate corresponding filter volumes includes: performing construction-guided filtering on the frequency-division data volumes to generate corresponding filter volumes.
[0051] Structure-guided filtering is based on the geological structural features of seismic data and employs an anisotropic diffusion smoothing algorithm. This means that smoothing is only applied to information parallel to the seismic phase axis, while information perpendicular to the phase axis remains unsmoothed. If a lateral discontinuity is detected in the seismic phase axis, smoothing is not performed; that is, the smoothing operation does not extend beyond the seismic reflection termination (fault and lithological boundary). Therefore, this filtering method preserves fault and lithological boundary information. After structure-guided filtering, discontinuous reflections in seismic data become stable, forming continuous and traceable phase axes, while the reflection termination patterns at faults are preserved.
[0052] It should be noted that guided filtering or other filtering methods can be used alone to filter the frequency division data. Alternatively, guided filtering and other filtering methods can be used in combination to filter the frequency division data, with multiple filtering techniques complementing each other.
[0053] In some embodiments, filtering the plurality of frequency-divided data volumes to generate corresponding filtered volumes includes: averaging the frequency-divided data volumes using a nonlinear filtering algorithm. Specifically, the nonlinear filtering algorithm is a standard edge detection filter related to graphics processing, characterized by discarding a portion of data considered outliers when calculating the average value, even if the data is clean. Among these, the nonlinear filtering algorithm includes "Weighted majority with minimum range" (WMMR).
[0054] In some embodiments, the averaging effect can also be simulated by designing specific nonlinear filters.
[0055] Step S103: Calculate the maximum likelihood volume based on the filter volume, and divide the maximum likelihood volume based on the fault plane direction scale and the fault dip angle direction scale to generate multiple fracture data volumes.
[0056] The maximum likelihood volume refers to the three-dimensional data volume extracted from seismic data using the maximum likelihood estimation method. Maximum likelihood estimation is a statistical method used to estimate probabilistic model parameters given observational data. Maximum likelihood estimation can be used to estimate various seismic attributes, such as seismic wave velocity, attenuation, and lithology; the maximum likelihood volume can represent the three-dimensional distribution of these parameters.
[0057] In some embodiments, calculating the maximum likelihood based on the filter volume includes:
[0058] Perform a first iterative operation on the filter body to generate a first intermediate result; and
[0059] Based on the first intermediate result, a second iteration is performed to generate the maximum likelihood volume. Multiple iterations can make the discontinuities of the fractures on the maximum likelihood volume more prominent.
[0060] Furthermore, the maximum likelihood volume can be calculated using fault likelihood seismic attribute detection. The standard procedure for fault likelihood attribute analysis is based on conventional 3D seismic data, first calculating the similarity attribute `semblance`, using the following formula:
[0061]
[0062] Where: image represents 2D or 3D image sample values (i.e., image samples for visualizing seismic data volumes); s This represents a structure-oriented average of 2D or 3D image sample values. However, when the numerator and denominator are small, the similarity ratio calculated in this way will vary greatly, so the numerator and denominator should be smoothed before calculating the similarity.
[0063] The expression for calculating the fault likelihood volume properties is:
[0064] f = 1 - semblance8
[0065] Where f represents the fracture likelihood property.
[0066] The maximum likelihood algorithm uses a similarity-based oriented average window. Its principle is to calculate similarity within a narrow oriented window and hold the result where the similarity is minimized. For a set of discrete dip directions, the fault likelihood is calculated within a window covering a certain azimuth range, referring to... Figure 5a , Figure 5b as well as Figure 5c and in Figure 5c Find the maximum likelihood among them. Figure 5a The blue line represents faults, and the black dots indicate the locations for calculating likelihood. Figure 5b This indicates that the likelihood is calculated within a window covering a certain orientation range.Figure 5c This represents the direction with the minimum similarity that is finally determined. This calculation is performed for each point in the image, and the maximum likelihood and the direction of the maximum likelihood are saved. In 3D, instead of calculating on a 2D rectangular shape, an azimuth range is tested for each point; it is now a 3D rectangular volume oriented with various tilt and azimuth angles.
[0067] In some embodiments, the maximum likelihood volume can be divided according to different fault plane orientation scales and different fault dip angle orientation scales to generate multiple fracture data volumes. The fracture data volumes are then visualized, and the fracture data volume with the best fracture identification effect is selected. The target fracture region is then determined using this fracture data volume.
[0068] Figure 6a , Figure 6b This is a fracture identification effect diagram in the generated fracture data volume after dividing the fault into maximum likelihood volumes with different fault plane orientation scale parameters and different fault dip orientation scale parameters. (Refer to...) Figure 6a , Figure 6b After comparison, it was found that the fault plane orientation scale parameter of 32 and the fault dip angle orientation scale parameter of 8 had the best fault identification effect.
[0069] Step S104: Determine the target fracture region based on the fracture data volume.
[0070] Specifically, Figure 7 The attribute map obtained by slicing the seismic data volume after filtering and calculating the maximum likelihood volume at a frequency of 20Hz clearly shows the direction of the main fractures and faults. Although some details such as small fractures are lost, the noise is effectively suppressed. For cases where the requirements for detail are not very high, the attribute map is clear, easy to read and understand at a glance.
[0071] Figure 8 The image is an attribute map obtained by slicing the maximum likelihood volume after filtering the seismic data volume at a frequency of 40Hz. The image shows that it is rich in detail, and while suppressing noise, it also highlights the fine cracks. Larger fractures are displayed very clearly, which is very helpful in interpreting large-scale fractures.
[0072] Figure 9 This is an attribute map obtained after filtering and calculating the maximum likelihood volume of seismic data at a frequency of 60Hz. The map shows that small geological details and structures are well represented, with clearer and richer details of fractures. The structural features are relatively well-represented, capable of characterizing the size, length, and other geometric features of fractures, as well as their developmental degree, which is beneficial for the precise identification and interpretation of structures. However, there is also a significant increase in noise.
[0073] The comparison shows that the coherence attributes of the seismic data volumes at several typical frequencies, after frequency division, exhibit distinct characteristics and emphases. Combining these attributes allows for multi-scale analysis and interpretation of seismic data. The structural features that easily identify and pinpoint target layers retain rich structural details required at smaller scales. This not only reduces interference but also makes structural features more apparent, characterizing the size, length, and other geometric features of faults and fractures, as well as their developmental degree, thus facilitating precise structural identification and interpretation.
[0074] This invention also provides a device for identifying fractures, the device including a processor, a memory, and a program stored in the memory and executable on the processor, wherein the processor executes the above-described fracture identification and characterization method based on frequency and maximum likelihood attributes when executing the program.
[0075] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described fracture identification and characterization method based on frequency and maximum likelihood attributes.
[0076] The frequency- and maximum likelihood-based fracture identification and characterization method provided in this invention first processes the seismic data corresponding to the target area using frequency-division phase analysis technology, then further smooths the data through filtering, calculates the maximum likelihood volume, and divides the maximum likelihood volume based on the fault plane direction scale and the fault dip direction scale. Finally, the target fracture area is determined based on the division results. On the one hand, this method improves the clarity of the identified fault plane and the resolution of fractures and other discontinuities; on the other hand, it meets the accuracy requirements for fracture characterization in hydrocarbon accumulation demonstration. Moreover, the technical process of this method is easy to implement in industry and can improve work efficiency.
[0077] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. A method for fracture identification characterization based on frequency and maximum likelihood properties, characterized by, The method comprises: processing seismic data corresponding to a target region by using a frequency division phase analysis technique to generate a plurality of frequency division data volumes; filtering the plurality of frequency division data volumes to generate corresponding filter volumes; calculating a maximum likelihood volume according to the filter volumes, and dividing the maximum likelihood volume based on a fault plane direction scale and a fault dip direction scale to generate a plurality of fault data volumes; and determining a target fault region based on the fault data volumes.
2. The method of claim 1, wherein, The plurality of frequency division data volumes comprise seismic data volumes corresponding to different frequencies.
3. The method of claim 2, wherein, The seismic data volumes corresponding to different frequencies are used to identify fault surfaces of different scales, wherein a low-frequency corresponding seismic data volume is used to identify large-scale fault surfaces; a medium-frequency corresponding seismic data volume is used to identify large-scale fault surfaces with large fault throws; a high-frequency corresponding seismic data volume is used to identify small-scale fault surfaces.
4. The method of claim 1, wherein, The processing of seismic data corresponding to a target region by using a frequency division phase analysis technique to generate a plurality of frequency division data volumes comprises: converting seismic data in a time window to a frequency domain by discrete Fourier transform or maximum entropy spectrum estimation, removing the influence of a wavelet, and then using a frequency-phase spectrum to identify the boundaries of discontinuous geological bodies.
5. The method of claim 1, wherein, The filtering of the plurality of frequency division data volumes to generate corresponding filter volumes comprises: performing structure-oriented filtering processing on the frequency division data volumes to generate corresponding filter volumes.
6. The method of claim 1, wherein, The filtering of the plurality of frequency division data volumes to generate corresponding filter volumes comprises: performing average processing on the frequency division data volumes by using a nonlinear filtering algorithm.
7. The method of claim 1, wherein, The calculation of a maximum likelihood volume according to the filter volumes comprises: performing a first iteration operation on the filter volumes to generate a first intermediate result; and performing a second iteration operation based on the first intermediate result to generate a maximum likelihood volume.
8. An apparatus for identifying a break, comprising: The device comprises a processor, a memory, and a program stored on the memory and executable on the processor, and the processor executes the program to implement the fault identification and characterization method based on frequency and maximum likelihood attributes according to any one of claims 1 to 7.
9. A computer program product comprising a computer program, characterized in that, The computer program, when executed by the processor, implements the fault identification and characterization method based on frequency and maximum likelihood attributes according to any one of claims 1 to 7.