Breakjoint type dessert reservoir prediction method and device based on three-dimensional seismic data

Abstract

The invention provides a broken joint type dessert reservoir prediction method and device based on three-dimensional seismic data, and belongs to the technical field of oil and gas geophysical exploration. The method comprises the following steps: denoising three-dimensional seismic data to obtain denoised three-dimensional seismic data, and extracting fault data from the denoised three-dimensional seismic data; then further extracting breaking joint body data from the fault data, and performing morphological expansion operation on the breaking joint body data to generate breaking joint zone data; performing three-dimensional seismic porosity parameter simulation on the three-dimensional seismic data to obtain a porosity data volume; and finally, carrying out conditional random field attribute fusion on the breaking joint zone data and the porosity data volume to obtain a breaking joint type dessert reservoir data volume, and carrying out data visualization on the data volume to obtain a breaking joint type dessert reservoir image. According to the method, the geologic features of the reservoir can be more comprehensively explained by fusing the breaking joint zone and the porosity data, and the sweet spot area in the reservoir can be identified.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas geophysical exploration, and in particular to a method and device for predicting fracture-type sweet spot reservoirs based on three-dimensional seismic data. Background Art

[0002] In tight oil and gas reservoirs, matrix reservoirs often have lower porosity and permeability, so the exploration and development benefits are low. Fault-type sweet spot reservoirs refer to reservoir areas with superior storage and permeability properties formed on the basis of matrix reservoirs through the transformation of faults and fractures. The detailed characterization of fault-type sweet spot reservoirs is of great significance for tight oil and gas reservoirs.

[0003] At present, a series of studies have been conducted on the prediction of fault sweet spot reservoirs, forming fracture detection methods based on coherence volume and gradient structure tensor algorithm, fracture prediction technology methods based on seismic rock physics drive, and shale oil reservoir sweet spot prediction methods based on logging data. These methods have solved the problem of fault sweet spot reservoir prediction to a certain extent.

[0004] However, there are still some deficiencies in the prediction of fault-type sweet spot reservoirs: on the one hand, traditional fracture prediction technology can only provide linear fault and fracture information, which limits the understanding of the complexity of the fracture network, especially in areas with highly developed fracture systems such as fault zones, where linear information is difficult to reflect the true distribution and connectivity of fractures. On the other hand, the commonly used method for merging faults and matrix reservoirs is mainly the weighted fusion method, which may be too simple and cannot fully reflect the complex relationship between fractures, faults and porosity. It is impossible to reflect the connectivity of the fracture network and the complexity of the spatial configuration of fractures and porosity.

[0005] Therefore, there is an urgent need for a fault-type sweet spot reservoir prediction method based on three-dimensional seismic data. Summary of the invention

[0006] In view of this, the present invention provides a method and device for predicting fault-type sweet spot reservoirs based on three-dimensional seismic data, so as to solve the problem that the sweet spot reservoir prediction method is difficult to fully reflect the complex relationship between cracks, faults and porosity, and the fault zone structure is not accurately portrayed.

[0007] The technical solution adopted by the present invention is: In a first aspect, the present invention provides a method for predicting a fracture-type sweet spot reservoir based on three-dimensional seismic data, comprising: The 3D seismic data is denoised using the dip-guided coherence enhancement denoising method to obtain denoised 3D seismic data. The key eigenvalues ​​in the seismic waveform are extracted from the denoised 3D seismic data using the intrinsic coherence technology, and the faults of the strata are identified based on the key eigenvalues ​​to obtain the fault data. Performing approximate derivative calculation on the fault data, and obtaining corresponding fault volume data according to the approximate derivative calculation result; Perform morphological expansion operation on the fracture volume data to generate fracture zone data; The seismic waveform indication inversion technology is used to simulate the 3D seismic porosity parameters of the 3D seismic data to obtain the porosity data volume; The fault zone data and porosity data are fused with conditional random field attributes to obtain the fault sweet spot reservoir data volume; The data volume of the fault-type sweet spot reservoir is input into the preset three-dimensional visualization model for data visualization conversion to obtain the image of the fault-type sweet spot reservoir.

[0008] Furthermore, the performing of morphological expansion operation on the fracture volume data to generate fracture zone data includes: Based on the fracture image in the fracture data, define the input fracture image and structural elements ; Represents the input image The pixel value in is the sampling point of the fracture body. and Represents the input image The horizontal and vertical coordinates of the pixels in ; Represents structural elements The pixel value in and Represents the structural elements The horizontal and vertical coordinates of the pixels in ; The shape of the structural element is designed to be circular, and the direction of the structural element extending outward is perpendicular to the fracture body direction. Dilate and fill, for each pixel coordinate The pixel value at which the morphological dilation operation is calculated :

[0009] in, Represents structural elements Pixel coordinates in ; Repeat the fracture image The expansion operation is performed until the expanded image meets the preset image standard, and the corresponding fracture zone data is obtained.

[0010] Furthermore, the three-dimensional seismic porosity parameter simulation is performed on the three-dimensional seismic data using the seismic waveform indication inversion technology to obtain the porosity data volume, including: Using the singular value decomposition method, an efficient dynamic clustering analysis of the waveform of the wellbore seismic trace is performed on the 3D seismic data to establish a mapping relationship between the seismic waveform structure and the porosity curve structure, and generate a set of porosity curve samples representing different seismic facies types. By analyzing the distribution of porosity curve sample sets of different seismic facies types, corresponding Bayesian inversion frameworks are established respectively; under each Bayesian framework, the common part of the sample set is selected as the initial model, and iterative simulation is performed to obtain a high-resolution porosity data volume.

[0011] Furthermore, the conditional random field attribute fusion is performed on the fault zone data and the porosity data volume to obtain the fault sweet spot reservoir data volume, including: A conditional random field model is defined, which takes the geological characteristics of the fault zone and the reservoir parameters of the porosity data volume as input to predict the reservoir characteristics of each pixel or sampling point; Selecting reservoir-related features from the fault zone data and porosity data volume to construct a potential function to quantify the value of a single variable or the relationship between variables; The maximum likelihood estimation algorithm is used to learn the parameters of the potential function. The learned parameters are used to control the conditional random field model to perform parameter fusion inference and prediction on the input fault zone data and porosity data volume to generate the corresponding fault-type sweet spot reservoir data volume.

[0012] In a second aspect, the present invention provides a fracture-type sweet spot reservoir prediction device based on three-dimensional seismic data, comprising: A data denoising module is used to denoise the three-dimensional seismic data using a dip-guided coherence enhancement denoising method to obtain denoised three-dimensional seismic data; The fault identification module is used to extract key characteristic values ​​in the seismic waveform from the denoised three-dimensional seismic data using the intrinsic coherence volume technology, and to identify faults in the strata according to the key characteristic values ​​to obtain fault data; A fracture body extraction module, used for performing approximate derivative calculation on the fault data, and obtaining corresponding fracture body data according to the approximate derivative calculation result; A fracture zone generation module is used to perform morphological expansion operations on fracture volume data to generate fracture zone data; A seismic inversion module is used to simulate the three-dimensional seismic porosity parameters of three-dimensional seismic data using seismic waveform indication inversion technology to obtain a porosity data volume; The attribute fusion module is used to perform conditional random field attribute fusion on the fault zone data and porosity data volume to obtain the fault sweet spot reservoir data volume; The reservoir prediction module is used to input the fault-type sweet spot reservoir data body into a preset three-dimensional visualization model for data visualization conversion to obtain a fault-type sweet spot reservoir image.

[0013] In summary, the beneficial effects of the present invention are as follows: According to the present invention, a method for predicting a fault-type sweet spot reservoir based on three-dimensional seismic data is provided. The method first performs denoising on the three-dimensional seismic data to obtain denoised three-dimensional seismic data, and extracts fault data from the denoised three-dimensional seismic data; then further extracts fault body data from the fault data, and performs morphological expansion operation on the fault body data to generate fault zone data. By applying intrinsic coherence technology and gradient analysis technology, faults and cracks can be more accurately identified and extracted, and morphological expansion operation is also used to refine the properties of the fault body, which is no longer limited to providing linear fault and crack information, and the complex structure of the fault zone can be finely portrayed. Then, three-dimensional seismic porosity parameter simulation is performed on the three-dimensional seismic data to obtain a porosity data body; finally, conditional random field attribute fusion is performed on the fault zone data and the porosity data body, which can more comprehensively explain the geological characteristics of the reservoir, including the development degree of faults and cracks and the coupling relationship with high-porosity reservoirs, which helps to identify the "sweet spot" area in the reservoir and can be used to better assess the risks of exploration and development. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the following is a brief introduction to the drawings required for use in the embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work, and these are all within the protection scope of the present invention.

[0015] Figure 1 It is a schematic diagram of the process of the fracture-type sweet spot reservoir prediction method based on three-dimensional seismic data of the present invention; Figure 2 is a schematic diagram of a seismic profile before denoising processing of the present invention; Figure 3 It is a schematic diagram of a seismic profile after denoising processing according to the present invention; Figure 4 It is a schematic cross-sectional diagram of the superposition of fault data and seismic data of the present invention; Figure 5 It is a schematic cross-sectional diagram of fault data of the present invention; Figure 6 It is a schematic cross-sectional diagram of the fracture volume data of the present invention; Figure 7 It is a schematic cross-sectional diagram of the fracture zone data of the present invention; Figure 8This is a cross-sectional schematic diagram of the superposition of the fracture volume data and the fracture zone data of the present invention; Fig. 9 is a schematic diagram of the porosity cross section of the present invention; Fig.10 It is a schematic cross-sectional diagram of the fracture-type sweet spot reservoir data of the present invention; Fig.11 It is a cross-sectional schematic diagram of the superposition of the fracture type sweet spot reservoir data and the fracture volume data of the present invention; Fig.12 This is a schematic diagram of the three-dimensional visualization effect of the fractured sweet spot reservoir of the present invention; Fig.13 This is a functional module diagram of the fracture-type sweet spot reservoir prediction device based on three-dimensional seismic data of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. If there is no conflict, the various features of the present invention and the embodiments can be combined with each other, all within the scope of protection of the present invention.

[0017] Example 1: Reference Figure 1 As shown, Figure 1 This is a flow chart of a method for quantitative characterization of seismic waveforms of the present invention. Figure 1 As shown, a method for predicting fracture-type sweet spot reservoirs based on three-dimensional seismic data of the present invention comprises: S1: De-noising the 3D seismic data using the dip-guided coherence enhancement denoising method to obtain de-noised 3D seismic data; S2: Using the intrinsic coherence volume technology to extract the key eigenvalues ​​in the seismic waveform from the denoised 3D seismic data, and identifying the faults of the strata according to the key eigenvalues ​​to obtain the fault data; S3: performing approximate derivative calculation on the fault data, and obtaining corresponding fault volume data according to the approximate derivative calculation result; S4: Perform morphological expansion operation on the fracture volume data to generate fracture zone data; S5: Using seismic waveform indication inversion technology to simulate 3D seismic porosity parameters on 3D seismic data to obtain a porosity data volume; S6: Fault zone data and porosity data are fused with conditional random field attributes to obtain fault sweet spot reservoir data; S7: Input the fracture-type sweet spot reservoir data volume into a preset three-dimensional visualization model for data visualization conversion to obtain a fracture-type sweet spot reservoir image.

[0018] Specifically, in the embodiment of the present invention, Figure 1 and Figure 2 This is a schematic diagram of the seismic profile before and after denoising the 3D seismic data using the dip-guided coherent enhancement denoising method. Denoising is an important step in seismic data processing because random noise can mask or blur seismic reflection information, especially for small geological structures such as faults. Intrinsic coherence technology improves the quality of seismic signals through specific algorithms, among which the dip-guided coherent enhancement denoising method is an effective technical means.

[0019] The method first estimates the dip information of the seismic body using plane wave destruction filtering technology. The dip information reflects the geometric characteristics of the strata and helps to identify the energy direction of the seismic reflection. By calculating the similarity between seismic traces, the coherence volume attribute is generated, which can highlight the continuity and consistency of the seismic signal, thereby enhancing the effective signal. Combined with the dip information, the seismic signal is denoised in a directional manner. This method does not simply suppress noise in the horizontal direction, but is carried out along the inclination and strike of the seismic reflection interface, which is more in line with the actual situation of the geological body. While denoising, the dip-guided filtering method can better protect important geological information such as faults in the seismic signal, avoiding the loss of these key details in the denoising process. Through the above processing, the signal-to-noise ratio of the seismic data is improved, making the geological structure such as faults more clearly visible, thereby providing a more reliable data basis for subsequent geological interpretation and reservoir prediction.

[0020] Specifically, intrinsic coherence technology is an advanced technology used in seismic data interpretation. It characterizes the lateral heterogeneity of the strata by analyzing the similarity of adjacent seismic trace signals in the seismic data volume, thereby identifying faults in the strata. The core of this technology is to calculate the coherence value of the seismic data volume to highlight discontinuities in the seismic data, such as faults and cracks.

[0021] When processing seismic data, the intrinsic coherence technology defines the three-dimensional seismic data as a matrix and extracts the key eigenvalues ​​in the seismic waveform through mathematical methods to characterize the characteristics of the formation (as shown in the following formula 1).

[0022] (1) In the formula, represents the intrinsic coherence value, represents a trace in seismic data, represents the total number of traces in the seismic data, represents the number of sampling points in the seismic data, and Represents the first and second eigenvalues, For the The jth sampling point.

[0023] The tomographic data obtained by intrinsic coherence calculation are as follows Figure 4 and Figure 5 As shown, Figure 4 For the superposition of fault data and seismic data, Figure 5 It is the section of fault data.

[0024] Furthermore, based on the intrinsic coherence to extract fault information, further extracting the properties of the fault body is a key technology in oil and gas exploration. The fault body usually refers to an area of ​​a certain width around the fault line affected by fault activity. This area may contain fracture networks, porosity changes and fluid flow channels, which are crucial to the storage and flow of oil and gas.

[0025] The application of gradient analysis algorithm in seismic data processing, especially in the detection of faults and fractures, is an important technical means. As a common method for calculating gradients, the finite difference method is a numerical method based on Taylor series expansion, which is used to approximate the derivative of a continuous function at discrete data points. For seismic data, this means that the derivative can be approximately calculated by the value of the data point without an analytical function form. Commonly used methods include forward difference method, backward difference method and central difference method.

[0026] The forward difference method calculates the derivative of a function at a certain point, using the difference between the function value at that point and the point after it. The backward difference method is the opposite of the forward difference method, using the value of the function at the point before a certain point to calculate the derivative. The central difference method combines the advantages of the forward and backward difference methods, using the values ​​of the function at two points before and after a certain point to approximate the derivative (Formula 2), which is expressed as this method is more stable and more accurate than the previous two methods. Therefore, the embodiment of the present invention uses the central difference method to calculate the approximate derivative of the fault data: (2) In the formula, Represents the location of a sampling point in seismic data; Represents the distance between two adjacent sampling points, that is, the step length; Representation function At the point The derivative at and Respectively exist The value at the sampling point after and before the point.

[0027] Finally, according to the approximate derivative calculation results, we finally get Figure 6 The fracture volume data section is shown.

[0028] Further, in the embodiment of the present invention, the morphological dilation operation is an image processing technology based on a structural element, which is used to expand and fill a binary image. In the process of expanding the fault line body to form a fault zone, the morphological dilation operation can be used to extend the influence of the fault point to the surrounding area. Therefore, in the embodiment of the present invention, the morphological dilation operation is performed on the fault body data to generate the fault zone data, including the following process: The input fracture body is an approximate binary image, assuming that , the structural element is , the result of the expansion operation is .

[0029] First, based on the fracture image in the fracture data, define the input fracture image and structural elements ; Represents the input image The pixel value in is the sampling point of the fracture body. and Represents the input image The horizontal and vertical coordinates of the pixels in ; Represents structural elements The pixel value in and Represents the structural elements The horizontal and vertical coordinates of the pixels in ; The shape of the structural element is designed to be circular, and the direction of the structural element extending outward is perpendicular to the fracture body direction. Dilate and fill, for each pixel coordinate The pixel value at which the morphological dilation operation is calculated (Formula 3): (3) in, Represents structural elements Pixel coordinates in ; Repeat the fracture image The expansion operation is performed until the expanded image meets the preset image standard, and the corresponding fracture zone data is obtained.

[0030] Among them, considering that the process of fault zone formation is uniform outward expansion, the shape of the structural element is designed to be circular, and the direction of outward expansion of the structural element is perpendicular to the direction of the fault body, so as to ensure that the direction of the fault zone formed after expansion is consistent with the direction of the fault body.

[0031] The faults in the fault body are larger structural elements, which will make the expansion effect more obvious, and the cracks are smaller structural elements, which will produce a finer expansion effect. Repeat the expansion operation on the entire fault body image, and keep observing the expansion results until the results that geologists consider reasonable are met. Then stop the expansion and get the final fault zone prediction result, that is, get Figure 7 and Figure 8 The fault zone data section shown. Figure 7 is the data section of the fault zone. Figure 8 The fault volume data is superimposed on the fault zone data profile.

[0032] Furthermore, in 3D seismic reservoir prediction, the seismic waveform indication simulation method is a commonly used technology. It can directly simulate the porosity body by analyzing the correspondence between the seismic waveform and the porosity curve. Compared with other reservoir inversion methods, this method has higher accuracy and reliability. In the embodiment of the present invention, the 3D seismic porosity parameter simulation is performed on the 3D seismic data using the seismic waveform indication inversion technology to obtain the porosity data body, which specifically includes the following process: The singular value decomposition method is used to perform efficient dynamic clustering analysis of the waveform of the wellbore seismic trace of 3D seismic data to establish a mapping relationship between the seismic waveform structure and the porosity curve structure, and generate a set of porosity curve samples representing different seismic facies types.

[0033] By analyzing the distribution of porosity curve sample sets of different seismic facies types, the corresponding Bayesian inversion frameworks are established respectively; in each Bayesian framework, the common part of the sample set is selected as the initial model, and iterative simulation is performed to obtain a high-resolution porosity data volume. Among them, the porosity profile in the porosity data volume is as follows Fig. 9 shown.

[0034] Furthermore, attribute fusion technology is a method that integrates information from different data sources to improve prediction accuracy. In seismic reservoir prediction, conditional random fields (CRF) is a commonly used attribute fusion technology, which can combine the geological characteristics of the fault zone with the reservoir parameters of the porosity data body to generate a data body that comprehensively reflects the reservoir characteristics. In the embodiment of the present invention, conditional random field attributes are fused on the fault zone data and the porosity data body to obtain the fault-type sweet spot reservoir data body, which specifically includes the following processes: A conditional random field model is defined. The conditional random field model takes the geological characteristics of the fault zone and the reservoir parameters of the porosity data volume as input and is used to predict the reservoir characteristics of each pixel or sampling point.

[0035] The reservoir characteristics-related features are selected from the fault zone data and the porosity data to construct a potential function to quantify the value of a single variable or the relationship between variables. Potential functions usually include two types of potential functions: one is the node potential, which describes the distribution of a single variable; the other is the edge potential, which describes the relationship between adjacent variables. Selecting the edge potential in the process of fusing the fault zone data and the porosity data can better handle the matching relationship between the fault zone data and the porosity data.

[0036] The maximum likelihood estimation algorithm is used to learn the parameters of the potential function. The learned parameters are used to control the conditional random field model to perform parameter fusion inference and prediction on the input fault zone data and porosity data volume, and the corresponding fault sweet spot reservoir data volume is generated. Finally, Fig.10 and Fig.11 The fractured sweet spot reservoir data section shown in Figure 1 shows a fractured sweet spot reservoir data section, where Fig.10 This is the data section of the fault-type sweet spot reservoir. Fig.11 This is a section of fault-type sweet spot reservoir data superimposed on fault volume data.

[0037] Specifically, in the embodiments of the present invention, the 3D visualization model converts complex underground structures into intuitive images by integrating different geological and geophysical attributes (such as faults, fractures, porous reservoirs, etc.) into a single 3D model, helping geologists identify the geometric morphology of reservoirs, such as layers, fracture networks, or irregular sweet spots, and helps to show the spatial relationship between reservoirs and faults, fracture networks, and other geological structures, which is crucial for understanding the fluid flow path within the reservoir. Fig.12 The schematic diagram of the 3D visualization effect of the fractured sweet spot reservoir is shown. The intuitive 3D model can guide the formulation of development strategies, such as determining the best drilling location, evaluating potential oil and gas production, and predicting long-term development effects.

[0038] The method of the embodiment of the present invention has the following technical effects: First, by applying intrinsic coherence technology and gradient analysis technology, faults and cracks can be more accurately identified and extracted, and morphological dilation operations can be further used to refine the properties of the fault body. Compared with traditional prediction technology, this technology is no longer limited to providing linear fault and crack information, but can finely depict the complex structure of the fault zone. This detailed analysis method is closer to the actual occurrence of underground faults.

[0039] Second, the core advantage of using seismic waveform indication inversion technology to carry out three-dimensional seismic porosity parameter simulation is that it can obtain high-resolution inversion results and provide more accurate underground reservoir pore structure information.

[0040] Third, conditional random field attribute fusion technology can integrate information from different data sources and provide a unified framework to process and fuse multi-attribute data. By fusing fault zone and porosity data, the geological characteristics of the reservoir can be more comprehensively explained, including the development degree of faults and fractures and the coupling relationship with high-porosity reservoirs, which helps to identify the "sweet spot" area in the reservoir and can be used to better assess the risks of exploration and development.

[0041] Example 2: Reference Fig.13 As shown, the embodiment of the present invention also provides a fracture-type sweet spot reservoir prediction device based on three-dimensional seismic data, comprising: A data denoising module is used to denoise the three-dimensional seismic data using a dip-guided coherence enhancement denoising method to obtain denoised three-dimensional seismic data; The fault identification module is used to extract key characteristic values ​​in the seismic waveform from the denoised three-dimensional seismic data using the intrinsic coherence volume technology, and to identify faults in the strata according to the key characteristic values ​​to obtain fault data; A fracture body extraction module, used for performing approximate derivative calculation on the fault data, and obtaining corresponding fracture body data according to the approximate derivative calculation result; A fracture zone generation module is used to perform morphological expansion operations on fracture volume data to generate fracture zone data; A seismic inversion module is used to simulate the three-dimensional seismic porosity parameters of three-dimensional seismic data using seismic waveform indication inversion technology to obtain a porosity data volume; The attribute fusion module is used to perform conditional random field attribute fusion on the fault zone data and porosity data volume to obtain the fault sweet spot reservoir data volume; The reservoir prediction module is used to input the fault-type sweet spot reservoir data body into a preset three-dimensional visualization model for data visualization conversion to obtain a fault-type sweet spot reservoir image.

[0042] The prediction device of the embodiment of the present invention first performs denoising on the three-dimensional seismic data to obtain the denoised three-dimensional seismic data, and extracts the fault data from the denoised three-dimensional seismic data; then further extracts the fracture body data from the fault data, and performs morphological expansion operation on the fracture body data to generate the fracture zone data. By applying the intrinsic coherence technology and the gradient analysis technology, the faults and cracks can be more accurately identified and extracted, and the morphological expansion operation is also used to refine the properties of the fracture body, which is no longer limited to providing linear fault and crack information, and can finely depict the complex structure of the fault zone. Then the three-dimensional seismic data is simulated by three-dimensional seismic porosity parameters to obtain the porosity data body; finally, the fault zone data and the porosity data body are fused with conditional random field attributes, which can more comprehensively explain the geological characteristics of the reservoir, including the development degree of faults and cracks and the coupling relationship with the high-porosity reservoir, which helps to identify the "sweet spot" area in the reservoir and can be used to better assess the risks of exploration and development.

[0043] Specifically, an embodiment of the present invention further provides a computer storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method for quantitative characterization of seismic waveforms as described in Example 1 is implemented.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting fracture-type sweet spot reservoirs based on three-dimensional seismic data, characterized in that: include: The 3D seismic data is denoised using the dip-guided coherence enhancement denoising method to obtain denoised 3D seismic data. The key eigenvalues ​​in the seismic waveform are extracted from the denoised 3D seismic data using the intrinsic coherence technology, and the faults of the strata are identified based on the key eigenvalues ​​to obtain the fault data. Performing approximate derivative calculation on the fault data, and obtaining corresponding fault volume data according to the approximate derivative calculation result; Perform morphological expansion operation on the fracture volume data to generate fracture zone data; The seismic waveform indication inversion technology is used to simulate the 3D seismic porosity parameters of the 3D seismic data to obtain the porosity data volume; The fault zone data and porosity data are fused with conditional random field attributes to obtain the fault sweet spot reservoir data volume; The data volume of the fault-type sweet spot reservoir is input into the preset three-dimensional visualization model for data visualization conversion to obtain the image of the fault-type sweet spot reservoir.

2. The method for predicting fracture-type sweet spot reservoir according to claim 1, characterized in that: The step of performing a morphological expansion operation on the fracture volume data to generate fracture zone data includes: Based on the fracture image in the fracture data, define the input fracture image and structural elements ; Represents the input image The pixel value in is the sampling point of the fracture body. and Represents the input image The horizontal and vertical coordinates of the pixels in ; Represents structural elements The pixel value in and Represents the structural elements The horizontal and vertical coordinates of the pixels in ; The shape of the structural element is designed to be circular, and the direction of the structural element extending outward is perpendicular to the fracture body direction. Dilate and fill, for each pixel coordinate The pixel value at which the morphological dilation operation is calculated : in, Represents structural elements Pixel coordinates in ; Repeat the fracture image The expansion operation is performed until the expanded image meets the preset image standard, and the corresponding fracture zone data is obtained.

3. The method for predicting fracture-type sweet spot reservoir according to claim 1, characterized in that: The method of using the seismic waveform indication inversion technology to simulate the three-dimensional seismic data for three-dimensional seismic porosity parameters to obtain a porosity data volume includes: Using the singular value decomposition method, an efficient dynamic clustering analysis of the waveform of the wellbore seismic trace is performed on the 3D seismic data to establish a mapping relationship between the seismic waveform structure and the porosity curve structure, and generate a set of porosity curve samples representing different seismic facies types. By analyzing the distribution of porosity curve sample sets of different seismic facies types, corresponding Bayesian inversion frameworks are established respectively; under each Bayesian framework, the common part of the sample set is selected as the initial model, and iterative simulation is performed to obtain a high-resolution porosity data volume.

4. The method for predicting fracture-type sweet spot reservoir according to claim 1, characterized in that: The conditional random field attribute fusion is performed on the fault zone data and the porosity data volume to obtain the fault sweet spot reservoir data volume, including: A conditional random field model is defined, which takes the geological characteristics of the fault zone and the reservoir parameters of the porosity data volume as input to predict the reservoir characteristics of each pixel or sampling point; Selecting reservoir-related features from the fault zone data and porosity data volume to construct a potential function to quantify the value of a single variable or the relationship between variables; The maximum likelihood estimation algorithm is used to learn the parameters of the potential function. The learned parameters are used to control the conditional random field model to perform parameter fusion inference and prediction on the input fault zone data and porosity data volume to generate the corresponding fault-type sweet spot reservoir data volume.

5. A fracture-type sweet spot reservoir prediction device based on three-dimensional seismic data, characterized in that: include: A data denoising module is used to denoise the three-dimensional seismic data using a dip-guided coherence enhancement denoising method to obtain denoised three-dimensional seismic data; The fault identification module is used to extract key characteristic values ​​in the seismic waveform from the denoised three-dimensional seismic data using the intrinsic coherence volume technology, and to identify faults in the strata according to the key characteristic values ​​to obtain fault data; A fracture body extraction module, used for performing approximate derivative calculation on the fault data, and obtaining corresponding fracture body data according to the approximate derivative calculation result; A fracture zone generation module is used to perform morphological expansion operations on fracture volume data to generate fracture zone data; A seismic inversion module is used to simulate the three-dimensional seismic porosity parameters of three-dimensional seismic data using seismic waveform indication inversion technology to obtain a porosity data volume; The attribute fusion module is used to perform conditional random field attribute fusion on the fault zone data and porosity data volume to obtain the fault sweet spot reservoir data volume; The reservoir prediction module is used to input the fault-type sweet spot reservoir data body into a preset three-dimensional visualization model for data visualization conversion to obtain a fault-type sweet spot reservoir image.

Images