A method and device for automatically grading fracture based on high-frequency volume space features

Abstract

The application relates to the technical field of oilfield exploitation, in particular to a fracture automatic grading description method and device based on high-frequency body space features, which comprises the following steps: arranging excitation points and receiving point grids in a study area, and obtaining a set of seismic signals collected by each receiving point in the study area; performing static correction, dynamic correction and denoising processing on each seismic signal; determining the peak position distribution consistency coefficient between the seismic signals; obtaining the first residual coefficient of each seismic signal to determine the confidence of the peak position distribution consistency coefficient between the seismic signals, and obtaining the second residual coefficient of each seismic signal; calculating the stacking weight of each seismic signal when performing stacking processing on all seismic signals in the set of seismic signals, obtaining the seismic trace signal of each receiving point after the stacking processing, constructing the original three-dimensional seismic data body of the study area, and performing automatic fracture grading according to the fracture multidimensional dynamic grading standard. Therefore, the description accuracy of fracture grading is improved.

Description

Technical Field

[0001] This application relates to the field of oilfield development technology, specifically to an automatic fracture classification description method and apparatus based on high-frequency volume spatial characteristics. Background Technology

[0002] When using seismic data from oil reservoirs to classify and describe faults in the reservoir, existing technologies typically employ a combination of seismic detectors to receive multiple seismic signals and then superimpose these signals as a single seismic trace signal in the final acquired seismic data. This is done to suppress regular and random interference introduced into the seismic data, thereby reducing the impact of regular and random interference on the classification and description of faults in the oil reservoir.

[0003] However, this technology ignores the differences in seismic signals caused by the different elevations and shot-receiver distances among the detectors in the seismic detector array, as well as the differences in random interference introduced into the seismic signals. This makes it easy for the multiple seismic signals received by the seismic detector array to be superimposed into a single seismic trace signal, which may lead to distortion of the superimposed seismic trace signal and fail to reflect the true geological information of the reservoir, thereby affecting the accurate classification and description of reservoir faults. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method and apparatus for automatic fracture classification and description based on high-frequency volumetric spatial characteristics. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide an automatic fracture classification description method based on high-frequency volume spatial characteristics, the method comprising the following steps:

[0006] A grid of excitation and receiving points was set up in the study area to obtain the set of seismic signals collected at each receiving point in the study area;

[0007] Each seismic signal is sequentially subjected to static correction, dynamic correction, and denoising; the synchronization degree of peak positions among seismic signals within the processed seismic signal set is analyzed to determine the peak position distribution consistency coefficient among seismic signals;

[0008] By using the waveform similarity between seismic signals within the seismic signal set, a first residual coefficient is obtained to evaluate the degree of noise residue of each seismic signal after denoising processing, in order to determine the confidence level of the peak distribution consistency coefficient between seismic signals, and a second residual coefficient is obtained for each seismic signal.

[0009] By combining the first and second residual coefficients, the superposition weight of each seismic signal is calculated when all seismic signals in the seismic signal set are superimposed, and the seismic trace signals of each receiving point after superposition processing are obtained. The original three-dimensional seismic data volume of the study area is constructed, and it is decomposed and reconstructed to obtain the high-frequency three-dimensional seismic data volume.

[0010] Extract the advantageous structural attributes of high-frequency 3D seismic data volumes along layer slices, establish a 3D spatial model of faults, extract fault spatial distribution parameters, and automatically classify faults according to multi-dimensional dynamic grading standards.

[0011] In one embodiment, the seismic signal set includes seismic signals received by seismic detectors at the same receiving point.

[0012] In one embodiment, determining the peak distribution consistency coefficient includes:

[0013] Identify each wave peak and its corresponding time in each seismic signal, calculate the metric distance between the time series of all wave peak corresponding times of any two seismic signals in the seismic signal set, and determine the peak distribution consistency coefficient between any two seismic signals based on the metric distance, wherein the peak distribution consistency coefficient is negatively correlated with the metric distance.

[0014] In one embodiment, determining the first residual coefficient includes:

[0015] Calculate the cross-correlation coefficient between any seismic signal in the seismic signal set and all other seismic signals, and perform forward fusion on the absolute values ​​of all cross-correlation coefficients corresponding to any seismic signal to obtain the first residual coefficient.

[0016] In one embodiment, the mean of the first residual coefficients of any two seismic signals is calculated as the confidence level of the peak distribution consistency coefficient between the two seismic signals.

[0017] In one embodiment, determining the second residual coefficient includes:

[0018] Using the normalized confidence level of the peak distribution consistency coefficients among seismic signals as weights, the peak distribution consistency coefficients between any seismic signal and all other seismic signals in the seismic signal set are weighted and summed to obtain the second residual coefficient of the given seismic signal.

[0019] In one embodiment, the superposition weight is the average of the normalized results of the first residual coefficient and the normalized results of the second residual coefficient.

[0020] In one embodiment, the extraction of the advantageous structural properties along layer slices includes:

[0021] Based on high-frequency 3D seismic data, structural attribute volumes are extracted, and then slices of different structural attributes along the target segment are extracted. The accuracy of different attribute slices in identifying faults is analyzed, and the structural attribute slices that can identify fault information of different levels are selected as the superior structural attribute slices.

[0022] In one embodiment, the RGB fusion method is used to assign red, green and blue color channels to the advantageous structural attributes along the layer slices for information fusion, and then three-dimensional modeling technology is used to establish a fracture three-dimensional space model.

[0023] The fracture space distribution parameters include geometric parameters, connectivity parameters, seepage and conductivity parameters, and parameters of adjacent well test wells.

[0024] Secondly, embodiments of this application also provide an automatic fracture grading description device based on high-frequency volume spatial characteristics, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0025] This application has at least the following beneficial effects:

[0026] This application eliminates signal offsets caused by differences in elevation and shot-receiver distance by performing static and dynamic corrections on the seismic signals received by each geophone. This process ensures that signals from different geophones can be compared on the same benchmark, making subsequent analysis more consistent and effectively reducing the distortion probability during signal superposition. Furthermore, denoising significantly reduces the impact of random interference on the seismic signals. By calculating the first and second residual coefficients, the superposition weight of each seismic signal is obtained when superimposing all seismic signals received by the same geophone combination. This effectively reduces the signal distortion phenomenon caused by differences in the propagation time of reflected waves due to differences in geophone elevation and shot-receiver distance, as well as random interference caused by environmental noise, in the seismic trace signals obtained after signal superposition. This allows the original three-dimensional seismic data volume established from the seismic trace signals to more accurately reflect the true reservoir geological information of the study area, thereby improving the accuracy of subsequent identification and description of reservoir faults in the study area using the original three-dimensional seismic data volume. Furthermore, this application decomposes high-frequency single-frequency volumes from conventional 3D seismic data volumes, breaking through traditional bandwidth limitations. It then reconstructs new high-frequency 3D seismic data volumes through data fusion. Based on this, it extracts structural attributes and selects corresponding attributes for RGB three-color fusion to perform fine fracture characterization. A high-precision 3D spatial model is established based on the fine fracture characterization results to clarify the spatial distribution parameters of fractures, thereby establishing a multi-dimensional dynamic grading standard. Finally, based on the standard, automatic fracture grading of the study area is completed. This solves the problems of single grading dimensions and insufficient spatial feature correlation in existing technologies, and is particularly suitable for fine characterization and dynamic evolution analysis of fracture systems in fracture-dense reservoirs, improving the accuracy of fracture grading description. Attached Figure Description

[0027] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 A flowchart illustrating the steps of an automatic fracture classification description method based on high-frequency volume spatial features, provided in one embodiment of this application;

[0029] Figure 2 The result of single-frequency data volume spectral decomposition of the original 3D seismic data volume;

[0030] Figure 3 The high-frequency three-dimensional seismic data volume obtained from reconstruction;

[0031] Figure 4 Slice along the layer for similarity;

[0032] Figure 5 Slice along the layer with maximum curvature;

[0033] Figure 6 This is a coherent slice along the layer;

[0034] Figure 7 This is a multi-attribute RGB fusion image;

[0035] Figure 8 A three-dimensional spatial model of the fracture. Detailed Implementation

[0036] To further illustrate the technical means and effects adopted by this application to achieve the intended inventive objective, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an automatic fracture classification description method and apparatus based on high-frequency volume spatial characteristics proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0038] The following description, in conjunction with the accompanying drawings, details the specific scheme of the automatic fracture classification description method and apparatus based on high-frequency volume spatial characteristics provided in this application.

[0039] Please see Figure 1 The diagram illustrates a flowchart of an automatic fracture grading description method based on high-frequency volume spatial features according to an embodiment of this application. The method includes the following steps:

[0040] S1, arranging a grid of excitation and reception points in the study area to obtain the set of seismic signals collected at each reception point in the study area.

[0041] This embodiment takes the Fuyu oil layer in the Daqing Chaoyanggou oilfield as an example, using this oil layer as the study area for automatic fault classification and description. Raw seismic data for the study area is obtained by arranging a grid of excitation and reception points. The excitation points use explosive detonation to generate seismic waves, and each reception point in the reception point grid uses a combination of seismic detectors to receive seismic signals. The set of all seismic signals received by all detectors in the seismic detector combination is denoted as the seismic signal set. The raw seismic data consists of the set of all seismic signals received by all reception points in the reception point grid.

[0042] S2 performs static correction, dynamic correction, and denoising on each seismic signal in sequence; analyzes the synchronization degree of the peak positions among the seismic signals in the processed seismic signal set, and determines the peak position distribution consistency coefficient among the seismic signals.

[0043] When using individual detectors in a seismic detector array to receive seismic signals, the elevations of all detectors in the same array may not be completely consistent due to the undulations of the surface topography. This can cause the same reflected wave generated by the excitation point in the study area to have different propagation times when it reaches detectors at different elevations after being reflected by the strata underground. Consequently, directly superimposing all the seismic signals received by the same seismic detector array will result in distortion of the reflected waveform in the superimposed signal.

[0044] Therefore, this embodiment first uses the reference plane static correction method to perform static correction processing on each seismic signal in each seismic signal set in the acquired raw seismic data. This reduces the impact of reflected wave distortion caused by the difference in reflection wave propagation time due to the different elevations of the detectors in the same seismic detector array when all seismic signals received by the same seismic detector array are superimposed, which affects the subsequent automatic fault classification description of the study area. The reference plane static correction method is a well-known technique, and the specific process will not be described in detail.

[0045] Secondly, when using individual detectors in a seismic detector array to receive seismic signals, the same reflected wave generated by the seismic wave at the excitation point in the study area will have different propagation times when it reaches different detectors in the same seismic detector array due to the different source-receiver distances (horizontal distance between the detector and the excitation point). This will cause distortion of the reflected waveform in the superimposed signal when all the seismic signals received by the same seismic detector array are directly superimposed.

[0046] Therefore, the dynamic correction method based on third-order cumulants is continued to be used to perform dynamic correction processing on each seismic signal after static correction. This reduces the impact of reflected wave distortion caused by the difference in reflection wave propagation time due to the different source-receiver offsets of the detectors in the same seismic detector array on the subsequent automatic fault classification description of the study area after all seismic signals received by the same seismic detector array are superimposed. The dynamic correction method based on third-order cumulants is a well-known technique, and the specific process will not be described in detail.

[0047] However, static and dynamic correction processing cannot completely eliminate the differences in reflected wave propagation time caused by variations in detector elevation and source-receiver distance among all seismic signals received by the same seismic detector array. Therefore, this embodiment needs to further evaluate the residual reflected wave propagation time differences in each seismic signal after static and dynamic correction processing. The aim is to assign different weights when superimposing all seismic signals received by the same seismic detector array, thereby reducing the impact of reflected wave distortion caused by residual reflected wave propagation time differences in the pre-superposition seismic signals on subsequent automatic fault classification descriptions of the study area.

[0048] Ideally, if the differences in reflected wave propagation time caused by variations in detector elevation and shot-receiver distance could be completely eliminated, then different seismic signals received by different detectors in the same seismic detector array should receive the same reflected wave at the same time. This would ensure that the peaks corresponding to the same reflected wave (reflected waves with the same reception order) in all seismic signals received by the same seismic detector array should have the same peak position after static and dynamic correction. Therefore, for any seismic signal received by any seismic detector array after static and dynamic correction, the greater the difference in the peak position distribution of reflected waves between a particular seismic signal and the others, the greater the residual degree of the reflected wave propagation time difference caused by variations in detector elevation and shot-receiver distance in that seismic signal.

[0049] However, the detectors in the seismic detector array are inevitably subject to noise interference when receiving seismic signals, which introduces random interference into the received seismic signals. This random interference affects the accuracy of obtaining the peak positions of each reflected wave in the seismic signal, and consequently affects the accuracy of assessing the differences in the propagation time of the residual reflected waves in the seismic signal.

[0050] Therefore, to reduce the impact of random interference in the received seismic signals on the assessment results of residual reflected wave propagation time differences, a seismic signal denoising method based on Radon transform is used to denoise each seismic signal after static and dynamic correction processing. This aims to suppress random interference in the seismic signals to the greatest extent possible, resulting in a statically corrected, dynamically corrected, and denoised seismic signal. This denoised signal is then used as data for superimposing all seismic signals received by the same seismic detector array, thereby reducing the differences in reflected wave propagation time caused by differences in detector elevation and source-receiver distance, as well as the impact of random interference introduced by environmental noise. The Radon transform-based seismic signal denoising method is a known technique, and implementers can choose other feasible existing denoising algorithms; this embodiment does not impose any restrictions on this.

[0051] Taking any two seismic signals B1 and B2 from any seismic signal set 'a' in the original seismic data as an example, the peak positions of all wave peaks in seismic signals B1 and B2 are extracted using a second-order difference peak-valley identification algorithm. This means the peak point corresponds to the sampling time in the seismic signal. All peak positions of seismic signals B1 and B2 are then sorted in ascending order according to their numerical values. The resulting two sequences represent the distribution of all reflected wave peak positions received in seismic signals B1 and B2, respectively. The metric distance between the two sequences is calculated, and the peak position distribution consistency coefficient between seismic signals B1 and B2 is determined based on this metric distance. The peak position distribution consistency coefficient is negatively correlated with the metric distance. The peak position distribution consistency coefficient is used to assess the degree of difference in the distribution of reflected wave peak positions between seismic signals B1 and B2; a larger coefficient indicates a smaller difference. The second-order difference peak-valley identification algorithm is a known existing technology, and implementers can choose other existing peak detection algorithms, such as automatic multi-scale peak detection algorithms.

[0052] In this embodiment, the reciprocal of the measured distance is used as the peak distribution consistency coefficient between seismic signals B1 and B2. It should be noted that when taking the reciprocal, if the denominator is 0, a parameter adjustment factor can be added to the denominator to prevent the denominator from being zero, ensuring the calculation result is meaningful. This parameter adjustment factor is a very small positive number. For example, the value of this parameter adjustment factor can be 0.01. Its specific value can be set by the implementer according to the actual situation; this embodiment does not impose specific limitations.

[0053] In another embodiment, the negative of the measured distance is used as the exponent of an exponential function with the natural constant as the base, and the result of the calculation of the exponential function is used as the peak distribution consistency coefficient between seismic signal B1 and seismic signal B2.

[0054] In this embodiment, the distance measurement is calculated using DTW distance. Implementers can choose other feasible distance measurement methods, such as Euclidean distance, Manhattan distance, etc.

[0055] S3, by using the waveform similarity between seismic signals within the seismic signal set, obtain the first residual coefficient to evaluate the noise residue level of each seismic signal after denoising processing, in order to determine the confidence level of the peak distribution consistency coefficient between seismic signals, and obtain the second residual coefficient of each seismic signal.

[0056] Denoising processing cannot completely eliminate all random interference in the seismic signals received by the geophones. Therefore, the smaller the degree of residual random interference in two seismic signals, the better the peak position distribution consistency coefficient calculated from these two seismic signals reflects the true difference in the distribution of reflected wave peak positions between the two seismic signals. Consequently, the assessment of the residual reflected wave propagation time difference in the seismic signals by this peak position distribution consistency coefficient is more accurate. Furthermore, all geophones in the same seismic detector array will receive reflected waves from the same geological layer due to their relatively close placement, resulting in similar waveforms in the received seismic signals. Random interference in the received seismic signals can disrupt this similarity.

[0057] Based on the above analysis, taking seismic signal B1 as an example, the absolute value of the cross-correlation coefficient between seismic signal B1 and each of the other seismic signals in seismic signal set a is calculated. The smaller the absolute value, the less similar the waveforms of seismic signal B1 are to each seismic signal in seismic signal set a. Therefore, the forward fusion result of all the absolute values ​​obtained from seismic signal B1 is used as the first residual coefficient of seismic signal B1, which is used to evaluate the degree of residual noise in the original seismic signal corresponding to seismic signal B1 after denoising processing. The larger the first residual coefficient, the more similar the waveforms of seismic signal B1 are to the other seismic signals in seismic signal set a, indicating a smaller degree of residual noise. The calculation of the cross-correlation coefficient between signals is a well-known technique, and the specific process will not be elaborated further.

[0058] It should be noted that forward fusion means combining multiple variables, which can be done by adding, multiplying, or averaging. In this embodiment, the mean of all the absolute values ​​obtained from the seismic signal B1 is used as the first residual coefficient of the seismic signal B1.

[0059] In the seismic signal set a, taking seismic signals B1 and B2 as examples, the mean of the first residual coefficients of seismic signals B1 and B2 is used as the confidence level of the peak position distribution consistency coefficient between seismic signals B1 and B2. The higher the confidence level, the smaller the residual degree of random interference in seismic signals B1 and B2. Then, the peak position distribution consistency coefficient calculated from seismic signals B1 and B2 can better reflect the difference in the actual distribution of reflected wave peak positions between the two seismic signals.

[0060] Calculate the peak position distribution consistency coefficient and the confidence level of the peak position distribution consistency coefficient between seismic signal B1 and each other seismic signal in seismic signal set a. Normalize the confidence level of the peak position distribution consistency coefficient between seismic signal B1 and all other seismic signals in seismic signal set a using the Softmax function, so that the sum of the confidence levels of all peak position distribution consistency coefficients corresponding to seismic signal B1 is 1.

[0061] Furthermore, using the normalized confidence level as the weight of the peak distribution consistency coefficient, the peak distribution consistency coefficients between seismic signal B1 and each other seismic signal in seismic signal set a are weighted and summed. The result is used as the second residual coefficient of seismic signal B1, which is used to evaluate the degree of residual of the difference in propagation time of reflected waves caused by the difference in detector elevation and shot-receiver distance in the original seismic signal corresponding to seismic signal B1 after static and dynamic correction. The larger the second residual coefficient, the smaller the corresponding residual degree.

[0062] S4. Combining the first residual coefficient and the second residual coefficient, calculate the superposition weight of each seismic signal when superimposing all seismic signals in the seismic signal set, obtain the seismic trace signal of each receiving point after superposition processing, construct the original three-dimensional seismic data volume of the study area, and decompose and reconstruct it to obtain the high-frequency three-dimensional seismic data volume.

[0063] When superimposing all seismic signals received by the same seismic detector array, the resulting signal will also exhibit reflected wave distortion due to random interference introduced into the seismic signals. Therefore, in this embodiment, when superimposing all seismic signals received by the same seismic detector array, an appropriate superposition weight is assigned to each seismic signal to effectively reduce the impact of reflected wave distortion caused by seismic signals with significant differences in residual reflected wave propagation time and large degrees of random interference on the subsequent automatic fault classification description of the study area.

[0064] Taking seismic signal set a as an example, the first and second residual coefficients of all seismic signals in seismic signal set a are normalized using the Min-Max normalization method. Taking seismic signal B1 as an example, the mean of the normalized results of the first and second residual coefficients of seismic signal B1 is used as the superposition weight of seismic signal B1. This weight is used to evaluate the magnitude of the superposition weight of seismic signal B1 when superimposing all seismic signals in seismic signal set a. The larger the superposition weight, the higher the accuracy of seismic signal B1, and the cleaner the removal of reflection wave propagation time differences and random interference in seismic signal B1, resulting in a lower residual degree.

[0065] The Softmax function is used to map the stacking weights of all seismic signals in the seismic signal set a to the range (0, 1), and the sum of the mapping results of all stacking weights is 1. Taking seismic signal B1 as an example, the stacking weights of seismic signal B1 after being mapped by the Softmax function are used as the final stacking weights of seismic signal B1.

[0066] Taking seismic signal set 'a' as an example, all seismic signals in seismic signal set 'a' are subjected to signal superposition processing. The final superposition weight value of each seismic signal is used as its weight in the signal superposition process. The signal obtained after superposition processing is then used as the seismic trace signal of the receiving point of the seismic detector array corresponding to seismic signal set 'a'. The signal superposition processing is a well-known technique, and its specific process will not be elaborated further.

[0067] Using the same method as for seismic signal set 'a', all seismic signals in each seismic signal set within the original seismic data were superimposed to obtain the seismic trace signals at the receiving points of the corresponding seismic detector arrays for each seismic signal set. These signals, combined with the geographic coordinates of the receiving points, were then used to construct the original 3D seismic data volume for the study area. In this volume, the horizontal dimension (X / Y axis) corresponds to the geographic coordinates of the receiving points, the vertical dimension (Z axis) corresponds to the seismic wave propagation time of the seismic trace signals at the receiving points, and the values ​​of the data cells (grid points in the 3D seismic data volume) correspond to the seismic wave reflection amplitude.

[0068] Furthermore, the acquired raw 3D seismic data volume is decomposed into multiple single-frequency data volumes. Specifically, wavelet transform is used to decompose the acquired raw 3D seismic data volume into equally spaced single-frequency data volumes. In this embodiment, the dominant frequency range of the raw 3D seismic data volume is 30~60Hz, and the set equal interval is 10Hz. The spectral decomposition result of the single-frequency data volume of the raw 3D seismic data volume is as follows: Figure 2 As shown. Wavelet transform is a well-known technique, and its specific process will not be described in detail.

[0069] Utilizing the characteristic of high-frequency information to enhance fracture signals, the high-frequency components of all single-frequency data volumes obtained from the decomposition of the original 3D seismic data volume are fused to reconstruct a new high-frequency 3D seismic data volume. The fused and reconstructed high-frequency 3D seismic data volume is as follows: Figure 3 As shown, in this embodiment, the frequency range of the high-frequency components is 60~150Hz. The signal reconstruction in the wavelet transform method is a well-known technique, and the specific process will not be described in detail.

[0070] S5 extracts the advantageous structural attributes of high-frequency 3D seismic data volumes along layer slices, establishes a 3D spatial model of faults, extracts fault spatial distribution parameters, and automatically classifies faults according to the multi-dimensional dynamic grading standard of faults.

[0071] Based on the reconstructed high-frequency 3D seismic data volume, the Petrel platform is used to extract different structural attribute volumes from the high-frequency 3D seismic data volume. The structural attribute volumes include similarity attributes, curvature attributes, ant-like volume attributes, coherence attributes, variance attributes, and edge detection-type attributes. Then, bedding slices of the target segment are extracted. From these bedding slices, structural attribute bedding slices that can identify different levels of fault information are selected as dominant structural attribute bedding slices. In this embodiment, structural attribute bedding slices with high fault identification accuracy are selected as dominant structural attribute bedding slices. Specifically, dominant structural attribute bedding slices include similarity bedding slices, maximum curvature bedding slices, and coherence bedding slices. Similarity bedding slices are as follows: Figure 4 As shown, the maximum curvature along the layer slice is as follows Figure 5 As shown, coherence along the slice is as follows Figure 6 As shown.

[0072] Based on the three advantageous constructed attributes obtained, slices are created along the layer. The RGB fusion method is then used to assign red, green, and blue color channels to these three advantageous constructed attributes along the layer slices for information fusion, resulting in a multi-attribute RGB fusion image, as shown below. Figure 7 As shown, this achieves fine depiction of fractures.

[0073] Based on the obtained detailed fracture characterization results, a detailed three-dimensional spatial model of the fracture is established using three-dimensional modeling technology, such as... Figure 8 As shown, the fracture spatial distribution parameters are extracted from the three-dimensional fracture spatial model based on the Paraview open-source platform. The fracture spatial distribution parameters include geometric parameters, connectivity parameters, seepage and conductivity parameters, and parameters of adjacent wells. The geometric parameters include spatial extension length, fault displacement, dip angle, and spatial curvature radius. The spatial connectivity parameters include whether it is connected to the oil source and whether it is active during the hydrocarbon generation and discharge period. The seepage and conductivity parameters are the cement content, and the parameters of adjacent wells are whether they are industrial oil flow wells.

[0074] Based on the obtained fracture spatial dynamic distribution parameters, a multi-dimensional dynamic classification standard for fractures was established, as shown in Table 1.

[0075] Table 1. Multidimensional Dynamic Grading Standards for Fracture

[0076] Based on the multi-dimensional dynamic grading standard for fractures, the spatial distribution parameters of the actual extracted fractures are determined, thereby realizing the automatic grading and description of fractures in different strata in the study area.

[0077] Based on the same inventive concept as the above method, this application embodiment also provides an automatic fracture classification and description device based on high-frequency volume spatial characteristics, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described methods for automatic fracture classification and description based on high-frequency volume spatial characteristics.

[0078] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, specific embodiments of this specification have been described above. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0079] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0080] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for automatic fracture classification and description based on high-frequency volumetric spatial features, characterized in that, The method includes the following steps: A grid of excitation and receiving points was set up in the study area to obtain the set of seismic signals collected at each receiving point in the study area; Each seismic signal is sequentially subjected to static correction, dynamic correction, and denoising; the degree of synchronization of peak positions among seismic signals within the processed seismic signal set is analyzed to determine the peak position distribution consistency coefficient among seismic signals; By using the waveform similarity between seismic signals within the seismic signal set, a first residual coefficient is obtained to assess the degree of noise residue in each seismic signal after denoising processing, thereby determining the confidence level of the peak distribution consistency coefficient between seismic signals and obtaining a second residual coefficient for each seismic signal. By combining the first and second residual coefficients, the superposition weight of each seismic signal is calculated when all seismic signals in the seismic signal set are superimposed, and the seismic trace signals of each receiving point after superposition processing are obtained. The original three-dimensional seismic data volume of the study area is constructed, and it is decomposed and reconstructed to obtain the high-frequency three-dimensional seismic data volume. Extract the advantageous structural attributes of high-frequency 3D seismic data volumes by layer slicing, establish a 3D spatial model of faults, extract the spatial distribution parameters of faults, and automatically classify faults according to the multi-dimensional dynamic classification standard of faults. The determination of the peak distribution consistency coefficient includes: Identify each wave peak and its corresponding time in each seismic signal, calculate the metric distance between the time series of all wave peak corresponding times of any two seismic signals in the seismic signal set, and determine the peak distribution consistency coefficient between any two seismic signals based on the metric distance, wherein the peak distribution consistency coefficient is negatively correlated with the metric distance; The determination of the first residual coefficient includes: Calculate the cross-correlation coefficient between any seismic signal in the seismic signal set and all other seismic signals, and perform forward fusion on the absolute values ​​of all cross-correlation coefficients corresponding to any seismic signal to obtain the first residual coefficient; The mean of the first residual coefficients of any two seismic signals is calculated and used as the confidence level of the peak distribution consistency coefficient between any two seismic signals. The determination of the second residual coefficient includes: Using the normalized confidence level of the peak distribution consistency coefficients among seismic signals as weights, the peak distribution consistency coefficients between any seismic signal and all other seismic signals in the seismic signal set are weighted and summed to obtain the second residual coefficient of the given seismic signal.

2. The automatic fracture classification description method based on high-frequency volume spatial characteristics as described in claim 1, characterized in that, The set of seismic signals includes seismic signals received by seismic detectors at the same receiving point.

3. The automatic fracture classification description method based on high-frequency volume spatial characteristics as described in claim 1, characterized in that, The superposition weight is the average of the normalized results of the first residual coefficient and the normalized results of the second residual coefficient.

4. The automatic fracture classification description method based on high-frequency volume spatial characteristics as described in claim 1, characterized in that, The extraction of the advantageous structural properties along the layer slice includes: Based on high-frequency 3D seismic data, structural attribute volumes are extracted, and then slices of different structural attributes along the target segment are extracted. The accuracy of different attribute slices in identifying faults is analyzed, and the structural attribute slices that can identify fault information of different levels are selected as the superior structural attribute slices.

5. The automatic fracture classification description method based on high-frequency volume spatial characteristics as described in claim 1, characterized in that, The RGB fusion method is used to assign red, green and blue color channels to the advantageous structural attributes along the layer slices for information fusion. Then, three-dimensional modeling technology is used to establish a fracture three-dimensional space model. The fracture space distribution parameters include geometric parameters, connectivity parameters, seepage and conductivity parameters, and parameters of adjacent well test wells.

6. An automatic fracture grading and description device based on high-frequency volume spatial characteristics, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-5.

Images