A method for identifying fractures based on post-stack seismic data

Abstract

The application relates to a fracture identification method based on post-stack seismic data, and belongs to the technical field of petroleum exploration interpretation. The application is characterized in that: through dip angle imaging enhancement processing, random noise is inhibited, original boundary information is reserved, then intrinsic coherence, fault enhancement and ant tracking technology are used to continuously enhance fracture information and suppress non-fracture information, so that a data body with low noise and clear fracture traces is obtained. The method is suitable for identification and interpretation of fractures with different scales under different geological conditions, and through in-depth mining of information contained in the seismic data, small and medium-sized fractures can be effectively identified, and the fracture identification precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to a fracture identification method based on post-stack seismic data, and belongs to the technical field of petroleum exploration interpretation. BACKGROUND

[0002] Fracture identification is an important part of petroleum exploration interpretation, and the accuracy of fracture identification is of great significance to correctly understanding the oil and gas distribution law of the fracture development area in the oil and gas basin, and even directly affects the decision of oil and gas exploration and development. The commonly used fracture identification is based on the interpretation technology of conventional seismic data, and through extracting seismic attributes sensitive to fractures, such as coherence, dip angle, curvature, ant body, etc., the fractures with obvious fault throw and clear seismic response characteristics can be effectively identified. However, the actual underground fracture is relatively complex, and there are different types and different scales of fractures. Especially for small-scale buried fractures, the seismic response characteristics are not obvious, and it is difficult to identify them with a single seismic attribute volume.

[0003] At present, the commonly used fracture identification method based on post-stack seismic data mainly includes the following:

[0004] (1) Coherent body fracture identification: the main basis is that the seismic waves between adjacent channels exist similarity, and the existence of fault will produce a sharp response on the seismic wave, affecting the similarity of seismic waves between channels. By using coherent body technology, three-dimensional seismic data volume can be operated according to a certain algorithm, and the coherence intensity can be divided according to the response degree of seismic wave in the stratum. The strong incoherent part often represents the distribution area of the fault.

[0005] (2) Dip angle body fracture identification: the main method is to obtain the local dip angle data and amplitude variance data of any point of the corresponding stratum by processing the seismic data volume. The existence of fault will increase the amplitude variance value, so as to determine the fault development area.

[0006] (3) Curvature body fracture identification: the main basis is that with the increase of the structural curvature of the stratum, the stress concentration phenomenon is intensified, and the stratum rupture degree is increased accordingly, so the structural curvature method can be used to predict the planar distribution characteristics of the fault.

[0007] (4) Ant body fracture identification: the principle is to set a large number of electronic "ants" in the seismic volume, and let each "ant" move forward along the possible fault plane while emitting "pheromone". The "ants" moving along the fault can track the fault plane. If it meets the expected fault, it will make a very obvious mark with "pheromone", and it will not make a mark or only make a less obvious mark for those planes that are not likely to be faults.

[0008] In summary, the above methods can realize the identification of fractures, but for different types and different scales of fractures, especially small-scale buried fractures, the above methods are difficult to accurately identify the fractures, and the fracture identification effect is poor. SUMMARY

[0009] The application aims to provide a fracture identification method based on post-stack seismic data to solve the problem of poor identification effect of small-scale fractures by a single identification method.

[0010] The application provides a fracture identification method based on post-stack seismic data to solve the above technical problems, which comprises the following steps:

[0011] 1) Obtain post-stack seismic data, and perform dip imaging enhancement processing on the post-stack seismic data;

[0012] 2) Perform intrinsic coherence calculation on the seismic data subjected to the dip imaging enhancement processing to obtain intrinsic coherence volume data;

[0013] 3) Perform linear enhancement on each time slice of the intrinsic coherence volume data to enhance the linear features of the horizontal direction slice of the intrinsic coherence volume data;

[0014] 4) Perform ant body calculation on the filtered intrinsic coherence volume data, and perform fracture tracking through ant search to realize the identification of fractures.

[0015] The application can suppress random noise, retain original boundary information, and continuously enhance fracture information and suppress non-fracture information by using intrinsic coherence, fault enhancement and ant tracking technology, so as to obtain a data volume with low noise and clear fracture traces. The method is suitable for the identification and interpretation of fractures of different scales under different geological conditions, and can effectively identify small and medium-sized fractures by deeply mining the information contained in seismic data, thereby improving the accuracy of fracture identification.

[0016] Further, in order to adapt to different terrains and better suppress noise, the dip angle is adjustable in the dip imaging enhancement processing in step 1).

[0017] Further, in order to obtain higher coherence volume, the intrinsic coherence calculation in step 2) adopts a coherence algorithm based on the eigenvalues of the covariance matrix of multi-channel data.

[0018] Further, the process of the intrinsic coherence calculation in step 2) is as follows:

[0019] Extract multi-channel seismic data from a given analysis time window to generate a sample vector matrix;

[0020] Calculate the covariance matrix of the sample vector matrix, and calculate the eigenvalues and eigenvectors of the covariance matrix, find the largest eigenvalue, and calculate the coherence value.

[0021] Further, the analysis time window comprises a vertical time window and a lateral time window, the vertical time window refers to the time window size when a certain point coherence calculation is performed; the lateral time window refers to the number of surrounding seismic traces selected when the coherence calculation is performed.

[0022] Further, in order to remove the influence of non-structural factors, the step 3) adopts a linear filter to filter the intrinsic coherence volume data before linear enhancement is performed on each time slice of the intrinsic coherence volume data, so as to remove the stripe noise therein.

[0023] Further, the number of sample points of the ant boundary in the ant volume calculation process of the step 4) is 3-7.

[0024] Further, the number of sample points of the ant tracking deviation in the ant volume calculation process of the step 4) is 0-3.

[0025] Further, the number of sample points of the ant step length set in the ant volume calculation process of the step 4) is 2-10.

[0026] Further, the number of allowed illegal steps set in the ant volume calculation process of the step 4) is 0-3, and the number of must legal steps is 0-3. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 is a post-stack fracture fine identification method flow chart in the method embodiment of the present application;

[0028] Figure 2 is a profile comparison chart before and after the dip imaging enhancement processing of the step one in the method embodiment of the present application;

[0029] Figure 3 is a slice chart of the layer intrinsic coherence volume in the step two in the method embodiment of the present application;

[0030] Figure 4 is a slice chart of the layer fault enhancement volume in the step three in the method embodiment of the present application;

[0031] Figure 5 is a slice chart of the layer ant volume in the step four in the method embodiment of the present application;

[0032] Figure 6 is a fracture fine identification and interpretation profile chart in the method embodiment of the present application. DETAILED DESCRIPTION

[0033] The specific implementation of the present application will be further described below in combination with the drawings.

[0034] Through the dip imaging enhancement processing, the random noise is suppressed, the original boundary information is reserved, the fracture information is continuously enhanced by using the intrinsic coherence, the fault enhancement and the ant tracking technology, and the non-fracture information is suppressed, so that a data volume with low noise and clear fracture trace is obtained. The implementation process of the method is as shown in Figure 1 The following will be described in combination with a specific seismic data, and the specific implementation process of the method is as follows.

[0035] 1. The post-stack seismic data is subjected to the dip imaging enhancement processing.

[0036] The post-stack seismic data is processed by using the structure-oriented filtering based on the dip control, the original boundary information is reserved, the random noise is suppressed, and the continuity of the seismic events in the seismic image is enhanced, so that the processed seismic data and the attributes are more conducive to the identification and interpretation of the fractures.

[0037] The dip control enhancement technology calculates the similarity of adjacent traces by using the changes of the dip and the azimuth, improves the seismic lateral signal-to-noise ratio, and since the dip and the azimuth are calculated, the ability to depict the faults is obviously enhanced, which is conducive to the subsequent research on the faults and the fractures. The dip control enhancement module can improve the seismic data quality to facilitate better horizon and fault interpretation. In the calculation process, the dip change of each sampling point is fully considered, and the comparison time window is selected according to the center of the calculation sampling point. Each sampling point involved in the calculation is defined as a corresponding calculation weight. In order to ensure that all the results are positive values, each correlation initial value is defined as 1, and a perfect correlation will produce a weight of 2. For the weighted average, the amplitude and the weight of each dip control sampling point are calculated first, and then the sampling points of all the comparison time windows are averaged and processed.

[0038] The two parameters of the lateral smoothing window and the vertical time window are crucial to the result of the dip control imaging enhancement processing, and need to be tested according to the actual data. Generally, a larger smoothing window and a larger vertical time window can obtain better noise reduction effect, and the values are generally 3-5. On the contrary, smaller parameters have weaker noise reduction effect, and the default parameter 1 can be generally used. For this embodiment, the result of the dip imaging enhancement processing on the original post-stack seismic data is as shown in Figure 2 The original boundary is more prominent, the signal-to-noise ratio is better, and the continuity of the seismic events is better, which provides a good data basis for the subsequent fracture identification.

[0039] 2. The intrinsic coherence volume is calculated for the seismic data subjected to the dip imaging enhancement processing.

[0040] The eigenvalue-based coherence algorithm is used to improve the lateral resolution of the fault while ensuring the anti-noise ability. In the algorithm analysis, first, the sample point vectors are extracted from the given analysis window to generate the matrix, and then the eigenvalues and eigenvectors of the covariance matrix are calculated to obtain the ellipse composed of the eigenvectors of the covariance matrix. The maximum eigenvalue is calculated to obtain the eigen-coherence volume of the seismic data.

[0041] Eigen-coherence can highlight the unrelated abnormal information. When there is a specific geological phenomenon, such as fault, joint, large-scale cleavage, stratigraphic pinch-out, lenticular sand body, fracture, etc., the similarity will be destroyed, which is manifested as the non-similarity of edge mutation. By plotting the longitudinal and transverse continuity of the similar or non-similar parts, the specific geological targets related to specific sedimentary environment, sedimentation and tectonic action that cannot be continuously identified by artificial can be distinguished. Figure 3 As shown in the figure, the unrelated abnormal information on the coherence slice has multiple solutions, part of which is the response of the fracture, and part of which is the response of the sinkhole, curved river and other specific geological bodies, which requires subsequent attribute discrimination.

[0042] When performing eigen-coherence calculation, the dip angle parameter needs to be set. The dip angle value has a greater impact on the results. If the dip angle value is too small, the detailed change characteristics can be highlighted, and the macroscopic view will be more chaotic. For steeply dipping strata, the horizon can be processed into discontinuous fault characteristics. The larger the dip angle value, the longer the time required for calculation. The definition of dip angle includes No. Dip Bins Multiplier and minimum and maximum dip angle parameters, both of which will affect the results. Therefore, when specifying the dip angle, both need to be considered. Generally, the default parameters of No. Dip Bins Multiplier are recommended first, and only the maximum and minimum dip angle parameters are adjusted. The definition unit of the dip angle value is ms / trace, not degree, so it can be measured from the seismic profile. Generally, a smaller dip angle can highlight the detailed characteristics, and a larger dip angle can obtain smooth results.

[0043] When performing eigen-coherence calculation, the analysis window includes vertical aperture and spatial aperture. The vertical aperture is the size of the time window when performing coherence calculation at a certain point. A smaller time window will have a more chaotic result, and the finer the detailed characteristics are described, the more noise there will be, so it is difficult to clearly see the macroscopic law. A larger time window will clearly reflect larger fractures. The spatial aperture refers to how many traces are used for coherence calculation and the optimal value is taken. The more traces, the smoother the results.

[0044] When performing intrinsic coherence calculations, a scaling parameter needs to be set: coherence output values ​​are typically between 0 and 1.0, where 1.0 represents the same trajectory. The Scaling function allows data to be linearly scaled from -128.0 to +127.0, eliminating the need to readjust the data when loading it into the interpretation system. Furthermore, typical values ​​are between 0.5 and 1.0, and the result can be "debiased" to maintain the optimal dynamic range. For example, 0.5 corresponds to -128, 1.0 to 127.0, and values ​​below 0.5 will be clipped to -128.0. This parameter is generally used at its default value during calculations. If a different range of coherence values ​​is required, such as setting the coherence value range to -1 to +1, then the corresponding -128 and 127 values ​​should be changed to -1 and +1.

[0045] When performing intrinsic coherence calculations, adaptive processing parameters also need to be set: the adaptive processing parameter options are high, medium, low, and no adaptive processing calculation. High parameters allow for strong noise reduction while maintaining important low coherence characteristics, while low parameters have weaker noise reduction capabilities. By default, there is no adaptive processing. Select the "Apply Adaptive Processing" checkbox to enable parameter selection. To minimize noise, select high adaptive smoothing and high fitness rate; the adaptive sampling parameters rarely change their default values. Higher values ​​make it more sensitive to tilt characteristics, but the results are usually noisier.

[0046] 3. Based on the obtained intrinsic coherence volume, fault enhancement processing is performed on the seismic data.

[0047] The resulting coherence volume typically contains stripes in both vertical and horizontal directions. These are non-structural anomalies generated during coherence volume calculation and need to be filtered out using a linear filter to remove the influence of non-structural factors. Then, linear enhancement is applied to each time slice in the coherence volume to strengthen the linear characteristics of the horizontal slices, producing a linearly enhanced data volume, thereby enhancing the planar features of the fault.

[0048] In this embodiment, the result after fault enhancement processing is as follows: Figure 4 As shown, linear anomalies are enhanced, which can be considered as responses to fractures; while nonlinear anomalies, such as the responses to sinkholes and meandering river channels, are suppressed.

[0049] 4. Perform ant-body calculations on the fault-enhanced seismic data.

[0050] A large number of "ants" are scattered in the seismic data volume processed above, and the "ants" that find the fracture traces meeting the preset fracture conditions in the seismic attribute volume will "release" a certain signal to call other "ants" in the region to concentrate on the fracture for tracking until the tracking and identification of the fracture are completed, and other fracture traces not meeting the fracture conditions will not be marked. The "ant tracking" algorithm can further enhance the fracture information and suppress the non-fracture information on the basis of the fault-enhanced attribute volume. The algorithm is called ant tracking because it simulates the behavior of natural ants marking the crawling track in the process of searching for food paths to optimize the search path. The "artificial ants" are put into the seismic volume as seed points to search for fractures during calculation, and the "ant colony" captures the fracture information, and the fracture response of the obtained attribute volume is clearer. The present application needs to set related parameters when calculating the ant volume, and the related parameters include the ant boundary, the ant tracking deviation, the ant step length, the number of allowed illegal steps, the number of required legal steps, and the termination condition. The specific meanings and setting methods of the parameters are as follows:

[0051] (1) Ant boundary (sample point number 1-30): This parameter is used as the control radius (defined by the sample point number) of each "ant", which determines the initial distribution state of the "ant colony". Since this parameter defines the total number of "ants" in the data volume, it has a great influence on the calculation time. For tracking large regional faults, the parameter should be larger (5-7 sample points). For tracking small fractures and cracks, it is recommended to use 3-4 sample points. In general, the parameter is less than 3, which has no practical significance, and if it is less than 3, it will cause multiple ants to track the same fracture, without increasing more details. This parameter does not mean that the "ant colony" appears in the data volume at the same time, but is used to ensure that the initial position of each ant searching for a local maximum does not overlap with the control range of other ants. The ant boundary is defined by the sample point radius, and if a "ant" cannot find a local maximum or calculate the direction within the radius, the "ant" will die.

[0052] (2) Ant tracking deviation (sample point number 0-3): This parameter controls the maximum allowed deviation of the local maximum value during tracking, and the maximum deviation from the initial direction is 15°. The algorithm allows the ant to accept the local maximum points on both sides of the predicted direction node, and the tracking deviation parameter will be considered if the distance from the maximum point exceeds the tracking step length. If the deviation is too large, the ant will not be able to continue tracking. If the parameter is 1, it means that the ant is allowed to search for a local maximum within 1 sample point on both sides of the position point. If the maximum value cannot be searched, an illegal step will be recorded. If the maximum value is searched, the current position point to the maximum point is a legal step.

[0053] (3) Ant step size (sample number 2-10): This parameter defines the search step size of the ant with sample number, which determines the single step length of each ant in searching for local maximum. Increasing this value will make each ant search further, but will reduce the accuracy.

[0054] (4) Number of allowed illegal steps (0-3): This parameter is how many steps of the ant step size cannot search for maximum. If an ant is in a valid position and cannot search for a maximum by taking one more step, it is called an illegal step. Using the initial estimated orientation, the ant can continue to take one more step. If there is still no search for a maximum, it is the second illegal step. At this time, if the parameter is set to 1, the ant will interrupt the search in this direction. If a valid sample is searched in the second step, the valid position and illegal position will be recorded (and whether the number of legal steps meets the condition, parameter 5). This parameter is how many consecutive illegal steps are allowed.

[0055] (5) Must legal step number (0-3): This parameter controls whether the illegal gap of the search result is connected, which is used in combination with the above-mentioned number of allowed illegal steps; the meaning of this parameter is the number of legal steps that must be included in the search path of each ant. For example, after two consecutive legal steps are searched by the ant, if the must legal step number parameter is set to 2 at this time, the search result is valid; if the must legal step number parameter is set to 3, and the next step is an illegal step, the broken tracking result will be invalid.

[0056] (6) Termination condition (0-50%): This parameter is the total number of illegal steps allowed by each ant during the tracking process; when the number of illegal steps reaches the parameter limit, the tracking of the ant stops.

[0057] For this embodiment, after the ant body calculation parameters are set in the above-mentioned manner, the ant body data obtained is as shown in Figure 5 On the slice of the layer ant body, the fracture information is rich and clear in hierarchy, and the fracture characteristics of different scales and different directions are clear, especially for the identification of small fractures.

[0058] Through the above-mentioned four steps, the fracture data identified by the present application is as shown in Figure 6 According to the fine fracture identification results, the flat section combined method is used to carry out fine fracture interpretation. The black solid line indicates the fracture with obvious fault throw, which is easy for manual interpretation, and the dotted line indicates the small fracture with non-obvious fault throw, which is difficult for manual interpretation, and through the fine fracture identification method, these small fractures can be effectively identified.

Claims

1. A method for fracture identification based on poststack seismic data, characterized in that, The identification method comprises the following steps: 1) obtaining post-stack seismic data, and performing dip imaging enhancement processing on the post-stack seismic data; the dip imaging enhancement processing adopts adjustable dip angle; 2) performing intrinsic coherence calculation on the seismic data subjected to the dip imaging enhancement processing to obtain intrinsic coherence volume data; The intrinsic coherence calculation adopts a coherence algorithm based on eigenvalues of a multi-channel data covariance matrix, and the process of the intrinsic coherence calculation is as follows: extracting multi-channel seismic data to generate a sample vector matrix from a given analysis time window; calculating a covariance matrix of the sample vector matrix, and calculating eigenvalues and eigenvectors of the covariance matrix, and finding the largest eigenvalue to calculate a coherence value; The analysis time window comprises a vertical time window and a horizontal time window, the vertical time window refers to the size of the time window when a point coherence calculation is performed, and the horizontal time window refers to the number of surrounding seismic channels selected when the coherence calculation is performed; 3) performing linear enhancement on each time slice of the intrinsic coherence volume data to enhance the linear feature of the horizontal direction slice of the intrinsic coherence volume data; before performing the linear enhancement on each time slice of the intrinsic coherence volume data, a linear filter is used to filter the intrinsic coherence volume data to remove stripe noise therein; 4) performing ant body calculation on the filtered intrinsic coherence volume data, and performing fracture tracking through ant search to realize fracture identification.

2. The method of claim 1, wherein, The number of sample points of the ant boundary in the ant body calculation process of the step 4) is 3-7.

3. The method of claim 1, wherein, The number of sample points of the ant tracking deviation in the ant body calculation process of the step 4) is 0-3.

4. The method of claim 1, wherein, The number of sample points of the ant step length set in the ant body calculation process of the step 4) is 2-10.

5. The method of claim 1, wherein, The number of allowed illegal steps set in the ant body calculation process of the step 4) is 0-3, and the number of must legal steps is 0-3.

Images