A finite seed point-based reservoir architecture unit automatic detection method

Abstract

The application discloses a reservoir configuration unit automatic detection method based on limited seed points, which comprises the following steps: S10, extracting three-dimensional seismic body attributes of the configuration unit and optimizing the attributes; S20, selecting a starting seed point for configuration unit tracking; S30, generating a tracking path according to the seed point; S40, automatically searching for a configuration unit boundary; S50, generating a single configuration unit body; S60, when a configuration unit body tracking is completed, inputting a seed point of a next configuration unit, and obtaining a next configuration unit body according to the same steps; after all configuration units are identified, a plurality of configuration unit bodies are obtained through post-processing. The application automatically identifies the spatial distribution of different grade reservoir configuration units according to three-dimensional seismic attributes based on a search path and gradient descent technology, accurately detects the boundaries of the configuration units, and realizes efficient and accurate and fine characterization of the reservoir configuration units.

Description

Technical Field

[0001] This invention relates to an automatic detection method for reservoir configuration units based on a finite number of seed points, belonging to the field of oil and gas field development geology technology. Background Technology

[0002] Reservoir configuration refers to the morphology, scale, orientation, and stacking relationships of reservoir constituent units at different levels. Methods and techniques based on Miall fluvial facies configuration analysis are effective means to reveal the distribution patterns of seepage barriers within reservoirs and improve oilfield development efficiency. With the advancement of oilfield development, increasingly higher demands are being placed on the detailed study of reservoir configuration. Well logging information allows for the detailed subdivision of well-point configuration units, while comparisons of inter-well configuration units are mostly based on speculative interpretations guided by sedimentary models. Three-dimensional seismic analysis is undoubtedly the most effective information for achieving detailed characterization of inter-well configuration units. In recent years, the integration of well and seismic analysis to interpret interfaces at different levels of configuration and predict the distribution of configuration units has received significant attention in the industry. While seismic profiles can often be used to manually interpret the boundaries of level 5 configuration units, spatial closure is difficult and the workload is enormous. For level 4-3 configuration units, because they are often below seismic resolution, it is difficult to generate stable responses on seismic data, making accurate interpretation of the spatial distribution of these configuration units challenging. Some scholars have proposed image recognition-based technologies that can automatically search for the center point of local configuration units. However, these technologies are mostly based on two-dimensional matrices and cannot achieve good three-dimensional closure. The acquired images do not have geological significance for reservoir configuration units, and therefore have little value for industrial application. Summary of the Invention

[0003] To overcome the problems in the prior art, this invention provides an automatic detection method for reservoir configuration units based on finite seed points. This invention automatically identifies the spatial distribution of reservoir configuration units of different levels based on search path and gradient descent techniques and three-dimensional seismic attributes, accurately detects the boundaries of each level of configuration unit, and achieves efficient, accurate and detailed characterization of reservoir configuration units.

[0004] The technical solution provided by this invention to solve the above-mentioned technical problems is: an automatic detection method for reservoir configuration units based on finite seed points, comprising the following steps:

[0005] Step S10: Extract the seismic three-dimensional volume properties of the configuration elements and optimize the properties;

[0006] Step S20: Select the starting seed point for configuration unit tracing;

[0007] Step S30: Generate a tracing path based on the seed points;

[0008] Step S40: Automatically search for configuration unit boundaries;

[0009] Based on the set of tracing paths for each seed point, for each tracing path, starting from the starting point, search for the boundary of the configuration unit point by point in the direction of the end point. If the boundary of the configuration unit is found, the search process of this tracing path is terminated, and all points from the starting point to the boundary point are classified as points within the configuration unit, and these point sets are stored.

[0010] Step S50: Generate a single configuration unit;

[0011] For a fixed seed point, after the above steps have traversed all other meshes of the model, the set of points belonging to the configuration unit on each search path can be obtained. Then, the set of points on each search path is merged into a set of points of a configuration unit, which is a single configuration unit body.

[0012] Step S60: After one configuration unit is tracked, the seed point of the next configuration unit is input, and the next configuration unit is obtained by following the same steps; after all configuration units are identified, multiple configuration units are obtained through post-processing.

[0013] A further technical solution is that the seismic three-dimensional volume attributes of the configuration unit extracted in step S10 are extracted from the three-dimensional data volume within the time window, which can be used to reveal seismic features that are difficult to identify in other profile images, or to intuitively observe the spatial changes of geological bodies.

[0014] A further technical solution is that, in step S10, the preferred criterion is that the closer to the center of the configuration unit, the higher the attribute value of the seismic three-dimensional volume attribute.

[0015] A further technical solution is that, in step S20, the local high value of the seismic attribute value is used as the starting seed point for tracing the configuration unit.

[0016] A further technical solution is that the specific process of step S30 is as follows: taking the input seed point grid as the starting point, traversing each grid in the seismic attribute volume of the configuration unit as the ending point, generating multiple signal tracking paths; wherein each tracking path is directional, the starting point is the grid where the seed point is located, and the ending point is a certain grid that has been traversed. At the same time, the spatial coordinates, attribute values ​​and gradient values ​​of the three-dimensional feature signal volume grids passed through will be stored on the tracking path.

[0017] A further technical solution is that the boundary identification of the configuration unit in step S40 has two termination cases:

[0018] The first case is that if there are no gradient inversion points on a certain search path from the starting point to the ending point, then the signal value cutoff value needs to be input at the same time as the spatial coordinates of the seed point. If there is no gradient inversion in the seismic signal value on this search path, but the signal value is lower than the signal value cutoff value, then the tracking is terminated in advance and the coordinates corresponding to the point are stored.

[0019] The second scenario involves searching for the first point on the tracking path where the gradient value is reversed, stopping the search, and using that point as the boundary point of the configuration unit on this tracking path, storing the coordinates corresponding to that point.

[0020] A further technical solution is that the post-processing in step S60 involves local correction through a smoothing algorithm and manual interaction.

[0021] The present invention has the following beneficial effects:

[0022] 1. It employs advanced algorithms, including search path and gradient descent, which are highly automated, accurate, and capable of 3D tracking;

[0023] 2. The human-computer interaction method of providing seed points guides the algorithm, overcoming the problem of unclear seismic response in interlayer structures and reducing uncertainty;

[0024] 3. This invention has good generalization ability and is applicable not only to fluvial reservoirs but also to the extraction of configurations in various sedimentary environments. Attached Figure Description

[0025] Figure 1 This is a flowchart of the configuration unit extraction algorithm proposed in this invention;

[0026] Figure 2 This is a schematic diagram of the configuration unit cell effect obtained by the present invention based on search path and gradient descent techniques;

[0027] Figure 3 This is a cross-sectional view of the final configuration unit obtained in this invention;

[0028] Figure 4 This is a three-dimensional display diagram of the final configuration unit obtained in this invention. Detailed Implementation

[0029] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] like Figure 1As shown, the present invention provides an automatic detection method for reservoir configuration units based on finite seed points, which specifically includes the following steps:

[0031] Step 1: Seismic attribute volume extraction and feature optimization of configuration units;

[0032] Seismic attribute extraction is the process of extracting reservoir-related information from seismic data volumes using mathematical methods. Seismic 3D volume attribute extraction involves extracting 3D attribute volumes within a suitable time window from the 3D data volume. This can be used to reveal seismic features that are difficult to identify from other profile images, or to visually observe spatial variations in geological bodies. The process of selecting attributes with better correlation to the features of the structural unit from among the extracted seismic attributes is called attribute optimization. The basic criterion for optimization is that the closer the attribute is to the center of a single structural unit, the higher its value.

[0033] Step 2: Select the configuration unit to trace the starting seed point;

[0034] Configurational units are geological units smaller than sedimentary microfacies, therefore their seismic responses are often weak, and conventional identification and analysis frequently rely on expert guidance. To ensure the quality of automatic configurational unit tracking, users need to input the coordinates of the starting seed point. The physical meaning of a seed point is its extreme point in a spatial scale. The coordinates of a seed point are often represented by its spatial location within the seismic signal volume. During use, the location of the seed point can be selected based on the variation characteristics of the seismic attribute profile signal, often using a local high value of the seismic attribute as the starting seed point for configurational unit tracking.

[0035] Step 3: Generate a tracing path based on the seed points;

[0036] For each search process of the boundary of the configuration element given a seed point, the input seed point grid is used as the starting point, and each grid in the seismic attribute volume of the configuration element is traversed as the ending point to generate multiple signal tracing paths.

[0037] Each tracking path is directional, starting from the grid where the seed point is located and ending at a certain grid that has been traversed. At the same time, the tracking path stores the spatial coordinates, attribute values ​​and gradient values ​​of the 3D feature signal volume grids it passes through.

[0038] Step 4: Automatically search for configuration unit boundaries;

[0039] Based on the set of tracing paths for each seed point, for each tracing path, starting from the starting point, search for the boundary of the configuration unit point by point in the direction of the ending point. If the boundary of the configuration unit is found, the search process of this tracing path is terminated. Then, all points from the starting point to the boundary point are classified as points within the configuration unit, and these point sets are stored.

[0040] There are two termination cases for boundary identification of configuration units, such as Figure 1 As shown:

[0041] The first case is that if there are no gradient inversion points on a certain search path from the starting point to the ending point, then the signal value cutoff value needs to be input at the same time as the spatial coordinates of the seed point. If there is no gradient inversion in the seismic signal value on this search path, but the signal value is lower than the signal value cutoff value, then the tracking is terminated in advance and the coordinates corresponding to the point are stored.

[0042] The second case is to search for the first point on the tracking path where the gradient value is reversed, stop the search, and use that point as the boundary point of the configuration unit on this tracking path, and store the coordinates of that point.

[0043] Step 5: Generate a single configuration unit;

[0044] For a fixed seed point, after the above steps have traversed all other meshes of the model, the set of points belonging to the configuration unit on each search path can be obtained. Then, the set of points on each search path is merged into a set of points of a configuration unit, and a single configuration unit body is obtained.

[0045] Step 6: After one configuration element is tracked, input the seed point for the next configuration element. The same steps will be followed to obtain the next configuration element. After all configuration elements have been identified, and after some post-processing, the cross-sectional configuration styles of multiple configuration elements are as follows: Figure 2 As shown, the 3D display effect is as follows: Figure 3 As shown;

[0046] On the other hand, due to the noise in the seismic data, the searched structural units may not conform to geological understanding, thus requiring post-processing. For individual singular points and points with abnormal signals that do not conform to geological understanding, local correction can be performed through smoothing algorithms and manual interaction.

[0047] Figure 1 The algorithm for extracting configuration units proposed in this invention follows this process: After the algorithm starts executing, the user inputs the coordinates of the configuration unit seed point in three-dimensional space and the three-dimensional feature signal volume. The algorithm automatically calculates the gradient volume of the feature signal volume based on the seed point. Based on this gradient volume, the algorithm traverses each sample point in the three-dimensional feature signal volume, calculating the path from the seed point to each sample point, resulting in multiple paths. Next, for each path, the algorithm searches and calculates the change pattern of the signal gradient volume. If a gradient reversal point is found, the position of the gradient reversal is recorded and stored (equivalent to the configuration unit boundary) in the reversed IJK path list. This same operation is performed on each path for each seed point, and iterative calculation yields a single configuration unit.

[0048] Figure 2 This diagram illustrates the effect of the configuration unit volume obtained by the present invention based on the search path and gradient descent techniques. During the initial execution of the algorithm, the user needs to input different configuration unit seed points sequentially. Inputting the seed points triggers automatic tracking of the configuration unit. In this example, five points with the strongest local signals on the configuration unit's feature signal volume are selected as seed points. The algorithm can then calculate the corresponding three-dimensional volume of the configuration unit.

[0049] Figure 3 This is a cross-sectional view of the final configuration unit obtained by this invention. By improving the quality of the seismic attribute feature signal of the configuration unit, optimizing the seed points of the configuration unit, and correcting with expert experience, a configuration unit that conforms to geological understanding can be obtained. The application effect of this invention on the actual river facies point dam profile is shown in the figure below. The configuration unit identification result conforms to geological understanding and has a high degree of matching with the feature signal of the configuration unit.

[0050] Figure 4 This is a three-dimensional display of the final configuration unit obtained by this invention. Based on the configuration unit calculated by the algorithm above, and further corrected by expert opinion, the three-dimensional bodies of different colors in the figure below represent different configuration units. The hollowed-out parts between configuration units belong to the configuration interfaces between configuration units.

[0051] This invention proposes a method for automatically identifying 3D structural elements based on the concepts of seed points, search paths, and gradient descent. Addressing the challenges of extracting structural elements from 3D seismic attributes in industry applications, the method utilizes seed point guidance, effectively integrating expert experience and reducing the uncertainty of identifying structural elements based on 3D seismic attributes. The search path technique overcomes the challenge of automatically tracking structural element boundaries in 3D space, and the gradient descent algorithm accurately identifies structural element boundaries, significantly improving work efficiency and accuracy.

[0052] The above description is not intended to limit the present invention in any way. Although the present invention has been disclosed through the above embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the present invention shall still fall within the scope of the present invention.

Claims

1. An automatic detection method for reservoir configuration units based on finite seed points, characterized in that, Includes the following steps: Step S10: Extract the seismic three-dimensional volume properties of the configuration elements and optimize the properties; Step S20: Select the starting seed point for configuration unit tracing; Step S30: Generate a tracing path based on the seed points; Step S40: Automatically search for configuration unit boundaries; Based on the set of tracing paths for each seed point, for each tracing path, starting from the starting point, search for the boundary of the configuration unit point by point in the direction of the end point. If the boundary of the configuration unit is found, the search process of this tracing path is terminated, and all points from the starting point to the boundary point are classified as points within the configuration unit, and these point sets are stored. In step S40, the boundary identification of the configuration unit has two termination scenarios: The first case is that if there are no gradient inversion points on a certain search path from the starting point to the ending point, then the signal value cutoff value needs to be input at the same time as the spatial coordinates of the seed point. If there is no gradient inversion in the seismic signal value on this search path, but the signal value is lower than the signal value cutoff value, then the tracking is terminated in advance and the coordinates corresponding to the point are stored. The second case is to search for the first point on the tracking path where the gradient value is reversed, stop the search, and use that point as the boundary point of the configuration unit on this tracking path, and store the coordinates of that point. Step S50: Generate a single configuration unit; For a fixed seed point, after the above steps have traversed all other meshes of the model, the set of points belonging to the configuration unit on each search path can be obtained. Then, the set of points on each search path is merged into a set of points of a configuration unit, which is a single configuration unit body. Step S60: After one configuration unit is tracked, the seed point of the next configuration unit is input, and the next configuration unit is obtained by following the same steps; after all configuration units are identified, multiple configuration units are obtained through post-processing.

2. The automatic detection method for reservoir configuration units based on finite seed points according to claim 1, characterized in that, The step S10 of extracting the seismic three-dimensional volume attributes of the configuration unit is to extract the three-dimensional attribute volume within the time window in the three-dimensional data volume. This can be used to reveal seismic features that are difficult to identify in other profile images, or to intuitively observe the spatial changes of geological bodies.

3. The automatic detection method for reservoir configuration units based on finite seed points according to claim 1, characterized in that, The preferred criterion in step S10 is that the closer to the center of the configuration unit, the higher the attribute value of the seismic three-dimensional volume property.

4. The automatic detection method for reservoir configuration units based on finite seed points according to claim 1, characterized in that, In step S20, the local high value of the seismic attribute value is used as the starting seed point for tracing the configuration element.

5. The automatic detection method for reservoir configuration units based on finite seed points according to claim 1, characterized in that, The specific process of step S30 is as follows: taking the input seed point grid as the starting point, traversing each grid in the seismic attribute volume of the configuration unit as the ending point, generating multiple signal tracking paths; each tracking path has a direction, the starting point is the grid where the seed point is located, and the ending point is a certain grid that has been traversed. At the same time, the tracking path will store the spatial coordinates, attribute values ​​and gradient values ​​of the three-dimensional feature signal volume grids that have been passed.

6. The automatic detection method for reservoir configuration units based on finite seed points according to claim 1, characterized in that, In step S60, the post-processing involves local correction using a smoothing algorithm and manual interaction.

Images