Geological model reconstruction method based on seismic data multi-horizon simultaneous tracking
By dividing the seismic data volume into three-dimensional computational units and using seed points and similarity thresholds for multi-level synchronous tracking, a high-resolution geological model is reconstructed, solving the problem of insufficient accuracy in seismic interpretation and modeling, and achieving efficient and refined geological modeling.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIAN RES INST OF CHINA COAL TECH & ENG GRP CORP
- Filing Date
- 2026-01-22
- Publication Date
- 2026-05-15
AI Technical Summary
Existing geological modeling methods suffer from insufficient accuracy and distortion in seismic interpretation and modeling, making it difficult to meet the needs of meter-level and decimeter-level geological modeling. Furthermore, fine-grained stratigraphic tracking in small areas is prone to topological errors such as interlayer crossover.
A geological model reconstruction method using multi-level simultaneous tracking of seismic data is adopted. This method divides the seismic data volume into three-dimensional computational units according to a preset surface size, uses seed points and similarity thresholds for multi-level simultaneous tracking, and combines dense point fitting surface method to reconstruct a high-resolution geological model.
It improves the accuracy and computational efficiency of geological models, solves the problem of geological model distortion, and realizes rapid linkage from interpretation to modeling, making it suitable for large-scale industrial use and promotion.
Smart Images
Figure CN122043618A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of geophysical exploration technology, specifically relating to a geological model reconstruction method based on multi-level simultaneous tracking of seismic data. Background Technology
[0002] In current mainstream geological modeling methods, the establishment of geometric geological models using seismic interpretation horizons is typically carried out in a linear and asynchronous manner. That is, after seismic interpretation is completed, relevant modeling work is carried out based on the interpretation data. First, seismic profiles are calibrated one by one using artificially synthesized records to determine the amplitude, phase, and other wave group characteristics of the target horizon. Then, the horizon is traced using both manual and machine-assisted methods, and finally, the geological model is constructed using the horizon information.
[0003] Constructing 3D geological models based on seismic data faces dual technical limitations in seismic interpretation and modeling methods. On the one hand, the interpretation accuracy of 10-meter and 100-meter seismic data differs significantly from the requirements of meter-level and decimeter-level geological modeling. Mainstream seismic interpretation methods are primarily designed for large sets of sedimentary stratigraphic boundaries with significant seismic wave group characteristics, and are easily affected by anomalous geological bodies such as collapse columns and faults, leading to tracking interruptions. Furthermore, tracking small-scale, fine-grained stratigraphic layers one by one is prone to topological errors such as interlayer cross-linking, making it difficult to meet the requirements of fine-grained modeling.
[0004] On the other hand, the geometric reconstruction representation of mainstream surface model methods is insufficient. The internal top and bottom interfaces rely on the parallelization assumption to perform proportional grid subdivision, and only use the top and bottom plate interface data interpreted by seismic analysis. It fails to effectively integrate the fine geometric sedimentary features contained in the seismic information, making it difficult to construct a watertight solid model, which leads to model distortion. Summary of the Invention
[0005] The purpose of this invention is to provide a geological model reconstruction method that simultaneously tracks multiple seismic data layers, in order to solve the problems of insufficient accuracy and distortion in existing geological models.
[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:
[0007] A geological model reconstruction method based on simultaneous tracking of multiple seismic data layers includes the following steps: Step 1: Based on the geological model obtained from the seismic exploration project, normalize the amplitude values of all seismic traces in the seismic model to obtain a normalized three-dimensional data volume. The three-dimensional data volume includes multiple main sequence stratigraphic layers, and each main sequence stratigraphic layer includes multiple secondary sequence stratigraphic layers that require further interpretation. The three-dimensional data volume is divided into multiple three-dimensional computational units along the main survey line and connecting survey line at fixed intervals; multiple three-dimensional computational units are distributed within each secondary sequence stratum, and each three-dimensional computational unit includes multiple seismic traces. Step 2: Arbitrarily select a secondary sequence stratigraphy, and assign multiple seed points along the main survey line and connecting survey line of the secondary sequence stratigraphy; ensure that each seed point falls on the seismic trace within the three-dimensional computational unit body of the secondary sequence stratigraphy. Select another secondary sequence stratigraphic stratum without a given seed point and repeat the above operation until every secondary sequence stratigraphic stratum has been traversed. Step 3: Arbitrarily select a primary sequence stratigraphy, and within the selected primary sequence stratigraphy, arbitrarily select a secondary sequence stratigraphy. For all 3D computational units with seed points in the selected secondary sequence stratigraphy, perform the following sub-steps, specifically including: Step 3.1: In the selected secondary sequence stratigraphy, arbitrarily select a three-dimensional computational unit with a seed point; Step 3.2: Use the seed points within the 3D computational unit as reference points; Pick the point with the highest similarity to the reference point on the seismic trace adjacent to the reference point, and use it as the similarity point; Step 3.3: Select a seismic trace adjacent to the seismic trace where the similar point is located, and use the similar point as the new reference point; Pick the point on the adjacent seismic trace that has the highest similarity to the new reference point, and use it as the new similar point; Step 3.4: Repeat the operation of step 3.3 until all seismic traces in the 3D computational unit where the seed point is located are traversed to obtain multiple reference points; Step 3.5: Use the dense point fitting surface method to construct a surface from multiple reference points; Step 3.6: Repeat the operations of steps 3.1-3.5 until every three-dimensional computational unit with a seed point in the secondary sequence stratigraphy selected in step 3.1 is traversed; Complete the interpretation of each three-dimensional computational unit with a seed point in the secondary sequence strata selected in step 3.1; Step 4, for the secondary sequence stratigraphy selected in Step 3, performs the following sub-steps, specifically including: Step 4.0: Set the similarity threshold; Step 4.1: Randomly select a three-dimensional computational unit with a seed point and use this three-dimensional computational unit as the reference three-dimensional computational unit; Step 4.2: Use the multi-channel stacking method to stack all seismic traces within the reference 3D calculation unit into a single seismic trace, which serves as the reference seismic trace. Step 4.3: Select each neighboring 3D computational unit of the reference 3D computational unit; Each seismic trace in a neighboring 3D computational unit is superimposed into a single seismic trace using a multi-trace stacking method, and these traces are denoted as neighboring seismic traces. Step 4.4: Calculate the similarity value between the reference seismic trace and each neighboring seismic trace, and compare it with the similarity threshold respectively; All nearby seismic traces with similarity values greater than the similarity threshold are considered as similar seismic traces, and the nearby three-dimensional computational units corresponding to the similar seismic traces are considered as similar three-dimensional computational units. Step 4.4: Take each similar 3D computational unit as a new reference 3D computational unit. For each new reference 3D computational unit, repeat the operations of steps 4.2-4.3 until all 3D computational units in the secondary sequence stratigraphy selected in step 3 have been traversed, and obtain all similar 3D computational units of the 3D computational unit with seed point selected in step 4.1. Step 4.5: Using the seed points in the 3D computational cell selected in Step 4.1 as reference points, repeat the operations of Steps 3.2-3.4 until every seismic trace in all similar 3D computational cell cells is traversed to obtain a surface composed of multiple reference points. Step 4.6: Select another 3D computational cell with a seed point and repeat the operations of steps 4.1-4.5 until all 3D computational cells with seed points in the secondary sequence stratigraphy selected in step 3 are traversed to complete the interpretation of the secondary sequence stratigraphy. Step 5: For the main sequence stratigraphy selected in Step 3, select another uninterpreted secondary sequence stratigraphy and repeat Steps 3-4 until all secondary sequence stratigraphy within the main sequence stratigraphy are traversed, thus completing the interpretation of the main sequence stratigraphy. Step 6: For the 3D data volume, select another uninterpreted main sequence stratigraphy and repeat steps 3-5 until all main sequence stratigraphy within the 3D data volume has been traversed, thus completing the interpretation of all main sequence stratigraphy within the 3D data volume and reconstructing the geological model corresponding to the 3D data volume.
[0008] The present invention also has the following feature L: Furthermore, the dense point fitting surface method described in step 3 employs B-Spline.
[0009] Furthermore, in step 3.2, when determining the point on an adjacent seismic trace with the highest similarity to the selected seed point, the following steps are specifically adopted: Step 3.2.1: Select multiple points at fixed intervals on adjacent seismic traces; Step 3.2.2: Calculate the similarity between the seed point and each point on the adjacent seismic trace, as shown in the following formula:
[0010] in, Represents a similarity function; Indicates the reference point; It represents any point on a seismic trace adjacent to the reference point; express and Similarity between them; This represents the natural exponential function, which in mathematics uses the natural constant. Exponential operations with base 0; Indicates the reference point The absolute value of the amplitude of the seismic trace; Point The absolute value of the amplitude of the seismic trace; This represents the standard deviation parameter; Representing the reference points and points The corresponding spatial vector; This indicates the calculation of the similarity of spatial vectors; Representing a spatial vector Similarity; Step 3.2.3: Compare each similarity value and pick the point with the highest similarity to the reference point on the seismic trace adjacent to the reference point, and use it as the similar point.
[0011] Furthermore, step 3.2.2 specifically uses the following formula to calculate the spatial vector. Similarity:
[0012] in, and Represent spatial vectors respectively The 2-norm.
[0013] Furthermore, the similarity threshold mentioned in step 4.0 is 0.6.
[0014] Compared with the prior art, the present invention has the following technical effects: (I) The geological model reconstruction method for simultaneous tracking of multiple seismic data layers in this invention first constructs a stratigraphic framework model based on existing seismic interpretation, dividing the seismic data volume into independent computational units according to a preset surface area scale. This divides the entire seismic data volume from Jiaotong University into multiple three-dimensional computational units, improving computational efficiency and fault tolerance, and also enabling the utilization of more subtle waveform changes in the seismic data. Using historical seismic interpretation data, secondary strata within a large stratigraphic set are analyzed based on the waveform characteristics of the seismic data. A multi-objective optimization algorithm is employed to achieve simultaneous tracking of multiple layers to reconstruct a high-resolution geological model. This method can extract secondary stratigraphic information contained in the seismic data, significantly improving the efficiency and accuracy of geological model building, and solving the problem of precision distortion in geological models.
[0015] (II) The geological model reconstruction method for simultaneous multi-layer tracking of seismic data in this invention selects several seed points with significant characteristics within the large-layer internal waveform features, and determines the seismic wave polarity corresponding to each small layer through the seed points. Subsequently, the correlation between adjacent seismic traces within the seismic trace unit is analyzed, and the marker points of each small layer are connected based on strong correlation to form a seismic interpretation layer within the computational unit. This layer is gradually expanded to the entire seismic data volume, and finally all small layers are output. Through iterative calculation, a refined small-layer geometric model reconstruction is achieved, realizing rapid linkage and iteration from interpretation to modeling, suitable for large-scale industrial use and promotion. Attached Figure Description
[0016] Figure 1 This is a process diagram of one embodiment of the present invention; Figure 2 This is a schematic diagram of a given seed point in one embodiment of the present invention; Figure 3 This is a comparison diagram of geological model cross-sections in one embodiment of the present invention; Figure 4 This is a comparison diagram of the cross-sectional reconstruction effects of the method of this invention and conventional surface modeling methods; Figure 5 This is a schematic diagram illustrating the calculation of similarity in one embodiment of the present invention. Detailed Implementation
[0017] It should be noted that all methods in this invention, unless otherwise specified, are methods known in the prior art. For example, the normalization method is a known method in the prior art.
[0018] The following are specific embodiments of the present invention. It should be noted that the present invention is not limited to the following specific embodiments. All equivalent modifications made based on the technical solutions of this application fall within the protection scope of the present invention.
[0019] A geological model reconstruction method based on simultaneous tracking of multiple seismic data layers includes the following steps: Step 1: Based on the geological model obtained from the seismic exploration project, normalize the amplitude values of all seismic traces in the seismic model to obtain a normalized three-dimensional data volume; wherein, the three-dimensional data volume includes multiple main sequence strata, and each main sequence stratum includes multiple secondary sequence strata that require further interpretation. In this step, the geological model is obtained directly through the pre-seismic exploration engineering process. Specifically, it consists of a large set of seismic traces arranged in a regular grid, reflecting the acoustic impedance differences of underground geological interfaces and forming the basis for stratigraphic imaging. This data volume is specifically coded on the plane with inline and cross-line (Xline) and their corresponding geodetic coordinate positions, and vertically indexed by time or depth.
[0020] Each coded location is called a seismic trace, and its corresponding vertical index is arranged according to a fixed sampling interval. For example, when the total index length is 2000 ms and the sampling interval is 2 ms, the trace will contain 1001 sampling points vertically. Each sampling point stores a floating-point or integer value to represent the amplitude value of the seismic wave propagation at the current time (or depth). Multiple seismic traces form a three-dimensional data volume in space. Seismic interpretation involves locating marker points using manual or automated methods based on the amplitude values of multiple seismic traces, i.e., the arrangement rules of seismic wave groups, connecting them into lines and surfaces, and comparing them with geological elements to understand subsurface geological phenomena (such as sequence stratigraphy, structures, etc.).
[0021] Normalization is a commonly used data preprocessing technique in existing technologies. It scales the data proportionally to make it fall into a specific standard range to facilitate subsequent numerical calculations and comparisons. In this step, since the amplitude variation range of each channel is large, all amplitude values are first normalized so that the subsequent automated methods will not be affected by excessively large or small absolute amplitude values. It is only necessary to obtain the normalized three-dimensional data volume and study the relative amplitude changes between channels in the normalized three-dimensional data volume.
[0022] A specific three-dimensional data volume diagram, such as Figure 1 As shown in S110.
[0023] Next, the three-dimensional data volume is divided into multiple three-dimensional calculation units along the main survey line and connecting survey line at fixed intervals; multiple three-dimensional calculation units are distributed within each secondary sequence stratum, and each three-dimensional calculation unit includes multiple seismic traces; In this embodiment, as Figure 1 As shown in S111, the normalized three-dimensional data volume is divided into 128x128 three-dimensional unit volumes according to the survey lines and connecting lines.
[0024] Step 2, as follows Figure 2 As shown, arbitrarily select a secondary sequence stratigraphy, and assign multiple seed points along the direction of the main survey line and connecting survey line of the secondary sequence stratigraphy; ensure that each seed point falls on the seismic trace within the three-dimensional computational cell of the secondary sequence stratigraphy; then select another secondary sequence stratigraphy without assigned seed points, and repeat the above operation until every secondary sequence stratigraphy has been traversed. Because seismic exploration teams deliver preliminary interpretation results of the seismic bodies (i.e., large-scale stratigraphic interpretation) along with the 3D data volume, they generally make a preliminary identification of strata with a large lateral range, good continuity of seismic wave groups, and clear overall boundaries.
[0025] Using seismic data to interpret the boundaries of the main sequence strata, such as the Permian, Triassic, and Jurassic, is called seismic interpretation of a large set of strata. However, the construction team generally does not conduct seismic interpretation of the secondary sequence strata of the Upper, Middle, and Lower Jurassic. These secondary sequence strata are the ones that need further interpretation in this embodiment.
[0026] This embodiment provides another method for determining secondary sequence stratigraphy that requires further interpretation. Specifically, when determining secondary sequence stratigraphy that requires further interpretation, the interpretation of the large set of stratigraphy is projected onto a three-dimensional data volume, and the secondary sequence stratigraphy that has relatively good wave group continuity and no interpretation is manually checked. These stratigraphy are then identified as secondary sequence stratigraphy that requires further interpretation.
[0027] Large-scale stratigraphic interpretations can be input into geological modeling software; in this embodiment, ModelGeo is used. Known modeling algorithms in the prior art, such as Kriging interpolation, are used to establish the corresponding geological model. Parameters such as range, sill value, and nugget value used during the Kriging interpolation process are recorded. The same applies to other modeling algorithms.
[0028] In this step, for ease of subsequent processing, this embodiment selects a main sequence stratigraphy from the three-dimensional data volume and numbers each secondary sequence stratigraphy within the main sequence stratigraphy according to the order of sedimentary deposition. Numbering these secondary sequence stratigraphy according to sedimentary patterns is a well-known practice in the field. For example, the Lower Jurassic strata can be numbered according to their sedimentary patterns from ancient times to the present, denoted as J1. 01 J1 02 J1 o3 In this context, J represents the Jurassic system, subscript 1 represents the lower Jurassic system, and superscripts 01-03 represent the secondary sequence strata that can be identified from ancient times to the present, which need to be interpreted using seismic data in subsequent steps.
[0029] When using the method of this embodiment, it is not limited to this form. Those skilled in the art can determine the secondary sequence strata that need further interpretation based on the actual working conditions or actual needs.
[0030] In this step, the act of assigning a seed point is a standard method in the field. The seismic waveform corresponding to the seed point location can represent the seismic wave characteristics of the secondary sequence, such as positive polarity, negative polarity, or zero phase. The planar location of the seed point is the same as that of the seismic trace, while its vertical location relies on the vertical marker of the seismic trace. It is a point in three-dimensional space of the secondary sequence, and this location has specified seismic waveform characteristics, such as wave crests, wave troughs, positive phase zero crossings, or negative phase zero crossings.
[0031] These secondary sequence strata are crucial for improving the resolution of geological models, but their reflection signals are often weak and lack continuity, making them susceptible to noise interference or confusion with the strong reflections of larger strata. Therefore, precise initial control is needed to guide the tracing.
[0032] Based on existing technologies, this embodiment further specifies multiple seed points in the directions of the main survey line and the connecting survey line. The purpose is to provide a small number of initial control points in two spatial directions that are absolutely accurate, feature-clear and standardized, to ensure that the direction is correct and the target is clear, thereby efficiently and accurately completing the fine layer extraction in the full three-dimensional space.
[0033] The reason this embodiment provides multiple seed points along the main survey line and connecting survey line is that selecting points on the main survey line profile (e.g., north-south) only controls the morphology of the thin layer in the north-south direction; selecting points on the connecting line profile (e.g., east-west) controls the morphology in the east-west direction. Only by combining both can the attitude (dip angle, dip direction) of the thin layer in three-dimensional space be uniquely determined. This is also a preliminary measure to avoid "closure errors." If points are only selected on the main survey line, the stratigraphic level tracked by the algorithm may not close in the connecting line direction, leading to contradictions. Placing points in both directions is equivalent to setting up landmarks at the intersections of meridians and parallels in the three-dimensional grid, greatly improving the consistency of global tracking.
[0034] Through the above operations, the seismic trace waveform corresponding to the seed point location can represent the seismic wave characteristics of the secondary sequence, such as positive polarity, negative polarity, or zero phase.
[0035] This step provides a small number of precise, well-defined, and standardized initial control points in two spatial directions to ensure that the direction is correct and the target is clear when building the model later, thereby efficiently and accurately completing the fine layer extraction in the full three-dimensional space.
[0036] Step 3: Randomly select a main sequence stratigraphy, and within the selected main sequence stratigraphy, randomly select a secondary sequence stratigraphy. Perform the following sub-steps, specifically including: Step 3.1: In the selected secondary sequence stratigraphy, arbitrarily select a three-dimensional computational unit with a seed point; Step 3.2: Use the seed points within the 3D computational unit as reference points; Pick the point with the highest similarity to the reference point on the seismic trace adjacent to the reference point, and use it as the similarity point; Step 3.3: Select the adjacent seismic traces of the similar point and use the similar point as the reference point; Pick the point on the adjacent seismic trace that has the highest similarity to the reference point, and use it as the similarity point; Step 3.4: Repeat the operation of step 3.3 until all seismic traces in the 3D computational unit where the seed point is located are traversed to obtain multiple reference points; Step 3.5: Use the dense point fitting surface method to construct a surface from multiple reference points; Step 3.6: Repeat the operations of steps 3.1-3.5 until every three-dimensional computational unit with a seed point in the secondary sequence stratigraphy selected in step 3.1 is traversed; Complete the interpretation of each three-dimensional computational unit with a seed point in the secondary sequence strata selected in step 3.1; It should be noted that the seismic trace where the seed point is located usually has two adjacent seismic traces. Therefore, in this embodiment, the points with the highest similarity to the selected seed point on the two adjacent seismic traces are determined and these two points are respectively regarded as similar points. Next, the similarity point is again designated as the reference point. The seismic trace containing this reference point will also have two adjacent seismic traces, but one of them is the seismic trace containing the seed point, and similarity calculation has already been performed. Therefore, when calculating the similarity of the seismic trace containing this reference point, only one adjacent seismic trace needs to be calculated.
[0037] like Figure 1 In S112 and S113, strongly similar nodes are linked to form the tracking results of each seismic trace within a three-dimensional computational unit with seed points; Step 3.5 employs a dense point fitting surface method, specifically the B-Spline dense point fitting surface method, to construct a surface from all reference points; Furthermore, when determining the point on an adjacent seismic trace that has the highest similarity to the selected seed point, the following steps are specifically adopted: Step 3.2.1: Select multiple points at fixed intervals on adjacent seismic traces; Step 3.2.2: Calculate the similarity between the seed point and each point on the adjacent seismic trace, as shown in the following formula:
[0038] in, Represents a similarity function; Indicates the reference point; It represents any point on a seismic trace adjacent to the reference point; express and Similarity between them; This represents the natural exponential function, which in mathematics uses the natural constant. Exponential operations with base 0; Indicates the reference point The absolute value of the amplitude of the seismic trace; Point The absolute value of the amplitude of the seismic trace; This represents the standard deviation parameter; Representing the reference points and points The corresponding spatial vector; This indicates the calculation of the similarity of spatial vectors; Representing a spatial vector Similarity; Step 3.2.3: Compare each similarity value and pick the point with the highest similarity to the reference point on the seismic trace adjacent to the reference point, and use it as the similar point.
[0039] Furthermore, step 3.2.2 specifically uses the following formula to calculate the spatial vector. Similarity:
[0040] in, and They represent The 2-norm.
[0041] Step 4, for the secondary sequence stratigraphy selected in Step 3, performs the following sub-steps, specifically including: Step 4.0: Set the similarity threshold; in this embodiment, the similarity threshold is set to 0.8. Step 4.1: Randomly select a three-dimensional computational unit with a seed point and use this three-dimensional computational unit as the reference three-dimensional computational unit; Step 4.2: Use the multi-channel stacking method to stack all seismic traces within the reference 3D calculation unit into a single seismic trace, which serves as the reference seismic trace. In this step, the multi-channel stacking method is a commonly used method in this field, and will not be described in detail here.
[0042] Step 4.3: Select each neighboring 3D computing unit cell of the reference 3D computing unit cell; in this step, the neighboring 3D computing unit cell, as the name suggests, refers to the 3D computing unit cell adjacent to the 3D computing unit cell where the seed point is located.
[0043] The seed point is located in a 3D computational unit with 8 faces, so there are usually 8 neighboring 3D computational units. Although the neighboring 3D computational units also have 8 faces and theoretically also have 8 neighboring 3D computational units, the specific number of neighboring 3D computational units can be determined based on the actual situation, since they may be located at the boundary of a secondary sequence stratum.
[0044] Each seismic trace in a neighboring 3D computational unit is superimposed into a single seismic trace using a multi-trace stacking method, and these traces are denoted as neighboring seismic traces. Step 4.4: Calculate the similarity value between the reference seismic trace and each neighboring seismic trace, and compare it with the similarity threshold respectively; All nearby seismic traces with similarity values greater than the similarity threshold are considered as similar seismic traces, and the nearby three-dimensional computational units corresponding to the similar seismic traces are considered as similar three-dimensional computational units. In this step, since the three-dimensional computational units have already been superimposed into a single seismic trace, the similarity calculation can be performed directly by calculating the similarity value between the superimposed seismic traces. The similarity calculation formula in step 3 can be directly applied.
[0045] Step 4.4: Take each similar 3D computational unit as a reference 3D computational unit. For each reference 3D computational unit, repeat the operations of steps 4.2-4.3 until all 3D computational units in the secondary sequence stratigraphy selected in step 3 have been traversed, and obtain all similar 3D computational units of the 3D computational unit with seed point selected in step 4.1. Step 4.5: Using the seed points selected in Step 4.1 within the 3D computational cell containing seed points as reference points, repeat steps 3.2-3.4 until every seismic trace within all similar 3D computational cell cells is traversed, obtaining a surface composed of multiple reference points; such as Figure 1 As shown in S113.
[0046] Step 4.6: Select another 3D computational cell with a seed point and repeat the operations of steps 4.1-4.5 until all 3D computational cells with seed points in the secondary sequence stratigraphy selected in step 3 are traversed to complete the interpretation of the secondary sequence stratigraphy. In this embodiment, the tracking data of each secondary sequence stratigraphy is synchronized to the sequence database in the geological modeling software. Database triggers are established using the database CDC tool to monitor for additions, deletions, and modifications to the interpreted data. For secondary sequence stratigraphy, when the interpreted data changes, a message queue is sent, and the geometric modeling process is repeated along with the main stratigraphic data, using the same modeling algorithm and parameters. This completes the reconstruction of the geological model. Because a new secondary sequence stratigraphic model is added to the main sequence stratigraphic model, the geometric model of the entire region is reconstructed, reflecting more geological information and richer internal patterns.
[0047] This embodiment employs a unit cell subdivision and multi-level sequence stratigraphic synchronous tracing method. Under the constraints of a regional stratigraphic framework model, multi-level data obtained through synchronous tracing are integrated through a database to reconstruct a high-resolution geometric geological model. This approach maintains the stability of the geometric geological model while significantly improving its accuracy.
[0048] like Figure 3 As shown in the diagram (in which secondary sequence strata are labeled as "small layers"), the geometric models before and after reconstruction show that the reconstruction model controlled by multiple small layers has significantly improved the accuracy of stratigraphic geometric features compared to a large stratigraphic model that only applies seismic interpretation.
[0049] Compared to conventional methods, the method in this embodiment not only overcomes the closure error problem caused by traditional layer-by-layer tracing methods, but also extracts more geological information from seismic data. The geometric model is more consistent with the characteristics of actual seismic data in terms of spatial rationality and geological regularity. The reconstructed geometric model better matches the actual morphology of the strata, providing a reliable foundation for subsequent statistical analysis of data and research on the lateral variation patterns of attributes in attribute modeling. Figure 4 As shown, the model profiles reconstructed by conventional methods and our method are compared. Traditional methods based on surface model scaling suffer from inconsistencies between the strike and dip of sub-layers and seismic characteristics, easily leading to step effect errors. Our method, by simultaneously tracking layers, can overcome the interruption effect of faults, resulting in a higher degree of consistency between the reconstructed geological model morphology and seismic data.
[0050] Here is a more specific example: like Figure 5As shown, under the constraint of the large set of strata T2~T3, the secondary sequence strata T2-1 is determined and traced. A red seed point is given, and the spatial coordinates of the seed point are (1068,642,-830). Among them, (1068,642) are the labels of the horizontal position survey line and the connecting survey line, which have a clear correspondence with the geodetic coordinates, and -830 is the vertical depth mark.
[0051] Select the seismic wave amplitude of the trace where the seed point is located. The seismic data matrix 'a' in the red matrix window is represented as follows: a= ; Taking the key adjacent trace location (1068, 642) as an example, the seismic data of this trace forms a data matrix b within the blue rectangular window: b= ; Using similarity function Calculate the correlation between two matrices. The value is taken as 0.4 to 1 times the root mean square amplitude of the data in the trace where the seed point is located, and then the correlation matrix is obtained: Cor= ; This leads to the position with the highest correlation: -832.
[0052] In summary, this embodiment utilizes historical seismic interpretation data to analyze secondary strata within a large stratigraphic suite based on seismic data waveform characteristics. A multi-objective optimization algorithm is employed to achieve simultaneous tracking of multiple strata, and the multi-strata data obtained from this simultaneous tracking are integrated into the modeling system to reconstruct a high-resolution geometric geological model. Using the method of this invention, secondary stratigraphic information contained in seismic data can be extracted, and the efficiency and accuracy of geometric modeling can be significantly improved.
Claims
1. A geological model reconstruction method based on multi-level simultaneous tracking of seismic data, characterized in that, Includes the following steps: Step 1: Based on the geological model obtained from the seismic exploration project, normalize the amplitude values of all seismic traces in the seismic model to obtain a normalized three-dimensional data volume. The three-dimensional data volume includes multiple main sequence stratigraphic layers, and each main sequence stratigraphic layer includes multiple secondary sequence stratigraphic layers that require further interpretation. The three-dimensional data volume is divided into multiple three-dimensional computational units along the main survey line and connecting survey line at fixed intervals; multiple three-dimensional computational units are distributed within each secondary sequence stratum, and each three-dimensional computational unit includes multiple seismic traces. Step 2: Arbitrarily select a secondary sequence stratigraphy, and assign multiple seed points along the main survey line and connecting survey line of the secondary sequence stratigraphy; ensure that each seed point falls on the seismic trace within the three-dimensional computational unit body of the secondary sequence stratigraphy. Select another secondary sequence stratigraphic stratum without a given seed point and repeat the above operation until every secondary sequence stratigraphic stratum has been traversed. Step 3: Arbitrarily select a primary sequence stratigraphy, and within the selected primary sequence stratigraphy, arbitrarily select a secondary sequence stratigraphy. For all 3D computational units with seed points in the selected secondary sequence stratigraphy, perform the following sub-steps, specifically including: Step 3.1: In the selected secondary sequence stratigraphy, arbitrarily select a three-dimensional computational unit with a seed point; Step 3.2: Use the seed points within the 3D computational unit as reference points; Pick the point with the highest similarity to the reference point on the seismic trace adjacent to the reference point, and use it as the similarity point; Step 3.3: Select a seismic trace adjacent to the seismic trace where the similar point is located, and use the similar point as the new reference point; Pick the point on the adjacent seismic trace that has the highest similarity to the new reference point, and use it as the new similar point; Step 3.4: Repeat the operation of step 3.3 until all seismic traces in the 3D computational unit where the seed point is located are traversed to obtain multiple reference points; Step 3.5: Use the dense point fitting surface method to construct a surface from multiple reference points; Step 3.6: Repeat the operations of steps 3.1-3.5 until every three-dimensional computational unit with a seed point in the secondary sequence stratigraphy selected in step 3.1 is traversed; Complete the interpretation of each three-dimensional computational unit with a seed point in the secondary sequence strata selected in step 3.1; Step 4, for the secondary sequence stratigraphy selected in Step 3, performs the following sub-steps, specifically including: Step 4.0: Set the similarity threshold; Step 4.1: Randomly select a three-dimensional computational unit with a seed point and use this three-dimensional computational unit as the reference three-dimensional computational unit; Step 4.2: Use the multi-channel stacking method to stack all seismic traces within the reference 3D calculation unit into a single seismic trace, which serves as the reference seismic trace. Step 4.3: Select each neighboring 3D computational unit of the reference 3D computational unit; Each seismic trace in a neighboring 3D computational unit is superimposed into a single seismic trace using a multi-trace stacking method, and these traces are denoted as neighboring seismic traces. Step 4.4: Calculate the similarity value between the reference seismic trace and each neighboring seismic trace, and compare it with the similarity threshold respectively; All nearby seismic traces with similarity values greater than the similarity threshold are considered as similar seismic traces, and the nearby three-dimensional computational units corresponding to the similar seismic traces are considered as similar three-dimensional computational units. Step 4.4: Take each similar 3D computational unit as a new reference 3D computational unit. For each new reference 3D computational unit, repeat the operations of steps 4.2-4.3 until all 3D computational units in the secondary sequence stratigraphy selected in step 3 have been traversed, and obtain all similar 3D computational units of the 3D computational unit with seed point selected in step 4.
1. Step 4.5: Using the seed points in the 3D computational cell selected in Step 4.1 as reference points, repeat the operations of Steps 3.2-3.4 until every seismic trace in all similar 3D computational cell cells is traversed to obtain a surface composed of multiple reference points. Step 4.6: Select another 3D computational cell with a seed point and repeat the operations of steps 4.1-4.5 until all 3D computational cells with seed points in the secondary sequence stratigraphy selected in step 3 are traversed to complete the interpretation of the secondary sequence stratigraphy. Step 5: For the main sequence stratigraphy selected in Step 3, select another uninterpreted secondary sequence stratigraphy and repeat Steps 3-4 until all secondary sequence stratigraphy within the main sequence stratigraphy are traversed, thus completing the interpretation of the main sequence stratigraphy. Step 6: For the 3D data volume, select another uninterpreted main sequence stratigraphy and repeat steps 3-5 until all main sequence stratigraphy within the 3D data volume has been traversed, thus completing the interpretation of all main sequence stratigraphy within the 3D data volume and reconstructing the geological model corresponding to the 3D data volume.
2. The geological model reconstruction method based on multi-level simultaneous tracking of seismic data as described in claim 1, characterized in that, The dense point fitting surface method described in step 3 uses B-Spline.
3. The geological model reconstruction method based on multi-level simultaneous tracking of seismic data as described in claim 2, characterized in that, In step 3.2, when determining the point on an adjacent seismic trace with the highest similarity to the selected seed point, the following steps are specifically adopted: Step 3.2.1: Select multiple points at fixed intervals on adjacent seismic traces; Step 3.2.2: Calculate the similarity between the seed point and each point on the adjacent seismic trace, as shown in the following formula: in, Represents a similarity function; Indicates the reference point; It represents any point on a seismic trace adjacent to the reference point; express and Similarity between them; This represents the natural exponential function, which in mathematics uses the natural constant. Exponential operations with base 0; Indicates the reference point The absolute value of the amplitude of the seismic trace; Point The absolute value of the amplitude of the seismic trace; This represents the standard deviation parameter; Representing the reference points and points The corresponding spatial vector; This indicates the calculation of the similarity of spatial vectors; Representing a spatial vector Similarity; Step 3.2.3: Compare each similarity value and pick the point with the highest similarity to the reference point on the seismic trace adjacent to the reference point, and use it as the similar point.
4. The geological model reconstruction method based on multi-level simultaneous tracking of seismic data as described in claim 3, characterized in that, Step 3.2.2 specifically uses the following formula to calculate the spatial vector. Similarity: in, and Represent spatial vectors respectively The 2-norm.
5. The geological model reconstruction method based on multi-level simultaneous tracking of seismic data as described in claim 4, characterized in that, The similarity threshold mentioned in step 4.0 is 0.6.