Synthetic Seismic Record Alignment and Calibration Method Based on Forced Dynamic Time Warping

Abstract

The present invention discloses a method for aligning and calibrating synthetic seismic records based on forced dynamic time warping, belonging to the exploration technology field, which includes the following steps: S1, data preprocessing and seismic wavelet extraction; S2, setting geological constraints and determining alignment and misalignment points; S3, generating synthetic seismic records by using reflection coefficients and seismic wavelets; S4, applying the forced dynamic time warping (FDTW) algorithm to calculate the optimal alignment path. By adopting the above method for aligning and calibrating synthetic seismic records based on forced dynamic time warping, the present invention can fully integrate prior geological information, effectively avoid matching deviations caused by abnormal seismic records, and greatly improve the alignment accuracy between synthetic records and actual records.

Description

Technical Field

[0001] The present invention relates to the field of exploration technologies, and in particular to a method for aligning and calibrating synthetic seismograms based on forced dynamic time warping. Background Art

[0002] A synthetic seismogram refers to a seismic response record generated by simulating through a geological model and the principle of seismic wave propagation, corresponding to actual seismic exploration data. Synthetic seismograms are crucial for establishing an accurate time-depth relationship between seismic data and well logging data. Their quality directly affects the accuracy of geological models and the prediction accuracy of oil and gas reservoirs.

[0003] In practical applications, due to the complexity of formation conditions, the isophase axes between synthetic seismograms and actual seismic records often have misalignments. Although the traditional Dynamic Time Warping (DTW) algorithm performs well in time series alignment, it faces some specific challenges when dealing with complex seismic record alignment. Especially when the seismic record contains key reflection layers or other geological features that require strict alignment or avoidance of alignment, the traditional DTW is difficult to effectively handle these complex prior constraint information, resulting in the alignment results not meeting the actual requirements in some cases.

[0004] To address this problem, the present invention proposes a method for aligning and calibrating synthetic seismograms based on Forced Dynamic Time Warping (FDTW). Summary of the Invention

[0005] The object of the present invention is to provide a method for aligning and calibrating synthetic seismograms based on forced dynamic time warping, which can fully integrate prior geological information, effectively avoid matching deviations caused by abnormal seismic records, and greatly improve the alignment accuracy between synthetic records and actual records.

[0006] To achieve the above object, the present invention provides a method for aligning and calibrating synthetic seismograms based on forced dynamic time warping, including the following steps:

[0007] S1. Data preprocessing and seismic wavelet extraction;

[0008] S2. Set geological constraints and determine alignment and non-alignment points;

[0009] S3. Generate synthetic seismograms using reflection coefficients and seismic wavelets;

[0010] S4. Apply the forced dynamic time warping FDTW algorithm to calculate the optimal alignment path.

[0011] Preferably, in S1, the conventional seismic processing technology is applied to the target well to preprocess the data, extract the seismic records of the wellside traces related to the target well, establish the preliminary correspondence with the seismic data, and extract the seismic wavelet or generate the Ricker wavelet therefrom.

[0012] Preferably, in S2, a geological model is established based on the geological characteristics and logging data. Combining the geological model and the logging data, the key reflection layers that need to be strictly aligned and the information of abnormal areas that need to be avoided in the actual seismic records are determined, forced alignment points and prohibited alignment points are set, and the constraint conditions for the alignment process are set on the basis of the established geological model.

[0013] Preferably, the forced pairs in S2 are a set of point pairs that must be strictly matched between the synthetic seismic record and the actual seismic record. During the path search process, the forced alignment points are used as the starting and ending points of the segments to ensure that the path must pass through the forced alignment points. The path is expressed as:

[0014] ;

[0015] where, is the number of forced alignment points, and represent the indices in the wellside trace sequence and the synthetic trace sequence respectively;

[0016] The path is decomposed into K + 1 sub-segments, and the optimal path is independently calculated for each sub-segment between the starting and ending points;

[0017] The prohibited alignment points are a set of point pairs that need to be completely excluded during the alignment process. For any non-permissible point pair , its corresponding cost value is set to infinity, that is:

[0018] .

[0019] Preferably, during the dynamic programming process, the prohibited alignment points will not be included in the optimal path, which specifically includes the following steps:

[0020] S2.1. Initialize the cumulative cost matrix and set all values to infinity to exclude the default path;

[0021] S2.2. Gradually fill the cost matrix according to the input data:

[0022] Among them, for the starting point, if it does not belong to , then calculate its local cost and assign a value;

[0023] For the first row and the first column, gradually accumulate the local cost while checking whether it belongs to ;

[0024] For other positions, the minimum cumulative cost is calculated through the dynamic programming formula, but the prohibited points in are skipped.

[0025] Preferably, in S3, the reflection coefficient in the geological model is convolved with the extracted seismic wavelet to generate a preliminary synthetic seismic record, and the reflection coefficient and convolution model parameters are continuously adjusted so that the synthetic record preliminarily reflects the actual seismic response.

[0026] Preferably, the specific steps in S4 are as follows:

[0027] S4.1. Initialize the parameters of the FDTW algorithm;

[0028] S4.2. Use the FDTW algorithm to align the synthetic seismic record and the actual seismic record segment by segment;

[0029] S4.3. Based on the preset forced alignment points and prohibited alignment points, adjust and update the dynamic time warping (DTW) cost matrix;

[0030] S4.4. Combine the generated alignment path, precisely match the synthetic seismic record with the actual seismic record, and generate the final alignment result dataset.

[0031] Therefore, the present invention adopts the above-mentioned method for aligning and calibrating synthetic seismic records based on forced dynamic time warping, which can fully integrate prior geological information, introduce forced alignment points and prohibited alignment points on the basis of the traditional DTW algorithm. Compared with the traditional method, the FDTW algorithm can ensure the accurate correspondence of key geological markers while effectively avoiding the matching deviation caused by abnormal seismic records, and greatly improves the alignment accuracy between the synthetic record and the actual record.

[0032] Next, through the drawings and embodiments, the technical solution of the present invention will be further described in detail. Description of the Drawings

[0033] Figure 1 is a flowchart of the method for aligning and calibrating synthetic seismic records based on forced dynamic time warping of the present invention;

[0034] Figure 2 is a flowchart of the FDTW algorithm subroutine of the method for aligning and calibrating synthetic seismic records based on forced dynamic time warping of the present invention;

[0035] Figure 3 is a schematic diagram of the FDTW algorithm cost matrix and the optimal path when the constraint conditions are opened in the method for aligning and calibrating synthetic seismic records based on forced dynamic time warping of the present invention;

[0036] Figure 4 It is a schematic diagram of the cost matrix and the optimal path of the FDTW algorithm when the constraint conditions of the synthetic seismic record alignment and calibration method based on forced dynamic time warping of the present invention are turned off;

[0037] Figure 5 It is a schematic diagram of the generation of synthetic seismic records of the synthetic seismic record alignment and calibration method based on forced dynamic time warping of the present invention;

[0038] Figure 6 It is a graph of the cost matrix and the optimal path of the synthetic seismic record alignment and calibration method based on forced dynamic time warping of the present invention;

[0039] Figure 7 It is a corresponding connection diagram of synthetic traces and well-side trace sequences of the synthetic seismic record alignment and calibration method based on forced dynamic time warping of the present invention;

[0040] Figure 8 It is a comparison chart of synthetic seismic records before and after alignment of the synthetic seismic record alignment and calibration method based on forced dynamic time warping of the present invention. Detailed implementation manners

[0041] The technical solutions of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0042] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings understood by those of ordinary skill in the field to which the present invention belongs.

[0043] Embodiment

[0044] As Figure 1 shown, the present invention provides a synthetic seismic record alignment and calibration method based on forced dynamic time warping. Taking the Dongying Formation in Block C of the Jizhong Depression as an example, the method includes the following steps:

[0045] S1. Data preprocessing and seismic wavelet extraction;

[0046] Apply conventional seismic processing techniques to preprocess the data for the target well to ensure data clarity and consistency. Data preprocessing includes but is not limited to steps such as denoising, static correction, removal of direct waves, and frequency filtering to ensure the quality and consistency of the data. Extract the well-side trace seismic records related to the target well, establish a preliminary correspondence with the seismic data, and extract the seismic wavelet or generate a Ricker wavelet therefrom to provide a basis for generating synthetic seismic records in the subsequent steps.

[0047] S2. Set geological constraints and determine alignment and non-alignment points;

[0048] A geological model is established based on geological features and logging data. Combining the geological model and logging data, key reflection horizons that need to be strictly aligned in the actual seismic record and information on abnormal areas or discontinuous areas caused by logging noise that need to be avoided are determined. Forced alignment points and prohibited alignment points are set. Based on the established geological model, constraint conditions for the alignment process are set. Geological constraint conditions include, but are not limited to, geological features such as formation interfaces, faults, and salt domes, which are determined through analysis of the geological model and logging data.

[0049] Forced alignment points (forcedpairs) are a set of point pairs that must be strictly matched between the synthetic seismic record and the actual seismic record. During the path search process, forced alignment points are used as the starting and ending points of segments to ensure that the path must pass through the forced alignment points. The path is represented as:

[0050] ;

[0051] where is the number of forced alignment points, and represent the indices in the well-side trace sequence and the synthetic trace sequence, respectively;

[0052] The path is decomposed into K + 1 sub-segments, and the optimal path is independently calculated for each sub-segment between the starting and ending points; as shown in Figure 3 and Figure 4 This segmented solution method ensures that the global path not only meets the continuity requirements of dynamic time warping but also strictly follows the forced alignment constraints.

[0053] Prohibited alignment points are a set of point pairs that need to be completely excluded during the alignment process, and its implementation is achieved by modifying the cumulative cost matrix (cost matrix). For any point pair that is not allowed to be aligned, its corresponding cost value is set to infinity, i.e.:

[0054] .

[0055] During the dynamic programming process, prohibited alignment points are not included in the optimal path, as shown in Figure 3 and Figure 4 to effectively avoid the path. The specific steps are as follows:

[0056] S2.1. Initialize the cumulative cost matrix and set all values to infinity to exclude the default path;

[0057] S2.2. Gradually fill the cost matrix according to the input data:

[0058] Among them, for the starting point, if it does not belong to , calculate its local cost and assign a value;

[0059] For the first row and the first column, gradually accumulate the local cost while checking whether it belongs to ;

[0060] For other positions, calculate the minimum cumulative cost through the dynamic programming formula, but skip in the prohibited points.

[0061] S3. Generate a synthetic seismic record using the reflection coefficient and the seismic wavelet;

[0062] Convolve the reflection coefficient in the geological model with the extracted seismic wavelet to generate a preliminary synthetic seismic record. Continuously adjust the reflection coefficient and the convolution model parameters to gradually optimize the matching effect between the two, so that the synthetic record can initially reflect the actual seismic response, laying a foundation for further alignment and calibration.

[0063] Figure 5 As shown, the schematic diagram of the synthetic seismic record generation in the synthetic seismic record alignment and calibration method according to the embodiment of the present invention shows how to generate a synthetic seismic record through the convolution of the reflection coefficient in the geological model and the extracted seismic wavelet.

[0064] S4. Apply the forced dynamic time warping (FDTW) algorithm to calculate the optimal alignment path; as Figure 2 shown, the specific steps are as follows:

[0065] S4.1. Initialize the FDTW algorithm parameters; including forced alignment points (forced_pairs), prohibited alignment points (must_not_pairs), and window size (window_size). These parameters ensure that the alignment path can achieve the optimal match while following the geological constraints.

[0066] S4.2. Use the FDTW algorithm to align the synthetic seismic record and the actual seismic record segment by segment; first, perform dynamic time warping (DTW) calculation on the path from the starting point to the first forced alignment point, then calculate the paths between each pair of forced alignment points independently in sequence, and finally align the path from the last forced alignment point to the end point.

[0067] S4.3. Adjust and update the dynamic time warping (DTW) cost matrix based on the preset forced alignment points and prohibited alignment points; by dynamically calculating the minimum cost of the path, the FDTW algorithm generates the optimal alignment path and outputs the sequence of matching coordinate points.

[0068] S4.4. Combine the generated alignment paths to precisely match the synthetic seismic record with the actual seismic record and generate the final aligned result dataset.

[0069] Figure 6 As shown, the cost matrix and the optimal path graph of the synthetic seismic record alignment and calibration method according to the embodiment of the present invention show the distribution of the cost matrix and the optimal path. This figure clearly presents the structure of the cost matrix and the selection process of the optimal path, indicating that the present invention can effectively integrate prior constraint conditions under complex geological conditions to achieve high-precision alignment.

[0070] Figure 7 As shown, the corresponding connection graph of the synthetic trace and the near-well trace sequence of the synthetic seismic record alignment and calibration method according to the embodiment of the present invention shows the corresponding connection relationship between the synthetic seismic record and the near-well trace seismic record. This figure clearly reflects how the alignment path precisely matches the key points in the two sequences, providing a reliable basis for the subsequent optimization of the geological model and time-depth calibration.

[0071] Figure 8 As shown, the comparison graph of the synthetic seismic record before and after alignment of the synthetic seismic record alignment and calibration method according to the embodiment of the present invention shows the comparison effect before and after the alignment of the synthetic seismic record. This figure clearly reflects how the alignment path is optimized after introducing the forced constraint conditions, enabling the key points of the synthetic seismic record to match more accurately with the actual seismic record, thereby significantly improving the alignment accuracy and the reliability of the geological model.

[0072] Therefore, by adopting the above synthetic seismic record alignment and calibration method based on forced dynamic time warping, the present invention can fully integrate prior geological information, effectively avoid the matching deviation caused by abnormal seismic records, and greatly improve the alignment accuracy between the synthetic record and the actual record. This method optimizes the alignment path of the seismic record by introducing prior forced alignment points and forbidden alignment points and other constraint conditions in the traditional DTW. This improvement not only effectively solves the alignment problem under complex geological conditions, but also improves the alignment accuracy and stability, thus providing more reliable support for the refinement of the geological model and the accuracy of oil and gas reservoir prediction.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements do not make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A synthetic seismic record alignment and calibration method based on forced dynamic time warping, characterized by: The following steps are involved: S1. Data preprocessing and seismic wavelet extraction; S2, setting geological constraints and determining alignment and misalignment points; Forced alignment points forcedpairs are a set of point pairs that must be strictly matched between synthetic seismic records and actual seismic records. During the path search process, the forced alignment points are used as the starting and ending points of the segments to ensure that the path must be forced to align with the points. The path is expressed as: ； in, To force the number of alignment points, and Represent the indexes in the wellside channel sequence and synthetic channel sequence respectively; path It is decomposed into K+1 sub-segments, and the optimal path between the starting point and the end point is calculated independently for each sub-segment; Disable snap points must_not_pairs is the set of point pairs that need to be completely excluded during the alignment process. For any point pair that is not allowed to be aligned , and its corresponding cost value Set to infinity, that is: ； During the dynamic planning process, the prohibited alignment points will not be included in the optimal path, which includes the following steps: S2.1, initialize the cumulative cost matrix and set all values ​​to infinity to exclude the default path; S2.

2. Fill the cost matrix step by step according to the input data: Among them, for the starting point, if it does not belong to must_not_pairs , then calculate its local cost and assign a value; For the first row and column, gradually accumulate the local costs and check whether it belongs to must_not_pairs ; For other locations, the minimum cumulative cost is calculated using the dynamic programming formula, but skipping must_not_pairs Prohibited points in S3, generating synthetic seismic records using reflection coefficients and seismic wavelets; S4, applying the forced dynamic time warping FDTW algorithm to calculate the optimal alignment path; The specific steps are as follows: S4.1, initialize FDTW algorithm parameters; S4.2, using the FDTW algorithm to align the synthetic seismic records and the actual seismic records segment by segment; S4.3, adjusting and updating the dynamic time warping DTW cost matrix based on the preset mandatory alignment points and prohibited alignment points; S4.

4. Combine the generated alignment paths to accurately match the synthetic seismic records with the actual seismic records and generate the final alignment result dataset.

2. The synthetic seismic record alignment and calibration method based on forced dynamic time warping according to claim 1, characterized in that: In S1, conventional seismic processing technology is applied to preprocess the data of the target well, extract the well side channel seismic records related to the target well, establish a preliminary correspondence with the seismic data, and extract seismic wavelets or generate Ricker wavelets from them.

3. The synthetic seismic record alignment and calibration method based on forced dynamic time warping according to claim 1, characterized in that: In S2, a geological model is established based on geological features and logging data. The key reflection layers that need to be strictly aligned and the abnormal area information that needs to be avoided in actual seismic records are determined by combining the geological model and logging data. Mandatory alignment points and prohibited alignment points are set. On the basis of the established geological model, constraint conditions for the alignment process are set.

4. The synthetic seismic record alignment and calibration method based on forced dynamic time warping according to claim 1, characterized in that: In S3, the reflection coefficient in the geological model is convolved with the extracted seismic wavelet to generate a preliminary synthetic seismic record. The reflection coefficient and convolution model parameters are continuously adjusted so that the synthetic record preliminarily reflects the actual seismic response.

Images