Multi-well synthetic seismic record automatic calibration and structure interpretation and reservoir prediction method

Abstract

This invention provides an automatic calibration, structural interpretation, and reservoir prediction method for multi-well synthetic seismic records, belonging to the technical field of multi-well synthetic seismic record calibration in geophysical exploration and development. The method is based on preprocessed seismic well bypass data and synthetic seismic records, employing a combination of a sliding variable-length short-time window and a dynamic time-curving algorithm, using the average velocity of multi-well seismic marker layers as the criterion to achieve automatic calibration of multi-well synthetic seismic records. This invention can be applied in geophysical exploration and development, effectively achieving automatic calibration of multi-well synthetic seismic records, automatically stretching and compressing synthetic seismic records appropriately, improving the efficiency and accuracy of synthetic seismic record calibration, combining single-well and multi-well synthetic seismic record calibration, improving consistency among multiple wells, and providing a better foundation for subsequent seismic interpretation and well-seismic integration work.

Description

Technical Field

[0001] This invention relates to the field of multi-well synthetic seismic record calibration technology in geophysical exploration and development. Specifically, it relates to an automatic calibration method for multi-well synthetic seismic records, a structural interpretation method for oil and gas reservoirs, and a reservoir prediction method for oil and gas resources. Background Technology

[0002] In geophysical exploration and development, synthetic seismic record calibration, as a common method for seismic horizon calibration, effectively connects geological, logging, and seismic data, enabling the mapping from geological horizons to seismic horizons and the conversion of geological information from the depth domain to the time domain. Existing synthetic seismic record calibration methods are generally based on the correlation between synthetic seismic records and well-side seismic trace data waveforms, primarily relying on manual and single-well comparisons. Because manual synthetic seismic record calibration is affected by factors such as the individual's understanding of the geological, logging, and seismic data of the work area, and their personal experience, a rapid, accurate, and automated synthetic seismic record calibration method is of great significance for subsequent structural interpretation and reservoir prediction.

[0003] Currently, the patent applications involving synthetic record calibration include:

[0004] 1. Chinese patent document “A Well-Seismic Calibration Method Based on Time-Frequency Continuous Wavelet Transform” (Publication No. CN106324681A, Publication Date January 11, 2017) describes an invention that uses well logging data to create synthetic seismic records. It employs a time-frequency continuous wavelet transform method to perform time-frequency transformations on both the synthetic and well-side seismic records, obtaining their time-frequency distribution maps. The time-frequency distribution maps are then adjusted to align their time-frequency axes, and the synthetic and / or well-side seismic records are simultaneously adjusted. Well-seismic calibration is then performed, and the reliability of the calibration results is determined by the correlation coefficients between the synthetic and well-side seismic records.

[0005] 2. Chinese patent document “Method, apparatus, system and computer-readable medium for calibrating synthetic seismic records” (publication number CN108037532A, publication date May 15, 2018) describes an invention that uses density and acoustic curves corrected for well logging environment to calculate the reflection coefficient of the corresponding strata; calculates the initial synthetic seismic record based on the wavelet dominant frequency; locks the position of the standard layer relative to the time scale; and calibrates the synthetic seismic record by adjusting the velocity attenuation coefficient when the wavelet dominant frequency is constant.

[0006] 3. Chinese patent document “A method and device for well seismic calibration” (publication number CN109932749A, publication date June 25, 2019) describes an invention that, for each well to be calculated, determines a first sampling point based on seismic interpretation results; determines the time-depth relationship of sampling points for each sampling point within the well to be calculated based on the first sampling point; summarizes the time-depth relationship of the calculated well for each sampling point based on the time-depth relationship of the sampling points for each sampling point; and performs well seismic calibration on multiple wells to be calculated within the study area based on the time-depth relationship of the calculated well for each well to be calculated.

[0007] 4. Chinese patent document “A rapid well seismic calibration method under dense well network conditions” (publication number CN110579806A, publication date December 17, 2019) describes an invention that uniformly selects 1 / 10 to 1 / 4 of the total number of wells in the study area to create single-well depth domain synthetic seismic records; adjusts the single-well depth domain synthetic seismic records to establish the correspondence between time and depth at a single well; converts the geological strata of a single well from the depth domain to the time domain to establish a time domain geological strata model; and automatically matches the remaining well strata in the study area to the time domain geological strata model, thereby achieving rapid well seismic calibration for all wells.

[0008] 5. Chinese patent document “A Well-Seismic Calibration Method” (publication number CN110673209A, publication date January 10, 2020) This invention selects top and bottom standard layers based on the seismic interpretation scheme, and performs calibration and quality control of the top and bottom standard layers respectively; using the calibration results of the top and bottom standard layers as constraints, the interlayer uses acoustic velocity instead of acoustic conversion to synthesize seismic record velocity, and uses average velocity fitting and trend surface removal of low-frequency effects to eliminate spatial errors, thereby obtaining accurate velocity fields between the top and bottom and standard layers; generating synthetic seismic records and carrying out well-seismic interlayer small-layer calibration.

[0009] 6. Chinese patent document “A Comprehensive Time-Depth Calibration Method and Device” (Publication No. CN111458751A, Publication Date July 28, 2020) describes an invention that compares formations between wells based on logging curves to obtain the fault location and displacement of each well fault point; performs well-seismic calibration based on wave resistance characteristics and synthetic seismic records to obtain time-depth calibration results; inputs the time-depth calibration results, fault location, and displacement into a seismic profile to determine the degree of agreement between each well fault point and the seismic fault point, and to determine the degree of agreement between the displacement of each well fault point and the displacement of the seismic hanging wall and footwall; and adjusts the synthetic seismic records based on wave resistance characteristics until the degree of agreement between each well fault point and the seismic fault point, as well as the degree of agreement between the displacement of the seismic hanging wall and footwall and the displacement of each well fault point, all reach preset values.

[0010] After analysis and research, the inventors stated that the existing technologies represented by the aforementioned patent documents can be roughly divided into three categories: First, introducing other information, such as VSP data, seismic interpretation results, well faults (distances) and seismic faults (distances), to improve the calibration effect of synthetic records; second, calibrating synthetic records from non-time domains (depth domains or frequency domains); and third, correcting or discarding traditional well logging sonic curves, creating new sonic velocity fields and new synthetic seismic records to complete the calibration of synthetic records. Generally speaking, the specific calibration process of synthetic seismic records and seismic signals still mainly relies on single-well and manual comparison. Because manual synthetic seismic record calibration is affected by factors such as the individual's understanding of the geological conditions of the work area, well logging and seismic data, and personal experience, the inventors intend to propose a rapid, accurate, and automatic method for calibrating multi-well synthetic seismic records, which will play a significant role in promoting subsequent structural interpretation and reservoir prediction. Summary of the Invention

[0011] The purpose of this invention is to address at least one of the aforementioned shortcomings of the prior art. For example, one objective of this invention is to provide a method for rapidly, accurately, and automatically calibrating multi-well synthetic seismic records.

[0012] To achieve the above objectives, the present invention provides an automatic calibration method for multi-well synthetic seismic records. This method is based on preprocessed seismic well bypass data and synthetic seismic records, and combines a sliding variable-length short time window with a dynamic time curvature algorithm. The average velocity of the multi-well seismic marker layer is used as the discrimination criterion to achieve automatic calibration of multi-well synthetic seismic records.

[0013] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the automatic calibration method for multi-well synthetic seismic records can be implemented through the following steps: Step 1, collect measured seismic data volumes within the work area, and select multiple drilled wells within the seismic data area according to a predetermined well network density to obtain logging data and geological information corresponding to the seismic marker layer M; Step 2, for the drilled wells, calculate the synthetic seismic record S of a single well using logging curves; Step 3, based on the plane coordinates of the drilled wells, determine the corresponding plane position of each single well in the measured seismic data volume, denoted as (inline, xline), and extract m×m=m centered on that trace. 2 Earthquake data, and take m 2 The average value of the seismic data is used as the seismic well bypass data R, where m is a natural number and not less than 1; Step 4: Resampling and normalization preprocessing of the seismic well bypass data R and the synthetic seismic record S are performed to achieve consistency in sampling rate and data dynamic range; Step 5: Using a sliding variable-length short time window, the seismic well bypass data R is truncated to obtain multiple local seismic data segments R. kStep 6: Calculate R for each seismic data segment using the dynamic time warp algorithm. k Step 7: Based on the matching distance with the synthetic seismic record S, compare and find the seismic data segment with the best matching effect to realize the single-well synthetic seismic record calibration, so as to obtain a new single-well time-depth relationship pair (h, t); Step 8: According to the multi-well matching results, use the depth value h0 of the seismic marker layer M to read the corresponding seismic time t0, calculate the average velocity V = 2 × h0 / t0, and statistically analyze the average velocity change of the seismic marker layer M; Step 9: If the average velocity difference between multiple wells is less than the velocity threshold, the automatic calibration of multiple wells is completed; If there are wells with significant average velocity differences, use the average velocity of multiple wells as the constraint basis to re-estimate the time value corresponding to the seismic marker layer, and use this as a benchmark to narrow the search time range of the well-side seismic data, and repeat steps 5 to 7 until the average velocity difference between multiple wells is less than the velocity threshold, then the automatic calibration of multiple wells is completed.

[0014] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the synthetic seismic record S may have a number of sample points L1.

[0015] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the local seismic data segment R k It can have a sample number L2, |L1-L2|≤a, where a is a given threshold value, representing the allowed range of stretching and compression of the synthetic record.

[0016] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, step 5 may include: setting the seismic data time [T1, T2] as the calibration search interval, designing a sliding variable-length short time window accordingly based on the length of the synthetic seismic record S, extracting seismic well bypass data R, and obtaining multiple local data segments R. k .

[0017] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, step 6 may include: using a dynamic time-warping algorithm to calculate Rs of the synthetic seismic record and each local seismic data segment respectively. k The best-matching seismic data segment R is determined by the minimum distance between the matching distances (Dist) and all calculated results. opt This achieves single-well synthetic seismic record calibration, obtaining new single-well time-depth relationship pairs (h, t); the matching distance Dist is calculated using the formula: Dist=D(i,j)+min{D(i-1,j-1),D(i-1,j),D(i,j-1)},1≤i≤L1,1≤j≤L2, where D(i,j) represents the sampling points S(i) and R in the two signals. k The distance between sampling points S(i-1) and R(j-1) in the two signals is given by D(i-1, j-1).k The distance between sampling points S(i-1) and R(j-1) is given by D(i-1,j), where D(i-1,j) represents the distance between sampling points S(i-1) and R(j-1) in the two signals. k The distance between sampling points S(i) and R(j) is given by D(i,j-1), where D(i,j-1) represents the distance between sampling points S(i) and R(j) in the two signals. k The distance is (j-1).

[0018] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the predetermined well network density may be greater than 4km × 4km.

[0019] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the number of drilled wells in step 1 may be greater than 10.

[0020] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, m can be 3 to 5.

[0021] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the velocity threshold can vary from 0 to 400 m / s.

[0022] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the velocity threshold may be [-200m / s, 200m / s].

[0023] Another aspect of the present invention provides a structural interpretation method for oil and gas reservoirs, which includes the automatic calibration method for multi-well synthetic seismic records as described above.

[0024] In another aspect, the present invention provides a reservoir prediction method for oil and gas resources, which includes the automatic calibration method for multi-well synthetic seismic records as described above.

[0025] Compared with the prior art, the beneficial effects of the present invention include at least one of the following:

[0026] (1) The automatic calibration method for multi-well synthetic seismic records proposed in this invention is an attempt at intelligent seismic interpretation technology. By designing a sliding variable-length short time window and a dynamic time bending algorithm, it can automatically search for the local data segment that best matches the synthetic seismic record, realize the calibration of single-well synthetic seismic records, and use the statistical analysis of the average velocity of multiple wells of the seismic marker layer as the criterion to screen out wells with poor calibration results, narrow the search range, and thus improve the calibration effect of synthetic seismic records. This method is a new automatic calibration scheme for multi-well synthetic seismic records.

[0027] (2) When applied to geophysical exploration and development, this invention can effectively realize the automatic calibration of multi-well synthetic seismic records, automatically realize the appropriate stretching and compression of synthetic seismic records, improve the efficiency and accuracy of synthetic seismic record calibration, realize the combination of single-well and multi-well synthetic seismic record calibration, improve the consistency between multiple wells, and provide a better foundation for subsequent seismic interpretation and well-seismic integration work.

[0028] (3) The calibration efficiency and accuracy of the present invention are far superior to the manual single-well synthetic record calibration method. Attached Figure Description

[0029] The above and other objects and / or features of the present invention will become clearer from the following description taken in conjunction with the accompanying drawings, in which:

[0030] Figure 1 A flowchart illustrating an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention is shown.

[0031] Figure 2A The illustration shows logging data and geological information of seismic marker layer M from one well in an exemplary embodiment of the multi-well synthetic seismic record automatic calibration method of the present invention. Figure 2B This illustration shows logging data and geological information of seismic marker layer M from another well in an exemplary embodiment of the multi-well synthetic seismic record automatic calibration method of the present invention.

[0032] Figure 3A This illustration shows seismic well bypass data from three wells in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention. Figure 3B The diagram shows synthetic seismic records of three wells in an exemplary embodiment of the multi-well synthetic seismic record automatic calibration method of the present invention.

[0033] Figure 4A A preprocessed synthetic seismic record map is shown in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention; Figure 4B A preprocessed local seismic data segment diagram is shown in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention; Figure 4C The diagram illustrates the dynamic time curvature calculation of two signals after preprocessing in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention.

[0034] Figure 5A The figure shown is an automatic calibration result of the synthetic seismic record of one well in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention; Figure 5BThe diagram shows the automatic calibration result of the synthetic seismic record of another well in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention.

[0035] Figure 6A This illustrates a histogram of statistical analysis of average velocity in the seismic standard layer of multiple wells in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention. Figure 6B A point plot of statistical analysis of average velocity of seismic standard layers from multiple wells is shown in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention.

[0036] Figure 7A and Figure 7B The images show the calibration diagrams of the synthetic seismic record before and after using the mean velocity of multiple wells as constrained in an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention. Detailed Implementation

[0037] The present invention will be described in detail below with reference to exemplary embodiments and accompanying drawings, including the automatic calibration method for multi-well synthetic seismic records, its structural interpretation method, and its reservoir prediction method.

[0038] In general, the key improvement of this invention is that after preprocessing the seismic well bypass data and synthetic seismic records, a sliding variable-length short time window is designed within the seismic data time range to extract local data segments from the seismic well bypass. The Dynamic Time Warping (DTW) algorithm is used to automatically search for the local data segment that best matches the synthetic seismic record, thereby achieving single-well synthetic seismic record calibration. The average velocity statistical analysis of multiple wells of the seismic marker layer is used as the evaluation criterion to screen wells with poor calibration results, narrow the search range, and further improve the calibration effect of synthetic seismic records.

[0039] To achieve the above objectives, the present invention provides an automatic calibration method for multi-well synthetic seismic records.

[0040] In an exemplary embodiment of the automatic calibration method for multi-well synthetic seismic records of the present invention, the automatic calibration method for multi-well synthetic seismic records is based on preprocessed seismic well bypass data and synthetic seismic records, and adopts a combination of sliding variable-length short time window and dynamic time bending (DTW) algorithm, using the average velocity of multi-well seismic marker layers as the discrimination criterion to realize the automatic calibration of multi-well synthetic seismic records.

[0041] Specifically, the automatic calibration method for multi-well synthetic seismic records can be achieved through the following steps.

[0042] Step 1: Collect measured seismic data volumes within the work area. Within the seismic data area, select multiple drilled wells according to the predetermined well network density to obtain logging data and geological information corresponding to the seismic marker layer M. The aforementioned predetermined well network density can be greater than 4km × 4km, and the number of drilled wells can be more than 10.

[0043] Step 2: For the selected drilled well, calculate the synthetic seismic record S of the single well using the well logging curve. Additionally, the sampling rate of the synthetic seismic record S can be set to 1 ms (or 2 ms).

[0044] Step 3: Based on the selected well plane coordinates, determine the corresponding plane position of each well in the measured seismic data volume, denoted as (inline, xline). Extract m×m = m centered on this trace (i.e., (inline, xline)). 2 Earthquake data, and take its (i.e., m) 2 The average value of the seismic well sidetrack data is used as the seismic well sidetrack data R, where m is a natural number and not less than 1. For example, m can be 3 to 5.

[0045] Step 4: Complete the resampling and normalization preprocessing of seismic well bypass data R and synthetic seismic records S to achieve consistency in sampling rate and data dynamic range.

[0046] The specific process of resampling preprocessing is as follows: analyze the sampling rate of the synthetic seismic record S and the seismic well bypass data R. If the two are inconsistent, the data with the smaller sampling rate is used as the standard for resampling.

[0047] The specific process of normalization preprocessing is as follows: Analyze the dynamic range of the synthetic seismic record S and the seismic well bypass data R. If the two are inconsistent, then take a single well as the benchmark and perform data normalization on the synthetic seismic record S and the seismic well bypass data R respectively.

[0048] Step 5: Using a sliding variable-length short time window, extract the seismic well bypass data R to obtain multiple local seismic data segments R. k .

[0049] Specifically, by setting the seismic data time [T1, T2] as the calibrated search interval, and designing a sliding variable-length short-time window according to the length of the synthetic seismic record S, and using this sliding variable-length short-time window to repeatedly extract seismic well bypass data R, multiple local data segments R can be obtained. k .

[0050] The synthetic seismic record S can have L1 sample points. Local seismic data segment R k It can have a sample number L2, |L1-L2|≤a, where a is a given threshold value representing the allowable range of stretching and compression of the synthetic record. For example, a can be 1 to 5.

[0051] Step 6: Calculate R for each seismic data segment using the dynamic time warp algorithm. k The matching distance with the synthetic seismic record S is compared to find the seismic data segment with the best matching effect, and the single-well synthetic seismic record is calibrated to obtain a new single-well time-depth relationship pair (h, t).

[0052] Specifically, the processing steps of the dynamic time warp algorithm include: using the matching distance calculation formula to calculate the synthetic seismic record S and the local seismic data segments R respectively. k The best-matching seismic data segment R is determined by the minimum distance between the matching distances (Dist) and all calculated results. opt This enables the calibration of single-well synthetic seismic records and the acquisition of new single-well time-depth relationship pairs (h, t).

[0053] The formula for calculating the matching distance Dist is: Dist=D(i,j)+min{D(i-1,j-1),D(i-1,j),D(i,j-1)},1≤i≤L1,1≤j≤L2.

[0054] In the formula, D(i,j) represents the sampling points S(i) and R(i) in the two signals. k The distance between sampling points S(i-1) and R(j-1) in the two signals is given by D(i-1, j-1). k The distance between sampling points S(i-1) and R(j-1) is given by D(i-1,j), where D(i-1,j) represents the distance between sampling points S(i-1) and R(j-1) in the two signals. k The distance between sampling points S(i) and R(j) is given by D(i,j-1), where D(i,j-1) represents the distance between sampling points S(i) and R(j) in the two signals. k The distance is (j-1).

[0055] Step 7: Based on the multi-well matching results, use the depth value h0 of the seismic marker layer M to read the corresponding seismic time t0, calculate the average velocity V = 2 × h0 / t0, and statistically analyze the changes in the average velocity of the seismic marker layer M.

[0056] In other words, based on the multi-well matching results, using the depth value h0 of the seismic marker layer M, the corresponding seismic time t0 in the time-depth relationship pair (h, t) of each single well can be read, and the average velocity V = 2 × h0 / t0 can be calculated to statistically analyze the changes in the average velocity of the seismic marker layer M. For example, the above statistical analysis can be achieved by calculating the mean, variance, and quantiles, and plotting histograms or scatter plots.

[0057] Step 8: If the average velocity difference between multiple wells is less than the velocity threshold, the automatic calibration of multiple wells is completed. If there are wells with significant average velocity differences, the average velocity of multiple wells is used as a constraint basis to re-estimate the time value corresponding to the seismic marker layer. Based on this, the search time range of the seismic data in the well-side channel is narrowed, and steps 5 to 7 are repeated until the average velocity difference between multiple wells is less than the velocity threshold. Then the automatic calibration of multiple wells is completed.

[0058] The velocity threshold can vary from 0 to 400 m / s. For example, the velocity threshold can be [-200 m / s, 200 m / s], [-100 m / s, 100 m / s], [-100 m / s, 300 m / s], etc.

[0059] Another aspect of the present invention provides a structural interpretation method for oil and gas reservoirs, which includes the automatic calibration method for multi-well synthetic seismic records as described above.

[0060] In another aspect, the present invention provides a reservoir prediction method for oil and gas resources, which includes the automatic calibration method for multi-well synthetic seismic records as described above.

[0061] To better understand the exemplary embodiments of the present invention described above, further explanation is provided below in conjunction with the accompanying drawings and specific examples.

[0062] like Figure 1 As shown, an automatic calibration method for multi-well synthetic seismic records includes the following steps:

[0063] Step S1: Collect seismic, well logging, and seismic marker layer information.

[0064] Specifically, the process involves collecting measured seismic data within the work area. Within this area, based on a specific well network density (5km x 5km), 12 drilled wells are selected to obtain logging data and geological information corresponding to the seismic marker layer M (characteristically strong impedance and stably developed within the region). For example, Figure 2A and Figure 2B The well logging data and geological information of seismic marker layer M collected from two wells are shown respectively. Figure 2A The first horizontal axis from left to right represents natural gamma (GR), measured in API; the second horizontal axis represents resistivity (RXO), measured in Ω·m; and the third horizontal axis represents density (DEN), measured in g / cm³. 3 The fourth horizontal axis represents the longitudinal wave velocity V. p The unit is m / s; the fifth horizontal axis represents porosity (POR), in %; the vertical axis represents depth (MD), in m.

[0065] Step S2: Seismic well bypass data extraction.

[0066] Specifically, based on the plane coordinates of the selected wells, their corresponding plane positions in the measured seismic data volume are determined and denoted as (inline, xline). 3×3=9 seismic data points centered on this position are extracted, and their average value is taken as the seismic well sidetrack data R.

[0067] For example, Figure 3A and Figure 3B The seismic well bypass data R and synthetic seismic records S of three of the wells are shown respectively; Figure 3A and Figure 3B The vertical axis represents time, in milliseconds; for emphasis, each data point is repeated 5 times. Figure 3A and Figure 3B It can be seen that the length of the synthetic seismic record S is much shorter than that of the seismic well bypass data R, and the temporal correspondence between the synthetic seismic record S and the seismic well bypass data R is uncertain.

[0068] It should be noted that, Figure 3A It is seismic well bypass data R. Figure 3B The composite seismic record S is calculated using well logging data. Both have time as their vertical axis, measured in milliseconds. However, the time of the seismic well-bypass data R has geological time significance, while the time of the composite seismic record S is currently related to the number of sampling points and velocity, and does not have specific geological time significance. Therefore, well-seismic comparison is required.

[0069] Step S3: Calculation of synthetic seismic records.

[0070] Specifically, for the selected drilled wells, the synthetic seismic record S of each well is calculated using the logging curve, and the sampling rate of the synthetic seismic record is set to 1ms (or 2ms).

[0071] Step S4: Preprocessing of seismic well bypass data and synthetic seismic records.

[0072] Specifically, the sampling rates of the synthetic seismic record S and the seismic well bypass data R are analyzed. If the two are inconsistent, the data with the smaller sampling rate is used as the standard for resampling.

[0073] The dynamic range of the synthetic seismic record S and the seismic well bypass data R is analyzed. If the two are inconsistent, the data of the synthetic seismic record S and the seismic well bypass data R are normalized separately based on a single well.

[0074] By completing the resampling and normalization preprocessing of seismic well bypass data R and synthetic seismic records S, consistency in sampling rate and data dynamic range is achieved.

[0075] Step S5: Automatic calibration of single-well synthetic seismic records.

[0076] Specifically, the automatic calibration process for single-well synthetic seismic records can be divided into the following steps:

[0077] (a) Set the seismic data time [T1, T2] as the calibrated search interval. Based on the length of the synthetic seismic record S (with L1 sample points), design a sliding variable-length short time window and repeatedly extract seismic well bypass data R to obtain multiple local data segments R. k (The number of sample points is L2, |L1-L2|≤a, where a is a given threshold value, representing the allowable range of stretching and compression of the synthetic record).

[0078] (b) Using the Dynamic Time Warp (DTW) algorithm, the synthetic seismic record S and the data segments R of each seismic data segment are calculated. k The matching distance Dist between the two is used to determine the seismic data segment R with the minimum distance. opt This enables the calibration of single-well synthetic seismic records and the acquisition of new single-well time-depth relationship pairs (h, t).

[0079] The formula for calculating the matching distance Dist is: Dist=D(i,j)+min{D(i-1,j-1),D(i-1,j),D(i,j-1)},1≤i≤L1,1≤j≤L2.

[0080] In the formula, D(i,j) represents the sampling points S(i) and R(i) in the two signals. k The distance between sampling points S(i-1) and R(j-1) in the two signals is given by D(i-1, j-1). k The distance between sampling points S(i-1) and R(j-1) is given by D(i-1,j), where D(i-1,j) represents the distance between sampling points S(i-1) and R(j-1) in the two signals. k The distance between sampling points S(i) and R(j) is given by D(i,j-1), where D(i,j-1) represents the distance between sampling points S(i) and R(j) in the two signals. k The distance is (j-1).

[0081] For example, Figure 4A , Figure 4B and Figure 4C The preprocessed synthetic seismic record S from one well and a local seismic data segment R are shown respectively. k And the dynamic time bending calculation diagrams of the two sets of signals, in which, Figure 4C Curve A in the diagram marks the synthetic seismic record S and the seismic data segment R. k The distance between them is calculated along the path. Curve B represents the synthetic seismic record, with its vertical axis representing the amplitude value (Amp). Curve C represents a local seismic data segment, with its vertical axis also representing the amplitude value (Amp). Figure 4A and Figure 4B It can be seen that the synthetic seismic record S and the seismic data segment R kWhen local sample points are stretched or compressed, the distance between sample points can be reduced, resulting in better matching of two data points.

[0082] Figure 5A and Figure 5B The automatically calibrated composite seismic records of two wells are shown in the figure. Figure 5A and Figure 5B The horizontal axis represents the amplitude value, which is dimensionless, and the vertical axis represents time, in milliseconds (ms). Figure 5A and Figure 5B The composite seismic record S (curve A), seismic data R (curve C), and optimal calibration segment R are shown respectively. opt (i.e., curve B), and displayed in an overlapping manner. (From...) Figure 5A and Figure 5B It can be seen that for these two wells, automatically finding the location of the wellside seismic trace data that best matches the synthetic seismic record S can make the time of the synthetic seismic record consistent with the time of the seismic data.

[0083] Step S6: Statistical analysis of average velocity of seismic marker layers in multiple wells.

[0084] Specifically, based on the time-depth relationship of each single well (h, t) and the depth value h0 of the seismic marker layer M, the corresponding seismic time t0 is read, and the average velocity V = 2 × h0 / t0 is calculated. The average velocity change of the seismic marker layer M is then statistically analyzed.

[0085] Then, based on the average velocity V of the seismic marker layer M in the drilled wells, a velocity interval of 400 m / s [-200, +200] is selected, and the mean V1 of the average velocity of all wells within this interval is calculated. If the average velocity of all drilled wells is within the selected velocity interval, the multi-well automatic calibration is completed; if there are wells with average velocity values ​​outside the selected velocity interval, the time t0' = 2 × h0 / t0 corresponding to the seismic marker layer M in these wells is re-estimated using the mean V1 of the average velocity and the depth h0 of the seismic marker layer M, and the calibration search interval is narrowed to [T1', T2'] based on time t0', and step S5 is repeated. Where T1' = t0' - L1, T2' = t0' + L1.

[0086] For example, Figure 6A The histogram of average velocity statistical analysis of seismic standard layer M in multiple wells is shown; Figure 6B A point plot showing the statistical analysis of the average velocity of seismic standard layer M in multiple wells is presented. Figure 6B The two dashed lines in the diagram represent the defined speed range. Figure 6A and Figure 6B It can be seen that the average velocity distribution of the seismic standard layer M in most wells is relatively concentrated, but there are some velocity anomalies.

[0087] Figure 7A and Figure 7B The diagrams show the synthetic seismic record calibration maps of the same well before and after constraining it using the average velocity value for a well whose average velocity in seismic standard layer M is outside the velocity range. Figure 7A and Figure 7B The horizontal axis represents the amplitude value, and the vertical axis represents the seismic data SEI. Figure 7A and Figure 7B Curve A represents the synthetic seismic record S, and curve B represents the optimal calibration segment R. opt Curve C represents the seismic data R. (From...) Figure 7A and Figure 7B It can be seen that after constraining by the average velocity value, the optimal calibration segment automatically selected from the seismic data is more reasonable and has better consistency with other wells.

[0088] This invention can be applied to geophysical exploration and development, effectively realizing automatic calibration of multi-well synthetic seismic records, automatically stretching and compressing synthetic seismic records appropriately, improving the efficiency and accuracy of synthetic seismic record calibration, combining single-well and multi-well synthetic seismic record calibration, improving consistency among multiple wells, and providing a better foundation for subsequent seismic interpretation and well-seismic integration work.

[0089] In summary, the beneficial effects of the present invention include at least one of the following:

[0090] (1) The automatic calibration method for multi-well synthetic seismic records proposed in this invention is an attempt at intelligent seismic interpretation technology. By designing a sliding variable-length short time window and a dynamic time bending algorithm, it can automatically search for the local data segment that best matches the synthetic seismic record, realize the calibration of single-well synthetic seismic records, and use the statistical analysis of the average velocity of multiple wells of the seismic marker layer as the criterion to screen out wells with poor calibration results, narrow the search range, and thus improve the calibration effect of synthetic seismic records. This method is a new automatic calibration scheme for multi-well synthetic seismic records.

[0091] (2) When applied to geophysical exploration and development, this invention can effectively realize the automatic calibration of multi-well synthetic seismic records, automatically realize the appropriate stretching and compression of synthetic seismic records, improve the efficiency and accuracy of synthetic seismic record calibration, realize the combination of single-well and multi-well synthetic seismic record calibration, improve the consistency between multiple wells, and provide a better foundation for subsequent seismic interpretation and well-seismic integration work.

[0092] (3) The calibration efficiency and accuracy of the present invention are far superior to the manual single-well synthetic record calibration method.

[0093] Although the present invention has been described above in conjunction with exemplary embodiments and accompanying drawings, those skilled in the art should understand that various modifications can be made to the above embodiments without departing from the spirit and scope of the claims.

Claims

1. An automatic calibration method for multi-well synthetic seismic records, characterized in that, The automatic calibration method for multi-well synthetic seismic records is based on preprocessed seismic well bypass data and synthetic seismic records. It combines a sliding variable-length short time window with a dynamic time bending algorithm and uses the average velocity of multi-well seismic marker layers as the discrimination criterion to achieve automatic calibration of multi-well synthetic seismic records. The automatic calibration method for multi-well synthetic seismic records is implemented through the following steps: Step 1: Collect measured seismic data within the work area. Within the area where the seismic data is located, select multiple drilled wells according to the predetermined well network density to obtain well logging data and geological information corresponding to the seismic marker layer M. Step 2: For the drilled wells, calculate the synthetic seismic record S of a single well using well logging curves; Step 3: Based on the planar coordinates of the drilled wells, determine the corresponding planar position of each well in the measured seismic data volume, denoted as (inline, xline), and extract m×m=m centered on that trace. 2 Earthquake data, and take m 2 The average value of the seismic data is taken as the seismic well bypass data R, where m is a natural number and not less than 1; Step 4: Complete the resampling and normalization preprocessing of seismic well bypass data R and synthetic seismic records S to achieve consistency in sampling rate and data dynamic range; Step 5: Using a sliding variable-length short time window, extract the seismic well bypass data R to obtain multiple local seismic data segments R. k ; Step 6: Calculate R for each seismic data segment using the dynamic time warp algorithm. k The matching distance with the synthetic seismic record S is compared to find the seismic data segment with the best matching effect, and the single-well synthetic seismic record is calibrated to obtain a new single-well time-depth relationship pair (h, t). Step 7: Based on the multi-well matching results, use the depth value h0 of the seismic marker layer M to read the corresponding seismic time t0, calculate the average velocity V=2×h0 / t0, and statistically analyze the change of the average velocity of the seismic marker layer M. Step 8: If the average velocity difference between multiple wells is less than the velocity threshold, the automatic calibration of multiple wells is completed. If there are wells with significant average velocity differences, the average velocity of multiple wells is used as a constraint basis to re-estimate the time value corresponding to the seismic marker layer. Based on this, the search time range of the seismic data in the well-side channel is narrowed, and steps 5 to 7 are repeated until the average velocity difference between multiple wells is less than the velocity threshold. Then the automatic calibration of multiple wells is completed.

2. The automatic calibration method for multi-well synthetic seismic records according to claim 1, characterized in that, The synthetic seismic record S has a sample number L1.

3. The automatic calibration method for multi-well synthetic seismic records according to claim 2, characterized in that, The local seismic data segment R k It has a sample number L2, |L1-L2|≤a, where a is a given threshold value, representing the allowed range of stretching and compression of the synthetic record.

4. The automatic calibration method for multi-well synthetic seismic records according to claim 3, characterized in that, Step 5 includes: setting the seismic data time [T1, T2] as the calibrated search interval, designing a sliding variable-length short time window according to the length of the synthetic seismic record S, extracting seismic well bypass data R, and obtaining multiple local data segments R. k .

5. The automatic calibration method for multi-well synthetic seismic records according to claim 3, characterized in that, Step 6 includes: using a dynamic time-warping algorithm to calculate Rs for the synthetic seismic record and each local seismic data segment. k The best-matching seismic data segment R is determined by the minimum distance Dist among all calculated results. opt This enables the calibration of single-well synthetic seismic records and the acquisition of new single-well time-depth relationship pairs (h, t). The formula for calculating the matching distance Dist is: Dist=D(i,j)+min{D(i-1,j-1),D(i-1,j),D(i,j-1)}, 1≤i≤L1, 1≤j≤L2, In the formula, D(i,j) represents the sampling points S(i) and R(i) in the two signals. k The distance between sampling points S(i-1) and R(j-1) in the two signals is given by D(i-1, j-1). k The distance between sampling points S(i-1) and R(j-1) is given by D(i-1,j), where D(i-1,j) represents the distance between sampling points S(i-1) and R(j-1) in the two signals. k The distance between sampling points S(i) and R(j) is given by D(i,j-1), where D(i,j-1) represents the distance between sampling points S(i) and R(j) in the two signals. k The distance is (j-1).

6. The automatic calibration method for multi-well synthetic seismic records according to claim 1, characterized in that, The predetermined well network density is greater than 4km × 4km.

7. The automatic calibration method for multi-well synthetic seismic records according to claim 1, characterized in that, The number of wells drilled in step 1 is greater than 10.

8. The automatic calibration method for multi-well synthetic seismic records according to claim 1, characterized in that, The value of m is 3 to 5.

9. The automatic calibration method for multi-well synthetic seismic records according to claim 1, characterized in that, The speed threshold ranges from 0 to 400 m / s.

10. The automatic calibration method for multi-well synthetic seismic records according to claim 9, characterized in that, The speed threshold is [-200m / s, 200m / s].

11. A method for interpreting the structure of an oil and gas reservoir, characterized in that, The structural interpretation method for oil and gas reservoirs includes the automatic calibration method for multi-well synthetic seismic records as described in any one of claims 1 to 10.

12. A reservoir prediction method for oil and gas resources, characterized in that, The reservoir prediction method for oil and gas resources includes the automatic calibration method for multi-well synthetic seismic records as described in any one of claims 1 to 10.

Images