A synthetic record calibration method and device based on three-dimensional model forward modeling

Abstract

The present application belongs to the technical field of oil and natural gas exploration and development, and specifically discloses a synthetic record calibration method and device based on three-dimensional model forward. The method of the present application firstly obtains initial depth-time curves of all wells, then carries out first time shift correction on the depth-time curves to obtain all preliminary calibrated wells and establish a parameter model, carries out forward to obtain three-dimensional forward records of different incident angles, then corrects the depth-time curves after the first time shift according to the obtained three-dimensional forward records to realize fine calibration of the target layer of the wells, obtains fine calibration of all wells, and finally removes the wells whose fine calibration results and calibration results for structural interpretation differ by more than one phase. The device of the present application comprises a depth-time curve calculation module, a preliminary calibration module, a model construction module, a forward module, a fine calibration module and a screening module. The present application is used for improving the quality of well-seismic calibration.

Description

Technical Field

[0001] The present invention belongs to the technical field of oil and natural gas exploration and development, and in particular relates to a synthetic record calibration method and device based on three-dimensional model forward modeling. Background Art

[0002] Seismic calibration bridges geological and seismic data, serving as a prerequisite and foundation for seismic tectonic interpretation and reservoir prediction. It addresses several challenges: first, assigning accurate stratigraphic age to seismic reflection events facilitates the selection of target horizons for wave group comparison, interpretation, and structural mapping; second, leveraging drilling stratification data to control horizon tracking and comparison in areas with significant lateral variation in seismic wave group characteristics; third, providing accurate geological horizon and well data control for seismic inversion; and fourth, providing a basis for selecting analysis time windows in seismic attribute analysis. Depending on the data used, three main seismic calibration methods exist: vertical seismic profile (VSP) calibration, acoustic synthetic record calibration, and outcrop calibration. Acoustic synthetic record calibration is the most widely used, particularly in seismic inversion-based reservoir prediction, where it remains the only currently used seismic calibration method.

[0003] Seismic inversion is currently the most commonly used reservoir prediction method. Its process includes seismic wavelet extraction, acoustic wavelet synthesis record calibration, geological framework model establishment, inversion parameter testing, inversion job execution, and inversion result verification. Compared to structural interpretation, inversion-based reservoir prediction places higher demands on acoustic wavelet synthesis record calibration, moving away from simple geological horizon calibration in structural interpretation toward detailed reservoir calibration. Currently, mainstream acoustic wavelet synthesis record calibration is still largely manual and performed well by well. Specifically, it involves the following steps: First, an initial depth-time relationship is generated using acoustic wavelet logging curves. Then, a one-dimensional seismic record is synthesized using correlation curves such as P-wave, S-wave, and density, combined with seismic wavelets. Finally, based on the wave group characteristics of the synthesized record and the actual seismic data, the synthetic record is manually dragged up and down, with the marker layer as the basis, to continuously adjust the matching relationship between the synthetic record and the wellbore seismic trace to ensure the best match. This process is repeated until the acoustic wavelet synthesis record calibration is completed for all wells in the study area.

[0004] For high-density well network blocks, if reservoir prediction is carried out through seismic inversion, the time required to complete the calibration of acoustic wave synthetic records for a large number of drillings will inevitably increase exponentially compared to other inversion links. At the same time, due to human factors, the quality of well-seismic calibration may be uneven, thereby affecting the accuracy of subsequent reservoir prediction and even the accuracy of oil and gas reservoir description and remaining oil and gas prediction, ultimately making it difficult to meet the fast-paced, high-precision production needs of oil fields. Summary of the Invention

[0005] To address the above-mentioned deficiencies in the known technology, the present invention aims to provide a synthetic record calibration device based on three-dimensional model forward modeling, which aims to save the time spent on completing the calibration of acoustic synthetic records for large-scale drilling, while reducing the influence of human factors and improving the quality of well seismic calibration.

[0006] Another object of the present invention is to provide a synthetic record calibration method based on three-dimensional model forward modeling. This method needs to be implemented using the above-mentioned device. This method saves the time spent on completing the calibration of acoustic wave synthetic records for large-scale drilling, while reducing the influence of human factors and improving the quality of well seismic calibration, thereby improving the accuracy of subsequent reservoir prediction, oil and gas reservoir description, and remaining oil and gas prediction.

[0007] To achieve the above object, the technical solutions adopted by the present invention are as follows:

[0008] A synthetic record calibration method based on three-dimensional model forward modeling, the method comprising the following steps performed in sequence:

[0009] S1. Calculate the initial depth-time curves of all wells;

[0010] S2. Perform the first time shift correction on the initial depth-time curves of all wells to obtain the depth-time curves after the first time shift, thereby achieving preliminary calibration of the target layers of all wells;

[0011] S3. Using the preliminarily calibrated wells, under the constraints of the stratigraphic framework, a high-precision elastic parameter model in the time domain is established through inter-well interpolation and extrapolation;

[0012] S4. Based on the high-precision elastic parameter model in the time domain, perform 3D forward modeling of amplitude variation with offset or amplitude variation with incident angle, i.e., 3D AVO / AVA forward modeling, and obtain 3D forward modeling records at different incident angles.

[0013] S5. Selecting actual partial incident angle stacked 3D seismic data as reference seismic data, and using the 3D forward modeling record corresponding to each actual partial incident angle stacked 3D seismic data as adjusted seismic data, obtaining the optimal time shift and maximum correlation 2D plane data for all adjusted seismic data traces, performing a second time shift correction on the deep-time curve after the first time shift, and obtaining the deep-time curve after the second time shift, thereby achieving fine calibration of the target layers of all wells;

[0014] The actual partial incident angle stacked 3D seismic data is known from the calibration results for structural interpretation in the early stage, and the calibration results for structural interpretation in the early stage are obtained before executing step S1;

[0015] S6. Based on the optimal time shift and maximum correlation of each well, the wells whose fine calibration results differ from the calibration results for structural interpretation by more than one phase are discarded to complete the final calibration.

[0016] As a limitation, obtaining the initial deep-time curve in step S1 includes the following steps performed in sequence:

[0017] a1) Using the acoustic time difference curve and the sound velocity conversion formula, the acoustic velocity curves of all wells in the study area are calculated. The velocity of each sampling point on the acoustic velocity curve is the formation velocity. The double-layer traveltime curves of adjacent sampling points of all wells are obtained in batches based on the double-layer traveltime formula.

[0018] a2) Based on the double-layer travel time curve, the sampling points on the double-layer travel time curve are accumulated and summed to obtain the initial depth-time curves of all wells.

[0019] As a second limitation, the process of performing the first time-shift correction on the initial deep-time curve in step S2 includes the following steps performed in sequence:

[0020] b1) Based on the calibration results for structural interpretation, determine the vertical marker layer and seismic reflection layer interpretation results, and extract the double-layer travel time of the marker layer in the seismic traces of all well target layers;

[0021] Determine the double-layer travel time of the marker layer of the well according to the initial depth-time curve obtained in step S1;

[0022] b2) The time displacement of each well is obtained by subtracting the double-layer travel time of the marker layer of the seismic trace where the target layer is located from the double-layer travel time of the marker layer of the well, and converting it into a single-layer time displacement curve with a constant value;

[0023] b3) Adding the initial depth-time curve obtained in step S1 to the primary time-shift curve to obtain the primary time-shifted depth-time curve, thereby achieving preliminary calibration of the target layer of each well.

[0024] As a third limitation, in step S3, two or three seismic reflection horizons including a marker layer are used, and a stratigraphic framework is established according to the sedimentary pattern and stratigraphic contact relationship of the study area.

[0025] As a fourth limitation, in step S4, the specific process of performing the three-dimensional amplitude variation with offset AVO forward modeling or the amplitude variation with incident angle AVA forward modeling is as follows:

[0026] By using a high-precision elastic parameter model in the time domain, combined with the actual partial incident angle stacking three-dimensional seismic data and the seismic wavelet corresponding to the actual partial incident angle stacking three-dimensional seismic data, forward modeling is carried out to obtain four-dimensional gather data. The four-dimensional gather data is then subjected to dimensionality reduction conversion to obtain three-dimensional forward modeling records corresponding to the partial stacking three-dimensional seismic data at different incident angles; among them, the added dimension is the incident angle dimension.

[0027] As a fifth limitation, in the step S5, the process of obtaining the best time shift amount of all adjusted seismic data traces and the two-dimensional plane data of maximum correlation includes the following steps in sequence:

[0028] c1) opening a time window containing the target interval as an analysis time window, the analysis time window containing a marker layer, and setting a maximum time shift amount;

[0029] c2) for the reference seismic data and the adjusted seismic data in the analysis time window, recording the time shift amount and calculating the correlation between the reference seismic data traces and the adjusted seismic data traces in the analysis time window for each time shift of the adjusted seismic data traces;

[0030] wherein, within the range of the maximum time shift amount, when the correlation between the reference seismic data traces and the adjusted seismic data traces in the analysis time window corresponding to a certain time shift amount is the maximum, the time shift amount is the best time shift amount.

[0031] As a further limitation, in the step S5, the second time shift correction of the deep-time curve after the first time shift includes: selecting the best time shift amount of the seismic trace marker layer of the target interval of all wells as the second time shift amount, converting the second time shift amount into a second time shift amount curve, and then adding it to the deep-time curve after the first time shift to obtain the deep-time curve after the second time shift.

[0032] The application also discloses a synthetic record calibration device based on three-dimensional model forward, which is applied to any one of steps S1-S6, and the device comprises:

[0033] a deep-time curve calculation module, a preliminary calibration module, a model construction module, a forward module, a fine calibration module and a screening module;

[0034] The deep-time curve calculation module is used to calculate the initial deep-time curve of the well and output to the preliminary calibration module;

[0035] The preliminary calibration module is used to perform the first time shift correction on the initial deep-time curve output by the deep-time curve calculation module to obtain the preliminary calibrated well and output to the model construction module;

[0036] The model construction module establishes a time domain high-precision elastic parameter model through inter-well difference and extrapolation under the constraint of the stratigraphic framework according to the preliminary calibrated well output by the preliminary calibration module, and outputs the time domain high-precision elastic parameter model to the forward module;

[0037] The forward module carries out three-dimensional amplitude versus offset forward or amplitude versus incidence angle forward according to the time domain high-precision elastic parameter model output by the model construction module to obtain three-dimensional forward records of different incidence angles and output to the fine calibration module;

[0038] The fine calibration module obtains the optimal time shift and maximum correlation two-dimensional plane data of all adjusted seismic data channels based on the three-dimensional forward modeling records of different incident angles obtained by the forward modeling module. It performs a second time shift correction on the deep-time curve after the first time shift to achieve fine calibration of the target layer of the well.

[0039] The screening module, based on the wells of each finely calibrated target layer output by the fine calibration module, discards the wells whose calibration results differ from the known calibration results for structural interpretation by more than one phase, and completes the final calibration.

[0040] Due to the adoption of the above technical solution, the present invention has the following beneficial effects compared with the prior art:

[0041] (1) The method of the present invention saves the time spent on completing the calibration of acoustic wave synthetic records for a large number of drillings, reduces the influence of human factors, improves the quality of well seismic calibration, and thus improves the accuracy of subsequent reservoir prediction, oil and gas reservoir description, and remaining oil and gas prediction;

[0042] (2) The method of the present invention effectively improves the quality of well seismic calibration by taking two time-shift corrections, thereby improving the accuracy of subsequent reservoir prediction, oil and gas reservoir description, and remaining oil and gas prediction.

[0043] In summary, the present invention can improve the quality of well seismic calibration, thereby improving the accuracy of subsequent reservoir prediction, oil and gas reservoir description, and remaining oil and gas prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The present invention will be described in further detail below with reference to the accompanying drawings and specific embodiments.

[0045] Figure 1 (a) is the single well velocity curve of Example 1 of the present invention;

[0046] Figure 1 (b) is the double-layer travel time curve of adjacent sampling points in Example 1 of the present invention;

[0047] Figure 2 (a) is the double-layer travel time curve of adjacent sampling points in Example 1 of the present invention;

[0048] Figure 2 (b) is the travel time curve based on the reference reference point according to Example 1 of the present invention;

[0049] Figure 3 This is a graph showing the time shift data for a single well in Example 1 of the present invention;

[0050] Figure 4 (a) is a calibration diagram of a synthetic recording before a time shift according to Example 1 of the present invention;

[0051] Figure 4(b) is a calibration comparison diagram of the synthesized record after a time shift in Example 1 of the present invention;

[0052] Figure 5 The stratigraphic framework established in Example 1 of the present invention;

[0053] Figure 6 The density model section in the time-domain high-precision elastic parameter model established in Example 1 of the present invention;

[0054] Figure 7 A longitudinal wave impedance model section in the time domain high-precision elastic parameter model established in Example 1 of the present invention;

[0055] Figure 8 A cross section of a shear wave impedance model in a time-domain high-precision elastic parameter model established in Example 1 of the present invention;

[0056] Figure 9 This is the three-dimensional AVO / AVA forward modeling record of Example 1 of the present invention;

[0057] Figure 10 This is a schematic diagram of data volume correlation analysis in Example 1 of the present invention;

[0058] Figure 11 (a) is a schematic diagram of the optimal time shift amount according to Example 1 of the present invention;

[0059] Figure 11 (b) is a schematic diagram of the maximum correlation coefficient corresponding to the optimal time shift amount in Example 1 of the present invention;

[0060] Figure 12 This is a calibration comparison diagram of the synthetic record after two time shifts in Example 1 of the present invention. DETAILED DESCRIPTION

[0061] In order to better explain the present invention and facilitate understanding, preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings through specific implementation methods.

[0062] Example 1 Synthetic Record Calibration Method Based on 3D Model Forward Modeling

[0063] This embodiment provides a synthetic log calibration method based on 3D model forward modeling. This method has been applied in Block B of the Western Basin of China. This embodiment includes the following steps performed in sequence:

[0064] S1. Calculate the initial depth-time curves of all wells;

[0065] S2. Perform the first time shift correction on the initial depth-time curves of all wells to obtain the depth-time curves after the first time shift, thereby achieving preliminary calibration of the target layers of all wells;

[0066] S3. Using the preliminarily calibrated wells, under the constraints of the stratigraphic framework, a high-precision elastic parameter model in the time domain is established through inter-well interpolation and extrapolation;

[0067] S4. Based on the high-precision elastic parameter model in the time domain, perform 3D forward modeling of amplitude variation with offset or amplitude variation with incident angle, i.e., 3D AVO / AVA forward modeling, and obtain 3D forward modeling records at different incident angles.

[0068] S5. Selecting actual partial incident angle stacked 3D seismic data as reference seismic data, and using the 3D forward modeling record corresponding to each actual partial incident angle stacked 3D seismic data as adjusted seismic data, obtaining the optimal time shift and maximum correlation 2D plane data for all adjusted seismic data traces, performing a second time shift correction on the deep-time curve after the first time shift, and obtaining the deep-time curve after the second time shift, thereby achieving fine calibration of the target layers of all wells;

[0069] The actual partial incident angle stacked 3D seismic data is known from the calibration results for structural interpretation in the early stage, and the calibration results for structural interpretation in the early stage are obtained before executing step S1;

[0070] S6. Based on the optimal time shift and maximum correlation of each well, the wells whose fine calibration results differ from the calibration results for structural interpretation by more than one phase are discarded to complete the final calibration.

[0071] In step S1 of this embodiment, obtaining the initial deep-time curve includes the following steps performed in sequence:

[0072] a1) Using the acoustic time difference curve and the sound velocity conversion formula, calculate the acoustic velocity curves for all wells in the study area. The velocity of each sampling point on the acoustic velocity curve is the formation velocity. Based on the double-layer traveltime formula, batch obtain the double-layer traveltime curves of adjacent sampling points for all wells. The double-layer traveltime of each sampling point can be obtained based on the formation velocity and sampling interval of each sampling point. Figure 1 (a) is the velocity curve of a single well, Figure 1 (b) is the double-layer travel time curve of adjacent sampling points. Since the sampling interval of the logging curve is small, usually 0.125m, it can be approximately assumed that the formation velocity between adjacent sampling points is constant. The formation velocity of the sampling point located below is used as the formation velocity between adjacent sampling points.

[0073] a2) Based on the double-layer travel time curve, the sampling points on the double-layer travel time curve are accumulated and summed to obtain the initial depth time curve of all wells. Figure 2 (a) is the double-layer travel time curve of adjacent sampling points, Figure 2 (b) is the travel time curve based on the reference point. In the figure, the point before the start point of the acoustic logging is used as the reference point. Figure 2The sampling points in (a) are accumulated and summed, and the accumulation and summation process is as follows: the double-layer travel time of the nth sampling point (n>1) relative to the reference point is the accumulation of all sampling points before the sampling point.

[0074] In step S2 of the embodiment, the process of the first time shift correction of the initial deep time curve includes the following steps performed in sequence:

[0075] b1) According to the calibration results for the structural interpretation, determine the interpretation results of the marker layer and the seismic reflection horizon adjacent to the target layer in the vertical direction, and extract the double-layer travel time T a of the marker layer of the seismic trace of the target layer segment of the well;

[0076] According to the initial deep time curve obtained in step S1, determine the double-layer travel time T b of the marker layer of the well;

[0077] b2) T a minus T b , to obtain the primary time shift amount of the well, Figure 3 is a single-well primary time shift amount data graph, in which ch73_horizon_time is T a and ch73_top_time is T b The primary time shift amount of each sampling point constitutes a primary time shift amount curve.

[0078] b3) Add the initial deep time curve obtained in step S1 to the primary time shift amount curve to obtain the deep time curve after the primary time shift, and realize the preliminary calibration of the target layer of the single well, Figure 4 (a) is a synthetic record calibration graph before the primary time shift, Figure 4 (b) is a synthetic record calibration comparison graph after the primary time shift, the horizontal axis corresponds to the plane position, and the vertical axis corresponds to the time amount.

[0079] In step S3 of the embodiment, two or three seismic reflection horizons including the marker layer are used, and a stratigraphic framework is established according to the sedimentary pattern and stratigraphic contact relationship of the study area, Figure 5 is the established stratigraphic framework, the five lines in the vertical direction are seismic interpretation horizons, and different colors in the graph represent different stratigraphic segments, Figure 6 is the rock density model profile in the established time-domain high-precision elastic parameter model, Figure 7 is the rock P-wave impedance model profile in the established time-domain high-precision elastic parameter model, Figure 8 is the rock S-wave impedance model profile in the established time-domain high-precision elastic parameter model.

[0080] In step S4 of the embodiment, the specific process of the three-dimensional amplitude versus offset (AVO) forward or amplitude versus angle (AVA) forward is as follows:

[0081] The four-dimensional gather data is obtained by using the time-domain high-precision elastic parameter model, combining the actual partial angle stack three-dimensional seismic data and the seismic wavelet corresponding to the actual partial angle stack three-dimensional seismic data, and carrying out forward, and the three-dimensional forward record corresponding to the different angle partial stack three-dimensional seismic data is obtained by reducing the dimension of the four-dimensional gather data; wherein, the added dimension is the angle dimension. Figure 9 The three-dimensional AVO / AVA forward record is obtained, wherein the near trace, the middle trace and the far trace correspond to different angles.

[0082] The process of obtaining the best time shift amount of all adjusted seismic data traces and the maximum correlation two-dimensional plane data in step S5 of the embodiment includes the following steps performed in sequence:

[0083] c1) opening a time window containing the target layer as an analysis time window, the analysis time window containing a marker layer, and setting a maximum time shift amount;

[0084] c2) for the reference seismic data and the adjusted seismic data in the analysis time window, recording the time shift amount and calculating the correlation between the reference seismic data trace and the adjusted seismic data trace in the analysis time window each time the adjusted seismic data trace is time shifted, wherein the downward shift relative to the initial position is positive, and the upward shift is negative;

[0085] Within the maximum time shift amount range, when the correlation between the reference seismic data trace and the adjusted seismic data trace in the analysis time window corresponding to a certain time shift amount is the maximum, the time shift amount is the best time shift amount, Figure 10 The data body correlation analysis diagram is shown, wherein the reference horizon is the marker layer, and the correlation corresponding to the T3 time is the maximum, so the T3 time shift amount is the best time shift amount, Figure 11 (a) is the best time shift amount, and the warm color tone represents a large time shift amount, and the cold color tone represents a small time shift amount, Figure 11 (b) is the maximum correlation coefficient corresponding to the best time shift amount, and the warm color tone represents a large correlation coefficient, and the cold color tone represents a small correlation coefficient.

[0086] The second time shift correction of the deep-time curve after the first time shift includes: selecting the best time shift amount of the seismic trace marker layer of the target layer of all wells as the second time shift amount, converting the second time shift amount into a second time shift amount curve, and then adding the deep-time curve after the first time shift to obtain the deep-time curve after the second time shift, Figure 12 The synthesized record calibration comparison diagram after the second time shift is shown, wherein the warm color tone represents a high correlation, and the cold color tone represents a low correlation.

[0087] After obtaining the optimal time shift and maximum correlation for all wells and achieving fine calibration of the target layers of all wells, the final calibration is completed by discarding wells whose fine calibration results differ from the known calibration results for structural interpretation by more than one phase based on the optimal time shift and maximum correlation of each well.

[0088] This example completed fine calibration for 2,458 wells in Block B of the Western China Basin. Using traditional synthetic log calibration methods, completing fine reservoir calibration for all wells would take an estimated 15-20 working days. Furthermore, due to human factors, calibration quality can vary, impacting the accuracy of subsequent seismic inversion. This example significantly improves calibration quality while reducing the time required to just 3-4 working days. This significantly increases the efficiency of well-seismic calibration to 70%, providing strong support for detailed characterization of clastic oil and gas reservoirs and rapid well deployment.

[0089] Example 2 Synthetic Record Calibration Device Based on 3D Model Forward Modeling

[0090] This embodiment is applied to embodiment 1, and includes a deep-time curve calculation module, a preliminary calibration module, a model building module, a forward modeling module, a fine calibration module, and a screening module.

[0091] The depth-time curve calculation module is used to calculate the initial depth-time curve of the well and output it to the preliminary calibration module;

[0092] The preliminary calibration module is used to perform the first time shift correction on the initial deep-time curve output by the deep-time curve calculation module, obtain the preliminary calibrated well, and output it to the model construction module;

[0093] The model building module, based on the preliminarily calibrated wells output by the preliminarily calibrated module, establishes a high-precision elastic parameter model in the time domain through inter-well difference and extrapolation under the constraints of the stratigraphic grid, and outputs the high-precision elastic parameter model in the time domain to the forward modeling module;

[0094] The forward modeling module performs forward modeling of 3D amplitude variation with offset or amplitude variation with incident angle based on the time-domain high-precision elastic parameter model output by the model building module, obtains 3D forward modeling records at different incident angles, and outputs them to the fine calibration module.

[0095] The fine calibration module obtains the optimal time shift and maximum correlation two-dimensional plane data of all adjusted seismic data channels based on the three-dimensional forward modeling records of different incident angles obtained by the forward modeling module. It performs a second time shift correction on the deep-time curve after the first time shift to achieve fine calibration of the target layer of the well.

[0096] The screening module screens the wells of each fine calibration target layer output by the fine calibration module, removes the wells with a difference greater than one phase between the calibration result and the known calibration result for the structure interpretation, and completes the final calibration.

Claims

1. A synthetic record calibration method based on 3D model forward modeling, characterized in that: The method comprises the following steps performed in sequence: S1. Calculate the initial depth-time curves of all wells; S2. Perform the first time shift correction on the initial depth-time curves of all wells to obtain the depth-time curves after the first time shift, thereby achieving preliminary calibration of the target layers of all wells; S3. Using the preliminarily calibrated wells, under the constraints of the stratigraphic framework, a high-precision elastic parameter model in the time domain is established through inter-well interpolation and extrapolation; S4. Based on the high-precision elastic parameter model in the time domain, perform 3D forward modeling of amplitude variation with offset or amplitude variation with incident angle, i.e., 3D AVO / AVA forward modeling, and obtain 3D forward modeling records at different incident angles. S5. Selecting actual partial incident angle stacked 3D seismic data as reference seismic data, and using the 3D forward modeling record corresponding to each actual partial incident angle stacked 3D seismic data as adjusted seismic data, obtaining the optimal time shift and maximum correlation 2D plane data for all adjusted seismic data traces, performing a second time shift correction on the deep-time curve after the first time shift, and obtaining the deep-time curve after the second time shift, thereby achieving fine calibration of the target layers of all wells; The actual partial incident angle stacked 3D seismic data is known from the calibration results for structural interpretation in the early stage, and the calibration results for structural interpretation in the early stage are obtained before executing step S1; S6. Based on the optimal time shift and maximum correlation of each well, the wells whose fine calibration results differ from the calibration results for structural interpretation by more than one phase are discarded to complete the final calibration.

2. The synthetic record calibration method based on 3D model forward modeling according to claim 1, characterized in that: The step S1 of obtaining the initial deep-time curve includes the following steps performed in sequence: a1) Using the acoustic time difference curve and the sound velocity conversion formula, the acoustic velocity curves of all wells in the study area are calculated. The velocity of each sampling point on the acoustic velocity curve is the formation velocity. The double-layer traveltime curves of adjacent sampling points of all wells are obtained in batches based on the double-layer traveltime formula. a2) Based on the double-layer travel time curve, the sampling points on the double-layer travel time curve are accumulated and summed to obtain the initial depth-time curves of all wells.

3. The synthetic record calibration method based on 3D model forward modeling according to claim 1 or 2, characterized in that: The process of performing the first time shift correction on the initial deep time curve in step S2 includes the following steps performed in sequence: b1) Based on the calibration results for structural interpretation, determine the vertical marker layer and seismic reflection layer interpretation results, and extract the double-layer travel time of the marker layer in the seismic traces of all well target layers; Determine the double-layer travel time of the marker layer of the well according to the initial depth-time curve obtained in step S1; b2) The time displacement of each well is obtained by subtracting the double-layer travel time of the marker layer of the seismic trace where the target layer is located from the double-layer travel time of the marker layer of the well, and converting it into a single-layer time displacement curve with a constant value; b3) Adding the initial depth-time curve obtained in step S1 to the primary time-shift curve to obtain the primary time-shifted depth-time curve, thereby achieving preliminary calibration of the target layer of each well.

4. The synthetic record calibration method based on 3D model forward modeling according to claim 1 or 2, characterized in that: In step S3, two or three seismic reflection horizons including the marker layer are used, and a stratigraphic framework is established according to the sedimentary pattern and stratigraphic contact relationship of the study area.

5. The synthetic record calibration method based on 3D model forward modeling according to claim 1 or 2, characterized in that: In step S4, the specific process of carrying out the three-dimensional amplitude variation with offset AVO forward modeling or the amplitude variation with incident angle AVA forward modeling is as follows: By using a high-precision elastic parameter model in the time domain, forward modeling is carried out in combination with the actual partial incident angle stacking 3D seismic data and the seismic wavelet corresponding to the actual partial incident angle stacking 3D seismic data to obtain 4D gather data. The 4D gather data is then subjected to dimensionality reduction conversion to obtain 3D forward modeling records corresponding to the partial stacking 3D seismic data at different incident angles; the added dimension is the incident angle dimension.

6. The synthetic record calibration method based on 3D model forward modeling according to claim 1 or 2, characterized in that: In step S5, the process of obtaining the optimal time shift and maximum correlation two-dimensional plane data of all adjusted seismic data traces includes the following steps performed in sequence: c1) Set a time window that includes the target layer segment as the analysis time window, include the marker layer in the analysis time window, and set the maximum time shift; c2) For the reference seismic data and the adjusted seismic data within the analysis time window, each time the adjusted seismic data channel is time-shifted, the time-shift amount is recorded and the correlation between the reference seismic data channel and the adjusted seismic data channel within the analysis time window is calculated; Among them, within the maximum time shift range, when the correlation between the reference seismic data trace and the adjusted seismic data trace in the corresponding analysis time window for a certain time shift is the largest, the time shift is the optimal time shift.

7. The synthetic record calibration method based on 3D model forward modeling according to claim 6, characterized in that: In step S5, performing a second time shift correction on the deep-time curve after the first time shift includes: selecting the optimal time shift of the seismic marker layer where the target layer segment of all wells is located as the second time shift, converting the secondary time shift into a secondary time shift curve, and then adding it to the deep-time curve after the first time shift to obtain the deep-time curve after the second time shift.

8. A synthetic record calibration device based on 3D model forward modeling, used to implement the calibration method according to any one of claims 1 to 7, characterized in that: The device comprises: a deep time curve calculation module, a preliminary calibration module, a model building module, a forward modeling module, a fine calibration module and a screening module; The depth-time curve calculation module is used to calculate the initial depth-time curve of the well and output it to the preliminary calibration module; The preliminary calibration module is used to perform the first time shift correction on the initial deep-time curve output by the deep-time curve calculation module, obtain the preliminary calibrated well, and output it to the model construction module; The model building module, based on the preliminarily calibrated wells output by the preliminarily calibrated module, establishes a high-precision elastic parameter model in the time domain through inter-well difference and extrapolation under the constraints of the stratigraphic grid, and outputs the high-precision elastic parameter model in the time domain to the forward modeling module; The forward modeling module performs forward modeling of 3D amplitude variation with offset or amplitude variation with incident angle based on the time-domain high-precision elastic parameter model output by the model building module, obtains 3D forward modeling records at different incident angles, and outputs them to the fine calibration module. The fine calibration module obtains the optimal time shift and maximum correlation two-dimensional plane data of all adjusted seismic data channels based on the three-dimensional forward modeling records of different incident angles obtained by the forward modeling module. It performs a second time shift correction on the deep-time curve after the first time shift to achieve fine calibration of the target layer of the well. The screening module, based on the wells of each finely calibrated target layer output by the fine calibration module, discards the wells whose calibration results differ from the known calibration results for structural interpretation by more than one phase, and completes the final calibration.

Images