Methods, apparatus, equipment and media for constructing velocity models of structurally heterogeneous strata

Abstract

This invention discloses a method, apparatus, equipment, and medium for constructing a velocity model of structurally heterogeneous strata. Based on well logging and seismic data, a time-domain geological structure model is established. Using tomographic phasing interpolation technology, based on pre-stack depth-migrated velocity volumes and well logging data of the target area, error coefficient volumes between the low-frequency well logging velocities and the three-dimensional pre-stack depth-migrated velocities of each segment in the geological structure model are obtained. These error coefficient volumes are then merged to obtain a low-frequency error coefficient volume. Based on this low-frequency error coefficient volume, the pre-stack depth-migrated velocity volume is corrected to obtain a low-frequency layer velocity volume. Using spectral simulation inversion technology, based on well logging and seismic data of the target area, a mid-to-high frequency layer velocity volume is obtained. Finally, the low-frequency and mid-to-high frequency layer velocity volumes are fused to establish a velocity model. This velocity model can effectively identify low-amplitude traps underlying strongly transversely heterogeneous strata.

Description

Technical Field

[0001] This invention relates to the field of petroleum exploration technology, and in particular to a method, apparatus, equipment and medium for constructing velocity models of structurally heterogeneous strata. Background Technology

[0002] Structural oil and gas reservoirs are an important type of oil and gas reservoir in oil and gas exploration, with great exploration potential, and numerous oil and gas fields have been discovered both domestically and internationally. The core of exploration for this type of reservoir is finding structural traps. However, with the continuous deepening of exploration and development, the exploration degree of the more significant and easily discoverable structural traps is already very high, and exploration is gradually shifting towards lower-profile structural oil and gas reservoirs that are not easily discovered. In recent years, exploration practices both domestically and internationally have shown that lower-profile structural oil and gas reservoirs have numerous exploration strata and broad exploration areas, containing abundant oil and gas resources, making them a hot area for oil and gas exploration with broad exploration prospects.

[0003] The core of low-amplitude structural oil and gas reservoir exploration is finding low-amplitude traps. These traps are highly sensitive to velocity models, and the accuracy of these models directly affects the characterization precision of low-amplitude traps. A precise and reliable velocity model is crucial for improving the accuracy of low-amplitude trap identification. As exploration and development in low-amplitude structural areas progresses, the geological conditions become increasingly complex, and the depth of the target strata gradually increases. Above the target strata often lie special geological bodies such as salt domes, mudstone wedges, and conglomerate bodies. These bodies typically differ significantly from the surrounding rocks in lithology, velocity, and thickness, resulting in strong lateral structural heterogeneity in the strata above the target strata. This strong lateral structural heterogeneity significantly increases the difficulty of establishing velocity models, thereby increasing the difficulty of identifying underlying low-amplitude traps. Conventional velocity modeling methods cannot accurately reflect the velocity variation characteristics of strongly laterally heterogeneous strata above the target strata, leading to large time-depth conversion errors and low accuracy in identifying low-amplitude structural features. This severely restricts the exploration and development process of low-amplitude traps beneath strongly structural heterogeneous strata.

[0004] Therefore, complex geological conditions pose challenges to the identification of low-amplitude traps. In order to improve the identification accuracy of low-amplitude traps underlying strong transverse structural heterogeneous strata and accelerate the exploration process of low-amplitude traps underlying strong transverse structural heterogeneous strata, it is urgent to establish a velocity model that can accurately characterize the velocity variation features of strong transverse structural heterogeneous strata and achieve a detailed description of the structural features of low-amplitude traps. Summary of the Invention

[0005] In view of the above problems, this invention is proposed to provide a method, apparatus, equipment and medium for constructing a velocity model of structurally heterogeneous strata that overcomes or at least partially solves the above problems. It takes tomographic phase-controlled interpolation technology, spectral simulation inversion technology and data fusion technology as the core, and adopts the method of low-frequency band correction, mid-to-high frequency band inversion and low-frequency and mid-to-high frequency band velocity data fusion to establish a high-precision velocity model with high geological characterization. It can effectively identify and finely characterize the low-amplitude traps under strongly transversely structurally heterogeneous strata.

[0006] Firstly, a method for constructing a velocity model for structurally heterogeneous layers is provided, including:

[0007] A time-domain geological structure model is established based on well logging and seismic data of the target area.

[0008] Based on the pre-stack depth migration layer velocity volume of the target area and the well logging data, the error coefficient volume between the well logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each layer in the geological structure model is obtained by using tomographic phase control interpolation technology.

[0009] The error coefficient volumes corresponding to each segment are merged to obtain the low-frequency band error coefficient volume;

[0010] Based on the low-frequency error coefficient volume, the pre-stack depth offset layer velocity volume is corrected to obtain the low-frequency layer velocity volume.

[0011] Based on well logging and seismic data of the target area, the mid-to-high frequency layer velocity volume is obtained using spectral simulation inversion technology.

[0012] The low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume are fused together, and a velocity model is established based on the fused layer velocity volume.

[0013] Optionally, the well logging data includes sonic transit time curves and density curves of each target well within the target area, and the seismic data is time-domain pre-stack depth migration seismic data. The establishment of a time-domain geological structure model based on the well logging data and seismic data of the target area includes:

[0014] Based on the sonic transit time curves and density curves of each target well within the target area, a synthetic seismic record is generated;

[0015] Seismic geological horizons were determined using synthetic seismic records, and the correspondence between geological horizons and seismic reflection phase axes was established.

[0016] Based on the correspondence between the geological strata and the seismic reflection phase axis, the faults and strata corresponding to the geological strata are tracked and interpreted to obtain time-domain seismic interpretation strata and fault data.

[0017] Based on the time-domain seismic interpretation horizon and fault data, a time-domain geological structure model is established.

[0018] Optionally, the pre-stack depth migration velocity volume based on the target area and the well logging data are used with tomographic phasing interpolation technology to obtain the error coefficient volume between the low-frequency well logging velocities and the three-dimensional pre-stack depth migration velocities corresponding to each segment in the geological structure model, including:

[0019] Based on the pre-stack depth migration layer velocity volume of the target area and the logging data, the error coefficient between the logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well in the target area is determined.

[0020] Based on the contact relationship between the top and bottom layers of each segment in the geological structure model, a set number of sub-layers are inserted between the top and bottom layers of each segment to obtain the interpolated lattice model corresponding to each segment.

[0021] Using the Kriging interpolation method, with the boundary data of each facies zone and the corresponding interpolation grid model of each facies zone as constraints, the error coefficient between the low-frequency logging velocity of the target well in each facies zone and the three-dimensional pre-stack depth migration velocity is determined based on the error coefficient between the low-frequency logging velocity of the target well in each facies zone and the three-dimensional pre-stack depth migration velocity.

[0022] By merging the error coefficient volumes corresponding to each phase zone within each layer, the error coefficient volume between the low-frequency logging layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each layer is obtained.

[0023] Optionally, the determination of the error coefficient between the low-frequency layer velocity of the well logging and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well in the target area, based on the pre-stack depth migration layer velocity volume of the target area and the well logging data, includes:

[0024] Based on the pre-stack depth migration velocity volume of the target area, the pre-stack depth migration velocity curves corresponding to each target well in the target area are extracted. The pre-stack depth migration velocity curves corresponding to each target well are the three-dimensional pre-stack depth migration velocity curves of the seismic traces closest to each target well.

[0025] Extract the sonic transit time curves of each target well within the target area from the logging data, and convert the sonic transit time curves into layer velocity curves;

[0026] The layer velocity curve is transformed in the time domain to obtain the time domain layer velocity curve;

[0027] The high-frequency end value of the effective frequency band of the three-dimensional pre-stack depth migration layer velocity volume is used as the high-pass frequency value of the low-pass filter to perform low-pass filtering on the time domain layer velocity curve to obtain the well logging low-frequency layer velocity curve.

[0028] Based on the three-dimensional pre-stack depth migration layer velocity curve and the well logging low-frequency layer velocity curve, the error coefficient between the well logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well in the target area is determined.

[0029] Optionally, the well logging data includes velocity and density curves of each target well within the target area, and the seismic data is time-domain pre-stack depth migration seismic data; based on the well logging data and seismic data of the target area, a mid-to-high frequency layer velocity volume is obtained using spectral simulation inversion technology, including:

[0030] Based on the density curves of each target well within the target area, the average density of each target well in each layer of the geological structure model in the time domain is determined.

[0031] Multiply the velocity curve value of each well by the average density of each layer to obtain the impedance curve of each target well in each layer.

[0032] Based on the impedance curves of each target well in each layer, the wave impedance curves of each target well in the target area are obtained.

[0033] The wave impedance curve is transformed from the depth domain to the time domain to obtain the time domain wave impedance curve.

[0034] Based on the time-domain impedance curves and the seismic data, the mid-to-high frequency layer velocity volumes corresponding to each layer within the target area are obtained through spectral simulation inversion.

[0035] The mid-to-high frequency segment layer velocity volumes corresponding to each segment within the target area are merged to obtain the mid-to-high frequency segment layer velocity volume.

[0036] Optionally, the step of fusing the low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume to obtain a velocity model includes:

[0037] Based on the spectral range of the low-frequency band layer velocity body, determine the high-pass frequency parameters and high-cutoff frequency parameters of the low-pass filter;

[0038] Based on the high-pass frequency parameters and high-cutoff frequency parameters of the low-pass filter, the low-frequency band layer velocity volume is subjected to low-pass filtering to obtain the filtered low-frequency band layer velocity volume.

[0039] Based on the spectral range of the mid-to-high frequency band layer velocity body, determine the low-pass frequency parameters and low-cutoff frequency parameters of the high-pass filter;

[0040] Based on the low-pass frequency parameters and low-cutoff frequency parameters of the high-pass filter, the mid-to-high frequency band layer velocity volume is subjected to high-pass filtering to obtain the filtered mid-to-high frequency band layer velocity volume.

[0041] According to the predetermined weight values ​​of low frequency and mid-high frequency, the filtered low-frequency band layer velocity volume and the mid-high frequency band layer velocity volume are fused to obtain the velocity model.

[0042] Optionally, the method further includes:

[0043] The velocity model is used to identify low-amplitude traps within the target area;

[0044] Based on the drilling data of the target wells within the target area, the reliability of the velocity model identification results is verified.

[0045] Secondly, a velocity model construction device for structurally heterogeneous layers is provided, comprising:

[0046] The structural model building module is used to build a time-domain geological structure model based on well logging data and seismic data of the target area.

[0047] The first error determination module is used to obtain the error coefficient volume between the low-frequency well logging velocity and the three-dimensional pre-stack depth migration velocity in each segment of the geological structure model based on the pre-stack depth migration layer velocity volume of the target area and the well logging data, using tomographic phase control interpolation technology.

[0048] The second error determination module is used to merge the error coefficient bodies corresponding to each segment to obtain the low-frequency error coefficient body.

[0049] The first velocity volume construction module is used to correct the pre-stack depth offset layer velocity volume based on the low-frequency band error coefficient volume to obtain the low-frequency band layer velocity volume.

[0050] The second velocity volume construction module is used to obtain the mid-to-high frequency layer velocity volume based on well logging data and seismic data of the target area using the spectrum simulation inversion method;

[0051] The velocity model building module is used to fuse the low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume, and to build a velocity model based on the fused layer velocity volume.

[0052] Thirdly, an electronic device is provided, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the speed model construction method as described in the first aspect.

[0053] Fourthly, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing the computer to perform the speed model construction method as described in the first aspect.

[0054] The technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages:

[0055] This invention provides a method, apparatus, device, and storage medium for constructing a velocity model of a structurally heterogeneous stratum. First, a time-domain geological structure model is established based on well logging and seismic data of the target area, providing an accurate geological framework for subsequent velocity model construction. Then, based on the pre-stack depth-migrated layer velocity volume and well logging data of the target area, tomographic phasing interpolation technology is used to calculate the error coefficient volume between the low-frequency well logging layer velocities and the three-dimensional pre-stack depth-migrated layer velocities corresponding to each segment in the geological structure model. The error coefficient volumes corresponding to each segment are merged to obtain a low-frequency error coefficient volume, which is used to correct the pre-stack depth-migrated layer velocity volume, resulting in a corrected low-frequency layer velocity volume, thereby improving the low-frequency characterization accuracy of the velocity model. Simultaneously, using spectral simulation inversion technology, this invention combines well logging and seismic data to obtain mid-to-high frequency layer velocity volumes, further enriching the information content of the velocity model. Finally, through data fusion technology, this invention effectively fuses the low-frequency layer velocity volume and the mid-to-high frequency layer velocity volumes to obtain a refined velocity model with high geological characterization. This velocity model not only accurately reflects the velocity variation characteristics of strongly transversely structured heterogeneous strata, but also significantly improves the identification accuracy of low-amplitude traps and the ability to finely characterize structural features, providing a scientific basis for oil and gas exploration and development decisions and deployments. It has important practical application value and prospects for promotion.

[0056] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0057] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0058] Figure 1 This is a flowchart of a method for constructing a velocity model of a structurally heterogeneous layer provided in an embodiment of the present invention;

[0059] Figure 2This is a schematic diagram of a time-domain pre-stack depth migration seismic profile provided in an embodiment of the present invention;

[0060] Figure 3 A cross-sectional view of a time-domain geological structure model is provided for an embodiment of the present invention;

[0061] Figure 4 This is a comparison chart of low-frequency correction of logging curves provided in an embodiment of the present invention;

[0062] Figure 5 This is a cross-sectional view of a low-frequency layer velocity body provided in an embodiment of the present invention;

[0063] Figure 6 This is a cross-sectional view of a mid-to-high frequency band layer velocity body provided in an embodiment of the present invention;

[0064] Figure 7 This is a cross-sectional view of a velocity model provided in an embodiment of the present invention;

[0065] Figure 8 This is a comparison diagram of the low-amplitude trap structure characteristics based on the conventional velocity model and the velocity model of this application;

[0066] Figure 9 This is a structural block diagram of a velocity model construction device for a structurally heterogeneous layer provided in an embodiment of the present invention. Detailed Implementation

[0067] To better understand the above technical solutions, the following will describe the above technical solutions in detail with reference to the accompanying drawings and specific implementation methods. It should be understood that the embodiments of this disclosure and the specific features in the embodiments are detailed descriptions of the technical solutions of this application, rather than limitations on the technical solutions of this application. Unless otherwise specified, the embodiments of this application and the technical features in the embodiments can be combined with each other.

[0068] In related technologies, there are currently three main types of velocity modeling for low-amplitude trap identification. The first type utilizes seismic migration velocities or stacking velocities to establish a velocity model. This method is suitable for relatively homogeneous media models, but its accuracy is lower for heterogeneous formations. The second type is numerically driven grid tomography velocity modeling. This method uses data-driven and computer-automated iterative solutions to obtain more refined velocity variations. However, this method has relatively low geological constraints and high ambiguity, limiting its application under complex geological conditions. The third type comprehensively utilizes seismic velocity, well logging velocity, and seismic interpretation data to establish a velocity model. Numerous methods have been developed, such as well fitting methods, well-logging average velocity models, well-seismic combined velocity modeling, and velocity inversion methods. This type of method is suitable for areas with relatively uniform control well distribution and can obtain high-accuracy velocity models, but its velocity prediction accuracy is lower in areas with few or no wells.

[0069] To address the aforementioned technical problems, this invention provides an innovative velocity model construction method. Utilizing well logging and seismic data, a time-domain geological structure model is established. With tomographic phasing interpolation, spectral simulation inversion, and data fusion as its core technologies, a high-resolution geological velocity model is constructed using low-frequency correction, mid-to-high-frequency inversion, and fusion of low-frequency and mid-to-high-frequency velocity data. This method enables effective identification and detailed characterization of low-amplitude traps underlying strongly transversely heterogeneous strata.

[0070] Figure 1 This is a flowchart of a method for constructing a velocity model of a structurally heterogeneous layer provided by an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:

[0071] Step S110: Based on well logging data and seismic data of the target area, establish a time-domain geological structure model.

[0072] Step S120: Based on the pre-stack depth migration layer velocity volume and logging data of the target area, tomographic phase control interpolation technology is used to obtain the error coefficient volume between the low-frequency logging layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each layer in the geological structure model.

[0073] Step S130: Merge the error coefficient volumes corresponding to each segment to obtain the low-frequency error coefficient volume.

[0074] Step S140: Based on the low-frequency error coefficient volume, the pre-stack depth offset layer velocity volume is corrected to obtain the low-frequency layer velocity volume.

[0075] Step S150: Based on the well logging data and seismic data of the target area, the mid-to-high frequency layer velocity volume is obtained by using spectral simulation inversion technology.

[0076] Step S160: Perform data fusion on the low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume, and establish a velocity model based on the fused layer velocity volume.

[0077] In some implementations, before performing step S110 above, the method further includes:

[0078] Acquire well logging data and seismic data. The well logging data shall include at least the sonic transit time curves and density curves of each target well in the target area, and the seismic data shall be time-domain pre-stack depth migration seismic data.

[0079] In some implementations, step S110 may include:

[0080] Step S111: Based on the sonic transit time curves and density curves of each target well within the target area, create a synthetic seismic record;

[0081] Step S112: Use the synthesized seismic records to perform seismic geological stratigraphy and establish the correspondence between geological stratigraphy and seismic reflection phase axes;

[0082] Step S113: Based on the correspondence between geological strata and seismic reflection phase axes, track and interpret faults and strata for the seismic reflection phase axes corresponding to the calibrated seismic geological strata to obtain time-domain seismic interpretation strata and fault data.

[0083] Step S114: Based on time-domain seismic interpretation of stratigraphic and fault data, establish a time-domain geological structure model.

[0084] For example, the target well can be a well within the target area that simultaneously possesses both sonic transit time and density curves. By creating synthetic seismic records, the depth-domain logging curves are converted to the time domain, seismic geological stratigraphy is performed, and the relationship between geological stratigraphy and seismic reflection phase axes is established, thereby obtaining the time-depth relationship of the target well within the target area. Time-domain seismic interpretation stratigraphic and fault data can include the combination patterns of various faults within the target area, the stratigraphic layers they cross, and the contact relationships between these layers. The combination patterns of various faults include grabens and horsts, step faults, ring faults, or radial faults, etc. The contact relationships between these layers can include six different stratigraphic contact modes: stratigraphic overlap, stratigraphic erosion, stratigraphic erosion, parallel bottom, parallel top, and parallel top-bottom.

[0085] In some implementations, the establishment of a time-domain geological structure model can be achieved in the following ways: Set the longitudinal time range of the geological structure model (in this embodiment, the top time Top is set to 0 milliseconds and the bottom time Base is set to 3000 milliseconds), and according to the interpretation mode of the seismic profile, in a bottom-up order, based on the time-domain seismic interpretation horizon and fault data, set the contact relationship of each stratum and the primary and secondary relationship of the faults, and establish the time-domain geological structure model.

[0086] In some implementations, the tomographic phase-controlled interpolation technique in step S120 is a kriging interpolation method that uses the phase zone boundary data of each segment and the corresponding interpolation lattice model of each segment as constraints. Step S120 may include:

[0087] Step S121: Based on the pre-stack depth migration layer velocity volume and logging data of the target area, determine the error coefficient between the low-frequency logging layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well in the target area.

[0088] In some implementations, step S121 includes:

[0089] Based on the pre-stack depth migration velocity volume of the target area, the pre-stack depth migration velocity curves corresponding to each target well in the target area are extracted. The pre-stack depth migration velocity curves corresponding to each target well are the three-dimensional pre-stack depth migration velocity curves of the seismic traces closest to each target well.

[0090] Extract the sonic transit time curves of each target well within the target area from the logging data, and convert the sonic transit time curves into layer velocity curves;

[0091] The layer velocity curve is transformed into the time domain to obtain the time domain layer velocity curve;

[0092] The high-frequency end value of the effective frequency band of the three-dimensional pre-stack depth migration layer velocity volume is used as the high-pass frequency value of the low-pass filter. The time-domain layer velocity curve is then low-pass filtered to obtain the well logging low-frequency layer velocity curve.

[0093] Based on the three-dimensional pre-stack depth migration layer velocity curve and the well logging low-frequency layer velocity curve, the error coefficient between the well logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well in the target area is determined.

[0094] For example, the acoustic transit time curve can be converted into a layer velocity curve according to the following formula (1):

[0095] Vel = 1000000 / DT; (1)

[0096] Wherein, DT is the acoustic time difference curve, with units of microseconds / meter, and Vel is the layer velocity curve, with units of meters / second.

[0097] Then, based on the correspondence between geological strata and seismic reflection phase axes established in step S112, the time-depth relationship of each target well within the target area is obtained. The depth-domain layer velocity curve is converted to the time domain to obtain the time-domain layer velocity curve. Next, based on the high-frequency end value H1 of the effective frequency band of the pre-stack migration velocity volume, H1 is used as the high-pass frequency value for low-pass filtering to perform low-pass filtering on the time-domain logging curve, obtaining the logging low-frequency layer velocity curve, denoted as Vel-Lowfreq. The three-dimensional pre-stack depth migration layer velocity curve is denoted as Psdm-Vel. The error coefficient between the logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity can be obtained by dividing the two, i.e.

[0098] C=Vel-Lowfreq / Psdm-Vel; (2)

[0099] The error coefficient C between the low-frequency layer velocity in well logging and the pre-stack depth offset layer velocity can be obtained using formula (2).

[0100] Step S122: Based on the contact relationship between the top and bottom layers of each segment in the geological structure model, insert a set number of sub-layers between the top and bottom layers of each segment to obtain the interpolated lattice model corresponding to each segment.

[0101] For example, taking a certain segment L5 in the time domain geological structure model as an example, according to the contact relationship between the top and bottom layers of segment L5, P sub-layers can be inserted between the top and bottom layers of segment L5 according to a certain ratio P (in this example, P = 30). The top layer, bottom layer and P inserted sub-layers of segment L5 constitute the interpolated lattice model L5-Model corresponding to segment L5.

[0102] Step S123: Using the Kriging interpolation method, with the boundary data of each facies zone and the corresponding interpolation grid model of each facies zone as constraints, and based on the error coefficient between the logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to the target well in each facies zone, determine the error coefficient body between the logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each facies zone.

[0103] Before performing step S123, the method may further include:

[0104] Based on seismic data, phase zone boundary data for each segment are obtained.

[0105] Step S123: Merge the error coefficient volumes corresponding to each phase zone within each layer segment to obtain the error coefficient volume between the low-frequency logging layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each layer segment.

[0106] For example, using the facies boundary data of the L5 section as the first constraint and the L5-Model as the second constraint for interpolation within the L5 section, the error coefficients corresponding to each facies zone in the L5 section can be obtained using the ordinary kriging interpolation method. Specifically, using the facies boundary data of f facies zones in this section as constraints, kriging interpolation is performed within each facies zone. For a certain facies zone Facies1 within the L5 section (in this example, the L5 section mainly develops three facies zones: Facies1, Facies2, and Facies3), the velocity error coefficients within the Facies1 facies zone of the L5 section are obtained based on the velocity error coefficients C of n wells located within this facies zone using the kriging interpolation method.

[0107] Specifically, for a specific facies zone Facies1 among the f facies zones in the L5 section (in this example, f = 3, meaning the L5 section has three facies zones: Facies1, Facies2, and Facies3), the boundary of this facies zone is used as the boundary constraint for the interpolation of the Facies1 facies zone in the L5 section. Within this facies zone, based on the velocity error coefficient C of the n wells located within this facies zone, for a specific HorizonM layer in the lattice model (the L5 section has P+2 layers), the error coefficient value at each CDP (Common Depth Point) along this layer in the three-dimensional work area is regarded as the regional variable V(x). For the n wells within this facies zone (in this example, n = 20), an error coefficient value can be read at each well along this layer. The error coefficient values ​​read from the n wells can be regarded as n observations, namely V(x1), V(x2), V(x3), ..., V(x4). n For the estimated value of the error coefficient at a certain CDP along the layer within the work area, denoted as V(x0), it can be obtained by weighting the observations of n known sampling points using formula (3):

[0108]

[0109] In the formula, α i (i = 1, 2, ..., n) are the weight coefficients to be determined. According to the principle of Kriging interpolation, when the function V(x) satisfies the assumption of second-order stationarity, based on the principle that the estimator is unbiased and the estimation variance is minimized, its weight coefficients are calculated by constructing Lagrange conditional extrema. The n weight coefficient values ​​of the optimal, linear and unbiased estimates can be solved using the system of equations (4):

[0110]

[0111] In the formula, γ(x) i ,x j ) represents the sampling point x i and x jThe linear variogram values ​​between the values, μ being the Lagrange daily number for minimizing variance, can be obtained from equation (4) as the weighting coefficient α. i The value of (i=1,2,…,n) can be substituted into equation (3) to obtain the estimated value of the error coefficient at a certain CDP, denoted as V(x0).

[0112] This yields an estimated error coefficient value for HorizonM along a certain CDP in the lattice model. Following this method, the estimated value of each CDP along HorizonM within the three-dimensional region is calculated sequentially, thus obtaining the error coefficient value along HorizonM. For each layer in the L5 interpolated lattice model—that is, a total of P+2 layers including the top layer, bottom layer, and interpolated sublayers of the L5 layer—the error coefficient values ​​for each layer in the L5 lattice model can be obtained sequentially using the method for obtaining the error coefficient values ​​along HorizonM. This yields the error coefficient volume of the L5 layer under the Facies1 constraint.

[0113] Step S123: Merge the error coefficient volumes corresponding to each phase zone within each layer segment to obtain the error coefficient volume between the low-frequency logging layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each layer segment.

[0114] For example, following the same method as obtaining the error coefficient body under the phase constraint of Facies1 in the L5 layer segment, the error coefficient bodies under the constraints of the remaining f-1 phase bands in the L5 layer segment are obtained in sequence, and the error coefficient bodies of each phase band obtained in the L5 layer segment are merged to obtain the error coefficient body under the phase constraint of the L5 layer segment.

[0115] In some implementations, step S130 merges the error coefficient volumes corresponding to each segment to obtain a low-frequency error coefficient volume, including:

[0116] For each segment (e.g., K+1 (K=13) segments) in the time-domain geological structure model, the error coefficient values ​​corresponding to each segment are merged in the vertical direction to obtain the low-frequency error coefficient volume Vel_Cor of the entire geological structure model.

[0117] In some implementations, step S140 corrects the pre-stack depth migration layer velocity volume based on the low-frequency error coefficient volume to obtain the low-frequency layer velocity volume, including:

[0118] The low-frequency error coefficient volume is multiplied by the pre-stack depth migration layer velocity volume to obtain the low-frequency layer velocity volume.

[0119] Specifically, for a certain CDP trace of a three-dimensional seismic body, the error coefficient value and the pre-stack depth migration layer velocity value of the i-th sampling point of the trace are Vel_Cor(i) and Interval_Vel(i), respectively. The corrected low-frequency layer velocity value LowFreq_Vel(i) of the i-th sampling point of the trace can be obtained by using formula (5).

[0120] LowFreq Vel(i) =Vel Cor(i) ×Interval Vel(i) (i = 1, 2, ..., Num) (5)

[0121] Where Num is the number of sampling points for each CDP trace, which is determined by the top time (Top), bottom time (Base), and sampling interval (m) of the time-domain geological structure model, Num = (Base - Top) / m + 1. For each CDP trace in the 3D seismic data volume, following the same calculation method as for the CDP traces above, the corrected low-frequency layer velocity data for each CDP trace can be obtained, thus completing the correction of the 3D low-frequency velocity and obtaining the corrected low-frequency layer velocity volume LowFreq_Vel.

[0122] In some implementations, step S150 uses spectral simulation inversion technology based on well logging data and seismic data of the target area to obtain the mid-to-high frequency layer velocity volumes corresponding to each segment in the time-domain geological structure model, including:

[0123] Based on the density curves of each target well within the target area, the average density of each target well in each layer of the time-domain geological structure model is determined.

[0124] Multiply the velocity curve value of each well by the average density of each layer to obtain the impedance curve of each target well in each layer.

[0125] Based on the impedance curves of each target well in each layer, the wave impedance curves of each target well in the target area are obtained.

[0126] The wave impedance curve is transformed from the depth domain to the time domain to obtain the time domain wave impedance curve;

[0127] Based on time-domain impedance curves and seismic data, the mid-to-high frequency layer velocity volumes corresponding to each layer within the target area are obtained through spectral simulation inversion.

[0128] By merging the mid-to-high frequency layer velocity volumes corresponding to each layer segment within the target area, a mid-to-high frequency layer velocity volume is obtained.

[0129] Among them, spectral simulation inversion is a frequency domain logging constrained wave impedance inversion technique. It performs spectral analysis on seismic data and well logging data respectively, finds the frequency domain matching operator that can achieve the best match between the seismic spectrum and the wave impedance spectrum of the well, applies the frequency domain matching operator to the seismic data, and inversely calculates the time domain operator to participate in the inversion operation to obtain the inversion data volume.

[0130] In some implementations, based on time-domain impedance curves and seismic data, spectral simulation inversion is used to obtain the mid-to-high frequency layer velocity volumes corresponding to each layer within the target area, including:

[0131] Spectral analysis of seismic data yields the amplitude spectrum of the seismic traces near each target well. Spectral analysis of the time-domain impedance curves yields the amplitude spectrum of the impedance corresponding to each target well. Based on the relationship between impedance and reflection coefficient, the amplitude spectrum of the reflection coefficient is obtained using the impedance amplitude spectrum. Based on the reflection coefficient amplitude spectrum and the seismic trace amplitude spectrum, the frequency domain matching operator that achieves the highest matching degree between the reflection coefficient amplitude spectrum and the seismic trace amplitude spectrum is obtained. The frequency domain matching operator is then converted into a time domain matching operator. ； The time-domain matching operator is applied to the seismic data for inversion, yielding the reflection coefficients in the time domain. Based on the relationship between wave impedance and reflection coefficients, the inverted reflection coefficients are converted into wave impedance values, resulting in the wave impedance inversion results for the seismic traces near each target well. The obtained wave impedance inversion results are then supplemented with a low-frequency background to obtain the absolute wave impedance of the seismic traces near each target well. The absolute wave impedances of the seismic traces near all target wells are merged to obtain the spectral simulated wave impedance inversion volume. Layer by layer, the spectral simulated wave impedance inversion volume is divided by the average density of each layer to obtain the layer velocity volume in the mid-to-high frequency range for each layer within the target area. The specific process is as follows:

[0132] First, m wells within the work area that simultaneously possess both acoustic and density curves are selected as preferred wells. Based on the velocity curves of these m preferred wells, acoustic impedance curves are constructed and used as input data for spectral simulation inversion. To facilitate the conversion of the inverted impedance data volume to the velocity data volume, the density value of each segment is taken as the average density logging value of each segment, Den_Avg, during impedance curve construction. That is, the density value of each segment is taken as a constant. Taking segment L5 as an example, the average density of the m wells in segment L5 is calculated. The velocity curve value of each well is multiplied by the constant density value Den_Avg to obtain the impedance curve of the m wells in segment L5. The same method is used to construct the impedance curves of the remaining K segments, thus obtaining the acoustic impedance curves of the m wells within the work area.

[0133] Then, using the obtained time-depth relationship of each well, the impedance curve is transformed from the depth domain to the time domain to obtain the time-domain impedance curve Impedance_time. Using the time-domain impedance curves of m wells in the work area and 3D pre-stack depth migration seismic data, the high-frequency layer velocity data volume of the target layer can be obtained through spectral simulation inversion. Specifically, based on the m selected wells in the work area, m seismic traces are extracted near each well, and each trace is denoted as s. i (t), (i=1,2,…,m), let the seismic wavelet be denoted as b(t), and the reflection coefficient be denoted as r. i (t)(i=1,2,…,m), then the convolution model of the seismic trace can be expressed as:

[0134] s i (t)=r i (t)*b(t); (6)

[0135] Transforming it to the frequency domain, we get the following equation:

[0136]

[0137] In the formula, S i (ω) represents the amplitude spectrum of the seismic trace, φ i (ω) represents the phase spectrum of the seismic trace, R i (ω) is the amplitude spectrum of the reflection coefficient, Ψ i (ω) is the phase spectrum of the reflection coefficient, B i (ω) represents the amplitude spectrum of the wavelet. Let be the phase spectrum of the wavelet, then according to formula (7), we can obtain:

[0138] S i (ω)=R i (ω)B i (ω); (8)

[0139] Taking the logarithm of formula (8), we get:

[0140] lnS i (ω)=lnR i (ω)+lnB i (ω); (9)

[0141] Assume the wave impedance at each preferred well is z i (t), (i = 1, 2, ..., m), then transforming it to the frequency domain yields:

[0142]

[0143] In the formula, Z i (ω) represents the wave impedance amplitude spectrum at the preferred well location, θ i(ω) represents the phase spectrum of the preferred well's wave impedance. Based on the industry-standard relationship between wave impedance and reflection coefficient, the amplitude spectrum of the reflection coefficient R can be obtained using the wave impedance amplitude spectrum. i (ω), the amplitude spectrum R of the reflection coefficient i (ω) and the amplitude spectrum S of the seismic trace i Substituting (ω) into formula (9), we can obtain the frequency matching operator B. i (ω).

[0144] Applying the matching operator obtained in the frequency domain to the seismic data yields the inversion results of the spectral simulation. Let A... i (ω)=1 / B i (ω), according to formula (8), we can obtain:

[0145]

[0146] Taking the inverse Fourier transform of formula (11) to the time domain yields the formula for calculating the reflection coefficient:

[0147] r i (t)=a i (t)*s i (t); (12)

[0148] In the formula, a i (t) is A i The time-domain function of (ω) can be obtained through the frequency matching operator B. i (ω) is obtained.

[0149] For a seismic trace at a CDP within a 3D work area, using formula (12), through the time domain operator a i (t) and earthquake traces i The reflection coefficient inversion result can be obtained by convolution of (t). Then, using the industry-standard correspondence between reflection coefficient and wave impedance, the reflection coefficient is converted into wave impedance value to obtain the wave impedance inversion result of the seismic trace. The wave impedance obtained at this time is the relative wave impedance. In order to analyze the inversion effect of the spectral simulation wave impedance more intuitively and facilitate the inversion quality control using wells, it is necessary to add low-frequency background to the obtained wave impedance inversion result to obtain the final absolute wave impedance value of the spectral simulation.

[0150] Following this method, the above calculations are performed on each seismic trace within the three-dimensional work area to obtain the spectral simulated wave impedance inversion volume. Then, the wave impedance inversion volume is divided layer by layer by the average density of each layer, Den_Avg, to obtain the mid-to-high frequency band layer velocity volume, denoted as HighFreq_Vel.

[0151] Taking the L5 section as an example, the average density of m wells in the L5 section is statistically analyzed. The wave impedance inversion volume of the L5 section is divided by the constant density value Den_Avg to obtain the mid-to-high frequency layer velocity value of the L5 section. Using a similar method to the calculation of the mid-to-high frequency layer velocity curve of the L5 section, the mid-to-high frequency layer velocity curves of the remaining K sections are calculated to obtain the mid-to-high frequency layer velocity volume HighFreq_Vel.

[0152] In some implementations, step S160, fusing the low-frequency layer velocity volume and the mid-to-high frequency layer velocity volume, may include:

[0153] Based on the spectral range of the low-frequency layer velocity volume, determine the high-pass frequency parameters and high-cutoff frequency parameters of the low-pass filter;

[0154] Based on the high-pass frequency parameters and high-cutoff frequency parameters of low-pass filtering, the low-frequency layer velocity volume is low-pass filtered to obtain the filtered low-frequency layer velocity volume.

[0155] Based on the spectral range of the mid-to-high frequency layer velocity volume, determine the low-pass frequency parameters and low-cutoff frequency parameters of the high-pass filter;

[0156] Based on the low-pass frequency parameters and low-cutoff frequency parameters of the high-pass filter, the mid-to-high frequency band layer velocity volume is subjected to high-pass filtering to obtain the filtered mid-to-high frequency band layer velocity volume.

[0157] Data fusion is performed on the filtered low-frequency and mid-to-high-frequency layer velocity volumes according to the predetermined weight values ​​for low-frequency and mid-to-high-frequency layers.

[0158] For example, the weight values ​​of low-frequency W_Low and mid-to-high-frequency W_High can be determined by analyzing the fit with the well. Generally, they are between 1 and 1.5, with higher weights for higher fit and lower weights for lower fit. According to the predetermined weight values ​​of low-frequency and mid-to-high-frequency, the filtered low-frequency layer velocity volume and the mid-to-high-frequency layer velocity volume are fused according to formula (13) to obtain the layer velocity model Vel_Merge.

[0159] For any sampling point i in a CDP trace of a 3D seismic body, the corresponding filtered low-frequency layer velocity value LowFreq_Vel_Filter(i) and mid-to-high frequency layer velocity value HighFreq_Vel_Filter(i) are read sequentially. Then, the layer velocity value Vel_Merge(i) after data fusion at that point can be obtained using formula (13):

[0160] Vel_Mergel(i)=LowFreq_Vel_Filter(i)×W_Low+HighFreq_Vel_Filter(i)×W_High; (13)

[0161] Following the above method, the low-frequency and mid-to-high-frequency layer velocity data of each sampling point of the 3D seismic body are fused and calculated to obtain the fused layer velocity data of each sampling point, forming a fused layer velocity volume. A velocity model can be established based on the fused layer velocity volume.

[0162] In some implementations, after step S160 is executed, the method may further include:

[0163] A velocity model is used to identify low-amplitude traps within the target area;

[0164] Based on drilling data of target wells within the target area, the reliability of the velocity model identification results is verified.

[0165] For example, the velocity volume Vel_tradition obtained by conventional methods and the velocity volume Vel_Merge obtained by the frequency band inversion velocity modeling method in this invention can be used to perform time-depth conversion and velocity mapping on the underlying strata of the strongly transversely heterogeneous strata in the study area. Combined with the characteristics of drilled wells, the structural maps obtained based on the two different velocity models are compared. The analysis results show that the former cannot clearly represent the characteristics of low-amplitude traps and has low consistency with drilled wells; the latter, due to the enhanced geological constraints achieved by using methods such as tomography and phasing control during the velocity model establishment process, can accurately represent the velocity variation characteristics of strongly transversely heterogeneous strata. The geological consistency and accuracy of the velocity model are relatively high. The structural maps obtained from this can clearly represent the trap characteristics of low-amplitude structures and have a high consistency with drilled wells, thus verifying the reliability of the method of this patent.

[0166] To better understand this invention, the following example, using a low-amplitude trap exploration area with underlying strata of heterogeneous strata exhibiting strong lateral structure, will be used to illustrate the above-mentioned velocity model construction method in detail:

[0167] Step (1): Obtain well logging data and seismic data; select m wells in the study area that simultaneously have sonic transit time curves and density curves as preferred wells. In this embodiment, m = 86; obtain the time-depth relationship of the preferred m wells, that is, use the sonic transit time curves and density curves of the preferred wells in the study area to make synthetic seismic records, convert the depth domain well logging curves to the time domain, perform seismic geological stratigraphic calibration on K (K = 13) stratigraphic interfaces in the study area, establish a one-to-one correspondence between the geological strata in the depth domain and the seismic reflection phase axis in the time domain, and obtain the time-depth relationship of each well. Figure 2 This is a schematic diagram of a time-domain pre-stack depth migration seismic profile provided in an embodiment of the present invention, as shown below. Figure 2As shown, a one-to-one correspondence between K layered interfaces (K=13) and seismic reflection axes in the study area was established through synthetic records. Synthetic records were produced for 86 selected wells, obtaining detailed time-depth relationships for these 86 wells. Time-domain seismic interpretation fault and horizon data were acquired. Based on the calibration results of the synthetic records, seismic horizons and faults were traced and interpreted on the seismic reflection phase axes corresponding to the calibrated K layered interfaces within the time-domain pre-stack depth migration seismic data volume, thus obtaining time-domain seismic interpretation horizon and fault data. Based on this, a geological structure model was created. Specifically, firstly, the combination patterns of each fault and the horizons it traverses were clarified, such as... Figure 2 As shown, the faults are mainly alternating graben and horst faults. For example, F2 and F3 are geoembedded structures, while F3 and F4 are horst structures. F1 breaks through 10 strata from H4 to H13, F3 breaks through 7 strata from H3 to H9, F7 and F11 break through 4 strata from H3 to H6, and other faults break through 5 strata from H3 to H7. Furthermore, based on six different stratigraphic contact modes—overlap, erosion, erosion, parallel bottom, parallel top, and parallel top-bottom—the characteristics of each fault in the study area are clearly defined. Regarding the contact relationships of the strata, in this embodiment, H1, H2, and H12 represent parallel bottom contact patterns; H4, H5, H6, H7, H8, H9, and H11 represent parallel top-bottom contact patterns; H10 and H13 represent parallel top contact patterns; and H3 represents a stratum erosion pattern. Then, the longitudinal time range of the geological structure model is set (in this embodiment, the top time is set to 0 milliseconds, and the bottom time is set to 3000 milliseconds). Based on the interpretation mode of the seismic profile, such as... Figure 2 As shown, following a bottom-up order, the contact relationships and fault combination patterns of each stratum are set, and the K seismic interpretation horizons and their corresponding fault data in the study area are modeled sequentially to obtain a time-domain geological structure model, such as... Figure 3 As shown, Figure 3 This is a cross-sectional view of a time-domain geological structure model provided in an embodiment of the present invention, such as... Figure 3 As shown, the model consists of K+1 (K=13) strata, which are L1 to L14 from top to bottom.

[0168] Step (2): First, obtain the time-domain pre-stack depth migration layer velocity volume, denoted as Interval_Vel; second, based on the layer velocity volume, extract the pre-stack depth migration layer velocity curves of the well-side traces of m preferred wells. The pre-stack depth migration layer velocity curves of the well-side traces can be obtained by the following method: according to the coordinate position of the well location, find the seismic trace closest to the coordinates of the well location, which is the well-side trace; using the sampling interval of the time-domain pre-stack depth migration layer velocity volume as the reading interval, sequentially read the pre-stack depth migration layer velocity values ​​of each point of the well-side trace from the first sampling point to the last sampling point, and connect the values ​​of each point read into a line to obtain the pre-stack depth migration layer velocity curve of the well-side trace, denoted as curve Psdm-Vel. In this embodiment, taking Well1 well as an example, the nearest seismic trace Well1-CDP is found based on the coordinates of Well1 well. According to the sampling interval of the pre-stack depth migration velocity volume, the pre-stack depth migration velocity values ​​of each point of the Well1-CDP seismic trace from the first sampling point to the last sampling point are read sequentially from the pre-stack depth migration velocity volume. The values ​​of each point read are then connected to form a line, thus obtaining the pre-stack depth migration velocity curve Psdm-Vel of the well-side trace. Figure 4 This is a comparison chart of low-frequency correction of logging curves provided in an embodiment of the present invention, such as... Figure 4 As shown, Figure 4 (a) shows a well logging velocity curve. Figure 4 (b) shows the pre-stack depth migration velocity curve of a Well1 well bypass.

[0169] Step (3): Based on the sonic transit time curve, obtain the low-frequency layer velocity of the well logging, and calculate the error coefficient between the low-frequency layer velocity and the pre-stack depth migration layer velocity. First, use formula (1) to convert the sonic transit time curve into a layer velocity curve; then, based on the established time-depth relationship of m wells in the study area, convert the depth domain layer velocity curve into the time domain to obtain the time domain layer velocity curve; at the same time, based on the high-frequency end value H1 of the effective frequency band of the pre-stack migration velocity body, use H1 as the high-pass frequency value of the low-pass filter to perform low-pass filtering on the time domain well logging curve to obtain the low-frequency layer velocity curve, denoted as curve Vel-Lowfreq; based on this, use formula (2) to obtain the error coefficient C by dividing the low-frequency layer velocity of the well logging by the pre-stack depth migration layer velocity. In this embodiment, using the time-depth relationship of each well and formula (1), the logging layer velocity curves were obtained from the acoustic transit time curves of the selected m wells in the study area. Based on the effective frequency band range of the pre-stack depth-shifted layer velocity volume in the time domain, the value of its high-frequency end H1 was determined to be 5 Hz. H1 was used as the high-pass frequency value for low-pass filtering. Low-pass filtering was applied to the time-domain logging curves to obtain the logging low-frequency layer velocity curve Vel-Lowfreq, as shown below. Figure 4 As shown, Figure 4(c) shows the low-frequency layer velocity curve of Well1. Based on this, according to the layer velocity curve Vel-Lowfreq and the pre-stack depth migration layer velocity curve Psdm-Vel, the error coefficient C of the low-frequency layer velocity correction is calculated using formula (2), as follows: Figure 4 As shown, Figure 4 (d) shows an error coefficient curve between the low-frequency layer velocity from the Well1 well logging and the pre-stack depth-shifted layer velocity.

[0170] Step (4) involves acquiring the facies boundary data for each layer, and using tomographic phasing interpolation technology to calculate the error coefficients for each layer, thereby obtaining the low-frequency error coefficient body. This is based on the geological structure model and the error coefficients between the low-frequency logging velocity and the pre-stack depth migration velocity obtained in step (3) for each well. Phase-controlled kriging interpolation is then performed layer by layer in the geological structure model to obtain the error coefficients for each layer. In this embodiment, as shown... Figure 3 As shown, the geological structure model contains K+1 (K=13) layers (L1~L14). Phase-controlled kriging interpolation is performed on each of these 14 layers to obtain the error coefficient volume.

[0171] Taking a certain segment L5 as an example, firstly, based on the contact relationship between the top and bottom layers of segment L5, P sub-layers are interpolated between the top and bottom layers of segment L5 according to a certain ratio P (P = 30 in this example). The top layer, bottom layer, and P interpolated sub-layers of segment L5 constitute the interpolation lattice model L5-Model of segment L5. Using the phase zone of segment L5 as the first constraint condition and the lattice model L5-Model as the second constraint condition for interpolation of segment L5, the error coefficient of segment L5 can be obtained by ordinary kriging interpolation method. Using f phase zones of segment L5 as constraints, kriging interpolation is performed in each phase zone. In this embodiment, segment L5 mainly develops three phase zones: Facies1, Facies2, and Facies (f = 3). For a certain phase zone Facies1 in segment L5, the velocity error coefficient within the Facies1 phase zone of segment L5 is obtained by kriging interpolation method based on the velocity error coefficient C of n wells located in that phase zone.

[0172] Specifically, the boundary of the Facies1 facies zone in the L5 section is used as the boundary constraint for interpolation of the Facies1 facies zone in the L5 section. Within this facies zone, based on the velocity error coefficients C of n wells located within this facies zone, for a certain HorizonM layer in the lattice model (there are P+2 layers in the L5 section), the error coefficient value at each CDP along this layer in the three-dimensional work area is regarded as a regional variable V(x). For the n wells within this facies zone (n=20 in this embodiment), an error coefficient value can be read at each well along this layer. The error coefficient values ​​read from the n wells can be regarded as n observations, namely V(x1), V(x2), V(x3), ..., V(x4). n For the error coefficient at a certain CDP along the layer in the work area, the estimated value is denoted as V(x0). Formula (3) can be applied to obtain the weighted value of the observations of n known sampling points, that is, to obtain the error coefficient estimate of the layer HorizonM at a certain CDP in the lattice model. According to this method, the estimated value of each CDP along the layer HorizonM in the three-dimensional area is calculated in turn, and the error coefficient value along the layer HorizonM is obtained. For each layer in the interpolated lattice model of the L5 layer segment, that is, a total of P+2 layers including the top layer, bottom layer and interpolated sub-layer of the L5 layer segment, the error coefficient value of each layer in the L5 layer lattice model can be obtained in turn according to the method of obtaining the error coefficient value along the layer HorizonM, and thus the error coefficient body of the L5 layer segment under the constraint of phase band Facies1 is obtained.

[0173] Following the same method as obtaining the error coefficient body under the Facies1 phase band constraint of the L5 layer, the error coefficient bodies under the constraints of the remaining f-1 phase bands of the L5 layer are obtained in sequence, and the error coefficient bodies of each phase band obtained in the L5 layer are merged to obtain the error coefficient body under the phase control constraint of the L5 layer.

[0174] The error coefficients of each segment in the geological structure model are merged to obtain the low-frequency segment error coefficient body Vel_Cor of the entire geological structure model. In this embodiment, the error coefficient values ​​of each of the K+1 (K=13) segments in the geological structure model are merged in the vertical direction to obtain the low-frequency segment error coefficient body Vel_Cor of the entire geological structure model.

[0175] Step (5) uses the low-frequency error coefficient volume to correct the pre-stack depth migration velocity volume, thus obtaining the corrected low-frequency layer velocity volume. In this embodiment, the specific steps are as follows: for a certain CDP trace of a three-dimensional seismic body, the error coefficient value and the pre-stack depth migration velocity value of any sampling point i in the trace are Vel_Cor(i) and Interval_Vel(i), respectively. The corrected low-frequency layer velocity value LowFreq_Vel(i) of the sampling point in the trace can be calculated using formula (5). The corrected low-frequency layer velocity value of each sampling point in the seismic trace is obtained in the same way.

[0176] For each CDP trace in the 3D seismic data volume, following the same calculation method as the seismic traces above, the corrected low-frequency layer velocity data for each CDP trace can be obtained, thus completing the correction of the 3D low-frequency velocity and obtaining the corrected low-frequency layer velocity volume LowFreq_Vel. For example... Figure 4 As shown, Figure 4 (e) shows a low-frequency corrected layer velocity curve of Well1 well. Figure 5 This is a cross-sectional view of a low-frequency layer velocity volume provided in an embodiment of the present invention, such as... Figure 5 As shown, the pre-stack depth offset layer velocity volume has been corrected based on the low-frequency error coefficient volume, resulting in the corrected low-frequency layer velocity volume.

[0177] Step (6) involves using spectral simulation inversion to obtain the layer velocity volume in the mid-to-high frequency band. Simulation inversion is a frequency domain logging constrained wave impedance inversion technique, which can be implemented through the following methods.

[0178] First, based on the velocity curves of the m selected wells in the work area, wave impedance curves are constructed and used as input data for spectral simulation inversion. To facilitate the conversion of the inverted wave impedance data volume to the velocity data volume, the density of each segment is taken as the average value of the density logging values ​​of each segment, Den_Avg, when constructing the impedance curve. That is, the density value of each segment is taken as a constant. Taking the L5 segment as an example, the average density of m wells in the L5 segment, Den_Avg, is calculated (in this embodiment, the average value of the density logging curve of the L5 segment is 2.45, that is, Den_Avg = 2.45). The velocity curve value of each well is multiplied by the constant density value Den_Avg to obtain the wave impedance curve of m wells in the L5 segment. Using the same method as the L5 segment impedance curve, the impedance curves of the remaining K segments are constructed to obtain the wave impedance curves of m wells in the work area. Then, using the time-depth relationship of each well obtained in step (1), the wave impedance curves of m wells are converted from the depth domain to the time domain to obtain the time domain impedance curve Impedance_time.

[0179] Using the time-domain impedance curves and 3D pre-stack depth migration seismic data from m wells within the work area, the mid-to-high frequency layer velocity data volume is calculated through spectral simulation inversion. Specifically, this can be divided into two programs. The first program uses spectral analysis of well logging impedance data and well-side seismic traces to determine the frequency domain matching operator that achieves optimal matching between the seismic spectrum and the well logging impedance spectrum. Based on the selected m wells within the work area, m well-side seismic traces are extracted at the selected well locations, each trace denoted as s. i (t), (i = 1, 2, ..., m), the seismic wavelet is denoted as b(t), and the reflection coefficient is denoted as r. i (t)(i=1,2,…,m), then based on the convolution model of seismic data (formula (6)) and the time-frequency transformation formula (7), the amplitude spectrum S of the earthquake can be obtained. i (ω), amplitude spectrum of reflection coefficient R i (ω) and the amplitude spectrum of the wavelet B i The relationship between (ω) is used. The frequency matching operator is calculated using formula (9). Since the wave impedance of the preferred well is known, the amplitude spectrum of the reflection coefficient R can be obtained using the amplitude spectrum of the wave impedance, based on the industry-standard correspondence between wave impedance and reflection coefficient. i (ω), and the amplitude spectrum R of the reflection coefficient i (ω) and the amplitude spectrum S of the seismic trace i Substituting (ω) into formula (9), we can obtain the frequency matching operator B. i (ω).

[0180] The second procedure involves applying a frequency domain matching operator to the seismic data to obtain the spectral simulation inversion data volume. This is based on the seismic amplitude spectrum S... i (ω), the amplitude spectrum of the reflection coefficient R i (ω) and the amplitude spectrum of the wavelet B i The relationship between (ω) is given by taking the reciprocal of the frequency domain matching operator as A. i (ω), i.e., A i (ω)=1 / B i (ω), then the reflection coefficient can be calculated using formula (12), where a i (t) is A i (ω) is a time-domain function that can be obtained through the frequency matching operator B. i (ω) is obtained. Therefore, according to formula (12), for a seismic trace at a certain CDP within the three-dimensional work area, the time domain operator a is used to obtain the seismic trace. i (t) and earthquake traces iThe reflection coefficient of the seismic trace can be obtained by convolution of (t). Then, using the industry-standard correspondence between reflection coefficient and wave impedance, the reflection coefficient is converted to wave impedance value to obtain the wave impedance inversion result of the seismic trace. The wave impedance obtained at this time is the relative wave impedance. In order to analyze the inversion effect of the spectral simulation wave impedance more intuitively and facilitate the inversion quality control by well logging, it is necessary to add the low-frequency background to the obtained mid-frequency relative wave impedance inversion result. That is, add the low-frequency background value (a low-frequency impedance value is assigned to each layer according to the well logging analysis results) to the relative impedance inversion value obtained by spectral simulation to obtain the final absolute wave impedance value of spectral simulation.

[0181] Following this method, the above calculations are performed on each seismic trace within the three-dimensional work area to obtain the spectral simulated wave impedance inversion volume. Then, layer by layer, the wave impedance inversion volume is divided by the average density value Den_Avg of each segment to obtain the mid-to-high frequency layer velocity volume, denoted as HighFreq_Vel. Taking the L5 segment as an example, the average density of m wells in the L5 segment is statistically analyzed. The wave impedance inversion volume of the L5 segment is divided by the constant density value Den_Avg to obtain the mid-to-high frequency layer velocity value of the L5 segment. Using a similar method to the calculation of the mid-to-high frequency layer velocity curve of the L5 segment, the mid-to-high frequency layer velocity curves of the remaining K segments are calculated to obtain the mid-to-high frequency layer velocity volume.

[0182] In this embodiment, firstly, based on the density curves of the selected m wells in the work area, the average density value Den_Avg of each layer is statistically analyzed. The velocity curve value of the m wells is multiplied by the average density value Den_Avg of each layer to obtain the impedance curve of the selected wells. Then, using the time-depth relationship of each well obtained in step (1), the impedance curve of the selected wells is transformed from the depth domain to the time domain. Secondly, based on the industry-standard correspondence between reflection coefficient and wave impedance, the logging wave impedance amplitude spectrum of the m wells is converted into the reflection coefficient amplitude spectrum, and the logging wave impedance amplitude spectrum R of the m wells is extracted. i (ω), calculate the wellbore amplitude spectrum S of the preferred m wells. i (ω), and thus the frequency domain matching operator B for spectral simulation inversion is obtained according to formula (9). i (ω). Then, the reciprocal of the frequency domain matching operator is taken and inverse Fourier transformed to the time domain to obtain the time domain operator a. i (t), using formula (12), through the time domain operator a i (t) and each seismic trace in the three-dimensional seismic body iThe reflection coefficient is obtained by convolution of (t). Then, using the industry-standard correspondence between reflection coefficient and wave impedance, the reflection coefficient is converted to wave impedance value to obtain the relative wave impedance inversion result. Based on the statistical results of well logging data, a low-frequency background value is set for each layer (the low-frequency background value is taken as the statistical average of the well logging wave impedance curves of each layer, such as 10050 and 11250 for L5 and L6 layers respectively), to obtain the final absolute wave impedance volume of the spectrum simulation. Then, the wave impedance inversion volume is divided by the density mean of each layer to obtain the inversion layer velocity volume HighFreq_Vel in the mid-to-high frequency range. Figure 6 As shown, Figure 6 This is a cross-sectional view of a mid-to-high frequency band layer velocity body provided in an embodiment of the present invention.

[0183] Step (7) is to establish a layer velocity model by fusing data from low-frequency and mid-to-high-frequency layer velocity volumes. First, the low-frequency layer velocity volume LowFreq_Vel is low-pass filtered. Based on the spectral range of the low-frequency layer velocity volume, the high-pass frequency parameter HP and the high-cutoff frequency parameter HC of the low-pass filter are determined to obtain the filtered low-frequency layer velocity volume LowFreq_Vel_Filter. The mid-to-high frequency layer velocity volume is high-pass filtered. Based on the spectral range of the mid-to-high frequency layer velocity volume, the low-pass frequency parameter LP and the low-cutoff frequency parameter LC of the high-pass filter are determined to obtain the filtered mid-to-high frequency layer velocity volume HighFreq_Vel_Filter. Second, the velocity values ​​of the low-frequency and mid-to-high frequency bands are weighted. The values ​​of the low-frequency weight W_Low and the mid-to-high frequency weight W_High are determined by analyzing the fit with the well. Generally, they are between 1 and 1.5. The higher the fit, the higher the weight, and the lower the fit, the lower the weight. The filtered data are weighted and fused according to formula (13) to obtain the layer velocity model Vel_Merge. For any sampling point i in a CDP trace of a 3D seismic body, read its corresponding low-frequency layer velocity value LowFreq_Vel_Filter(i) and mid-to-high frequency layer velocity value HighFreq_Vel_Filter(i) in sequence, and then use the above formula (13) to obtain the layer velocity value Vel_Merge(i) after data fusion at that point.

[0184] Following the method described above, the low-frequency and mid-to-high-frequency layer velocity data of each sampling point of the 3D seismic body are fused and calculated to obtain the fused layer velocity data of each sampling point.

[0185] In this embodiment, based on the spectrum of low-frequency layer velocity and the spectrum of mid-to-high frequency layer velocity, the high-pass frequency parameter HP and high-cutoff frequency parameter HC of the low-frequency layer velocity low-pass filter are set to 5Hz and 8Hz, respectively, and the low-pass frequency parameter LP and low-cutoff frequency parameter LC of the mid-to-high frequency layer velocity high-pass filter are set to 8Hz and 5Hz, respectively. On this basis, through the analysis of well consistency, the low-frequency weight W_Low and the mid-to-high frequency weight W_High values ​​are determined to be 1 and 1.05, respectively. For each sampling point of the three-dimensional seismic body, the fused layer velocity value of each sampling point is calculated using formula (13), thereby realizing the data fusion of the low-frequency layer velocity body and the mid-to-high frequency layer velocity body. The required velocity model can be established based on the fused layer velocity body Vel_Merge. Figure 7 As shown, Figure 7 This is a cross-sectional view of a velocity model provided in an embodiment of the present invention.

[0186] Step (8) analyzes the reliability of the method in this patent by comparing the structural features after time-depth conversion based on different velocity models. The layer velocity volume Vel_tradition obtained by conventional methods and the layer velocity volume Vel_Merge obtained by the frequency-band inversion velocity modeling method in this invention are used to perform time-depth conversion and variable velocity mapping on the underlying strata of the strongly transversely heterogeneous strata in the study area. Combined with the characteristics of drilled wells, the structural maps obtained based on the two different velocity models are compared. Figure 8 This is a comparison diagram of the low-amplitude trap structure features based on the conventional velocity model and the velocity model of this application, such as... Figure 8 As shown, the analysis results indicate that: the former cannot clearly characterize the features of low-amplitude traps and has low consistency with drilled wells. In this example, the drilled producing well Well-1 is not included in the structural trap after time-depth conversion based on the traditional velocity model; the trap amplitude is only 7 meters. The depth of the drilled well in this layer is -2522 meters, while the depth after time-depth conversion based on the traditional velocity model is -2547 meters, with an error of 25 meters. The time-depth conversion error is large, and the trap characterization is low. The latter, due to the enhanced geological constraints achieved by using methods such as tomographic phasing during the velocity model establishment process, can accurately characterize the velocity variation features of strongly lateral structural heterogeneous formations. The velocity model... The geological fit and precision are both high, and the resulting structural map can clearly characterize the trap features of low-amplitude structures. It also has a high degree of fit with the drilled wells. In this embodiment, the drilled oil well Well-1 is located in the structural trap after time-depth conversion based on the method of this patent. The trap amplitude is 16 meters, the depth of the drilled well in this layer is -2522 meters, and the depth after time-depth conversion based on the method of this patent is -2520 meters, with an error of 2 meters. The time-depth conversion error is small, the characterization of low-amplitude traps is high, the identification and characterization of low-amplitude traps are significantly improved, and the time-depth conversion depth error is significantly reduced, thereby verifying the reliability of the velocity model constructed by this invention.

[0187] Based on the same inventive concept, this invention also provides a velocity model construction device for structurally heterogeneous layers, used to construct a velocity model for structurally heterogeneous layers. Figure 9 This is a structural block diagram of a velocity model construction device for a structurally heterogeneous layer provided in an embodiment of the present invention, as shown below. Figure 9 As shown, the velocity model building device 900 includes:

[0188] The structural model building module 901 is used to build a time-domain geological structure model based on well logging data and seismic data of the target area.

[0189] The first error determination module 902 is used to obtain the error coefficient volume between the low-frequency well logging velocity and the three-dimensional pre-stack depth migration velocity in each segment of the geological structure model based on the pre-stack depth migration layer velocity volume and well logging data of the target area, using tomographic phase control interpolation technology.

[0190] The second error determination module 903 is used to merge the error coefficient volumes corresponding to each segment to obtain the low-frequency error coefficient volume.

[0191] The first velocity volume construction module 904 is used to correct the pre-stack depth offset layer velocity volume based on the low-frequency band error coefficient volume to obtain the low-frequency band layer velocity volume.

[0192] The second velocity volume construction module 905 is used to obtain the mid-to-high frequency layer velocity volume based on well logging data and seismic data of the target area using the spectrum simulation inversion method;

[0193] The velocity model building module 906 is used to fuse the low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume, and to build a velocity model based on the fused layer velocity volume.

[0194] The specific details of the speed model construction method involved in the above control device can be understood by referring to the relevant descriptions and effects in the above-described speed model construction method embodiments, and will not be repeated here.

[0195] Based on the same inventive concept as the above method, the present invention also provides an electronic device, which may include a processor and a memory, wherein the processor and the memory can communicate with each other via a bus or other means. The processor may be a Central Processing Unit (CPU). The processor may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof. The memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the speed model construction method in the embodiments of the present invention. The processor executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory, thereby implementing the speed model construction method in the above method embodiments.

[0196] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. The one or more modules are stored in the memory and, when executed by the processor, perform actions such as... Figure 1 The velocity model construction method in the illustrated embodiment.

[0197] For specific details regarding the aforementioned electronic devices, please refer to the relevant documentation. Figure 1 The relevant descriptions and effects in the illustrated embodiments are for understanding purposes only and will not be repeated here.

[0198] Based on the same inventive concept as the speed model construction method, the present invention also provides a computer-readable storage medium storing computer instructions, which are used to cause a computer to execute the speed model construction method in the above embodiments.

[0199] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium can be read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0200] The technical solutions provided in the above embodiments of this application have at least the following technical effects or advantages:

[0201] This invention provides a method, apparatus, device, and storage medium for constructing a velocity model of a structurally heterogeneous stratum. First, a time-domain geological structure model is established based on well logging and seismic data of the target area, providing an accurate geological framework for subsequent velocity model construction. Then, based on the pre-stack depth-migrated layer velocity volume and well logging data of the target area, tomographic phasing interpolation technology is used to calculate the error coefficient volume between the low-frequency well logging layer velocities and the three-dimensional pre-stack depth-migrated layer velocities corresponding to each segment in the geological structure model. The error coefficient volumes corresponding to each segment are merged to obtain a low-frequency error coefficient volume, which is used to correct the pre-stack depth-migrated layer velocity volume, resulting in a corrected low-frequency layer velocity volume, thereby improving the low-frequency characterization accuracy of the velocity model. Simultaneously, using spectral simulation inversion technology, this invention combines well logging and seismic data to obtain mid-to-high frequency layer velocity volumes, further enriching the information content of the velocity model. Finally, through data fusion technology, this invention effectively fuses the low-frequency layer velocity volume and the mid-to-high frequency layer velocity volumes to obtain a refined velocity model with high geological characterization. This velocity model not only accurately reflects the velocity variation characteristics of strongly transversely structured heterogeneous strata, but also significantly improves the identification accuracy of low-amplitude traps and the ability to finely characterize structural features, providing a scientific basis for oil and gas exploration and development decisions and deployments. It has important practical application value and prospects for promotion.

[0202] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0203] Similarly, it should be understood that, in order to simplify this disclosure and aid in understanding one or more of the various aspects of the invention, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof. However, this method of disclosure should not be construed as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the following claims, inventive aspects lie in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into this detailed description, wherein each claim itself is a separate embodiment of the invention.

[0204] It should be noted that the above embodiments are illustrative of the invention and not restrictive of the invention, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims.

Claims

1. A method for constructing a velocity model of a structurally heterogeneous layer, characterized in that, include: A time-domain geological structure model is established based on well logging and seismic data of the target area. Based on the pre-stack depth migration layer velocity volume of the target area and the well logging data, the error coefficient volume between the well logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each layer in the geological structure model is obtained by using tomographic phase control interpolation technology. The error coefficient volumes corresponding to each segment are merged to obtain the low-frequency band error coefficient volume; Based on the low-frequency error coefficient volume, the pre-stack depth offset layer velocity volume is corrected to obtain the low-frequency layer velocity volume. Based on well logging and seismic data of the target area, the mid-to-high frequency layer velocity volume is obtained using spectral simulation inversion technology. The low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume are fused together, and a velocity model is established based on the fused layer velocity volume.

2. The method according to claim 1, characterized in that, The well logging data includes the sonic transit time curves and density curves of each target well within the target area. The seismic data is time-domain pre-stack depth migration seismic data. The establishment of a time-domain geological structure model based on the well logging data and seismic data of the target area includes: Based on the sonic transit time curves and density curves of each target well within the target area, a synthetic seismic record is generated; Seismic geological horizons were determined using synthetic seismic records, and the correspondence between geological horizons and seismic reflection phase axes was established. Based on the correspondence between the geological strata and the seismic reflection phase axis, the faults and strata corresponding to the geological strata are tracked and interpreted to obtain time-domain seismic interpretation strata and fault data. Based on the time-domain seismic interpretation horizon and fault data, a time-domain geological structure model is established.

3. The method according to claim 1, characterized in that, The pre-stack depth migration velocity volume based on the target area and the well logging data are used to obtain the error coefficient volume between the low-frequency well logging velocities and the three-dimensional pre-stack depth migration velocities in each segment of the geological structure model, using tomographic phasing interpolation technology. This includes: Based on the pre-stack depth migration layer velocity volume of the target area and the logging data, the error coefficient between the logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well in the target area is determined. Based on the contact relationship between the top and bottom layers of each segment in the geological structure model, a set number of sub-layers are inserted between the top and bottom layers of each segment to obtain the interpolated lattice model corresponding to each segment. Using the Kriging interpolation method, with the boundary data of each facies zone and the corresponding interpolation grid model of each facies zone as constraints, the error coefficient between the low-frequency logging velocity of the target well in each facies zone and the three-dimensional pre-stack depth migration velocity is determined based on the error coefficient between the low-frequency logging velocity of the target well in each facies zone and the three-dimensional pre-stack depth migration velocity. By merging the error coefficient volumes corresponding to each phase zone within each layer, the error coefficient volume between the low-frequency logging layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each layer is obtained.

4. The method according to claim 3, characterized in that, The method, based on the pre-stack depth migration layer velocity volume of the target region and the logging data, determines the error coefficient between the low-frequency logging layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well within the target region, including: Based on the pre-stack depth migration velocity volume of the target area, the pre-stack depth migration velocity curves corresponding to each target well in the target area are extracted. The pre-stack depth migration velocity curves corresponding to each target well are the three-dimensional pre-stack depth migration velocity curves of the seismic traces closest to each target well. Extract the sonic transit time curves of each target well within the target area from the logging data, and convert the sonic transit time curves into layer velocity curves; The layer velocity curve is transformed in the time domain to obtain the time domain layer velocity curve; The high-frequency end value of the effective frequency band of the three-dimensional pre-stack depth migration layer velocity volume is used as the high-pass frequency value of the low-pass filter to perform low-pass filtering on the time domain layer velocity curve to obtain the well logging low-frequency layer velocity curve. Based on the three-dimensional pre-stack depth migration layer velocity curve and the well logging low-frequency layer velocity curve, the error coefficient between the well logging low-frequency layer velocity and the three-dimensional pre-stack depth migration layer velocity corresponding to each target well in the target area is determined.

5. The method according to claim 1, characterized in that, The well logging data includes velocity and density curves of each target well within the target area, and the seismic data is time-domain pre-stack depth migration seismic data; based on the well logging data and seismic data of the target area, a mid-to-high frequency layer velocity volume is obtained using spectral simulation inversion technology, including: Based on the density curves of each target well within the target area, the average density of each target well in each layer of the geological structure model in the time domain is determined. Multiply the velocity curve value of each well by the average density of each layer to obtain the impedance curve of each target well in each layer. Based on the impedance curves of each target well in each layer, the wave impedance curves of each target well in the target area are obtained. The wave impedance curve is transformed from the depth domain to the time domain to obtain the time domain wave impedance curve. Based on the time-domain impedance curves and the seismic data, the mid-to-high frequency layer velocity volumes corresponding to each layer within the target area are obtained through spectral simulation inversion. The mid-to-high frequency segment layer velocity volumes corresponding to each segment within the target area are merged to obtain the mid-to-high frequency segment layer velocity volume.

6. The method according to claim 1, characterized in that, The process of fusing data from the low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume includes: Based on the spectral range of the low-frequency band layer velocity body, determine the high-pass frequency parameters and high-cutoff frequency parameters of the low-pass filter; Based on the high-pass frequency parameters and high-cutoff frequency parameters of the low-pass filter, the low-frequency band layer velocity volume is subjected to low-pass filtering to obtain the filtered low-frequency band layer velocity volume. Based on the spectral range of the mid-to-high frequency band layer velocity body, determine the low-pass frequency parameters and low-cutoff frequency parameters of the high-pass filter; Based on the low-pass frequency parameters and low-cutoff frequency parameters of the high-pass filter, the mid-to-high frequency band layer velocity volume is subjected to high-pass filtering to obtain the filtered mid-to-high frequency band layer velocity volume. According to the predetermined weight values ​​for low frequency and mid-to-high frequency, the filtered low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume are fused.

7. The method according to claim 1, characterized in that, The method further includes: The velocity model is used to identify low-amplitude traps within the target area; Based on the drilling data of the target wells within the target area, the reliability of the velocity model identification results is verified.

8. A device for constructing a velocity model of a structurally heterogeneous layer, characterized in that, include: The structural model building module is used to build a time-domain geological structure model based on well logging data and seismic data of the target area. The first error determination module is used to obtain the error coefficient volume between the low-frequency well logging velocity and the three-dimensional pre-stack depth migration velocity in each segment of the geological structure model based on the pre-stack depth migration layer velocity volume of the target area and the well logging data, using tomographic phase control interpolation technology. The second error determination module is used to merge the error coefficient bodies corresponding to each segment to obtain the low-frequency error coefficient body. The first velocity volume construction module is used to correct the pre-stack depth offset layer velocity volume based on the low-frequency band error coefficient volume to obtain the low-frequency band layer velocity volume. The second velocity volume construction module is used to obtain the mid-to-high frequency layer velocity volume based on well logging data and seismic data of the target area using the spectrum simulation inversion method; The velocity model building module is used to fuse the low-frequency band layer velocity volume and the mid-to-high frequency band layer velocity volume, and to build a velocity model based on the fused layer velocity volume.

9. An electronic device, characterized in that, include: A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the speed model construction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the speed model construction method according to any one of claims 1 to 7.

Images