Tomographic velocity model training method and device based on VSP data

Abstract

The present disclosure provides a method and apparatus for training a tomographic velocity model based on VSP data, belonging to the field of artificial intelligence technology. The method comprises: within the velocity field provided by the initial tomographic velocity model, using a depth-informed Fresnel ray tracing method to trace the paths of rays received at a detector point with depth information, obtaining the range of Fresnel path lengths in the depth direction within the velocity field; obtaining the inverted travel times of rays arriving at the detector point along different paths within the Fresnel path length range at different depths; and training the initial tomographic velocity model based on the inverted travel times of rays arriving at the detector point along different paths within the Fresnel path length range at different depths and an improved inversion objective function weighted by VSP data depth. The present disclosure improves the coverage density of first-arrival paths, resulting in better ray tomographic inversion results and further enhancing the accuracy of the constructed tomographic velocity model.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to a tomographic velocity model training method and device based on VSP data. Background Art

[0002] With the development of new geophysical exploration technologies, more refined velocity parameters are needed to address near-surface mountain static correction issues and provide more accurate velocity fields for migration imaging. To obtain these refined velocity parameters, ray theory and wave theory can be used to perform ray tomographic inversion on VSP (Vertical Seismic Profile) data and construct a tomographic velocity model.

[0003] However, due to the sparse data of the first arrival travel time of the VSP upgoing wave, the coverage density of the first arrival wave path is low when performing ray tomography inversion, resulting in poor ray tomography inversion effect, and further leading to low accuracy of the tomographic velocity model constructed based on the ray tomography inversion results. Summary of the Invention

[0004] The present disclosure provides a method and apparatus for training a tomographic velocity model based on VSP data, capable of constructing a tomographic velocity model with high accuracy. The technical solution is as follows:

[0005] In a first aspect, a method for training a tomographic velocity model based on VSP data is provided, the method comprising:

[0006] An initial tomographic velocity model is constructed based on the shot point elevation information, receiver point depth information and initial velocity values ​​in the vertical seismic profile VSP upgoing wave first arrival travel time data.

[0007] In the velocity field provided by the initial tomographic velocity model, a Fresnel body ray tracing method based on depth information is used to trace the path of the ray received by the detection point with depth information to obtain the Fresnel body path length range in the depth direction in the velocity field;

[0008] Performing an improved objective function inversion with VSP data depth weighting on the Fresnel body path length range to obtain the inverted travel time of the ray reaching the detection point along different paths within the Fresnel body path length range at different depths;

[0009] Based on the inverted travel times of the rays reaching the detection points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting, the model parameters of the initial tomographic velocity model are adjusted to obtain a trained tomographic velocity model.

[0010] In another embodiment of the present disclosure, constructing an initial tomographic velocity model based on the elevation information of the shot points, the depth information of the receiver points, and the initial velocity values ​​in the first arrival travel time data of the vertical seismic profile (VSP) includes:

[0011] Performing coordinate transformation on the depth information of the detection points to obtain elevation information of all points on the surface of the velocity field;

[0012] Performing interpolation calculation based on the elevation information of the detection point to obtain interpolation elevation information of the detection point;

[0013] The initial velocity model is constructed based on the elevation information of all points on the surface of the velocity field and the initial velocity values.

[0014] In another embodiment of the present disclosure, the Fresnel volume ray tracing method based on depth information performs path tracing on rays received by the detection point with depth information within the velocity field provided by the initial tomographic velocity model to obtain the Fresnel volume path length range in the depth direction within the velocity field, including:

[0015] performing path tracing on the ray received by the detection point at the position indicated by each depth information using a Fresnel volume ray tracing method based on depth information within the velocity field provided by the initial tomographic velocity model to obtain the path length of the ray received by the detection point at the position indicated by each depth information;

[0016] Calculating, based on the path length of the ray received by the detection point at the position indicated by each depth information, the inverted travel time of the ray with different path lengths at the position indicated by each depth information;

[0017] Determining the shortest path length of the detection point at the position indicated by each depth information based on the inverted travel time of rays of different path lengths at the position indicated by each depth information;

[0018] The shortest path lengths of the detection points at different positions constitute the Fresnel body path length range.

[0019] In another embodiment of the present disclosure, before obtaining the trained tomographic velocity model, the model parameters of the initial tomographic velocity model are adjusted based on the inverted travel times of the rays reaching the detection points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting, and the like, further comprising:

[0020] Determining optimal offset distances of the geophones at different depths based on the depth information of the geophones and the improved inversion objective function with VSP data depth weighting;

[0021] determining offset weights of Fresnel body paths of different depth information based on the distance between the detection point and the shot point emitting the ray and the optimal offset distance;

[0022] The inversion objective function is determined based on the Fresnel body paths of different depth information and their offset weights.

[0023] In another embodiment of the present disclosure, the model parameters of the initial tomographic velocity model are adjusted based on the inverted travel times of the rays reaching the detection points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting to obtain a trained tomographic velocity model, including:

[0024] Inputting the inverted travel time of the ray along different paths within the Fresnel body path length range at different depths to the detection point into the improved inversion objective function with VSP data depth weighting, and outputting the function value of the improved inversion objective function with VSP data depth weighting;

[0025] adjusting model parameters of the initial tomographic velocity model based on the function value;

[0026] Based on the tomographic velocity model after the model parameters are adjusted, continuing the Fresnel volume ray tracing method based on the depth information, tracing the path of the ray received by the detection point with the depth information until the training cutoff condition is met;

[0027] The tomographic velocity model obtained when the training cutoff condition is met is used as the trained tomographic velocity model.

[0028] In a second aspect, an improved tomographic velocity model training device based on VSP data is provided, the device comprising:

[0029] A construction module is used to construct an initial tomographic velocity model based on the shot point elevation information, receiver point depth information and initial velocity values ​​in the vertical seismic profile VSP upgoing wave first arrival travel time data;

[0030] a tracking module configured to perform path tracing on rays received by the detection points having depth information within the velocity field provided by the initial tomographic velocity model based on a Fresnel body ray tracing method using depth information, and obtain a range of Fresnel body path lengths in the depth direction within the velocity field;

[0031] an inversion module, configured to perform an improved VSP data depth-weighted objective function inversion on the Fresnel body path length range to obtain an inverted travel time of the ray reaching the detection point along different paths within the Fresnel body path length range at different depths;

[0032] An adjustment module is configured to adjust model parameters of the initial tomographic velocity model based on inversion travel times of the rays arriving at the detection points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting, so as to obtain a trained tomographic velocity model.

[0033] In another embodiment of the present disclosure, the construction module is configured to perform coordinate transformation on the depth information of the detection point to obtain elevation information of the detection point; perform interpolation calculation based on the elevation information of the detection point to obtain elevation information of all points on the surface layer of the velocity field; and construct the initial velocity model based on the elevation information of all points on the surface layer of the velocity field and the initial velocity value.

[0034] In another embodiment of the present disclosure, the tracking module is configured to perform path tracing on rays received by the detection point at the position indicated by each depth information within the velocity field provided by the initial tomographic velocity model using a Fresnel volume ray tracing method based on depth information to obtain a path length of the ray received by the detection point at the position indicated by each depth information; calculate, based on the path length of the ray received by the detection point at the position indicated by each depth information, an inverted travel time of rays with different path lengths at the position indicated by each depth information; determine, based on the inverted travel time of rays with different path lengths at the position indicated by each depth information, a shortest path length of the detection point at the position indicated by each depth information; and form the shortest path lengths of the detection point at different positions into a Fresnel volume path length range.

[0035] In another embodiment of the present disclosure, the apparatus further comprises:

[0036] a determination module, configured to determine optimal offset distances of detection points at different depths based on the depth information of the detection points and the improved inversion objective function with VSP data depth weighting;

[0037] The determining module is further configured to determine offset weights of Fresnel body paths of different depth information based on the distance between the detection point and the shot point emitting the ray and the optimal offset distance;

[0038] The determination module is further configured to determine the inversion objective function based on the Fresnel body paths of different depth information and their offset weights.

[0039] In another embodiment of the present disclosure, the adjustment module is configured to input the inverted travel times of the rays arriving at the detection points along different paths within a range of Fresnel body path lengths at different depths into the improved inversion objective function with VSP data depth weighting, and output a function value of the improved inversion objective function with VSP data depth weighting; adjust model parameters of the initial tomographic velocity model based on the function value; continue to perform path tracing on the rays received at the detection points with depth information using a Fresnel body ray tracing method based on depth information based on the tomographic velocity model after the model parameters are adjusted, until a training cutoff condition is met; and use the tomographic velocity model obtained when the training cutoff condition is met as the trained tomographic velocity model.

[0040] In a third aspect, an electronic device is provided, comprising a processor and a memory, wherein the memory stores at least one program code, and the at least one program code is loaded and executed by the processor to implement the tomographic velocity model training method based on VSP data as described in the first aspect.

[0041] In a fourth aspect, a computer-readable storage medium is provided, wherein the storage medium stores at least one program code, and the at least one program code is loaded and executed by a processor to implement the tomographic velocity model training method based on VSP data as described in the first aspect.

[0042] In a fifth aspect, a computer program product is provided, comprising computer program code, wherein the computer program code is stored in a computer-readable storage medium, a processor of an electronic device reads the computer program code from the computer-readable storage medium, and the processor executes the computer program code, so that the electronic device performs the tomographic velocity model training method based on VSP data as described in the first aspect.

[0043] The technical solutions provided by the embodiments of the present disclosure have the following beneficial effects:

[0044] Within the velocity field provided by the initial tomographic velocity model, a depth-informed Fresnel volume ray tracing method is used to trace the paths of rays received at depth-informed geophones. This method obtains the Fresnel volume path length range in the depth direction within the velocity field. This Fresnel volume path length range has a greater coverage density than a single ray path, resulting in better ray tomographic inversion results. The tomographic velocity model is trained using the inverted travel times of rays arriving at geophones along different Fresnel volume path lengths at different depths and an improved objective function weighted by VSP data depth, improving its accuracy. Furthermore, by adding an offset weight to the inversion objective function, the shadow region of the first-arrival ray path is reduced to almost zero, avoiding large-scale smoothing of the tomographic velocity model. This enhances the stability of the inversion solution from the data itself, effectively compensating for the shortcomings of the VSP data in inversion, and further ensuring the accuracy of the trained tomographic velocity model. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0046] Figure 1 This is a flow chart of a tomographic velocity model training method based on VSP data provided by an embodiment of the present disclosure;

[0047] Figure 2 This is a flow chart of a method for coordinate transformation of the elevation of a detection point in the first arrival traveltime data of a VSP upgoing wave provided by an embodiment of the present disclosure;

[0048] Figure 3 Schematic diagram of the position relationship between a detection point and a shot point provided by an embodiment of the present disclosure;

[0049] Figure 4 This is the effect diagram of direct interpolation of this related technology;

[0050] Figure 5 is an effect diagram after interpolation in the embodiment of the present disclosure;

[0051] Figure 6 is a schematic diagram of a ray path in the related art and a Fresnel path in the present disclosure;

[0052] Figure 7 is an effect diagram of static correction using the tomographic velocity model trained by the embodiment of the present disclosure;

[0053] Figure 8Schematic diagram of a tomographic velocity model training device based on VSP data provided by the present disclosure;

[0054] Figure 9 A structural block diagram of an electronic device provided by an exemplary embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0055] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0056] It should be understood that the terms "each," "plurality," and "any" used in the embodiments of the present disclosure include two or more, each refers to each of the corresponding plurality, and any refers to any one of the corresponding plurality. For example, if a plurality of words includes 10 words, each refers to each of the 10 words, and any refers to any one of the 10 words.

[0057] The information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this disclosure are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions.

[0058] Before implementing the embodiments of the present disclosure, the terms involved in the embodiments of the present disclosure are explained first.

[0059] VSP is a seismic observation method that observes seismic wavefields in a well. Geophones are placed at varying depths in the well to record seismic signals generated by surface sources. In vertical seismic profiles, because geophones can be placed deep within the formation through the well, they can receive both upgoing waves and downgoing waves. This is a key feature of vertical seismic profiles compared to horizontal ones.

[0060] Static Correction: During onshore seismic data processing, the data are typically corrected to a uniform datum, typically a horizontal plane. Seismic exploration and interpretation theories assume that the excitation and receiving points are aligned and that formation velocities are uniform. However, in reality, the ground is often uneven, the depths of individual excitation points can vary, and the wave velocities in the low-velocity zone differ significantly from those in the formation, inevitably affecting the shape of the measured time-distance curve. To eliminate these effects, the raw seismic data are corrected for terrain, excitation depth, and low-velocity zones. These corrections are invariant across different seismic interfaces at the same observation point and are therefore collectively referred to as static corrections.

[0061] First arrival wave: After an earthquake occurs, the first wave received by the earthquake observation point is called the first arrival wave.

[0062] Traveltime: The time it takes for a seismic wave to propagate from the shot point to the receiver. Calculating traveltime is a core concept in seismic migration and tomography, and a crucial factor in achieving amplitude-preserved prestack migration.

[0063] Elevation: also known as absolute elevation, refers to the distance from a point to the absolute base plane along the plumb line.

[0064] Fresnel Volume Ray Tracing: Compared to conventional rays, Fresnel Volumes can be considered a physical ray, representing an equivalent concept to 3D Fresnel zones. Fresnel Volume Ray Tracing builds on standard ray tracing by adding the calculations to define the parameters of the first Fresnel zone.

[0065] SIRT (simultaneous iterative reconstruction technique): Its principle is to utilize all rays passing through a pixel. The iterative update is a weighted average of the corrections to all projection lines according to their contribution factors, then back-projected. Unlike ART (Algebraic Reconstruction Technique), which updates the image once for each projection line, SIRT combines the contributions of all projection lines, preventing errors in a single projection line from significantly affecting the reconstruction result. This effectively suppresses noise in the reconstructed image.

[0066] The present disclosure provides a method for training a tomographic velocity model based on VSP data. Taking an electronic device as an example, the electronic device has strong computing capabilities and can be a terminal, such as a smart phone, a laptop, a desktop computer, etc. The electronic device can also be a server, including a single physical server, a cluster or a distributed system composed of multiple physical servers, etc. The present disclosure does not impose any specific restrictions on the product type of the electronic device. Figure 1 , the method process provided by the embodiment of the present disclosure includes:

[0067] 101. An initial tomographic velocity model is constructed based on the shot point elevation information, receiver point depth information and initial velocity values ​​in the vertical seismic profile (VSP) upgoing wave first arrival travel time data.

[0068] The tomographic velocity model is used to determine the velocity of seismic waves when they arrive at each detector point, providing a more accurate velocity field for migration imaging. The electronic device constructs an initial tomographic velocity model based on the shot point elevation information, the detector point depth information, and the initial velocity values ​​in the VSP upgoing wave first arrival travel time data. When constructing the initial tomographic velocity model, the following methods can be used:

[0069] 1011. The electronic device performs coordinate transformation on the depth information of the detection point to obtain the elevation information of the detection point.

[0070] The depth information of the detection point includes the true depth information of the detection point and the drilling depth information. The electronic device subtracts the drilling depth information from the true depth information of the detection point to obtain the elevation information of the detection point. For example, Figure 2 The true depth position of the detection point is S. Subtracting the drilling depth, the elevation position of the detection point can be obtained as S.

[0071] This process can be implemented using the following formula:

[0072] Z ele =Z' depth -Z depth

[0073] Among them, Z ele is the elevation information of the detection point, Z' depth To obtain the true depth information of the detection point, Z depth The drilling depth of the detection point.

[0074] The electronic device can obtain the VSP shot detection position relationship by transforming the depth information of the detection point into coordinates. The position relationship can be found in Figure 3 .

[0075] 1012. The electronic device performs interpolation calculation based on the elevation information of the detection point to obtain the elevation information of all points on the surface of the velocity field.

[0076] Interpolation is the process of interpolating a continuous function based on discrete data, so that the continuous curve passes through all given discrete data points. As an important method for approximating discrete functions, interpolation can estimate the approximate value of a function at a finite number of points based on the function's values ​​at other points. In the disclosed embodiments, the depth information of the detection points is interpolated together with the elevation information, avoiding the depth anomalies that can occur with direct modeling. Figure 4 The figure shows the interpolation effect of the related technology based on the depth information of the detection point. Figure 5 The interpolation effect diagram of the present disclosure based on the elevation information of the detection point is shown. Figure 4 and Figure 5 It can be seen that the interpolation method disclosed herein produces a smoother effect.

[0077] 1013. The electronic device constructs an initial velocity model based on the elevation information and initial velocity values ​​of all points on the surface in the velocity field.

[0078] The electronic device determines model parameters of the initial velocity model based on elevation information and initial velocity values ​​of all points on the surface of the velocity field, and then constructs the initial velocity model based on the model parameters of the initial velocity model.

[0079] 102. In the velocity field provided by the initial tomographic velocity model, a Fresnel body ray tracing method based on depth information is used to trace the path of rays received by detection points with depth information to obtain the Fresnel body path length range in the depth direction within the velocity field.

[0080] When the electronic device performs path tracing on the rays received by the detection point with depth information within the velocity field provided by the initial tomographic velocity model, and obtains the Fresnel body path length range in the depth direction within the velocity field, the following method can be used:

[0081] 1021. Within the velocity field provided by the initial tomographic velocity model, the electronic device uses the Fresnel body ray tracing method based on depth information to trace the path of the ray received by the detection point at the position indicated by each depth information, and obtains the path length of the ray received by the detection point at the position indicated by each depth information.

[0082] In the velocity field provided by the initial tomographic velocity model, the electronic equipment simulates the propagation process of rays from the shot point to the detection point. During the propagation process, the Fresnel volume ray tracing method based on the depth information is used to trace the path of the rays received by the detection point at the position indicated by each depth information, thereby obtaining the path length of the rays received by the detection point at the position indicated by each depth information.

[0083] 1022. The electronic device calculates the inversion travel time of rays with different path lengths at the position indicated by each depth information based on the path length of the rays received by the detection point at the position indicated by each depth information.

[0084] The electronic device obtains the propagation speed of the seismic wave in different media during the process of propagating from the shot point to the detection point, and based on the path length of the ray received by the detection point at the position indicated by each depth information, divides the path length of the ray in each medium by the propagation speed in the corresponding medium to obtain the propagation time of the seismic wave in each medium, and then adds the propagation time in each medium to obtain the inverted travel time of the ray with different path lengths at the position indicated by the detection point at each depth information.

[0085] 1023. The electronic device determines the shortest path length of the detection point at the position indicated by each depth information based on the inverted travel time of rays of different path lengths at the position indicated by each depth information.

[0086] The electronic device determines the shortest path length of the detection point at the position indicated by each depth information based on the inverted travel time of rays of different path lengths at the position indicated by each depth information and using the calculation range formula of the Fresnel body ray path.

[0087] The currently commonly used calculation range formula for the Fresnel body ray path is as follows:

[0088] t L (g,r)+t L (r,s)-t Lmin (g,s)≤1 / (2f)

[0089] Among them, t L (g, r) represents the time of any point in the travel time field at the detection point, t L (r,s) represents the time of any point in the shot travel time field, t Lmin (g,s) represents the minimum value of the sum of the travel time fields of the shot and detector, f represents the frequency of the seismic wave, g represents the location of the detector point, s represents the location of the shot point, and r represents an arbitrary point.

[0090] However, the above formula does not consider the influence of different depth positions of the detection points, resulting in the calculated shortest path length being a single ray path with a small coverage range. Based on the characteristics of VSP data, the disclosed embodiment innovatively improves the calculation range calculation formula of the Fresnel volume ray path based on VSP data. The improved formula is as follows:

[0091] t L (g depth ,r)+t L (r,s)-t Lmin (g depth ,s)≤1 / (2f)

[0092] Among them, g depth Represents a detection point with depth. As the depth changes, t L (g depth ,r) The travel time field calculated by the above method has a completely different meaning. Compared with the single ray path, the Fresnel body ray path calculated by the above method has obvious advantages of high coverage and better uniformity, such as Figure 6 The accompanying drawings shown.

[0093] 1024. The electronic device combines the shortest path lengths of the detection points at different positions into the Fresnel body path length range.

[0094] The electronic device combines the shortest path lengths of the detection points at different positions to obtain the Fresnel body path length range.

[0095] The disclosed embodiments use a fast Fresnel volume ray tracing method to calculate ray paths, which significantly improves the coverage of the VSP first arrival wave path and increases the uniformity of ray penetration, thereby increasing the stability of the inversion solution and obtaining a more accurate tomographic velocity model.

[0096] 103. The inversion of the objective function with VSP data depth weighting after improving the Fresnel body path length range is used to obtain the inverted travel time of the ray reaching the detection point along different paths within the Fresnel body path length range at different depths.

[0097] Based on the acquired Fresnel body path length range, the electronic device uses the SIRT algorithm to invert the Fresnel body path length range at different depths, and obtains the inverted travel time of the ray along different paths within the Fresnel body path length range to reach the detection point.

[0098] 104. Based on the inverted travel time of rays arriving at the detector point along different paths within the Fresnel path length range at different depths and the improved inversion objective function with VSP data depth weighting, the model parameters of the initial tomographic velocity model are adjusted to obtain the trained tomographic velocity model.

[0099] The inversion objective function is a function used to optimize the training process of the tomographic velocity model. The existing inversion objective function is:

[0100] Φ(m)=||dG(m)|| 2 +τ1||R1m|| 2 +τ2||R2m|| 2

[0101] Where, ||dG(m)|| 2 is the two-norm of the time difference between different paths, τ1||R1m|| 2 is the first-order regularization term, τ2||R2m|| 2 is the second-order regularization term. τ1||R1m|| 2 and τ2||R2m|| 2 They are all constraints of the inversion calculation and can achieve a stable solution.

[0102] Since VSP data in the well contains shallow return waves and a large number of direct wave paths, research has found that shallow return waves can strongly affect the velocity distribution in shallow and middle layers. Therefore, the method provided in the embodiment of the present disclosure dynamically introduces an offset distance weight that affects the offset distance of the shallow layer when performing tomographic inversion, so that the calculation is performed within a reasonable range.

[0103] In the embodiment of the present disclosure, the process of constructing the inversion objective function is:

[0104] In the first step, the electronic equipment determines the optimal offset distance of the detection points at different depths based on the depth information of the detection points and the improved inversion objective function with VSP data depth weighting.

[0105] In the embodiment of the present disclosure, the offset distance varies with the depth of the detection point in the VSP data. The variation pattern can be expressed by the following formula:

[0106] Offset=A*Z depth +B

[0107] Among them, Offset represents the offset distance, Z depth represents the drilling depth of the detection point. A and B are known parameters that can be determined based on empirical values. The function of A and B is to dynamically set the offset size according to the depth of the VSP detection point, separate the offset distances of the shallow layer from the middle layer and the deep layer, and thus dynamically constrain the inversion objective function.

[0108] In the second step, the electronic device determines the offset weight of the Fresnel body path of different depth information based on the distance between the detection point and the shot point of the emitting ray and the optimal offset distance.

[0109] The offset weight is used to select the receiver points for training as the depth changes. The offset weight depends on the relationship between the distance between the shot point and the receiver point and the optimal offset distance. When the distance between the shot point and the receiver point is less than or equal to the optimal offset distance, the offset weight is 1; when the distance between the shot point and the receiver point is greater than the optimal offset distance, the offset weight is 0.

[0110] In the third step, the electronic device determines the inversion objective function based on the Fresnel body path and its offset weight at different depth information.

[0111] The inversion objective function determined in the embodiment of the present disclosure is as follows:

[0112] Φ(m)=ω offset ||dG(m)|| 2 +τ1||R1m|| 2 +τ2||R2m|| 2

[0113] Among them, ωoffset is the offset weight.

[0114] In another embodiment of the present disclosure, the electronic device adjusts the model parameters of the initial tomographic velocity model based on the inversion travel time of rays arriving at the detection point along different paths within the Fresnel body path length range and the inversion objective function to obtain the trained tomographic velocity model using the following method:

[0115] In the first step, the electronic device inputs the inverted travel time of the ray along different paths within the Fresnel body path length range at different depths to the detection point into the improved inversion objective function with VSP data depth weighting, and outputs the function value of the improved inversion objective function with VSP data depth weighting.

[0116] In the second step, the electronic device adjusts the model parameters of the initial tomographic velocity model based on the function value.

[0117] In the third step, the electronic device continues to use the Fresnel volume ray tracing method based on depth information based on the tomographic velocity model after adjusting the model parameters, and traces the path of the rays received by the detection point with depth information until the training cutoff condition is met.

[0118] The cutoff condition includes the number of adjustments reaching a preset number or the function value of the target loss function being less than a preset threshold. The preset number of adjustments can be set based on the training accuracy of the model, for example, 5 or 6 times. The preset threshold can also be set based on the training accuracy of the model.

[0119] In the fourth step, the electronic device uses the tomographic velocity model obtained when the training cutoff condition is met as the trained tomographic velocity model.

[0120] The disclosed embodiments improve upon the conventional inversion objective function by adding a very important weighting method. Based on the relationship between the offset and depth during the initial arrival of the VSP, an effective method for automatically limiting the offset along with depth variations is proposed. Based on the characteristics of Fresnel ray tracing, the weights of the Fresnel ray paths corresponding to different depths are automatically calculated, thereby making more reasonable corrections to the values ​​of the constructed inversion coefficient matrix. Ultimately, the inversion calculation results are more reasonable and have higher computational accuracy.

[0121] Figure 7 A schematic diagram shows the comparative effects of the tomographic velocity model trained using the embodiment of the present disclosure before and after static correction. Through the comparison, it can be found that the tomographic velocity model trained using the embodiment of the present disclosure has a better effect after static correction.

[0122] The method provided by the disclosed embodiments employs a depth-informed Fresnel volume ray tracing method within the velocity field provided by the initial tomographic velocity model to trace the paths of rays received at depth-informed geophones. This method obtains the Fresnel volume path length range in the depth direction of the velocity field. This Fresnel volume path length range has a greater coverage density than a single ray path, resulting in better ray tomographic inversion results. The tomographic velocity model is trained using the inverted travel times of rays arriving at geophones along different paths within the Fresnel volume path length range at different depths and an improved objective function weighted by VSP data depth, thereby improving the accuracy of the tomographic velocity model. Furthermore, by adding an offset weight to the inversion objective function, the shadow area of ​​the first-arrival ray path is reduced to almost zero, avoiding large-scale smoothing of the tomographic velocity model. This enhances the stability of the inversion solution from the data itself, effectively compensating for the shortcomings of the VSP data in inversion, and further ensuring the accuracy of the trained tomographic velocity model.

[0123] See also Figure 8 The present disclosure provides a tomographic velocity model training device based on VSP data, the device comprising:

[0124] A construction module 801 is used to construct an initial tomographic velocity model based on the elevation information of the shot points, the depth information of the receiver points and the initial velocity values ​​in the vertical seismic profile VSP upgoing wave first arrival travel time data;

[0125] A tracking module 802 is configured to perform path tracing on rays received by a detection point having depth information within the velocity field provided by the initial tomographic velocity model using a Fresnel volume ray tracing method based on depth information, thereby obtaining a range of Fresnel volume path lengths in the depth direction within the velocity field.

[0126] Inversion module 803 is used to invert the improved objective function with VSP data depth weighting for the Fresnel body path length range, and obtain the inverted travel time of rays reaching the detection point along different paths within the Fresnel body path length range at different depths;

[0127] The adjustment module 804 is configured to adjust the model parameters of the initial tomographic velocity model based on the inverted travel times of the rays arriving at the detector points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting, thereby obtaining a trained tomographic velocity model.

[0128] In another embodiment of the present disclosure, construction module 801 is configured to perform coordinate transformation on depth information of detection points to obtain elevation information of the detection points; perform interpolation calculation based on the elevation information of the detection points to obtain elevation information of all surface points in the velocity field; and construct an initial velocity model based on the elevation information of all surface points in the velocity field and the initial velocity values.

[0129] In another embodiment of the present disclosure, the tracking module 802 is configured to perform path tracing of rays received by the detection point at the position indicated by each depth information within the velocity field provided by the initial tomographic velocity model using a Fresnel volume ray tracing method based on depth information to obtain a path length of the ray received by the detection point at the position indicated by each depth information; calculate, based on the path length of the ray received by the detection point at the position indicated by each depth information, the inverted travel time of rays with different path lengths at the position indicated by each depth information; determine, based on the inverted travel time of rays with different path lengths at the position indicated by each depth information, the shortest path length of the detection point at the position indicated by each depth information; and combine the shortest path lengths of the detection point at different positions into a Fresnel volume path length range.

[0130] In another embodiment of the present disclosure, the apparatus further comprises:

[0131] A determination module is used to determine the optimal offset distance of the receiver points at different depths based on the depth information of the receiver points and the improved inversion objective function with VSP data depth weighting;

[0132] The determination module is further used to determine the offset weights of the Fresnel body paths with different depth information based on the distance between the detection point and the shot point of the emitting ray and the optimal offset distance;

[0133] The determination module is also used to determine the inversion objective function based on the Fresnel body paths and offset weights of different depth information.

[0134] In another embodiment of the present disclosure, the adjustment module 804 is configured to input the inverted travel times of rays arriving at a detection point along different paths within a range of Fresnel body path lengths at different depths into an improved inversion objective function with VSP data depth weighting, and output a function value of the improved inversion objective function with VSP data depth weighting; adjust model parameters of an initial tomographic velocity model based on the function value; continue to use a Fresnel body ray tracing method based on depth information to trace the paths of rays received at the detection points with depth information based on the tomographic velocity model after the model parameters are adjusted, until a training cutoff condition is met; and use the tomographic velocity model obtained when the training cutoff condition is met as the trained tomographic velocity model.

[0135] In summary, the apparatus provided by the embodiments of the present disclosure employs a depth-informed Fresnel volume ray tracing method within the velocity field provided by the initial tomographic velocity model to trace the paths of rays received at depth-informed geophones. This method obtains the Fresnel volume path length range in the depth direction of the velocity field. This Fresnel volume path length range has a greater coverage density than a single ray path, resulting in better ray tomographic inversion results. The tomographic velocity model is trained using the inverted travel times of rays arriving at geophones along different paths within the Fresnel volume path length range at different depths and an improved objective function weighted by VSP data depth, thereby improving the accuracy of the tomographic velocity model. Furthermore, by adding an offset weight to the inversion objective function, the shadow region of the first-arrival ray path is reduced to almost zero, avoiding large-scale smoothing of the tomographic velocity model. This enhances the stability of the inversion solution from the data itself, effectively compensating for the shortcomings of the VSP data in inversion, and further ensuring the accuracy of the trained tomographic velocity model.

[0136] Figure 9 FIG. 1 is a block diagram of an electronic device 900 according to an exemplary embodiment of the present disclosure. Generally, the electronic device 900 includes a processor 901 and a memory 902 .

[0137] The processor 901 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 901 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 901 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 901 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 901 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0138] Memory 902 may include one or more computer-readable storage media, which may be non-transitory. Memory 902 may also include high-speed random access memory and non-volatile memory, such as one or more magnetic disk storage devices or flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in memory 902 is configured to store at least one instruction, which is configured to be executed by processor 901 to implement the tomographic velocity model training method based on VSP data provided in the method embodiments of the present disclosure.

[0139] In some embodiments, electronic device 900 may optionally include a peripheral device interface 903 and at least one peripheral device. Processor 901, memory 902, and peripheral device interface 903 may be connected via a bus or signal lines. Each peripheral device may be connected to peripheral device interface 903 via a bus, signal lines, or circuit boards. Specifically, the peripheral device includes a power supply 904.

[0140] The peripheral device interface 903 can be used to connect at least one I / O (Input / Output)-related peripheral device to the processor 901 and the memory 902. In some embodiments, the processor 901, the memory 902, and the peripheral device interface 903 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 901, the memory 902, and the peripheral device interface 903 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0141] Power supply 904 is used to power the various components of electronic device 900. Power supply 904 can be AC ​​power, DC power, a disposable battery, or a rechargeable battery. When power supply 904 includes a rechargeable battery, the rechargeable battery can be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery that is charged via a wired line, while a wireless rechargeable battery is a battery that is charged via a wireless coil. The rechargeable battery can also be used to support fast charging technology.

[0142] Those skilled in the art will understand that Figure 9 The structure shown in the figure does not constitute a limitation on the electronic device 900, and the electronic device 900 may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0143] In an exemplary embodiment, a computer-readable storage medium including instructions, such as a memory including instructions, is also provided. The instructions are executable by a processor of the electronic device 900 to implement the above-described VSP data-based tomographic velocity model training method. Alternatively, the storage medium may be a non-transitory computer-readable storage medium, such as a CD-ROM (Compact Disc Read-Only Memory), ROM, RAM (Random Access Memory), magnetic tape, floppy disk, or optical data storage device.

[0144] Those skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or by a program to instruct the relevant hardware, and the program may be stored in a computer-readable storage medium, which may be a read-only memory, a disk, or an optical disk, etc.

[0145] The above description is merely an optional embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the scope of protection of the present disclosure.

Claims

1. A tomographic velocity model training method based on VSP data, characterized in that: The method comprises: An initial tomographic velocity model is constructed based on the shot point elevation information, receiver point depth information and initial velocity values ​​in the vertical seismic profile VSP upgoing wave first arrival travel time data. In the velocity field provided by the initial tomographic velocity model, a Fresnel body ray tracing method based on depth information is used to trace the path of the ray received by the detection point with depth information to obtain the Fresnel body path length range in the depth direction in the velocity field; Performing an improved objective function inversion with VSP data depth weighting on the Fresnel body path length range to obtain the inverted travel time of the ray reaching the detection point along different paths within the Fresnel body path length range at different depths; Adjusting the model parameters of the initial tomographic velocity model based on the inverted travel times of the rays arriving at the detection points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting to obtain a trained tomographic velocity model; The Fresnel body ray tracing method based on depth information, within the velocity field provided by the initial tomographic velocity model, performs path tracing on the rays received by the detection point with depth information to obtain the Fresnel body path length range in the depth direction within the velocity field, including: performing path tracing on the ray received by the detection point at the position indicated by each depth information using a Fresnel volume ray tracing method based on depth information within the velocity field provided by the initial tomographic velocity model to obtain the path length of the ray received by the detection point at the position indicated by each depth information; Calculating, based on the path length of the ray received by the detection point at the position indicated by each depth information, the inverted travel time of the ray with different path lengths at the position indicated by each depth information; Determining the shortest path length of the detection point at the position indicated by each depth information based on the inverted travel time of rays of different path lengths at the position indicated by each depth information; The shortest path lengths of the detection points at different positions constitute the Fresnel body path length range.

2. The method according to claim 1, characterized in that The initial tomographic velocity model is constructed based on the elevation information of the shot points, the depth information of the receiver points and the initial velocity values ​​in the first arrival travel time data of the vertical seismic profile VSP, including: Performing coordinate transformation on the depth information of the detection point to obtain the elevation information of the detection point; Performing interpolation calculation based on the elevation information of the detection points and the shot points to obtain elevation information of all points on the surface of the velocity field; The initial tomographic velocity model is constructed based on the elevation information of all points on the surface of the velocity field and the initial velocity values.

3. The method according to claim 1, characterized in that Before obtaining the trained tomographic velocity model, the method further includes adjusting the model parameters of the initial tomographic velocity model based on the inverted travel times of the rays reaching the detection points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting, and further includes: Determining optimal offset distances of the geophones at different depths based on the depth information of the geophones and the improved inversion objective function with VSP data depth weighting; determining offset weights of Fresnel body paths of different depth information based on the distance between the detection point and the shot point emitting the ray and the optimal offset distance; The inversion objective function is determined based on the Fresnel body paths of different depth information and their offset weights.

4. The method according to claim 1, wherein The method adjusts the model parameters of the initial tomographic velocity model based on the inversion travel time of the ray reaching the detection point along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting to obtain a trained tomographic velocity model, including: Inputting the inverted travel time of the ray along different paths within the Fresnel body path length range at different depths to the detection point into the improved inversion objective function with VSP data depth weighting, and outputting the function value of the improved inversion objective function with VSP data depth weighting; adjusting model parameters of the initial tomographic velocity model based on the function value; Based on the tomographic velocity model after the model parameters are adjusted, continuing the Fresnel volume ray tracing method based on the depth information, tracing the path of the ray received by the detection point with the depth information until the training cutoff condition is met; The tomographic velocity model obtained when the training cutoff condition is met is used as the trained tomographic velocity model.

5. A tomographic velocity model training device based on VSP data, characterized in that: The device comprises: A construction module is used to construct an initial tomographic velocity model based on the shot point elevation information, receiver point depth information and initial velocity values ​​in the vertical seismic profile VSP upgoing wave first arrival travel time data; a tracking module configured to perform path tracing on rays received by the detection points having depth information within the velocity field provided by the initial tomographic velocity model based on a Fresnel body ray tracing method using depth information, and obtain a range of Fresnel body path lengths in the depth direction within the velocity field; an inversion module for performing an improved VSP data depth-weighted objective function inversion on the Fresnel body path length range to obtain an inverted travel time of the ray reaching the detection point along different paths within the Fresnel body path length range at different depths; an adjustment module, configured to adjust model parameters of the initial tomographic velocity model based on inversion travel times of the rays arriving at the detection points along different paths within the Fresnel body path length range at different depths and the improved inversion objective function with VSP data depth weighting, to obtain a trained tomographic velocity model; The tracking module is configured to perform path tracing on rays received by the detection point at the position indicated by each depth information within the velocity field provided by the initial tomographic velocity model using a Fresnel volume ray tracing method based on depth information to obtain a path length of the ray received by the detection point at the position indicated by each depth information; calculate, based on the path length of the ray received by the detection point at the position indicated by each depth information, the inverted travel time of rays with different path lengths at the position indicated by each depth information; determine, based on the inverted travel time of rays with different path lengths at the position indicated by each depth information, the shortest path length of the detection point at the position indicated by each depth information; and form the shortest path lengths of the detection point at different positions into the Fresnel volume path length range.

6. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the memory stores at least one program code, and the at least one program code is loaded and executed by the processor to implement the tomographic velocity model training method based on VSP data according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that The storage medium stores at least one program code, and the at least one program code is loaded and executed by the processor to implement the tomographic velocity model training method based on VSP data according to any one of claims 1 to 4.

8. A computer program product, characterized in that The computer program product includes computer program code, which is stored in a computer-readable storage medium. A processor of an electronic device reads the computer program code from the computer-readable storage medium, and the processor executes the computer program code, so that the electronic device performs the tomographic velocity model training method based on VSP data according to any one of claims 1 to 4.

Images