A method for predicting fracture-cavity filling degree based on prestack seismic data based on structural tensor constraints

Abstract

The invention discloses a method for predicting the filling degree of fractures and holes by prestack seismic based on structural tensor constraint, which belongs to the field of geophysical technology, and the steps are as follows: obtaining well logging data and mud logging data; performing fracture-hole model forward modeling, and selecting the most sensitive gradient structural tensor attribute body; constructing the volume constraint of the prestack elastic parameter body according to the well number statistical coincidence rate of the most sensitive gradient structural tensor attribute body; performing AVA gather forward modeling, and constructing a low-frequency model for prestack inversion to obtain the prestack elastic parameter body; determining the fracture-hole development boundary of the drilling sample point in the actual drilling area by using the volume constraint of the prestack elastic parameter body, and analyzing the elastic parameter value range of full filling, half filling and unfilling of the fracture-hole in combination with the leakage characteristics, and obtaining the rock physics interpretation model; predicting the filling degree of fractures and holes in the actual drilling area based on the rock physics interpretation model, and obtaining the fracture-hole filling degree prediction result. The invention solves the problem that it is difficult to accurately and effectively predict fractures and holes.

Description

Technical Field

[0001] The invention belongs to the field of geophysical technology, and in particular relates to a method for predicting the degree of fracture-cavity filling by pre-stack seismic based on structural tensor constraints. Background Art

[0002] Fracture-cavity reservoirs are the main reservoir type of carbonate fracture-cavity oil and gas reservoirs and are the key to high and stable production of oil and gas fields. The effectiveness of fracture-cavity reservoirs is mainly affected by the filling degree. Therefore, accurately predicting the filling degree of fracture-cavity reservoirs is of great significance to the exploration and development of fracture-cavity oil and gas reservoirs. Existing methods for predicting the filling degree through geophysics include: (1) logging-based methods, such as indicating the filling degree of fractures and holes through abnormal changes in drilling and logging parameters, identifying the filling degree of fractures and holes through logging, or predicting the filling degree of fractures and holes through electrical imaging; (2) seismic-based methods, such as directly predicting the filling degree of fractures and holes through post-stack seismic sensitivity attributes or post-stack inversion bodies, or directly predicting the filling degree of fractures and holes through pre-stack seismic sensitivity attributes or pre-stack elastic parameters; (3) forward modeling or simulation-based methods, such as directly predicting the filling degree of fractures and holes through forward modeling technology, or measuring the filling degree of fractures and holes through three-dimensional physical model filling design; (4) machine learning-based methods, such as predicting the filling degree of fractures and holes based on BP neural network, where the input targets mainly include underground river level, underground river type, tunnel style, relationship with underground river entrance and exit, relationship with hall cave, etc., so as to establish a filling degree prediction model and calculate the filling degree of fractures and holes based on this model.

[0003] Existing methods for predicting the filling degree through geophysics have the following main defects: (1) Methods based mainly on well logging usually analyze and predict a single well and cannot effectively predict the filling degree of the reservoir in the horizontal direction; (2) Methods based mainly on seismic usually directly use relevant seismic attributes or elastic parameters to predict the filling degree, ignoring the influence of non-fractures and holes, resulting in low reliability of prediction results; (3) Methods based mainly on forward modeling or simulation only simulate the numerical values ​​of the fracture and hole model and cannot directly predict the actual fracture and hole filling degree; (4) Methods based mainly on machine learning still need to verify the reliability of the data at their input end, making it difficult to guarantee the accuracy of the prediction results. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for predicting the degree of fracture-hole filling by pre-stack seismic based on structural tensor constraints. The fracture-hole body is predicted by the structural tensor attribute body, and the pre-stack elastic parameters such as the longitudinal-to-straight wave velocity ratio and the longitudinal wave impedance are further optimized based on this constraint to realize the prediction of the degree of fracture-hole filling, thereby evaluating the effectiveness of the fracture-hole and solving the problem of difficulty in accurately and effectively predicting the fracture-hole.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] The present invention provides a method for predicting the degree of fracture-cavity filling by prestack seismic based on structural tensor constraints, comprising the following steps:

[0007] S1, obtaining well logging data and mud logging data;

[0008] S2. Perform fracture-cavity model forward modeling based on well logging data and mud logging data, and select the most sensitive gradient structure tensor attribute volume;

[0009] S3, constructing the volume constraint of the prestack elastic parameter body according to the statistical coincidence rate of the well number of the most sensitive gradient structure tensor attribute body;

[0010] S4, performing AVA gather forward modeling based on well logging data and mud logging data, and constructing a low-frequency model for prestack inversion to obtain a prestack elastic parameter volume;

[0011] S5. Using the volume constraint of the prestack elastic parameter body, determine the fracture-cavity development boundary of the drilling sample point in the actual drilling area, and analyze the elastic parameter value range of the fracture-cavity full filling, half filling and unfilling in combination with the leakage characteristics to obtain the rock physical interpretation model;

[0012] S6. Based on the rock physics interpretation model, the fracture and cavity filling degree in the actual drilling area is predicted to obtain the fracture and cavity filling degree prediction result.

[0013] The beneficial effects of the present invention are as follows: the present invention provides a method for predicting the degree of fracture-cavity filling by pre-stack seismic based on structural tensor constraints, which optimizes the most sensitive gradient structural tensor attribute body that can accurately characterize the fracture-cavity body based on the forward modeling of the fracture-cavity model related to the fault, so as to use the fracture-cavity body within the fracture-cavity development boundary in the measured well area as the body constraint of the pre-stack elastic parameter body to determine the prediction range of the fracture-cavity filling degree; the feasibility of using pre-stack data to predict the degree of fracture-cavity filling is verified by forward modeling of the AVA data set, and then the pre-stack elastic parameter body that can characterize the degree of fracture-cavity filling is optimized, and finally the rock physics interpretation volume plate is established in combination with the leakage characteristics to predict the degree of fracture-cavity filling. The method of the present invention can accurately evaluate the effectiveness of fractures and caves.

[0014] Other advantages of the present invention will be analyzed in more detail in subsequent embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0016] Figure 1The present invention is a flowchart of a method for predicting the degree of fracture-cavity filling by prestack seismic based on structural tensor constraints in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0018] like Figure 1 As shown, in one embodiment of the present invention, the present invention provides a method for predicting the degree of fracture-hole filling based on pre-stack seismic based on structural tensor constraints, comprising the following steps:

[0019] S1, obtaining well logging data and mud logging data;

[0020] S2. Forward modeling of the fracture-cavity model is performed based on the well logging data and the mud logging data, and the most sensitive gradient structure tensor attribute body is selected; forward modeling of the fracture-cavity model refers to the process of constructing a geological model containing complex geological structures through numerical simulation methods, simulating the propagation and reflection characteristics of seismic waves in complex media, and generating theoretical seismic response data.

[0021] The S2 comprises the following steps:

[0022] S21. Based on the well logging data and the mud logging data, a fracture-cavity model related to the fault is established and the model is forward modeled to obtain the seismic response characteristics when the fracture-cavity is developed;

[0023] Model forward modeling refers to the use of existing logging data and mud logging data to establish an underground geological model. According to the propagation principle of seismic waves in underground media, the characteristics of the established seismic model are calculated through forward simulation such as ray tracing or wave equation migration. This scheme uses the forward modeling software Tesserial to set up a seismic acquisition observation system model, simulates the fracture-cavity model related to the fault based on the one-dimensional wave equation, obtains the seismic response characteristics when the fracture-cavity develops, and extracts a number of attribute bodies associated with the fracture-cavity on this basis;

[0024] S22. Based on the seismic response characteristics when fractures and caves are developed, the seismic data body is subjected to dip scanning and Gaussian filtering, and the structure tensor matrix is ​​decomposed, and the sensitive gradient structure tensor attribute body with the highest statistical coincidence rate of well numbers is selected as the most sensitive gradient structure tensor attribute body.

[0025] The S22 comprises the following steps:

[0026] S221, performing dip scanning on the seismic data volume based on the seismic response characteristics when fractures and caves are developed, and filtering the seismic data volume using a first three-dimensional Gaussian function to obtain a smoothed data volume, wherein the standard deviation of the first three-dimensional Gaussian function is a first standard deviation;

[0027] The first standard deviation is determined based on the size of the fracture-cavity body and the noise in the study area. In this embodiment, the first standard deviation is set to 0.6;

[0028] The calculation expression of the smoothed data volume is as follows:

[0029] ,

[0030] ,

[0031] in, represents a smooth data volume, represents the first three-dimensional Gaussian function, represents the seismic data volume, represents the first standard deviation, represents an exponential function with e as the base constant, represents the position of the Gaussian distribution on the X-axis, Represents the center point of the three-dimensional Gaussian distribution in the X-axis direction, represents the position of the Gaussian distribution on the Y axis, represents the center point of the three-dimensional Gaussian distribution on the Y axis, represents the position of the Gaussian distribution on the Z axis, Represents the center point of the three-dimensional Gaussian distribution on the Z axis;

[0032] S222, calculate the gradient vector of the smooth data body and the gradient tensor ,in, It means that the smooth data volume is differentiated with respect to the position on the X-axis. It means that the smooth data volume is differentiated with respect to the position on the Y axis. It means that the smoothed data volume is differentiated with respect to the position on the Z axis, and T means transpose;

[0033] S223, filtering the gradient tensor of the smoothed data volume using a second three-dimensional Gaussian function to obtain a filtered gradient tensor of the smoothed data volume, wherein the standard deviation of the second three-dimensional Gaussian function is a second standard deviation;

[0034] The calculation expression of the second three-dimensional Gaussian function is as follows:

[0035] ,

[0036] in, represents the second three-dimensional Gaussian function, represents the second standard deviation;

[0037] The principle of selecting the value of the second standard deviation is that it is inversely proportional to the accuracy of the fracture body and directly proportional to the noise suppression. In this scheme, the value of the second standard deviation is 1.6;

[0038] S224, calculating a gradient structure tensor matrix based on the gradient vector of the smoothed data volume and the filtered gradient tensor;

[0039] The calculation expression of the gradient structure tensor matrix is ​​as follows:

[0040] ,

[0041] in, Represents the gradient structure tensor matrix;

[0042] In this scheme, the gradient structure tensor matrix is ​​a real symmetric matrix with semi-positive quadratic form properties, three non-negative eigenvalues, and the corresponding three eigenvectors are orthogonal to each other. For the structure tensor matrix of any three-dimensional data, it can be decomposed into three eigenvectors and corresponding eigenvectors.

[0043] S225. Decompose the gradient structure tensor matrix to obtain three eigenvalues ​​and corresponding eigenvectors;

[0044] The calculation expressions of the three eigenvalues ​​and the corresponding eigenvectors are as follows:

[0045] ,

[0046] in, represents the first eigenvector, represents the second eigenvector, represents the third eigenvector, represents the first eigenvalue, represents the second eigenvalue, represents the third eigenvalue;

[0047] The three eigenvalues ​​are respectively the first eigenvalue, the second eigenvalue and the third eigenvalue; the first eigenvalue corresponds to the first eigenvector, the direction corresponding to the first eigenvector is the direction in which the local stratigraphic structure changes fastest, and is perpendicular to the local plane of the reflection event axis, and the corresponding change amount is the first eigenvalue; the second eigenvalue corresponds to the second eigenvector; the third eigenvalue corresponds to the third eigenvector; the plane formed by the second eigenvector and the third eigenvector is perpendicular to the first eigenvector, and represents the direction of lateral discontinuous changes, such as the seismic data changes caused by the fault reflection characteristics. The direction in which the local stratigraphic structure changes fastest on the plane is the direction corresponding to the second eigenvector, and the corresponding change rate is the second eigenvalue, and the direction in which the local stratigraphic structure changes slowest on the plane is the direction corresponding to the third eigenvector, and the corresponding change rate is the third eigenvalue. These three eigenvalues ​​and corresponding eigenvectors construct a structural change gradient characterization system in different dimensional directions in three-dimensional space.

[0048] S226. According to the three eigenvalues ​​and the corresponding eigenvectors, a number of sensitive gradient structure tensor attribute bodies describing the carbonate rock fracture-cavity body are selected;

[0049] In this embodiment, for carbonate rock fracture-cavity bodies, the first eigenvalue reflects the layered reflection characteristics, and the second and third eigenvalues ​​characterize the abnormal reflection area of ​​the fracture-cavity bodies. Therefore, the second and third eigenvalues ​​can be used as sensitive gradient structure tensor attribute bodies for characterizing carbonate rock fracture-cavity bodies.

[0050] S227, respectively compare the statistical agreement rates of the number of wells describing the fracture-cavity body based on each sensitive gradient structure tensor attribute body and the actual number of wells drilled, and select the sensitive gradient structure tensor attribute body with the highest statistical agreement rate of the number of wells as the most sensitive gradient structure tensor attribute body.

[0051] The S227 includes the following steps:

[0052] S2271. Use sensitive gradient structure tensor attribute bodies to finely characterize the fracture-cavity body, and obtain the number of wells corresponding to each sensitive gradient structure tensor attribute body;

[0053] S2272, respectively divide the number of wells corresponding to each sensitive gradient structure tensor attribute body by the actual number of wells drilled to obtain a statistical coincidence rate of the number of wells corresponding to each sensitive gradient structure tensor attribute body;

[0054] S2273. Select the sensitive gradient structure tensor attribute body with the highest statistical coincidence rate of well number as the most sensitive gradient structure tensor attribute body.

[0055] In this embodiment, for the same drilling data, if the total number of wells is 32, the number of wells that match the second eigenvalue as the sensitive gradient structure tensor attribute body with the actual drilling is 26, and the number of wells that match the third eigenvalue as the sensitive gradient structure tensor attribute body with the actual drilling is 22. Therefore, the statistical coincidence rate of the number of wells corresponding to the second eigenvalue is higher than the statistical coincidence rate of the number of wells corresponding to the third eigenvalue, and the second eigenvalue is selected as the most sensitive gradient structure tensor attribute body.

[0056] S3, constructing the volume constraint of the prestack elastic parameter body according to the statistical coincidence rate of the well number of the most sensitive gradient structure tensor attribute body;

[0057] The calculation expression of the volume constraint of the prestack elastic parameter body in S3 is as follows:

[0058] ,

[0059] in, Indicates the development status of the slit-cavity body. The statistical agreement rate of the number of wells representing the most sensitive gradient structure tensor attribute volume, represents the threshold value of the statistical coincidence rate of the well number, Indicates the state of suture development. Indicates that the suture hole is not developed.

[0060] In this embodiment, the threshold value of the statistical coincidence rate of the number of wells is set to 0.19, that is, if the statistical coincidence rate of the number of wells of the most sensitive gradient structure tensor attribute body is greater than 0.19, it indicates that the fractures and holes are in a developed state, and if the statistical coincidence rate of the number of wells of the most sensitive gradient structure tensor attribute body is less than or equal to 0.19, it indicates that the fractures and holes are in an undeveloped state. In this solution, the development of fractures and holes is used as a constraint of the structure tensor attribute body.

[0061] S4. Perform AVA gather forward modeling based on well logging data and mud recording data, and build a low-frequency model for pre-stack inversion to obtain pre-stack elastic parameter volume; Amplitude Variation with Angle (AVA) is an important technology used to analyze lithology and fluid properties in seismic exploration; Gather forward modeling refers to the process of calculating the reflection amplitude of seismic waves at different incident angles through numerical simulation methods to generate theoretical seismic gather data; AVA gather forward modeling refers to the process of generating theoretical AVA gather data by establishing a geological model and simulating the reflection characteristics of seismic waves at different incident angles;

[0062] The S4 comprises the following steps:

[0063] S41, based on the well logging data and the mud logging data, the AVA gathers are forward modeled using the real drilling data with shear wave data, and the AVA characteristics of the fractures and holes with different filling degrees are analyzed to obtain the AVA gathers with different filling degrees;

[0064] The AVA features of different filling degrees in S41 include no AVA features, first-class AVA features, second-class AVA features and third-class AVA features.

[0065] In this scheme, by analyzing the actual drilling data with shear wave data, it is found that: when the fracture hole is fully filled, there is no AVA feature; when the fracture hole is half filled, it is a type I AVA feature or a type II AVA feature, where the type I AVA feature refers to the amplitude of the wave crest event axis in the angle gather decreasing with the increase of the incident angle, and the type II AVA feature refers to the angle gather having a feature of a wave crest amplitude to a wave trough amplitude conversion from zero incident angle to maximum incident angle; when the fracture hole is not filled, it is a type III AVA feature, where the type III AVA feature refers to the amplitude of the wave trough event axis in the angle gather increasing with the increase of the incident angle. The above different degrees of AVA features depend on the P-wave velocity, S-wave velocity and density.

[0066] S42, denoising the AVA gathers with different filling degrees, and performing gather removal and residual amplitude compensation on the denoised AVA gathers to obtain optimized AVA gathers;

[0067] In this embodiment, the AVA gathers with different filling levels are denoised by filtering denoising, linear denoising, singular value denoising and other means to improve the gather signal-to-noise ratio;

[0068] S43, stacking the optimized AVA gathers to obtain near, middle and far stacked data volumes;

[0069] S44, repeating the well-seismic calibration and extracting the AVA wavelet of the near, middle and far stacked data volumes for a preset number of times to obtain the lateral constraints of the seismic interpretation layers;

[0070] In this embodiment, the purpose of well-seismic calibration is to establish the correspondence between logging data and seismic data, that is, the correspondence between near, middle and far stacked data bodies, so as to determine the geological layer corresponding to the seismic phase axis, realize the unification of time domain and depth domain, and provide a basis for geological interpretation and reservoir prediction. In this embodiment, well-seismic calibration and the wavelet extraction of near, middle and far stacked data bodies are mutually iterative, and multiple iterations are performed to fully extract the lateral constraints of the seismic interpretation layer; seismic data is continuous in space, and there is correlation between adjacent seismic traces. The lateral constraints of the layer use this characteristic to improve the interpretation accuracy in the following ways; by utilizing the lateral correlation of seismic data and combining geological prior information, key information such as geological structure and reservoir distribution can be effectively identified.

[0071] S45, interpolating the logging data along the layers using the lateral constraints of the seismic interpretation layers to construct a low-frequency model, wherein the low-frequency model includes a longitudinal-straight wave velocity ratio model and a longitudinal wave impedance model;

[0072] S46. Perform prestack inversion based on the low-frequency model to obtain a prestack elastic parameter volume.

[0073] In this scheme, the Zoeppritz equation in Jason software is used to perform prestack inversion to obtain the prestack elastic parameter volume.

[0074] The pre-stack elastic parameter bodies in S46 are the longitudinal and transverse wave velocity ratio body and the longitudinal wave impedance body.

[0075] S5. Using the volume constraint of the prestack elastic parameter body, determine the fracture-cavity development boundary of the drilling sample point in the actual drilling area, and analyze the elastic parameter value range of the fracture-cavity full filling, half filling and unfilling in combination with the leakage characteristics to obtain the rock physical interpretation model;

[0076] The S5 comprises the following steps:

[0077] S51, using the volume constraint of the prestack elastic parameter body to determine the fracture-cavity development of the drilling sample points in the actual drilling area, and obtain the fracture-cavity development boundary;

[0078] In this scheme, based on the prestack elastic parameter body, the area in the actual drilling area where the fractures and holes are developed and the area in the state of no fractures and holes can be distinguished, so as to obtain the fracture and hole development boundary; in this scheme, only the filling degree of the part in the state of fracture and hole development is predicted.

[0079] S52, according to the fracture-cavity development boundary, the filling state of the area in the fracture-cavity development state of the drilling sample point is matched with the value range of the prestack elastic parameter body, and the value range of the prestack elastic parameter body when the fracture-cavity is fully filled, when the fracture-cavity is half filled, and when the fracture-cavity is not filled is obtained, wherein the value range of the prestack elastic parameter body refers to the value range of the longitudinal-spherical wave velocity ratio body and the longitudinal wave impedance body;

[0080] S53, obtaining leakage characteristics in the logging data of the actual drilling area;

[0081] S54. Based on the leakage characteristics, the value ranges of the P-wave and S-wave velocity ratio bodies and the P-wave impedance bodies when the fractures and caves are fully filled, half filled, and unfilled are verified and adjusted to obtain a rock physics interpretation model.

[0082] In this embodiment, the value ranges of the P-wave and S-wave velocity ratio body and the P-wave impedance body under different filling conditions of the fractures and holes are adjusted based on the leakage characteristics, so that the adjusted value ranges can accurately match the filling state of the fractures and holes, that is, the value ranges of the adjusted P-wave and S-wave velocity ratio body and the P-wave impedance body are used as the rock physics interpretation model.

[0083] In this embodiment, taking Well W1 as an example, the rock physics interpretation model constructed based on Well W1 is: when the P-wave impedance threshold value is greater than 17200 / s*g / cm and the P-wave velocity ratio threshold value is greater than 1.91, it indicates that the fractures are fully filled; when the P-wave impedance threshold value is less than 17200m / s*g / cm and the P-wave velocity ratio threshold value is less than 1.91, it indicates that the fractures are not filled, and the rest indicate semi-filling. Due to the influence of prediction resolution, the logging scale can identify 11 fractures, of which 1-7 are unfilled, 8, 10 and 11 are fully filled, and 9 is semi-filled. The seismic scale predicts 2 fractures, of which 1 is unfilled and 2 is semi-filled.

[0084] S6. Based on the rock physics interpretation model, the fracture and cavity filling degree in the actual drilling area is predicted to obtain the fracture and cavity filling degree prediction result.

[0085] The S6 comprises the following steps:

[0086] S61. According to the rock physics interpretation model, the top and bottom of the target layer in the actual drilling area are used as time windows to extract the plane diagrams of unfilled fractures and holes, fully filled fractures and holes, and half-filled fractures and holes respectively;

[0087] S62, superimposing the plan views of the unfilled fractures and holes, the fully filled fractures and holes, and the half-filled fractures and holes to obtain the prediction result of the fracture and hole filling degree.

[0088] In this example, the filling degree prediction profile predicted by seismic prediction is consistent with the logging interpretation, in which the unfilled fracture cave 1 occurs during the drilling process at 1838.5m 3 Leakage. According to the range of the P-wave velocity ratio body and the P-wave impedance body in the rock physics interpretation model, the top and bottom of the target layer were used as the time window to extract the unfilled, fully filled and semi-filled plane maps, and then the superposition display showed that the fractures and holes were distributed along the faults, most of which were fully filled or semi-filled, with only a small number of fractures and holes unfilled. Moreover, the unfilled fractures and holes were mainly distributed in the range related to large-scale faults, and small-scale faults were mostly fully filled, which may be related to the stress direction, strength and crack opening of the fault.

[0089] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for predicting the degree of fracture-cavity filling based on prestack seismic data based on structural tensor constraints, characterized in that: The steps include: S1. Obtaining well logging data and mud logging data; S2. Perform fracture-cavity model forward modeling based on well logging data and mud logging data, and select the most sensitive gradient structure tensor attribute volume; S3, constructing the volume constraint of the prestack elastic parameter body according to the statistical coincidence rate of the well number of the most sensitive gradient structure tensor attribute body; S4, performing AVA gather forward modeling based on well logging data and mud logging data, and constructing a low-frequency model for prestack inversion to obtain a prestack elastic parameter volume; The pre-stack elastic parameter body is a longitudinal-straight wave velocity ratio body and a longitudinal wave impedance body; S5. Using the volume constraint of the prestack elastic parameter body, determine the fracture-cavity development boundary of the drilling sample point in the actual drilling area, and analyze the elastic parameter value range of the fracture-cavity full filling, half filling and unfilling in combination with the leakage characteristics to obtain the rock physical interpretation model; The S5 comprises the following steps: S51, using the volume constraint of the prestack elastic parameter body to determine the fracture-cavity development of the drilling sample points in the actual drilling area, and obtain the fracture-cavity development boundary; S52, according to the fracture-cavity development boundary, the filling state of the area in the fracture-cavity development state of the drilling sample point is matched with the value range of the prestack elastic parameter body, and the value range of the prestack elastic parameter body when the fracture-cavity is fully filled, when the fracture-cavity is half filled, and when the fracture-cavity is not filled is obtained, wherein the value range of the prestack elastic parameter body refers to the value range of the longitudinal-spherical wave velocity ratio body and the longitudinal wave impedance body; S53, obtaining leakage characteristics in the logging data of the actual drilling area; S54. Based on the leakage characteristics, the value ranges of the P-wave and S-wave velocity ratio bodies and the P-wave impedance bodies when the fractures and caves are fully filled, half filled, and unfilled are verified and adjusted to obtain a rock physics interpretation model; S6. Based on the rock physics interpretation model, the fracture and cavity filling degree in the actual drilling area is predicted to obtain the fracture and cavity filling degree prediction result.

2. The method for predicting the degree of fracture-cavity filling by prestack seismic prediction based on structural tensor constraints according to claim 1 is characterized in that: The S2 comprises the following steps: S21. Based on the well logging data and the mud logging data, a fracture-cavity model related to the fault is established and the model is forward modeled to obtain the seismic response characteristics when the fracture-cavity is developed; S22. Based on the seismic response characteristics when fractures and caves are developed, the seismic data body is subjected to dip scanning and Gaussian filtering, and the structure tensor matrix is ​​decomposed, and the sensitive gradient structure tensor attribute body with the highest statistical coincidence rate of well numbers is selected as the most sensitive gradient structure tensor attribute body.

3. The method for predicting the degree of fracture-cavity filling by prestack seismic prediction based on structural tensor constraints according to claim 2 is characterized in that: The S22 comprises the following steps: S221, performing dip scanning on the seismic data volume based on the seismic response characteristics when fractures and caves are developed and filtering the seismic data volume using a first three-dimensional Gaussian function to obtain a smoothed data volume, wherein the standard deviation of the first three-dimensional Gaussian function is a first standard deviation; The calculation expression of the smoothed data volume is as follows: ， ， in, represents a smooth data volume, represents the first three-dimensional Gaussian function, represents the seismic data volume, represents the first standard deviation, Indicates e is an exponential function of the basis constant, represents the position of the Gaussian distribution on the X-axis, Represents the center point of the three-dimensional Gaussian distribution in the X-axis direction, represents the position of the Gaussian distribution on the Y axis, represents the center point of the three-dimensional Gaussian distribution on the Y axis, represents the position of the Gaussian distribution on the Z axis, Represents the center point of the three-dimensional Gaussian distribution on the Z axis; S222, calculate the gradient vector of the smooth data body and the gradient tensor ,in, It means that the smooth data volume is differentiated with respect to the position on the X-axis. It means that the smooth data volume is differentiated with respect to the position on the Y axis. It means that the smoothed data volume is differentiated with respect to the position on the Z axis. T represents transpose; S223, filtering the gradient tensor of the smoothed data volume using a second three-dimensional Gaussian function to obtain a filtered gradient tensor of the smoothed data volume, wherein the standard deviation of the second three-dimensional Gaussian function is a second standard deviation; The calculation expression of the second three-dimensional Gaussian function is as follows: ， in, represents the second three-dimensional Gaussian function, represents the second standard deviation; S224, calculating a gradient structure tensor matrix based on the gradient vector of the smoothed data volume and the filtered gradient tensor; The calculation expression of the gradient structure tensor matrix is ​​as follows: ， in, Represents the gradient structure tensor matrix; S225. Decompose the gradient structure tensor matrix to obtain three eigenvalues ​​and corresponding eigenvectors; The calculation expressions of the three eigenvalues ​​and the corresponding eigenvectors are as follows: ， in, represents the first eigenvector, represents the second eigenvector, represents the third eigenvector, represents the first eigenvalue, represents the second eigenvalue, represents the third eigenvalue; S226. According to the three eigenvalues ​​and the corresponding eigenvectors, a number of sensitive gradient structure tensor attribute bodies describing the carbonate rock fracture-cavity body are selected; S227, respectively compare the statistical agreement rates of the number of wells describing the fracture-cavity body based on each sensitive gradient structure tensor attribute body and the actual number of wells drilled, and select the sensitive gradient structure tensor attribute body with the highest statistical agreement rate of the number of wells as the most sensitive gradient structure tensor attribute body.

4. The method for predicting the degree of fracture-cavity filling by prestack seismic prediction based on structural tensor constraints according to claim 3 is characterized in that: The calculation expression of the volume constraint of the prestack elastic parameter body in S3 is as follows: ， in, Indicates the development status of the slit-cavity body. The statistical agreement rate of the number of wells representing the most sensitive gradient structure tensor attribute volume, represents the threshold value of the statistical coincidence rate of the well number, Indicates the state of suture development. Indicates that the suture hole is not developed.

5. The method for predicting the degree of fracture-cavity filling by prestack seismic prediction based on structural tensor constraints according to claim 1, characterized in that: The S4 comprises the following steps: S41, based on the well logging data and the mud logging data, the AVA gathers are forward modeled using the real drilling data with shear wave data, and the AVA characteristics of the fractures and holes with different filling degrees are analyzed to obtain the AVA gathers with different filling degrees; S42, denoising the AVA gathers with different filling degrees, and performing gather removal and residual amplitude compensation on the denoised AVA gathers to obtain optimized AVA gathers; S43, stacking the optimized AVA gathers to obtain near, middle and far stacked data volumes; S44, repeating the well-seismic calibration and extracting the AVA wavelet of the near, middle and far stacked data volumes for a preset number of times to obtain the lateral constraints of the seismic interpretation layers; S45, interpolating the logging data along the layers using the lateral constraints of the seismic interpretation layers to construct a low-frequency model, wherein the low-frequency model includes a longitudinal-straight wave velocity ratio model and a longitudinal wave impedance model; S46. Perform prestack inversion based on the low-frequency model to obtain a prestack elastic parameter volume.

6. The method for predicting the degree of fracture-cavity filling by prestack seismic prediction based on structural tensor constraints according to claim 5 is characterized in that: The AVA features of different filling degrees in S41 include no AVA features, first-class AVA features, second-class AVA features and third-class AVA features.

7. The method for predicting the degree of fracture-cavity filling by prestack seismic prediction based on structural tensor constraints according to claim 1, characterized in that: The S6 comprises the following steps: S61. According to the rock physics interpretation model, the top and bottom of the target layer in the actual drilling area are used as time windows to extract the plane diagrams of unfilled fractures and holes, fully filled fractures and holes, and half-filled fractures and holes respectively; S62, superimposing the plan views of the unfilled fractures and holes, the fully filled fractures and holes, and the half-filled fractures and holes to obtain the prediction result of the fracture and hole filling degree.

Images