Fracture structure multi-scale modeling method based on multi-source data fusion

Abstract

The invention discloses a fracture structure multi-scale modeling method based on multi-source data fusion, and belongs to the field of geological modeling, and the method comprises the following steps: S1, collecting and preprocessing multi-source data; s13, aligning the data, and integrating the data from different sources together; s2, three-dimensional geologic modeling: S21, generating an initial three-dimensional grid by using geologic modeling software to obtain an initial three-dimensional geologic body model; s22, dynamically adjusting the density of the initial three-dimensional grid by using a self-adaptive grid division technology; s23, using a machine learning algorithm to identify and classify faults; s3, expanding the three-dimensional geologic body model macroscopically and microcosmically; and S4, performing dynamic simulation and prediction by using the three-dimensional geologic body model. By adopting the fracture structure multi-scale modeling method based on multi-source data fusion, the accuracy and reliability of the fracture structure model are improved by integrating multiple data sources and multi-scale analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of geological modeling, and in particular to a multi-scale modeling method of fracture structure based on multi-source data fusion. Background Art

[0002] Fault structure is a general term for rock fractures, including faults, joints, fissures and other structures, and sometimes fault zones appear in groups. Fault structure is one of the main structures that control the distribution of minerals. It is often a channel for the rise and migration of mineralizing solutions and a place for storing minerals, water (groundwater) and oil (petroleum). Fault structures can be divided into longitudinal faults (the two strike the same direction), transverse faults (the two strike orthogonal), and oblique faults (the two strike obliquely) according to their relationship with the overall strike of the regional structure; they can be divided into compressive faults, tensile faults and torsional (or shear) faults according to the nature of the force that produces the fault. Fault zones are large in scale, generally hundreds of kilometers to thousands of kilometers long, thousands of kilometers to tens of kilometers wide, and can reach tens of kilometers to hundreds of kilometers in depth. They are called deep faults or deep and large faults. Therefore, studying fault structures is of practical significance for prospecting and exploration, hydrogeology and engineering geology, as well as understanding regional structural characteristics.

[0003] Traditional methods often rely on a single data source, such as seismic data or drilling data, when constructing fault structure models. The limitations of a single data source lead to low accuracy and reliability of the model, which is difficult to meet the high requirements of modern geological research and engineering applications. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-scale modeling method of fault structure based on multi-source data fusion to solve the above technical problems.

[0005] To achieve the above object, the present invention provides a multi-scale modeling method of fault structure based on multi-source data fusion, comprising the following steps:

[0006] S1. Multi-source data collection and preprocessing:

[0007] S11. Collect multi-source data consisting of seismic data, drilling data, geological survey data, and remote sensing data;

[0008] S12. Clean and standardize the collected multi-source data;

[0009] S13, data alignment, and use data fusion technology to integrate data from different sources to form a unified geological model;

[0010] S2. 3D geological modeling:

[0011] S21, using geological modeling software to generate an initial three-dimensional grid to obtain an initial three-dimensional geological body model;

[0012] S22, using adaptive meshing technology to dynamically adjust the density of the initial three-dimensional grid according to the complexity of the geological structure;

[0013] S23, Apply machine learning algorithms to identify and classify faults;

[0014] S3. Expanding the 3D geological model at the macro and micro levels:

[0015] Microscopic scale: Use scanning electron microscopy and X-ray diffraction to obtain tomographic microscopic data;

[0016] Macroscale: using geomechanical models to simulate large-scale geological processes;

[0017] S4. Dynamic simulation and prediction using three-dimensional geological model:

[0018] Stress field analysis: Use the finite element method to simulate stress distribution in the crust and predict fault activity and potential earthquake risks;

[0019] Fluid dynamics simulation: Combined with fluid dynamics models, simulate the flow behavior of fluids in fault zones to evaluate resource distribution and environmental impact;

[0020] Time evolution: Use time evolution models to simulate the changes of faults over time and predict future geological events.

[0021] Preferably, step S13 specifically includes the following steps:

[0022] S131, time consistency correction:

[0023] S1311. Time alignment: align the timestamps of different data sources:

[0024]

[0025] Where, t j represents the moment of the jth time point; represents the original timestamp; t0 represents the base time, and Δt represents the time interval;

[0026] S1312, linear interpolation: perform linear interpolation in time to fill in the gaps in time:

[0027]

[0028] In the formula, x(t1) and x(t2) are the data values ​​corresponding to time points t1 and t2 respectively; x(t) represents the data value corresponding to time point t;

[0029] S132, spatial consistency correction:

[0030] S1321, spatial alignment: convert spatial coordinates of different data sources into the same coordinate system:

[0031] (x′, y′)=(a 11 x+a 12 y+a 13 ,a 21 x+a 22 y+a 23 ) (3);

[0032] In the formula, (x, y) represents the original coordinates; (x′, y′) represents the transformed coordinates, and a 11 、a 12 、a 13 、a 21 、a 22 、a 23 All represent conversion parameters;

[0033] S1322, spatial interpolation: Perform bilinear interpolation in space to fill in the gaps in space:

[0034] z(x,y)=w1z1+w2z2+w3z3+w4z4 (4);

[0035] Where z(x,y) represents the gap filling value in space; z1, z2, z3, z4 represent the four data points closest to the gap, w1, w2, w3, w4 represent the weight coefficients corresponding to z1, z2, z3, z4, respectively;

[0036] S133, fusing the data corrected for temporal consistency and spatial consistency;

[0037] S1331, weighted average method: weighted average is performed based on the reliability and accuracy of the data source:

[0038]

[0039] In the formula, z fused (x, y, t) represents the weighted average result of the data values ​​of multiple data sources at position (x, y) and time t; z i (x, y, t) represents the data value of the i-th data source at position (x, y) and time t; n represents the total number of data sources; w i Represents the weight coefficient of the i-th data source;

[0040] S1332. Considering the uncertainty and dynamic changes of data, the Kalman filter algorithm is used for data fusion:

[0041]

[0042] K k=P k|k-1 H T (HP k|k-1 H T +R) -1 (7);

[0043] In the formula, represents the fused state estimate; Indicates the predicted state; K k represents the Kalman gain; H represents the observation matrix; P k|k-1 represents the predicted covariance matrix; R represents the observed noise covariance matrix;

[0044] S134. Use K-fold cross validation method to verify the fused data.

[0045] Preferably, step S21 specifically includes the following steps:

[0046] S211. Generate the initial 3D mesh using GOCAD, Petrel or SKUA;

[0047] S212, Stratigraphic division: Use drilling logs and logging data to determine the boundaries of different strata, and use seismic profile data to identify the reflection interface of the strata;

[0048] S213, Fault Identification: Identify faults on the seismic profile and confirm the location and direction of the faults in combination with the geological map;

[0049] S214, projecting the stratum boundary onto the initial three-dimensional grid to generate a stratum model, and projecting the fault position onto the initial three-dimensional grid to generate a fault model;

[0050] S215, rock type modeling:

[0051] S2151. Determine the rock types of different strata based on geological surveys and obtain the physical and chemical properties of rock samples based on experimental analysis;

[0052] S2152. Use Kriging interpolation method to interpolate attributes, expand rock type attributes to the entire three-dimensional space, and obtain a three-dimensional geological model.

[0053] Preferably, step S22 specifically includes the following steps:

[0054] S221, calculate the residual R of each grid cell of the three-dimensional initial grid q , to determine the refinement area:

[0055] R q =f(x i )-Lu h (x i ) (8);

[0056] In the formula, f(x i ) represents the source term; L represents the partial differential operator; u h represents the control input; x i Represents state variables;

[0057] S222, dynamically adjusting the threshold according to the global error distribution;

[0058] ∈ q =α×max(E j ) (9);

[0059] In the formula, ∈ q represents the error threshold of the qth grid unit; α is a constant; max(E j ) represents the maximum error among all grid cells;

[0060] S223, when the residual R calculated in step S221 q Greater than the error threshold ∈ calculated in step S222 q When , the qth grid unit is refined by binary tree;

[0061] S224, iterate step S221-step S223 until the residual R q Not greater than the error threshold ∈ calculated in step S222 q , execute step S225;

[0062] S225. Use Laplacian smoothing method to smooth the mesh:

[0063]

[0064] In the formula, represents the value of the pth node after the k+1th iteration; N p Represents the number of adjacent nodes of node p; Represents the value of the g-th neighbor node at the k-th iteration.

[0065] Preferably, the macroscopic scale described in step S3 specifically includes the following steps:

[0066] S31. Assign physical parameters to each grid cell according to rock type;

[0067] S32. Define the constitutive relationship using the linear elastic model:

[0068] σ=E∈ (11);

[0069] In the formula, σ represents stress; ∈ represents strain; E represents elastic modulus;

[0070] S33, setting fixed boundary conditions and initial stress conditions;

[0071] u(x b ,t)=0 (12);

[0072] σ n =P (13);

[0073] In the formula, u(x b ,t) represents displacement, x b represents the boundary position; σ n represents normal stress; P represents applied pressure;

[0074] S34. According to Newton's second law, the dynamic control equation is established:

[0075]

[0076] Where ρ represents density; represents acceleration; represents stress divergence, and f represents body force;

[0077] S35. Use the finite element method to solve the dynamic control equations.

[0078] Preferably, step S4 further includes step S5, intelligent optimization and verification:

[0079] S51, Model optimization: Use genetic algorithm or particle swarm optimization algorithm to optimize model parameters;

[0080] S52. Model verification: The model is verified through actual geological data and historical earthquake data to evaluate the reliability and accuracy of the model.

[0081] Therefore, the present invention adopts the above-mentioned multi-scale modeling method of fault structure based on multi-source data fusion, which has the following beneficial effects:

[0082] 1. Improve model accuracy: Through multi-source data fusion and multi-scale analysis, the accuracy and reliability of the fault structure model are significantly improved;

[0083] 2. Enhanced practicality: This method has broad application prospects in the fields of oil and gas exploration, groundwater management, earthquake disaster assessment, etc., and can provide strong support for resource development and disaster prevention;

[0084] 3. Intelligence and automation: The introduction of machine learning and intelligent optimization algorithms improves the efficiency of model building and prediction;

[0085] 4. Multiscale modeling technology analyzes fault structures at different scales from micro to macro, further improving the accuracy and practicality of the model. That is, at the micro scale, scanning electron microscopy (SEM) and X-ray diffraction (XRD) are used to study the microcracks and microstructures inside the rock to understand their impact on the macroscopic fault behavior; at the macro scale, geomechanical models are used to simulate regional geological structures and analyze the overall behavior of the fault system.

[0086] In summary, the combination of multi-source data fusion and multi-scale modeling technology not only significantly improves the accuracy and reliability of fault structure modeling, but also provides strong support for its application in oil and gas exploration, groundwater management, earthquake disaster assessment and other fields.

[0087] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 The present invention is a flowchart of a multi-scale modeling method of fault structures based on multi-source data fusion. DETAILED DESCRIPTION

[0089] In the description of the present invention, it should be noted that the terms "upper", "lower", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the invented product is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In the description of the present invention, it should also be noted that, unless otherwise clearly specified and limited, the terms "setting", "installation", and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0090] The embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0091] like Figure 1 As shown, a multi-scale modeling method of fault structure based on multi-source data fusion includes the following steps:

[0092] S1. Multi-source data collection and preprocessing:

[0093] S11. Collect multi-source data consisting of seismic data, drilling data, geological survey data, and remote sensing data;

[0094] In this embodiment, the sources of earthquake data are: (1) Seismic stations: by setting up seismic stations, the propagation of earthquake waves is recorded. (2) Earthquake database: using public data provided by national or international earthquake data centers. (3) On-site monitoring: temporary earthquake monitoring in a specific area, using portable seismographs for data collection.

[0095] Sources of drilling data: (1) Drilling log: records the lithology, thickness, color and other information of the geological layer during the drilling process. (2) Core samples: obtain samples of the physical and chemical properties of underground rocks through coring drilling. (3) Well logging data: use various logging instruments (such as resistivity logging, sonic logging, natural gamma logging, etc.) to obtain the physical parameters of the formation. (4) Production data: records the production parameters of oil wells and gas wells, such as production and pressure.

[0096] Sources of geological survey data: (1) Field survey: Recording surface geological phenomena such as rock outcrops, faults, folds, etc. through field investigation. (2) Geological map: Draw geological maps and mark geological units, structural features and other information. (3) Geological drilling: Drilling at key locations to obtain information on underground geological structures. (4) Geochemical analysis: Collect soil and rock samples, analyze their chemical composition, and understand geochemical characteristics.

[0097] Sources of remote sensing data: (1) Satellite images: Use satellite remote sensing technology to obtain optical images, radar images, etc. of the earth's surface. (2) Aerial photography: Use aircraft, drones and other platforms to conduct aerial photography to obtain high-resolution image data. (3) Multispectral and hyperspectral data: Use multispectral and hyperspectral sensors to obtain spectral information of the earth's surface. (4) Synthetic aperture radar (SAR): Use SAR technology to obtain radar images of the earth's surface, especially in areas covered by clouds and fog.

[0098] S12. Clean and standardize the collected multi-source data;

[0099] In this embodiment, data cleaning includes missing value processing, outlier processing and duplicate value processing, and minimum-maximum normalization is used for standardization processing.

[0100] S13, data alignment, and use data fusion technology to integrate data from different sources to form a unified geological model;

[0101] Step S13 specifically includes the following steps:

[0102] S131, time consistency correction:

[0103] S1311. Time alignment: align the timestamps of different data sources:

[0104]

[0105] Where, t j represents the moment of the jth time point; represents the original timestamp; t0 represents the base time, and Δt represents the time interval;

[0106] S1312, linear interpolation: perform linear interpolation in time to fill in the gaps in time:

[0107]

[0108] In the formula, x(t1) and x(t2) are the data values ​​corresponding to time points t1 and t2 respectively; x(t) represents the data value corresponding to time point t;

[0109] S132, spatial consistency correction:

[0110] S1321, spatial alignment: convert spatial coordinates of different data sources into the same coordinate system:

[0111] (x′,y′)=(a 11 x+a 12 y+a 13 ,a 21 x+a 22 y+a 23 ) (3);

[0112] In the formula, (x, y) represents the original coordinates; (x′, y′) represents the transformed coordinates, and a 11 、a 12 、a 13 、a 21 、a 22 、a 23 All represent conversion parameters;

[0113] S1322, spatial interpolation: perform bilinear interpolation in space to fill in the gaps in space:

[0114] z(x,y)=w1z1+w2z2+w3z3+w4z4 (4);

[0115] Where z(x,y) represents the gap filling value in space; z1, z2, z3, z4 represent the four data points closest to the gap, w1, w2, w3, w4 represent the weight coefficients corresponding to z1, z2, z3, z4, respectively;

[0116] S133, fusing the data corrected for temporal consistency and spatial consistency;

[0117] S1331, weighted average method: weighted average is performed based on the reliability and accuracy of the data source:

[0118]

[0119] In the formula, z fused (x, y, t) represents the weighted average result of the data values ​​of multiple data sources at position (x, y) and time t; z i (x, y, t) represents the data value of the i-th data source at position (x, y) and time t; n represents the total number of data sources; w i Represents the weight coefficient of the i-th data source;

[0120] S1332. Considering the uncertainty and dynamic changes of data, the Kalman filter algorithm is used for data fusion:

[0121]

[0122] K k =P k|k-1 H T (HP k|k-1 H T +R) -1 (7);

[0123] In the formula, represents the fused state estimate; Indicates the predicted state; K k represents the Kalman gain; H represents the observation matrix; P k|k-1 represents the predicted covariance matrix; R represents the observed noise covariance matrix;

[0124] S134. Use K-fold cross validation method to verify the fused data.

[0125] S2. 3D geological modeling:

[0126] S21, using geological modeling software to generate an initial three-dimensional grid to obtain an initial three-dimensional geological body model;

[0127] Step S21 specifically includes the following steps:

[0128] S211. Generate the initial 3D mesh using GOCAD, Petrel or SKUA;

[0129] S212, Stratigraphic division: Use drilling logs and logging data to determine the boundaries of different strata, and use seismic profile data to identify the reflection interface of the strata;

[0130] S213, Fault Identification: Identify faults on the seismic profile and confirm the location and direction of the faults in combination with the geological map;

[0131] S214, projecting the stratum boundary onto the initial three-dimensional grid to generate a stratum model, and projecting the fault position onto the initial three-dimensional grid to generate a fault model;

[0132] S215, rock type modeling:

[0133] S2151. Determine the rock types of different strata based on geological surveys and obtain the physical and chemical properties of rock samples based on experimental analysis;

[0134] S2152. Use Kriging interpolation method to interpolate attributes, expand rock type attributes to the entire three-dimensional space, and obtain a three-dimensional geological model.

[0135] S22, using adaptive meshing technology to dynamically adjust the density of the initial three-dimensional grid according to the complexity of the geological structure;

[0136] Step S22 specifically includes the following steps:

[0137] S221, calculate the residual R of each grid cell of the three-dimensional initial grid q , to determine the refinement area:

[0138] R q =f(x i )-Lu h (x i ) (8);

[0139] In the formula, f(x i ) represents the source term; L represents the partial differential operator; u h represents the control input; x i Represents state variables;

[0140] S222, dynamically adjusting the threshold according to the global error distribution;

[0141] ∈ q =α×max(E j ) (9);

[0142] In the formula, ∈ q represents the error threshold of the qth grid unit; α is a constant; max(E j ) represents the maximum error among all grid cells;

[0143] S223, when the residual R calculated in step S221 q Greater than the error threshold ∈ calculated in step S222 q When , the qth grid unit is refined by binary tree;

[0144] S224, iterate step S221-step S223 until the residual R qNot greater than the error threshold ∈ calculated in step S222 q , execute step S225;

[0145] S225. Use Laplacian smoothing method to smooth the mesh:

[0146]

[0147] In the formula, represents the value of the pth node after the k+1th iteration; N p Represents the number of adjacent nodes of node p; Represents the value of the g-th neighbor node at the k-th iteration.

[0148] S23, Apply machine learning algorithms to identify and classify faults;

[0149] S3. Expanding the 3D geological model at the macro and micro levels:

[0150] Microscopic scale: Use scanning electron microscopy and X-ray diffraction to obtain tomographic microscopic data;

[0151] Macroscale: using geomechanical models to simulate large-scale geological processes;

[0152] The macroscopic method described in step S3 specifically includes the following steps:

[0153] S31. Assign physical parameters to each grid cell according to rock type;

[0154] S32. Define the constitutive relationship using the linear elastic model:

[0155] σ=E∈ (11);

[0156] In the formula, σ represents stress; ∈ represents strain; E represents elastic modulus;

[0157] S33, setting fixed boundary conditions and initial stress conditions;

[0158] u(x b ,t)=0 (12);

[0159] σ n =P (13);

[0160] In the formula, u(x b ,t) represents displacement, x b represents the boundary position; σ n represents normal stress; P represents applied pressure;

[0161] S34. According to Newton's second law, the dynamic control equation is established:

[0162]

[0163] Where ρ represents density; represents acceleration; represents stress divergence, and f represents body force;

[0164] S35. Use the finite element method to solve the dynamic control equations.

[0165] S4. Dynamic simulation and prediction using three-dimensional geological model:

[0166] Stress field analysis: Use the finite element method to simulate stress distribution in the crust and predict fault activity and potential earthquake risks;

[0167] Fluid dynamics simulation: Combined with fluid dynamics models, simulate the flow behavior of fluids in fault zones to evaluate resource distribution and environmental impact;

[0168] Time evolution: Use time evolution models to simulate the changes of faults over time and predict future geological events.

[0169] Step S4 also includes step S5, intelligent optimization and verification:

[0170] S51, Model optimization: Use genetic algorithm or particle swarm optimization algorithm to optimize model parameters;

[0171] S52. Model verification: The model is verified through actual geological data and historical earthquake data to evaluate the reliability and accuracy of the model.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A multi-scale modeling method for fault structures based on multi-source data fusion, characterized by: The following steps are involved: S1. Multi-source data collection and preprocessing: S11. Collect multi-source data consisting of seismic data, drilling data, geological survey data, and remote sensing data; S12. Clean and standardize the collected multi-source data; S13, data alignment, and use data fusion technology to integrate data from different sources to form a unified geological model; S2. 3D geological modeling: S21, using geological modeling software to generate an initial three-dimensional grid to obtain an initial three-dimensional geological body model; S22, using adaptive meshing technology to dynamically adjust the density of the initial three-dimensional grid according to the complexity of the geological structure; S23, Apply machine learning algorithms to identify and classify faults; S3. Expanding the 3D geological model at the macro and micro levels: Microscopic scale: Use scanning electron microscopy and X-ray diffraction to obtain tomographic microscopic data; Macroscale: using geomechanical models to simulate large-scale geological processes; S4. Dynamic simulation and prediction using three-dimensional geological model: Stress field analysis: Use the finite element method to simulate stress distribution in the crust and predict fault activity and potential earthquake risks; Fluid dynamics simulation: Combined with fluid dynamics models, simulate the flow behavior of fluids in fault zones to evaluate resource distribution and environmental impact; Time evolution: Use time evolution models to simulate the changes of faults over time and predict future geological events.

2. The multi-scale modeling method of fault structure based on multi-source data fusion according to claim 1 is characterized in that: Step S13 specifically includes the following steps: S131, time consistency correction: S1311. Time alignment: align the timestamps of different data sources: Where, t j represents the moment of the jth time point; represents the original timestamp; t0 represents the base time, and Δt represents the time interval; S1312, linear interpolation: perform linear interpolation in time to fill in the gaps in time: In the formula, x(t1) and x(t2) are the data values ​​corresponding to time points t1 and t2 respectively; x(t) represents the data value corresponding to time point t; S132, spatial consistency correction: S1321, spatial alignment: convert spatial coordinates of different data sources into the same coordinate system: (x′,y′)=(a 11 x+a 12 and+a 13 ,to 21 x+a 22 and+a 23 ) (3); In the formula, (x, y) represents the original coordinates; (x′, y′) represents the transformed coordinates, and a 11 、a 12 、a 13 、a 21 、a 22 、a 23 All represent conversion parameters; S1322, spatial interpolation: perform bilinear interpolation in space to fill in the gaps in space: z(x,y)=w1z1+w2z2+w3z3+w4z4 (4); Where z(x,y) represents the gap filling value in space; z1, z2, z3, z4 represent the four data points closest to the gap, w1, w2, w3, w4 represent the weight coefficients corresponding to z1, z2, z3, z4, respectively; S133, fusing the data corrected for temporal consistency and spatial consistency; S1331, weighted average method: weighted average is performed based on the reliability and accuracy of the data source: In the formula, z fused (x, y, t) represents the weighted average result of the data values ​​of multiple data sources at position (x, y) and time t; z i (x, y, t) represents the data value of the u-th data source at position (x, y) and time t; n represents the total number of data sources; w i Represents the weight coefficient of the i-th data source; S1332. Considering the uncertainty and dynamic changes of data, the Kalman filter algorithm is used for data fusion: K k =P k|k-1 H T (HP k|k-1 H T +R) -1 (7); In the formula, represents the fused state estimate; Indicates the predicted state; K k represents the Kalman gain; H represents the observation matrix; P k|k-1 represents the predicted covariance matrix; R represents the observed noise covariance matrix; S134. Use K-fold cross validation method to verify the fused data.

3. The multi-scale modeling method of fault structure based on multi-source data fusion according to claim 2 is characterized by: Step S21 specifically includes the following steps: S211. Generate the initial 3D mesh using GOCAD, Petrel or SKUA; S212, Stratigraphic division: Use drilling logs and logging data to determine the boundaries of different strata, and use seismic profile data to identify the reflection interface of the strata; S213, Fault Identification: Identify faults on the seismic profile and confirm the location and direction of the faults in combination with the geological map; S214, projecting the stratum boundary onto the initial three-dimensional grid to generate a stratum model, and projecting the fault position onto the initial three-dimensional grid to generate a fault model; S215, rock type modeling: S2151. Determine the rock types of different strata based on geological surveys and obtain the physical and chemical properties of rock samples based on experimental analysis; S2152. Use Kriging interpolation method to interpolate attributes, expand rock type attributes to the entire three-dimensional space, and obtain a three-dimensional geological model.

4. The multi-scale modeling method of fault structure based on multi-source data fusion according to claim 3 is characterized by: Step S22 specifically includes the following steps: S221, calculate the residual R of each grid cell of the three-dimensional initial grid q , to determine the refinement area: R q =f(x i )-Luh(x i ) (8); In the formula, f(x i ) represents the source term; L represents the partial differential operator; u h represents the control input; x i Represents state variables; S222, dynamically adjusting the threshold according to the global error distribution; ∈ q =α×max(E j ) (9); In the formula, ∈ q represents the error threshold of the qth grid unit; α is a constant; max(E j ) represents the maximum error among all grid cells; S223, when the residual R calculated in step S221 q Greater than the error threshold ∈ calculated in step S222 q When , the qth grid unit is refined by binary tree; S224, iterate step S221-step S223 until the residual R q Not greater than the error threshold ∈ calculated in step S222 q , execute step S225; S225. Use Laplacian smoothing method to smooth the mesh: In the formula, represents the value of the pth node after the k+1th iteration; N p Represents the number of adjacent nodes of node p; Represents the value of the g-th neighbor node at the k-th iteration.

5. The multi-scale modeling method of fault structure based on multi-source data fusion according to claim 4 is characterized by: The macroscopic method described in step S3 specifically includes the following steps: S31. Assign physical parameters to each grid cell according to rock type; S32. Define the constitutive relationship using the linear elastic model: σ=E∈ (11); In the formula, σ represents stress; ∈ represents strain; E represents elastic modulus; S33, setting fixed boundary conditions and initial stress conditions; u(x b ,t)=0 (12); s n =P (13); In the formula, u(x b ,t) represents displacement, x b represents the boundary position; σ n represents normal stress; P represents applied pressure; S34. According to Newton's second law, the dynamic control equation is established: Where ρ represents density; represents acceleration; represents stress divergence, and f represents body force; S35. Use the finite element method to solve the dynamic control equations.

6. The multi-scale modeling method of fault structure based on multi-source data fusion according to claim 5 is characterized by: Step S4 also includes step S5, intelligent optimization and verification: S51, Model optimization: Use genetic algorithm or particle swarm optimization algorithm to optimize model parameters; S52. Model verification: The model is verified through actual geological data and historical earthquake data to evaluate the reliability and accuracy of the model.

Images